221 Medicine Chaisson et al FIGURE 2. Confirmation of anti-MDA-5 antibodies in the sera of Cases 1 and 2 using an immunoprecipitation assay. 35 S-methionine-labeled MDA-5, generated by in vitro transcription/translation, was immunoprecipitated using serum from Cases 1 and 2 (lanes 2 and 3) or 2 different normal controls (lanes 4 and 5). Input 35S-methionine-labeled MDA-5 (not subjected to immunoprecipitation) is shown in lane 1. These data demonstrate that Cases 1 and 2 have anti-MDA-5 antibodies. ventilation failed. He was transitioned to comfort care and died shortly after ventilator support was withdrawn. Autopsy revealed bilateral diffuse alveolar damage and acute interstitial pneumonitis. There was also focal intraalveolar hemorrhage, edema, and acute inflammation in the left upper lobe. No pulmonary emboli were identified. Paratracheal lymphadenopathy was noted with the largest lymph node measuring 4.5 cm in greatest dimension, benign on histologic examination. No other significant pathologic findings were seen on autopsy. Postmortem, immunoprecipitation was performed on the patient‚Äôs serum to evaluate the presence of MDA-5 antibody, given the constellation of clinical findings and the absence of another cause of rapidly progressive lung disease. Immunoprecipitations performed in our research laboratory using in vitro transcription/translated 35S-methionine-labeled MDA-5 confirmed the presence of MDA-5 antibodies in serum from this patient (Figure 2, Case 1). Cause of death was ultimately determined to be the result of rapidly progressive ILD, likely secondary to MDA-5-associated dermato-pulmonary syndrome. Case 2 A 60-year-old healthy white woman began noticing increasing fatigue, followed 2 weeks later by acute onset of night & Volume 91, Number 4, July 2012 sweats, chills, nonproductive cough, and dyspnea. She was treated with 2 courses of antibiotics over 3 weeks, with improvement in the dyspnea and chills, but continued fatigue. Two months later, she noticed skin changes including periorbital edema, and erythema on her forehead, nose, chin, elbows, and axillae, and she developed symmetric inflammatory arthritis involving her hands, wrists, shoulders, knees, and ankles. She was started on a steroid taper by her primary care physician and referred for further evaluation. Complete blood count and electrolyte panel were unremarkable. ESR was 19 mm/h. TSH, CK, and aldolase were also unremarkable. Serologies, including ANA, ANCA, RF, antiphospholipid antibody, antihistidyl tRNA synthetase (Jo-1) antibody, cryofibrinogen, and cryoglobulins, were negative. Mammogram, Pap smear, and colonoscopy were normal. Steroids were reinitiated at this point with improvement in the arthralgia, but persistence of the rash. Two months later, the patient was admitted to the hospital with worsening dyspnea and chest tightness. Chest CT showed multifocal areas of ground-glass infiltrates. A third course of antibiotics was prescribed with improvement in her dyspnea. A prednisone taper starting at 40 mg daily was prescribed given the persistent rash, arthralgia, myalgia, and fatigue. Only the arthralgia improved with prednisone, and pulmonary symptoms worsened. Repeat CT of the chest showed improved bilateral ground-glass infiltrates compared to the initial scan, but worsening peripheral, patchy airspace opacities. She also had developed new papules of the palms and painful fissures at the lateral aspects of the interphalangeal joints of the hands. Painful ulcers developed on her wrists, forearms, arms and buttocks. She continued to have active arthritis. She was started on methotrexate. Six months after symptom onset, she had lost 60 lb, and her exercise tolerance was limited to walking a few steps in her house. She was hospitalized and transferred to our institution for progressive dyspnea. A third chest CT scan showed multilobar consolidation (Figure 3A). Pulmonary function tests demonstrated a restrictive pattern with severely impaired gas transfer (forced vital capacity EFVC^ 2.34 L, 69% predicted; forced expiratory volume in the first second EFEV1^ 1.87 L, 70% predicted; total lung capacity ETLC^ 3.79 L, 69.6% predicted; diffusing capacity of lung for carbon monoxide EDLCO^ 8.75, 41.3% predicted). On examination the patient was afebrile, normotensive, breathing 18 breaths/min, and was saturating 95% on room air. Her pulse was 98 beats/min. She had diffuse nonscarring alopecia, upper eyelid edema, livedoid rash on her FIGURE 3. A, Thoracic CT images with intravenous contrast enhancement showing scattered peripheral interstitial thickening throughout both lungs and multiple subcentimeter ground-glass pulmonary nodules throughout all lobes of both lungs. B, Cutaneous ulceration of the dorsal third metacarpophalangeal joint of the left hand. 222 www.md-journal.com * 2012 Lippincott Williams & Wilkins Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine & Volume 91, Number 4, July 2012 chin and nasolabial folds, and V-neck erythema. Erythema was present over the dorsum of the second to fifth metacarpophalangeal joints bilaterally with an ulceration on the dorsum of the third left metacarpophalangeal joint (Figure 3B). She had exquisitely painful ulcers at the tips and lateral borders of the fingers, as well as a livedoid palmar rash. Palmar papules were found overlying the second metacarpophalangeal joint as well as thenar and hypothenar eminences. She had wrist and elbow ulcers and bilateral painful ulcers on her buttocks. She had no clinical evidence of weakness. An electromyogram was not performed. White blood cell count was 4270. Electrolyte panel was unremarkable. Sedimentation rate was elevated at 45 mm/h. Complement C3, complement C4, serum protein electrophoresis and urine protein to creatinine ratio were normal. Repeat serologies, including a myositis panel (Quest Diagnostics assay), were unremarkable with exception of ANA of 1:80 in speckled pattern and a positive anti-Ro antibody (ELISA assay). Ferritin was elevated at 713 ng/mL. Skin biopsies from multiple sites (chest, ulcers, palms, buttock ulcer) were nonspecific. A skin biopsy from the chest rash showed rare dying keratinocytes, diffuse pigment-laden macrophages, and a mild, superficial perivascular and interstitial infiltrate. Wedge biopsies of the lung (lingula, left superior lower lobe, left inferior lower lobe) showed interstitial fibrosis with honeycombing. She was treated with prednisone taper starting at 60 mg daily and azathioprine 75 mg daily, escalated to 150 mg daily. After 5 months on therapy her energy level improved and she returned to activities in her home. Pulmonary function tests showed preserved lung function (FVC 2.29 L, 69% predicted; FEV1 1.9 L, 71% predicted; TLC 3.55 L, 65% predicted; DLCO 10.75, 44% predicted). The lesions on her fingers and the ulcerative lesions of her wrists and elbows persisted despite therapy, and were complicated by wound infection necessitating surgical debridement and intravenous antibiotic therapy. No new skin lesions appeared. The patient gave written informed consent for further research antibody testing given her constellation of symptoms and progressive skin and lung disease. She was found to have antiMDA-5 antibody by immunoprecipitation assay (see Figure 2, Case 2). METHODS We reviewed the MEDLINE database (National Library of Medicine, Bethesda, MD) for publications between 1966 and July 6, 2011, using the following search terms: ‚Äò‚Äòamyopathic dermatomyositis,‚Äô‚Äô ‚Äò‚Äòclinically amyopathic dermatomyositis,‚Äô‚Äô ‚Äò‚Äòdermatomyositis,‚Äô‚Äô ‚Äò‚Äòinterstitial lung disease,‚Äô‚Äô ‚Äò‚ÄòMDA-5,‚Äô‚Äô ‚Äò‚Äòmelanoma differentiation-associated protein 5,‚Äô‚Äô ‚Äò‚Äòanti-CADM-140,‚Äô‚Äô and ‚Äò‚Äòanti-CADM-140 antibody.‚Äô‚Äô We reviewed only articles written in English, and limited our search to humans. Additionally, articles were excluded if they did not contain discussion of clinical features of MDA-5 antibody, and/or only reviewed prior studies. One study was excluded for insufficient data.36 Abstract and full text reviews were performed independently by 3 authors (NFC, JP, A-MO). Data extraction was performed by the same 3 authors. Where possible, we attempted to identify publications that may have used the same patients in duplicate studies. Immunoprecipitations to confirm the presence of MDA-5 antibodies were performed as described2 using 35S-methioninelabeled MDA-5 that was generated by coupled in vitro transcription/translation from full-length cDNA encoding human MDA-5 (Origene, Rockville, MD). One KL of serum was used in each immunoprecipitation, and the immunoprecipitates were * 2012 Lippincott Williams & Wilkins MDA-5 Autoantibody and Dermato-Pulmonary Syndrome visualized by fluorography after electrophoresis on 10% SDSpolyacrylamide gels. RESULTS We identified 11 studies that met our inclusion criteria (Table 1). Most studies were case series or cross-sectional analyses. One case report was included. No clinical trials or prospective analyses were identified. Two pairs of studies likely included data from similar patient cohorts. All studies but 1 evaluated Japanese or Korean patients; the other study involved American patients. In all studies but 1, the presence of MDA-5 antibodies was identified only in patients with DM or CADM. In the remaining study, a patient with systemic sclerosis was MDA-5 antibody positive.15 Among MDA-5 antibody-positive patients, 8 studies included data regarding presence of myositis.8,11,13,15,23,27,28,32 Accounting for studies that included overlapping patient cohorts, 80% of MDA-5-positive individuals were clinically amyopathic (CK G300 and clinically insignificant muscle weakness). It is notable that MDA-5 is a cytoplasmic protein, whose expression is interferon (IFN)-regulated. To our knowledge there are no systematic analyses of whether patients with MDA-5 autoantibodies also express other autoantibodies. The first patient described here was ANA negative, and the second patient had low ANA titer and a positive anti-Ro antibody. There was a significant prevalence of ILD (60%Y100%) among MDA-5 antibody-positive individuals, although criteria (radiographic changes or restriction on pulmonary function test) for defining ILD varied by study (see Table 1). A significant portion of patients (22%Y100%) was classified as having acute or subacute ILD, characterized by evidence of progression of ILD within 1 or 3 months, respectively. This was significantly higher than MDA-negative patients (4%Y20%).8,23,32 One study evaluated differences in chest CT scan findings between MDA-5 antibody-positive and -negative patients with DM-associated ILD.32 Patients with MDA-5 antibody tended to have regions of ground-glass attenuation in random distribution or focally in the lower lobes of the lung, whereas MDA-5 antibody-negative patients had predominately reticular changes in the lower lobes. Significant skin findings in patients with MDA-5 antibodies included skin ulceration, arthritis, mechanic‚Äôs hands, palmar papules, heliotrope rash, Gottron papules, periungual erythema, and periungual telangiectasias (Table 2). Evidence from a systematic study of patients referred to a dermatology clinic7 indicates that the first 4 clinical findings are significantly more prevalent in MDA-5 antibody-positive individuals than in MDA5 antibody-negative DM patients. Six studies reported data on treatment, although most did not include specific dosing or mention treatment algorithms employed (Table 3). In all but 1 of the 6 studies,15 prednisone or prednisolone was used. Other immunosuppressants included cyclophosphamide, cyclosporine, azathioprine, IVIG, and, in 1 study, methotrexate. Only 1 study reported individual treatment data.13 In that study, all MDA-5 antibody-positive patients received prednisolone initially, but no specific preference for treatment could be identified beyond this. Mortality data were provided in 8 studies.8,10,11,13,15,18,23,32 Accounting for studies that included overlapping patient cohorts, overall mortality was 38% for MDA-5 antibody-positive patients, over follow-up periods ranging from 6 months to 14 years. Two studies reported mortality data at 6 months8,23 that were much higher in MDA-5 antibody-positive patients compared with MDA-5 antibody-negative patients (57% Erange, 46%Y75%^ vs. 9% Erange, 8%Y9%^). One study that followed patients an www.md-journal.com Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 223 224 www.md-journal.com Nested case-control Multicenter cross-sectional Case series Case series Case series Cross-sectional 27 13 23 11 15 8 Japanese Japanese Japanese Japanese Japanese Japanese Korean Japanese American Japanese Japanese DM, CTD DM DM, CTD, SSc with ILD DM DM and PM MDA-5 DM DM with ILD DM, CTD, IPF, healthy controls DM, CTD, IPF, healthy controls DM, CTD Population Studied 30‚Ć 192 65 135 376‚Ć 255‚Ć 49 2 77 25 294‚Ć 30‚Ć 13 65 82 376‚Ć 42‚Ć 38 2 77 25 67‚Ć 8‚Ć (1/7) 13 (3/10) 14 (3/11) 21 (4/17) 43‚Ć (9/34) 8‚Ć (2/6) 9 (4/5) 2 (0/2) 10 (3/7) 12 (4/8) 27‚Ƃİ (NA) 100 (NA) NA NA 100 (50) 100 (92) 100 (100) 100 (77) 100 (84) 100 (53) 95 (91) 100 (75) 45.8 T 16 47.5 T 24.7 51.5 T 8.8 53.5 T 9.4 NA 44.5 T 12.7 53 T NA 57.3 T 9.4 43.6 T 14.6 47 T 15 60.5 T 10.9 75 46 63 14 44 NA 33 50 NA 58 NA 100 92 100* 95 93 88 66 100 60 100* NA Total Total Patients No. Patients Age at % With % Patients With DM With MDA-5 Onset (yr) CADM or DM With Studied or CADM (M/F) (Mean T SD) (CADM Only) Deaths (%) ILD 100 54 71 79 NA 50 66 100 22 42 NA % With ILD and A/SIP 20 4 NA NA NA NA NA NA 5 8 NA % With ILD and A/SIP MDA-5-Negative Patients Abbreviations: A/SIP = acute or subacute interstitial pneumonitis, CTD = connective tissue disease, IPF = idiopathic pulmonary fibrosis, NA = not available; PM = polymyositis, SD = standard deviation, SSc = systemic sclerosis/scleroderma. *All patients had ILD as part of the inclusion criteria. ‚ĆIndicates possible duplication of patients between studies. ‚Ä°Number of patients with positive result by immunoprecipitation. Cross-sectional Case report Case series Cross-sectional Cross-sectional Study Type 18 10 7 32 28 Reference Population Demographic Anti-MDA-5-Positive Patients TABLE 1. Demographic Characteristics for Both Anti-MDA-5 Antibody-Positive and -Negative Patients, Previous Reports Chaisson et al Medicine & Volume 91, Number 4, July 2012 * 2012 Lippincott Williams & Wilkins Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine & Volume 91, Number 4, July 2012 MDA-5 Autoantibody and Dermato-Pulmonary Syndrome TABLE 2. Skin Features and Biopsy Results Among Patients Who Were Anti-MDA-5 Positive by Immunoprecipitation Findings Gottron Periungual Periungual Skin Palmar Compatible Heliotrope Papules Erythema Telanciectasias Ulceration Arthritis Papules Mechanic‚Äôs Fever With DM on Reference Rash (%) (%) (%) (%) (%) (%) (%) Hands (%) (%) Skin Biopsy (%) 10 7 27 13 23 15 8 100 70 50 56 15 68 38 100 70 75 86 92 68 25 NA NA NA 70 NA NA 63 NA 90 NA NA NA NA NA 50 80 NA 30 NA NA NA NA 70 50 42 69 NA NA NA 60 NA NA NA NA 63 NA 60 NA NA NA NA NA NA NA NA 74 69 NA NA 0 NA 100 NA NA NA NA Abbreviations: See previous table. average of 588 days reported 7 of 12 patients with MDA-5 antibody died compared to none in the MDA-5 antibody-negative cohort. In all 8 studies, the majority of deaths in MDA-5 antibodypositive patients were attributable to progressive ILD. DISCUSSION DM encompasses a wide spectrum of disease, almost universally involving the skin, and frequently affecting muscle and lung. The disease phentoypes are complex, with the combination of tissues involved in different individuals varying considerably.30 The disease associated with MDA-5 autoantibodies is characterized by striking ILD and a distinct cutaneous phenotype (especially ulcers and vasculopathy), but minimal to absent muscle inflammation and damage. This constellation of damage is suggestive of a common vascular target. The cutaneous and pulmonary involvement in the absence of myositis leads us to posit that this unusual dermato-pulmonary syndrome has distinct underlying causes and mechanisms. Although recent studies have suggested that muscle disease can occur in MDA-5 antibodypositive patients, it is unusual to have significant muscle weakness or elevation of muscle enzymes in this setting.29,32 Detection of commercially available autoantibodies is also rare in patients with anti-MDA-5. In the 2010 study by Gono et al,11 only 21% of patients with MDA-5 antibodies had a positive ANA. However, the presence of cytoplasmic staining by immunofluorescence was noted in both of our patients and is TABLE 3. Referenced Therapies Used in Patients With Anti-MDA-5 Antibodies* Reference 10 32 13 11 15 8 Therapy Pulsed MP, MP, CsA, Tac, IVIG, AZA MP, CsA, Cyc Pulsed MP, MP, CsA, Cyc, IVIG MP, Cyc, CI Pulsed MP, CsA, Cyc, AZA, MTX MP, immunosuppressants not otherwise defined Abbreviations: AZA = azathioprine, CI = calcineurin inhibitor not otherwise specified, CsA = cyclosporin A, Cyc = cyclophosphamide, IVIG = intravenous immunoglobulin, MP = methylprednisolone, MTX = methotrexate, Tac = tacrolimus. *Specific algorithms were not addressed in any study, and individual data were provided in only 1 study (reference 13). * 2012 Lippincott Williams & Wilkins characteristic in patients with MDA-5 antibody.27 In a patient who presents with ILD, palmar papules, painful skin ulcers and absent myopathy, absent nuclear staining and presence of cytoplasmic immunofluorescence pattern should raise suspicion for MDA-5 antibodies. Cutaneous Manifestations of MDA-5 The hallmark cutaneous manifestations of the DM spectrum include violaceous papules or plaques located on the dorsal aspect of the metacarpophalangeal or interphalangeal joints, called Gottron papules; erythematous or violaceous macules in the same distribution, called Gottron macules; and heliotrope rash, which is a periorbital violaceous erythema. These findings, which are present in over 70% of individuals with DM,20 occur with similar prevalence in both MDA-5 antibody-positive and -negative patients. MDA-5 specific skin manifestations include 1) painful, erythematous papules, macules, or both on the palmar surfaces of the metacarpophalangeal and interphalangeal joints that can have a central, ivory coloration, ulceration, and sometimes manifest as 2 distinct papules on either side of the joint; 2) cutaneous ulceration, found in 80% of these patients compared to G20% of anti-MDA-5-negative DM patients. Ulcers can be found on the elbows and knees, lateral nailfolds, and within Gottron papules. Patients with MDA-5 antibodies also have a higher prevalence of tender gums and/or oral erosions compared with MDA-5 antibody-negative patients.7,13 Our 2 patients both had palmar papules, and Case 2 had painful ulceration of the hands, wrists, elbows, and buttocks. In neither case could a clinical diagnosis of classic DM be made with certainty. Additionally, skin biopsies in both patients were nondiagnostic. MDA-5-Associated Interstitial Lung Disease The spectrum of ILD in patients with DM can range from the asymptomatic individual with incidental radiographic or pulmonary function abnormalities, to rapidly progressive and fatal disease. The reported prevalence of ILD among all patients with an idiopathic inflammatory myopathy varies widely.5,21 In all studies of MDA-5-associated disease we identified, the lung was frequently involved and was often severely affected, highlighting the lung as a central focus in this syndrome. Studies indicate that ILD progresses more rapidly in those patients with MDA-5 antibodies, with approximately 40% of MDA-5 antibody-positive patients manifesting rapidly progressive ILD. In the largest group of MDA-5 antibody-positive patients reported to our knowledge to date,13 93% had ILD, using results from chest radiography, chest CT, and pulmonary function tests. Patients were classified www.md-journal.com Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 225 Medicine Chaisson et al as having rapidly progressive ILD if they had worsening dyspnea, hypoxemia, and progression of interstitial changes on radiography within 1 month of onset of symptoms. In general, MDA-5 antibody-positive patients who have coexistent muscle disease have a lower mortality than those without muscle disease.11 Both of our patients initially described pulmonary symptoms close to the time of onset of skin manifestations, with pulmonary symptoms preceding skin disease in 1 patient, and following skin disease in the other. We found only 1 study that looked at specific radiographic findings in patients with MDA-5 antibodies.32 The findings of ground-glass attenuation in MDA-5 antibody-positive patients versus reticular patterns in MDA-5 antibody-negative patients remains very nonspecific and is unlikely to change management without further verification of underlying disease pathology by the clinician. Additionally, these data need to be validated in a large, blinded cohort before firm conclusions can be made. Pathologic data in patients with MDA-5 antibodies have not been well described to date either. In 1 of our patients, wedge biopsy revealed usual interstitial pneumonia pattern, while in the other a diffuse alveolar damage pattern was present. Findings of diffuse alveolar damage are uncommon in patients with CADM and DM, and in these cases, patients generally have poor response to therapy and dismal prognosis.22,25 MDA-5 Autoantibodies and Disease Pathogenesis MDA-5 belongs to the family of retinoic acid-inducible gene I (RIG-I)-like receptors (RLRs), a family of RNA helicases that functions as cytoplasmic sensors of pathogen-associated molecular patterns within viral RNA. To date, 3 RLR members have been identified: RIG-I, MDA-5, and LGP2 (laboratory of genetics and physiology 2). While all are broadly expressed in most tissues, levels are typically low in resting cells but are greatly increased with IFN exposure and after virus infection.17,34,35 The RLRs drive type 1 IFN production and antiviral gene expression, which mediates the intracellular immune response to control virus infection. Recent studies suggest that RLRs may also activate other inflammatory signaling pathways, including the inflammasome.26 Different RLRs appear to detect infection preferentially with different viruses. Interestingly, MDA-5 detects several RNA viruses, including picornaviruses (like polio, coxsackie, and rhinovirus), flaviviruses (Dengue and West Nile), as well as vaccinia (DNA virus). The striking vascular character of the skin phenotype in MDA-5 antibody-positive patients and the rapidly progressive lung disease suggest that blood vessels are the primary target of the immune response, potentially initially targeted during virus infection, and driven later by type I IFN pathways. Defining the mechanisms underlying the initiation and propagation of the anti-MDA-5 immune response is a high priority. Elucidating the sites of MDA-5 expression in the target tissue (and especially blood vessels at these sites) will be important in this regard. It has been noted that resistance to type 1 diabetes in humans is associated with polymorphisms that impair the function of the mda5 gene, suggesting that the relative strength of induction of type I IFNs may be important in establishing and propagating autoimmunity.3,24 Since uncontrolled viral infection on the one hand (inadequate viral sensing by MDA-5) or heightened type I IFN-induced inflammation on the other (strong MDA-5 function) both have the capacity to render an individual susceptible to tissue damage, it will be of interest to define whether mda5 polymorphisms are present in these patients, which either augment or impair MDA-5 expression and function, and thereby regulate downstream type I IFN effects. If augmented MDA-5 signaling and type I IFN activity play a role in 226 www.md-journal.com & Volume 91, Number 4, July 2012 the amplified ILD and vascular phenotype seen in this dermatopulmonary syndrome, this might have therapeutic implications for newly available agents that inhibit type I IFN signaling. Alternatively, if these pathways are decreased in MDA-5 antibodypositive patients, the possibility that increased virus-induced damage might play a role in initiating lung disease in these patients could prompt a different set of approaches. Response to Therapy Effective therapeutic modalities to treat the MDA-5-associated dermato-pulmonary syndrome have not yet been identified, and no data exist comparing the effectiveness of 1 therapy over another. This is likely because this entity has only recently been described, and it is difficult to diagnose given the current lack of a widely available MDA-5 antibody assay. While no firm guidelines can be drawn from the literature regarding treatment of MDA-5 antibody-positive individuals, many experts continue to treat patients similarly to those who have DM and are MDA-5 antibody negative, usually beginning with corticosteroid therapy. Limited reports in the literature of patients with DM and rapidly progressive ILD suggest cyclophosphamide and cyclosporine early in the disease may be beneficial.14,33 Cyclosporine has also been used to treat ulcerative lesions in a patient with CADM and intractable skin ulceration,31 but these observations were made before the recognition of the MDA-5-associated entity. The poor outcomes associated with MDA-5 antibodies strongly suggest that the optimal therapy has not yet been defined. It is possible that rapid, aggressive institution of immunosuppression might avoid irreversible lung injury if diagnosis is prompt. On the other hand, standard approaches to DM and associated conditions may not be appropriate for this condition; instead, more specific therapy targeting type I IFNs, or potentially antiviral therapy if a pathogen is defined, may be more appropriate. In the current study, 1 patient died (Case 1) despite a trial of IVIG, pulse dose prednisone, and cyclophosphamide. The other patient (Case 2) received high-dose steroids and was maintained on prednisone and azathioprine since her diagnosis, but continued to have significant disease activity. Outcomes Data from previous studies show that the presence of MDA-5 antibody is associated with increased mortality in patients with DM and CADM compared to those who are MDA-5 negative.8,23,32 Although many studies did not describe the cause of death, those that did suggested that the majority of individuals succumbed to ILD refractory to immunosuppressive therapy. The presence of MDA-5 antibodies clearly has prognostic value for those who present with this characteristic dermatopulmonary syndrome. Although the numbers remain small, approximately 40% of MDA-5 antibody-positive patients died, mostly within the first year. This was very different from MDA-5 antibody-negative patients, where 1-year mortality was 0%, and overall mortality was G10%. In 2010, Gono et al9 reported findings in individuals with CADM and ILD suggesting that increased ferritin levels correlated with faster progression of ILD. Presence of MDA-5 antibody was not measured in that study, but studies of individuals with MDA-5 antibodies support the observation that the higher ferritin levels might be associated with increased ILD progression.10,11,32 Conclusion In the current report, we describe 2 patients with a progressive dermato-pulmonary syndrome that defied diagnosis using clinically available modalities. Unusual skin features included * 2012 Lippincott Williams & Wilkins Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine & Volume 91, Number 4, July 2012 violaceous papules on the palmar surfaces of the metacarpophalangeal and interphalangeal joints, and painful ulcerative skin lesions. Inflammatory muscle disease was absent. Standard diagnostic studies, including autoantibodies, skin biopsy, and VATS lung biopsy, failed to reveal the underlying process. Antibodies to MDA-5 were present, strongly suggesting that autoimmunity to MDA-5 plays a role in generating this specific phenotype. In these settings, when extensive evaluation is otherwise nondiagnostic, clinical suspicion for MDA-5 antibodies should be high, and further testing should be aggressively pursued. REFERENCES 1. Anonymous. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med. 2000;18:1301Y1308. 2. Casciola-Rosen LA, Pluta AF, Plotz PH, Cox AE, Morris S, Wigley FM, Petri M, Gelber AC, Rosen A. The DNA mismatch repair enzyme PMS1 is a myositis-specific autoantigen. Arthritis Rheum. 2001;44:389Y396. 3. Downes K, Pekalski M, Angus KL, Hardy M, Nutland S, Smyth DJ, Walker NM, Wallace C, Todd JA. Reduced expression of IFIH1 is protective for type 1 diabetes. PLoS One. 2010;5:e12646. 4. Euwer RL, Sontheimer RD. Amyopathic dermatomyositis (dermatomyositis sine myositis). Presentation of six new cases and review of the literature. J Am Acad Dermatol. 1991;24:959Y966. 5. Fathi M, Dastmalchi M, Rasmussen E, Lundberg IE, Tornling G. Interstitial lung disease, a common manifestation of newly diagnosed polymyositis and dermatomyositis. Ann Rheum Dis. 2004;56:297Y301. 6. Fathi M, Lundberg IE, Tornling G. Pulmonary complications of polymyositis and dermatomyositis. Semin Respir Crit Care Med. 2007;28:451Y458. 7. Fiorentino D, Chung L, Zwerner J, Rosen A, Casciola-Rosen L. The mucocutaneous and systemic phenotype of dermatomyositis patients with antibodies to MDA5 (CADM-140): a retrospective study. J Am Acad Dermatol. 2011;65:25Y34. 8. Fujikawa K, Kawakami A, Kaji K, Fujimoto M, Kawashiri S, Iwamoto N, Aramaki T, Ichinose K, Tamai M, Kamachi M, Nakamura H, Ida H, Origuchi T, Ishimoto H, Mukae H, Kuwana M, Kohno S, Takehara K, Sato S, Eguchi K. Association of distinct clinical subsets with myositis-specific autoantibodies towards anti-155/140-kDa polypeptides, anti-140-kDa polypeptides, and anti-aminoacyl tRNA synthetases in Japanese patients with dermatomyositis: a single-centre, cross-sectional study. Scand J Rheumatol. 2009;38:263Y267. 9. Gono T, Kawaguchi Y, Hara M, Masuda I, Katsumata Y, Shinozaki M, Ota Y, Ozeki E, Yamanaka H. Increased ferritin predicts development and severity of acute interstitial lung disease as a complication of dermatomyositis. Rheumatology (Oxford). 2010;49:1354Y1360. 10. Gono T, Kawaguchi Y, Ozeki E, Ota Y, Satoh T, Kuwana M, Hara M, Yamanaka H. Serum ferritin correlates with activity of anti-MDA5 antibody-associated acute interstitial lung disease as a complication of dermatomyositis. Mod Rheumatol. 2011;21:223Y227. 11. Gono T, Kawaguchi Y, Satoh T, Kuwana M, Katsumata Y, Takagi K, Masuda I, Tochimoto A, Baba S, Okamoto Y, Ota Y, Yamanaka H. Clinical manifestation and prognostic factor in anti-melanoma differentiation-associated gene 5 antibody-associated interstitial lung disease as a complication of dermatomyositis. Rheumatology (Oxford). 2010;49:1713Y1719. 12. Gunawardena H, Betteridge ZE, McHugh NJ. Myositis-specific autoantibodies: their clinical and pathogenic significance in disease expression. Rheumatology (Oxford). 2009;48:607Y612. * 2012 Lippincott Williams & Wilkins MDA-5 Autoantibody and Dermato-Pulmonary Syndrome 13. Hamaguchi Y, Kuwana M, Hoshino K, Hasegawa M, Kaji K, Matsushita T, Komura K, Nakamura M, Kodera M, Suga N, Higashi A, Ogusu K, Tsutsui K, Furusaki A, Tanabe H, Sasaoka S, Muro Y, Yoshikawa M, Ishiguro N, Ayano M, Muroi E, Fujikawa K, Umeda Y, Kawase M, Mabuchi E, Asano Y, Sodemoto K, Seishima M, Yamada H, Sato S, Takehara K, Fujimoto M. Clinical correlations with dermatomyositis-specific autoantibodies in adult Japanese patients with dermatomyositis: a multicenter cross-sectional study. Arch Dermatol. 2011;147:391Y398. 14. Hirakata M, Nagai S. Interstitial lung disease in polymyositis and dermatomyositis. Curr Opin Rheumatol. 2000;12:501Y508. 15. Hoshino K, Muro Y, Sugiura K, Tomita Y, Nakashima R, Mimori T. Anti-MDA5 and anti-TIF1-gamma antibodies have clinical significance for patients with dermatomyositis. Rheumatology (Oxford). 2010;49: 1726Y1733. 16. Kameda H, Nagasawa H, Ogawa H, Sekiguchi N, Takei H, Tokuhira M, Amano K, Takeuchi T. Combination therapy with corticosteroids, cyclosporin A, and intravenous pulse cyclophosphamide for acute/ subacute interstitial pneumonia in patients with dermatomyositis. J Rheumatol. 2005;32:1719Y1726. 17. Kang DC, Gopalkrishnan RV, Lin L, Randolph A, Valerie K, Pestka S, Fisher PB. Expression analysis and genomic characterization of human melanoma differentiation associated gene-5, mda-5: a novel type I interferon-responsive apoptosis-inducing gene. Oncogene. 2004;23:1789Y1800. 18. Kang EH, Nakashima R, Mimori T, Kim J, Lee YJ, Lee EB, Song YW. Myositis autoantibodies in Korean patients with inflammatory myositis: anti-140-kDa polypeptide antibody is primarily associated with rapidly progressive interstitial lung disease independent of clinically amyopathic dermatomyositis. BMC Musculoskelet Disord. 2010;11:223. 19. Kawachi Y, Maruyama H, Furuta J, Fujisawa Y, Nakamura Y, Takahashi T, Otsuka F. Cutaneous deep necrosis with dermatomyositis: correlation with interstitial pneumonia. Eur J Dermatol. 2007;17:345Y346. 20. Kovacs SO, Kovacs SC. Dermatomyositis. J Am Acad Dermatol. 1998;39:899Y922. 21. Labirua A, Lundberg IE. Interstitial lung disease and idiopathic inflammatory myopathies: progress and pitfalls. Curr Opin Rheumatol. 2010;22:633Y638. 22. Lee CS, Chen TL, Tzen CY, Lin FJ, Peng MJ, Wu CL, Chen PJ. Idiopathic inflammatory myopathy with diffuse alveolar damage. Clin Rheumatol. 2002;21:391Y396. 23. Nakashima R, Imura Y, Kobayashi S, Yukawa N, Yoshifuji H, Nojima T, Kawabata D, Ohmura K, Usui T, Fujii T, Okawa K, Mimori T. The RIG-I-like receptor IFIH1/MDA5 is a dermatomyositis-specific autoantigen identified by the anti-CADM-140 antibody. Rheumatology (Oxford). 2010;49:433Y440. 24. Nejentsev S, Walker N, Riches D, Egholm M, Todd JA. Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes. Science. 2009;324:387Y389. 25. Parambil JG, Myers JL, Ryu JH. Diffuse alveolar damage: uncommon manifestation of pulmonary involvement in patients with connective tissue diseases. Chest. 2006;130:553Y558. 26. Poeck H, Bscheider M, Gross O, Finger K, Roth S, Rebsamen M, Hannesschlager N, Schlee M, Rothenfusser S, Barchet W, Kato H, Akira S, Inoue S, Endres S, Peschel C, Hartmann G, Hornung V, Ruland J. Recognition of RNA virus by RIG-I results in activation of CARD9 and inflammasome signaling for interleukin 1 beta production. Nat Immunol. 2010;11:63Y69. 27. Sato S, Hirakata M, Kuwana M, Suwa A, Inada S, Mimori T, Nishikawa T, Oddis CV, Ikeda Y. Autoantibodies to a 140-kd polypeptide, CADM-140, in Japanese patients with clinically amyopathic dermatomyositis. Arthritis Rheum. 2005;52:1571Y1576. www.md-journal.com Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 227 Medicine Chaisson et al 28. Sato S, Hoshino K, Satoh T, Fujita T, Kawakami Y, Fujita T, Kuwana M. RNA helicase encoded by melanoma differentiation-associated gene 5 is a major autoantigen in patients with clinically amyopathic dermatomyositis: association with rapidly progressive interstitial lung disease. Arthritis Rheum. 2009;60:2193Y2200. 29. Sato S, Kuwana M. Clinically amyopathic dermatomyositis. Curr Opin Rheumatol. 2010;22:639Y643. 30. Shamim EA, Rider LG, Pandey JP, O‚ÄôHanlon TP, Jara LJ, Samayoa EA, Burgos-Vargas R, Vazquez-Mellado J, Alcocer-Varela J, Salazar-Paramo M, Kutzbach AG, Malley JD, Targoff IN, Garcia-De la Torre I, Miller FW. Differences in idiopathic inflammatory myopathy phenotypes and genotypes between Mesoamerican Mestizos and North American Caucasians: ethnogeographic influences in the genetics and clinical expression of myositis. Arthritis Rheum. 2002;46:1885Y1893. 31. Shimojima Y, Ishii W, Kato T, Hoshi K, Matsuda M, Hashimoto T, Tanaka Y, Ikeda S. Intractable skin necrosis and interstitial pneumonia in amyopathic dermatomyositis, successfully treated with cyclosporin A. Intern Med. 2003;42:1253Y1258. 32. Tanizawa K, Handa T, Nakashima R, Kubo T, Hosono Y, Watanabe K, Aihara K, Oga T, Chin K, Nagai S, Mimori T, Mishima M. 228 www.md-journal.com & Volume 91, Number 4, July 2012 HRCT features of interstitial lung disease in dermatomyositis with anti-CADM-140 antibody. Respir Med. 2011;105:1380Y1387. 33. Yamasaki Y, Yamada H, Yamasaki M, Ohkubo M, Azuma K, Matsuoka S, Kurihara Y, Osada H, Satoh M, Ozaki S. Intravenous cyclophosphamide therapy for progressive interstitial pneumonia in patients with polymyositis/dermatomyositis. Rheumatology (Oxford). 2007;46:124Y130. 34. Yoneyama M, Kikuchi M, Matsumoto K, Imaizumi T, Miyagishi M, Taira K, Foy E, Loo YM, Gale M Jr, Akira S, Yonehara S, Kato A, Fujita T. Shared and unique functions of the DExD/H-box helicases RIG-I, MDA5, and LGP2 in antiviral innate immunity. J Immunol. 2005;175: 2851Y2858. 35. Yoneyama M, Kikuchi M, Natsukawa T, Shinobu N, Imaizumi T, Miyagishi M, Taira K, Akira S, Fujita T. The RNA helicase RIG-I has an essential function in double-stranded RNA-induced innate antiviral responses. Nat Immunol. 2004;5:730Y737. 36. Zahn S, Barchet W, Rehkamper C, Hornung T, Bieber T, Tuting T, Wenzel J. Enhanced skin expression of melanoma differentiation-associated gene 5 (MDA5) in dermatomyositis and related autoimmune diseases. J Am Acad Dermatol. 2011;64:988Y989. * 2012 Lippincott Williams & Wilkins Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. ˇ˛Abstract European Journal of Internal Medicine 18 (2007) 314 320 Original article www.elsevier.com/locate/ejim Anti- and pro-oxidant factors and endothelial dysfunction in chronic cigarette smokers with coronary heart disease E. Rocchi a,N , F. Bursi c, P. Ventura b, A. Ronzoni a, C. Gozzi a, G. Casalgrandi a, L. Marri a, R. Rossi a, M.G. Modena a a Dept. Medicine e Specialita Mediche, Chair of Terapia Medica, Universita di Modena e Reggio Emilia, Italy b Dept. Medicine e Specialita Mediche, Chair of Medicina II, Universita di Modena e Reggio Emilia, Italy c Dept. Emergenza-Urgenza, Chair of Cardiologia, Universita di Modena e Reggio Emilia, Italy Received 22 April 2006; received in revised form 18 September 2006; accepted 12 October 2006 Background: Endothelial dysfunction in cigarette smokers has been ascribed to increased oxidative damage. The aims of the present study were to compare the endothelial function of normotensive smokers with that of non-smokers and to examine its relation to some parameters representative of oxidative damage and of antioxidant capacity. Methods: We investigated 32 chronic smokers (15 30 cigarettes daily) affected by coronary heart disease, ranging from acute myocardial infarction to instable angina pectoris, and 28 matched non-smokers without any definite risk factors. All subjects underwent assessment of nitric oxide (NO)-dependent endothelial function, measured as brachial artery vasodilatation in response to reactive ischemia, using a standardized echographic method. Plasma and urinary levels of NO were also measured in all subjects, as were urinary 15-isoprostane F2t, plasma serum lipids, homocysteine (Hcy), ascorbic acid, retinol, tocopherol, and alpha- and beta-carotene (by high-performance liquid chromatography). Results: Smokers showed a significantly lower NO-mediated vasodilatation response (3.50% vs. 6.18%, pb0.001) and higher levels of urinary NO metabolites and 15-isoprostane F2t. They also had higher levels of Hcy ( p b 0.001); these values were significantly and inversely related to NO serum levels (r="0.512, pb0.001). Moreover, smokers had a significant and corresponding reduction in circulating levels of ascorbic acid, tocopherol, and alpha- and beta-carotene. Conclusions: The present study shows a clear relation between endothelial dysfunction (NO production impairment) and cigarette smoking, especially in the presence of high levels of LDL-cholesterol. It also defines some markers of both oxidative damage and antioxidant protective capacity in this condition. The monitoring of these factors may be advisable in order to assess the amount of endothelial damage. © 2007 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. Keywords: Antioxidants; Endothelial dysfunction; Cigarette smoking; Homocysteine; Isoprostane; Nitric oxide 1. Introduction Cigarette smoking represents an independent risk factor for cardiovascular disease [1]. However, the relative role of nicotine [2 5] and of other factors (hypertension, hypercho- lesterolemia) [6,7] in inducing vascular damage, and partic- N Corresponding author. University of Modena and Reggio Emilia, Dept. Medicine e Specialita Mediche, Chair of Terapia Medica, Via del Pozzo 71, Policlinico di Modena, Italy. Tel.: +39 059 4222151; fax: +39 059 4224363. E-mail address: rocchi.emilio@unimore.it (E. Rocchi). ularly endothelial dysfunction, is not well-defined. The latter, which is widely considered to be an important hallmark of atherosclerosis, is often the final result of different complex mechanisms, involving the production of reactive oxygen species (free radicals and peroxylradicals) [8] and leading to reduced nitric oxide (NO) endothelial availability. In the vasculature, endothelial NO (EDNO) plays a key role in maintaining the blood flow by modulating physio- logical vasodilatation [9 12]. Another important physiolog- ical reaction of NO is the formation of thionitrites or S- nitrosothiols (RSNOs) [13]. These adducts form in the 0953-6205/$ - see front matter © 2007 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.ejim.2006.10.006 Groups Non-smokers (n = 28) Smokers (n=32) Poultry 1.92 ± 0.15 1.66±0.11 Meat 1.64 ± 0.15 1.75±0.16 Leafy vegetables 2.31 ± 0.16 2.17±0.13 Tomatoes and carrots 2.08 ± 0.16 2.03±0.14 Nuts 0.95 ± 0.15 1.05±0.14 Milk 2.24 ± 0.23 2.00±0.20 Dairy products 2.31 ± 0.12 2.55±0.14 E. Rocchi et al. / European Journal of Internal Medicine 18 (2007) 314 320 315 Table 1 Dietary report in the groups studied, based on an empirical scoring system for the different classes of nutrients Fruit 2.54 ± 0.16 2.52±0.16 Values from 0 to 3, corresponding, respectively, to an absence of the class in the weekly diet to an nutritional board. Mean ± S.E. optimal dietary intake, were assigned according to a local presence of a thiol compound and oxygen and have, in turn, an important role in reducing vascular tone and platelet aggregation [14,15]. Among the different kinds of RSNOs, the transformation of homocysteine (Hcy), a well-known independent risk factor for atherothrombosis, into S-nitroso- homocysteine seems relevant. This substance, unlike  native Hcy, does not support hydrogen peroxide genera- tion and, hence, cell toxicity, but shows, on the contrary, well-defined and important vasodilating and inhibiting properties of platelet aggregation and adhesion of leuko- cytes to vessel walls [16,17]. In the presence of an excess of reactive oxygen species, there is a reduction in the availability of endothelial NO, due to a greater formation of peroxynitrites, which induces many different damaging effects to lipid, protein, and nucleic acid metabolism. The decrease in NO availability induces different pro-atherogenic events, such as hyper-expression of vascular cell adhesion molecules (V-CAM; I-CAM), an increase in Hcy in  native (and hence toxic) form, and a reduction in vasodilator response [17 19]. For these reasons, an assessment of pro- and antioxidant factors that are able to influence endothelial function, and particularly NO availability, seems relevant. In the present study of patients affected by coronary heart disease (CHD) and stratified by the presence/absence of smoking habit, we compared the flow-mediated dilatation (FMD) in response to reactive ischemia and the corresponding biochemical mar- kers of both oxidative damage and antioxidant status, in order to establish which of these markers may be more useful for clinical, diagnostic use and control interventions [20]. 2. Materials and methods 2.1. Patients The study was conducted in accordance with the prin- ciples of the Declaration of Helsinki. All participants gave their written informed consent before the study. Sixty subjects (all Caucasian) were recruited for the study. Thirty-two (aged 50 ± 9 years, mean ± SD, 13 females) were chronic cigarette smokers and 28 (aged 52 ± 8 years, 15 females) were non-smokers. All of the subjects had pre- viously been admitted to the cardiology unit of our institu- tion and diagnosed as having coronary heart disease (CHD), ranging from acute myocardial infarction (AMI) to instable angina pectoris (IAP). Smoking habit (15 30 cigarettes daily) was the single or largely predominant risk factor in the first group; no other definite risk factors were present in the second group. No other disease (diabetes in particular) was observed in either group and all subjects had stopped using drugs (diuretics, nitrates, ACE inhibitors, when present, in some cases) in the last 48 h prior to the study. The dietary habits of the subjects are presented in Table 1. 2.2. Assessment of brachial arterial reactivity A previously validated ultrasound study of brachial artery reactivity was performed in all patients upon entering the study and 6 months later using an Acuson 128 XP/10 mainframe (Acuson, Mountain View, California, USA) with a 7.0 MHz linear-array transducer. Images were stored on a super VHS videotape recorder for further analysis. The technique for assessing brachial artery FMD has already been described in detail elsewhere [21,22]. The methodology has an inter-observer variability in diameter measurements of 0.45 ± 0.25%, yielding a coefficient of variation of 1.34% and a coefficient of repeatability of 0.8%. The study was performed early in the morning and the smokers were ad- vised not to smoke. 2.3. Chemicals All chemicals, of HPLC grade, were purchased from Farmitalia (Milan, Italy). Ascorbic acid, retinol, D-alpha- tocopherol, alpha-carotene, beta-carotene, tocopherol ace- tate, retinol acetate (the last two as internal standards), Table 2 Main laboratory analytes in each group investigated Groups Age (years) Non-smokers 52 ± 1.5 (n=28) Smokers 50 ± 1.7 (n=32) Mean ± S.E. RBC◊106 4.57 ± 0.10 4.49 ± 0.09 Hb g/dl 13.32 ± 0.29 13.29 ± 0.34 LDL mg/dl 143.9 ± 10.6 174.7 ± 6.9 HDL mg/dl 50.1 ± 3.4 48.9 ± 3.25 TGL mg/dl 135.6 ± 23.4 134.8 ± 14.3 TSL mg/dl 519 ± 35 572 ± 23 Glucose mg/dl 110.6 ± 5.8 109.3 ± 6.1 Creatinine mg/dl 1.0 ± 0.04 0.85 ± 0.03 Albumin g/dl 4.05 ± 0.10 4.03 ± 0.07 316 E. Rocchi et al. / European Journal of Internal Medicine 18 (2007) 314 320 Table 3 Plasma ascorbic acid, tocopherol (vitamin E) to total serum lipids ratio (E/TSL), retinol, alpha- and beta-carotene in the groups investigated Groups Non-smokers (n=28) Age years EGTA, gluthatione, ortho-phenylendiamine, nitrate reduc- tase, and NADH were obtained from Sigma (Milan, Italy). Ascorbate oxydase was obtained from Roche (Mannheim, West Germany). Cadmium granules were obtained from Aldrich (Milan, Italy). Echinenone (as an internal standard for carotenoids) was a generous gift from Hoffmann-La Roche (Basel, Switzerland). 2.4. Sample preparation All subjects, after an overnight fasting, underwent blood withdrawal and urine collection. The blood samples were analyzed for liposoluble vitamins (retinol, tocopherol), alpha- and beta-carotene, ascorbic acid, nitric oxide, Hcy, and other routine analytes. 2.5. Determination of vitamins and carotenoids Ascorbic acid in plasma was analyzed following the method of Speek et al. [23]. The ascorbic acid concentra- tion was calculated from peak heights with the standard working solution as a reference. Plasma vitamins and carotenoids, as well as the ratio between vitamin E and the sum of total serum lipids, usually expressed in mg/dl (E/ TSL), were assessed following a previously published method [24]. 2.6. Determination of nitric oxide NO was determined using Green's method, which relies on the measurement of nitrites and nitrates since they are obviously the most stable metabolites [25]. 2.7. Determination of urine 15-isoprostane F2t The determination of 15-isoprostane F2t (also known as 8- iso-prostaglandin F2± or 8-epi-prostaglandin F2± ) in urine was made according to a commercially available enzyme- linked immunoassay (ELISA; MED.DIA, S. Germano Vercellese, VL, Italy). Briefly, urine samples are mixed with an enhancing reagent that essentially eliminates interferences with 15-isoprostane F2t, conjugated to horse- radish peroxydase (HRP) for binding to a polyclonal antibody specific for 15-isoprostane F2t coated on the microplates. The HRP activity results in color development when substrate is added, with the intensity of the color proportional to the amount of 15-isoprostane F2t in the samples or standards [26 30]. 52±1.5 50 ± 1.7 Ascorbic acid mmol/L 41.2±3.1 E/TSL ratio 2.51±0.25 1.92N N N ± 0.11 Retinol mmol/L 3.08±0.19 2.76 ± 0.10 Alpha-carotene nmol/L 34.0±6.8 22.3N N ± 2.9 Beta-carotene nmol/L 173±29.1 105N N N ± 11.8 Smokers (n = 32) Mean±S.E. Statistical analysis (ANOVA): N N pb0.01, N N N pb0.001. 36.5N N ± 2.6 2.8. Determination of homocysteine Briefly, bound Hcy is reduced by dithiothreitol to free Hcy, which is then enzymatically converted to S-adenosyl-L- homocysteine (SAH) by SAH-hydrolase. Sample and tracer (SAH fluorescein) compete for a mouse monoclonal anti- SAH antibody, and the intensity of polarized light is mea- sured. The concentration of Hcy in plasma is inversely re- lated to the fluorescence polarization [31]. Hcy was measured both at baseline (fasting) and after an oral methionine load (PMLHcy; 6 h after 100 mg/kg methionine). 2.9. Laboratory and statistical analysis Other biochemical indices were determined by standard laboratory methods. Results are expressed as mean±S.E.M. Both inter-group comparisons and comparisons between smokers and non-smokers were performed by one-way ANOVA and linear coefficient correlation (r), employing an IBM PC computer (IBM Corp. Armonk, New York, USA). 3. Results The results are presented in Tables 1 3 and Figs. 1 3. No significant difference in dietary habits was found between the smoking and non-smoking groups (Table 1). A complete overview of the routine analyses performed is presented in Table 2. The biochemical results of the subjects studied, who were stratified according to smoking habit, overlapped quite a bit in both groups. With regard to the parameters expressing the biological antioxidant potential, we observed that plasma ascorbic acid, E/TSL ratio, and alpha- and beta-carotene were significantly lower in the smokers (Table 3). Moreover, the FMD by Fig. 1. Percentage of brachial artery flow-mediated dilatation after upper arm occlusion in the non-smoking and smoking groups (analysis of variance between the two groups). vascular reactivity was almost halved in the smoking group (Fig. 1), and the difference between groups was highly significant ( p b 0.001). In order to confirm the hypothesis of an EDNO-dependent impaired vascular response, we measured NO metabolites in blood and urine, obtaining higher values of NO in plasma and lower values in urine of non-smokers with respect to smokers, although without a clear significant difference (Fig. 2). Moreover, the increase in urinary NO in the smoking group seemed quantitatively related to the number of cigarettes smoked daily. The isoprostane levels were also higher in the urine of smokers (pb0.001), supporting the idea of a possible effect of free oxygen radicals in inducing arachidonic acid metabolites, such as isoprostane. Finally, Hcy, which was not signifi- cantly different between the two groups at baseline, showed a highly significant increase ( p b 0.001) after the methionine load (PMLHcy) in the smoking group (Fig. 2). When plasma PMLHcy was plotted against plasma NO levels, a highly Fig. 2. A) Urinary and serum nitric oxide (NO); B) urinary isoprostane; and C) plasma homocysteine (both base value: Hcy and post-methionine load; PMLHcy) in the non-smoking and smoking groups (analysis of variance between the two groups). Fig. 3. Different linear regression indices between plasma homocysteine post-methionine load (PMLHcy) and nitric oxide (NO), total serum lipids (TSL), and vitamin E to total serum lipid ratio (E/TSL), respectively. significant negative correlation was observed (Fig. 3). Similarly, a significant direct correlation between PMLHcy and TSL and a trend towards an inverse correlation between PMLHcy and E/TSL were observed (Fig. 3). Finally, in order to evaluate the relative role of LDL-cholesterol levels with respect to smoking habit, we stratified both the smokers and non-smokers according to normal or high LDL-cholesterol serum levels, obtaining a significant difference only in plasma NO, which was reduced in proportion to the LDL increase, but not in FMD (as percentage) or other significant parameters, within both the smoking and non-smoking groups. 4. Discussion We compared two groups that were significantly homogeneous with regard to the type of illness and life E. Rocchi et al. / European Journal of Internal Medicine 18 (2007) 314 320 317 318 E. Rocchi et al. / European Journal of Internal Medicine 18 (2007) 314 320 and alimentary habits, with smoking considered as the discriminating parameter. FMD, by vascular reactivity, was significantly reduced in smokers. Similar observations were made with regard to NO, levels of which were higher in non- smokers than in smokers. On the other hand, both the decrease in NO in plasma and the increase in plasma nitrites, that was observed in the smokers, seem to suggest that plasma NO is more quickly metabolized as a result of higher oxidative damage (greater production of reactive oxygen species capable of inactivating NO) in this last group, and this seems quantitatively related to the number of cigarettes smoked daily. We chose to assess isoprostane (15-isoprostane F2t, also known as 8-iso-prostaglandin F2±) in order to better evaluate the amount of oxidative damage, as it is the final and stable product formed as a result of the activation of an alternative pathway (i.e., different from the one leading to thromboxane and prostaglandin production) of arachidonic acid metabo- lism in the presence of reactive oxygen species [26]. Urinary 15-isoprostane F2t was significantly increased (Fig. 3) in smokers. A recent study, addressing this topic indeed, con- firmed its role as an independent risk marker of CHD [32]. With regard to Hcy, which is also considered an independent risk factor of vascular damage, we did not observe any significant difference in the baseline values of this parameter between the two groups; nevertheless, PMLHcy values were significantly higher in smokers, suggesting the existence of a significant alteration in Hcy metabolism in these subjects (Fig. 3). This observation is in agreement with the data concerning a linkage between smoking and derangement of Hcy metabolism [33,34] and that not only moderately high base levels of Hcy, but also high levels of PMLHcy alone, represent an independent risk factor for vascular damage [35]. It is likely that in smokers, where NO availability is reduced, this results in a failure of NO-induced  detoxifica- tion of Hcy and of other Hcy metabolites with toxic (i.e., vasoconstrictive) effects [36,37]. This consideration is supported by the inverse and significant correlation we observed between PMLHcy and NO (Fig. 3). NO availability protects from endothelial damage induced by Hcy, as NO supplies a metabolic pathway (through the formation of S- nitroso-homocysteine) that is able to protect against Hcy- induced endothelial oxidative damage. This pathway may become exhausted rather quickly in the presence of ex- cessively high levels of Hcy or excessively low levels of NO, or both [16,38]. The parameters of oxidative damage (particularly urinary isoprostane and PMLHcy) showed significant and corresponding results. Thus, the present study confirms what is known from the literature, but what about the antioxidant capacity? In order to evaluate this issue, which has not been extensively explored in conjunction with the data presented, we assessed the most representative parameters of antioxidant activity, expressing the antioxidant power of both the intra- and extracellular lipid phase (carotenoids and liposoluble vitamins) and of the interstitial aqueous phase (vitamin C). Vitamin A levels seemed barely reduced in smokers, as expected, for it is known that plasma circulating vitamin A, which is bound to retinol-binding protein (RBP), is present in small amounts. In contrast, the vitamin A stored in liver, as palmitic ester (about 80% of the total amount of vitamin A in the body), increases [24]. Circulating vitamin A is probably more indicative of liver protein synthesis capacity (as RBP) than of vitamin A body content [39]. This is different for vitamin A precursors (alpha- and beta-carotene), as these compounds are not stored in tissues. Their plasma levels are good indicators of their real body availability. Our data show a significant decrease in vitamin A precursors in the smoking group, especially with regard to beta-carotene (Table 3). Also, vitamin C levels were reduced in smokers. It is well-known that this compound not only has important, direct antioxidant activity in the interstitial aqueous district [40] but it also exerts an important protective effect on vitamin E, probably the most important antioxidant in the lipid phase [24,33,41,42]. Vitamin E (tocopherol) levels were similar in the two groups; yet, the antioxidant effect of tocopherol is directly related to lipoproteins, by which it is carried, protecting them from oxidation [43]. For this reason, the circulating levels of vitamin E should be considered as referring to circulating lipid levels and expressed as E/TSL ratio. As shown in Table 3, the E/TSL ratio was lower in smokers than in non-smokers [44,45]. The observation of a significant correlation between PMLHcy levels and total lipids (Fig. 3) supports the results of the present study, concerning a significant linkage between different well- known cardiovascular risk factors (Hcy, LDL-cholesterol). NO consumption and, hence, endothelial dysfunction repre- sent, in our opinion, the final common mechanism of damage of all these factors, through free radical generation [36,46]. On the other hand, our findings also suggest a role for antioxidant compounds in the prevention of damage itself, particularly from the inverse trend of vitamin E (expressed as E/TSL ratio) with respect to Hcy levels (Fig. 3). Obviously, further cases could better highlight such a finding. However, the present study does, in fact, confirm the protective effects of both vitamin C and vitamin E against the oxidative damage promoted by Hcy [47,48]. As a practical conse- quence, all of these vitamin factors need to be evaluated in conjunction with the markers of oxidative damage. In summary, the present study, while confirming that the endothelial dysfunction, induced by chronic exposure to cig- arette smoking, is clearly due to oxidative damage, outlines as well the investigations that are of clinical value. This damage may, in fact, be assessed both directly by specific biochemi- cal markers (particularly urinary NO and isoprostane) and indirectly by Hcy levels after a methionine load. It is also associated with a reduction in antioxidant capacity (both hydro- and liposoluble factors). Both of these parameters should be measured in an accurate clinical evaluation in order to define the level of damage and possible strategies for its prevention. 5. Learning points " Thepreviouslyvalidatedultrasoundstudyofbrachialartery reactivity to ischemia appears to be a suitable method for studying endothelial nitric oxide (NO) availability. " Endothelial dysfunction (NO production impairment) has been confirmed in cigarette-smoking patients and in the presence of an increase in LDL-cholesterol. " Urinary isoprostane and urinary levels of homocysteine post-methionine load (PMLHcy) are proposed as useful markers of oxidative damage in clinical studies. " Both hydro- and liposoluble antioxidants are shown to be depleted in cigarette-smoking patients with coronary heart disease. Among them, the ratio between tocopherol (vitamin E) and the sum of total circulating lipids (E/TSL) may be chosen as a sensitive marker of endothelial oxidative damage in order to define preventive strategies in clinical practice. Acknowledgement This work was made possible by financial support from the Italian Ministry of University and Scientific and Technological Research (MURST), Grant no. 60 ROCC02, to the University of Modena and Reggio Emilia. ˇ˛Antibiotic resistance Geoff Scott In an environment containing vast numbers of micro-organisms saturated with or repeatedly exposed to antibiotics, Darwinian theory predicts the inevitable selection of resistant organisms. Fol- lowing the introduction of a new antibiotic, resistant strains may be reported within as little as 1 year, and after a latent interval of some years, resistance increases dramatically and the value of the antibiotic for empirical therapy is reduced. Staphylococcus aureus is now almost universally resistant to penicillin  a phenomenon first observed in hospital outbreaks of surgical sepsis in the late 1940s. This organism has shown a remarkable facility to become resistant to every antibiotic intro- duced, even vancomycin. Reduced sensitivity to vancomycin (vancomycin-intermediate S. aureus) is associated with a thick peptidoglycan cell wall. The complex gene cassette for vancomycin resistance in Enterococcus spp. (vanA) can easily be transferred to S. aureus in vitro and has now been detected in methicillin- resistant S. aureus (MRSA), making the organism fully resistant to glycopeptides. It is only a matter of time before strains are seen that are sensitive to only a small number of novel antibiotics in development and perhaps some of the older antibiotics (e.g. tetracycline, chloramphenicol). In contrast, some organisms have not yet acquired resistance despite huge pressures; these include Streptococcus pyogenes, Gram-positive anaerobes such as Clostri- dium spp. and Peptostreptococcus spp. and Neisseria meningitidis (to penicillin). Almost all anaerobes were considered to be sensitive to metro- nidazole until recent reports from Spain, France and the USA suggested that resistant Bacteroides fragilis will soon become a serious problem. Ecology Resistance is driven by antibiotic use in human and agricultural/ veterinary practice. Resistance selection occurs by spontaneous mutations occuring at a rate of 10 9 10 6 driven by the presence of antimicrobials. Resistance elements appear in saprophytic bacteria as a result of antibiotic use for growth promotion and protecting crops. When these are eaten, resistance may be transferred to human bacteria that are occasionally pathogenic. The likelihood of finding resistant organisms in the gut is related to the tonnage of antibiotic use in the country in which the individual resides, Geoff Scott is Consultant in Clinical Microbiology at University College London Hospitals, London, UK. He qualified from the Royal Free Hospital, London, and trained at Northwick Park Hospital, London and University College Hospital. His interests include control of tuberculosis and resistant infections, and changing habits. and depends on the ease with which a resistance mechanism can arise. Novel resistance is usually detected in different places in the world at about the same time, implying that antibiotic pressures are similar worldwide and that bacteria have limited means of dealing with the problem of survival. Following the appearance of one mutant resistant progeny, rapidly replicating bacteria can recolonize carrier sites in less than 24 hours. A resistant mutant of Mycobacterium tuberculosis, which replicates once every 24 hours, can recolonize diseased lung within 2 weeks. Once resistance has been selected, organisms transferred from person to person continue to be resistant. When the antibiotic pressure is removed, novel colonizing and infecting flora tend to revert to the sensitive phenotype. Some bacteria containing large resistance plasmids are considered unfit and seem to disappear from the clinical environment more easily than others. Resistance and choice of antibiotics The term  antibiotic resistance implies that a particular antibiotic is ineffective in a clinical infection. This may be because the organism is inherently resistant to the antibiotic or because it is inacces- sible. In vitro, resistance is defined by measuring the minimum inhibitory concentration (MIC) and the minimum bactericidal concentration (MBC) of the antibiotic against an organism, under ideal laboratory conditions, using appropriate controls to define the cut-off points between  resistant ,  intermediate and  sensitive . Methods of testing in the laboratory are problematic, however, and any result that does not correlate with clinical experience should be challenged. Low MIC and high MBC imply that an antibiotic is bacteristatic; that is, it inhibits the growth of an organism but is unable to kill it. Even with bactericidal antibiotics, killing in vivo normally requires an intact immune system. Thus, though bacteristatic antibiotics are ineffective in neutropenic sepsis, bactericidal antibiotics may suppress infection only for it to recrudesce when the antibiotic is removed. This observation governs strategies for the management of such infections. Paradoxically, even when an organism is declared resistant to an antibiotic by in vitro tests, it may appear to be effective in clinical use; this may be because MIC/MBC can be exceeded in vivo by giving a sufficiently large dose. This is often seen in lower urinary tract infection and in the treatment of, for example, non-meningeal penicillin-resistant pneumococcal infection with high-dose penicillin. Empirical treatment: management of an infection before bacteri- ology results are available depends on making the correct diagnosis and assessing the antibiotic sensitivities of the suspected organ- ism or organisms from local epidemiological knowledge. Once culture results are known, it is usually another day before in vitro antibiotic sensitivities are available. This is a critical period in the management of an infected patient and explains why, in severe infections, broad-spectrum antibiotics are often chosen initially when the definitive diagnosis is uncertain " and why withholding treatment may be life-threatening. This period is generally more protracted when an organism such as M. tuberculosis is slow to grow in vitro. Delay in the treatment of tuberculous meningitis may lead to permanent neurological damage. Whether the correct antibiotics were chosen initially is known only weeks later. ANTIBIOTICS MEDICINE 33:3 47 © 2005 The Medicine Publishing Company Ltd Mechanisms of resistance Resistance may be inherent (e.g. vancomycin against Gram- negative organisms, nitrofurantoin against Proteus spp.) or acquired via genetic elements encoding three fundamental mechanisms: " production of inactivating enzymes " change in the target site " exclusion of the antibiotic from the target site. The latter may occur by restriction of access through porins (only in Gram-negatives) or by active excretion. Antibiotic-inactivating enzymes may be produced in vast excess, surrounding the organism (e.g. ≤-lactamase from S. aureus), or in limited amounts in the periplasmic space of Gram-negatives. The effect in vivo is similar, but the latter organisms may appear sensi- tive in vitro. Classically, this is seen in Enterobacter spp. resistant to extended-spectrum ≤-lactam antibiotics such as piperacillin and cefotaxime; exposure of the organism to inducers such as penicillin, clavulanate, certain cephalosporins and imipenem may  switch on the production of chromosomal (AmpC) ≤-lactamases. The classical TEM (Escherichia coli) and SHV (Klebsiella spp.) ≤- lactamases (which inactivate ampicillin) have shown a remarkable ability to mutate to extended-spectrum forms (which inactivate cefoxitin and cefotetan) and to inhibitor resistance (not inhibited by clavulanate or tazobactam). CTX-M cefotaximase arose by escape from the chromosome of Kluyvera spp. and is now widespread in Enterobacteriaceae including E. coli. Other classical inactivating enzymes include chloramphenicol acetyltransferase and aminoglycoside-modifying enzymes. Target site change may be a structural alteration preventing binding of an antibiotic, or a mechanism whereby the metabolic pathway that is normally inhibited is bypassed by an alternative one. This is seen in sulphonamide resistance and in MRSA, which has acquired a novel penicillin-binding protein from Staphylococcus scuiri. Important target site changes include topoisomerases II and IV (quinolones), ≤ subunit of DNA-dependent RNA polymerase (rifamycins) and methylation of 23S target (14/15-membered macrolides). Exclusion of antibiotic: porins are protein structures embedded in the outer bilipid membrane of Gram-negative organisms. They control what passes into and out of the cell on the basis of molec- ular size and charge. Mutations in the genetic elements encoding porins (permeability mutations) may exclude one antibiotic or, more commonly, multiple antibiotics. Pseudomonas aeruginosa has two outer lipid membranes, produces ≤-lactamases constitutively, and therefore tends to be more resistant than coliforms. Alternatively, organisms may actively excrete antibiotics, notably tetracyclines, macrolides and quinolones. Efflux pumps may be up-regulated. S. aureus resistant to erythromycin by active excretion remains sensitive to clindamycin, but the more common target site mutation affects susceptibility to both antibiotics. Acquisition of resistance Direct mutation of chromosomal genes may lead to resistance, and non-fatal mutations are passed to all progeny. Alternatively, small, mobile, circular pieces of DNA termed  plasmids may be passed from one bacterium to another (even to bacteria of different species) by various mechanisms such as direct transfer by type II pili (Figure 1) and phage transfer. Plasmids reproduce each time the organism divides and can probably be lost as easily as gained, given the correct environment. Transposons are large genetic elements often containing mul- tiple genes necessary to confer phenotypic resistance. They may encode pheromones, which attract bacteria to each other and can be transferred on plasmids, generally being incorporated into the chromosomal DNA of the new host. Free DNA from dead bacteria and even mammalian cells can be incorporated into bacterial chromosomes. A remarkable mechanism of resistance is seen in penicillin-resistant Streptococcus pneumoniae, which have incor- porated a mosaic of genes into the chromosome from other resis- tant ±-haemolytic streptococci. Problematic resistant bacteria Gram-positive organisms MRSA has replaced methicillin-sensitive S. aureus as the more common cause of hospital-associated sepsis in most parts of the world. The targets of the ≤-lactam antibiotics are penicillin-binding proteins, carboxypeptidases and transpeptidases, which catalyse bridging of pentapeptide subunits of peptidoglycan. All strains of MRSA already produce penicillinase, but they now have a new penicillin-binding protein (PBP2') that is not inhibited by methi- cillin and its congeners oxacillin and flucloxacillin. They are resist- ant to all ≤-lactam antibiotics. The mecA gene encodes PBP2', but requires expression of several ancillary genes within a transposon for full expression. Thus, some strains with mecA may appear to be sensitive to methicillin in vitro. In addition, a plasmid encodes resistance to a variable number of other antibiotics. Strains with a tendency to spread easily and to predominate in hospitals are termed  epidemic MRSA (EMRSA). The current epidemic of EMRSA15 or EMRSA16 in the UK started in 1994. In the laboratory, these are detected by their antibiogram, but other typing methods (e.g. pulsed-field gel electrophoresis of chromo- somal DNA) are needed to show that two strains are indistinguish- able, thus implying that cross-infection has occurred. When a few patients in one or two wards acquire a novel strain, such an outbreak can easily be tracked and then controlled. When EMRSA strains become endemic (Figure 2), more general measures are needed to reduce the risk to patients admitted to the hospital. MRSA may become endemic in nursing homes, creating a pool for novel introduction into hospital; transfer of patients from ward to ward and colonization of staff leads to continued exposure of patients to new strains and helps maintain endemicity. Infection with MRSA is not untreatable. Strains are often (though not predictably) susceptible to gentamicin or other aminoglycosides, rifampicin, co-trimoxazole, chloramphenicol or ciprofloxacin, and sometimes to tetracyclines, macrolides, fusidic acid and pseudomonic acid. (Note that rifampicin and fusidic acid should never be used alone because resistant mutants are selected very rapidly.) MRSA is almost always susceptible to glycopeptides, though strains with reduced sensitivity to vancomycin have been occasionally described in patients with chronic colonization or infection who have been treated for several weeks.1 Some rare strains of S. aureus are dependent on vancomycin to allow them to grow. MEDICINE 33:3 48 © 2005 The Medicine Publishing Company Ltd ANTIBIOTICS Penicillin-resistant S. pneumoniae strains have shown a gradual phase-shift increase in MIC to penicillin over several years. MIC is about 0.001 mg/litre in sensitive strains and 1 mg/litre in resistant strains. This is achieved by changes in penicillin-binding proteins with lower affinities for ≤-lactams. In the UK, such strains currently represent about 4% of those causing invasive infection, but in some areas (e.g. Spain) the rate is as high as 50%. These strains are usually resistant to many other useful antibiotics, including cephalosporins, chloramphenicol and erythromycin. Treatment with high-dose penicillin is effective in pneumonia but not in meningitis caused by resistant strains, some of which are sensitive to second-generation and third-generation cephalo- sporins. The mortality from severe pneumococcal pneumonia remains relatively constant regardless of whether the strain is resistant to penicillin, but meningitis caused by a resistant strain is more likely to be fatal. Splenectomized patients are advised to take life-long low-dose oral penicillin prophylaxis against the rare possibility of over- whelming pneumococcal sepsis. In the future, such prophylaxis may become less effective, but it is difficult to identify an alterna- tive simple and safe regimen. Glycopeptide-resistant enterococci: enterococci have become important causes of nosocomial postoperative infection in the USA. They are likely to cause sepsis and pneumonia only in severely ill patients in the UK, and occasionally strains are resistant to vanco- mycin with (vanA) or without (vanB) teicoplanin. These genes require the action of complex accessory genes for full expression. vanA encodes a structural change in the terminal amino acid of the pentapeptide chain of peptidoglycan, from D-ala to D-lac. This substitution prevents binding of vancomycin, enabling construction of the dipeptide bridges in peptidoglycan to continue. Typing indicates wide heterogeneity of strains, and that carriers usually harbour more than one strain. This implies that the trans- poson encoding vanA and the accessory genes is promiscuous and easily able to enter the host s enterococci. These organisms are of low virulence. Infections may respond to simple antibiotics such as amoxicillin if they appear sensitive (Enterococcus faecalis), but are generally resistant to all but a few new antibiotics. Colonization with enterococci is driven by cephalosporins and fluoroquinolones, to which enterococci are constitutively resistant. Gram-negative organisms In contrast to resistant Gram-positives, against which old anti- biotics or antibiotics in development are occasionally active, some Gram-negative organisms (particularly non-glucose fermenters such as Pseudomonas, Stenotrophomonas and Acinetobacter spp.), especially in ICUs, are resistant to every useful antibiotic and there are no new agents in development. Glucose fermenters: in terms of resistance and endemic/epidemic problems in ill, hospitalized patients, Klebsiella, Enterobacter and Serratia spp. are more troublesome than E. coli. They produce SHV-type ≤-lactamases constitutively, and mutations lead to resist- ance to all ≤-lactams other than the carbapenems (or occasionally aztreonam). They often lose susceptibility to quinolones and ANTIBIOTICS ab cd 1 DNA can be transferred a on direct contact, b,cviatypeIIpilior d by phages (bacterial viruses). (b and c by courtesy of Zeneca.) MEDICINE 33:3 49 © 2005 The Medicine Publishing Company Ltd aminoglycosides and become essentially untreatable. Some strains of Salmonella enterica (particularly typhimurium), and the enteric salmonellae (e.g. S. typhi) have acquired stable broad resistance to all useful antibiotics. These are now pandemic. Non-glucose fermenters: P. aeruginosa has been replaced as a troublesome cause of nosocomial infection by other environ- mental and skin bacteria such as Acinetobacter baumanii var. calcoaceticus. Some strains are resistant to all available antibiotics, others are sensitive only to carbapenems and some aminoglyco- sides (e.g. amikacin). In response to long-term antibiotic use over many years, Burkholderia cepacia tends to replace S. aureus and P. aeruginosa as colonizing flora in patients with cystic fibrosis. It may also show resistance to many antibiotics and can cause troublesome cross-infection. Neisseria spp.: N. gonorrhoeae is interesting in that different strains may exhibit the three mechanisms of resistance to penicil- lin (reduced permeability, changes in penicillin-binding proteins and ≤-lactamase production). Each confers a stepwise increase in MIC, and there is no clear cut-off between sensitive and resistant strains. Many strains have become resistant to other oral drugs such as ciprofloxacin, though these are currently rare in the UK. N. meningitidis has been slow to acquire resistance to penicillin, though a few strains isolated recently in Spain have slightly higher MICs than predicted. The value of sulphonamides was largely lost by the 1970s, and though chloramphenicol remains useful, second- generation cephalosporins (cefotaxime or ceftriaxone) seem to give the best results in clinical treatment. M. tuberculosis: resistance to first-line antituberculosis agents has been a major problem in the re-emergence of tuberculosis. In some countries (e.g. the states of the former USSR), multi-drug- resistant strains (resistant at least to rifampicin and isoniazid) are extremely common, and more so in patients who have been treated previously. In the UK, lone resistance to isoniazid occurs in 5% (except in London, where there is currently an epidemic of one clone resistant to isoniazid that started in 1995 and has led to a resistance rate of 15% in affected areas), and to rifampicin, ethambutol and pyrazinamide in 1% or fewer. Multi-drug-resistant strains account for 1.2%. Strains resistant to all are found through- out the country, but are more commonly seen in London. Most patients can be treated satisfactorily with a standard 6-month regimen, but the results of routine antibiotic sensitivity tests are unknown for 5"16 weeks after sending specimens to the laboratory, and patients with resistant strains need more toxic second-line drugs and prolonged courses of treatment. Rapid liquid culture and detection of genes encoding resistance may improve the speed of laboratory diagnosis. Strategies for reducing the impact of resistance Antibiotic resistance is driven by antibiotic use. When antibiotics are superseded and therefore used less, strains resistant to these tend to disappear. In the community: in the UK, more than 80% of human use of antibiotics occurs in the community, mostly for respiratory tract infections. The Standing Medical Advisory Committee, in its report The path of least resistance, made recommendations to reduce inappropriate prescribing.2 " No antibiotics should be given for simple coughs and colds. " Antibiotics should not be routinely prescribed for sore throats, unless there is evidence of streptococcal infection. " Antibiotics are not routinely required for acute otitis media and sinusitis-like symptoms; if given, courses can be limited to 3 days. In addition, 3 days treatment should suffice in otherwise healthy women with uncomplicated cystitis. This strategy has been useful in reducing prescribing by GPs, but anxiety has resulted from anecdotal reports of an increase in bacterial respiratory infections. In hospital, antibiotic use can be reduced by several means. New patients with methicillin-resistant Staphylococcus aureus at University College London Hospitals, 1991 1999 450 400 350 300 250 200 150 100 50 0 Infections Carriers ANTIBIOTICS 1991 1992 1993 1994 1995 1996 1997 1998 1999 Year MEDICINE 33:3 50 © 2005 The Medicine Publishing Company Ltd 2 Annual incidence ANTIBIOTICS Sites of action of antibiotics and some resistance mechanisms Glycopeptide Glycopeptide too large Aminoglycosides ≤-lactams Mutation in porin Porin Aminoglycosides Macrolides Chloramphenicol RNA Gram- Gram-positive negative Peptidoglycan Supercoiled DNA Topoisomerases Quinolones Folate synthesis Sulphonamides Trimethoprim Substrate change Mutation ≤-lactams Periplasmic space (≤-lactamase activity) DNA ≤-lactamases Chloramphenicol acteyltransferase Aminoglycoside-modifying enzymes Inactivating enzymes Active excretion Macrolides Quinolones Tetracyclines " Routine use of antibiotics for surgical prophylaxis should be reduced to a minimum. " Antibiotics should not be started immediately in all suspected infections. Certain patients (e.g. those with febrile neutropenia or evidence of septicaemia) require urgent antibiotic therapy, but in many other cases (e.g. mild pyrexia postoperatively), it is safe and ultimately preferable to withhold antibiotics until culture results are known or there is clear evidence of bacterial infection. " An alternative is to discontinue antibiotics as soon as infor- mation is available suggesting that the problem is not bacterial or has resolved itself. This is termed an  antibiotic-stop policy, and the aim is to encourage doctors to actively review the need for antibiotics after, say, the second day, given information on cultures and surrogate markers that has by then become available. " Certain antibiotics can be withheld from the hospital formulary. This is the main benefit of an agreed antibiotic policy. The choice of restricted antibiotics may be decided on the basis of cost rather than likely selection of resistance. (These antibiotics must be used occasionally, however, when resistance to all other available drugs has been selected.) " In theory, antibiotics can be rotated such that, for example, predominantly penicillins are used at some times, and cephalo- sporins or quinolones at others. There is little clear scientific evidence that this has any effect, and major, complicated studies would be needed to determine the effect of change in use on both normal and infecting flora. However, it has been shown that, in a setting of heavy cephalosporin use, discontinuation of use of this class of drugs leads to a reduction in the risk of colonization with glycopeptide-resistant enterococci and Clostridium difficile- associated diarrhoea. Other strategies: development of resistance in human pathogens might be delayed if antibiotics were not used so widely in animal husbandry, particularly for growth promotion. The future Resistance to antibiotics is one of the greatest threats to the success of modern medicine. It has recently become more serious because we can no longer be sure that any antibiotic chosen empirically will work, and because of the emergence of totally resistant bac- teria. How to reduce resistance without simply discontinuing use of all antibiotics is a dilemma. We do not know to what degree antibiotic use must be reduced to decrease the selective pressure and allow reversion to sensitivity, nor do we know how to protect the few remaining drugs that can be used to treat resistant infec- tions. Doctors and vets should overcome the view that they have an inalienable right to prescribe empirical antibiotics, and should set targets for reduction of their personal prescribing. ˇ˛Antibiotics for acute otitis media: a meta-analysis with individual patient data Maroeska M Rovers, Paul Glasziou, Cees L Appelman, Peter Burke, David P McCormick, Roger A Damoiseaux, Isabelle Gaboury, Paul Little, Arno W Hoes Summary Background Individual trials to test effectiveness of antibiotics in children with acute otitis media have been too small for valid subgroup analyses. We aimed to identify subgroups of children who would and would not benefit more than others from treatment with antibiotics. Methods We did a meta-analysis of data from six randomised trials of the effects of antibiotics in children with acute otitis media. Individual patient data from 1643 children aged from 6 months to 12 years were validated and re-analysed. We defined the primary outcome as an extended course of acute otitis media, consisting of pain, fever, or both at 3 7 days. Findings Significant effect modifications were noted for otorrhoea, and for age and bilateral acute otitis media. In children younger than 2 years of age with bilateral acute otitis media, 55% of controls and 30% on antibiotics still had pain, fever, or both at 3 7 days, with a rate difference between these groups of "25% (95% CI "36% to "14%), resulting in a number-needed-to-treat (NNT) of four children. We identified no significant differences for age alone. In children with otorrhoea the rate difference and NNT, respectively, were "36% ("53% to "19%) and three, whereas in children without otorrhoea the equivalent values were "14% ("23% to "5%) and eight. Interpretation Antibiotics seem to be most beneficial in children younger than 2 years of age with bilateral acute otitis media, and in children with both acute otitis media and otorrhoea. For most other children with mild disease an observational policy seems justified. Lancet 2006; 368: 1429 35 See Comment page 1397 Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht, the Netherlands (M M Rovers PhD, C L Appelman MD, R A Damoiseaux MD, Prof A W Hoes MD); Departments of Paediatrics and Otolarynglogy, Wilhelmina Children s Hospital, University Medical Centre Utrecht, the Netherlands (M M Rovers); University of Oxford, Department of Primary Health Care, Institute of Health Sciences, Oxford, UK (Prof P Glasziou MD, P Burke FRCGP); Department of Pediatrics, University of Texas Medical Branch Galveston, Texas, USA (Prof D P McCormick MD); Chalmers Research Group, Children s Hospital of Eastern Ontario Research Institute, Ottawa, Ontario, Canada (I Gaboury); Primary Medical Care, Community Clinical Sciences Division, University of Southampton, Aldermoor Health Centre, Southampton, UK (Prof P Little FRCGP) Correspondence to: Dr Maroeska M Rovers Julius Centre for Health Sciences and Primary Care, Stratenum 7∑109, PO Box 85060, 3508 AB Utrecht, the Netherlands M.Rovers@umcutrecht.nl Introduction Acute otitis media is one of the most common childhood infections, the leading cause of doctors consultations, and the most frequent reason for children to take antibiotics.1 Evidence from systematic reviews, however, suggests that antibiotics provide only marginal benefit.2,3 Furthermore, prescribing antibiotics is known to encourage clinic visits for subsequent episodes, intensify pressure on clinicians to prescribe, increase antibiotic use, and promote antibiotic resistance.4 6 Guidelines therefore recommend selective use of anti- biotics for acute otitis media, especially in children aged 2 years or older. In children younger than 2 years, no consensus has been reached. Some guidelines recommend antibiotics for all these children,7,8 whereas others advise antibiotics only for children under 2 years if they are severely affected or have persistent signs of disease or related comorbidity.9,10 Reliable identification of subgroups of children who do, and do not, benefit from treatment with antibiotics has not been straightforward, because individual trials have been too small for valid and reliable subgroup analyses. A meta-analysis of the individual data from original trials enables the opportunity to identify subgroups that are most likely to benefit. We therefore aimed to identify subgroups that might benefit most from such treatment. Methods Selection of trials We did a systematic search of the Cochrane library, PubMed database, EMBASE, and the proceedings of the international symposia on recent advances in otitis media. We selected trials that (1) used random allocation of children, (2) included children aged 0 12 years with acute otitis media, (3) compared antibiotics with placebo or no treatment, and (4) had pain and fever as an outcome. All trials were assessed for four major quality criteria: proper randomisation methods; degree of follow-up; and blinding of the outcome assessor, patient, and care giver. All trials obtained informed consent and ethics approval. The primary investigators of all selected trials were asked for the raw data of their trials. The data thus obtained were thoroughly checked for consistency, plausibility, integrity of randomisation, and follow-up. A few issues were queried with the responsible trial investigator or statistician, and all were resolved. Outcome variables The primary outcome was an extended course of acute otitis media, which was defined as pain, fever, or both at 3 7 days. We used this composite endpoint since both factors are relevant from clinical and patients (or parental) perspectives. Fever was defined as temperature of 38∞C or higher, and pain was assessed by parents and recorded in diary form (as either yes or no). Both outcome measures were dichotomised, since several trials measured them in this way. Fever and pain were also studied separately (as secondary outcomes). Additionally, the adverse effects of antibiotic treatment mentioned in every trial were analysed. Independent predictors of an extended course of disease had been established in an earlier study within www.thelancet.com Vol368 October21,2006 1429 Articles Articles the same setting (unpublished data).11 We used these independent baseline predictors ie, age (<2 vs e"2 years), fever (yes vs no), and bilateral acute otitis media (yes vs no) to investigate whether those at risk of an extended course had enhanced benefits from treatment with antibiotics. We also examined the effects of concurrent otorrhoea at baseline (yes vs no), both alone and in combination with the identified predictors, since this condition seems to be a clinically relevant outcome that occurs too infrequently to be identified as an independent predictor. Statistical analyses Information was available for 72% of the potential subgroups (range 28 100%) and for 90% of the outcome variables (range 81 98%). To reduce bias and to increase statistical efficiency, we imputed the missing data for all trials using the linear regression method (multivariate analyses) available in SPSS (version 12.0).11 Regression was based on the correlation between individual variables with missing values and all other variables, as estimated from the complete set of data. We imputed missing values only within trials. To decide whether pooling of data for analysis was justified, we assessed heterogeneity between studies using I2, which describes the percentage of variation between studies due to heterogeneity rather than chance.12 The range for I2 lies between between 0% (ie, no observed heterogeneity) and 100%. The resulting I2 was lower than 25% (p>0∑30) indicating that studies were sufficiently similar to justify pooling of data. We calculated relative risks (RR), rate differences (RD), and NNT, with their 95% CI, for both the primary and secondary outcomes. To assess whether the effect of antibiotics was modified by age, bilateral acute otitis media, fever, otorrhoea, or a combination of these factors, we did a fixed-effect logistic regression analysis. In this model, the independent variables were: treatment with antibiotics (yes vs no); the potential-effect modifiers (age, bilateral acute otitis media, fever, otorrhoea, or com- binations of these); and an interaction term (defined as use of antibiotics times potential-effect modifier). We also used a binary dummy variable to identify each study within the regression analysis. Dependent variables were an extended course (primary outcome), fever, and pain at 3 7 days (secondary outcomes). We calculated the c-index (area under the receiver operating curve) to measure the accuracy of each model. If a significant interaction effect was identified, we did stratified analyses of the rate ratios and rate differences within each stratum of the subgroups. The percentages of children with an extended course during each consecutive day within each of the identified subgroups were calculated for the five trials that asked parents to fill out diaries noting signs of the disease. Finally, we did sensitivity analyses, including only those trials that measured the outcomes on the same day, used the same dose regimen, or included placebo. All analyses were performed according to the intention-to-treat principle. Role of the funding source This study was sponsored by the Dutch College of General Practitioners and the Netherlands Organisation for Health Research and Development (grant number 4200.0010). This sponsor had no role in study design, Ref 22 Ref 24 Ref 26 Number of patients 121 240 512 Participants Children aged 6 months to 12 years visiting a GP with recurrent AOM Children aged 6 months to 2 years visiting a GP with AOM Children aged 6 months to 5 years presenting to clinics or the emergency department with AOM Interventions Amoxicillin with clavulanate vs placebo Amoxicillin vs placebo Amoxicillin vs placebo Duration of intervention 7 days 10 days 10 days Outcomes Fever after 3 days Pain after 3 days Otorrhoea Otoscopy and tympanometry after 1 month Symptoms at day 4 assessed by a GP (including fever and earpain) Otoscopy and tympanometry after 6 weeks and 3 months Telephone follow-up at day 1, 2, 3, and between 10 and 14 days (including fever) Tympanometry at 1 and 3 months Ref 23 232 Children aged 3 to 10 years with AOM Amoxicillin vs placebo 7 days Symptoms noted by parents (including fever and ear pain) Home visits by researcher after 24 h and 5 7 days Otoscopy and tympanometry after 1 and 3 months Ref 25 315 Children aged 6 months to 10 years visiting a GP with AOM Immediate antibiotics (amoxicillin) vs delayed treatment 7 days Symptoms noted by parents (including fever and earpain) Absence from school Consumption of paracetamol Ref 27 223 Children aged 6 months to 12 years with AOM Immediate antibiotics (amoxicillin) vs delayed treatment 10 days Symptoms noted by parents (including fever and earpain) Analgesic consumption Nasopharyngeal carriage Adverse events Absence from school Tympanometry after 12 and 30 days AOM=acute otitis media; GP=general practitioner. Table 1: Characteristics of the six trials included in our meta-analysis 1430 www.thelancet.com Vol368 October21,2006 Articles Age <2 years Recurrent AOM Winter season Passive smoking! Coughing! Ear pain Bilateral AOM** Perforation! ! Bulging tympanic membrane 287 (35%) 429 (52%) 620 (75%) 218 (33%) 476 (72%) 724 (88%) 220 (33%) 19 (7%) 342 (42%) 567 (35%) 831 (51%) 1243 (76%) 432 (34%) 936 (72%) 1447 (88%) 456 (34%) 39 (7%) 685 (42%) Antibiotics (n=819) 280 (34%) 402 (49%) 623 (76%) 214 (34%) 460 (72%) 723 (88%) 236 (35%) 20 (8%) 343 (42%) Controls Total (n=824) (n=1643) Male sex 411 (50%) 411 (50%) 822 (50%) Siblings* 455 (76%) 472 (78%) 927 (77%) Being breastfed 244 (64%) 255 (64%) 499 (64%) Cryingß 407 (83%) 413 (83%) 820 (83%) Runny nose∂ 428 (77%) 429 (78%) 857 (78%) Fever|| 282 (40%) 287 (41%) 569 (40%) Otorrhoea 51 (19%) 65 (23%) 116 (21%) Red tympanic membrane 751 (92%) 754 (92%) 1505 (92%) Data are number (%). AOM=acute otitis media. Overall number=*1207, 778, 1299, ß984, ∂1105, ||1411, **1328, ! ! 555. Percentages do not always add to 100% because of missing data for some characteristics. Table 2: Baseline characteristics of patients in the six trials data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. Results Our search strategy identified nineteen trials that investigated the effectiveness of antibiotics in children with acute otitis media. After screening, nine trials were excluded, because randomisation was inadequate, the control group received another treatment, information about our selected outcomes was not available, or because they focused on special study populations, such as children with ventilation tubes.13 21 Of the ten eligible trials, six research groups provided us with their data22 27 and four did not.28 31 The methodological quality of the six remaining studies was generally high. Five used adequate concealed allocations (blinded randomisations) and outcome assessments. Loss to follow-up was less than 10%. Table 1 shows the main characteristics of the six trials. The mean age of the children was 3∑4 years (range 0 11 ); half were boys; about half had recurrent acute otitis media; and about a third had bilateral acute otitis media (table 2). Our meta-analysis showed that, relative to placebo, overall RR for an extended course of acute otitis media at 3 7 days with antibiotics was 0∑83 (95% CI 0∑78 0∑89). The rate difference between the control group and the antibiotics group was 13% (9 17), resulting in a NNT of eight children. Overall RR of fever at 3 7 days was 0∑95 (0∑92 0∑98); the rate difference was 5% (2 8) and NNT was 20 children. The corresponding figures for children who had pain at 3 7 days were 0∑86 (0∑81 0∑91); 11% (7 15); and ten children, respectively. Our analyses showed that the effect of antibiotics was modified by age and bilateral disease, and by otorrhoea, notably for the primary outcome of pain, fever, or both at 3 7 days (table 3). In children aged less than 2 years with bilateral acute otitis media, more than half the control group and less than a third of the antibiotics group still had pain, fever, or both at 3 7 days, with a rate difference of about 25%. In children aged 2 years or older with bilateral disease the rate difference was about 12%. For age alone no differences were identified. The c-indices, calculated to gauge the accuracy of each model, were 0∑63, 0∑58 and 0∑61, respectively, for age and bilaterality, age alone, and bilaterality alone. About 60% of children with otorrhoea in the control group had pain, fever, or both at 3 7 days, whereas only about 25% of those given antibiotics had protracted illnesses. The rate difference, of about 36%; was much greater than that for those without otorrhoea, which was about 14%. Other factors, in combination with otorrhoea, such as age, bilateral disease, or both did not substantially alter this pattern ie, children with otorrhoea seemed to benefit most from treatment with antibiotics, irrespective of other characteristics. With pain alone as the primary outcome, the effect of antibiotics was modified by age and bilateral disease together (p-value for interaction 0∑01) (table 3). For children aged less than 2 years with bilateral acute otitis media, twice as many controls still had pain at 3 7 days, compared with those given antibiotics. For age alone no differences were identified. Figure 1 shows the proportion of children with an extended course of disease in the subgroups for which antibiotics were of most benefit ie, children younger than 2 years of age with bilateral disease, and those with otorrhoea. For both these subgroups, symptoms resolved faster in children who received antibiotics than in children randomised to the control group, but this difference disappeared after 4 5 days. Sensitivity analyses, including only those trials that measured the outcome at the same time during follow-up, used the same dose of antibiotics, or included a placebo, were in agreement with the overall results. The most commonly described adverse effect of antibiotic treatment was diarrhoea, which ranged from 2% to 14% in controls and from 4% to 21% in those given antibiotics in each of the six trials that we analysed. Occurrence of rash ranged from 2% to 6% in the control groups, and from 1% to 8% in the antibiotic groups. One child from the control group developed meningitis at day 3,24 but seemed to have received antibiotics at day 2 because of deterioration. No mastoiditis or other serious complications were mentioned in these six trials. www.thelancet.com Vol368 October21,2006 1431 Articles Number (%) Group given antibiotics Control RD (95% CI) group (n=824) NNT RR (95% CI) p value for interaction* (n=819) Pain, fever, or both at 3 7 days Age <2 years 567 (35%) 91 (33%) 137 (48%) "15% ("23 to "7) 7 0∑77 (0∑68 0∑89) e"2 years 1076 (65%) 107 (20%) 166 (31%) "11% ("16 to "6) 10 0∑86 (0∑80 0∑93) 0∑83 Bilateral AOM No 872 (66%) 104 (24%) 132 (30%) "6% ("12 to 0) 17 0∑92 (0∑85 1∑00) Yes 456 (34%) 64 (27%) 104 (47%) "20% ("28 to "11) 5 0∑72 (0∑62 0∑84) 0∑021 Age and bilateral AOM <2 years+bilateral AOM 273 (20%) 42 (30%) 74 (55%) "25% ("36 to "14) 4 0∑64 (0∑62 0.80) <2 years+unilateral AOM 261 (20%) 45 (35%) 53 (40%) "5% ("17 to 7) 20 0∑92 (0∑76 1∑11) e"2 years+bilateral AOM 183 (14%) 20 (23%) 30 (35%) "12% ("25 to 1) 9 0∑84 (0∑70 1∑02) e"2 years+unilateral AOM 611 (46%) 59 (19% 79 (26%) "7% ("14 to 0) 15 0∑92 (0∑85 1∑01) 0∑022 Otorrhea Yes 116 (21%) 12 (24%) 39 (60%) "36% ("53 to "19%) 3 0∑52 (0∑37 0∑73) 0∑039 No 439 (89%) 61 (28%) 94 (42%) "14% ("23 to "5%) 8 0∑80 (0∑70 0∑92) Pain at 3 7 days Age, years < 2 years 567 (35%) 77 (28%) 115 (40%) "12% ("20 to -4%) 9 0.83 (0∑73 0∑93) e" 2 years 1076 (65%) 86 (16%) 142 (26%) "10% ("15 to -5%) 10 0.88 (0∑82 0∑93) 0∑76 Bilateral AOM No 872 (66%) 85 (20%) 102 (23%) "3% ("8 to -2%) 34 0.96 (0∑89 1∑03) Yes 456 (34%) 48 (20%) 88 (40%) "20% ("28 to -12%) 5 0.75 (0∑66 0∑85) 0∑005 Age and bilateral AOM < 2 years+bilateral AOM 273 (20%) 32 (23%) 62 (46%) "23% ("34 to -12%) 5 0.70 (0∑58 0∑84) < 2 years+unilateral AOM 261 (20%) 41 (31%) 42 (33%) "2% ("13 to 9%) 50 0.99 (0∑84 1∑17) e" 2years+bilateralAOM 183 (14%) 16 (17%) 26 (30%) "13% ("25 to 1%) 8 0.83 (0∑71 0∑99) e" 2years+unilateralAOM 611 (46%) 44 (15%) 59 (19%) "4% ("10 to 2%) 25 0.95 (0∑88 1∑02) 0∑009 AOM=acute otitis media. RD=rate difference. RR=rate ratio. NNT=number needed to treat. *p value for the interaction term (antibiotics x subgrouping variable) in the fixed effect regression analysis. Table 3: Subgroup analyses with both the rate differences and rate ratios 1432 www.thelancet.com Vol368 October21,2006 Discussion Our meta-analyses of individual patient data showed that antibiotics are more beneficial in children aged less than 2 years with bilateral acute otitis media, and in those with both acute otitis media and otorrhoea ie, in these groups three to four children have to be treated to prevent an extended course of the disease in one child. Although none of the trials included in this meta-analysis have had adequate power to produce precise effect estimates in clinically relevant subgroups, both McCormick27 and Appelman22 and their colleagues had suggested that children younger than 2 years might benefit most from antibiotics for otitis media. The results of our fixed-effect logistic regression analysis, however, showed that the effects of antibiotic treatment were not significantly modified by either age or bilateral disease alone. Additionally, the NNT was lower for the combined model than for individual components, indicating that targeting of both age and bilaterality would increase the benefits of antibiotic therapy. Moreover, the subgroups studied were based on a multivariate prognostic model, which showed that age and bilaterality were both independent predictors of an extended course of disease (unpublished data). Although we need to understand the causal mech- anism of the subgroups effects before final conclusions can be drawn, we can postulate that, in children aged less than 2 years with bilateral acute otitis media and in those with otorrhoea, the infection is more often bacterial than viral. Indeed, Palmu and co- workers32 have shown that culture-positive cases of acute otitis media are more often bilateral than are culture-negative events; middle ear effusion samples obtained through tympanic membranes with known pre-existing perforations were more likely to be culture- positive than were samples obtained through an intact membrane. Furthermore, perforations are more often caused by an infection with Streptococcus pneumoniae than with Haemophilus influenzae or Moraxella catar- rhalis.32 S pneumoniae is most common in young children.32 Figure 1: Proportion of children with an extended course of acute otitis media A: d"2 years with bilateral disease; B: e"2 years with unilateral disease; C: with otorrhoea; and D: without otorrhoea. The main strength of our study was that, by re- analysing the data of six trials, we were able to include 1643 children, which gave us the power to identify subgroups that could benefit most from treatment with antibiotics. Nevertheless, some of our findings deserve further discussion. First, only six of the ten eligible randomised, controlled trials could be included in our meta-analysis. The main characteristics of the four trials for which individual patient data were not available were, however, much the same as those in the six included trials. Moreover, the overall results of our subset of six trials are very similar to the overall results reported by the Cochrane review3 that did include all trials. A funnel plot of the included studies (data not shown) also indicated that publication bias was unlikely. Second, we could not do a pooled analysis with respect to failure rates since these rates were defined and measured differently in each of the six included trials. We did, however, undertake subgroup analyses of failure rate within each trial, and subsequently pooled these results for the six trials. The results were in accord with the pooled results for the subgroups ie, the largest effect of antibiotics was in children aged younger than 2 years with bilateral acute otitis media (rate difference "8%, 95% CI "17% to 0%), and the smallest effect was in children aged 2 years or older with unilateral acute otitis media (rate difference "3%, "7% to 1%). Third, the severity of the pain was estimated by parents and not further quantified in the trials, which could have resulted in an incorrect estimation of the real pain. Analysis with fever alone, however, showed much the same trend. Moreover, the fact that in many children the complaints at days 3 7 were mild should be taken into account in interpretation of reported NNTs and in the decision to initiate antibiotic therapy in individual patients. Fourth, the results are based on child participants, who might not be representative of those visiting general practitioners. For example, the most severely affected children might be under-represented. However, because we had access to raw data from six trials, we had high numbers of children from specific high-risk groups, which are often under-represented in single trials. Furthermore, the children we included seem represen- tative of those with acute otitis media visiting general practitioners, since the percentages of those aged less than 2 years and 2 years or older were much the same as those from a national survey in Netherlands of children with acute otitis media in primary care (ie, 35% vs 33%, and 65% vs 67%, respectively).33 Fifth, the rate of mastoiditis was so low that we could not obtain a precise estimate for risk of this complication. The trials done so far, however, showed that initially withholding antibiotics from children with acute otitis media does not increase suppurative complications. Whether restrictive antibiotic use increases acute mastoiditis at the population level remains unresolved, but the potential increase is only two cases per 100 000 person-years and should be weighed against potential adverse effects.1 Sixth, since not all trials used the most objective diagnostic methods (eg, pneumatic otoscopy or tympanometry) some children in our meta-analysis might not have had ear infections. Sensitivity analyses www.thelancet.com Vol368 October21,2006 1433 Articles 100 90 80 70 60 50 40 30 20 10 0 100 90 80 70 60 50 40 30 20 10 0 AC Control group Antibiotic group BD 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 Days of follow-up Proportion of children (%) 1434 www.thelancet.com Vol368 October21,2006 Articles with the three trials that did use these diagnostic methods were, however, in accord with the overall results. Seventh, we did not study all possible subgroups. We selected established predictors of an extended course of disease (unpublished data) and some clinically relevant variables, and did stratified analyses only for those variables that showed a significant p value for the interaction in the fixed regression model. We might therefore have missed a subgroup. Our approach is, however, in agreement with recommendations for study of subgroups.34 The strength of this approach is that our prognostic analyses revealed only a few relevant sub- groups, limiting the number of subgroup analyses and subsequent false-positive findings (type I error) that could be caused by multiple testing. Furthermore, other subgroups that might benefit more from treatment with antibiotics (eg, children with Down syndrome or cleft palate) could not be studied in this meta-analysis of individual patient data, because these subgroups were excluded in the individual trials. The experience of many clinicians that these subgroups of children benefit more from treatment with antibiotics has not yet been evidenced in randomised controlled trials. Eighth, we did not adjust for potential confounding due to differences between trials. We did, however, examine whether such confounding had occurred in our study, and noted that children aged less than 2 years were most likely to have fever and an abnormal tympanic membrane at baseline. We therefore used the Mantel Haenszel technique to adjust for these potential confounders in our subgroup analyses. Since the effect estimates were not altered by adjustments, crude effect estimates are presented. We conclude that antibiotics are beneficial in relieving residual pain or fever at 3 7 days in children younger than 2 years of age with bilateral acute otitis media, and in children with acute otitis media and otorrhoea. For most other children with mild disease an observational policy seems justified. Contributors M M Rovers designed and planned the study, and gathered, analysed, and interpreted the data. P Glasziou and A W Hoes contributed to the initial idea and design of the study, interpreted the data, and super- vised the study. C L Appelman, P Burke, D McCormick, R A Damoiseaux, I Gaboury, and P Little provided the data of the original trials, contributed to the protocol, and interpreted the data. The manuscript was prepared by M M Rovers, and all authors have seen and approved the final version. Conflict of interest statement We declare that we have no conflict of interest. Acknowledgments We thank Anne Schilder for commenting on an earlier version of this paper. CLINICAL RESEARCH STUDY Anticoagulation-associated Adverse Drug Events Gregory Piazza, MD,a Thanh Nha Nguyen, PharmD,b Deborah Cios, PharmD,c Matthew Labreche, PharmD,b Benjamin Hohlfelder,b John Fanikos, RPh, MBA,c Karen Fiumara, PharmD,d Samuel Z. Goldhaber, MDa a Cardiovascular Division, Department of Medicine, Brigham and Women‚Äôs Hospital, Harvard Medical School, Boston, Mass; bVenous Thromboembolism Research Group, cDepartment of Pharmacy, and dCenter for Clinical Excellence, Brigham and Women‚Äôs Hospital, Boston, Mass. ABSTRACT PURPOSE: Anticoagulant drugs are among the most common medications that cause adverse drug events (ADEs) in hospitalized patients. We performed a 5-year retrospective study at Brigham and Women‚Äôs Hospital to determine clinical characteristics, types, root causes, and outcomes of anticoagulant-associated ADEs. METHODS: We reviewed all inpatient anticoagulant-associated ADEs, including adverse drug reactions (ADRs) and medication errors, reported at Brigham and Women‚Äôs Hospital through the Safety Reporting System from May 2004 to May 2009. We also collected data about the cost associated with hospitalizations in which ADRs occurred. RESULTS: Of 463 anticoagulant-associated ADEs, 226 were medication errors (48.8%), 141 were ADRs (30.5%), and 96 (20.7%) involved both a medication error and ADR. Seventy percent of anticoagulantassociated ADEs were potentially preventable. Transcription errors (48%) were the most frequent root cause of anticoagulant-associated medication errors, while medication errors (40%) were a common root cause of anticoagulant-associated ADRs. Death within 30 days of anticoagulant-associated ADEs occurred in 11% of patients. After an anticoagulant-associated ADR, most hospitalization expenditures were attributable to nursing costs (mean $33,189 per ADR), followed by pharmacy costs (mean $7451 per ADR). CONCLUSION: Most anticoagulant-associated ADEs among inpatients result from medication errors and are, therefore, potentially preventable. We observed an elevated 30-day mortality rate among patients who suffered an anticoagulant-associated ADE and high hospitalization costs following ADRs. Further quality improvement efforts to reduce anticoagulant-associated medication errors are warranted to improve patient safety and decrease health care expenditures. ¬© 2011 Elsevier Inc. All rights reserved. ‚Ä¢ The American Journal of Medicine (2011) 124, 1136-1142 KEYWORDS: Adverse drug events; Adverse drug reactions; Anticoagulation; Medication errors Adverse drug events (ADEs), which comprise medication errors and adverse drug reactions (ADRs), represent a major source of harm among hospitalized patients and have been a driving force behind implementation of electronic health Funding: Dr Piazza is supported by a Research Career Development Award (K12 HL083786) from the National Heart, Lung, and Blood Institute (NHLBI). This study was funded, in part, by a clinical research grant from Johnson & Johnson (New Brunswick, NJ). Conflict of Interest: None. Authorship: All authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to Gregory Piazza, MD, Cardiovascular Division, Brigham and Women‚Äôs Hospital, 75 Francis St., Boston, MA 02115. E-mail address: gpiazza@partners.org 0002-9343/$ -see front matter ¬© 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2011.06.009 records, including computerized provider order entry.1-6 One study reported that 6.5 medication-related adverse events occurred per 100 hospitalizations and estimated that more than one quarter of these events were the result of a medication error.7 Medication errors are potentially preventable causes of ADRs that can occur at all stages of the medication process, including prescription, transcription, dispensing, and administration.8-12 Compared with nonpreventable events, potentially preventable ADEs have been shown to double the incremental additional length of stay and health care costs attributable to the event.13 Anticoagulant drugs, including warfarin, unfractionated heparin, and low-molecular-weight heparin, are among the most commonly implicated medications that cause ADEs in Piazza et al Anticoagulation-associated Adverse Drug Events 1137 was noxious and unintended.17 A medication error was hospitalized patients.1,4,6,11,14 Despite implementation of computerized provider order entry, electronic medication defined as an event that caused or led to inappropriate administration records, and improved infusion pump techmedication use or patient harm.17 A medication error that nology (‚Äúsmart pumps‚Äù), medication errors involving antiwas discovered and corrected before reaching the patient coagulant medications remain common.11,12 Elderly15,16 was classified as a near miss. All isolated medication errors and cardiac patients14 represent and ADRs that resulted from medpopulations at particularly high ication errors were considered porisk for suffering anticoagulanttentially preventable. CLINICAL SIGNIFICANCE associated ADRs. We searched the Safety ReTo determine the clinical charporting System using Risk Mon‚óè Most anticoagulant-associated adverse acteristics, types, severity, root itor Pro (rL Solutions, Inc., drug events (70%) are potentially causes, and outcomes of anticoagCambridge, Mass). We evalupreventable. ulant-associated ADEs, we perated the patient characteristics, ‚óè Transcription errors comprise the most formed a 5-year retrospective ADE type and severity, root frequent root cause of anticoagulantstudy of the Safety Reporting Syscause, and outcomes of all antitem at Brigham and Women‚Äôs coagulation-associated medicaassociated medication errors, and medHospital. We reviewed discrete tion errors and ADRs. All patient ication errors are a common root cause ADEs that originated during hosrecords were reviewed for eviof anticoagulant-associated adverse dence of treatment with an antipitalization at Brigham and Womdrug reactions. coagulant medication. Hemoren‚Äôs Hospital. We also conducted ‚óè After anticoagulant-associated adverse rhagic events were classified an analysis of the cost associated drug reactions, most hospitalization exwith hospitalizations in which according to the Global Use of penditures are attributable to nursing ADRs occurred. Strategies to Open Occluded Coronary Arteries (GUSTO) criand pharmacy costs. teria for severe or life-threatenMETHODS ‚óè Efforts to minimize anticoagulant-assoing, moderate, or mild bleedciated medication errors are warranted ing.18 We obtained 30-day Patient Population to improve patient safety and decrease follow-up for 100% of patients Brigham and Women‚Äôs Hospital is hospitalization costs. included in the registry. a 793-bed acute tertiary care facilWe included all anticoagulant ity providing medical and surgical medications administered for procare for patients with general phylaxis and treatment of thrommedical, cardiothoracic, orthopedic, oncologic, neurologic, boembolic events. ADEs associated with the use of the obstetric and gynecologic, neonatal, urologic, and gastroinfollowing anticoagulants were reviewed: unfractionated testinal conditions. Brigham and Women‚Äôs Hospital utilizes heparin, low-molecular-weight heparin, fondaparinux, wara Medical Informatics System that integrates an online medfarin, and direct thrombin inhibitors, such as argatroban, ical record, computerized provider order entry, an electronic bivalirudin, and lepirudin. A multidisciplinary team, includmedication administration record, and an electronic safety ing physicians, pharmacists, and a hospital patient safety reporting system. The Safety Reporting System is a volunofficer, reviewed all reported anticoagulant-related events tary computerized reporting system for ADRs and medicaand evaluated the root case of all ADEs. tion errors, which was instituted in May 2004. We also collected data about the cost associated with hospitalizations in which ADEs occurred. Because medData Collection ication errors that do not culminate in an ADR do not add We reviewed all inpatient anticoagulant-associated ADEs, significant incremental cost to hospitalization, we inincluding ADRs and medication errors, reported at Brigham cluded only ADRs in our cost analysis. ADRs that added and Women‚Äôs Hospital through the Safety Reporting Syssignificant incremental cost to hospitalization were distem from May 2004 to May 2009. Inpatient events for the tinguished from those that did not by consensus of a purpose of our study were required to have originated durphysician (GP), pharmacist (TNN), and patient safety ing hospitalization at Brigham and Women‚Äôs Hospital. We officer (KF). While patients may have suffered multiple excluded reported anticoagulant-associated ADEs that took ADRs during a hospital admission, we analyzed costs place in the Emergency Department, Operating Room, Carincurred after the initial event. We utilized a software diac Catheterization Laboratory, Cardiovascular Recovery program for cost accounting that is capable of tabulating Room, Post-Anesthesia Care Unit, and Neonatal Intensive the cost associated with a patient admission (Transitions Care Unit because medication administration records were System Incorporated, Waltham, Mass). We sorted costs not consistently computerized in these units during the regby the following categories: nursing, pharmacy, blood istry period. We defined an ADR as a response to a drug products, clinical laboratory, radiology, surgery, nonsurnormally used for prophylaxis or therapy of disease and that gical intervention, and anesthesia. We compared total and 1138 The American Journal of Medicine, Vol 124, No 12, December 2011 component costs of hospitalization for admissions in which an ADR occurred and added incremental expense with those for admissions in which an ADR occurred but was not determined to incur incremental cost. Statistical Methods Descriptive statistics including baseline characteristics; variables related to ADEs and those pertaining to outcomes were stratified as continuous or binary. Continuous variables were presented as medians with interquartile ranges. Binary variables were presented as numbers and proportions. Cost variables were presented as means with standard deviations and were compared using a 2-sample t-test. All reported P-values were 2-sided, and a P-value of œΩ.05 was considered statistically significant. All statistical analyses were performed using STATA version 9.2 (STATA Corp., College Station, Tex). Table 2 Characteristics of Adverse Drug Events (ADEs) in Patients Receiving Anticoagulation (n œ≠ 463)* Characteristic of ADE Adverse drug reaction (ADR), n (%) Medication error, n (%) Combined medication error and ADR, n (%) Near miss, n (%) Potentially preventable ADE, n (%) Anticoagulant associated with ADE Unfractionated heparin, n (%) Warfarin, n (%) Low-molecular weight heparin, n (%) Argatroban, n (%) Bivalirudin, n (%) Lepirudin, n (%) Fondaparinux, n (%) 141 226 96 33 322 (30.5) (48.8) (20.7) (7.1) (69.5) 270 96 44 29 18 3 3 (58.3) (20.7) (9.5) (6.3) (3.9) (0.7) (0.7) ADR œ≠ adverse drug reaction. *Patients could have had more than 1 ADE including both medication errors and ADRs. RESULTS Baseline Characteristics Patients who suffered anticoagulant-associated ADEs had a median age of 62 years (Table 1). Their median body mass Table 1 Baseline Characteristics and Medical Conditions in Patients with Anticoagulation-associated Adverse Drug Events (ADEs) (n œ≠ 463)* Baseline characteristic Median age on admission, years (interquartile range) Age œæ75 years, n (%) Male, n (%) Median body mass index, kg/m2 (interquartile range) Median length of stay, d (interquartile range) Medical condition Hypertension, n (%) Surgery in past 2 months, n (%) Ischemic heart disease, n (%) Atrial fibrillation, n (%) Serum creatinine œæ1.5 mg/dL, n (%) Deep vein thrombosis or pulmonary embolism, n (%) Heart failure, n (%) Diabetes, n (%) Active cancer without metastases, n (%) Status post heart valve surgery, n (%) Chronic obstructive pulmonary disease, n (%) History of stroke, n (%) Active cancer with metastases, n (%) Other thromboembolism, n (%) History of cancer, n (%) Dialysis, n (%) Thrombophilia, n (%) *Patients could have had more than 1 ADE. 62 (49-72) 92 (19.9) 246 (53.1) 27.4 (23.8-33.1) 13 (7-23) 243 213 140 126 125 121 (52.5) (46) (30.2) (27.2) (27) (26.1) 106 104 74 68 66 (22.9) (22.5) (16) (14.7) (14.3) 45 42 40 34 30 16 (9.7) (9.1) (8.6) (7.3) (6.5) (3.5) index was 27.4 kg/m2. Patients who suffered anticoagulantassociated ADEs had a median length of stay of 13 days. Patients with anticoagulant-associated ADEs had a high frequency of medical conditions, such as atrial fibrillation, history of deep vein thrombosis or pulmonary embolism, heart failure, ischemic heart disease, chronic kidney disease, and stroke, that increase the risk of thromboembolism. Characteristics of Anticoagulant-associated Adverse Drug Events In 250,725 admissions over the 5-year study period, there were 463 anticoagulant-associated ADEs reported. Of these anticoagulant-associated ADEs, 226 were medication errors (48.8%), 141 were ADRs (30.5%), and 96 (20.7%) involved both a medication error and ADR. Seventy percent of anticoagulant-associated ADEs were potentially preventable (Table 2). Unfractionated heparin (58%) and warfarin (21%) were the most commonly implicated drugs in anticoagulant-associated ADEs. An average of 1.4 anticoagulant-associated ADEs occurred per patient over the 5-year study period. Characteristics of Anticoagulant-associated Medication Errors The most frequent type of anticoagulant-associated medication errors were missed medication doses (25%) and wrong rate or frequency (23%) (Table 3). Transcription errors (48%) were the most frequent root cause of anticoagulantassociated medication errors, followed by memory lapses (16%). Characteristics of Anticoagulant-associated Adverse Drug Reactions The most frequent anticoagulant-associated ADRs were abnormal coagulation studies (72%), any bleeding (25%), and Piazza et al Anticoagulation-associated Adverse Drug Events Table 3 Types and Root Causes of Anticoagulant-associated Medication Errors (n œ≠ 323) Type of medication error* Missed dose, n (%) Wrong rate or frequency, n (%) Medication not discontinued when ordered, n (%) Extra dose, n (%) Wrong dose, n (%) Wrong time of administration, n (%) Failure to act on laboratory result, n (%) Known allergy or contraindication, n (%) Wrong route, n (%) Medication expired, n (%) Wrong patient, n (%) Wrong drug, n (%) Preparation error, n (%) Wrong technique of administration, n (%) Medication administered without an order, n (%) Inadequate monitoring, n (%) Root cause of medication error Transcription error, n (%) Memory lapse, n (%) Infusion or parenteral administration problem, n (%) Rule violation, n (%) Lack of knowledge about drug, n (%) Faulty drug identity checking, n (%) Drug preparation error, n (%) Drug stocking or delivery problem, n (%) Faulty interaction between services, n (%) Lack of information about the patient, n (%) Faulty dose checking, n (%) Known allergy or contraindication, n (%) 79 75 31 26 23 17 17 13 10 8 7 7 6 3 2 1 (24.5) (23.2) (9.6) (8.1) (7.1) (5.3) (5.3) (4) (3.1) (2.5) (2.2) (2.2) (1.9) (0.9) (0.6) (0.3) 155 (48) 52 (16.1) 28 (8.7) 19 18 17 12 9 7 4 1 1 (5.9) (5.6) (5.3) (3.7) (2.8) (2.2) (1.2) (0.3) (0.3) *Patients may have had more than 1 medication error type. thrombocytopenia (18%) (Table 4). Seventy-two percent of all ADRs were associated with at least one occurrence of excessive anticoagulation, such as a super-therapeutic international normalized ratio or activated partial thromboplastin time. Seventeen percent of anticoagulant-associated ADRs culminated in transfusion of at least one unit of packed red blood cells. Undetected predisposing conditions (58%), such as a previously unknown drug allergy, and medication errors (40%) comprised the most common root causes of anticoagulant-associated ADRs. Outcomes of Anticoagulant-associated Adverse Drug Events Death within 30 days of anticoagulant-associated ADEs occurred in 11% of patients. In-hospital death occurred after 5.6% of anticoagulant-associated ADEs, while death after discharge but within 30 days of the ADE was observed in 5% (Table 5). A rehabilitation center stay was frequently required (40%) after an admission in which an anticoagulant-associated ADE occurred. Patients who suffered anticoagulant-associated ADEs had a high rate of re-hospitalization within 30 days of the ADE (17.5%). 1139 Cost Analysis of Anticoagulant-associated Adverse Drug Reactions Most hospitalization expenditures after an anticoagulantassociated ADR were attributable to nursing costs (mean $33,189 per ADR) followed by pharmacy costs (mean $7451 per ADR) (Table 6). ADRs that were determined to add incremental expense were associated with significant increases in total hospitalization cost (mean $118,429 vs $54,858, P œ≠ .02) as well as cost after the ADR (mean Table 4 Characteristics and Root Causes of Adverse Drug Reactions (ADRs) (n œ≠ 238) Characteristic of ADR Abnormal coagulation studies, n (%) 172 (72.3) Occurrences of excessive anticoagulation during 171 (71.9) hospitalization, n (%) Any bleeding event, n (%) 59 (24.8) Occult bleed 21 (8.8) Bleeding event related to surgery 14 (5.9) Hematoma 8 (3.4) Any gastrointestinal bleeding 6 (2.5) Decrease in hematocrit 6 (2.5) Retroperitoneal hemorrhage 3 (1.3) Median length of stay post bleeding event, days 9 (5-20) (interquartile range) Blood transfusion administered, n (%) 40 (16.8) Blood transfusion administered with 48 hours 13 (5.5) after surgery, n (%) Treatment of bleeding ADR, n (%) Vitamin K 10 (4.2) Protamine 2 (0.8) Fresh frozen plasma 11 (4.6) Surgery 6 (2.5) Catheterization laboratory 1 (0.4) GUSTO bleeding classification, n (%) I 19 (8) II 32 (13.5) III 8 (3.4) Thrombocytopenia, n (%) 43 (18.1) Heparin-induced thrombocytopenia, n (%) 31 (13) Any thromboembolic event, n (%) 16 (6.7) Deep vein thrombosis 6 (2.5) Pulmonary embolism 6 (2.5) Myocardial infarction 1 (0.4) Stroke 2 (0.8) Other arterial thromboembolic event 1 (0.4) Thromboembolic ADR requiring treatment, n (%) 8 (3.4) Treatment of thromboembolic ADR, n (%) Catheterization laboratory 3 (1.3) Interventional radiology procedure 1 (0.4) Surgery 2 (0.8) Root cause of ADR Undetected predisposing condition, n (%) 138 (58) Medication error, n (%) 94 (39.5) Drug-drug interaction, n (%) 3 (1.3) Patient medication non-adherence, n (%) 2 (0.8) GUSTO œ≠ Global Use of Strategies to Open Occluded Coronary Arteries study. 1140 Table 5 The American Journal of Medicine, Vol 124, No 12, December 2011 Outcomes of Adverse Drug Events (ADEs) (n œ≠ 463) Outcome In-hospital death, n (%) Death after discharge but within 30 days of ADE, n (%) Death due to thromboembolism, n (%) Death due to bleeding event, n (%) Discharge status, n (%) Home Rehabilitation center Hospice Acute care facility Deceased Readmission within 30 days of ADE, n (%) 26 (5.6) 23 (5) 4 (0.9) 1 (0.2) 236 186 13 2 26 81 (51) (40.2) (2.8) (0.4) (5.6) (17.5) $89,733 vs $23,680, P œ≠ .004) compared with ADRs in which no incremental cost was determined to be incurred (Table 7). ADRs that were determined to add incremental cost to hospitalization were associated with significant increases in costs of nursing, pharmacy, blood products, clinical laboratory, and radiology. DISCUSSION We found that 48.8% of anticoagulant-associated ADEs were medication errors, 30.5% were ADRs, and 20.7% involved both medication errors and ADRs. Seventy percent of all reported anticoagulant-associated ADEs were potentially preventable. Transcription errors were the most common root cause of medication errors, while undetected predisposing conditions and medication errors were the most frequent root causes of ADRs. We noted high 30-day mortality (11%) in patients who experienced an anticoagulantassociated ADE during hospitalization. We observed high post-ADE hospitalization costs, largely attributable to nursing and pharmacy expenditures. We had implemented computerized provider order entry, an electronic medication administration record, ‚Äúsmart‚Äù infusion pumps, and barcode technology before the registry was instituted. Therefore, we were surprised to observe that 70% of anticoagulant-associated ADEs were potentially preventable and that 40% of related ADRs were due to medication errors. Our data corroborate previous reports demonstrating that a substantial proportion of ADEs are due to medication errors and therefore potentially preventable.4,5 Computerized provider order entry,5 improved infusion pump technology,11 and implementing barcode technology8,9 reduce medication errors. However, our root cause analysis suggests that further improvements can be made to reduce anticoagulant-associated medication errors, particularly those due to transcription errors. While computerized provider order entry, ‚Äúsmart‚Äù infusion pump technology, electronic medication administration records, and barcode technology reduce the frequency of some transcription errors, human error may occur at transition points among these clinical tools and result in mistakes at any number of steps in medication transcription. The emergence of novel oral anticoagulants that are administered in fixed doses may further reduce dosing and infusion pump errors.19 We observed high 30-day mortality (11%) in patients suffering an anticoagulant-associated ADE during hospitalization. We speculate that this may be related to a combination of high medical acuity of patients suffering anticoagulant-associated ADEs and complications of the ADEs themselves. Patients with anticoagulant-associated ADEs had a high frequency of medical conditions such as heart failure, ischemic heart disease, chronic kidney disease, and stroke, all of which may increase patient vulnerability to the complications of anticoagulant-associated ADEs. Furthermore, anticoagulant-associated ADEs may be a marker for fragile patients with complicated hospital courses and long lengths of stay. The contribution of medical errors, including those associated with ADEs, to deaths occurring during hospitalization, may be overestimated in studies relying on physician review.20 Regardless of whether anticoagulantassociated ADEs contribute to increased mortality or are Table 6 (ADRs)* Cost Associated with Adverse Drug Reactions Mean ($) œÆ SD‚Ć Overall and departmental costs Total hospitalization Hospitalization pre-ADR Hospitalization post-ADR Nursing Pharmacy Blood products Clinical laboratory Radiology Surgical Interventional procedure Anesthesia Drug-related costs Argatroban Bivalirudin Thrombolytic therapy Low-molecular weight heparin Unfractionated heparin Lepirudin Warfarin Fondaparinux Aprotinin Vitamin K Desmopressin Protamine Aminocaproic acid 88,842 œÆ 166,708 29,851 œÆ 49,264 58,991 œÆ 146,065 33,189 œÆ 81,236 7451 œÆ 21,878 2318 œÆ 9235 2001 œÆ 5069 1864 œÆ 4175 225 œÆ 1349 210 œÆ 1485 46 œÆ 257 1570 œÆ 7097 1102 œÆ 5879 87 œÆ 579 49 œÆ 154 35 œÆ 64 34 œÆ 327 23 œÆ 45 14 œÆ 76 4 œÆ 51 2œÆ6 2 œÆ 12 1œÆ6 0œÆ1 ADR œ≠ adverse drug reaction. *Patients may have suffered multiple ADRs during a hospital admission. Costs incurred were analyzed after the initial event. ‚ĆCosts are rounded to the nearest US dollar. Piazza et al Anticoagulation-associated Adverse Drug Events Table 7 1141 Incremental Cost Associated with Adverse Drug Reactions (ADRs)* Overall and departmental costs (Mean œÆ SD)‚Ć Total hospitalization Hospitalization pre-ADR Hospitalization post-ADR Nursing Pharmacy Blood products Clinical laboratory Radiology Surgical Interventional procedure Anesthesia Drug-related costs (Mean œÆ SD)‚Ć Argatroban Bivalirudin Thrombolytic therapy Lepirudin Low-molecular weight heparin Unfractionated heparin Warfarin Fondaparinux Aprotinin Vitamin K Desmopressin Protamine Aminocaproic acid No Incremental Cost (n œ≠ 85) Incremental Cost (n œ≠ 74) P-Value 54,858 œÆ 115,659 31,178 œÆ 58,172 23,680 œÆ 65,283 14,008 œÆ 42,457 2830 œÆ 9243 548 œÆ 1703 902 œÆ 2469 798 œÆ 2228 0œÆ0 0œÆ0 0œÆ0 118,429 œÆ 196,840 28,696 œÆ 40,265 89,733 œÆ 185,395 49,888 œÆ 101,194 11,473 œÆ 28,123 3858 œÆ 12,359 2957 œÆ 6406 2792 œÆ 5158 421 œÆ 1828 393 œÆ 2019 87 œÆ 348 .02 .75 .004 .005 .01 .02 .01 .002 .0496 .1 .03 231 œÆ 1441 70 œÆ 366 24 œÆ 205 0œÆ0 53 œÆ 183 27 œÆ 67 15 œÆ 43 0œÆ4 0œÆ0 1œÆ5 0œÆ0 0œÆ0 0œÆ0 2735 œÆ 9486 2001 œÆ 7946 142 œÆ 766 63 œÆ 446 45 œÆ 125 41 œÆ 60 30 œÆ 47 26 œÆ 102 8 œÆ 69 3œÆ7 3 œÆ 17 3œÆ8 0œÆ1 .03 .04 .2 .23 .72 .14 .04 .03 .35 .01 .17 .003 .21 *Patients may have suffered multiple ADRs during a hospital admission. Costs incurred were analyzed after the initial event. ‚ĆCosts are rounded to the nearest US dollar. markers for higher medical acuity, patients who suffer these events represent a particularly vulnerable population. Accordingly, Quality Improvement initiatives to reduce anticoagulant-associated ADEs are critical. Most hospitalization costs after an anticoagulant-associated ADE were attributable to nursing and pharmacy expenses. ADRs that incurred incremental cost doubled the total hospitalization costs and nearly quadrupled postevent costs compared with ADRs that did not incur any incremental expense. Our findings are similar to other reports that demonstrate the heavy financial burden of ADEs.13,21,22 Based on these data, efforts to reduce anticoagulant-associated ADEs have the potential to reduce hospitalization costs as well as improve patient safety. Our data were obtained from a voluntary patient safety reporting system and therefore might be limited by underreporting and selective reporting, in which only the most severe ADEs or those with the greatest consequences are reported. Because we excluded patient care areas in which medication administration records were not consistently computerized, we may have omitted subgroups of patients who had suffered a higher proportion of anticoagulant-associated ADEs. Because anticoagulant medications are given throughout the hospital but not all patient care areas have electronic medication administration records, it is not possible to calculate the denominator of patients exposed to anticoagulants. Therefore, we could not calculate the incidence of anticoagulantassociated ADEs and could not identify an appropriate comparison population of patients exposed to anticoagulant drugs who did not suffer ADEs. In addition, our analysis did not encompass community-acquired ADEs that resulted in hospitalization. While 5 deaths were clearly attributable to ADE-related bleeding (n œ≠ 1) or thromboembolism (n œ≠ 4), we were not able to quantify the contribution of ADEs to other deaths in this cohort. Finally, because of limitations in our cost accounting software, we were unable to separate costs incurred after the initial ADR from costs associated with subsequent events in a particular patient during the same hospitalization. The methodology utilized in this analysis is consistent with published criteria for evaluating the scientific value of clinical data registries.23 We utilized a large database generated from a tertiary care center that is representative of similar acute care facilities. Our study provides real-world insights into anticoagulant-associated ADEs from a medical center that has implemented an electronic health record that 1142 The American Journal of Medicine, Vol 124, No 12, December 2011 integrates computerized provider order entry, an electronic medication administration record, ‚Äúsmart‚Äù infusion pumps, and barcode technology. We utilized a multidisciplinary group of physicians, pharmacists, and a patient safety officer to provide the highest accuracy for classifying events and determining root causes of ADEs. Finally, we provided outcomes data for patients who suffered anticoagulantassociated ADEs with 100% follow-up. CONCLUSIONS Most anticoagulant-associated ADEs among inpatients result from medication errors and are therefore potentially preventable. Transcription errors are the most common cause of medication errors which, in turn, are a frequent cause of anticoagulant-associated ADRs. We observed elevated 30-day mortality among patients who suffered an anticoagulant-associated ADE and high hospitalization costs following ADRs. Further efforts to reduce anticoagulant-associated medication errors are warranted to improve patient safety and decrease health care expenditures. References 1. Zaidenstein R, Eyal S, Efrati S, et al. Adverse drug events in hospitalized patients treated with cardiovascular drugs and anticoagulants. Pharmacoepidemiol Drug Saf. 2002;11:235-238. 2. Silverman JB, Stapinski CD, Churchill WW, et al. Multifaceted approach to reducing preventable adverse drug events. Am J Health Syst Pharm. 2003;60:582-586. 3. Piazza G, Goldhaber SZ. Computerized decision support for the cardiovascular clinician: applications for venous thromboembolism prevention and beyond. Circulation. 2009;120:1133-1137. 4. Morimoto T, Sakuma M, Matsui K, et al. Incidence of adverse drug events and medication errors in Japan: the JADE Study. J Gen Intern Med. 2010;26:148-153. 5. Hug BL, Witkowski DJ, Sox CM, et al. Adverse drug event rates in six community hospitals and the potential impact of computerized physician order entry for prevention. J Gen Intern Med. 2009;25:31-38. 6. Classen DC, Jaser L, Budnitz DS. Adverse drug events among hospitalized Medicare patients: epidemiology and national estimates from a new approach to surveillance. Jt Comm J Qual Patient Saf. 2010;36: 12-21. 7. Bates DW, Cullen DJ, Laird N, et al. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group. JAMA. 1995;274:29-34. 8. Poon EG, Keohane CA, Yoon CS, et al. Effect of bar-code technology on the safety of medication administration. N Engl J Med. 2010;362: 1698-1707. 9. Poon EG, Cina JL, Churchill W, et al. Medication dispensing errors and potential adverse drug events before and after implementing bar code technology in the pharmacy. Ann Intern Med. 2006;145:426-434. 10. Leape LL, Bates DW, Cullen DJ, et al. Systems analysis of adverse drug events. ADE Prevention Study Group. JAMA. 1995;274:35-43. 11. Grissinger MC, Hicks RW, Keroack MA, et al. Harmful medication errors involving unfractionated and low-molecular-weight heparin in three patient safety reporting programs. Jt Comm J Qual Patient Saf. 2010;36:195-202. 12. Fanikos J, Stapinski C, Koo S, et al. Medication errors associated with anticoagulant therapy in the hospital. Am J Cardiol. 2004;94:532-535. 13. Bates DW, Spell N, Cullen DJ, et al. The costs of adverse drug events in hospitalized patients. Adverse Drug Events Prevention Study Group. JAMA. 1997;277:307-311. 14. Fanikos J, Cina JL, Baroletti S, et al. Adverse drug events in hospitalized cardiac patients. Am J Cardiol. 2007;100:1465-1469. 15. Hanlon JT, Pieper CF, Hajjar ER, et al. Incidence and predictors of all and preventable adverse drug reactions in frail elderly persons after hospital stay. J Gerontol A Biol Sci Med Sci. 2006;61:511-515. 16. Hajjar ER, Hanlon JT, Artz MB, et al. Adverse drug reaction risk factors in older outpatients. Am J Geriatr Pharmacother. 2003;1:82-89. 17. Nebeker JR, Barach P, Samore MH. Clarifying adverse drug events: a clinician‚Äôs guide to terminology, documentation, and reporting. Ann Intern Med. 2004;140:795-801. 18. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. The GUSTO investigators. N Engl J Med. 1993;329:673-682. 19. Eikelboom JW, Weitz JI. New anticoagulants. Circulation. 2010;121: 1523-1532. 20. Hayward RA, Hofer TP. Estimating hospital deaths due to medical errors: preventability is in the eye of the reviewer. JAMA. 2001;286: 415-420. 21. Jennings HR, Miller EC, Williams TS, et al. Reducing anticoagulant medication adverse vents and avoidable patient harm. Jt Comm J Qual Patient Saf. 2008;34:196-200. 22. Eckman MH, Levine HJ, Pauker SG. Making decisions about antithrombotic therapy in heart disease. Decision analytic and cost-effectiveness issues. Chest. 1995;108:457S-470S. 23. Alpert JS. Are data from clinical registries of any value? Eur Heart J. 2000;21:1399-1401. CLINICAL RESEARCH STUDY Antidepressant Use and Cognitive Decline: The Health and Retirement Study Jane S. Saczynski, PhD,a,b,c Allison B. Rosen, MD,b,c Ryan J. McCammon, AB,d Kara Zivin, PhD,e,f,g,h Susan E. Andrade, PhD,b Kenneth M. Langa, MD, PhD,d,f,h Sandeep Vijan, MD,d,f Paul A. Pirraglia, MD,i Becky A. Briesacher, PhDj a Department of Medicine, University of Massachusetts Medical School, Worcester; bMeyers Primary Care Institute, Worcester, Mass; Department of Quantitative Health Sciences, University of Massachusetts Medical School, Worcester; dDivision of General Medicine, Department of Medicine, University of Michigan Medical School, Ann Arbor; eDepartment of Veterans Affairs, National Serious Mental Illness Treatment Resource and Evaluation Center, Ann Arbor, Mich; fDepartment of Veterans Affairs, Center for Clinical Management Research, Ann Arbor, Mich; gDepartment of Psychiatry, University of Michigan Medical School, Ann Arbor; hInstitute for Social Research, University of Michigan, Ann Arbor; iDepartment of Medicine, Alpert School of Medicine at Brown University, Providence, RI; jDepartment of Pharmacy and Health Systems Sciences, Northeastern University, Boston, Mass. c ABSTRACT BACKGROUND: Depression is associated with cognitive impairment and dementia, but whether treatment for depression with antidepressants reduces the risk for cognitive decline is unclear. We assessed the association between antidepressant use and cognitive decline over 6 years. METHODS: Participants were 3714 adults aged 50 years or more who were enrolled in the nationally representative Health and Retirement Study and had self-reported antidepressant use. Depressive symptoms were assessed using the 8-item Center for Epidemiologic Studies Depression Scale. Cognitive function was assessed at 4 time points (2004, 2006, 2008, 2010) using a validated 27-point scale. Change in cognitive function over the 6-year follow-up period was examined using linear growth models, adjusted for demographics, depressive symptoms, comorbidities, functional limitations, and antidepressant anticholinergic activity load. RESULTS: At baseline, cognitive function did not differ signiÔ¨Åcantly between the 445 (12.1%) participants taking antidepressants and those not taking antidepressants (mean, 14.9%; 95% conÔ¨Ådence interval, 14.315.4 vs mean, 15.1%; 95% conÔ¨Ådence interval, 14.9-15.3). During the 6-year follow up period, cognition declined in both users and nonusers of antidepressants, ranging from √Ä1.4 change in mean score in those with high depressive symptoms and taking antidepressants to √Ä0.5 change in mean score in those with high depressive symptoms and not taking antidepressants. In adjusted models, cognition declined in people taking antidepressants at the same rate as those not taking antidepressants. Results remained consistent across different levels of baseline cognitive function, age, and duration of antidepressant use (prolonged vs short-term). CONCLUSIONS: Antidepressant use did not modify the course of 6-year cognitive change in this nationally representative sample. √ì 2015 Elsevier Inc. All rights reserved.  The American Journal of Medicine (2015) 128, 739-746 KEYWORDS: Antidepressants; Cognition; Depression; Epidemiology Funding: See last page of article. ConÔ¨Çict of Interest: See last page of article. Authorship: See last page of article. Requests for reprints should be addressed to Jane S. Saczynski, PhD, Division of Geriatric Medicine, Department of Medicine, 377 Plantation St, Suite 315, Worcester, MA 01605. E-mail address: Jane.saczynski@umassmed.edu 0002-9343/$ -see front matter √ì 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjmed.2015.01.007 Depression is associated consistently with cognitive impairment and an increased risk for dementia in clinical and epidemiologic studies of older adults.1-8 A meta-analysis of case-control and prospective studies concluded that a history of depression approximately doubled the risk for dementia.1 However, whether treatment for depression can reduce the rate of cognitive impairment and dementia is unclear.9-17 740 The American Journal of Medicine, Vol 128, No 7, July 2015 Research to date on the effects of antidepressants on the that includes assessments of depressive symptoms and rate of decline in cognitive function has been limited by cognitive function. The HRS began in 1992, and particismall sample sizes (<100 patients), short (<12 months) pants are re-interviewed every 2 years with high follow-up follow-up, lack of a comparison group, and selected subrates (90%-95%).22 The PDS was a mail survey distribgroups of patients, such as those who responded to therapy uted to a subsample of the HRS, drawn from respondents to (ie, those whose depressive symptoms decreased after the 2004 HRS wave, designed to track changes in pretreatment).11-17 A number of scription drug use among beneÔ¨Åciaries as Medicare Part D was studies have also examined CLINICAL SIGNIFICANCE phased in. The PDS sample inwhether antidepressant treatment cludes HRS respondents born in can slow the rate of cognitive  Cognitive function did not differ at 1942 or earlier (age !65 years in decline or progression of dementia baseline between antidepressant users 2007) or those who were already among patients with preexisting and nonusers. covered by Medicare or Medicaid cognitive impairments or demen Rate of decline in cognitive function between 2002 and 2004. This tia.15,18-20 From these studies, it analysis starts with the 2004 data has been difÔ¨Åcult to disentangle over 6 years did not differ between anto correspond with the Ô¨Årst wave the association between antidetidepressant users and nonusers. of the PDS. This study was pressant treatment and long-term  Treatment with antidepressant medicaapproved by the University of cognitive outcomes in a broadly tion use does not seem to modify the Michigan Institutional Review representative population of inwell-established association between Board; participants provided dividuals with a wide range of depression and cognitive decline. informed consent at enrollment. cognitive function, from nonThe current study included all impaired to moderately impaired, PDS respondents born before when depression is assessed. 1943 who were community dwelling, self-respondents Little work has been done on the relationship between providing cognitive function and depressive symptom asantidepressant use and cognition in population-based samsessments in the 2004 HRS interview. Participants‚Äô cogniples. One study of 595 patients found that antidepressant use tive function was assessed at each wave through the 2010 was associated with an increased risk of cognitive decline survey, with up to 4 total assessments. Our primary analyses over 4.5 years among depressed patients without cognitive focus on change in cognitive function in patients who were, impairment.10 Within the Women‚Äôs Health Initiative Memcompared with those who were not, taking antidepressants ory Study, antidepressant use was associated with a 70% according to the PDS. increased risk of incident mild cognitive impairment over 7.5 years.9 Of note, although the Women‚Äôs Health Initiative Memory Study collected information on the type of antiAssessment of Cognitive Function depressant taken, neither study assessed the anticholinergic activity of antidepressants. Medications with anticholinergic At each HRS wave, cognitive function was assessed using a effects, including amitriptyline, doxepin, paroxetine, and previously described and validated 27-point scale based on a nortriptyline among others, can block muscarinic receptors battery of tests that included tests of memory, serial 7 causing impairment in various cognitive functions, subtractions, and naming.23 This battery is a subset of an including memory, executive function, and processing expanded battery (range, 0-35) administered to participants speed.21 aged !65 years in the HRS; the expanded battery includes measures of orientation.24 Participants requiring proxy inWe examined whether antidepressant use was associated with cognitive decline over a 6-year period using data from terviews (because of cognitive or physical impairments that the nationally representative Health and Retirement Study limited their ability to self-respond) were excluded from this (HRS) and the HRS Prescription Drug Study (PDS), which analysis. Cut points for cognitive function were based on include serial assessments of cognitive function, depressive prior studies with the HRS data,25,26 as well as methods symptoms, and antidepressant treatment. We hypothesized used for the Aging, Demographics, and Memory Study, a that treatment with antidepressant medications would be supplemental study of dementia in the HRS.27 These cut associated with slower rates of cognitive decline. points deÔ¨Åned a level of cognitive function that was generally consistent with normal function (12-27 points), cognitive impairment but no dementia (7-11 points), and dementia (0-6 points).28 MATERIALS AND METHODS Data Source Study data were drawn from the 2004, 2006, 2008, and 2010 waves of the HRS, and from the 2005 and 2007 waves of the PDS. The HRS is a longitudinal, nationally representative survey of US residents aged 51 years and older Assessment of Depression Baseline depression status was based on the 2004 wave of the HRS using an 8-item version of the Center for Epidemiologic Studies Depression Scale.29,30 This version of the Saczynski et al Antidepressant Use and Cognitive Decline 741 Center for Epidemiologic Studies Depression Scale included the following questions with ‚Äúyes/no‚Äù response options: Much of the time during the last week, I felt depressed. I felt everything I did was an effort. My sleep was restless. I was happy (reverse coded). I felt lonely. I enjoyed life (reverse coded). I felt sad. I could not ‚Äúget going.‚Äù The total number of ‚Äúyes‚Äù responses were summed to calculate an overall depression score (range, 0-8); participants with scores !4 were considered to have high depressive symptomatology.30 We conducted several sensitivity analyses: one in which we stratiÔ¨Åed the analysis on the basis of cognitive function at baseline (no impairment, moderate impairment, dementia) and a second one in which we stratiÔ¨Åed the analysis on the basis of age at baseline (61-74 vs !75 years) to examine whether the association between antidepressant use and change in cognitive function varied according to level of cognitive function or age. In additional sensitivity analyses, we restricted the sample to those taking antidepressants at baseline (n ¬º 445) and examined the effect of self-reported antidepressant use >1 year at baseline compared with shorter-term use and the effect of continuous antidepressant use over the entire follow-up period compared with discontinuation. All analyses were conducted using SAS version 9.3 (SAS Institute Inc, Cary NC) and MPlus version 7.1. Assessment of Antidepressant Use Antidepressant use was assessed using the 2005 and 2007 PDS. Respondents were asked to list all medications prescribed and directed to provide the information on drug name and dosage directly from the prescription bottle label. Medications were then matched to the American Hospital Formulary System, a database of all prescription drugs. Respondents were considered to be taking an antidepressant if any of their medications were matched to therapeutic class code 28:16.04, which includes selective serotonin reuptake inhibitors, serotonin norepinephrine reuptake inhibitors, other novel antidepressants (eg, bupropion), tricyclic antidepressants, and monoamine oxidase inhibitors. Other Study Variables The following sociodemographic variables from the 2004 HRS survey were included as covariates: age, sex, education (less than high school vs high school or more), race (nonHispanic-white, non-Hispanic-black, Hispanic, and other), chronic conditions (count of 7 self-reported conditions, including high blood pressure, diabetes, heart disease, stroke, lung disease, cancer, and arthritis), self-reported visual or hearing impairment, dependence in Activities of Daily Living,31 and self-reported duration of antidepressant use. We also included a measure of baseline anticholinergic burden (based on the Anticholinergic Drug Scale),32 because anticholinergic activity is high in some antidepressant medications and is associated with cognitive decline.33 Analyses Baseline (2004) characteristics were compared for participants who were taking antidepressants as assessed by the 2005 PDS compared with those who were not taking antidepressants using the chi-square test for categoric variables and analysis of variance analyses for continuous variables. A linear latent growth model was used to examine change on the 27-point cognitive battery over the 6-year follow-up. In this model, time is measured as years since baseline (2004), and random effects are used to capture individual differences in baseline levels (intercepts) and rates of change (slopes). We adjusted for depression, age, sex, education, race, hearing/visual impairment, comorbidity burden, and Activities of Daily Living impairment. We also adjusted for anticholinergic activity load of medications reported in the 2005 PDS. RESULTS Medication data from the 2005 PDS were available for 4320 individuals. Participants born after 1942 (n ¬º 185), those who required a proxy respondent (n ¬º 373), and those with incomplete data on the Center for Epidemiologic Studies Depression Scale (n ¬º 48) were excluded from the study sample, resulting in an analytic sample of 3714 participants. Of the 3714 eligible respondents, 445 (12%) were taking antidepressants in 2005. Participants taking antidepressants were more likely to be female, to drink alcohol daily, and to be white in contrast to participants not taking antidepressants (Table 1). Those taking antidepressants had a higher burden of comorbidity; they were more likely to need assistance with Activities of Daily Living and Instrumental Activities of Daily Living and more likely to have hearing and visual impairments. As expected, participants taking antidepressants were more likely to report a higher level of depressive symptoms than participants not taking antidepressants. Age, education, marital status, and smoking status did not differ according to antidepressant use (Table 1). Antidepressant Use and Cognitive Function Table 2 shows unadjusted cognitive performance by baseline antidepressant use and depressive symptoms. At baseline, no signiÔ¨Åcant differences were detected in average cognition scores between the 445 (12.1%) participants taking antidepressants and those not taking antidepressants (mean 14.9, 95% conÔ¨Ådence interval [CI], 14.3-15.4 vs mean 15.1%; 95% CI, 14.9-15.3). During the 6-year follow-up period, both users and nonusers of antidepressants experienced cognitive decline ranging from √Ä1.4 change in mean score (95% CI, √Ä1.1 to √Ä1.7) in people with high depressive symptoms and taking antidepressants to √Ä0.5 change in mean score (95% CI, √Ä0.6 to √Ä0.3) in people with high depressive symptoms and not taking antidepressants. In our initial linear growth model, adjusted for age and depressive symptoms, there was a nonsigniÔ¨Åcant negative association between antidepressant use and baseline cognitive 742 Table 1 The American Journal of Medicine, Vol 128, No 7, July 2015 Sample Characteristics by 2005 Antidepressant Use Antidepressant Use Measure Baseline age, mean, y (95% CI) Sex, % female Education less than high school, % White Black Hispanic/other Married, % Current smoker, % Intake of 1√æ drinks per day, % CES-D, mean (95% CI) CES-D depression* % No. of chronic conditions, % 0 1 2 3√æ Chronic conditions, mean (95% CI) Visual impairment Hearing impairment 2005 AA load 0 1 or 2 3√æ 2005 AA load, mean (95% CI) ADL, % any dependence IADL, % any dependence Duration antidepressant use, % !1 y No (n ¬º 3269) % (95% CI) Yes (n ¬º 445) % (95% CI) 72.0 (71.7-72.4) 56.7 23.6 85.0 8.1 6.9 57.4 10.1 12.3 1.3 (1.2-1.4) 12.0 71.9 (70.4-72.8) 69.3 20.7 93.9 2.3 3.8 52.3 13.5 10.8 2.3 (2.0-2.6) 26.6 13.4 26.3 29.8 30.5 1.9 (1.9-2.0) 19.2 23.2 8.3 20.8 30.1 40.8 2.3 (2.1-2.5) 25.8 25.7 64.8 26.1 9.1 0.8 (0.7-0.8) 15.2 11.9 ‚Äî 14.3 52.8 33.0 2.1 (2.0-2.3) 28.9 23.0 89.8 P Value .72 <.001 .91 <.001 .09 .042 .41 <.001 <.001 <.001 <.001 .006 .34 <.001 <.001 <.001 <.001 AA ¬º anticholinergic activity; ADL ¬º Activities of Daily Living; CES-D ¬º Center for Epidemiologic Studies Depression Scale; CI ¬º conÔ¨Ådence interval; IADL ¬º Instrumental Activities of Daily Living. *CES-D !4. function (Intercept; Table 3, Model 1), but a signiÔ¨Åcant decline in cognitive function over the 6-year study period in antidepressant users (Slope; Table 3, Model 1). Depression and age were both negatively associated with baseline level of function, but only age was associated with change (decline) in cognitive function over time. In subsequent models, we found a nonsigniÔ¨Åcant negative association between antidepressant use and change in cognitive function when we further adjusted for sex, education, race, Activities of Daily Living impairments, comorbidity burden (Table 3, Model 2), and anticholinergic load (Table 3, Model 3). Figure 1 presents unadjusted model-based estimates by baseline depressive symptom status (high/low) and antidepressant use corresponding to Model 1 (Table 3) for ages 72 years (2004) to 78 years (2010). Subgroup and Sensitivity Analyses We conducted several sensitivity analyses. First, we examined whether the association between antidepressant use and change in cognitive function varied by baseline level of cognitive function or by age. In linear growth models adjusted for age and depression, antidepressant use was associated with a signiÔ¨Åcant decline in cognitive function among participants with normal cognitive function at baseline (n ¬º 2817; Slope ¬º √Ä0.094 [standard error ¬º 0.04], P < .05) but nonsigniÔ¨Åcant decline in cognitive function among participants with cognitive impairment but no dementia (n ¬º 721; Slope ¬º √Ä0.095 [0.10]) and those who were demented (n ¬º 176; Slope ¬º √Ä0.133 [0.17]). Of note, although the association between antidepressant use and cognitive decline was signiÔ¨Åcant in the participants with normal cognitive function at baseline and not in the 2 cognitively impaired groups, we observed a negative association between antidepressant use and cognitive change across levels of baseline cognitive function. In additional models and similar to Ô¨Åndings in the overall cohort, there was a nonsigniÔ¨Åcant association between antidepressant use and cognitive decline in younger participants (age 61-74 years, n ¬º 2398; Slope ¬º √Ä0.065 [standard error ¬º 0.05]) and older participants (age 75√æ years, n ¬º 1316; Slope ¬º √Ä0.133 [standard error ¬º 0.08]). We also conducted sensitivity analyses among participants taking antidepressants at baseline (n ¬º 414) to examine Saczynski et al Table 2 Antidepressant Use and Cognitive Decline 743 Changes in Cognitive Functioning Over 6-Year Period by Baseline Antidepressant Use Baseline (95% CI) Time 1 (95% CI) Time 2 (95% CI) Time 3 (95% CI) Low depressive symptoms and not taking antidepressants Normal, % Borderline, % Impaired, % Mean score n 81.3 (79.6-83.0) 15.6 (13.8-17.4) 3.0 (2.4-3.6) 15.3 (15.1-15.6) 2832 79.8 (77.6-81.9) 15.0 (13.3-16.7) 5.3 (4.4-6.1) 15.1 (14.9-15.4) 2673 78.2 (75.9-80.5) 16.3 (14.4-18.3) 5.5 (4.4-6.5) 15.0 (14.7-15.2) 2424 73.2 (70.9-75.4) 17.9 (16.2-19.5) 9.0 (7.7-10.2) 14.5 (14.3-14.8) 2071 Low depressive symptoms and taking antidepressants Normal, % Borderline, % Impaired, % Mean score n 80.6 14.8 4.6 15.1 324 (75.0-86.3) (9.8-19.7) (1.8-7.4) (14.4-15.8) 74.4 16.6 9.0 14.6 302 (69.7-79.0) (12.5-20.7) (5.4-12.6) (14.0-15.2) 72.4 15.8 11.8 14.5 262 (65.8-79.0) (10.4-21.3) (7.8-15.8) (13.8-15.3) 68.7 20.6 10.6 14.4 219 (61.8-75.7) (15.4-25.8) (6.3-14.9) (13.8-15.1) High depressive symptoms and not taking antidepressants Normal, % Borderline, % Impaired, % Mean score n 64.0 29.3 6.7 13.1 437 (57.9-70.1) (24.3-34.4) (3.5-9.8) (12.6-13.6) 62.6 25.8 11.7 13.1 401 (56.9-68.2) (20.8-30.7) (7.7-15.6) (12.5-13.6) 60.1 28.1 11.8 13.2 341 (52.5-67.7) (21.8-34.5) (8.2-15.3) (12.5-13.8) 58.0 25.1 16.9 12.6 257 (51.8-64.3) (19.7-30.5) (11.6-22.1) (11.9-13.3) High depressive symptoms and taking antidepressants Normal, % Borderline, % Impaired, % Mean score n 74.1 19.5 6.4 14.2 121 (65.2-83.0) (11.9-27.1) (1.7-11.1) (13.3-15.1) 67.4 23.1 9.4 13.6 114 (56.7-78.2) (14.0-32.2) (3.7-15.2) (12.4-14.7) 69.8 19.7 10.4 13.7 93 (58.1-81.5) (10.3-29.2) (3.6-17.2) (12.5-14.8) 54.2 29.9 15.9 12.8 71 (40.6-67.9) (19.3-40.6) (7.0-24.7) (11.6-14.0) CI ¬º conÔ¨Ådence interval. the effect of duration and discontinuation of antidepressant use. In linear growth models adjusted for depressive symptoms and age, duration of antidepressant use at baseline and continuity of antidepressant use over the follow-up period were nonsigniÔ¨Åcantly associated with cognitive change (Slope for >1 year use ¬º 0.103 [standard error ¬º 0.10]) (Slope for continuous use ¬º 0.026 [standard error ¬º 0.09]). DISCUSSION In a nationally representative cohort of older adults, we found that antidepressant use was not associated with changes in cognitive function over a 6-year period, adjusting for depressive symptoms, comorbidity burden, and anticholinergic load. Findings were consistent across levels of baseline cognitive function (normal function, mild cognitive impairment, and dementia). The results suggest that pharmacologic treatment for depression does not modify the negative association observed between depression and cognitive decline in clinical and epidemiologic studies. Many clinical trials (mostly small) have been conducted to assess the association between antidepressant treatment and cognitive decline, and Ô¨Åndings have been mixed. When responders and nonresponders are examined separately, responders to treatment tend to show improvements in cognitive function,11-13 whereas nonresponders show no improvement12 or decline.13 Other studies have shown only modest improvements in cognition after treatment with antidepressants, equal to the practice effects observed in cognitive performance in control groups.14-17 One study found that treatment with antidepressants decreased the risk of dementia, but only among patients with mild cognitive impairment at baseline.15 Little work has been published on treatment with antidepressants and cognitive trajectories in population-based studies. However, the Ô¨Åndings suggest that, among nonimpaired participants, antidepressant use is associated with an increased risk of incident mild cognitive impairment in up to 7.5 years of follow-up.9,10 We found a nonsigniÔ¨Åcant decline in cognitive function over a 6-year follow-up period among participants taking antidepressants at baseline. Our sample was not restricted to cognitively healthy participants at baseline, and results were similar across strata of baseline cognitive function (normal, mild cognitive impairment, and dementia). To our knowledge, the effect of anticholinergic activity load has not been controlled or accounted for in previous studies of the longitudinal association between antidepressant use and cognitive decline. High anticholinergic activity 744 The American Journal of Medicine, Vol 128, No 7, July 2015 Table 3 Linear Latent Growth Models of 2004-2008 Health and Retirement Study 27-Point Cognition Measure (n ¬º 3714) Model Intercept Antidepressant use CES-D 4√æ Age Female <12 y of school Black Hispanic Impaired vision Impaired hearing Chronic conditions (range, 0-7) 1 ADL impairment 2√æ ADL impairments AA load of 1 or 2 AA load of 3√æ Mean Residual variance Slope Antidepressant use CES-D 4√æ Age Female <12 y of school Black Hispanic Impaired vision Impaired hearing Chronic conditions 1 ADL impairment 2√æ ADL impairments AA load of 1 or 2 AA load of 3√æ Mean Residual variance 1 CoefÔ¨Åcient (SE) 2 CoefÔ¨Åcient (SE) 3 CoefÔ¨Åcient (SE) √Ä0.11 (0.25) √Ä1.72‚Ä° (0.23) √Ä0.18‚Ä° (0.01) √Ä0.39 √Ä0.54‚Ä° √Ä0.16‚Ä° 0.42‚Ć √Ä2.72‚Ä° √Ä2.84‚Ä° √Ä2.44‚Ä° √Ä0.28 √Ä0.68‚Ä° √Ä0.11 √Ä0.65‚Ć √Ä1.04‚Ć √Ä0.36 √Ä0.54‚Ä° 0.16‚Ä° √Ä0.43‚Ć √Ä2.72‚Ä° √Ä2.85‚Ä° √Ä2.45‚Ä° √Ä0.27 √Ä0.67‚Ä° √Ä0.10 √Ä0.64‚Ć √Ä1.03‚Ć √Ä0.03 √Ä0.13 17.40‚Ä° (0.16) 11.33‚Ä° (0.42) 18.39‚Ä° (0.16) 8.10‚Ä° (0.37) 18.40‚Ä° (0.16) 8.10‚Ä° (0.37) √Ä0.09* (0.04) 0.002 (0.04) √Ä0.02‚Ä° (<0.01) √Ä0.8 √Ä0.01 √Ä0.02‚Ä° √Ä0.02 √Ä0.01 0.02 0.16‚Ć √Ä0.02 0.04 0.02 0.01 0.01 √Ä0.08 √Ä0.01 √Ä0.02‚Ä° √Ä0.02 √Ä0.01 0.02 0.16‚Ć √Ä0.02 0.04 0.02 <0.01 0.01 √Ä0.01 √Ä0.02 √Ä0.07‚Ć (0.02) 0.07‚Ä° (0.01) √Ä0.12‚Ä° (0.03) 0.06‚Ä° (0.01) (0.24) (0.20) (0.01) (0.14) (0.16) (0.27) (0.26) (0.19) (0.16) (0.07) (0.21) (0.32) (0.04) (0.05) (<0.01) (0.03) (0.03) (0.04) (0.05) (0.03) (0.03) (0.01) (0.04) (0.05) (0.26) (0.20) (0.01) (0.14) (0.16) (0.27) (0.27) (0.18) (0.16) (0.06) (0.21) (0.32) (0.16) (0.20) (0.05) (0.05) (<0.01) (0.03) (0.03) (0.04) (0.05) (0.03) (0.03) (0.01) (0.04) (0.06) (0.03) (0.04) √Ä0.10‚Ä° (0.03) 0.06‚Ä° (0.01) AA ¬º anticholinergic activity; ADL ¬º Activities of Daily Living; CES-D ¬º Center for Epidemiologic Studies Depression; SE ¬º standard error. *P < .05. ‚ĆP < .01. ‚Ä°P < .001. load is recognized increasingly for its association with negative cognitive effects, and anticholinergic activity is present, often in high doses, in many antidepressants. We were surprised by the small effect of anticholinergic activity load on baseline cognitive function and cognitive decline in our study. A number of population-based studies have examined anticholinergic activity load and cognitive decline, with signiÔ¨Åcant associations observed in some,33-36 but not all,37-39 studies. The association between anticholinergic activity and cognitive decline may be related to level of cognitive function at baseline. For example, in a study of more than 13,000 patients, the strongest association between anticholinergic activity load and decline on the Mini Mental State Exam was in participants with scores in the ‚Äúnormal‚Äù range (26-30).33 We found that antidepressant use was associated with a decline in cognitive function across all levels of baseline cognitive function (normal, cognitive impairment no dementia, and dementia) and that the association was strongest among participants with normal cognitive function at baseline and before and after adjustment for anticholinergic activity load of antidepressants. Study Strengths and Limitations The present study has a number of strengths. The HRS is a nationally representative cohort, and our analysis was not Saczynski et al Antidepressant Use and Cognitive Decline 745 results were consistent across baseline levels of cognitive function and age, and did not change when we controlled for the anticholinergic activity load of antidepressants. Because depression is a well-known risk factor for cognitive decline and dementia, consideration of nonpharmacologic approaches to treatment for depression that also are related to cognitive function (eg, social engagement, physical activity, and diet) may be used in combination with pharmacologic approaches to address cognitive function and depressive symptoms. References Figure 1 Linear growth model of antidepressant use, depression, and cognition. Estimates are for participant who is aged 72 years at baseline (2004). CESD ¬º Center for Epidemiologic Studies Depression Scale. restricted to select subgroups of patients, such as those who responded to therapy or those without preexisting cognitive impairment. The PDS medication data, although selfreported, collects detailed information on medications used, so therapeutic class code can be determined by matching with the American Hospital Formulary System. As a result, we were able to code medications for anticholinergic activity load and examine the effect of high anticholinergic activity load on cognitive trajectories that has been missing from previous studies. There are several potential limitations to our study. We did not look at speciÔ¨Åc classes of antidepressants and do not know whether patients were taking antidepressants for depression or other indications, such as weight loss. Our study was not designed to examine responsiveness to antidepressant treatment; however, we controlled for depressive symptoms in our analyses. Duration of antidepressant use was self-reported and thus could have resulted in misclassiÔ¨Åcation, particularly among cognitively impaired participants. The battery of cognitive tests allowed us to examine global cognitive function but not individual cognitive domains, and it did not provide a clinical diagnosis of dementia. However, we used cut points that have been validated against clinical diagnosis of dementia.26 Likewise, the Center for Epidemiologic Studies Depression Scale assesses depressive symptoms rather than providing a clinical diagnosis of depression; however, we used a wellestablished cut point that has been validated against clinical diagnosis of depression.30 CONCLUSIONS In this large nationally representative cohort of older adults, we found that antidepressant use was not signiÔ¨Åcantly associated with 6-year change in cognitive function. These 1. Jorm AF. History of depression as a risk factor for dementia: an updated review. Aust N Z J Psychiatry. 2001;35:776-781. 2. Barnes DE, Alexopoulos GS, Lopez OL, Williamson JD, Yaffe K. Depressive symptoms, vascular disease, and mild cognitive impairment: Ô¨Åndings from the Cardiovascular Health Study. Arch Gen Psychiatry. 2006;63:273-279. 3. Jorm AF, van Duijn CM, Chandra V, et al. Psychiatric history and related exposures as risk factors for Alzheimer‚Äôs disease: a collaborative re-analysis of case-control studies. EURODEM Risk Factors Research Group. Int J Epidemiol. 1991;20:S43-S47. 4. Speck CE, Kukull WA, Brenner DE, et al. History of depression as a risk factor for Alzheimer‚Äôs disease. Epidemiology. 1995;6:366-369. 5. Green RC, Cupples LA, Kurz A, et al. Depression as a risk factor for Alzheimer disease: the MIRAGE Study. Arch Neurol. 2003;60: 753-759. 6. Saczynski JS, Beiser A, Seshadri S, Auerbach S, Wolf PA, Au R. Depressive symptoms and risk of dementia: the Framingham Heart Study. Neurology. 2010;75:35-41. 7. Dotson VM, Beydoun MA, Zonderman AB. Recurrent depressive symptoms and the incidence of dementia and mild cognitive impairment. Neurology. 2010;75:27-34. 8. Goveas JS, Espeland MA, Woods NF, Wassertheil-Smoller S, Kotchen JM. Depressive symptoms and incidence of mild cognitive impairment and probable dementia in elderly women: the Women‚Äôs Health Initiative Memory Study. J Am Geriatr Soc. 2012;59:57-66. 9. Goveas JS, Hogan PE, Kotchen JM, et al. Depressive symptoms, antidepressant use, and future cognitive health in postmenopausal women: the Women‚Äôs Health Initiative Memory Study. Int Psychogeriatr. 2012;24:1252-1264. 10. Ravaglia G, Forti P, Lucicesare A, et al. Prevalent depressive symptoms as a risk factor for conversion to mild cognitive impairment in an elderly Italian cohort. Am J Geriatr Psychiatry. 2008;16:834-843. 11. Butters MA, Becker JT, Nebes RD, et al. Changes in cognitive functioning following treatment of late-life depression. Am J Psychiatry. 2000;157:1949-1954. 12. Mandelli L, Serretti A, Colombo C, et al. Improvement of cognitive functioning in mood disorder patients with depressive symptomatic recovery during treatment: an exploratory analysis. Psychiatry Clin Neurosci. 2006;60:598-604. 13. Culang ME, Sneed JR, Keilp JG, et al. Change in cognitive functioning following acute antidepressant treatment in late-life depression. Am J Geriatr Psychiatry. 2009;17:881-888. 14. Nebes RD, Pollock BG, Houck PR, et al. Persistence of cognitive impairment in geriatric patients following antidepressant treatment: a randomized, double-blind clinical trial with nortriptyline and paroxetine. J Psychiatr Res. 2003;37:99-108. 15. Reynolds CF 3rd, Butters MA, Lopez O, et al. Maintenance treatment of depression in old age: a randomized, double-blind, placebocontrolled evaluation of the efÔ¨Åcacy and safety of donepezil combined with antidepressant pharmacotherapy. Arch Gen Psychiatry. 2011;68: 51-60. 16. Herrera-Guzman I, Herrera-Abarca JE, Gudayol-Ferre E, et al. Effects of selective serotonin reuptake and dual serotonergic-noradrenergic 746 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. reuptake treatments on attention and executive functions in patients with major depressive disorder. Psychiatry Res. 2010;177:323-329. Gallassi R, Di Sarro R, Morreale A, Amore M. Memory impairment in patients with late-onset major depression: the effect of antidepressant therapy. J Affect Disord. 2006;91:243-250. Rosenberg PB, Mielke MM, Han D, et al. The association of psychotropic medication use with the cognitive, functional, and neuropsychiatric trajectory of Alzheimer‚Äôs disease. Int J Geriatr Psychiatry. 2012;27:1248-1257. Devanand DP, Pelton GH, Marston K, et al. Sertraline treatment of elderly patients with depression and cognitive impairment. Int J Geriatr Psychiatry. 2003;18:123-130. Lopez OL, Wisniewski SR, Becker JT, Boller F, DeKosky ST. Psychiatric medication and abnormal behavior as predictors of progression in probable Alzheimer disease. Arch Neurol. 1999;56:1266-1272. Mulsant BH, Pollock BG, Kirshner M, Shen C, Dodge H, Ganguli M. Serum anticholinergic activity in a community-based sample of older adults: relationship with cognitive performance. Arch Gen Psychiatry. 2003;60:198-203. Health and Retirement Study. Sample Sizes and Response Rates. Available at: http://hrsonline.isr.umich.edu/sitedocs/sampleresponse. pdf. Accessed August 16, 2012. Ofstedal M, Fisher GG, Herzog AR. Documentation of Cognitive Functioning Measures in the Health and Retirement Study. Ann Arbor, MI: University of Michigan; 2005. Herzog AR, Wallace RB. Measures of cognitive functioning in the AHEAD Study. J Gerontol B Psychol Sci Soc Sci. 1997;52B:37-48. Langa KM, Chernew ME, Kabeto MU, et al. National estimates of the quantity and cost of informal caregiving for the elderly with dementia. J Gen Intern Med. 2001;16:770-778. Langa KM, Larson EB, Karlawish JH, et al. Trends in the prevalence and mortality of cognitive impairment in the United States: is there evidence of a compression of cognitive morbidity? Alzheimers Dement. Mar 2008;4:134-144. Langa KM, Plassman BL, Wallace RB, et al. The Aging, Demographics, and Memory Study: study design and methods. Neuroepidemiology. 2005;25:181-191. Crimmins EM, Kim JK, Langa KM, Weir DR. Assessment of cognition using surveys and neuropsychological assessment: the Health and Retirement Study and the Aging, Demographics, and Memory Study. J Gerontol B Psychol Sci Soc Sci. 2011;66(Suppl 1):i162-i171. Radloff L. The CES-D Scale: a self-report depression scale for research in the general population. Appl Psychol Methods. 1977;1:385-401. The American Journal of Medicine, Vol 128, No 7, July 2015 30. StefÔ¨Åck D. Documentation of Affective Functioning Measures in the Health and Retirement Study. Ann Arbor, MI: University of Michigan; 2000. 31. Katz S, Downs TD, Cash HR, Grotz RC. Progress in development of the index of ADL. Gerontologist. 1970;10:20-30. 32. Carnahan RM, Lund BC, Perry PJ, Pollock BG, Culp KR. The anticholinergic drug scale as a measure of drug-related anticholinergic burden: associations with serum anticholinergic activity. J Clin Pharmacol. 2006;46:1481-1486. 33. Fox C, Richardson K, Maidment ID, et al. Anticholinergic medication use and cognitive impairment in the older population: the Medical Research Council Cognitive Function and Ageing study. J Am Geriatr Soc. 2011;59:1477-1483. 34. Campbell NL, Boustani MA, Lane KA, et al. Use of anticholinergics and the risk of cognitive impairment in an African American population. Neurology. 2010;75:152-159. 35. Carriere I, Fourrier-Reglat A, Dartigues JF, et al. Drugs with anticholinergic properties, cognitive decline, and dementia in an elderly general population: the 3-city study. Arch Intern Med. 2009;169: 1317-1324. 36. Ancelin ML, Artero S, Portet F, Dupuy AM, Touchon J, Ritchie K. Nondegenerative mild cognitive impairment in elderly people and use of anticholinergic drugs: longitudinal cohort study. BMJ. 2006;332:455-459. 37. Fox C, Livingston G, Maidment ID, et al. The impact of anticholinergic burden in Alzheimer‚Äôs Dementia-the Laser-AD study. Age Ageing. 2011;40:730-735. 38. Low L-F, Anstey KJ, Sachdev P. Use of medications with anticholinergic properties and cognitive function in a young-old community sample. Int J Geriatr Psychiatry. 2009;24:578-584. 39. Bottiggi KA, Salazar JC, Yu L, et al. Long-term cognitive impact of anticholinergic medications in older adults. Am J Geriatr Psychiatry. 2006;14:980-984. Funding: The Health and Retirement Study is funded by the National Institute on Aging (U01 AG09740). JSS was supported in part by funding from the National Institute on Aging (K01AG33643) and the National Heart, Lung, and Blood Institute (U01HL105268). This research was supported in part by P01AG031098 from the National Institute on Aging and the Department of Veterans Affairs, Health Services Research and Development (CD2 07-206-1 and VA IIR 10-176-3). ConÔ¨Çict of Interest: None. Authorship: All authors had access to the data and played a role in writing this manuscript. ˇ˛Antidiabetic drugs W stephen Waring Abstract over recent years, there has been rapid expansion of different classes of antihyperglycaemic drugs. each possesses a unique pharmacological mechanism of action and, correspondingly, they have diverse toxicological profiles. antidiabetic drugs have the potential to exert severe toxic effects, and patients normally require urgent medical assessment and treatment. Keywords acarbose; antidiabetic; antihyperglycaemic; DPP-iV inhibitor; GLP-1; insulin; metformin; overdose; sulphonylureas; thiazolidinediones Mechanisms of toxicity Insulin Insulin modifies glucose transporter function via its tyrosine kinase-linked receptor. Parenteral administration may cause pro- found hypoglycaemia and activation of neurohormonal counter- regulatory mechanisms. Overdose may be associated with significantly longer duration of action than therapeutic doses. Inhaled insulin has limited bioavailability, and significant hypogly- caemia is less likely. Insulin is rapidly degraded in the gastrointes- tinal tract, and ingestion is not expected to cause systemic effects. Sulphonylureas Sulphonylureas bind to a specific islet cell receptor and stimulate insulin secretion. Acute ingestion may cause significant hypogly- caemia; onset is normally within 8 hours of ingestion.1 Biguanides Metformin suppresses gluconeogenesis from lactate and pyru- vate, and enhances tissue sensitivity to insulin. Typically, met- formin ingestion does not cause hypoglycaemia, but may cause severe lactic acidosis.2 Thiazolidinediones Thiazolidinediones enhance sensitivity to insulin by stimulating peroxisomal proliferator-activated receptor gamma (PPAR-≥). Ingestion may cause mild hypoglycaemia in a small number of W Stephen Waring FRCPE PhD is a Consultant Toxicologist at the Royal Infirmary of Edinburgh, and honorary Senior Lecturer in Clinical Pharmacology at the University of Edinburgh. He has previously worked within the pharmaceutical industry, and research interests include early drug development, assessment of drug safety, and clinical aspects of poisoning. Conflicts of interest: none declared. patients.3 Hepatic injury is a recognized adverse effect of regular treatment, but risk after single acute ingestion is likely to be low. ±-glucosidase inhibitors These inhibit hydrolysis of complex carbohydrates and impair absorption from the small intestine. Gastrointestinal effects are common, including diarrhoea and abdominal pain. Hypoglycaemia is not a recognized feature. Glucagon-like peptide-1 analogues and dipeptidyl peptidase-IV inhibitors Glucagon-like peptide-1 (GLP-1) enhances insulin responsive- ness, and is degraded by dipeptidyl peptidase-IV (DPP-IV). Both GLP-1 analogues and DPP-IV antagonists are relatively new drug classes, and their toxicity is poorly characterized. Given that their effects are insulin-dependent, ingestion alone would not be expected to cause significant hypoglycaemia. Toxicity and acute features Poisoning by sulphonylureas, biguanides and insulin are asso- ciated with moderate-to-severe poisoning in 4.6%, 12.2%, and 14.7% respectively, and fatal poisoning in 0.9%, 6.1% and 3.6% respectively.4 Insulin and sulphonylurea overdose may cause significant hypoglycaemia. Features are due to both hypoglycaemia and activation of neurohormonal mechanisms, and include agitation, altered behaviour, excess sweating, slurred speech, tachycardia, seizures, and coma (Table 1). Metformin ingestion can be associated with abdominal pain, vomiting, and diarrhoea due to drug accumulation in the Poisonous substances Summary of toxic features associated with antidiabetic drug classes Class Insulin  subcutaneous (short and long acting) Sulphonylureas, e.g. gliclazide, glipizide Biguanides, e.g. metformin Thiazolidinediones, e.g. pioglitazone, rosiglitazone ±-glucosidase inhibitors, e.g. acarbose, voglibose GLP-1 analogues, e.g. exenatide, liraglutide DPP-IV inhibitors, e.g. sitagliptin, vildagliptin Risk of hypoglycaemia Profound, may be prolonged for several days Profound, may be prolonged for hours to days negligible Low risk, mild negligible Low risk Low risk Other features Local injection site (parenteral bolus) severe lactic acidosis, gastrointestinal symptoms Possible hepatic dysfunction gastrointestinal tract and, in patients who develop severe lactic acidosis, agitation, confusion, tachypnoea and hypotension. Management Hypoglycaemia should be reversed by administration of intra- venous dextrose or glucose, aiming to achieve a target plasma glucose concentration of e"4 mmol/L (higher target concentra- tions may be needed to avoid symptoms in some patients with poorly controlled diabetes). Prolonged intravenous infusions can be required, and this is associated with the risk of hepatocel- lular glycogen accumulation and reversible liver dysfunction.5 Lactate provides an alternative substrate for brain metabolism, and lessens the risk of neuroglycopenia. Glucagon and octreotide antagonize the effects of insulin, and may reduce the quantity of carbohydrate administration needed to maintain euglycaemia.1,6 Surgical excision should be considered only in cases where very large amounts of insulin have been identified in a single depot and other approaches to management have not proved success- ful. There are insufficient data to establish whether any of these treatments confer a better outcome than systemic administration of dextrose alone. Patients who develop severe lactic acidosis require close observation of haemodynamic status, urinary output, serum electrolytes and acid base balance and, where appropriate, cor- rection of any fluid or electrolyte disturbance. Haemodialysis or filtration may be required for severe acidosis and renal failure in the context of metformin poisoning. Complications Profound hypoglycaemia may lead to persisting neurological deficits, for example cognitive impairment and delayed reaction times.7 Risks are greatest in those exposed to profound hypogly- caemia for a prolonged duration. Therefore, prompt restoration and maintenance of euglycaemia are indicated to minimize the risks of neuropsychiatric sequelae. Anti-HMGCR Autoantibodies in European Patients With Autoimmune Necrotizing Myopathies Inconstant Exposure to Statin Yves Allenbach, MD, PhD, Laurent Drouot, MSc, Aude Rigolet, MD, Jean Luc Charuel, PharmD, Fabienne Jouen, MD, Norma B. Romero, MD, Thierry Maisonobe, MD, Odile Dubourg, MD, Anthony Behin, MD, Pascal Laforet, MD, PhD, Tania Stojkovic, MD, Bruno Eymard, MD, PhD, Nathalie Costedoat-Chalumeau, MD, PhD, Emmanuelle Campana-Salort, MD, Anne Tournadre, MD, Lucile Musset, MD, Brigitte Bader-Meunier, MD, Isabelle Kone-Paut, MD, PhD, Jean Sibilia, MD, PhD, Laurent Servais, MD, PhD, Olivier Fain, MD, Claire Larroche, MD, Elisabeth Diot, MD, PhD, Benjamin Terrier, MD, PhD, Raphael De Paz, MD, Antoine Dossier, MD, Dominique Menard, MD, Chafika Morati, MD, Marielle Roux, MD, Xavier Ferrer, MD, Jeremie Martinet, PharmD, Sophie Besnard, MD, Remi Bellance, MD, Patrice Cacoub, MD, PhD, Laurent Arnaud, MD, PhD, Bernard Grosbois, MD, PhD, Serge Herson, MD, Olivier Boyer, MD, PhD, and Olivier Benveniste, MD, PhD, on behalf of the French Myositis Network Abstract: Necrotizing autoimmune myopathy (NAM) is a group of acquired myopathies characterized by prominent myofiber necrosis with little or no muscle inflammation. Recently, researchers identified autoantibodies (aAb) against 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR) in patients with NAM, especially in statin-exposed patients. Here we report what is to our knowledge the first European cohort of patients with NAM. The serum of 206 patients with suspicion of NAM was tested for detection of anti-HMGCR aAb using an addressable laser bead immunoassay. Forty-five patients were found to be anti-HMGCR positive. Their mean age was 48.9 ¬± 21.9 years and the group was predominantly female (73.3%). Statin exposure was recorded in 44.4% of patients. Almost all patients had a muscular deficit (97.7%), frequently severe (Medical Research Council [MRC] 5 ‚â§3 in 75.5%). Subacute onset (<6 mo) was noted for most of them (64.4%). Nevertheless, 3 patients (6.6%) had a slowly progressive course over more than 10 years. Except for weight loss (20%), no extramuscular sign was observed. The mean CK level was high (6941 ¬± 8802 IU/L) and correlated with muscle strength evaluated by manual muscle testing (r = ‚àí0.37, p = 0.03). Similarly, anti-HMGCR aAb titers were correlated with muscular strength From the AP-HP, H√¥pital Piti√©-Salp√™tri√®re, Department of Internal Medicine 1 and Inflammation-Immunopathology-Biotherapy Department (I2B), East Paris Neuromuscular Diseases Reference Center, Inserm U974, Universit√© Pierre et Marie Curie, Paris 6, Paris (YA, AR, SH, O. Benveniste); Inserm, U905, Immunology Department, Normandie Univ, IRIB and Rouen University Hospital (LD, FJ, JM, O. Boyer); AP-HP, H√¥pital Piti√©-Salp√™tri√®re, Department of Immunochemistry, Universit√© Pierre et Marie Curie, Paris 6, Paris (JLC, LM); AP-HP, H√¥pital Piti√©-Salp√™tri√®re, Institute of Myology, Unite de Morphologie Musculaire, CNRS-UMR7215, Paris (NBR); APHP, H√¥pital Piti√©-Salp√™tri√®re, Department of Neuropathology, Universit√© Pierre et Marie Curie, Paris 6, Paris (TM, OD); AP-HP, H√¥pital Piti√©Salp√™tri√®re, Department of Neurology, East Paris Neuromuscular Diseases Reference Center, Universit√© Pierre et Marie Curie, Paris 6, Paris (AB, PL, TS, BE); AP-HP, H√¥pital Cochin Centre de R√©f√©rence Maladies AutoImmunes et Syst√©miques Rares, Service de M√©decine Interne P√¥le M√©decine, Universit√© Ren√© Descartes Paris V, Paris (NCC, BT); APHM, H√¥pital la Timone, Centre de R√©f√©rence des Maladies Neuromusculaires, Marseille (ECS); Clermont-Ferrand University Hospital, Department of Rheumatology, Clermont-Ferrand (AT); AP-HP, H√¥pital Necker, Department of Paediatric Rheumatology, Paris (BBM); AP-HP, H√¥pital Bic√™tre, Department of Paediatric Rheumatology, Kremlin-Bic√™tre, Paris (IKP); CHU Strasbourg, Department of Rheumatology, Strasbourg (JS); H√¥pital Piti√©Salp√™tri√®re, Institute of Myology, Paris (LS); AP-HP, H√¥pital Jean Verdier, Department of Internal Medicine, La Seine-Saint-Denis (OF); AP-HP, 150 www.md-journal.com (r = ‚àí0.31; p = 0.03) and CK level (r = 0.45; p = 0.01). Mean duration of treatment was 34.1 ¬± 40.8 months, and by the end of the study no patient had been able to stop treatment. This study confirms the observation and description of anti-HMGCR aAb associated with NAM. The majority of patients were statin naive and needed prolonged treatments. Some patients had a dystrophic-like presentation. Anti-HMGR aAb titers correlated with CK levels and muscle strength, suggesting their pathogenic role. (Medicine 2014;93: 150‚Äì157) Abbreviations: aAb = autoantibodies, ALBIA = addressable laser bead immunoassay, AU = arbitrary units, C5b‚Äì9 = membranolytic attack complex, CK = creatine kinase, DMARD = disease-modifying antirheumatic drugs, HMGCR = 3-hydroxy-3-methylglutaryl-coenzyme A reductase, IVIg = intravenous immunoglobulin, MHC = major histocompatibility complex class I antigen up-regulation, MMT = manual muscular testing, MRC = Medical Research Council, NAM = necrotizing autoimmune myopathy, SRP = signal recognition particle. H√¥pital Avicenne, Department of Internal Medicine, Bobigny (CL); CHRU Tours, Department of Internal Medicine, Tours (ED); Fondation A. de Rothschild, Department of Neurology, Paris (RDP); AP-HP, H√¥pital Bichat, Department of Internal Medicine, Paris (AD); CHU Rennes, Department of Neurology, Rennes (DM); Centre Hospitalier de la Region d‚ÄôAnnecy, Department of Internal Medicine, Annecy (CM); H√¥pital Pierre Oudot, Department of Internal Medicine, Bourgouin (MR); H√¥pital du Haut Lev√™que, Department of Neurology, Bordeaux (XF); CHU Rennes, Department of Internal Medicine, Rennes (SB, BG); CHU Fort de France, Department of Neurology, Fort de France (RM); AP-HP, H√¥pital Piti√©-Salp√™tri√®re, Departement HospitaloUniversitaire I2B, UPMC Univ Paris 06, UMR 7211, INSERM, UMRS 959, Department of Internal Medicine 2, Paris (PC, LA); France. Financial support and conflicts of interest: The authors have no funding or conflicts of interest to disclose. Reprints: Yves Allenbach, MD, PhD, AP-HP, H√¥pital Piti√©-Salp√™tri√®re, Department of Internal Medicine 1 and Inflammation-Immunopathology-Biotherapy Department (I2B), East Paris Neuromuscular Diseases Reference Center, Inserm U974, Universit√© Pierre et Marie Curie, Paris 6, 75013, Paris, France (e‚Äêmail: yvesallenbach@gmail.com). Copyright ¬© 2014 by Lippincott Williams & Wilkins ISSN: 0025-7974 DOI: 10.1097/MD.0000000000000028 Medicine ‚Ä¢ Volume 93, Number 3, May 2014 Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine ‚Ä¢ Volume 93, Number 3, May 2014 INTRODUCTION N ecrotizing myopathies are characterized by predominant muscle fiber necrosis and regeneration with little or no inflammation. Among the inflammatory myopathies, this histologic pattern defined a new subgroup called immune-mediated necrotizing myopathy7 or necrotizing autoimmune myopathy (NAM). The immune-mediated nature of these myopathies was first suggested by their response to immunosuppressive treatments4,8,14 and their frequent association with anti-signal recognition particle (anti-SRP) autoantibodies (aAb).8,14 Nevertheless, anti-SRP aAb are observed in only 16% of necrotizing myopathies.5 Remarkably, Mammen and colleagues5,17 recently identified a new specific autoantibody recognizing 200 and 100 kDa proteins, initially in 16 United States patients (61%) with pathologic features of necrotizing myopathy without any known myositis-specific antibody. This United States cohort was then expanded by 29 patients (n = 45) in whom the 100 kDa protein was identified as the target of statins 3-hydroxy-3methylglutaryl-coenzyme A reductase (HMGCR).9 AntiHMGCR-positive (Anti-HMGCR+) patients were in their fifth decade of life9,11 and were very frequently exposed to statin (72.7% of cases).11 They presented proximal and symmetric muscular deficit with high creatine kinase (CK) levels (mean, 9718 ¬± 7383 IU/L) and responded to immunosuppressants.9 Since anti-HMGCR aAb were not observed in a large cohort of statin-exposed patients, including patients with self-limited statin intolerance,10 anti-HMGCR aAb are considered specific for NAM.10 To validate this concept in a different cohort of patients by using an alternative independent testing assay, we report herein a second cohort, equivalent in size, of anti-HMGCR+ patients with their clinico-pathologic features. METHODS Anti-HMGCR Detection and Quantitative Titration For anti-HMGCR aAb detection and quantification of serial titers, we recently developed a new addressable laser bead immunoassay (ALBIA), using the same strategy as that previously used for anti-SRP aAb quantification.2 The technical description and diagnostic value of this assay are detailed elsewhere.5a Briefly, the catalytic domain of recombinant human HMGCR protein fused to a glutathione S transferase tag was obtained from Sigma (Saint Louis, MO) and coupled to Bio-PlexR COOH beads with the Bio-PlexR amine coupling kit according to the manufacturer‚Äôs protocol (Bio-Rad, Hercules, CA); 1250 HMGCR-coated beads were added to 100 ŒºL of patient or control serum for 2 hours under shaking. Biotinylated mouse anti-human IgG-specific secondary Ab (Southern Biotech, Birmingham, AL) was added, and, after washing, beads were further incubated with streptavidin-R-phycoerythrin (Qiagen). Blank (no serum, secondary antibody only), negative controls (anti-HMGCR Ab-negative serum), and positive controls (highly positive human anti-HMGCR Ab serum) were included in every assay. Mean fluorescence intensity (MFI) was determined on a Bio-Plex apparatus using the Bio-Plex Manager Software 4.0 (Bio-Rad Laboratories, Hercules, CA). Anti-HMGCR Ab titers were determined at a 1/D dilution using the following formula: [MFI serum/MFI calibrator] √Ç [titer of calibrator] √Ç D/500. The calibrator is a highly positive human anti-HMGCR Ab serum (the same throughout the study) whose titer was arbitrarily set to 100 arbitrary units (AU/mL). ¬© 2014 Lippincott Williams & Wilkins Anti-HMGCR Autoimmune Necrotizing Myopathies We tested the first 37 sera that were found positive with this assay, and all 37 sera also scored positive for anti-HMGCR aAb using the recently developed Myositis Euroline 7 line immunoassay from Euroimmun (L√ºbeck, Germany). The assay was always negative when testing control sera from healthy blood donors (n = 100) or from patients with different inflammatory/ autoimmune diseases (n = 142), according to established classification criteria: American College of Rheumatology revised criteria for systemic lupus erythematosus13 with anti-dsDNA aAb, American Rheumatism Association criteria for rheumatoid arthritis1 with anti-CCP antibodies and/or rheumatoid factor, revised European criteria for primary Sj√∂gren syndrome16 with anti-SSA and/or anti-SSB aAb, Bohan and Peter criteria3 for dermatomyositis, Troyanov criteria for overlap myositis15 with anti-tRNA-synthetase Ab, and Griggs criteria for inclusionbody myositis.6 Patients and Data Collection We tested first the sera of 23 patients having the diagnosis of NAM based on clinical and histologic criteria. These patients were recruited by clinicians (YA, AR, AB, PL, TS, BE, SH, and OB) from the neuromuscular diseases reference center of East Paris. The positive first results were communicated to the French Myositis Network. Sera of 183 patients (including 8 pediatric patients) selected by clinicians of the network were sent to what was until June 2013 the sole French immunology laboratory able to detect anti-HMGCR. These 183 patients had a suspicion of NAM based on clinical and/or pathologic criteria. All these patients were also selected because of their negativity for myositis-specific aAb anti-Jo-1 and anti-SRP using a commercial kit (dTek or Euroimmun). Furthermore, all patients diagnosed as anti-HMGCR aAb positive were also tested for the presence of the following myositis-specific aAb: anti-PL-7, anti-PL-12, anti-SRP, anti-PMSCl, and anti-Mi2 (using a commercial kit, either dTek or Euroimmun). We controlled negativity for anti-SRP aAb detection using analytical sensitivity and specificity of ALBIA‚ÄìSRP.2 All patients determined to be anti-HMGCR aAb positive by ALBIA-HMGCR from July 2012 to June 2013 were recorded for this study. For all patients, the following parameters were recorded: age, sex, past medical history, statin exposure, characteristics of myopathy including date of diagnosis, clinical manifestation with muscular deficit evaluation (by muscle manual testing [MMT] using the Medical Research Council [MRC] scale on 5 points), CK level, and reports of muscle biopsy. Biopsies of 20 patients were performed following standardized procedures and were analyzed by the same pathologists exclusively specialized in muscle pathology (OD and TM) permitting semiquantitative analyses and standardized immunohistochemistry analyses. In those cases, muscle tissue was frozen and stored at ‚àí80¬∞C. On 8 Œºm-thick cryosections, morphologic analysis (hematin and eosin counter staining) and immunohistochemistry analyses were performed. Features such as the presence of necrosis/regeneration fibers, inflammation, major histocompatibility complex (MHC) class I antigen upregulation, membranolytic attack complex (C5b‚Äì9) deposits and vacuoles were recorded. First-line treatments including corticosteroid and diseasemodifying anti-rheumatic drugs (DMARD) were recorded as well as treatment intensification during the follow-up. Treatment intensification was defined as an introduction or an increase of more than 50% of the dose of corticosteroids and/or the introduction of DMARD. www.md-journal.com Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 151 Medicine ‚Ä¢ Volume 93, Number 3, May 2014 Allenbach et al TABLE 1. Characteristics of 45 Anti-HMGCR+ Patients Characteristic Age, yr; mean Sex ratio (M:F) Statin exposure Muscular involvement Muscular deficit Subacute onset Progressive onset Severe deficit (‚â§3) Myalgia Dysphagia CK level Extraskeletal muscular involvement Weight loss Interstitial lung disease Cardiac insufficiency Arthralgia Raynaud phenomenon 48.9 ¬± 21.9 0.36 44.4% (n = 20) 97.7% (n = 44) 64.4% (n = 29) 33.3% (n = 15) 75.5% (n = 34) 53.3% (n = 24) 26.7% (n = 12) 6941 ¬± 8802 IU/L 20% (n 2.2% (n 2.2% (n 11.1% (n 11.1% (n = 9) = 1) = 1) = 5) = 5) For this medical research where we used identifiable human sera and data, we obtained agreement from the French Ministry of Research (AC 2008-87) for the collection, analysis, storage, and reuse. All samples were obtained for diagnostic purposes. In this situation of retrospective study, patients‚Äô consents were impossible or impractical to obtain. We conducted the research after consideration and approval of the research ethics committee of Piti√©-Salp√™tri√®re Hospital. Statistical Analysis Categorical variables are reported as numbers and/or percentages and were compared using a chi-square or, when appropriate, Fisher exact test. Quantitative variables are reported as mean (¬± standard deviation) and compared using a nonparametric test. For all statistical analyses, p < 0.05 was considered significant. Correlation analysis was performed using the Spearman test. Statistical analyses were conducted using Prism software. RESULTS CK level. Four other young girls were initially diagnosed as having limb girdle muscular dystrophy (LGMD) because of a slowly progressive muscle deficit. Autoimmune diseases were noted in the past medical history of 5 patients (thyroiditis, n = 4, and type I diabetes, n = 1). Statin exposure was present in 20 (44.4%) patients. These patients received mainly atorvastatin (n = 10/20) and rosuvastatin (n = 4/20). Statin-exposed patients were older than statin-naive patients (mean age, 64.4 ¬± 6.8 yr vs. 36.6 ¬± 21.7 yr, respectively; p = 0.001). Muscular Deficit and CK Level in Anti-HMGCR+ Patients The circumstances of discovery of NAM were mainly signs related to muscle weakness (n = 37). Otherwise, 3 patients declared having isolated myalgia, and 4 consulted for increased CK level without any other symptoms. For 1 patient, antiHMGCR presence was suggested by a cytoplasmic fluorescence pattern on antinuclear screening test for the etiology of deep venous thrombosis. During the course of the disease, with a mean follow-up of 49 ¬± 22 months, 53.3% of patients (n = 24/45) suffered from myalgia, and 97.7% of patients (n = 44/45) had a muscle weakness. Subacute onset (<6 mo) was noted for most of them (see Table 1); nevertheless, 33.3% of patients (n = 15/45) had a slow progressive muscular deficit ranging from 10 months to years. Of note, 3 cases had been considered to be muscular dystrophy for more than 10 years prior to the detection of antiHMGCR aAb. In most of the cases (n = 34/45, 75.5%), MMT showed a severe proximal deficit (MRC score ‚â§3/5 for the weakest muscular group), and 5 patients were bedridden at diagnosis (see Table 1). Axial weakness and dysphagia were found in 35.5% (n = 16/45) and 26.7% (n = 12/45) of patients, respectively. No facial deficit was observed. Marked muscle atrophy was noted for 22.2% of patients (n = 10/45), and 2 patients also had scapular winging. All patients had increased CK levels (mean, 6941 ¬± 8802 IU/L). Furthermore, CK level correlated with muscular strength (r = ‚àí0.37, p = 0.03) according to the weakest muscle groups (Figure 1). Except for the onset age, no significant difference was observed in statin-exposed patients versus nonexposed patients Characteristics and Statin Exposure in Anti-HMGCR+ Patients Forty-five anti-HMGCR+ patients, coming from 17 different French hospitals in the myositis network, were included in the study. This represented 21.8% of the total of tested patients (n = 45/206), 26% of the first series of patients (n = 6/23), and 21.3% of the second series of patients (n = 39/183). Most patients were white, except 3 African and 1 Asian patients. Patients had a mean age of 48.9 ¬± 21.9 years at the time of the first signs of the disease. They were predominantly female (n = 33/45; 73%) (Table 1). One patient had onset of the disease at 3 months of pregnancy, and a second at immediate childbirth. Eight pediatric cases (age range, 4-16 yr) were observed. Of note, only 1 patient among the 8 was tested and found to be anti-HMGCR+ before the age of 16 years. For the 7 others, the diagnosis was performed in adulthood. Four girls were initially diagnosed as having an inflammatory myopathy (polymyositis, n = 1; dermatomyositis, n = 2; NAM, n = 1) because of rapid progressive scapular and pelvic deficit with high 152 www.md-journal.com FIGURE 1. Correlation between CK level and muscular deficit in patients with NAM. For each patient the muscular deficit is given as the Medical Research Council score of the weakest muscular group using manual muscular testing (MMT MRC). ¬© 2014 Lippincott Williams & Wilkins Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine ‚Ä¢ Volume 93, Number 3, May 2014 Anti-HMGCR Autoimmune Necrotizing Myopathies concerning percentage of myalgia, severity of muscular deficit, percentage of dysphagia, and mean CK level (data not shown). Histologic Necrotizing Myopathy Pattern All but 2 patients had a muscle biopsy (n = 43) performed for the diagnosis of the myopathy. Morphologic analysis showed for 42 patients (97.6%) a necrotizing myopathy pattern defined by the presence of fibers with necrosis and/or regeneration and no or few inflammatory infiltrates. Results of 20 muscular biopsies performed in our reference center for neuromuscular disorders following standardized procedures and analyzed by the same myo-pathologists are represented in Table 2. Necrosis and/or regenerative phenomena varied from 1 patient to another, ranging from few to intense muscular necrosis (see Table 2 and Figure 2A and B). This pattern was associated with irregularity of the size of fibers in 90% of cases (n = 18/20) due to presence of atrophic fibers with nonspecific repartition, notably in the perimysium (see Table 2 and Figure 2C). Most patients (60%, n = 12/20) did not have muscular infiltrates. Six patients had small muscle infiltrates, and only 2 had mild (but obvious) inflammatory infiltrates (see Figure 2C). C5b9 deposit was frequently observed on necrotic fibers but also at the membrane of few normal fibers (65%, n = 13/20) (see Table 2 and Figure 2D). Deposits of C5b9 decorating a few muscle capillaries were occasionally observed (25%; n = 5/20) but not as dramatically as is typically seen in dermatomyositis. Inflammation, when present, was usually in perivascular areas. Overexpression of MHC class I by fibers was mostly observed on regenerative or necrotic fibers (Figure 2E). Nevertheless, sometimes (n = 8) a few normal fibers, in focal area, had MHC class I overexpression. Only 2 patients (including the 1 with the most obvious inflammatory infiltrates) had diffuse and intense MHC class I overexpression (Figure 2F), as can be observed in polymyositis or inclusion body myositis. Extramuscular Involvement Some patients (20%; n = 9/45) had fatigue associated with weight loss (range, 2‚Äì18 kg). Twenty-one percent of patients (n = 6/28) had increased C-reactive protein serum levels (range, 6‚Äì79 mg/mL). Among the 27 patients who had a computed thoracic scan, none had an interstitial lung disease, except 1 patient in whom an idiopathic interstitial lung disease was diagnosed by a lung biopsy 8 years before the onset of the myopathy. Among the 27 patients who had pulmonary function tests, 7 (25.9%) TABLE 2. Histologic Muscle Analysis* Patient Fiber Necrosis and Disparity of Muscular Regeneration Fiber Size Infiltrates Topography of Infiltrates C5b9 Deposits ‚àí Capillaries and fibers Fibers 1 ++ Yes ‚àí 2 ++ Yes + 3 4 5 6 7 8 + +++ + + ++ No Yes No Yes Yes Yes ‚àí + ‚àí ‚àí ‚àí ‚àí Pericapillary and perinecrotic ‚àí Pericapillary ‚àí ‚àí ‚àí ‚àí 9 +++ Yes + Pericapillary 10 11 12 ++ ++ +++ Yes Yes Yes ‚àí + ++ 13 ++ Yes ++ 14 15 16 17 ++ + ++ +++ Yes Yes Yes Yes ‚àí + ‚àí + ‚àí Pericapillary Endomysial and pericapillary Endomysial and pericapillary ‚àí Pericapillary ‚àí Endomysial 18 19 +++ +++ Yes Yes ‚àí ‚àí ‚àí ‚àí 20 +++ Yes ‚àí ‚àí ‚àí ‚àí Fibers Diffuse Fiber MHCI Endomysial Rimmed Overexpression Fibrosis Vacuoles No No No No No No No No No No No No No Yes No No No No No No No No No No ‚àí ‚àí Capillaries and fibers Capillaries and fibers ‚àí ‚àí Fibers No No No No No Yes Yes No Yes No No Yes Fibers No No No No No No No Yes No No No No No No No Yes No No No Yes No No No No Fibers ‚àí Fibers Capillaries and fibers Fibers Capillaries and fibers Fibers *Table shows the results of 20 patients with muscular biopsy following the same standardized procedures performed in 1 center. Semiquantitative analysis of muscular necrosis and regeneration were performed (+ corresponds to <10 necrosis and/or regenerative fibers, ++ corresponds to 10‚Äì30, and +++ corresponds to >30 per biopsy). C5b9 deposits were considered only on non-necrotic fibers or capillaries. ¬© 2014 Lippincott Williams & Wilkins www.md-journal.com Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 153 Medicine ‚Ä¢ Volume 93, Number 3, May 2014 Allenbach et al FIGURE 2. Histologic and immunohistologic analysis of muscle biopsies from anti-HMGCR+ patients. A, Muscle biopsy from a NAM patient showing rare necrotizing fiber (arrow) with some atrophic fibers (arrowheads) (hematin eosin stain). B, Muscle biopsy from another patient showing a high number of necrotized fibers (arrow) with important numbers of regenerative fibers (arrowhead). C, Muscular infiltrates are sometimes present (arrow) and in low intensity as represented here (arrow) for 1 patient. D, Immunohistologic analysis showing C5b9 deposits decorating some necrotic and non-necrotic fibers (arrows). E, Immunohistologic analysis of MHC class I expression showing a representative case of MHC overexpression on 2 fibers (arrows). F, A few patients may have diffuse and intense MHC overexpression not only on necrotic fibers (arrows) but also on non-necrotic fibers (arrowheads). had a restrictive syndrome with a decrease of forced vital capacity (range, 33%‚Äì42% of predicted volumes). On electrocardiogram, 4 patients (12.9%; n = 4/31) had conduction abnormalities (right branch block, n = 2; left hemi branch block, n = 2). One patient had an auricular fibrillation, and 1 an atrial flutter. For these 2 patients, 1 (aged 72 yr) was diagnosed 3 years before the onset of the myopathy and the other in the context of venous thromboembolism. Among 25 patients who had an echocardiogram, only 1 showed a decrease of ejection fraction (measured at 20% of normal); this patient was known to have an idiopathic dilated cardiomyopathy for more than 20 years before the onset of the skeletal myopathy. Seven patients had Raynaud phenomenon (n = 5/45) or acrosyndrome (n = 2/45), and 5 patients had arthralgia without synovitis. No patient had any association with another connective tissue disease, nor presented other specific aAb. Five patients had cancer during their medical history. The cancers were diagnosed 19 years (breast adenocarcinoma), 4 years (breast adenocarcinoma), 12 months (melanoma), 6 months 154 www.md-journal.com (renal adenocarcinoma), and 2 months (pulmonary adenocarcinoma) after the first signs of the NAM. Treatment and Evolution Thirty-nine of the 45 patients (86.6%) received immunosuppressive drugs and/or immune-modulatory treatment. Six patients (statin-exposed, n = 4) were not treated. One non-statinexposed patient without muscular deficit was not treated, but follow-up time by the end of the study was less than 6 months. Two patients refused the treatment. Three remaining statinexposed patients did not receive any treatment. They slowly improved after statin withdrawal without any other intervention. One had complete remission (normal strength and CK level) 8 months after statin withdrawal, and because she had severe familial hypercholesterolemia, statins were reintroduced without any muscular symptom or CK increase. Another patient improved in strength (with a MRC score at 4 vs. 3 concerning the weakest muscular group) and decreased CK level (500 IU/L vs. ¬© 2014 Lippincott Williams & Wilkins Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine ‚Ä¢ Volume 93, Number 3, May 2014 Anti-HMGCR Autoimmune Necrotizing Myopathies TABLE 3. Treatments Used for NAM Patients Patient Number First-Line Treatment 2 3 4 6 7 8 9 11 12 13 14 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 CT + IVIg + MTX CT + MTX MTX + CT CT + MTX + PE + IVIg CT + MTX CT + MTX CT + MTX + PE + IVIg CT CT + IVIg CT + MTX CT + IVIg + MTX CT + IVIg CT + MTX CT + MTX CT + IVIg + MTX CT + EDX CT + MTX CT CT CT CT CT + MTX CT CT + IVIg + MTX CT CT + IVIg CT IVIg IVIg CT + MTX CT CT CT + MTX + IVIg CT + MTX CT CT CT CT + MTX CT + MTX Number of Treatment Intensification Disease-Modifying Antirheumatic Drugs 0 0 0 1 0 2 0 5 4 4 2 4 1 0 0 0 3 5 6 1 10 1 1 2 2 1 1 0 0 3 3 3 0 1 1 1 1 0 1 none none none CT/MTX/RTX none CT/MMF/TACRO/PE/MTX/IVIg none CT/AZA/MTX/CYC/MMF/IVIg CT/AZA/MTX/RTX/PE/IVIg CT/IVIg/RTX CT/AZA/IVIg/RTX CT/MTX/IVIg/MMF/RTX CT/MTX none none none CT/AZA CT/IVIg/AZA/MTX/RTX CT/IVIg/MTX/AZA/MMF CT/AZA CT/IVIg/AZA/CYC/TACRO/RTX IVIg CT/IVIg/PE/MTX IVIg MTX/PE/IVIg/RTX CT/EDX IVIg none none CT/MTX/IVIg/MMF/RTX/AZA CT/MMF CT/EDX/IVIg CT/MTX/IVIg CT/MTX/IVIg IVIg CT/AZA MTX none CT/MTX Treatment Duration (Months) 3 1 3 3 3 24 18 114 41 42 40 62 36 4 1 3 80 180 90 34 120 18 5 8 11 23 12 7 7 42 89 36 4 3 50 6 84 11 12 Abbreviations: AZA = azathioprine, CT = corticosteroid, CYC = cyclosporine, EDX = cyclophosphamide, IVIg = intravenous immunoglobulin, MMF = mycophenolate mofetil, MTX = methotrexate, PE = plasmapheresis, RTX = rituximab, TACRO = tacrolimus. 5000 IU/L) at month 4. The remaining patient was diagnosed 18 months after statin withdrawal. At this period, muscle strength was 3 for the weakest muscular group and CK level was 500 IU/mL, whereas it was 2500 IU/mL 18 months before. Four months later, the patient‚Äôs strength spontaneously increased from 3 to 4, myalgia disappeared, and CK level decreased to 240 IU/L. For the other statin-exposed patients (n = 16), statin was removed and immunosuppressive drugs were initiated. Of note for 2 patients, statin was reintroduced without significant flare. Mean duration of treatment for 39 treated patients was 34.1 ¬± 40.8 months (range, 1‚Äì180 mo). By the end of the study, no ¬© 2014 Lippincott Williams & Wilkins patient had been able to stop the treatment for a prolonged time over 1 year because of relapse. For the first line of treatment all patients but 2 were treated with corticosteroids, frequently associated with immunosuppressive drugs (methotrexate, n = 20; cyclophosphamide, n = 1) or intravenous immunoglobulin (IVIg, n = 10, Table 3). One bedridden patient was also treated by plasmapheresis as firstline treatment. All patients but 1 who were treated for more than 1 year required intensification by DMARD (see Table 3). Because of flares or insufficient control of the NAM, the mean number of treatment intensifications was 1.8 ¬± 2.1 (range, 1‚Äì10). www.md-journal.com Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 155 Medicine ‚Ä¢ Volume 93, Number 3, May 2014 Allenbach et al FIGURE 3. Correlation between muscular deficit and CK level with anti-HMGCR aAb titer. For each patient, we analyzed the correlation between the anti-HMGCR aAb titer and muscular deficit given as the MRC score of the weakest muscular group using manual muscular testing (MMT MRC) (A) or CK level (B). DMARD were methotrexate (n = 15), azathioprine (n = 10), mycophenolate mofetil (n = 6), cyclosporine (n = 2), tacrolimus (n = 2), cyclophosphamide (n = 2), rituximab (n = 9), plasmapheresis (n = 3), IVIg (n = 17). The mean treatment duration and the number of treatment intensifications was not significantly different between statin-exposed and statin-naive patients (32.2 ¬± 39.4 mo vs. 30.8 ¬± 32.3, respectively; p = 0.83; and 1.9 ¬± 1.8 vs. 1.1 ¬± 1.5, respectively; p = 0.31). By the end of the study only 1 patient had died, due to aspiration pneumonia; this patient was bedridden. Titer of Anti-HMGCR Finally, in an attempt to indirectly investigate a possible physiopathogenic role of anti-HMGCR, we looked for correlation between aAb titer and the muscular parameters: strength and CK level. We observed a significant correlation (r = ‚àí0.31; p = 0.03) between the anti-HMGCR aAb titer and the MRC score of the weakest muscle groups (Figure 3A). Similarly, we observed a significant correlation (r = 0.45; p = 0.01) between the aAb titers and CK level tested the same day (Figure 3B). Of note, no significant difference was observed between HMGCR titer in statin-exposed and statin-naive patients (data not shown). Differentiating statin-exposed and statin-naive patients, we did not observe a significant correlation between CK levels and aAb titers (r = 0.6; p = 0.06 and r = 0.38; p = 0.09, respectively). But a significant correlation between muscle strength and aAb titers was observed in the statin-naive group (r = 0.56; p = 0.01), which we did not observe in the statin-exposed group (r = 0.06; p = 0.6). DISCUSSION In this study describing the second case series of antiHMGCR NAM patients (n = 45), we found that both children and adults may be affected by a severe muscular deficit with an acute or a more slowly progressive onset, without extramuscular involvement in most cases. Most patients required several lines of treatment over a prolonged period. The high CK level correlates with muscular strength, and the titer of antiHMGCR aAb correlates with muscular strength and CK level. We observed a marked female predominance (73%) that the previous study did not observe (57.8%).9 Of note, the onset of disease in 2 patients (5%) was during or just after pregnancy, which is a rare situation in other forms of autoimmune myopathies.11 The 2 newborns did not present any sign of myopathy. 156 www.md-journal.com Eight of the 45 cases involved patients aged younger than 16 years. This observation and the 2 pediatric cases reported by Mammen et al9 suggest that anti-HMGCR NAM should be considered in the differential diagnosis in children with myopathy, as was previously shown for anti-SRP NAM.12 Dermatomyositis is not the sole inflammatory myopathy occurring during childhood. We observed that the onset of the muscular deficit in antiHMGCR+ patients may be slow, and 3 patients had a disease progression for more than 10 years before the diagnosis. Typically, these patients were initially considered to have a limb girdle muscular dystrophy with no molecular diagnosis. Indeed, muscle histologic analysis may show a dystrophic pattern with necrosis/regeneration, associated with irregular size of fibers and endomysial fibrosis. This pitfall was also described in antiSRP NAM.2 It led us and others12 to recommend testing for anti-SRP and anti-HMGCR aAb in patients presenting clinical and pathologic features compatible with limb girdle muscular dystrophy with no molecular diagnosis. Of note, muscular inflammation may also be observed on muscle biopsy. Yet, our cohort of anti-HMGCR+ subjects did not include patients with prominent inflammatory cell infiltrates, since anti-HMGCR aAb were only looked for in patients suspected of having NAM and, presumably, patients with significant degrees of inflammation were not screened for anti-HMGCR antibodies. Contrary to the case of other overlap syndromes associated with myositis-specific antibodies, such as antisynthetases, we did not observe clear extramuscular involvement in antiHMGCR+ patients. For example, no patient had interstitial lung disease. This is in line with the observations of Mammen et al.9 Only 44% of patients were statin exposed in the current study, compared to the 72.7% frequency reported in the study by Mammen et al.17 While we also observed that statin-naive patients were younger, we did not find any other difference between the statin-naive and statin-exposed groups, whereas Mammen et al9 found that mean CK levels were lower in statin-exposed patients. The difference in the prevalence of statin exposure in our study compared to Mammen‚Äôs study may be explained by the fact that we studied a different population. For example, our study is a collaborative work involving 2 pediatric and 2 dystrophic centers. Thus, our cohort is younger on average (aged 44 ¬± 19 vs. 52 ¬± 16 yr) and includes subjects who were initially considered to have an inherited myopathy. Another difference is that Mammen and colleagues9 screened their entire cohort of subjects suffering from ¬© 2014 Lippincott Williams & Wilkins Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine ‚Ä¢ Volume 93, Number 3, May 2014 inflammatory myopathies for anti-HMGCR antibodies, whereas we screened mainly those suspected of having NAM. Three statin-exposed patients improved within months after statin withdrawal without any other intervention. Such cases were not previously described. It may suggest that some muscular statin-intolerant patients are in fact real NAM cases. Mammen et al10 tried to address this point and showed that none of 51 patients with self-limited statin intolerance was antiHMGCR aAb positive. Nevertheless, anti-HMGCR+ NAM diagnosis cannot be ruled out in patients diagnosed as ‚Äústatinintolerant‚Äù especially if a high CK level is observed. This difference may be due to the technology used for the detection of anti-HMGCR aAb and/or patient recruitment, since statinintolerant had a mild increase in CK level (131.44 ¬± 71.07 IU/L) as compared to our 3 patients with CK level above 1000 IU/L. Anti-HMGCR treated patients had long treatment duration, in line with those previously reported.17 At the study‚Äôs end, all patients continued their treatment and almost everyone required DMARD (ranging from 1 to 10) for intensification, suggesting that anti-HMGCR NAM is a severe disease justifying prolonged treatment. The correlation between CK level and muscular strength suggests that CK level monitoring may be a good surrogate biomarker of disease activity. Nevertheless, the correlation coefficient was not high. This is not surprising since other parameters may be involved in muscular deficit, such as uncontrolled disease duration and/or muscular adipose involution. We also confirmed what Mammen and colleagues17 first reported, that anti-HMGCR titers correlated with muscle strength and CK levels. Together, these results argue for a possible pathogenic role of anti-HMGCR in the pathophysiology of this condition, and mimic what we also observed in anti-SRP+ patients.2 To conclude, to our knowledge this is the first study outside the United States to confirm the observation and description of anti-HMGCR+ NAM. We showed a marked female predominance, the existence of pediatric cases, and that disease onset may be insidious, leading to an erroneous diagnosis of muscle dystrophy. The majority of our patients were not exposed to statin. The CK level may be a good biomarker for the follow-up of patients who most often need prolonged treatments because of the risk of disease flare, in a disease where antiHMGCR aAb may have a role in the physiopathology. ACKNOWLEDGMENT The authors are indebted to Andrew Mammen for critical reading of the manuscript and helpful discussions. REFERENCES 1. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 1988;31:315‚Äì324. 2. Benveniste O, Drouot L, Jouen F, Charuel JL, Bloch-Queyrat C, Behin A, Amoura Z, Marie I, Guiguet M, Eymard B, Gilbert D, Tron F, Herson S, Musset L, Boyer O. Correlation of anti-signal recognition particle autoantibody levels with creatine kinase activity in patients with necrotizing myopathy. Arthritis Rheum. 2011;63:1961‚Äì1971. 3. Bohan A, Peter JB. Polymyositis and dermatomyositis (first of two parts). N Engl J Med. 1975;292:344‚Äì347. 4. Bronner IM, Hoogendijk JE, Wintzen AR, van der Meulen MF, 9Linssen WH, Wokke JH, de Visser M. Necrotising myopathy, an ¬© 2014 Lippincott Williams & Wilkins Anti-HMGCR Autoimmune Necrotizing Myopathies unusual presentation of a steroid-responsive myopathy. J Neurol. 2003;250:480‚Äì485. 5. Christopher-Stine L, Casciola-Rosen LA, Hong G, Chung T, Corse AM, Mammen AL. A novel autoantibody recognizing 200-kd and 100-kd proteins is associated with an immune-mediated necrotizing myopathy. Arthritis Rheum. 2010;62:2757‚Äì2766. 5a. Drouot L, Allenbach Y, Jouen F, Charuel JL, Martinet J, Meyer A, Hinschberger O, Bader-Meunier B, Kone-Paut I, Campana-Salort E, Eymard B, Tournadre A, Musset L, Sibilia J, Marie I, Benveniste O, Boyer O. Exploring necrotizing autoimmune myopathies with a novel immunoassay for anti-3-hydroxy-3-methyl-glutaryl-CoA reductase autoantibodies. Arthritis Res Ther. 2014;16:R39. Epub ahead of print. 6. Griggs RC, Askanas V, DiMauro S, Engel A, Karpati G, Mendell JR, Rowland LP. Inclusion body myositis and myopathies. Ann Neurol. 1995;38:705‚Äì713. 7. Hoogendijk JE, Amato AA, Lecky BR, Choy EH, Lundberg IE, Rose MR, Vencovsky J, de Visser M, Hughes RA. 119th ENMC international workshop: trial design in adult idiopathic inflammatory myopathies, with the exception of inclusion body myositis, 10‚Äì12 October 2003, Naarden, The Netherlands. Neuromuscul Disord. 2004;14:337‚Äì345. 8. Kao AH, Lacomis D, Lucas M, Fertig N, Oddis CV. Anti-signal recognition particle autoantibody in patients with and patients without idiopathic inflammatory myopathy. Arthritis Rheum. 2004;50:209‚Äì215. 9. Mammen AL, Chung T, Christopher-Stine L, Rosen P, Rosen A, Doering KR, Casciola-Rosen LA. Autoantibodies against 3-hydroxy-3-methylglutaryl-coenzyme A reductase in patients with statin-associated autoimmune myopathy. Arthritis Rheum. 2011;63:713‚Äì721. 10. Mammen AL, Pak K, Williams EK, Brisson D, Coresh J, Selvin E, Gaudet D. Rarity of anti-3-hydroxy-3-methylglutaryl-coenzyme A reductase antibodies in statin users, including those with self-limited musculoskeletal side effects. Arthritis Care Res (Hoboken). 2012;64:269‚Äì272. 11. Silva CA, Sultan SM, Isenberg DA. Pregnancy outcome in adult-onset idiopathic inflammatory myopathy. Rheumatology (Oxford). 2003;42:1168‚Äì1172. 12. Suzuki S, Satoh T, Sato S, Otomo M, Hirayama Y, Sato H, Kawai M, Ishihara T, Suzuki N, Kuwana M. Clinical utility of anti-signal recognition particle antibody in the differential diagnosis of myopathies. Rheumatology (Oxford). 2008;47:1539‚Äì1542. 13. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, Schaller JG, Talal N, Winchester RJ. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 1982;25:1271‚Äì1277. 14. Targoff IN, Johnson AE, Miller FW. Antibody to signal recognition particle in polymyositis. Arthritis Rheum. 1990;33:1361‚Äì1370. 15. Troyanov Y, Targoff IN, Tremblay JL, Goulet JR, Raymond Y, Senecal JL. Novel classification of idiopathic inflammatory myopathies based on overlap syndrome features and autoantibodies: analysis of 100 French Canadian patients. Medicine (Baltimore). 2005;84:231‚Äì249. 16. Vitali C, Bombardieri S, Jonsson R, Moutsopoulos HM, Alexander EL, Carsons SE, Daniels TE, Fox PC, Fox RI, Kassan SS, Pillemer SR, Talal N, Weisman MH; European Study Group on Classification Criteria for Sjogren‚Äôs Syndrome. Classification criteria for Sjogren‚Äôs syndrome: a revised version of the European criteria proposed by the American-European Consensus Group. Ann Rheum Dis. 2002;61:554‚Äì558. 17. Werner JL, Christopher-Stine L, Ghazarian SR, Pak KS, Kus JE, Daya NR, Lloyd TE, Mammen AL. Antibody levels correlate with creatine kinase levels and strength in anti-3-hydroxy-3-methylglutaryl-coenzyme A reductase-associated autoimmune myopathy. Arthritis Rheum. 2012;64:4087‚Äì4093. www.md-journal.com Copyright ¬© 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 157 ˇ˛Articles Lancet 2010; 37: 664 72 Published Online February 4, 2010 DOI:10.1016/S0140- 6736(09)61962-0 See Comment page 618 Department of Internal Medicine (W A Nienhuis MD, Y Stienstra MD), Department of Epidemiology (J P Schouten MSc), and Infectious Diseases Service and Tuberculosis Unit, Department of Internal Medicine and Department of Pulmonary Diseases and Tuberculosis (Prof T S van der Werf MD), University Medical Centre Groningen, University of Groningen, Netherlands; Agogo Presbyterian Hospital, Agogo,Ghana (W A Thompson MD, K M Abass); Nkawie-Toase Government Hospital, Nkawie, Ghana (P C Awuah MD, W Tuah); Kumasi Centre for Collaborative Research in Tropical Medicine, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana (N Y Awua-Boateng, Prof O Adjei PhD); National Buruli Ulcer Program, Accra, Ghana (E O Ampadu MD); and Department of Infectious Diseases and Tropical Medicine, Ludwig-Maximilians University of Munich, Munich, Germany (V Siegmund PhD, G Bretzel MD) Correspondence to: Prof Tjip S van der Werf, University Medical Centre Groningen, PO Box 30 001, 9700 RB Groningen, Netherlands t.s.van.der.werf@int.umcg.nl Antimicrobial treatment for early, limited Mycobacterium ulcerans infection: a randomised controlled trial Willemien A Nienhuis, Ymkje Stienstra, William A Thompson, Peter C Awuah, K Mohammed Abass, Wilson Tuah, Nana Yaa Awua-Boateng, Edwin O Ampadu, Vera Siegmund, Jan P Schouten, Ohene Adjei, Gisela Bretzel, Tjip S van der Werf Summary Background Surgical debridement was the standard treatment for Mycobacterium ulcerans infection (Buruli ulcer disease) until WHO issued provisional guidelines in 2004 recommending treatment with antimicrobial drugs (streptomycin and rifampicin) in addition to surgery. These recommendations were based on observational studies and a small pilot study with microbiological endpoints. We investigated the efficacy of two regimens of antimicrobial treatment in early-stage M ulcerans infection. Methods In this parallel, open-label, randomised trial undertaken in two sites in Ghana, patients were eligible for enrolment if they were aged 5 years or older and had early (duration <6 months), limited (cross-sectional diameter <10 cm), M ulcerans infection confirmed by dry-reagent-based PCR. Eligible patients were randomly assigned to receive intramuscular streptomycin (15 mg/kg once daily) and oral rifampicin (10 mg/kg once daily) for 8 weeks (8-week streptomycin group; n=76) or streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin (7∑5 mg/kg once daily), both orally, for 4 weeks (4-week streptomycin plus 4-week clarithromycin group; n=75). Randomisation was done by computer-generated minimisation for study site and type of lesion (ulceration or no ulceration). The randomly assigned allocation was sent from a central site by cell-phone text message to the study coordinator. The primary endpoint was lesion healing at 1 year after the start of treatment without lesion recurrence or extensive surgical debridement. Analysis was by intention-to-treat. This trial is registered with ClinicalTrials.gov, number NCT00321178. Findings Four patients were lost to follow-up (8-week streptomycin, one; 4-week streptomycin plus 4-week clarithromycin, three). Since these four participants had healed lesions at their last assessment, they were included in the analysis for the primary endpoint. 73 (96%) participants in the 8-week streptomycin group and 68 (91%) in the 4-week streptomycin plus 4-week clarithromycin group had healed lesions at 1 year (odds ratio 2∑49, 95% CI 0∑66 to infinity; p=0∑16, one-sided Fisher s exact test). No participants had lesion recurrence at 1 year. Three participants had vestibulotoxic events (8-week streptomycin, one; 4-week streptomycin plus 4-week clarithromycin, two). One participant developed an injection abscess and two participants developed an abscess close to the initial lesion, which was incised and drained (all three participants were in the 4-week streptomycin plus 4-week clarithromycin group). Interpretation Antimycobacterial treatment for M ulcerans infection is effective in early, limited disease. 4 weeks of streptomycin and rifampicin followed by 4 weeks of rifampicin and clarithromycin has similar efficacy to 8 weeks of streptomycin and rifampicin; however, the number of injections of streptomycin can be reduced by switching to oral clarithromycin after 4 weeks. Funding European Union (EU FP6 2003-INCO-Dev2-015476) and Buruli Ulcer Groningen Foundation. 664 www.thelancet.com Vol375 February20,2010 Introduction Buruli ulcer is a necrotising infection of subcutaneous tissue caused by Mycobacterium ulcerans.1 The name Buruli ulcer comes from a region near the Nile River delta in Uganda, named Buruli County, where the disease was highly endemic in the 1960s.2 Today, the disease is emerging in west African countries with thousands of cases every year, mainly in children.3,4 A plasmid of M ulcerans encodes the production of mycolactone,4,5 an immunomodulatory macrolide toxin that causes tissue necrosis.6 M ulcerans is acquired near slow-flowing and stagnant water in tropical and subtropical environments. The natural reservoir and mode of transmission of the infection remain largely obscure and might differ between endemic foci around the world.7,8 However, skin injury9 and insect bites10 have been proposed as modes of transmission. M ulcerans infection usually starts as a nodule, papule, plaque, or oedema. When left alone, the lesion breaks open and a typical painless ulcer with undermined edges appears, which can progress to a large necrotic lesion. WHO has defined lesions with a cross-sectional diameter of less than 5 cm as category I, 5 15 cm as category II, and more than 15 cm, lesions on important sites (eye, breast, and genitalia), or multiple lesions as category III. M ulcerans infection can be self-limiting, but scar tissue and contractures in joints leave patients Articles 250 patients screened 70 ineligible 18 lesion too large 16 too young 7 pregnant 2 hearing impairment 7 recurrent disease 20 doubtful diagnosis and PCR and AFB negative 26 with clinical disease but without PCR confirmation* 154 enrolled and started on treatment 151 randomised 143 with infection confirmed by PCR 5 with infection confirmed by other method! 3 without confirmation of infection 3 withdrew/lost before randomisation 1 accidentally operated 1 died 1 non-compliant 75 assigned to streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin for 4 weeks 1 died (in week 16) ß 2 migrated (in week 32 and week 36)ß 75 assessed for primary endpoint 76 assigned to streptomycin and rifampicin for 8 weeks 1 withdrew (in week 6)ß 76 assessed for primary endpoint with functional limitations and can result in social stigma.11,12 The diagnosis can be made clinically but culture is the gold standard. However, this method is difficult and has low sensitivity.1,3,4,13 Since the develop- ment of PCR targeting insertion sequence 2404 (IS2404) a repetitive oligonucleotide unit with more than 200 copies in the genome of M ulcerans14 diagnostic confirmation has improved substantially.13,15,16 Buruli ulcer is one of 19 neglected tropical diseases addressed by WHO in its Global plan to combat neglected tropical diseases 2008 2015.17 In this plan, the organisation describes Buruli ulcer as a disease for which there are no cost-effective control methods. Since the disease s first description in 1948,18 different treatments have been investigated. Extensive surgical debridement, with or without subsequent skin grafting, is standard treatment. However, surgery cannot completely remove all bacilli19 and recurrence is common, with reported rates varying between 6% and 47%.20 22 Although larger excisions might be more effective, they can increase chances of residual functional limitations. In the first of two randomised controlled trials for M ulcerans infection, clofazimine did not show a significant benefit compared with placebo.23 In individuals with small (<5 cm), non-ulcerated lesions, recurrence-free healing without surgery was reported in five of eight participants who were treated with clofazimine compared with five of 17 who were treated with placebo. In ten patients with larger and ulcerated lesions, all except one (in the placebo group) needed surgery. A second study compared the effect of dapsone plus rifampicin with placebo. Of 41 randomised patients, 30 completed the 2-month trial. Rate of healing did not differ between groups. Uneven baseline characteristics might partly explain why patients assigned to active treatment had a larger reduction in lesion size than did patients assigned to placebo.24 Many antimycobacterial agents show activity against M ulcerans in vitro, and experiments in animals, such as the mouse footpad model, show that streptomycin in combination with rifampicin is highly bactericidal. In a pilot study sponsored by WHO, 31 patients clinically diagnosed with pre-ulcerative M ulcerans infection were treatedwithstreptomycinandrifampicinfor0,2,4,8,or 12 weeks.25 All lesions were excised; M ulcerans infection was confirmed by PCR in 21 cases. In ten patients who were treated for 2 weeks or less, viable bacilli could be isolated from excised tissues, whereas M ulcerans could not be cultured from tissue taken from 11 patients who were treated for 4 weeks or longer. Lesions either reduced or stabilised in size in all patients.25 On the basis of these findings, preliminary guidelines were issued by WHO recommending streptomycin in combination with rifampicin as standard treatment for M ulcerans infection,26 with or without additional surgical debridement or skin grafting.27 When our study was designed, clarithromycin was believed to have only bacteriostatic activity in vivo.3,28 We assessed the efficacy of antibiotic therapy with oral Figure 1: Trial profile AFB=acid-fast bacilli.*Patients not enrolled but given 8 weeks of treatment with streptomycin and rifampicin. Participant died of cause unrelated to M ulcerans infection. ! See text for details. ßHealed at time of last assessment, included in the final analysis. rifampicin and intramuscular streptomycin given for 8 weeks for treatment of early M ulcerans infection in patients from Ghana. This regimen was compared with rifampicin and streptomycin given for 4 weeks, followed by an oral combination of clarithromycin and rifampicin for 4 weeks. Our aim was to identify an effective alternative treatment to extensive surgical debridement, and to explore possibilities to keep the use of injectable antimicrobial treatment to a minimum. Methods Participants The study design was partly based on discussions within a WHO expert group on Buruli ulcer that took place between 2001, and 2003. Between April, 2006, and January, 2008, patients were recruited at two sites (Nkawie- Toase Government Hospital, Nkawie, and Agogo Presbyterian Hospital, Agogo) in Ghana. Patients clinically diagnosed with M ulcerans disease were recruited by active case finding. Patients were eligible for enrolment www.thelancet.com Vol375 February20,2010 665 Articles Sex (male) Body-mass index (kg/m2) Tribe Akan Other Lesion surface area (cm2) Type of lesion No ulceration Ulceration HIV infection 8-week streptomycin group (n=76) 27 (36%) 16∑6 (14∑9 19∑4) 20 (26%) 56 (74%) 29 (9 55) 49 (64%) 27 (36%) 0 4-week streptomycin plus 4-week clarithromycin group (n=75) 19 (25%) 17∑2 (15∑4 19∑4) 18 (24%) 57 (76%) 26 (10 46) 43 (57%) 32 (43%) 3 (4%) Age (years) 12 (8 18) 12 (9 22) Study site Agogo 54 (71%) 53 (71%) Nkawie 22 (29%) 22 (29%) Duration of disease (weeks) 4 (2 6) 3 (2 4) Category of lesion I 29 (38%) 29 (39%) II or III 47 (62%) 46 (61%) Lesion distribution (side of body) Left 27 (36%) 28 (37%) Right 49 (64%) 47 (63%) Data are n (%) or median (IQR). Patients in the 8-week streptomycin group were assigned to receive intramuscular streptomycin and oral rifampicin for 8 weeks. Patients in the 4-week streptomycin plus 4-week clarithromycin group were assigned to receive streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin, both orally, for 4 weeks. Table 1: Patient baseline characteristics 666 www.thelancet.com Vol375 February20,2010 if they were aged 5 years or older, had a reported disease duration of less than 6 months, and had lesions with a cross-sectional diameter (indurated area) of 10 cm or less. M ulcerans infection was confirmed by IS2404 dry-reagent- based PCR.29 Exclusion criteria were pregnancy, drug in- tolerance, and renal, hepatic, and acoustic impairment. The protocol and consent forms were approved by the Committee on Human Research, Publication, and Ethics of the School of Medical Science, Kwame Nkrumah University of Science and Technology, Kumasi, and the Komfo Anokye Teaching Hospital, Kumasi (CHRPE/07/01/05), and by the Ethical Review Committee of Ghana Health Services (GHS-ERC-01/01/06). The Medical Ethics Review Committee of the University Medical Centre Groningen, Netherlands, reviewed the protocol before ethics clearance in Ghana. Written and verbalinformedconsentwasobtainedfromallparticipants aged 12 years or older, and from parents, carers, or legal representatives of participants aged 18 years or younger. Procedures After participants had given informed consent, we obtained demographic and clinical information and took blood samples. We undertook pregnancy tests in female participants aged 10 years or older, and hearing tests in all participants (AS208 portable equipment; Inter- acoustics, Assens, Denmark) to obtain baseline audio- metric data. HIV antibody testing was done with cold-stored sera after completion of the study. Lesions were photographed and traced onto acetate sheets. Three 3 mm punch biopsy samples were taken under local anaesthesia; two swabs of ulcerated lesions were also taken. All samples were transported to the Kumasi Centre for Collaborative Research in Tropical Medicine laboratory in Kumasi, Ghana, for IS2404 dry- reagent-based PCR and Ziehl-Neelsen staining to detect acid-fast bacilli; mycobacterial culture was done on Lowenstein-Jensen slopes at 32oC.13 One punch biopsy was reserved for histopathological examination. Participants started streptomycin (15 mg/kg once daily intramuscularly) and rifampicin (10 mg/kg once daily orally) after the diagnostic procedures. After assess- ments and start of treatment at the hospital, most parti- cipants were treated as outpatients. Once a week, participants were given study drugs to take to the nearest health facility to receive directly observed treatment (DOT) for the subsequent days, with daily wound care. Only participants that had extensive oedema or lesions at difficult sites (joints, eye, or genitalia), or lesions with suspected secondary infection were admitted to hospital; participants who could not receive DOT or wound care at home were also admitted to hospital. DOT was recorded on forms by the health-care worker or helper who was observing the treatment. Participants were followed up at weekly intervals during the first 8 weeks. At these visits, clinical assessments and digital photographs were taken, DOT forms were checked, and participants were invited to report any adverse events. Once every 2 weeks, the size of the lesion was traced onto an acetate sheet and blood cell counts were taken; we also undertook liver and kidney function tests and hearing tests in all participants, and pregnancy tests in female participants aged 10 years or older. Randomisation and masking Before the end of week 4, participants with M ulcerans infection confirmed by PCR were randomly assigned to receive streptomycin intramuscularly and rifampicin orally for 4 more weeks (8-week streptomycin group) or rifampicin and clarithromycin (7∑5 mg/kg once daily), both orally, for another 4 weeks (4-week streptomycin plus 4-week clarithromycin group). Randomisation was done with minimisation for study site and type of lesion (ulceration or no ulceration). The study coordinator (WAN) forwarded the information of every enrolled participant by cell-phone text messaging to a statistician (JPS) at the Department of Epidemiology, University Medical Centre Groningen, Netherlands. There, a computer-generated randomisation program was used, and the randomly assigned allocation was then sent by text message to the study coordinator. Individuals who were clinically diagnosed with M ulcerans disease but Articles Study Treatment Sex Age Category of Stage Size of lesion Additional Timepoint Additional diagnostics Diagnostic results site group (years) lesion (mm) information (weeks*) (timepoint, weeks*) Participants with treatment failure recorded before week 52 1 Agogo 8-week streptomycin Male 6 II Ulcer 120◊98 Extensive debridement 6 ∑∑ ∑∑ 2 Agogo 8-week streptomycin Male 10 III Ulcer and oedema, critical site 126◊79 Extensive debridement 20 ∑∑ ∑∑ 3 Agogo 4-week streptomycin plus 4-week clarithromycin Female 12 II Plaque 73◊60 Lesion progression, extensive debridement 8 Surgical resected tissue PCR negative, ZN negative, culture positive 4 Nkawie 4-week streptomycin plus 4-week clarithromycin Female 11 I Nodule 30◊24 Ulceration of lesion; 4 additional weeks of streptomycin and rifampicin 12 Punch biopsy and swab PCR negative, ZN negative, culture negative 5 Nkawie 4-week streptomycin plus 4-week clarithromycin Female 5 I Plaque 48◊47 New ulceration after initial closure 34 Swab and surgical resected tissue PCR negative, ZN negative, culture negative Participants who were not healed at time of primary endpoint (week 52) 6 Agogo 8-week streptomycin Male 22 II (with oedema leading to III) Ulcer 112◊80 ∑∑ ∑∑ ∑∑ ∑∑ 7 Agogo 4-week streptomycin plus 4-week clarithromycin Female 20 III Ulcer and multiple nodules ∑∑ Ulcer healed; multiple nodules with ulceration before healing ∑∑ Swab (32), swab (52) 32 weeks: PCR positive, ZN negative, culture postive; 52 weeks: PCR positive, ZN positive, culture negative 8 Nkawie 4-week streptomycin plus 4-week clarithromycin Male 12 II Ulcer 95◊95 Inadequate wound care ∑∑ Swab (29), swab (72), surgically resected tissue (80) 29 weeks: PCR negative, ZN negative, culture negative; 72 weeks: PCR negative, ZN negative, culture positive; 80 weeks: PCR negative, ZN negative, culture negative 9 Nkawie 4-week streptomycin plus 4-week clarithromycin Female 7 II Ulcer 122◊100 ∑∑ ∑∑ ∑∑ ∑∑ 10 Agogo 4-week streptomycin plus 4-week clarithromycin Female 27 III Plaque and two small ulcers 113◊86 HIV positive ∑∑ ∑∑ ∑∑ Treatment failure was recorded if a participant s lesion had not healed by week 52, lesion recurrence occurred within 1 year, or lesion size increased to 150% or more at any timepoint compared with baseline with surgical debridement undertaken as deemed necessary by the attending doctor in the hospital. *Weeks after start of treatment. Results for insertion sequence 2404 dry-reagent-based PCR, Ziehl-Neelsen (ZN) staining to identify acid-fast bacilli, and M ulcerans culture. Patients in the 8-week streptomycin group were assigned to receive intramuscular streptomycin and oral rifampicin for 8 weeks. Patients in the 4-week streptomycin plus 4-week clarithromycin group were assigned to receive streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin, both orally, for 4 weeks. Table 2: Characteristics of ten participants with treatment failure who did not have confirmation by PCR continued treatment with streptomycin plus rifampicin and were not randomised; these individuals were followed up and analysed separately. This was an open-label trial. Follow-up and study outcomes After 8 weeks of antimicrobial treatment, missed doses were not supplemented. Participants were followed up at week 10 and week 12 after start of treatment, and then monthly to week 36, and bimonthly to week 52. Study visits included clinical assessment with reporting of adverse effects, measurement of lesion size (if not healed) by tracing onto an acetate sheet, and photography of the lesion. Participants travel costs werereimbursedandsmallmonthlyincentives(sugar, condensed milk, and cocoa powder) were offered for time spent in the study. Treatment failure was recorded if a participant s lesion had not healed by week 52, lesion recurrence occurred within 1 year, or lesion size increased to 150% or more at any timepoint compared with baseline with surgical debridement undertaken as deemed necessary by the attending doctor in the hospital. Neither the investigators who took measurements of the lesions, nor the attending doctor in the hospital making the final decision for extensive surgical debridement were masked to treatment assignment. Removal of necrosis and slough is part of normal wound care and skin grafting speeds up healing www.thelancet.com Vol375 February20,2010 667 Articles 8-week streptomycin group 4-week streptomycin plus 4-week clarithromycin group 95% CI 0∑0067 0∑10 0∑046 0∑19 0∑084 0∑25 0∑16 0∑36 0∑34 0∑57 0∑51 0∑73 0∑78 0∑93 0∑88 0∑99 Week 1 Week 4 Week 6 Week 8 Weeks 10 12 Weeks 16 20 Weeks 24 28 Weeks 32 36 Weeks 44 52 95% CI 0∑0019 0∑090 0∑028 0∑15 0∑036 0∑17 0∑054 0∑20 0∑094 0∑26 0∑28 0∑49 0∑52 0∑74 35 0∑82 0∑96 21 0∑94 1∑00 7 Total Healed (n) (n) 76 1 72 1 70 0 69 2 66 3 59 12 37 10 18 11 3 2 Treatment failure (n) 0 0 1 0 0 1 0 0 1 Cumulative proportion healed 0∑013 0∑066 0∑076 0∑11 0∑16 0∑37 0∑63 0∑90 0∑99 Total Healed Treatment Cumulative (n) (n) failure (n) proportion healed 75 0 0 74 1 0 0∑027 72 4 0 0∑093 66 2 1 0∑15 60 5 0 0∑24 50 10 0 0∑45 Week 2 75 3 0 0∑053 0∑020 0∑13 75 1 0 0∑013 0∑019 0∑091 Week 5 71 1 0 0∑076 0∑036 0∑17 73 1 0 0∑040 0∑013 0∑12 Week 7 70 0 0 0∑076 0∑036 0∑17 68 2 0 0∑12 0∑064 0∑22 Weeks 8 10 67 1 0 0∑12 0∑064 0∑22 63 2 1 0∑17 0∑11 0∑28 Weeks 12 16 63 4 0 0∑21 0∑14 0∑32 55 5 0 0∑32 0∑22 0∑43 Weeks 20 24 46 9 0 0∑50 0∑39 0∑61 40 5 0 0∑52 0∑41 0∑64 7 0 0∑61 11 1 0∑87 3 4 0∑96 Weeks 28 32 27 9 0 0∑76 0∑65 0∑85 28 7 0 0∑71 0∑60 0∑81 Weeks 36 44 7 4 0 0∑96 0∑90 0∑99 9 2 0 0∑90 0∑81 0∑95 Patients in the 8-week streptomycin group were assigned to receive intramuscular streptomycin and oral rifampicin for 8 weeks. Patients in the 4-week streptomycin plus 4-week clarithromycin group were assigned to receive streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin, both orally, for 4 weeks. Table 3: Actuarial life table for cumulative proportion of healing for both treatment groups, by time interval 668 www.thelancet.com Vol375 February20,2010 but does not affect bacterial load. These interventions were therefore not regarded as evidence of treatment failure. The primary clinical endpoint was lesion healing (complete re-epithelialisation) at 1 year after the start of treatment without recurrence or extensive surgical debridement. Secondary outcomes were time to wound healing and time to complete wound coverage by a crust. Daily sterile dressings were only applied at the health facility if lesions were open and discharging. Before final healing occurs, lesions might turn dry with a crust. At this stage, participants could cover the lesions for protection at home, without visiting the health facility to receive wound care and sterile dressings. Since participants reported this stage of wound healing as beneficial, we also measured time to complete wound coverage by a crust without complete re-epithelialisation as a secondary endpoint. The safety outcome measure was occurrence of adverse events. Statistical analysis When the study was designed, there was no information available about healing rates for the proposed regimens; therefore, we assumed a healing rate of 80% in the 8-week streptomycin group. We calculated that a sample size of 148 randomised and fully assessable participants (74 in each group) would be needed to detect a difference in healing rate of 20% or more (<60% in the 4-week streptomycin plus 4-week clarithromycin group) with a one-sided alpha of 0∑05 and a power of 80%. We calculated an odds ratio for the primary clinical endpoint by use of Fisher s exact test. Because secondary outcome data were interval-censored, we analysed the cumulative incidence of healing by use of actuarial life table analysis and weighted log-rank tests for interval- censored data, in particular the group proportional hazards model30 and a generalised Wilcoxon-Mann- Whitney test,31 which emphasises early events. We calculated the exact permutation p value for the scores of the group proportional hazards model and Wilcoxon- Mann-Whitney tests and the non-parametric maximum likelihood estimate of the survival distribution function.32 Other secondary outcome measures were assessed by actuarial life table analysis. All analyses were by intention to treat. Statistical analysis was done with SPSS version 16.0, R version 2.9.2, and Stata version 10.1. An independent data safety monitoring board reviewed the data for safety purposes after inclusion of 57 and 115 participants. Interim reports were discussed at the annual WHO meeting on Buruli ulcer in Geneva in 2007 and 2008, and presented at the 2008 combined ICAAC/IDSA Annual Meeting in Washington, DC, USA.33 After the trial had been completed, two independent wound experts from University Medical Centre Groningen, who were masked to treatment assignment, assessed the primary study endpoint (healing at 1 year) using the digital photographs of the lesions taken during the trial. This trial is registered with ClinicalTrials.gov, number NCT00321178. Role of the funding source The sponsors of the study had no role in study design, data collection, data analysis, data interpretation, writing Articles 1∑0 0∑8 0∑6 0∑4 0∑2 0 Number at risk 8-week streptomycin group 4-week streptomycin plus 4-week clarithromycin group Group proportional hazards model p=0∑26 (99% CI 0∑22 0∑29) Generalised Wilcoxon-Mann-Whitney test p=0∑60 (99% CI 0∑56 0∑64) 0 4 8 12 16 20 24 28 32 36 40 44 48 52 Weeks 8-weekstreptomycingroup 76 72 71 69 67 66 63 59 46 37 27 18 7 3 4-weekstreptomycinplus 75 74 73 68 63 60 55 50 40 35 28 21 9 7 4-week clarithromycin group Figure 2: Non-parametric maximum likelihood estimates for time to healing The non-parametric maximum likelihood estimates for each treatment group are plotted with shaded rectangles denoting the indeterminate rises in the proportion healed during each time interval. Linear interpolation lines of healing within these indeterminate regions are also shown. Patients in the 8-week streptomycin group were assigned to intramuscular streptomycin and oral rifampicin for 8 weeks. Patients in the 4-week streptomycin plus 4-week clarithromycin group were assigned to receive streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin, both orally, for 4 weeks. ofthereport,orindecisionsaboutsubmissionofresults for publication. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. Results Figure 1 shows the trial profile. 180 eligible patients started treatment. 26 patients with suspected but unconfirmed M ulcerans infection received streptomycin and rifampicin for 8 weeks. Of 151 participants who were enrolled and randomised, eight had a clinical diagnosis without confirmation of M ulcerans infection by PCR. Five of these eight participants had infection later confirmed by one or more diagnostic tests (Ziehl-Neelsen staining, two; culture, one; histopathology, two). Three randomised participants did not have diagnosis confirmed by any test. Table 1 shows baseline characteristics of study participants. Lesions were more frequently seen on the right side of the body (64%) than on the left side (36%; p<0∑0001). Three (2%) participants were HIV positive; these individuals had initial lesions and clinical presentations that were indiscernible from those of HIV-negative participants. One participant in the 8-week streptomycin group withdrew from the study at week 6. In the 4-week streptomycin plus 4-week clarithromycin group, two participants moved out of the study area and were lost to follow-up (week 32 and 36) and one participant, who later tested positive for HIV infection, died in week 16 of urosepsis. Since these four participants had healed lesions at their last assessment, they were included in the analysis for the primary clinical endpoint and in the analyses for time to healing. Compliance to study treatment was assessed by use of DOT forms, signed by health personnel at the health facilities. Adherence to treatment protocol was 98% in the 8-week streptomycin group and 99% in the 4-week streptomycin plus 4-week clarithromycin group. Treatment failure was recorded in ten participants, three in the 8-week streptomycin group and seven in the 4-week streptomycin plus 4-week clarithromycin group. Table 2 shows the characteristics of these individuals. Five participants were not healed at week 52, all of whom had a substantial decrease in lesion size. One participant had several lesions, and four had large lesions at the start of treatment (one of whom had HIV infection). Of the five participants with treatment failure before week 52, two had large lesions, one had a pre-ulcerative lesion that ulcerated later, one had a progressive lesion, and one had a lesion that www.thelancet.com Vol375 February20,2010 669 Cumulative proportion of participants with healed lesions Articles Patient 1 Patient 3 Patient 5 Study Sex site Agogo Female Nkawie Female Nkawie Male Age Category Stage (years) of lesion 12 II Plaque 5 II Ulcer 12 II Ulcer Indication for culture Lesion progression Pus collection No complete healing at week 52 Timepoint Additional information (weeks)* 11 Extensive surgical debridement 18 Lesion healed without further intervention 72 Inadequate wound care; surgical debridement Diagnostic specimen Surgically resected tissue Swab Swab Patient 2 Patient 4 Nkawie Agogo Female Female 12 20 II III Oedema Ulcer and multiple nodules Development of a second lesion Confirmation of nodule This analysis was not specified by the protocol. All five participants were in the 4-week streptomycin plus 4-week clarithromycin group (streptomycin and rifampicin for 4 weeks followed by rifampicin and clarithromycin for 4 weeks). *Time of tissue specimen collection (weeks after start of treatment). Participant with treatment failure. Table 4: M ulcerans isolated by culture of tissue specimens after 8-week treatment period 14 32 Lesion healed without further intervention Ulcer healed; multiple nodules ulcerated before healing Swab Swab 670 www.thelancet.com Vol375 February20,2010 almost healed, but opened up again. No participants with healed lesions had a recurrence at week 52. 73 (96%) participants in the 8-week streptomycin group and 68 (91%) participants in the 4-week streptomycin plus 4-week clarithromycin group had healed lesions at week 52 (odds ratio for failure in healing for 4-week streptomycin plus 4-week clarithromycin vs 8-week streptomycin 2∑49, 95% CI 0∑66 to infinity, p=0∑16). We obtained consistent findings when the four participants who were not followed up to week 52 were excluded from the analysis and when two wound experts masked to treatment assignment assessed the primary endpoint by use of photographs available at the different timepoints (data not shown). Table 3 shows the actuarial life table for cumulative proportion of healing. The estimated cumulative proportion of patients healed at week 52 was 0∑99 (95% CI 0∑94 1∑00) in the 8-week streptomycin group and 0∑96 (95% CI 0∑88 0∑99) in the 4-week streptomycin plus 4-week clarithromycin group; a difference of 0∑034 (95% CI  0∑024 to 0∑091) between groups. Figure 2 shows the non-parametric maximum likelihood estimates for healing in the intention-to-treat population. Neither the group proportional hazards model (p=0∑26; 99% CI 0∑22 0∑29) nor the generalised Wilcoxon-Mann-Whitney test (p=0∑60; 99% CI 0∑56 0∑64) showed a significant difference in time to healing between groups. The group proportional hazards model suggested a shorter time to healing in the 8-week streptomycin group whereas the Wilcoxon-Mann-Whitney test suggested that time to healing was shorter in the 4-week streptomycin plus 4-week clarithromycin group. Adjustment for study site and type of lesion (ulceration or no ulceration) did not affect the results (data not shown). Five participants received skin grafts, four in the 8-week streptomycin group (at week 16, 24, 24, and 28), and one in the 4-week streptomycin plus 4-week clarithromycin group (at week 20). Time to healing of category I lesions (median 18 weeks, 95% CI 14 22) was significantly shorter than that for category II and III lesions (30 weeks, 95% CI 26 34, p=0∑002; data pooled for the two treatment groups; five participants with skin grafts not included). Time to complete wound coverage by a crust was also significantly shorter for category I lesions than category II and III lesions (14 weeks, 95% CI 11 18, vs 22 weeks, 95% CI 22 26; p=0∑002). Three participants had vestibulotoxic events, one in the 8-week streptomycin group (aged 49 years, starting after 7 weeks of treatment) and two in the 4-week streptomycin plus 4-week clarithromycin group (aged 24 years and 38 years, starting after 4 weeks and 3 weeks of treatment, respectively). Analysis of digital photographs showed that three participants had mild to moderate functional limitations at the end of the study: one had a contracture with substantial decrease in range of movement of the thumb and index finger (4-week streptomycin plus 4-week clarithromycin group); two had ulcers on the back of the hand and wrist that resulted inclaw-hands(oneineachgroup).Noliverorkidney function test abnormalities or audiological deterioration occurred that necessitated termination of streptomycin treatment. One participant developed an injection abscess (4-week streptomycin plus 4-week clarithro- mycin group) and two participants (both in the 4-week streptomycin plus 4-week clarithromycin group) developed an abscess close to the initial lesion which was incised and drained. One participant in the 8-week streptomycin group and two participants in the 4-week streptomycin plus 4-week clarithromycin group reported abdominal discomfort. Some participants had additional diagnostic tests not specified in the protocol. Table 4 shows the characteristics of the five participants in whom M ulcerans was isolated by culture after treatment; all were in the 4-week streptomycin plus 4-week clarithromycin group. Three of these five participants had treatment failure: in two, surgical debridement was done; in the third, multiple nodules ulcerated successively over 52 weeks before final healing. Two participants had lesion healing without further intervention within the study period. Discussion Our study has shown that early, limited M ulcerans infection can be safely and effectively managed by antimicrobial treatment alone, without surgical debridement. The drug regimen proposed by WHO, Articles consisting of 8 weeks of streptomycin and rifampicin, seemed effective and was not associated with deterioration requiring subsequent surgical debridement. Treatment with oral clarithromycin plus rifampicin during the second 4-week period resulted in similar outcomes to continuation of treatment with streptomycin and rifampicin. Our findings are important for patients with M ulcerans infection who live in remote, resource- poor areas in west Africa, where people often need to walk for several hours to reach health-care facilities, skilled personnel are scarce, and patients tend to refrain from treatment because of fear of surgery. Our results also support the use of antimicrobial treatment in individuals who are unable to receive streptomycin eg, pregnant women or those who cannot tolerate aminoglycosides. With few reported side-effects, the treatment regimens used in this trial seemed well tolerated, although vestibulotoxicity remains a concern. The rate of lesion recurrence in our study at 52 weeks was lower than that reported in retrospective studies assessing the effect of surgery, in which rates of between 6% and 47% were reported.20 22 Time to healing was a median of 18 weeks for category I lesions and 30 weeks for larger lesions. The length of this healing period might have obscured the potential of antimicrobial treatment in earlier studies that either looked at healing after 2 months,24 or assigned participants to surgery when early healing was not seen during follow- up.23 HIV was not an important confounder; most case- control studies from west Africa report a low incidence of HIV in patients with M ulcerans infection.34,35 One strength of this study is the large proportion of participants (147 of 151) who were followed up to week 52. Second, most participants (148 of 151) had laboratory- confirmed M ulcerans infection, which contrasts with previous trials that were partly undertaken before PCR- based diagnostic confirmation tests were available.23,24 Finally, the sample size in our study was substantially larger than that in earlier studies. A potential weakness of our study is the open-label design. However, masking would have substantially increased costs, and a trial in which children can be assigned to placebo injections is not justified for safety reasons. Moreover, although only one injection abscess was recorded, intramuscular injections in rural Africa are not the preferred option. Another limitation of our study is that no formal external monitoring was done. However, limited auditing was organised. Additionally, consistent results were obtained when two wound experts who were masked to treatment assignment reviewed all digital photographs available at the different timepoints. We therefore believe that the study was robust. One concern is that healing took a fairly long time. Additionally, we could not address the issue of prevention of disabilities in a formal way, although our analysis of digital photographs combined with clinical assessment showed that only three participants had mild to moderate functional limitations at week 52 (all three involving hand function). Contractures and functional limitations are common in ulcers that are close to joints.11 Future studies should assess prevention of disabilities, include all categories of lesions, and investigate oral drug regimens. Thus, antimicrobial treatment is highly effective for treatment of early, limited M ulcerans infection, and the number of intramuscular injections of streptomycin can be reduced without compromising efficacy. Contributors TSvdW and YS designed and supervised the study. WAN coordinated the study. WAN, WAT, PCA, and EOA were responsible for patient screening and enrolment. KMA, WT, and WAN provided patient care and requested informed consent from participants, participants parents, or legal representatives, and collected the clinical and laboratory data. GB, VS, NYA-B, and OA were responsible for the laboratory confirmation. JPS, WAN, and YS did the statistical analyses. TSvdW, WAN, and YS contributed to the interpretation of the results and the writing and critical review of the report. All authors have seen and approved the final version of the report. ˇ˛European Journal of Internal Medicine 15 (2004) 451 459 Original article Antineutrophil cytoplasmic antibody-associated neutropenia Paul Coppoa,1, David Gheza,1, Vincent Fuentesb, Djaouida Bengoufac, Eric Oksenhendlera, Bruno Triboutd, Jean-Pierre Clauvela, Kaiss Lassouedb,* aService d Immuno-He matologie, Ho∆pital Saint-Louis, Paris bService d Immunologie, CHU d Amiens, France cLaboratoire d Immunopathologie, Ho∆pital Saint-Louis, Paris dService de Pathologie Vasculaire, CHU d Amiens, France Received 14 January 2004; received in revised form 1 July 2004; accepted 31 August 2004 www.elsevier.com/locate/ejim Abstract Background: Antineutrophil cytoplasmic antibodies (ANCA) can be associated with various disorders. However, their association with neutropenia has never been reported. Methods: Nine patients with chronic unexplained neutropenia and ANCA were studied. Clinical charts were extensively analyzed and all patients underwent hematological and immunological investigations. Results: All patients (6 women and 3 men) were Caucasian and had a mean age of 49 years (range 16 67 years). All presented with a neutropenia below 1.5ˇ€‹109/L for more than 6 months. The neutropenia was b0.5ˇ€‹109/L in six cases and moderate in three. There was no evidence of toxic- or drug-related neutropenia or of a hematological malignancy. Autoimmune anemia and/or thrombocytopenia were present in five patients. ANCA, with various specificities, were present in all patients. ANCA were associated with various other autoantibodies in eight patients, including antisurface-neutrophil antibodies in three cases. Four of the six patients with severe neutropenia experienced infections. Five patients were treated with hematopoietic growth factors, steroids, intravenous immunoglobulins, splenectomy, methotrexate and/or cyclophosphamide, allowing the neutrophil count to be restored transiently or permanently. Conclusions: A subset of patients with neutropenia of possible autoimmune origin may develop ANCA. Their detection would provide strong evidence of an autoimmune mechanism. Neutropenia should be added to the list of ANCA-associated diseases. D 2004 Elsevier B.V. All rights reserved. Keywords: Neutropenia; Antineutrophil cytoplasmic antibodies; Autoimmunity 1. Introduction Antineutrophil cytoplasmic antibodies (ANCA) are autoantibodies that are directed against different neutrophil antigens. When detected by immunofluorescence on etha- nol-fixed human neutrophils, ANCA usually displays three major patterns: cytoplasmic (cANCA), perinuclear (pANCA), and atypical (xANCA). The main targets of * Corresponding author. Service d Immunologie, Faculte  de Me decine, 3 rue des Louvels, F-80036 Amiens Cedex 1, France. Tel.: +33 3 22 82 79 06; fax: +33 3 22 82 79 07. E-mail address: kaiss.lassoued@sa.u-picardie.fr (K. Lassoued). 1 Contributed equally to this work. 0953-6205/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ejim.2004.08.009 ANCA are either myeloperoxydase (MPO) or proteinase 3 (PR3) [1 4], but they may also be directed against lactoferrin, elastase, cathepsin, lysozyme [5], bactericidal permeability increasing protein (BPI), and azurocidin [6]. ANCA are observed in a large spectrum of diseases [7 19]. Their specificity can make them a helpful tool in the diagnosis of primary systemic vasculitides [20]. Anti-PR3 antibodies (Abs) are strongly associated with Wegener s disease and, to a lesser extent, with microscopic polyarteritis and necrotizing crescentic glomerulonephritis [7 9]. Anti- MPO antibodies are often associated with systemic vascu- litis but can also be found in various autoimmune disorders and connective tissue diseases without evidence of vasculitis [10,11,13,14,16 18]. Other specificity is not considered to 452 P. Coppo et al. / European Journal of Internal Medicine 15 (2004) 451 459 be a helpful serological marker of any diseases. We report herein nine patients with chronic peripheral neutropenia associated with ANCA. To our knowledge, this association has not previously been described in the literature. 2. Material and methods 2.1. Patients Nine adult patients were referred to the Hematology Departments of Ho∆pital Saint-Louis (Paris, France) and Ho∆pital Sud (Amiens, France) for chronic neutropenia defined as a neutrophil count below 1.5ˇ€‹109/L for more than 6 months and the absence of a toxic or infectious etiology. All patients had a complete physical examination. Medications were systematically recorded, as well as any toxic usage or exposure. All patients had serologies for acute viral infections [EBV, CMV, parvovirus B19, hepatitis B (HBV) and C (HCV) viruses], as well as for HIV and HTLV- 1. In one patient with CVID, a blood sample was examined for the presence of CMV antigenemia, HBS antigen, and HCV and HIV transcripts using PCR. A blood smear examination was done on all patients. Analysis of bone marrow aspirates and/or biopsy, immunophenotyping of blood and/or bone marrow lymphocytes, and an analysis of blood lymphocyte clonality were performed when warranted. 2.2. ANCA detection and characterization Sera samples were stored at ˇ€‹20 8C until analysis. In patients 8 and 9, ANCA were sought and found to be positive 2 years prior to the onset of neutropenia; in all other patients, ANCA were sought for the first time in the setting of neutropenia. In all patients, ANCA detection was repeatedly performed during the follow-up. ANCA were detected by indirect immunofluorescence (IIF) using healthy human neutrophils (INOVA Diagnostics, San Diego, CA, USA) [21]. To rule out false-positive results, artefact results due to the presence of antinuclear antibodies (ANA) in five patients, freshly isolated granulocytes from healthy donors were fixed in different ways [98% ethanol (15 min, ˇ€‹20 8C), 95% methanol (15 min, ˇ€‹20 8C), 4% formaldehyde (10 min, +20 8C), 45% acetone 4% formaldehyde (15 min, +20 8C)] and tested by IIF for the presence of ANCA [22]. The sera were examined independently for their immunofluor- escence patterns by two investigators. ANCA specificity was determined using an enzyme-linked immunosorbent assay (ELISA; Euroimmune, Lqbeck, Germany) and by Western blot. For Western blot analysis, white blood cells were isolated from normal blood donors by centrifugation in Ficoll-Hypaque gradients, followed by hypotonic lysis [0.2% NaCl in distilled water, 1 mM phenylmethylsulfonyl- fluoride (PMSF), 1 mM EDTA]. Granulocytes were resuspended in 1.6% NaCl solution, washed, and lysed in 1% Nonidet P-40 (NP-40), 150 mM NaCl, 50 mM Tris (pH 7.4), 0.2% sodium azide, 1 mM di-isopropylfluorophos- phate, 10 mM EDTA, 20 mM iodoacetamide, and 1 mM PMSF. Cell lysates were centrifuged (20,000ˇ€‹g, 20 min, +4 8C) and the supernatant resuspended in 2ˇ€‹Laemmli s buffer [23]. Granulocyte cytoplasmic proteins were separated using SDS-10% polyacrylamide gel electrophoresis (PAGE) and electrotransferred onto nitrocellulose filters, as previ- ously described [24]. Immobilized proteins were incubated with patients sera diluted at 1/50, then with goat antihuman antibody coupled to horseradish peroxydase, and revealed using an enhanced chemiluminescence (ECL) kit (Amer- sham Arlington Heights, IL, USA). Purified MPO (Sigma, France) was also used in some Western blot experiments. 2.3. Other biological investigations The patients were evaluated with routine laboratory tests, including blood cell count, bone marrow aspirates and/or bone marrow biopsy. All patients had extensive immuno- logical investigations. Detection of ANA and identification of their specificities were performed as reported elsewhere [25]. Antisurface neutrophil autoantibodies were detected with granulocyte immunofluorescence (GIFT) and with a granulocyte agglu- tination test (GAT) and antiplatelet antibodies (Abs) by monoclonal Ab immobilization of platelet antigens (MAIPA), as described elsewhere [26 28]. Lymphocyte immunophenotyping and blood lymphocyte clonality analysis were performed as previously described [26]. Control sera were obtained from 8 children with anti- NA1 or anti-NA2 Abs-autoimmune neutropenia, 14 patients with drug-induced neutropenia, 11 patients with isolated immune thrombocytopenia [8] or autoimmune hemolytic anemia [3], 6 patients with Sjfgren s syndrome, 2 patients with Hashimoto s thyroiditis and 8 intravenous immuno- globulin (IVIG)-treated CVID patients with no neutropenia, and 22 healthy individuals. Positive controls for anti-PR3 and MPO Abs were provided with the commercial kits. 3. Results 3.1. Clinical features All nine patients were Caucasians. Three patients were male and six female, and they ranged in age from 16 to 67 years (mean 49 years). In four patients (patients 1, 2, 5, and 6), the past medical history was unremarkable (Table 1). Patient 3 had been diagnosed with Hashimoto s thyroiditis 13 years earlier and had a history of recurrent miscarriages associated with antiphospholipid Abs. Patients 4 and 8 had a long-standing history of non-destructive seronegative poly- arthritis of unknown origin. Patients 8 and 9 had salivary gland infiltration suggestive of Sjfgren s syndrome. In the latter, ANCA were found 2 years before the onset of Table 1 Main clinical and hematological findings P. Coppo et al. / European Journal of Internal Medicine 15 (2004) 451 459 453 Patient Sex/age Organomegaly 1 M/16 Peripheral adenopathies 2 F/46  3 F/56 Splenomegaly 4 M/60 Splenomegaly Hepatomegaly 5 F/37  6 M/44 Peripheral Adenopathies 7 F/67  8 F/62  9 F/54  Infections Anal abcess, folliculitis None Furunculosis None Clonal CD8+gy+ T cell expansion None Pneumonia None Recurrent bronchitis Other conditions Raynaud s phenomenon Livedo Evan s syndrome  Hashimoto s thyroiditis Livedo Necrotizing vasculitis Multiple miscarriages Hemolytic anemia Non destructive polyarthritis Hemolytic anemia Thrombocytopenia Myalgias Good s syndrome Evan s syndrome Polyclonal CD8+ T-cell Expansion Neutropenia (ˇ€‹109/L) Lower value b100 1200 b100 b100 1200 1150 b100 Bone marrow Myeloid maturational arrest (myelocyte/ metamyelocyte stage) ND Normal Lymphocyte expansion ND Normal Myeloid maturational arrest (myelocyte/ metamyelocyte stage) Lymphocyte expansion Myeloid maturational arrest (myelocyte/ metamyelocyte stage) Normal Treatment IV Ig Steroids G-CSF No treatment G-CSF Splenectomy G-CSF Methotrexate Steriods No treatment No treatment IV Ig G-CSF Steroids Cyclophosphamide Steroids Methotrexate Steroids G-CSF Methotrexate Non-destructive 200 polyarthritis Sjfgren s syndrome Polyclonal CD8+ ah+T-cell expansion (LGL) Sjfgren s syndrome 150 Male (M), female (F), large granular lymphocytes (LGL), not done (ND), intravenous immunoglobulins (IVIG), granulocyte colony-stimulating factor (G-CSF). neutropenia. Patient 7 had had recurrent upper respiratory tract infections related to a common variable immunodefi- ciency (CVID) diagnosed 2 years earlier. She was being treated monthly with IVIG (O.4 g/kg). Moderate splenomegaly was present in patients 3 and 4 and was associated with hepatomegaly in patient 3. This patient underwent a splenectomy early in the course of the disease. Peripheral lymph node enlargement was found in two patients (patients 1 and 6). In one (patient 1), a lymph node biopsy showed a non-specific lymphoid hyperplasia. Cutaneous manifestations were seen in two patients (patients 1 and 3) and consisted of a livedo in both cases, associated with a Raynaud s phenomenon in patient 1 and with necrotizing vasculitis in patient 3. Infections occurred in four patients with severe neutro- penia (b0.5ˇ€‹109/L) and consisted of folliculitis with an anal abscess in one case (patient 1), diffuse furunculosis in one case (patient 3), pneumonia in the patient with CVID (patient 7), and recurrent bronchitis in patient 9, who had Sjfgren s syndrome. Aside from these two patients, there was no patient with respiratory manifestations. There was no evidence of sinus involvement in any of the nine patients. Neither drug intake nor exposure to toxic compounds that could induce neutropenia could be recorded. 3.2. Hematological findings (Table 1) All patients had neutropenia of variable severity (Table 1). Non-periodic fluctuations in the neutrophils were observed in patients 3 and 4. Four patients (patients 2, 6, 8, and 9) had neutropenia with no other cytopenia. In three patients (patients 1, 5, and 7), the neutropenia was associated with a peripheral thrombocytopenia that was severe (b10ˇ€‹109/L in patients 1 and 7) or moderate (100ˇ€‹109/L in patient 5). Four patients had hemolytic anemia (patients 1, 3, 4, and 7). In patient 1, Evan s syndrome was diagnosed 5 years after the neutropenia, and in patient 4, hemolytic anemia appeared 7 years later. Thrombocytopenia and/or hemolytic anemia were present at the time of diagnosis in three patients. The monocyte count was normal in all patients. 454 P. Coppo et al. / European Journal of Internal Medicine 15 (2004) 451 459 Blood smear examination showed large granular lym- phocytes (LGL) in patient 8, accounting for 40% of the total lymphocyte count (i.e., 0.8ˇ€‹109/L). This LGL population consisted of polyclonal TCR ah+ CD3+, CD4ˇ€‹, and CD8+ lymphocytes. Patient 4 had persistent clonal TCR gy+ CD3+, CD4ˇ€‹, CD8+ lymphocyte expansion in peripheral blood, accounting for 30% of the blood lymphocyte count, with no LGL. The cytological pattern did not suggest any hematological malignancies. Blood lymphocyte count and immunophenotyping were normal in the remaining patients. Blood lymphocyte clonality was also assessed in patients 3 and 9 and did not reveal any predominant lymphocyte clones. A bone marrow analysis was performed in six patients; in all cases, the bone marrow aspirate showed normal or increased cellularity. Myeloid cell count was normal in three cases (patients 3, 6, and 9) and showed a myeloid maturational arrest at the myelocyte/metamyelocyte stage in three others (patients 1, 7, and 8). Among these, one (patient 7) had a polyclonal TCR ah+ CD3+, CD8+ cell expansion (38%) and another (patient 4) had a TCR gy+ CD3+, CD4ˇ€‹, CD8+ lymphocyte expansion that was also found in blood. 3.3. ANCA detection and characterization Using an IIF assay and ethanol-fixed neutrophils, ANCA were found to be positive in all patients (Table 2). The fluorescence pattern was cytoplasmic (cANCA) in four cases (patients 1, 2, 5, and 6), perinuclear (pANCA) in two cases (patients 3 and 7), and atypical (xANCA) in three (patients 4, 8, and 9; Fig. 1). Sera were further tested by IIF using ethanol-, methanol-, formaldehyde-, acetone-form- aldehyde-fixed neutrophils and found to label neutrophil cytoplasm in all cases. Using ELISA, the ANCA specificity was anti-PR 3 (patient 2), anti-MPO plus anti-elastase (patient 3), anti- MPO (patient 8), and anti-lactoferrin (patient 7). Reac- tivity of patients 3 and 8 with MPO was confirmed by Western blot using the purified enzyme (data not shown). Fig. 1. Fluorescence photomicrograph of neutrophils after indirect immunofluorescence staining using serum from patient 1. Magnification ˇ€‹40. In the remaining patients, the specificity was undetermined by ELISA using MPO, PR3, cathepsin G, elastase, BPI, and lactoferrin as substrates and by Western blot using MPO. ANCA were not detected by IIF or ELISA in the control sera obtained from 22 healthy individuals, from 8 children with anti-NA1 or anti-NA2-associated neutrope- nia, from patients with drug-induced neutropenia or various autoimmune disorders, or from 8 intravenous Ig- treated CVID patients, ruling out the possibility that ANCA were passively transmitted through IVIG prepara- tions in patient 7. Three sera (patients 1, 4, and 5), negative by ELISA, were tested by Western blot using cytoplasmic extracts of freshly isolated polymorphonuclear cells. As shown in Fig. 2, these sera displayed a heterogeneous pattern of reactivity. The serum of patient 1 showed a strong reactivity with two proteins of 60 and 33 kD and a weaker reactivity with two proteins of 45 and 38 kDa (lane 3). A strong reactivity with 61 and 52 kDa proteins and a weaker reactivity with additional proteins (45, 43, and 27 kDa) were obtained with patient 4 serum (lane 4). Patient 5 serum revealed two major bands of 59 and 53 kDa, and one additional faint band of 45 kDa (lane 5). The serum of patient 8 (lane 6), which Table 2 Main immunological findings Immunological data Patients ANCA Titer (at diagnosis)a Fluorescence pattern Specificity Titer 100 400 400 Fluorescence pattern speckled homogeneous homogeneous 123 45 67 80 640 cANCA pANCA n.i. lactoferrin ˇ€‹ˇ€‹ ˇ€‹+ 89 80 160 ANCA xANCA MPO n.i. ˇ€‹ˇ€‹ ++ 160 320 speckled homogeneous ˇ€‹ˇ€‹ ˇ€‹ ND 320 80 cANCA cANCA NI PR3 +ˇ€‹ˇ€‹ ++ Antisurface neutrophil Abs AntinuclearAbs ˇ€‹+ˇ€‹ ˇ€‹+ Coombs test +ˇ€‹+ +ˇ€‹ Antiplatelet Abs ˇ€‹ ˇ€‹ ˇ€‹ ˇ€‹ + Antibodies (Abs), not identified (n.i.), not done (ND). a Positive if N20. ˇ€‹+ ND+ 80 5120 80 pANCA xANCA cANCA MPO/elastase n.i. n.i. P. Coppo et al. / European Journal of Internal Medicine 15 (2004) 451 459 455 3.5. Other findings Fig. 2. Neutrophil cytoplasmic proteins were electrophoretically separated by SDS 10% PAGE, transferred onto nitrocellulose, and probed with normal serum (lane 1), control serum with anti-MPO Abs (lane 2), and sera from patients 1 (lane 3), 4 (lane 4), 5 (lane 5), and 8 (lane 6). Molecular weights are indicated on the right. contained anti-MPO Abs, recognized the 58 kDa heavy chain of MPO, as did the anti-MPO control serum (lane 2), each of which reacted with additional proteins. 3.4. Other biological findings (Table 2) Three patients had antisurface-neutrophil Abs detected by GIFT, the specificity of which could not be identified (patients 1, 4, and 5); anti-NA1, NA2, NB1, and NB2 specificities were ruled out in all three patients. ANA were detected in five patients, displaying a homogeneous pattern or a speckled pattern. No specificity could be identified in any of the cases; there were neither antidouble-stranded DNA nor anti-ENA autoantibodies. Rheumatoid factor with decreased C3, C4, CH 50 was found in two patients associated with a type III (patient 3) or a type II (patient 8) mixed cryoglobulinemia. Antikeratin and antifilaggrin Abs were not detected in the two patients with non-destructive arthritis (patients 4 and 8). Patients 1, 3, 4, and 7 had a positive Coombs test associated with hemolytic anemia. Cold agglutinins (polyclonal IgM with an anti-I specificity) were also present in patient 3 at a titer of 125. Antiplatelet glycoprotein IIb/IIIa Abs was found in the two patients who had peripheral thrombocytopenia. The gamma globulin level was normal in all patients and none of them had a deficiency in serum IgA or IgG subclasses, except for the patient with the CVID. Serum creatinine level was normal and the search for proteinuria and hematuria was consistently negative. Serol- ogies for acute viral infections (EBV, CMV, parvovirus B19), as well as for HIV and HTLV-1, were negative. CMV antigenemia, HBS antigen, and HCV and HIV-1 transcripts were absent in the CVID patient. Computerized abdominal and chest tomography was performed in six patients (1, 3, 4, 7, 8, and 9). It was normal in patients 1, 8, and 9 and confirmed the hepatomegaly in patient 4 and the spleno- megaly in patients 3 and 4. In patient 7, it revealed a thymoma, consistent with the diagnosis of Good s syn- drome, since this patient had a CVID. 3.6. Treatment and follow-up In this retrospective analysis of patients from two separate hospitals with various underlying conditions and different degrees of severity of neutropenia, treatment varied. Patients 2, 5, and 6, who had neutropenia between 1.15ˇ€‹109/L and 1.5ˇ€‹109/L, were not treated. The neutro- penia remained stable and none of them experienced any infections within a median follow-up of 2 years (range 1 7 years). They have remained in good health without any evolution to a characterized disease. Patients 1, 3, 4, and 9 were variably treated with IVIG, steroids, G-CSF, splenectomy, and/or oral methotrexate, resulting in a complete recovery in all cases (long-term in patients 1 and 9). Patients 3 and 4 required intermittent therapy with G-CSF in order to maintain neutrophil counts above 1ˇ€‹109/L. When neutropenia appeared in patient 7, who had a CVID, she was initially treated with G-CSF (250 Ag/day) for 8 days with little effect. She was then given prednisolone (0.5 mg/kg per day) for 5 days and two courses of cyclophosphamide, at a dose of 800 mg every month, allowing for an increase in the neutrophil count. Yet, she subsequently died of pneumonia. Patient 8 was treated with steroids (1 mg/kg per day for 1 month) with no effect; neutrophil counts have remained below 0.5ˇ€‹109. Because he never experienced infectious complications after 3 years of follow-up, he did not receive further treatment. The evolution under treatment of the neutrophil count in patients 3, 4, and 7 is shown in Fig. 3. None of the patients, in particular patient 2, who had anti-PR3 Abs, developed manifestations suggestive of Wegener s disease, microscopic polyangiitis, or crescentic glomerulonephritis after a follow-up of 7 years. During follow-up, ANCA remained present with fluctuating titers in the sera of all patients, and there was no clear correlation between ANCA titers and neutrophil counts. None of the patients, in particular patient 2, who had anti-PR3 Abs, developed manifestations suggestive of 456 P. Coppo et al. / European Journal of Internal Medicine 15 (2004) 451 459 Fig. 3. Evolution of the neutrophil count in three patients. Cyclo- phosphamide (CPM), methotrexate (MTX), granulocyte colony-stimulating factor (G-CSF). Wegener s disease, microscopic polyangiitis, or crescentic glomerulonephritis after a follow-up of 7 years. During follow-up, ANCA remained present with fluctuating titers in the sera of all patients, and there was no clear correlation between ANCA titers and neutrophil counts. 4. Discussion ANCA is considered to be important serological markers for primary systemic vasculitides. When detected by IIF assays using the patient s serum and ethanol-fixed human neutrophils, ANCA display three different fluorescent staining patterns: a cytoplasmic (cANCA) pattern, a perinuclear (pANCA) pattern and, more rarely, an atypical (x-ANCA) pattern. cANCA are usually directed to the 29- kDa neutral serine protease PR3, whereas pANCA are directed to MPO or, occasionally, to other components of azurophilic granules, such as lysozyme, elastase, cathepsin G, lactoferrin, azurocidin, or BPI. ANCA specificity can be routinely assessed with Western blot or ELISA using specific antigens. Anti-PR3 serum reactivity is present in up to 90% of patients with Wegener s granulomatosis [9,29,30], but it can also be detected in patients with microscopic polyarteritis and necrotizing crescentic glomer- ulonephritis [7,8,12]. The detection of ANCA is, however, very rare in the general population, as opposed to ANA or rheumatoid factor [31], and false-positive results have been chiefly described in patients with polyclonal hypergamma- globulinemia and HIV infection [32]. We report herein nine patients who share both non- cyclical chronic neutropenia and a positive ANCA test. ANCA specificity could be identified in only four patients and was different from one patient to the next (MPO, elastase, and lactoferrin). Sera from the five other patients did not exhibit any reactivity with PR3, MPO, elastase, cathepsin G, BPI, or lactoferrin according to the ELISA. With Western blot, three of these five ELISA-negative sera showed reactivity to several granulocyte cytoplasmic proteins whose precise identification will require further investigation. However, we cannot rule out the possibility that some other antigenic targets may not have been detected with Western blot. In keeping with this hypothesis, ANCA was often reported to react poorly by immunoblot- ting with either crude neutrophil extracts or azurophilic granule extracts [33]. There was no evidence of either a toxic or immunoaller- gic mechanism. Hypersplenism does not seem probable since only two patients had moderate splenomegaly and the neutropenia relapsed in the one who underwent a splenec- tomy. There was no paroxysmal nocturnal hemoglobinuria clone detected in our patients (data not shown). Bone marrow examination showed a myeloid maturational arrest at the myelocyte/metamyelocyte stage with no evidence of myelodysplasia or blast excess. Chronic T-cell proliferation was observed in the periph- eral blood and/or in bone marrow in three patients. One was a polyclonal T CD8+ bone marrow cell proliferation associated with Good s syndrome. This case is reminiscent of those of two other patients previously reported who had an IgG2 subclass deficiency associated with a polyclonal T- lymphocyte bone marrow infiltration and severe neutropenia [34]. It is important to mention that, although agranulocy- tosis has been described in Good s syndrome, it is not usually associated with bone marrow lymphoid infiltration [35]. Two patients had chronic T-cell expansion that consisted of clonal T gy+ lymphocytes both in peripheral blood and in bone marrow, or polyclonal CD3+ CD8+ ah+ LGL in peripheral blood only, associated with or without splenomegaly. Since both patients suffered from a sero- negative non-destructive polyarthritis, Felty s syndrome or a Felty s-like syndrome might have been envisaged. More- over, ANCA have been reported at very early stages of rheumatoid arthritis, albeit without any association with neutropenia, and seem to be predictive of rapid radiographic joint destruction [36]. That our patients never developed any joint destruction or serological markers of rheumatoid arthritis pleads strongly against this diagnosis. The mech- anism of neutropenia in patients with T-cell proliferation is unknown. While LGL are known to have multiple effects on the hematopoietic system, in vitro studies have generally failed to demonstrate any suppressive effect of LGL on granulopoiesis [37,38]. In our six other patients, an autoimmune mechanism was strongly suggested by the association with manifestations of autoimmunity such as Hashimoto s thyroiditis, Sjfgren s syndrome, autoimmune hemolytic anemia or thrombocyto- penia, and the presence of a large variety of autoantibodies. Furthermore, the presence of circulating autoantibodies directed against surface neutrophil antigens in three cases provides additional evidence for the autoimmune origin of the neutropenia. The absence of detectable surface neutro- phil autoantibodies in the remaining six patients may be explained by the nonexpression of the target antigen by the neutrophils used in the assays, the very low titer of the circulating autoantibodies, and/or a very high affinity, causing the Abs to be essentially bound to their targets. The pathogenic role of antineutrophil autoantibodies is still not completely clear, apart from the autoimmune neutropenia of children, which generally combines an isolated granulocytopenia and circulating autoantibodies usually directed against isoforms of CD16 such as NA1, NA2, NB1, and NB2 [39]. In adults, the neutropenia is often associated with other autoimmune manifestations, particu- larly autoimmune thrombocytopenia and hemolytic anemia [40,41]. However, the clinical significance of antineutrophil autoantibodies is far from unequivocal as they can often be found in various autoimmune disorders such as SLE or Sjfgren s syndrome in patients with a normal neutrophil count [41,42]. In diseases commonly associated with ANCA, in particular Wegener s granulomatosis, there is often a marked hyperleukocytosis [43]. Therefore, the presence of ANCA in neutropenic patients is unexpected and raises the question of their possible role in the pathogenesis of the neutropenia. One hypothesis is that the usually intracytoplasmic targets of ANCA might be redistributed on the patients neutrophil surface, thereby enabling their recognition by ANCA and neutrophil destruction, by either a complement-mediated or antibody-dependent cytotoxicity. Interestingly, it has been established that antigens like PR3 and MPO can be surface- expressed upon neutrophil activation [44 47]. Therefore, ANCA-associated neutropenia might represent an uncom- mon subset of autoimmune neutropenia with particular antineutrophil autoantibodies directed against targets that are normally located in the cytoplasm but could be induced or aberrantly expressed on the neutrophil surface of these patients. This would provide an explanation for the negativity of assays like GATT and GIFT in most of our cases since they use pooled neutrophils from healthy donors. In keeping with our hypothesis, sera from four of our patients transiently stained both the cytoplasm and the surface of all-trans retinoic acid-stimulated myelomonocytic HL60 cells (personal data). Experiments are currently being conducted to identify the target antigens recognized by these Abs. As has been suggested for other autoantibodies [48], ANCA have recently been shown to enter the PMN and cause cell damage, including apoptosis [49]. However, we cannot exclude the possibility that T lymphocytes may play a major role. They could either be self-reactive cytotoxic T lymphocytes, producing granulopoiesis inhibitory cyto- kines, or induce apoptosis of polymorphonuclear cells via Fas/Fas ligand interaction. In this latter setting, one hypothesis is that ANCA might be induced by a massive release of target antigens by apoptotic neutrophils [48] that could stimulate self-reactive T and B lymphocytes as a secondary event. This may explain the absence of a correlation between ANCA titers and neutrophil counts, as well as the large number of antigen targets detected by Western blot and ELISA. To summarize, we report a series of nine patients with a heterogeneous spectrum of underlying diseases but who all share the association of ANCA and neutropenia of unknown origin in which an autoimmune mechanism is strongly suggested. The responsibility of ANCA in the pathogenesis of the neutropenia cannot be affirmed. A systematic analysis of a larger series of patients is required to better define the clinical significance of neutropenia associated with ANCA. We believe that ANCA should now be sought in every neutropenia of undetermined origin. This would provide a strong argument for an autoimmune mechanism. Until a pathogenic role can clearly be determined, ANCA should at least be considered to be an interesting marker of auto- immune neutropenia. Treatment should be considered individually since ANCA-associated neutropenia may be of variable severity and can be associated with various other conditions. Acknowledgements The authors thank Muriel Bargis-Touchard, Sylvie Duflot,andBe ne dictePecquet for their expert secretarial assistance. ˇ˛European Journal of Internal Medicine 20 (2009) 403 406 Contents lists available at ScienceDirect European Journal of Internal Medicine journal homepage: www.elsevier.com/locate/ejim Original article Antioxidant enzyme levels in cases with gastrointesinal cancer Yalcin Kekec, Semra Paydas N , Abdullah Tuli, Suzan Zorludemir, Gurhan Sakman, Gulsah Seydaoglu Cukurova University Faculty of Medicine Departments of General Surgery, Medical oncology, Biochemistry, Pathology and Biostatistics, Turkey article info Article history: Received 26 August 2008 Received in revised form 20 November 2008 Accepted 18 December 2008 Available online 26 January 2009 Keywords: Gastrointestinal cancer Antioxidant system SOD G6PD GR 1. Introduction Oxygen free radicals, including highly reactive superoxide radicals termed as ROS have been of increasing interest in cancer researches in the past two decades. The effects of ROS can be harmful and beneficial [1,2]. At low levels, ROS can induce a mitogenic effect and participates intracellular signallng pathways [3,4]. The induction of ROS induces DNA damage and has been shown to be involved in the etiology of several diseases including cancer [2,4]. These harmful effects of ROS are protected by both enzymatic and non-enzymatic antioxidants [1]. ROS are generated during normal aerobic metabolism and increased levels are present during oxidative stress. It has been proposed that ROS is necessary for life and essential for the regulation of essential physiologic functions. However, at high concentrations, ROS are cytotoxic. ROS are important in cell differentiation, apoptosis, and cell proliferation. These functions are regulated by redox-sensitive signal transduction pathways. The amount of antioxidants in the cells is high and so cells prevent or repair itself from damages caused by ROS. ROS-induced damage can result in cell death, mutations, chromosomal aberrations and also carcinogenesis [5]. For this reason there is great interest to ROS in cancer in the last years. These N Corresponding author. Cukurova University Faculty of Medicine Department of Medical Oncology, Adana/Turkey. E-mail address: sepay@cu.edu.tr (S. Paydas). abstract The aim is to evaluate the antioxidant enzyme levels in tumoral tissues and accompanying normal tissues in gastrointestinal cancer; and compare the colorectal cancer (CRC) with gastric cancer (GC). Method: Antioxidant enzymes including glutathione reductase (GR), glutathione peroxidase (GPX), super- oxide dismutase (SOD), malondialdehyde (MDA) and glucose 6 phosphate dehyrogenase (G6PD) which are important for anti-oxidant functions were evaluated in fresh tumor tissues and adjacent normal tissues obtained from a total of 58 patients. Results: All the enzyme levels were higher in tumoral tissues compared to normal tissue from non-cancerous disease. There was not a significant difference for enzyme levels between CRC and GC groups except GPx. GPx activity tended to be higher in cases without serosal involvement (SI), and this activity was higher in cases without lymph node (LN) involvement in normal tissue (p = 0.012). MDA activity was higher in cases without serosal involvement compared to with SI groups in tumor tissue (p=0.050). G6PD activity in normal tissue was higher in cases with serosal involvement and LN involvement (p=0.064, 0.046, respectively). GR activity was higher in signet ring cell cancer (SRC) than adeno cancer. In GC, G6PD activity in tumor was tended to be higher in undifferentiated cancer (p=0.071). Conclusion: The antioxidant enzymes activities such as GPX, SOD, G6PD, MDA and GR were found to be related with malignant phenotype in gastrointestinal cancers. We need further studies to understand the biologic and clinical importance of these enzymes in GI cancers. © 2009 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. antioxidants are antioxidant enzymes and some vitamins. There are three major types of antioxidant enzymes in mammalian cells: superoxide dismutase (SOD), catalase (CAT), and peroxidase, of which glutathione peroxidase (GPx) is the most important component of these [6,7]. MDA is a naturally occurring product of lipid peroxidation and prostaglandin synthesis. Lipid hydroperoxides are formed in vivo through free radical pathways from the action of ROS on the polyunsaturated fatty acids. MDA is formed during decomposition of lipid hydropeoxides. The major adduct to DNA is a pyrimidopurinone called M1G (M1G is a secondary DNA damage product arising from primary reactive oxygen species (ROS) damage to membrane lipids or deoxyribose) and this appears to be a major endogenous DNA adduct in human tissues. This may contribute to cancer development [8,9]. In this study, antioxidant enzymes such as GR, GPX, SOD, MDA and G6PD activities were studied in tumoral tissues and normal tissue from non-cancerous disease in 58 cases with gastrointestinal cancer and compared them according to some clinical and histo-pathological characteristics of these tumors. 1.1. Method Antioxidant enzyme (AOE) system including glutathione reductase (GR), glutathione peroxidase (GPX), superoxide dismutase (SOD), malondialdehyde (MDA) and glucose 6 phosphate dehyrogenase 0953-6205/$  see front matter © 2009 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.ejim.2008.12.003 404 Y. Kekec et al. / European Journal of Internal Medicine 20 (2009) 403 406 (G6PD) was evaluated both in fresh tumor tissues and adjacent normal tissues obtained from a total of 58 patients with gastrointestinal cancer (33 had CRC and 25 were had GC). 2. Patients and methods 58 cases were included in this study; 25 of them had GC and 33 had CRC. Fresh tumor tissues and normal tissue from non-cancerous disease were used as study material. 2.1. Sample preparation Fresh tumor tissues and adjacent normal tissues obtained from patients with colorectal and gastric cancer immediately after surgery was placed in cold 0.9% NaCl solution. The tissues were blotted on filter paper, weighed and in 0.25 M sucrose and 0.15 M NaCl with addition of 6 ºl 250 mM butylated hydroxytoluene in ethanol, to prevent formation of new peroxides during the assay. The tissues were homogenized under standardized conditions and centrifuged at 10.000 ◊g for 10 min at 4 ∞C. The supernatant was kept in an ice- cold condition until assayed. 3. Biochemical assays Superoxide dismutase (SOD) enzyme was assayed according to the methods of McCord et al. [10]. The enzyme specific activity was expressed in units per milligram protein. Glutathione reductase (GR) activity was determined by the method of Mize and Langdon by monitoring the oxidation of nicotinamide adenine dinucleotide phosphate (NADPH) at 340 nm [11]. The activity was expressed in millimoles NADPH per minute per milligram protein. Glutathione peroxidase (GSH-Px) activity was measured spectro- photometrically using a technique based on the method of Paglia and Valentine [12]. The activity was expressed in universal units. One unit of the enzyme activity is defined as the amount of enzyme catalyzing the oxidation of ºmol NADPH/min/mg protein. Glucose 6-phosphate dehydrogenase enzyme activity was assayed by the method of Beutler and Aksoy et al. [13,14]. The enzyme activity was expressed in micromoles NADPH reduced per minute per mg protein. The lipid peroxide level in tissue was assayed in terms of malondialdehyde level (pmol/mg protein), the end product of lipid peroxidation, by using method described by Ohkawa et al. [15]. Protein was determined in diluted aliquots of the tissue homogenates by the method of Lowry et al. with bovine serum albumin as the standard [16]. 3.1. Statistical methods For each continuous variable, normality was checked. Student t test was chosen for normally distributed data. Appropriate non- parametric tests were chosen (Mann Whitney U and Kruskal Wallis test) for the data was not normally distributed. Since analysis of variance was significant, comparisons were applied using the Mann Whitney U test. Bonferroni's correction was applied (p b 0.05/n; where n=number of comparisons) when multiple comparison were made. Results were presented as mean±SD and median. 4. Results Thirty-three cases with CRC and 25 cases with GC were included. Mean age was 53.62±1.93 (median: 53 range 22 84) in total cases. Age range was 40 84 and 22 78 for GC and CRC, respectively. Female/ male ratio was 19/14 in CRC group and 10/15 in GC group. Histologically; 28 cases had adeno cancer, 2 had SRC and 3 had undifferentiated cancer in CRC group. In GC group; 15 cases had adeno cancer, 6 had SRC and 4 had undifferentiated cancer. Serosal involvement was detected in 25 cases with CRC and 20 cases with GC. Lymph node involvement was seen in 22 cases with CRC and 20 cases with GC. Table 1 shows the enzyme levels of CRC and GC cases both in tumoral tissues and accompanying normal tissues. All the enzyme levels, were found to be significantly higher in tumoral tissues comparing to normal tissues (pb0.001 for all, not stated in the table). There was not a significant difference for enzyme levels between CRC and GC groups except GPx in normal tissue from non-cancerous disease; GPx activity is significantly higher in CRC group than GC group (p=0.04). There were no significant differences between CRC and GC groups for enzyme levels in tumoral tissues. In Table 2 enzyme levels were evaluated between sex, serosal involvement, lymph node involvement and histologic type groups in non-tumoral and tumoral tissues. Enzyme levels were not different between males and females both in non-tumoral and tumoral tissues. G6PD activity in non-tumoral tissue was found to be higher in cases with lymph node involvement as compared without lymph node involvement (p = 0.046). GPx activity, both in tumor and non-tumoral tissue, tended to be higher in cases without serosal involvemet as compared with serosal involvement group (p = 0.05 and 0.047, respectively) and MDA activity was higher in cases without serosal involvement compared with serosal involvement group in tumor tissue (p = 0.05). According to histological types; GR activity was higher in SRC group than AC (p = 0.011) and GPx activity was higher in AC group than SRC group in tumoral tissue (p = 0.019). These comparisons performed by the subgroups (CRC and GC) of gastrointestinal cancer. However in cases with CRC, G6PD activity in non-tumoral and tumoral tissue was found to be higher in females than males (p = 0.034 and 0.043, respectively). GR activity in normal tissue was found to be higher in females than males (p=0.018). In CRC, GPx activity in normal tissue was higher in cases with without serosal involvement (p = 0.02), GPx activity in tumor tissue was higher in cases Table 1 The enzyme levels of colo-rectal cancer and gastric cancer cases both in tumoral tissues and accompanying normal tissues. Tissues Enzyme Colo-rectal cancer (n=33) Mean ± SD Median Gastric cancer (n=25) p value Mean ± SD Median Non-tumoral GR 0.19±0.04 0.20 ± 0.06 0.18 0.33 ± 0.10 0.31 10.21 ± 6.0 9.6 6.76 ± 4.9 5.0 15.20 ± 10.2 12.0 0.41 ± 0.11 0.38 0.21 ± 0.10 0.19 34.06 ± 10.0 34.2 34.28 ± 15.9 32.0 47.28 ± 26.5 45.0 0.577 0.040 0.455 1.000 0.545 0.132 0.113 0.666 0.666 0.338 Tumor GR 14.0 0.36 ± 0.11 0.37 0.18 GPX 0.38 ±0.12 0.34 SOD 8.95 ± 4.9 9.0 MDA 7.06 ± 5.4 5.0 G6PD 15.15 ± 7.1 GPX 0.23 ± 0.07 0.22 SOD 36.08 ± 13.7 36.0 MDA 35.11 ± 14.2 33.0 G6PD 52.70 ± 22.9 48.0 Superoxide dismutase (SOD): units per milligram protein. Malondialdehyde (MDA): units per milligram protein. Glutathione reductase (GR): millimoles NADPH per minute per milligram protein. Glutathione peroxidase (GSH-Px): ºmol NADPH/min/mg protein. Glucose 6-phosphate dehidrogenase (G6PD): micromoles NADPH reduced per minute per mg protein. Y. Kekec et al. / European Journal of Internal Medicine 20 (2009) 403 406 405 Table 2 The enzyme levels according to the sex, lymph node involvement, serosal invasion and histological subtype. Sex Lymph node involvement Serosal invasion Histological type Mean±SD Mean±SD Mean±SD Mean±SD Female Male No Yes No Yes AC SRC Other (n=29) (n=29) (n=16) (n=42) (n=13) (n=45) (n=43) (n=8) (n=7) Non-tumoral GR 0.20±0.06 0.18±0.03 0.19±0.04 0.19±0.05 0.18±0.03 0.19±0.05 0.19±0.04 0.21±0.10 0.18±0.04 GPX 0.35±0.11 0.36 ± 0.11 0.40 ± 0.13 0.34 ± 0.10 0.42 ± 0.15 0.34 ± 0.09* 0.36 ± 0.11 0.34 ± 0.12 0.35±0.11 SOD 9.2±5.1 9.7±5.7 9.2±3.3 9.6±6.0 8.5±3.2 9.7±5.9 9.5±5.5 11.6±5.9 7.0±3.5 MDA 7.2±5.0 6.6±5.4 5.9±4.7 7.3±5.3 6.5±5.4 7.0±5.2 7.0±5.2 7.6±6.4 5.3±3.4 G6PD 16.3±8.0 13.9±8.9 11.2±6.5 16.6±8.8* Tumor 12.5±5.9 15.9±9.1 14.9±8.5 13.8±7.1 17.8±10.6 GR 0.41±0.12 GPX 0.22±0.08 SOD 36.0±12.8 MDA 34.0±13.3 G6PD 53.2±24.2 0.36±0.10 0.36±0.15 0.22±0.09 0.21±0.07 34.3±11.7 36.6±11.8 35.4±16.4 34.9±17.7 47.4±24.8 45.1±29.2 0.39±0.10 0.22±0.09 34.6±12.4 34.6±13.8 52.3±22.5 0.36±0.11 0.27±0.10 40.1±11.3 42.3±16.7 42.9±23.1 0.39±0.11 0.37±0.12 0.21±0.08* 0.23±0.08 33.8±12.2 35.9±11.7 32.5±13.7* 36.3±15.3 52.5±24.7 49.6±25.6 0.46±0.10 0.17±0.13 33.8±11.2 28.9±13.7 43.2±18.1 0.40±0.06* 0.20±0.07* 32.2±17.2 31.8±12.6 63.1±21.7 *pb0.05 between groups. SRC: signet ring cell cancer. AC: adeno cancer. Superoxide dismutase (SOD): units per milligram protein. Glutathione reductase (GR): millimoles NADPH per minute per milligram protein. Glutathione peroxidase (GSH-Px): ºmol NADPH/min/mg protein. Glucose 6-phosphate dehidrogenase (G6PD): micromoles NADPH reduced per minute per mg protein. without serosal involvement (p = 0.016). In CRC, GPx activity was higher in cases without lymph node involvement (p = 0.012). In GC, G6PD activity in normal tissue was higher in cases with lymph node involvement (p = 0.012). However in CRC, GR activity in normal tissue was higher in AC than others (p = 0.05), GPx and SOD activities in tumor tissue was higher in AC than others (p = 0.004 and 0.029, respectively). In GC, G6PD activity in tumor tended to be higher in undifferentiated cancer (p=0.071). 5. Discussion It is very well known that the enzymatic antioxidant system includes SOD, CAT and GPX. Reactive oxygen metabolites are implicated in the initiation and promotion of cancer [17]. Numerous papers have been published about the altered levels of antioxidant enzymes in cancer cells. Physiologically SODs convert superoxide radical into hydrogen peroxide and molecular oxygen (O2). However the catalases and peroxidases convert hydrogen peroxide into water. In this way, two toxic species, superoxide radical and hydrogen peroxide, are converted to the water. These antioxidant enzymes are thought to be necessary for life in all oxygen metabolizing cells. There is no need to co-factors for SOD and CAT functions, but GPx requires several co-factors and proteins. Two proteins; GR and G-6-PD are considered secondary antioxidant enzymes, since they do not act on ROS, but they are helpers to the GPx to its function [7]. Cancer cells are nearly always low in MnSOD and CAT activity, and usually low in Cu/ZnSOD activity [18]. For this reason MnSOD initially has been proposed as a tumor suppressor gene on the basis of low SOD expression in malignant tumors [19,20]. However several studies have challenged this tumor suppressor concept due to the higher MnSOD expression in neoplasms [17,21,22]. We found higher SOD activity in tumoral tisues as compared with accompanying normal mucosae. However we found SOD activities in tumor tissue was higher in ACthan SRC which is more malignant phenotype. These results suggest and support the variability of this enzyme sytem in GI cancers. It has been shown that in some cancers, reduced expression of MnSOD is due to mutations in the promoter of the gene, while in other types of cancer, reduced levels ofMnSOD are due to abnormal methylation, loss of heterozygosity or mutation in the coding sequence [23 25]. These diferent mechanisms cause the chaotic results about SOD in malignant tumors including GC, CRC and other GI tumors [17,21,22,26 28]. GPx activity is variable [7]. In our study, the only enzyme found to be lower in tumors as compared with normal tissue from non- cancerous disease was GPx and we do not know the exact cause of this. Oxygen radicals react with polyunsaturated fatty acid residues in phospholipids. One of the most abundant carbonyl products of lipid peroxidation is MDA, which reacts with DNA to form adducts. Lipid peroxidation appears to be a major source of endogenous DNA damage in humans that may contribute significantly to the occurrence cancer and other genetic diseases [29]. We found higher MDA levels in tumor tissues as compared with accompanying normal mucosae and also MDA was found to be associated with more malignant phenotype. This is not surprise, because MDA is a product of lipid peroxidation and prostaglandin synthesis which is very important in gastrointest- inal malignant tumor development. The content of AOEs in human neoplastic and nonneoplastic tissues has been reported to be different with regard to the organs and enzyme forms It has not been found an association between SOD levels and clinic-pathological parameters in all studies [17]. The significance of these ROS in human malignant tumors is controversial and requires further studies. 6. Learning points " This study is about the pathogenetic role of the antioxidant system gastrointestinal system. There are many enzyme systems in this pathway and some are harmful for organism. It is important to detect their role and also opposite systems to prevent carcinogenesis. ˇ˛Antiphospholipid syndrome Steven Austin Hannah Cohen Abstract Antiphospholipid syndrome (APS) is an autoimmune disorder in which autoantibody production can lead to a hypercoagulable state, pregnancy failure and/or a multitude of other systemic manifestations. In recent times, research into the antiphospholipid antibodies has increased our understanding of the pathogenic process and encouraged improved detection of antiphospholipid antibodies. However, the precise nature of the pathology of APS remains challenging, as does the management of this disorder. This review outlines the key features of APS, including the laboratory tests and their interpretation, and offers advice regarding the management of patients with APS both in the medical and obstetric settings. Keywords anticardiolipin antibodies; anticoagulation; antiphospholipid syndrome; b2-glycoprotein-I antibodies; lupus anticoagulant; obstetric morbidity; thrombosis Antiphospholipid syndrome (APS) is characterized by throm- bosis (venous and/or arterial) and/or pregnancy failure in association with the persistent production of a group of hetero- geneous autoantibodies known as antiphospholipid antibodies (aPL). The primary targets of aPL are phospholipid-binding proteins, although antibodies directed against phospholipids and other proteins also occur. Additional features, particularly thrombocytopenia, are variably present. In the laboratory, the usual diagnostic tests for aPL are: lupus anticoagulants (LA), which cause prolongation of phospholipid-dependent clotting assays (e.g. activated partial thromboplastin time, dilute Russell s viper venom time [DRVVT]); this effect is abolished by the addition of excess phospholipid (e.g. from platelets) anticardiolipin antibodies (aCL) of immunoglobulin G (IgG) and IgM class, which are determined by enzyme-linked immunosorbent assay (ELISA) b2-glycoprotein-I (b2-GPI) antibodies of IgG and IgM class detected by ELISA The prevalence of aPL in the form of LA or aCL is 1e5% of healthy individuals. The prevalence increases in the elderly and in those with chronic disease. APS has conventionally been divided into primary and secondary forms; the latter is associated Steven Austin MBBS FRACP FRCPA is a Locum Consultant Haematologist at St George s Hospital, London, UK. Competing interests: none declared. Hannah Cohen MD FRCP FRCPath is Consultant Haematologist at University College London Hospitals NHS Foundation Trust, London, UK. Competing interests: none declared. with chronic inflammatory conditions, mainly systemic lupus erythematosus (SLE). However, this distinction was abandoned in the Sydney classification on the basis that it is unknown whether APS and SLE are two diseases coinciding in an indi- vidual, underlying SLE offers a setting for the development of APS, or APS and SLE represent two elements of the same process. Several studies have shown that the prevalence of aPL in SLE patients is variable (15e86%). The frequency of antibody posi- tivity is likely to be around 30%, with the wide variation found in the literature explained by study variations, ethnicity, and extent of autoimmune disease activity. Up to an estimated 40% of patients with SLE and aPL will eventually develop clinical features consistent with APS, whereas under 5% of patients with APS will develop SLE. Pathogenesis Recent progress has furthered our understanding of the mecha- nism behind the development of aPL. However, a precise path- ophysiological mechanism for the clinical features of APS remains elusive. Proposed prothrombotic mechanisms, all of which can result in a prothrombotic state, include: ˇ€‹ acquired resistance to activated protein C ˇ€‹ endothelial cell activation ˇ€‹ up-regulation of tissue factor ˇ€‹ reduced fibrinolysis ˇ€‹ oxidant-mediated endothelial injury ˇ€‹ ADAMTS13 (a disintegrin-like and metalloprotease with thrombospondin type 1 motif, member 13) dysfunction. Platelet and monocyte activation may also promote throm- bosis. Additional mechanisms of displacement of annexin V and complement activation have been linked with direct cellular damage and abnormal placentation, leading to implantation failure and foetal loss. The clinical significance of any one (or more) of these hypotheses remains unclear and reflects the likely multifactorial complex nature of this condition, as is generally the case in acute thrombosis. Animal studies have suggested that the mechanism of molecular mimicry plays an important role in experimental APS. OTHER AUTOIMMUNE DISORDERS What s new? C Antibodies directed against domain I of b2-GPI correlate better with thromboembolic complications than antibodies directed against other domains of b2-GPI C Low titre antiphospholipid antibodies appear to be implicated in recurrent miscarriage C High-intensity anticoagulation is appropriate for some patients e however, substantive evidence is lacking C Low-dose aspirin is ineffective for the prevention of thrombosis in patients with asymptomatic aPL C Plasmapheresis, intravenous immunoglobulin and immuno- suppression are non-evidence-based therapies gaining popu- larity for the management of patients with resistant disease C Statins and ACE inhibitors have been shown to possess anti- inflammatory properties in vitro and may have a role in the armamentarium of APS in the future MEDICINE 38:2 101 O 2009 Elsevier Ltd. All rights reserved. Bacterial peptides homologous to b2-GPI infused into mice have been reported to induce antibodies to b2-GPI along with APS manifestations. The development of aPL is probably only one step towards the development of APS, and it is likely that other factors play a role. Such  second hits or  triggers may tip haemostasis in favour of a prothrombotic state and include infection, endothelial injury, and other non-immunological pro- coagulant factors. The patient s genetic background (in relation to candidate genes of inflammatory mediators) may also be a critical variable for the development of clinical APS manifes- tations. It is now accepted that anti-b2-GPI are the major path- ological antibodies in APS, although there is no clear consensus on how the occurrence of these antibodies is associated with the various clinical features. Systematic review of published data suggests that LA are a stronger risk factor than aCL for throm- bosis (arterial or venous) and aCL are associated more with arterial than venous thrombosis. Anti-b2-GPI may confer a stronger thrombotic risk than do aCL especially for venous thrombosis. Positivity in more than one assay and immuno- globulin subclass (IgG being more significant for thrombosis) is also important. Low-titre antibodies appear to be implicated in recurrent miscarriage. Clinical features Thrombosis APS is associated with venous and arterial thrombotic events. Among patients with venous thromboembolism, 3e17% have aCL, and 3e14% have LA. Well-established risk factors for thrombosis (e.g. pregnancy and surgery) also increase the risk of thrombosis in APS patients, as do coexistent inherited thrombophilias such as factor V Leiden. The most common venous thrombotic manifestation of APS is lower limb deep venous thrombosis, which occurs in up to 55% of patients, half of whom also have pulmonary embolism. Occasionally, venous thromboses occur in unusual sites (e.g. cerebral venous sinuses), and indeed any part of the venous system can be involved (superficial, mesenteric, portal, intracranial, retinal). Neurological features Ischaemic stroke is the most common neurological feature (>50% of CNS complications), and the commonest type of arterial thrombosis in APS. Recurrent stroke can lead to multi- infarct dementia. More subtle cognitive dysfunction has been reported to be associated with aPL and may represent micro- thrombotic change in the cerebral vasculature or a direct effect of aPL on neuronal tissue. Some studies suggest a high rate of recurrent stroke in aPL-positive patients and a younger age of onset of symptomatic events. aPL have also been linked with Sneddon s syndrome (recurrent stroke and livedo reticularis). Other reported, but unproven, neurological manifestations include migraines, seizures, chorea, transverse myelitis and a multiple sclerosis-like syndrome. Pregnancy morbidity Persistent aPL are seen in about 15% of women with recurrent miscarriage in the first or second trimester. In these women, the fetal loss rate can be up to 90% without pharmacological treatment. Late fetal death or stillbirth can also occur. Other possible obstetric complications of APS include placental insuf- ficiency, HELLP syndrome (haemolysis, elevated liver enzymes and low platelets) and severe early pre-eclampsia. Fetal and neonatal complications include intrauterine growth restriction, prematurity and, rarely, thrombosis. Catastrophic APS (CAPS) Although most patients with APS suffer thrombosis in one area at a time, they occasionally present with life-threatening acute multiple organ failure from extensive microvascular thrombosis ( thrombotic storm ). Laboratory evidence of disseminated intravascular coagulation can occur. Suggested precipitants of CAPS include infection, oral contraceptives, surgery and with- drawal of anticoagulation. The mortality rate is approximately 50%. Seronegative antiphospholipid syndrome A subset of patients has been identified who exhibit clinical manifestations of APS, without any recognized aPL on labo- ratory testing. These individuals are said to have SNAP (seronegative antiphospholipid) syndrome and anticoagulation may be warranted. Subsequent repeat aPL testing can be positive. Other clinical associations of aPL These are listed in Table 1. Diagnosis APS patients present to a wide range of specialties. Two ques- tions are of fundamental importance. OTHER AUTOIMMUNE DISORDERS Clinical associations of antiphospholipid antibodies Conditions associated with production of aPL C Systemic lupus erythematosus, rheumatoid arthritis, systemic sclerosis, Beh 'cet s disease, temporal arteritis, Sjo gren s syndrome C Infections e HIV, varicella, hepatitis C, syphilis, malaria, leprosy C Drugs e phenothiazines, procainamide, phenytoin, quinidine, hydralazine C Lymphoproliferative disease (lymphoma, paraproteinaemia) Clinical manifestations in patients with aPL C Cardiovascular e venous/arterial thromboembolic disease, valvular heart disease, sterile endocarditis with embolism C Obstetric e recurrent miscarriage, intrauterine foetal death (IUFD), stillbirth, early severe pre-eclampsia, HELLP, placental insufficiency, prematurity, intrauterine growth restriction (IUGR) C Neurological e cerebral ischaemic events, chorea, dementia, psychiatric disorders, transverse myelopathy, seizures, GuillaineBarre  syndrome, Sneddon s syndrome C Haematological e autoimmune thrombocytopenia, autoimmune haemolytic anaemia C Dermatological e livedo reticularis aPL, antiphospholipid antibodies; HELLP, haemolysis, elevated liver enzymes and low platelets. MEDICINE 38:2 102 O 2009 Elsevier Ltd. All rights reserved. Table 1 What criteria are used to make the diagnosis? Table 2 illustrates the criteria for APS based on the 2006 Sydney update of the 1999 Sapporo classification. Although useful as a clinical tool, these criteria were devised primarily for research purposes, and are tailored to a high diagnostic specificity (in practice such criteria can exclude patients with features consis- tent with APS). It is essential that persistent aPL positivity is confirmed in two separate blood samples; the criteria recom- mend an interval of at least 12 weeks between sampling. Given that thrombotic disease, pregnancy loss and transient aPL posi- tivity are common, it is important to address all other causes of thrombosis and miscarriage during the initial evaluation of sus- pected APS. A careful drug history is also needed: compounds such as phenothiazines, phenytoin and hydralazine are associ- ated with aPL positivity. Who should be tested for APS? Table 3 lists the indications. Testing for aPL should be performed in a specialized haemostasis laboratory. Samples for LA should be taken with minimal venous stasis, rapid draw and immediate anticoagulation. Platelet-poor plasma should be prepared within 1 h of blood collection. Lack of adequate sampling and preparation may result in a false-negative result for LA. Anticoagulation may interfere with laboratory diagnosis of LA; however, in certain specialized laboratories aPL testing can be performed whilst the patient is taking anticoagulant treatment. Strategies include the use of the DRVVT where the international normalized ratio (INR) is only modestly elevated (<3.0) or the use of an alternative coagulation test employing a reagent that is less sensitive to the effects of warfarin (the Taipan snake venom time). Management The Haemostasis and Thrombosis Task Force of the British Society for Haematology (BCSH) has published guidelines on the management of APS and the American College of Chest Table 3 Physicians (ACCP) guidelines include advice on the management of patients with APS and a history of pregnancy morbidity. Use of specialist referral centres is important. Decisions about the commencement and duration of anticoagulant therapy rest on the accurate diagnosis of such events, so appropriate assessment and imaging are vital. The treatment of venous thromboembo- lism must never be delayed whilst awaiting aPL test results. An evidence-based approach to management is difficult because of the limited number of randomized controlled trials involving the APS population. Asymptomatic aPL Prophylactic intervention for the prevention of primary throm- bosis in asymptomatic individuals with persistent aPL is not supported by available evidence. However, it is prudent to observe general thrombotic risk reduction measures (e.g. avoidance of smoking, oestrogen-containing oral contraception and hormone replacement therapy) and give short-term heparin prophylaxis during high-risk periods (surgery, periods of immo- bilization and hospitalization). A recent prospective, randomized clinical trial has shown that prophylactic low-dose aspirin is ineffective for the prevention of thrombosis in individuals with asymptomatic aPL. Whether primary prophylaxis with aspirin is useful for some subsets of aPL patients at particularly high risk of thrombosis, such as those with SLE or with specific patterns of aPL positivity, remains to be established. Venous thromboembolism and stroke Initial management of venous thromboembolism involves stan- dard therapy with low-molecular-weight heparin (LMWH) fol- lowed by oral anticoagulation. The duration of therapy depends on any additional risk factors, the location, severity and conse- quences of previous thrombosis, patient age and the relative risk of further thrombosis versus haemorrhage due to anti- coagulation. In other words, therapeutic decisions must be individualized. Recurrent venous thromboembolic events are managed using long-term warfarin. The optimal intensity of long- term anticoagulation must balance the risk of thrombosis against that of bleeding (1% per year, 0.25% severe haemorrhage). OTHER AUTOIMMUNE DISORDERS Indications for antiphospholipid antibody testing Consider testing in all patients with C Apparently spontaneous venous thromboembolism, especially in the young or in those with thrombosis at unusual sites C Recurrent thrombosis C Stroke, myocardial infarction and peripheral arterial occlusive events presenting at a young age (<40 years), or in selected older patients with no obvious conventional cardiovascular risk factors C Systemic lupus erythematosus or those with autoimmune disease and thrombosis C Recurrent pregnancy loss or pregnancy complications with premature birth C Unexplained thrombocytopenia C Livedo reticularis Diagnostic criteria for antiphospholipid syndrome Clinical criteria C Vascular thrombosis e one or more episodes of arterial, venous or small vessel thrombosis in any tissue or organ (confirmed by imaging or histopathology) C Recurrent pregnancy loss (1 > 10 weeks gestation, or 3 < 10 weeks gestation) or one or more premature births due to preg- nancy complications Laboratory criteria C Lupus anticoagulant in plasma on two occasions at least 12 weeks apart C Anticardiolipin antibodies of IgG and/or IgM isotype on two occasions at least 12 weeks apart C Anti-b2-GPI antibody of IgG or IgM isotype on two occasions at least 12 weeks apart Antiphospholipid syndrome is considered to be definitely present when at least one clinical criterion and one laboratory criterion are met. Ig, immunoglobulin; GPI, glycoprotein-I. Table 2 MEDICINE 38:2 103 O 2009 Elsevier Ltd. All rights reserved. Recent prospective, randomized controlled trials suggest that high-intensity warfarin is not superior to moderate-intensity. However, patients with a high risk of recurrent thrombosis were not included in these trials, and it is likely that for some patients a target INR of 2.5 is insufficient. Although such a group is difficult to define, further events occurring despite therapeutic anticoagulation should alert one to the requirement for more intensive anticoagulation with a target INR of 3.5 (3.0e4.0). In patients with arteriothrombotic stroke associated with aPL, management should involve long-term warfarin. The optimal intensity remains disputed, as prospective studies have not shown a benefit of high-intensity compared with moderate- intensity anticoagulation (but the study population was not representative of an APS population). The addition of aspirin may also be of benefit, but with increased risk of bleeding. There is interest in the use of hydroxychloroquine and statins to reduce thrombotic risk in APS, but their use in this context remains empirical. In view of the association between APS and arterial thrombotic disease, every effort should be made to correct conventional arteriopathic risk factors in all patients with APS. Pregnancy morbidity All pregnant women with APS should be managed within a specialist obstetric unit with ready access to neonatal care. Given the teratogenicity of warfarin (6.4%), which is the highest in the first trimester, warfarinization is contraindicated in early pregnancy. In addition, foetal intracerebral haemorrhage can occur at any gestation. Women on long-term oral anticoagulation must be advised that warfarin should be stopped as early as possible and before 6 weeks gestation. The recommended treatment in women who experience recurrent miscarriage (and no history of thrombosis) associated with aPL is a combination of low-dose aspirin and heparin. However, the evidence for this is driven primarily by the results of a single randomized, controlled trial in which aspirin was used as the control. While some studies suggest aspirin alone or even supportive care only, in vitro data support an adjunctive role for heparin in this situation. At our institution we administer aspirin (75 mg daily) and low-dose LMWH therapy (dalteparin 2500 units daily subcutaneously (SC) to patients with APS associated with recurrent miscarriage, and, on a pragmatic basis, dalteparin 5000 units daily SC to patients with aPL and late obstetric complications). Aspirin should be started as soon as the urine pregnancy test becomes positive (preconceptual aspirin should also be considered), and LMWH when foetal heart activity is established. The ideal duration of such therapy has not been determined, but in women with miscarriage about 25% of treated pregnancies are associated with late obstetric complications. In these patients, a thrombotic basis may be contributory, and it is reasonable to continue therapy to 38 weeks gestation. Thrombotic risk is increased during pregnancy, and women with APS with thrombotic mani- festations require careful anticoagulation, which should be individualized. Catastrophic antiphospholipid syndrome Anecdotal reports suggest that the following agents may be beneficial: anticoagulation, intravenous immunoglobulin, plasma exchange and immunosuppressive therapy, including high-dose corticosteroids, cyclophosphamide and rituximab. Conclusions In recent years there has been significant progress in the under- standing of the aetiology, pathophysiology and diagnosis of APS. However, the precise mechanism of action of aPL remains elusive and more clarification is needed, particularly with regard to the neurological manifestations and late obstetric complications. Furthermore, we lack a solid evidence base to guide optimal management. Current therapy relies on a logical approach based on available evidence. Of paramount importance is recognition of clinical scenarios that warrant investigations for aPL, taking heed of the possibility of transient aPL, and application of an appro- priate protocol for laboratory testing. Interpretation of test results must occur in concert with the clinical history, paying particular attention to any additional thrombotic risk factors. Antith- rombotic agents e aspirin, LMWH and warfarin e are the main- stay of therapy of APS. Further large-scale randomized clinical trials using these agents are needed to increase the evidence base and optimize management of APS. Novel approaches to eliminate the autoantibody, which may offer the potential for cure in the future, are under investigation. Anti-PL-7 (Anti-Threonyl-tRNA Synthetase) Antisynthetase Syndrome Clinical Manifestations in a Series of Patients From a European Multicenter Study (EUMYONET) and Review of the Literature Ane Labirua-Iturburu, MD, Albert Selva-O‚ÄôCallaghan, MD, PhD, Melinda Vincze, MD, Katalin Danko¬¥, MD, PhD, Jiri Vencovsky, MD, PhD, Benjamin Fisher, MD, MRCP, Peter Charles, FRCPath, FIBMS, Maryam Dastmalchi, MD, PhD, and Ingrid E. Lundberg, MD, PhD Abstract: Autoantibodies against several aminoacyl-transfer-RNA synthetases have been described in patients with myositis; anti-threonyltRNA synthetase (anti-PL-7) is one of the rarest. We describe the clinical and laboratory characteristics of a cohort of European anti-PL-7 patients, and compare them with previously reported cases. This multicenter study of patients positive for anti-PL-7, identified between 1984 and 2011, derives from the EUMYONET cohort. Clinical and serologic data were obtained by retrospective laboratory and medical record review, and statistical analyses were performed with chi-squared and Fisher exact tests. Eighteen patients, 15 women, were anti-PL-7 antibody positive. Median follow-up was 5.25 years (interquartile range, 2.8Y10.7 yr), and 4 patients died. All patients had myositis (12 polymyositis, 5 dermatomyositis, and 1 amyopathic dermatomyositis), 10 (55.6%) had interstitial lung disease, and 9 (50%) had pericardial effusion. Occupational exposure to organic/inorganic particles was more frequent in patients with interstitial lung disease than in the remaining patients (5 of 10 vs. 1 of 7; p = 0.152), although the difference was not significant. Concurrent autoantibodies against Ro60 and Ro52 were seen in 8 of 14 (57%) patients studied. In the literature review the most common manifestations of anti-PL-7 antisynthetase syndrome were interstitial lung disease (77%), myositis (75%), and arthritis (56%). As in other subsets of the antisynthetase syndrome, myositis and interstitial lung disease are common features of the anti-PL-7 antisynthetase syndrome. In addition, we can add pericarditis as a possible manifestation related to anti-PL-7 antibodies. (Medicine 2012;91: 206Y211) From the Internal Medicine Department (AL-I, AS-O), Vall d‚ÄôHebro¬¥n General Hospital, Universitat Auto`noma de Barcelona and Vall d‚ÄôHebro¬¥n Research Institute, Barcelona, Spain; 3rd Department of Internal Medicine (MV, KD), Division of Immunology, Medical and Health Science Center, University of Debrecen, Debrecen, Hungary; Institute of Rheumatology (JV), Charles University, Prague, Czech Republic; Kennedy Institute of Rheumatology (BF, PC), Imperial College London, London, UK; Rheumatology Unit (MD, IEL), Department of Medicine, Karolinska University Hospital, Solna, Karolinska Institutet, Stockholm, Sweden. Funding and conflicts of interest: This study was supported in part by the European Union Sixth Framework Programme (project AutoCure; LSH-018661), and the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet. JV is supported by MSM 0021620812 from Ministry of Education, Youth and Sports in the Czech Republic. The authors have no conflicts of interest to disclose. Reprints: Ane Labirua-Iturburu, MD, A1/Urquijo 21, 4 izq, 48008, Bilbao, Spain (e-mail: ane.labirua)gmail.com). Copyright * 2012 by Lippincott Williams & Wilkins ISSN: 0025-7974 DOI: 10.1097/MD.0b013e318260977c 206 www.md-journal.com Abbreviations: ANA = antinuclear antibodies, DLCO = diffusing capacity of the lung for carbon monoxide, DM = dermatomyositis, FEV1 = forced expiratory volume in 1 second, FVC = forced vital capacity, ILD = interstitial lung disease, IQR = interquartile range, PM = polymyositis. INTRODUCTION I diopathic inflammatory myopathiesVpolymyositis (PM) and dermatomyositis (DM)Vare systemic autoimmune diseases characterized by skeletal muscle inflammation, but other organs are frequently involved such as skin in dermatomyositis and lungs and heart in both polymyositis and dermatomyositis. Up to 56% of patients with myositis are positive for various autoantibodies, which can be classified as associated (present in other rheumatic disorders) or specific (positive predominantly in myositis).4 Among the myositis-specific autoantibodies group, antisynthetase antibodies are the most commonly found, and are directed against aminoacyl-transfer-RNA synthetases, a group of cytoplasmic enzymes that catalyze binding of an amino acid to its cognate tRNA, a necessary step in the formation of polypeptides. To our knowledge, 8 autoantibodies against different synthetases have been described to date: anti-Jo-1, antiPL-7, anti-PL-12, anti-EJ, anti-OJ, anti-KS, anti-YRS, and antiZo.1,5,11,14,21,24,28 The clinical profiles associated with these various antisynthetase antibodies, in particular anti-Jo-1, have been investigated previously.7,22 The main associated clinical features are myositis, interstitial lung disease (ILD), arthritis, fever, Raynaud phenomenon, and mechanic‚Äôs hands, comprising the ‚Äò‚Äòantisynthetase syndrome.‚Äô‚Äô In addition, individual autoantibody specificities may be associated with distinctive clinical features. Non-Jo-1 antisynthetase antibodies seem to be markers of hypomyopathic forms with prominent lung involvement.10,13,15 Few studies have focused on the clinical manifestations of patients with anti-PL-7, an antibody directed against threonyl-tRNA synthetase.12,25,33 We conducted the current study to investigate the clinical and laboratory profiles of patients with anti-PL-7 antibody in a large cohort of European patients, and to compare them with previously reported cases. PATIENTS AND METHODS Patients Patients positive for anti-PL-7 recorded during the period of 1984 to 2011 were identified from the laboratory databases or rheumatology clinics of the following university hospitals: Vall Medicine & Volume 91, Number 4, July 2012 Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine & Volume 91, Number 4, July 2012 d‚ÄôHebro¬¥n General Hospital, Universitat Auto`noma de Barcelona, Barcelona, Spain; Medical and Health Science Center, University of Debrecen, Debrecen, Hungary; Karolinska University Hospital, Stockholm, Sweden; Institute of Rheumatology, Charles University, Prague, Czech Republic; and Imperial College Healthcare NHS Trust, London, United Kingdom. In total 964 myositis patients were identified. Clinical data were obtained by a retrospective review of medical records. Clinical findings included the presence or absence of inflammatory myopathy, ILD, arthritis, Raynaud phenomenon, mechanic‚Äôs hands, fever, skin rash, heart involvement, history of smoking and environmental exposures related to profession, and other relevant clinical features. The diagnosis of dermatomyositis and polymyositis was based on the criteria of Bohan and Peter,2,3 and only patients with definite or probable disease were included. The Sontheimer criteria27 were used to diagnose amyopathic dermatomyositis. The diagnosis of ILD was established when any of the following conditions were present: interstitial infiltrates on chest radiography or high-resolution computed tomography (ground-glass opacities, honeycombing, fibrosis, and/or interstitial thickening) and/or a restrictive pattern on pulmonary function testing (forced vital capacity EFVC^ G80%, forced expiratory volume in 1 second EFEV1^ G70%, diffusing capacity of the lung for carbon monoxide EDLCO^ G75%), and/or positive histology. Pericardial effusion was assessed by transthoracic echocardiography. Autoantibody Analyses Myositis-specific and -associated autoantibodies were identified by line immunoassay (Myositis Profile Euroline, Euroimmun, Lu¬®beck, Germany) or RNA and protein immunoprecipitation assay. Anti-PL-7 antibodies were confirmed by at least 2 of the following techniques: ELISA, line immunoassay (Myositis Profile Euroline), or RNA and protein immunoprecipitation assay. To exclude false-positive cases, we included only patients who repeatedly tested positive for anti-PL-7. Antinuclear antibodies (ANA) and extractable nuclear antigens were assessed by ELISA. Serum samples were obtained after patients provided oral informed consent. Human Leukocyte Antigen Typing Human leukocyte antigen (HLA) class II was detected with a sequence-specific primer and sequence-specific oligonucleotide polymerase chain reaction technique. Ethics The institutional review boards of all the participating centers approved the study. Literature Review We searched the English-language literature in the PubMed database (National Library of Medicine, Bethesda, MD) for related articles published up to March 2011, using the following key words: ‚Äò‚ÄòPL-7 antibody,‚Äô‚Äô ‚Äò‚Äòanti-PL-7,‚Äô‚Äô ‚Äò‚Äòanti-threonyl-tRNA synthetase,‚Äô‚Äô and ‚Äò‚Äòantisynthetase syndrome.‚Äô‚Äô Statistical Analyses Qualitative data are presented as numbers and percentage, and quantitative data as the median and interquartile range (IQR). Association between the presence of anti-PL-7 antibodies and qualitative variables was assessed using the chi-squared and Fisher exact tests. All statistical analyses were performed with SPSS 13.0 software (SPSS, Chicago, IL). Significance was set at a p value of less than 0.05. * 2012 Lippincott Williams & Wilkins Anti-PL-7 Antisynthetase Syndrome RESULTS Clinical Features Eighteen anti-PL-7-positive patients were identified out of 964 patients tested from the contributing centers (1.87%). The median age at diagnosis was 52.5 years (IQR, 40Y58.8 yr), and there was a predominance of women (15 patients, 83%). Eight (44.4%) patients were current or former smokers. PM was diagnosed in 12 (66.7%) patients, DM in 5 (27.8%), and amyopathic DM in 1 (5.6%) patient. Two patients had overlap syndromes: 1 patient with DM met the criteria for rheumatoid arthritis, with positive testing for rheumatoid factor and anticyclic citrullinated peptide, and the other had PM and Sjo¬®gren syndrome. Clinical features of the 18 patients are summarized in Table 1. Twelve (66.7%) patients had arthritis. Ten (55.6%) patients had ILD. Fever attributed to the disease was present in 10 (55.6%) patients, and Raynaud phenomenon in 11 (61.1%). Nine (50%) patients had pericardial effusion with no apparent underlying cause (for example, heart failure). Two of them developed massive effusion with pericardial tamponade: 1 resolved with medical treatment (glucocorticoids), and the other required pericardiocentesis. Mechanic‚Äôs hands were reported in 5 (27.8%) patients. Only 1 patient had a malignant disease, breast cancer, which was diagnosed 4 years after the PM diagnosis. The onset of ILD was before the start of myositis in 2 cases, at the same time as myositis in 7, and after myositis in 1. In all 10 patients with ILD, pulmonary involvement was confirmed by high-resolution computed tomography, which showed various radiologic patterns. Signs of fibrosis were documented in 9 of the 10 patients, interstitial thickening in 5, honeycombing in 4, and ground-glass opacities in 4. Pulmonary function tests were carried out in 5 cases and showed a restrictive pattern with a median FVC at diagnosis of 61% (IQR, 57%Y75.5%) and FEV1 of 69.5% (IQR, 51.4%Y74.3%). Transbronchial biopsy in 1 patient and autopsy findings in another disclosed alveolitis in both cases. One patient required orotracheal intubation due to respiratory failure caused by severe ILD, with later recovery. Six patients in the cohort had been exposed to various organic/inorganic particles at work, including 2 factory workers, a plumber, a farmer, a cleaning worker, and a miner (uranium mines). All but 1 (a factory worker) had underlying ILD, which appeared before or at the same time as myositis. Nonetheless, there was no significant association between working exposure and ILD (5 of 10 with ILD vs. 1 of the 7 remaining patients, p = 0.152). The median follow-up was 5.25 years (IQR, 2.8Y10.7 yr), and 4 patients died. The cause of death was attributed to progression of ILD in 2 patients (1 complicated with Pneumocystis jiroveci infection), acute myocardial infarction in 1 patient, and sepsis secondary to Salmonella species in 1 patient. The following treatment options were used either alone or in combination: glucocorticoids (100%), azathioprine (33%), cyclophosphamide (33%), methotrexate (22%), immunoglobulins (16%), and cyclosporine (5%). Four patients were treated with glucocorticoids alone, 5 patients with glucocorticoids plus 1 immunosuppressive drug, 7 patients with glucocorticoids and 2 immunosuppressive drugs, and 2 patients with 3 or more immunosuppressive drugs (see Table 1). Antibody Profile and HLA Seven patients were ANA-positive, with titers ranging from 1/320 to 1/2560. Anti-Ro/SSA antibody was found in 4 patients, and anti-Ro52 antibody in 5 patients. Anti-La/SSB was found in 2 patients, and anti-PM-Scl, anti-RNP, and anticardiolipin www.md-journal.com Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 207 Medicine Labirua-Iturburu et al & Volume 91, Number 4, July 2012 TABLE 1. Clinical and Serologic Features of 18 European Patients With Antisynthetase (PL-7) Syndrome Patient Age/Sex Smoking (yr) Status Job 1 65/F No Shop worker 2 3 4 5 52/F 56/F 37/F 45/F Yes No No No 6 7 8 9 10 11 12 13 40/F 55/F 32/F 40/F 38/F 55/F 53/F 62/M Yes No No No No No Yes Yes 14 61/M Yes 15 16 46/M 58/F 17 18 69/F 40/F Myositis AR Ray MH Fever ILD PD Ab DM Yes Yes Yes No Administration Housewife Cleaning worker Administration Administration Administration Waiter Toy factory Farmer Car factory Plumber DM ADM DM PM Yes No Yes No No No Yes Yes Yes Yes No No No No Yes Yes PM PM PM PM PM PM PM PM No Yes Yes Yes Yes No Yes No Yes Yes Yes No Yes Yes Yes Yes No No No No Yes No No Yes No No No Yes No No Yes Yes PM No No No Yes Yes Yes Miner (uranium) Song writer Administration No No PM-Scl No No No No RNP No No Ro, La No Yes Ro, La Yes No ANA, ACA Yes Yes Yes Yes ANA, Ro52, CL Yes No Ro52 PM DM No No Yes Yes No No Yes Yes Yes No No NA RF, CCP No Yes Administration Administration PM DM Yes Yes Yes No No No Yes Yes Therapy Yes Yes ANA, RF GC + CP+ MTX+AZA No Yes Ro, Ro52 GC+AZA+IG Yes Yes ANA, Ro52 GC+CP+AZA No Yes ANA, Ro GC+CYS+IG Yes Yes ANA GC+AZA Death Yes; sepsis Yes; MI GC GC GC GC+MTX GC+CP+AZA GC+CP GC+MTX GC+MTX GC GC+AZA+MF GC+AZA+ MTX Yes Yes GC+CP+MTX Yes No ANA, Ro52 GC+CP+AZA+ MTX+IG Yes; ILD Yes; ILD Abbreviations: Ab = antibodies, ADM = amyopathic dermatomyositis, AR = arthritis, AZA = azathioprine, CCP = cyclic citrullinated peptide, CL = cardiolipin, CP = cyclophosphamide, CYS = cyclosporin, GC = glucocorticoids, IG = intravenous immunoglobulin, MF = mycophenolate, MH = mechanic‚Äôs hands, MI = myocardial infarction, MTX = methotrexate, NA = not available, PD = pericardial disease, RF = rheumatoid factor, Ray = Raynaud phenomenon. antibodies were found in 1 patient each. Rheumatoid factor was positive in 2 cases, 1 of whom additionally had anticyclic citrullinated peptide. HLA typing was performed in 13 patients. TABLE 2. Immunogenetic Profile (HLA Class II Alleles) of 18 Patients With Antisynthetase (PL-7) Syndrome Patient HLA-DRB1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 208 HLA-DRB1*03,04 HLA-DRB1*04,15 HLA-DRB1*12 HLA-DRB1*03 HLA-DRB1*07,*07 HLA-DRB1*04,*11 HLA-DRB1*03;*04 HLA-DRB1*11 HLA-DRB1*04,*07 HLA-DRB1*03,*11 HLA-DRB1*01,*11 HLA-DRB1*03,*04 HLA-DRB1*08,*15 www.md-journal.com DRB haplotype is shown in Table 2. The most common haplotypes were DRB1*04 in 6 patients and DRB1*03 in 5 patients. Only 2 of the patients with ILD showed the DRB1*03 allele. Literature Review We identified 54 additional cases of anti-PL-7 antisynthetase syndrome in the literature. The main clinical manifestations in the overall group of 72 patients (present cohort and reported cases) were ILD in 56 (77.8%) patients and myositis in 54 (75%). The prevalence of other signs and symptoms of anti-PL-7 antisynthetase syndrome is summarized in Table 3. DISCUSSION Antisynthetase syndrome, to our knowledge first described in 1990, is recognized by its characteristic clinical manifestations associated with the presence of an anti-tRNA synthetase antibody.20 Anti-Jo-1 is the most frequent, followed by anti-PL-12 and anti-PL-7. It has been suggested that in addition to a common core syndrome, different clinical features may be associated with different antisynthetase antibodies, but the scarcity of published studies, most of which are case reports, precludes any definitive conclusions in this regard.16 In the present study, we analyzed the clinical and laboratory features of a series of anti-PL-7-positive patients from 5 European countries. In common with other reported antisynthetase syndromes, myositis and ILD were the most frequent manifestations of anti-PL-7-associated syndrome, and these common features were confirmed in a review of all cases published * 2012 Lippincott Williams & Wilkins Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine & Volume 91, Number 4, July 2012 Anti-PL-7 Antisynthetase Syndrome TABLE 3. Main Characteristics of Patients With Antisynthetase (PL-7) Syndrome, Previous and Present Reports Anti-PL-7 Report (First Author, Ref.) 12 Hervier Sato25 Fischer9 Yamasaki33 Mathews21 Dugar7 Marguerie20 Targoff 29 Lega17 Troyanov31 Labirua-Iturburu (PR) Total, no. (%) No. of Patients Myositis (No.) ILD (No.) Arth (No.) Ray (No.) MH (No.) Fever (No.) PE (No.) 12 7 7 6 5 4 4 4 3 2 18 72 6 6 3 4 4 3 3 4 1 2 18 54 (75) 12 7 7 6 1 2 3 3 3 2 10 56 (77.8) 4 6 3 3 1 4 4 2 1 12 40 (55.6) 1 4 2 2 2 4 1 1 1 11 29 (40.3) 4 2 1 1 5 13 (18) 8 4 1 10 22 (30.5) 4 1 1 1 9 16 (22.2) Abbreviations: Arth = arthritis/arthralgia, MH = mechanic‚Äôs hands, PE = pericardial effusion, PR = present report, Ray = Raynaud phenomenon. to date in the literature. An unexpected finding in the current cohort was pericardial effusion in half of our cases. Due to the low patient numbers, few studies have focused on the clinical and laboratory characteristics of patients with antiPL-7 antibodies. The most complete are 2 studies performed in Japan, comprising 7 and 6 patients, respectively,25,33 and a 2011 report by Hervier et al12 in France with 12 patients. In these studies, ILD was present in 100% of cases, while myositis appeared in 50%Y86%. In the present series, 6 of 18 patients did not have diagnosed ILD, and there was a much higher prevalence of myositis (all patients had idiopathic inflammatory myopathies except 1, who had amyopathic dermatomyositis). These results differ to some degree from those of previous studies, in which the prevalence of myositis in patients with non-Jo-1 antisynthetase autoantibodies was lower than in patients with anti-Jo1.10,15 However, a pooled analysis of our findings together with those from all published case reports and series in the literature7,9,12,17,20,21,25,29,31,33 (72 anti-PL-7-positive patients) showed that the frequency of ILD and myositis (77.8% and 75%, respectively; see Table 3) does not differ greatly from the clinical findings in patients with anti-Jo-1. One explanation for the discrepancies between individual studies could be the potential selection bias involved in the study of all rare diseases. In small series, some clinical manifestations might be overrepresented; hence to try to overcome these problems, we included a review of all cases published in the literature, in addition to our own cohort, and analyzed the pooled data set. Other possible reasons for the dissimilar results might be the racial/ethnic or genetic background of the patients, a different referral pattern (pulmonary, rheumatology, or internal medicine specialists), and the fact that in many cases, only patients with clinically evident myositis were tested for antisynthetase antibodies. Little is known about the etiopathogenesis of the antisynthetase syndrome, or for that matter, of myopathies in general. According to some authors, an unknown trigger (for example, viral infection) might enter the respiratory tract and lead to a conformational modification of aminoacyl-tRNA synthetase in the pulmonary alveoli. This could result in production of autoantibodies against the synthetase enzyme; this immune response (antibodies and specific T cells) would spread to other internal organs such as muscle and joints.18,19 An interesting factor that should be investigated in well-designed studies is the contribution of organic/inorganic dust inhalation at work, which could be * 2012 Lippincott Williams & Wilkins another etiologic trigger of autoimmunity. There has been some evidence supporting the idea of environmental exposure to antigens as an initial step in the pathogenesis of antisynthetase syndrome.30 In the current series, 6 patients had been occupationally exposed to different organic/inorganic particles and all but 1 had ILD. Nonetheless, statistical significance was not achieved for this factor, probably because of the small number of patients studied. It is noteworthy that 9 of 18 patients in the current series had pericardial effusion. In the study by Hervier et al,12 4 of 12 (33%) patients with anti-PL-7 had pericardial effusion, and 3 other single cases have been reported in different studies.17,21,33 Pericarditis is, compared to systemic lupus erythematosus, infrequent in patients with myositis. The reason patients with anti-PL-7 develop pericardial disease is unknown. It can be speculated that anti-PL-7 antibodies act against certain antigens in pericardial tissue, and thereby lead to pericardial inflammation. Nevertheless, some cases of pericarditis have also been reported in patients with anti-Jo-1.6,23 In a study and a review of literature by Schmidt et al,26 they found a prevalence of 18% of pericarditis in a total of 231 anti-Jo-1 patients, and we cannot exclude that pericarditis is underestimated in myositis patients because of low physical function due to muscle and lung involvement. Other relevant observations can be drawn from this study. First, it seems that clinical heterogeneity is a feature of the antisynthetase syndrome, irrespective of the autoantibody implicated. The results from the pooled analysis indicate that although antisynthetase syndrome can present with a single clinical manifestation (for example, myositis, ILD, Raynaud phenomenon, arthritis, mechanic‚Äôs hands), the most common features are myositis, ILD, and arthritis. The frequency of these symptoms is similar to that seen in patients with anti-Jo-1 antibodies. As for the prognosis of antisynthetase syndrome, which is currently poorly defined, ILD seems to be the major cause of death. It has been suggested that ILD associated with antisynthetase syndrome may have a better prognosis than antibody-negative ILD,8 but this has yet to be confirmed. As was seen in the current study and 1 previously reported study,12 ILD in patients with antisynthetase syndrome can be aggressive and refractory to treatment. It has been reported that coexisting Jo-Ro antibodies may predict more severe ILD,32 but only 1 of our patients who died of ILD was antiRo52-positive. Further study of lung disease in larger antisynthetase syndrome cohorts is needed to elucidate these findings. www.md-journal.com Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 209 Medicine Labirua-Iturburu et al One limitation of the present study is its retrospective nature and the fact that complete data sets were not available for all patients. Not all patients underwent pulmonary function tests or high-resolution computed tomography; hence, asymptomatic ILD could be more prevalent in our cohort than our results suggest. In addition, a heterogeneous referral pattern could have led to both over- and underdiagnosis of some clinical features. Lastly, the small sample size due to the relative rarity of the syndrome probably contributed to a lack of statistical power in some of the comparisons. In conclusion, our study of a European cohort of anti-PL-7 patients identified myositis as the most frequent clinical manifestation, with ILD being less prevalent than has been previously reported. Nevertheless, when these manifestations were evaluated in all the cases reported to date, including those from the present study, their prevalence did not vary greatly from the rates reported for other antisynthetase antibody syndromes. Occupational exposure should be taken into account in these patients, because environmental antigens could contribute to triggering the immune response and the development of the disease. The unexpected high percentage of patients with pericardial effusion in our cohort is a novel finding and needs to be confirmed in future series. It may be a characteristic of anti-PL-7 antisynthetase syndrome, which could warrant a recommendation of routine transthoracic echocardiography in patients presenting with this serotype. REFERENCES 1. Betteridge Z, Gunawardena H, North J, Slinn J, McHugh N. Anti-synthetase syndrome: a new autoantibody to phenylalanyl transfer RNA synthetase (anti-Zo) associated with polymyositis and interstitial pneumonia. Rheumatology (Oxford). 2007;46:1005Y1008. 2. Bohan A, Peter JB. Polymyositis and dermatomyositis (First of two parts). N Engl J Med. 1975;292:344Y347. 3. Bohan A, Peter JB. Polymyositis and dermatomyositis (Second of two parts). N Engl J Med. 1975;292:403Y407. 4. Brouwer R, Hengstman GJD, Vree Egberts W, Ehrfeld H, Bozic B, Ghirardello A, Gr√∏ndal G, Hietarinta M, Isenberg D, Kalden JR, Lundberg I, Moutsopoulos H, Roux-Lombard P, Vencovsky J, Wikman A, Seelig HP, van Engelen BG, van Venrooij WJ. Autoantibody profiles in the sera of European patients with myositis. Ann Rheum Dis. 2001;60:116Y123. & Volume 91, Number 4, July 2012 11. Hashish L, Trieu EP, Sadanandan P. Identification of autoantibodies to tyrosyl-tRNA synthetase in dermatomyositis with features consistent with antisynthetase syndrome Eabstract^. Arthritis Rheum. 2005;52:S312. 12. Hervier B, Uzunhan Y, Hachulla E, Benveniste O, Nunes H, Delaval P, Musset L, Dubucquoi S, Wallaert B, Hamidou M. Antisynthetase syndrome positive for anti-threonyl-tRNA synthetase (anti-PL7) antibodies. Eur Respir J. 2011;37:714Y717. 13. Hervier B, Wallaert B, Hachulla E, Adoue D, Lauque D, Audrain M, Camara B, Fournie B, Couret B, Hatron PY, Dubucquoi S, Hamidou M. Clinical manifestations of antisynthetase syndrome positive for anti-alanyl-tRNA synthetase (anti-PL12) antibodies: a retrospective study of 17 cases. Rheumatology (Oxford). 2010;49:972Y976. 14. Hirakata M, Suwa A, Nagai S, Kron MA, Tireu EP, Mimori T, Akizuki M, Targoff IN. Anti-KS: identification of autoantibodies to asparaginyl-transfer RNA synthetase associated with interstitial lung disease. J Immunol. 1999;162;2315Y2320. 15. Kalluri M, Sahn SA, Oddis CV, Gharib SL, Christopher-Stine L, Danoff SK, Casciola-Rosen L, Hong G, Dellaripa PF, Highland KB. Clinical profile of anti-PL12 autoantibody: cohort study and review of the literature. Chest. 2009;135:1550Y1556. 16. Labirua A, Lundberg IE. Interstitial lung disease and idiopathic inflammatory myopathies: progress and pitfalls. Curr Opin Rheumatol. 2010;22:633Y638. 17. Lega JC, Cotin V, Fabien N, Thivolet-Bejui F, Cordier JF. Interstitial lung disease associated with anti-PM/Scl or anti-aminoacyl-tRNA synthetase autoantibodies: a similar condition? J Rheumatol. 2010;37:1000Y1009. 18. Levine SM, Raben N, Xie D, Askin FB, Tuder R, Mullins M, Rosen A, Casciola-Rosen LA. A novel conformation of histidyl-transfer RNA synthetase in the lung: the target tissue in Jo-1 autoantibody-associated myositis. Arthritis Rheum. 2007;56:2729Y2739. 19. Lundberg IE, Grundtman C. Developments in the scientific and clinical understanding of inflammatory myopathies. Arthritis Res Ther. 2008;10:220. 20. Marguerie C, Bunn CC, Beynon HL, Bernstein RM, Hughes JM, So AK, Walport MJ. Polymyositis, pulmonary fibrosis and autoantibodies to aminoacyl-tRNA synthetase enzymes. Q J Med. 1990;77:1019Y1038. 21. Mathews MB, Reichlin M, Hughes GRV, Bernstein R. Anti-threonyl-tRNA synthetase, a second myositis-related autoantibody. J Exp Med. 1984;160:420Y434. 5. Bunn CC, Bernstein RM, Mathews AM. Autoantibodies against alanyl-tRNA synthetase and tRNA-ala coexist and are associated with myositis. J Exp Med. 1986;163:1281Y1291. 22. Mileti LM, Strek ME, Niewold TB, Curran JJ, Sweiss NJ. Clinical characteristics of patients with anti-Jo-1 antibodies. J Clin Rheumatol. 2009;15:254Y255. 6. Derk CT, Sandorfi N, Curtis MT. A case of anti-Jo1 myositis with pleural effusions and pericardial tamponade developing after exposure to a fermented Kombucha beverage. Clin Rheumatol. 2004;23:355Y357. 23. Mogulkoc N, Kabasakal Y, Ekren PK, Bishop PW. An unusual presentation of anti-Jo-1 syndrome, mimicking lung metastases, with massive pleural and pericardial effusions. J Clin Rheumatol. 2006;12:90Y92. 7. Dugar M, Cox S, Limaye V, Blumbergs P, Roberts-Thomson PJ. Clinical heterogeneity and prognostic features of South Australian patients with anti-synthetase autoantibodies. Intern Med J. 2011;41:674Y679. Epub 2010 Jan 4. 24. Nishikai M, Reichlin M. Heterogeneity of precipitating antibodies in polymyositis and dermatomyositis. Characterization of the Jo-1 antibody system. Arthritis Rheum. 1980;23:881Y888. 8. Fathi M, Vikgren J, Boijsen M, Tylen U, Jorfeldt L, Tornling G, Lundberg IE. Interstitial lung disease in polymyositis and dermatomyositis: longitudinal evaluation by pulmonary function and radiology. Arthritis Rheum. 2008;59:677Y685. 25. Sato S, Hirakata M, Kuwana M, Nakamura K, Suwa A, Inada S, Mimori T, Ikeda Y. Clinical characteristics of Japanese patients with anti-PL-7 (antithreonyl-tRNA synthetase) autoantibodies. Clin Exp Rheumatol. 2005;23:609Y615. 9. Fischer A, Swigris JJ, du Bois R, Lynch DA, Downey GP, Cosgrove GP, Frankel SK, Fernandez-Perez ER, Gillis JZ, Brown KK. Anti-synthetase syndrome in ANA and anti-Jo-1 negative patients presenting with idiopathic interstitial pneumonia. Respir Med. 2009;103:1719Y1724. 26. Schmidt WA, Wetzel W, Friedlander R, Lange R, Sorensen HF, Lichey HJ, Genth E, Mierau R, Gromnica-Ihle E. Clinical and serological aspects of patients with anti-Jo-1 antibodiesVan evolving spectrum of disease manifestations. Clin Rheumatol. 2000;19: 371Y377. 10. Friedman AW, Targoff IN, Arnett FC. Interstitial lung disease with autoantibodies against aminoacyl-tRNA synthetases in the absence of clinically apparent myositis. Semin Arthritis Rheum. 1996;26:459Y467. 210 www.md-journal.com 27. Sontheimer RD. Cutaneous features of classic dermatomyositis and amyopathic dermatomyositis. Curr Opin Rheumatol. 1999;11:475Y482. * 2012 Lippincott Williams & Wilkins Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Medicine & Volume 91, Number 4, July 2012 28. Targoff IN. Autoantibodies to aminoacyl-transfer RNA synthetases for isoleucine and glycine. Two additional synthetases are antigenic in myositis. J Immunol. 1990;144:1737Y1743. 29. Targoff IN, Arnett FC, Reichlin M. Antibody to threonyl-transfer RNA synthetase in myositis sera. Arthritis Rheum. 1988;31:515Y524. 30. Tillie-Leblond I, Wislez M, Valeyre D, Crestani B, Rabbat A, Israel-Biet D, Humbert M, Couderc LJ, Wallaert B, Cadranel J. Interstitial lung disease and anti-Jo-1 antibodies: differences between acute and gradual onset. Thorax. 2008;63:53Y59. 31. Troyanov V, Targoff IN, Tremblay JL, Goulet JR, Raymond Y, Senecal JL. Novel classification of idiopathic inflammatory * 2012 Lippincott Williams & Wilkins Anti-PL-7 Antisynthetase Syndrome myopathies based on overlap syndrome features and autoantibodies: analysis of 100 French Canadian patients. Medicine (Baltimore). 2005;84:231Y249. 32. Vancsa A, Csipo I, Nemeth J, Devenyi K, Gergely L, Danko K. Characteristics of interstitial lung disease in SS-A positive/Jo-1 positive inflammatory myopathy patients. Rheumatol Int. 2009;29:989Y994. 33. Yamasaki Y, Yamada H, Nozaki T, Akaogi J, Nichols C, Lyons R, Loy AC, Chan EK, Reeves WH, Satoh M. Unusually high frequency of autoantibodies to PL-7 associated with milder muscle disease in Japanese patients with polymyositis/dermatomyositis. Arthritis Rheum. 2006;54:2004Y2009. www.md-journal.com Copyright ¬© 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 211 ˇ˛Antipsychotics Kalliopi Vallianatou Abstract First generation (typical) and second generation (atypical) antipsychotics are the mainstay of psychosis treatment. Most antipsychotics block dopa- mine D2 receptors. Greater affinity for serotonin 5-HT2 receptors over D2 and fast dissociation from the D2 receptor are the two competing theories for the activity of atypical antipsychotics. Movement disorders, hyperpro- lactinaemia, cardiac and metabolic disorders are common antipsychotic adverse effects. Antipsychotic long-acting injections are also available in addition to oral formulations. Keywords antipsychotic; dopamine; psychosis; serotonin Antipsychotic drugs reduce the severity of psychotic symptoms and prevent relapse in schizophrenia1 and other psychotic disorders (see Medicine 2012; 40(11): 586e590). They are also used in mania, depression and delirium, amongst other condi- tions. There are two classes: older typical or conventional or first generation antipsychotics (FGAs); and newer atypical or second generation antipsychotics (SGAs). Basic principles of prescribing antipsychotics for psychotic disorders2e4  For first-episode schizophrenia, offer an oral antipsychotic at the lowest recommended therapeutic dosage. Agree the choice with the patient.  Increase the dosage after 2 weeks if the response is poor.  Aim for the lowest effective dosage, as many adverse effects are dose-related.  Prescribing  as required antipsychotics for a calming effect increases the risk of the patient exceeding the maximum dose, resulting in a greater burden of adverse effects including oversedation or cardiac disorders and greater risk of drug interactions.  Antipsychotic combinations should be avoided, apart from during short periods of switching between two antipsy- chotics, and considered only in consultation with a psychiatrist due to increased risk of QTc prolongation.  Continue treatment with the antipsychotic for at least 1e2 years, or longer for patients with multiple episodes. First generation antipsychotics (FGAs) FGAs are those associated with acute extrapyramidal symptoms (EPS) even at therapeutic dosages.1,5 FGAs share some class- specific adverse effects but differ in both chemical structure and receptor binding.5 Commonly prescribed FGAs in the UK Kalliopi Vallianatou MPharm Senior Clinical Pharmacist, South London and the Maudsley NHS Foundation Trust, Maudsley Hospital, London, UK. Conflicts of interest: none declared. include chlorpromazine, trifluoperazine, haloperidol, flu- pentixol, zuclopenthixol, pipotiazine and fluphenazine. Second generation antipsychotics (SGAs) SGAs are characterized by lower incidence of EPS and, with some exceptions (amisulpride, risperidone), a lack of sustained increase in plasma prolactin concentration (hyperprolactinaemia).6 The defining feature of atypicality for this group of drugs has not been concluded. Commonly prescribed SGAs in the UK include risper- idone, olanzapine, quetiapine, aripiprazole, amisulpride, sulpir- ide, clozapine, paliperidone and asenapine. Mode of action Most antipsychotics act as antagonists at D2 receptors in the mes- olimbic pathway (to reduce positive symptoms) or in the meso- cortical pathway (to reduce negative symptoms), but D2 receptor blockade in the striatum (causing extrapyramidal symptoms) and tuberoinfundibular pathway (causing hyperprolactinaemia) can also occur.1 D2 occupancy for antipsychotic effect may lie within 65%e70%, whereas above 72% adverse effects may emerge.7 Some antipsychotics also block serotonergic 5-HT2, a1-adre- noceptors, H1-histamine and muscarinic receptors.5 Some SGAs show lower affinity for D2 receptors, which is thought to account for decreased EPS and improvement in affective and cognitive effects.6 Unlike most other antipsychotics, aripiprazole is a partial D2 and 5-HT1a agonist. Aripiprazole binds to D2 recep- tors and both blocks and stimulates them.8 Efficacy Olanzapine, risperidone and amisulpride may be modestly more efficacious than FGAs.3,4 Olanzapine is second only to clozapine in its efficacy.9 About 30% of schizophrenia patients will be treatment resistant3 and more than half of these may respond to clozapine. Antipsychotic depots or long-acting injections (LAIs) also exist and are commonly prescribed where adherence to oral medication is problematic.10 Clozapine is indicated for use after the failure of two prior antipsychotics, including an SGA. It is associated with an increased risk of potentially fatal neutropenia or agranulocytosis and thus regular haematological monitoring is required. Trough clozapine plasma concentration should be checked following completion of dose titration, 2e3 days after any dose change or following a change in smoking habit (target 0.35e0.5 mg/litre). If concentrations over 0.5 mg/litre are found then seizure prophy- laxis with sodium valproate or lamotrigine should be considered.4 Adverse effects Antipsychotics are associated with a wide range of adverse effects, such as movement disorders, hyperprolactinaemia, postural hypotension, prolongation of the QTc interval and the metabolic syndrome (Table 1). The last of these comprises obesity, dyslipidaemia, diabetes and hypertension and has been a concern with the use of SGAs.11 Consequently, baseline tests should be carried out before an antipsychotic is prescribed and after treatment has started. Tests include full blood count, renal function, lipids, plasma glucose, liver function tests, creatine kinase, prolactin, ECG, blood pressure, pulse and weight.4 Full TREATMENT STRATEGY AND PSYCHOPHARMACOLOGY MEDICINE 40:12 676 ! 2012 Elsevier Ltd. All rights reserved. TREATMENT STRATEGY AND PSYCHOPHARMACOLOGY Adverse effects of antipsychotics4,11 Adverse effect Extrapyramidal symptoms4,11 Antipsychotics More common with FGAs. Incidence is higher with haloperidol, fluphenazine, trifluoperazine, perphenazine. Dystonias are reported for several SGAs but incidence is low. Incidence rates as for dystonia. Rates of akathisia are about 25% less for SGAs than FGAs. More likely with aripiprazole than other SGAs. More common in the elderly and in those with acute EPS at start of treatment. Risk is higher with FGAs than SGAs.4,11 Common with clozapine, risperidone, quetiapine.4 Is usually plasma drug concentration-dependent. Common with haloperidol. Moderate effect with amisulpride, quetiapine, chlorpromazine. Common with clozapine. Risk is highest with clozapine and olanzapine. Moderate with quetiapine and risperidone. Low with high-potency FGAs, aripiprazole and amisulpride.4 Common with FGAs and some SGAs including risperidone, paliperidone and amisulpride.4 Not seen with aripiprazole (which lowers plasma prolactin concentration), clozapine and quetiapine. More common with some FGAs and clozapine. Dystoniaa Pseudo-parkinsonism Akathisiab Tardive dyskinesiac Cardiac Postural hypotension QTc prolongation Tachycardia Other Metabolic Hyperprolactinaemia Anticholinergic Muscle spasm (e.g. eyes rolling, head and neck) Resting tremor, rigidity, bradykinesia, masked faces Restlessness, nervousness, compulsion to keep moving Lip-smacking, pill rolling Drop in blood pressure on standing, light headedness, blurred vision (QTc normal limits are: men <440 ms, women <470 ms) Rapid heart rate (often >100 bpm) Weight gain, waist circumference >35 inches for women or >40 inches for men, dyslipidaemia, diabetes Galactorrhoea, gynaecomastia, amenorrhoea, loss of libido, sexual dysfunction, reduction in bone mineral density Dry mouth, urinary retention, constipation and sedation a b c For treatment anticholinergic drugs are given (PO/IM/IV depending on severity). Not responsive to anticholinergics. Dose reduction, b-blockers or benzodiazepines may help. Switch to an SGA; stop any anticholinergics as they may worsen tardive dyskinesia. Table 1 blood counts are required for clozapine treatment, which follow a strict monitoring protocol. A rare, potentially fatal, reaction to antipsychotics is the neuroleptic malignant syndrome (NMS), which presents with fever, rigidity, confusion, autonomic instability, tachycardia, abnormal blood pressure, elevated creatine kinase (CK), leuco- cytosis and altered liver function tests.4 NMS constitutes a medical emergency and patients may require intensive care treatment.4 Drug interactions12,13  Clozapine: plasma levels are increased by paroxetine, cit- alopram, ciprofloxacin, erythromycin, caffeine and ritonavir and decreased by smoking and carbamazepine. Concomi- tant use of clozapine with carbamazepine should be avoi- ded due to increased risk of bone marrow suppression.12  Increased risk of ventricular arrhythmias when antipsy- chotics that prolong QTc are given with amiodarone.  Increased risk of bradycardia and respiratory depression when intramuscular (IM) olanzapine is given with IM benzodiazepines.  Antipsychotics antagonize the anticonvulsant effect of antiepileptics (by lowering the seizure threshold).  Plasma levels of quetiapine, olanzapine, risperidone and aripiprazole are reduced by carbamazepine. A Antipsychotics in vulnerable groups Pregnancy3,4 C In unplanned pregnancy reassure the patient and discuss the risks of stopping or switching medication or leaving the condition untreated. C Patients on antipsychotics considering future pregnancy should be informed that hyperprolactinaemia may reduce the chances of pregnancy. C Most data are for low-dose FGAs (e.g. chlorpromazine, halo- peridol, trifluoperazine). C If the patient is taking clozapine or olanzapine, monitor for gestational diabetes. C Use the lowest effective dose. The activity of some drug- metabolizing enzymes may be changed leading to higher or lower plasma drug concentration. C Avoid long-acting injections and anticholinergic drugs. MEDICINE 40:12 677 ! 2012 Elsevier Ltd. All rights reserved. Hepatic impairment4 C Start with low doses and increase slowly, increase dose interval, monitor LFTs weekly. C Sulpiride and amisulpride e no dose adjustment required if normal renal function. C Avoid chlorpromazine (hepatotoxic) and long-acting injections C Haloperidol may be preferred although it has a  caution in liver disease. Renal impairment4,14 C Start with small doses as increased CNS sensitivity and EPS in renal impairment. C Avoid long-acting injections and QTc-prolonging drugs. C Avoid sulpiride and amisulpride as they are primarily renally excreted. C Haloperidol (e.g. 2e6 mg/day) or olanzapine (5 mg/day) may be suggested. MEDICINE 40:12 678 ! 2012 Elsevier Ltd. All rights reserved. TREATMENT STRATEGY AND PSYCHOPHARMACOLOGY ˇ˛Antithrombotic management of patients with prosthetic heart valves: current evidence and future trends Jack C J Sun, Michael J Davidson, Andre Lamy, John W Eikelboom Over 4 million people worldwide have received a prosthetic heart valve, and an estimated 300000 valves are being implanted every year. Prosthetic heart valves improve quality of life and survival of patients with severe valvular heart disease, but the need for antithrombotic therapy to prevent thrombotic complications in valve recipients poses challenges for clinicians and patients. Here, we review antithrombotic therapies for patients with prosthetic heart valves and management of thromboembolic complications. Advances in antithrombotic therapy and valve technologies are likely to improve the management of patients with prosthetic heart valves in developed countries, but the most important unmet need and potential for benefit from these new therapies is in developing countries where a massive and rapidly increasing burden of valvular heart disease exists. Lancet 2009; 374: 565 76 Division of Cardiac Surgery, McMasterUniversity, Hamilton, ON, Canada (J C J Sun MD, A Lamy MD); Division of Cardiac Surgery, Harvard Medical School, Boston, MA, USA (M J Davidson MD); and Department of Medicine, McMaster University, Hamilton, ON, Canada (J W Eikelboom MD) Correspondence to: Dr Jack C J Sun, Division of Cardiac Surgery, Hamilton General Hospital, 237 Barton Street East, Hamilton, ON L8L 2X2, Canada sunjc2@mcmaster.ca Introduction Valvular heart disease affects more than 100 million people worldwide and is a growing problem because of the high incidence of rheumatic heart disease in developing countries and the increasing burden of degenerative valve disease in the ageing population.1,2 About 4 million prosthetic heart valve replacements have been done over the past 50 years,3 and this remains the only definitive treatment for most patients with severe valvular heart disease. 300000 prosthetic heart valve replacements are done every year worldwide, 100000 in North America;4 the total number of replacements is projected to be 850 000 per year by 2050.5 Two major types of prosthetic heart valves exist: mechanical and bioprosthetic. Mechanical prosthetic heart valves are more durable but also more thrombogenic than bioprosthetic valves. The advantages of bioprosthetic over mechanical valves are that they provide more physiological haemodynamics and do not need long-term anticoagulation. Recent developments in the design of bioprosthetic heart valves have improved their durability and resistance to structural deterioration, and these valves are now increasingly being used in younger patients than in the past.6,7 In this review, we discuss different types of prosthetic heart valves and assess the antithrombotic management of patients with prosthetic heart valves and the management of valve-related thrombotic complications. We also review new developments and emerging technologies, including genotype-based warfarin dosing, self-monitoring of oral anticoagulant therapy, novel antithrombotic drugs, and the potential benefit of transcatheter valve replacement. Prosthetic heart valves Mechanical valves Three main types of mechanical valves exist: caged-ball, single leaflet or tilting-disk, and bileaflet valves. Mechanical valves have three key components: occluder (closure mechanism), housing, and sewing ring.8 All have some degree of regurgitant flow (washing jet) that prevents thrombus formation on the surfaces of the valve. The first prosthetic heart valve was the Starr-Edwards caged-ball valve introduced in 1960.9 The original version of the Starr-Edwards valve had a silicone rubber (silastic) ball or poppet that freely moved within the confines of a three-strut alloy cage (figure 1A). Subsequent models had a metal ball and a four-strut cage. The free-ball design theoretically prevents thrombus that forms on the sewing ring from extending onto the occluder.10 However, the ball generates a wake of stagnant blood flow that might contribute to the high risk of thromboembolism reported with caged-ball valves.11 The Starr-Edwards valve was the gold standard against which new mechanical valves were compared for more than 20 years,12 and is still widely used in many developing countriesbecauseofitslowcost. Single-leaflet or tilting-disk valves consist of a major and a minor orifice. Because the tilting disk enables central flow of blood, the risk of thromboembolism is lower than that with caged-ball valves, which have circumferential blood flow. Tilting-disk valves seem to be associated with a slightly higher risk of thromboembolism Review Search strategy and selection criteria We searched Medline between 1966 and March, 2009, and the Cochrane electronic database (4th quarter 2008) for English-language articles that addressed the long-term management of patients with prosthetic heart valves, with the terms  heart valve prosthesis ,  heart valve prosthesis implantation ,  antithrombotic ,  antiplatelet ,  anticoagulation ,  aspirin ,  vitamin K antagonist ,  warfarin ,  acenocoumarol ,  phenprocoumon ,  thrombosis ,  randomised controlled trial ,  randomised ,  controlled trial , and  meta-analysis . We reviewed reference lists of relevant papers identified by our electronic search. We focused on randomised controlled trials and meta-analyses of randomised controlled trials because they provide the least biased and most robust evidence for treatment. When randomised controlled trials were not available, we included observational studies and took into consideration expert opinion. www.thelancet.com Vol374 August15,2009 565 Review A B C D E F G H Figure 1: Different models of prosthetic heart valves (A) Starr-Edwards caged-ball valve (courtesy of Edwards Lifesciences LLC, Irvine, CA, USA). (B) Bjork-Shiley tilting-disk valve (courtesy of Sorin Group of Canada Inc, Canada). (C) Medtronic Hall tilting-disk valve (with permission from Medtronic Inc, Canada). (D) St Jude Medical Regent bileaflet valve (courtesy of St Jude Medical Canada). (E) Medtronic HK II ultra porcine valve (with permission from Medtronic Inc). (F) Medtronic Freestyle porcine valve (with permission from Medtronic Inc). (G) Carpentier-Edwards Perimount bovine pericardial valve. (H) Edwards SAPIEN transcatheter pericardial aortic valve (courtesy of Edwards Lifesciences LLC, Irvine, CA, USA). than bileaflet mechanical valves, possibly because they have a region of stagnant blood flow adjacent to the aorta, immediately downstream from the minor orifice, and because the regurgitant washing-jet volume is lower than that of bileaflet valves.8 The first successful tilting-disk valve was the Bjork-Shiley valve introduced in 1969,13 which consists of a single leaflet of pyrolytic carbon held in place by large inflow and small outflow alloy struts encircled by a teflon sewing ring8 (figure 1B). The Bjork-Shiley convexo-concave valve was withdrawn in 1986 because of several cases of strut fracture and embolisation of the disk.8 The Medtronic-Hall tilting-disk valve was one of the most commonly implanted tilting-disk valves (figure 1C). The St Jude Medical valve introduced in 197714 was the first bileaflet valve and is the single most commonly implanted mechanical valve to date. It is made of pyrolytic carbon coated with graphite and consists of two leaflets hinged on a ring (figure 1D). Encircling this structure is a sewing ring. Bileaflet valves provide symmetric, non-turbulent, central blood flow.8 Numerous bileaflet valves modelled on the St Jude valve are commercially available. Bileaflet valves are currently the most commonly used mechanical valve. Bioprosthetic valves Most bioprosthetic valves are of porcine origin (an intact heart valve from a pig is sewn into the valve structure) or constructed from a sheet of bovine pericardium that is cut to form valve leaflets and sewn into the valve structure. Valves are preserved in glutaraldehyde and mounted on a frame or stent made of metal or plastic covered with fabric that acts as the sewing ring.15 Bioprosthetic valves mimic native heart valves more closely than mechanical valves because they have unobstructed central flow, although both types of valves provide excellent haemodynamics. Bioprosthetic heart valves are less thrombogenic than mechanical valves and do not require long-term anticoagulant therapy. Porcine and pericardial (bovine) valves have similar thrombogenicity. Porcine valves are the most widely used bioprosthetic valves.10 The first commercial porcine valve was the Hancock valve introduced in 1970. An example of a porcine bioprosthetic valve is shown in figure 1E. Bovine pericardial valves have several theoretical advantages over porcine valves. Valve leaflets are larger, which accommodates shrinkage during the life of the valve; leaflet opening is more complete and symmetric, which improves valve haemodynamics; and the collagen content is higher, which improves valve durability. The Carpentier-Edwards Perimount valve (figure 1G) is the only pericardial valve widely available in North America. Pericardial valves seem to be at least as durable as contemporary porcine valves,16,17 but it is unclear whether theoretical advantages of pericardial valves over porcine valves translate into improved outcomes for patients. Table 1 shows rates of valve deterioration according to 566 www.thelancet.com Vol374 August15,2009 Review 10 years 15 years Aortic valve Mitral valve 21 40 20% 70% 41 50 12% 46% 51 60 18% 42% 61 70 6% 33% >70 2% 5% 21 40 36% 90% 41 50 37% 61% ( at 12 years) 51 60 31% 47% (at 12 years) 61 70 31% 67% >70 18% 34% Data are based on 2943 patients (17 471 patient-years). *Data are from Jamieson and colleagues.18 Valve failure=intrinsic abnormality of the valve, not due to endocarditis or thrombosis, leading to an increase in New York Heart Association (NYHA) class symptoms or reoperation. Table 1: Bioprosthetic valve failure at 10 and 15 years according to age (years) of patient and valve position at implantation* age of the recipient and to the position of the valve (ie, in the aortic or mitral position).18 Stentless bioprosthetic valves have no stent or frame as part of their structure. The aim of these valves is to provide a larger effective orifice area and lower postoperative transvalvular gradients than stented valves, thereby facilitating left ventricular mass regression in patients with severe aortic stenosis. Several randomised trials have compared stented with stentless valves, and a meta-analysis of these studies found that left ventricular mass regression was significantly greater at 6 months in patients receiving stentless valves than in those receiving stented valves.19 By 12 months, however, left ventricular mass regression was equivalent in the two groups. Figure 1F shows a Medtronic freestyle valve consisting of a porcine valve housed within its native aorta. Other freestyle valves include the Sorin Freedom, Edwards Prima Plus, and St Jude Toronto. Transcatheter valves Two transcatheter aortic valves have been implanted in many patients in clinical trials: the Cribier-Edwards (now Edwards SAPIEN) (figure 1H) and the CoreValve System aortic valve prosthesis.20 The original Cribier-Edwards valve consisted of equine tissue, but the SAPIEN consists of bovine pericardium and a steel stent, and the CoreValve of porcine pericardium and a nitinol stent. The effect of crimping to enable transcatheter placement and subsequent re-expansion of the valve on their long-term durability and thrombogenicity is unknown. Valve prototypes of different materials, some of which can be repositioned or retrieved, are in early stages of clinical assessment.21 Antithrombotic therapy Figure 2 shows a suggested algorithm adapted from the 2006 American College of Cardiology (ACC) and American Heart Association (AHA),22 and the 2008 American College of Chest Physicians (ACCP) guide- lines23 for the antithrombotic management of patients with prosthetic heart valves. These guidelines are mostly based on observational data because only a few randomised studies have been done. Factors that contribute to the thrombogenicity of prosthetic heart valves include: altered blood flow and haemostatic activation caused by vessel-wall disruption during surgery or exposure of artificial surfaces (sutures, sewing ring, occluder, and valve housing) to the circulating blood.24 Because almost all prosthetic valves are stented, they have a smaller effective orifice area than native valves, which results in a transvalvular flow gradient. Stagnant flow can be caused by the valve occluder or growth of endocardial tissue (pannus) into the leaflets or valve mechanism. Endothelialisation of the valve stent occurs over about 3 months after valve implantation, after which the risk of thrombosis decreases.10,25 Short-term parenteral anticoagulation with unfrac- tionated heparin or low-molecular-weight heparin is often used until therapeutic concentrations of an oral vitamin K antagonist are reached. Aspirin and vitamin K antagonists, alone or in combination, are used for long-term management of patients with prosthetic heart valves. Vitamin K antagonists are the only oral anti- coagulants available for valve-implanted patients. Warfarin has a mean half-life of about 40 h and is the most widely used vitamin K antagonist in North America. Other antagonists commonly used in Europe include acenocoumarol (half-life 8 11 h), fluindione (half-life 30 h), and phenprocoumon (half-life 3 5 days). Vitamin K antagonists are difficult to use in clinical practice because they have a slow onset and offset, narrow therapeutic window, and variable dose response in individuals, and interact with several foods and drugs.24 These antagonists need to be closely monitored for their anticoagulant effect, which is inconvenient for patients and costly for health-care systems. The international normalised ratio is a standardised method of reporting the intensity of anticoagulant therapy with vitamin K antagonists.26,27 Antithrombotic treatment for mechanical valves Estimates of the risk of thromboembolism after mechanical prosthetic heart valve replacement in patients not treated with anticoagulants mainly come from small case-series of patients with a contraindication to vitamin K antagonists. A 1994 systematic review of these studies28 showing 1225 patient-years of follow-up reported rates of valve thrombosis of 1∑8 (95% CI 0∑9 3∑0) per 100 patient-years, major embolism of 4∑0 (2∑9 5∑2) per 100 patient-years, and total embolism of 8∑6 (7∑0 10∑4) per 100 patient-years. These data are mainly derived from www.thelancet.com Vol374 August15,2009 567 Review Mechanical Bioprosthetic Aortic Mitral Aortic Bileaflet or Medtronic Hall Caged ball or tilting disk Warfarin for 3 months INR 2∑0 3∑0 Aspirin 75 100 mg per day for 3 months Warfarin for 3 months INR 2∑5 3∑5 Warfarin life-long INR 2∑0 3∑0 ±aspirin 75 100 mg per day Risk factors Warfarin life-long INR 2∑5 3∑5 ±aspirin 75 100 mg per day Life-long aspirin 75 100 mg per day Mitral Recommended Alternative 568 www.thelancet.com Vol374 August15,2009 Figure 2: Algorithm for antithrombotic therapy for prosthetic heart valves* ACC=American College of Cardiology. ACCP=American College of Chest Physicians. AHA=American Heart Association. INR=international normalised ratio. *Based on the 2006 ACC/AHA guidelines22 and the 2008 ACCP guidelines.23 Risk factors: atrial fibrillation, previous thromboembolism, left ventricular ejection fraction less than 35%, and hypercoagulable condition. patients with caged-ball or tilting-disk valves in the aortic position; thromboembolism rates are 1∑5 to 2 times higher for mechanical valves in the mitral position.28 No reliable data exist for the risk of thromboembolism in patients with bileaflet mechanical valves who do not receive antithrombotic therapy. No randomised trials have compared vitamin K antagonists alone with vitamin K antagonists with initial heparin (either unfractionated heparin or low- molecular-weight heparin) immediately after valve surgery. A 2006 systematic review of observational and randomised trials comparing outcomes in patients with mechanical valves treated immediately after surgery with different intensities of vitamin K antagonists reported an absolute rate of thromboembolism of 0∑9% and bleeding of 3∑3% during the first 30 days.29 Patients who also received a combination therapy with unfractionated heparin or low-molecular-weight heparin started 6 24 h after surgery and continued until a therapeutic international normalised ratio was achieved had thromboembolism rates of 0∑6 1∑1% and bleeding rates of 4∑8 7∑2% during the first 30 days. An observational study has shown a reduction in the risk of valve thrombosis when low-molecular-weight heparin was added to oral anticoagulation until the international normalised ratio reached the target therapeutic range.30 No randomised controlled trials have compared initial bridging therapy with unfractionated heparin versus low-molecular-weight heparin, but observational studies found no difference between them for either thromboembolism or bleeding.31,32 On the basis of the scarce data available, it is reason- able to start treatment with vitamin K antagonists after mechanical heart valve replacement as soon as haemostasis is secure, usually within 6 24 h after surgery. No high-quality data exist to guide decisions for the use of unfractionated or low-molecular-weight heparin immediately after mechanical heart valve replacement until the international normalised ratio is therapeutic on vitamin K antagonist therapy. The ACC/ AHA and ACCP guidelines suggest that it is reasonable to start bridging therapy immediately after surgery if bleeding is not an issue.22,23 The rationale for long-term anticoagulation in patients with mechanical heart valves is based on their inherent thrombogenicity and the high rates of thromboembolism in the absence of anticoagulation. One randomised study33 showed that antiplatelet therapy alone compared with oral anticoagulation was associated with a three-fold increase in thromboembolic events during 18 24 months of follow-up, although antiplatelet drugs caused less bleeding. The optimum target international normalised ratio range for mechanical valves was analysed in a retrospective, observational study by Cannegieter and colleagues34 involving patients with different types of mechanical prosthetic heart valves in the aortic, mitral, or both, position who received varying intensities of vitamin K antagonist therapy. The lowest rates of a combination of bleeding and thromboembolic events occurred when the international normalised ratio was between 2∑5 and 4∑9 (absolute rate of 2 events per 100 patient-years).34 Rates of thromboembolism were higher for valves in the mitral compared with those in the aortic position (0∑9 vs 0∑5 per 100 patient-years). Caged-ball valves seemed to be the most thrombogenic (2∑5 thromboembolic events per 100 patient-years), followed by tilting-disk valves (0∑7 per 100 patient-years) and bileaflet valves (0∑5 per 100 patient-years). Two subsequent randomised studies35,36 assessed whether targeting the low end of the international normalised ratio range would give adequate protection from thromboembolic events while reducing the risk of Review bleeding. Results showed that mechanical bileaflet aortic valves can be anticoagulated to an international normalised ratio of 2∑5 (range 2∑0 3∑0) instead of a higher international normalised ratio without an increased risk of thromboembolism and with a reduction in bleeding.35,36 The number of patients with mechanical mitral valves in these randomised studies was small and, because mitral valves have a higher rate of thromboembolism than that of aortic valves, a target international normalised ratio of 3∑0 (range 2∑5 3∑5) is recommended for mechanical mitral valves. For patients with caged-ball or tilting-disk valves in the aortic position, the same target ratio is recommended because these valves are more thrombogenic than bileaflet valves. A Cochrane systematic review of 11 randomised controlled trials involving 2428 patients found that the addition of aspirin to oral anticoagulation reduced mortality and thromboembolic events compared with oral anticoagulation alone at the cost of increased bleeding.37 39 The addition of low-dose aspirin (d"100 mg per day) to warfarin did not increase bleeding, but increased bleeding with this combination of drugs has been shown in other clinical settings.40 The ACC/AHA guidelines22 recommend the use of aspirin and warfarin for all patients with mechanical valves, whereas the ACCP guidelines23 and the European Society of Cardiology (ESC) guidelines41 recommend the addition of aspirin only for those who have additional thromboembolic risk factors or other indications for antiplatelet therapy, such as coronary or peripheral arterial disease. No reliable evidence exists concerning the appropriate antithrombotic management of patients with mechanical heart valves who have other thromboembolic risk factors (eg, decreased left ventricular function [ejection fraction <35%], atrial fibrillation, previous thromboembolism, and hypercoagulable conditions42,43). A reasonable target international normalised ratio is 3∑0 (range 2∑5 3∑5) in patients deemed to be at increased risk or to treat with low-dose aspirin therapy (75 100 mg per day), in addition to a vitamin K antagonist, but these recommendations are based only on expert opinion. Antithrombotic treatment for bioprosthetic valves The rate of thromboembolic events for patients with bioprosthetic valves seems to be highest during the first 3 months after surgery.44 The largest reported groups of patients with bioprosthetic valves who were given no antithrombotic therapy during both the initial 3 months and long term consisted of only 15645 and 13646 patients with follow-up for 1 and 7 years, respectively. The rate of thromboembolism at 1 year was 1∑3% (aortic valves only) and at 7 years was 1∑5% and 1∑7% per patient-year for aortic and mitral valves, respectively. According to two small observational studies,47,48 patients with bioprosthetic heart valves who received anticoagulation therapy for 3 months after surgery followed by no antithrombotic therapy had a thromboembolism rate of 1∑5 5∑2% per patient-year after 3 7 years of follow-up. 5∑2% is likely to be an overestimation and has not been reproduced in other studies or case series. Because of the perceived increase in thromboembolic risk during the first 3 months after surgery in patients with bioprosthetic heart valves, most studies have treated patients with a vitamin K antagonist for the first 3 months. The only published randomised study49 that compared varying intensities of vitamin K antagonist therapy for the management of patients with bioprosthetic valves involved 108 patients (most of whom received an aortic bioprosthetic valve) and found no difference in major embolic events between vitamin K antagonist therapy that targeted international normalised ratio of 2∑0 2∑3 compared with that of 2∑5 4∑0 during the first 3 months after surgery (~2% in each group). However, the lower target group had 37% less bleeding than the higher target group. On the basis of this study, it seems reasonable to target an international normalised ratio of 2∑5 (range 2∑0 3∑0) for patients with bioprosthetic valves in the aortic position. The target for patients with a bioprosthetic valve in the mitral position is 3∑0 (range 2∑5 3∑5). Therapy with vitamin K antagonists is usually started immediately after surgery and continued for 3 months. Two small randomised trials50,51 and many small observational studies52 57 have assessed whether treatment with vitamin K antagonists could be replaced with antiplatelet treatment as initial therapy for patients with a bioprosthetic aortic valve. None of the studies found a difference in thromboembolic or bleeding events in patients treated with vitamin K antagonists compared with those treated with aspirin, but studies were small and underpowered, and observational studies are subject to confounding factors. Low-dose aspirin is regarded by the ACC/AHA22 and the ACCP,23 as an alternative to warfarin for the first 3 months after bioprosthetic aortic valve replacement. The ESC41 recommends initial treatment with vitamin K antagonists over antiplatelet therapy. Two ongoing, prospective international, multicentre registries ACTION58 and ANSWER59 are following patients after bioprosthetic valve replacement to assess whether there is a difference in clinical outcomes for different postoperative antithrombotic regimens. On the basis of available evidence, low-dose aspirin is a reasonable alternative to warfarin during the first 3 months after surgery in patients with bioprosthetic aortic valves. Because the risk of thromboembolism is low after 3 months,44 risk of bleeding is likely to outweigh any benefit if vitamin K antagonist therapy is continued beyond 3 months. Irrespective of the choice of initial antithrombotic therapy, patients should be treated with life-long, low-dose aspirin (d"100 mg per day) after the first 3 months. An observational study of 215 patients47 showed that those who received aspirin therapy had 75% fewer thromboembolic events than those receiving www.thelancet.com Vol374 August15,2009 569 Review 570 www.thelancet.com Vol374 August15,2009 no antithrombotic therapy at 36 months. Antithrombotic treatment for valvuloplasty bands and rings Annuloplasty bands and rings are used to repair mitral and tricuspid valves. They are sewn to the annulus to plicate it and prevent further dilatation. They consist of rubber covered in polytetrafluoroethylene similar to the stents of prosthetic valves.15 A systematic review60 identified 12 small observational studies that reported outcomes in patients receiving antithrombotic therapy after valve repair. Most patients in these studies were treated with vitamin K antagonists for the first 2 3 months, and had low rates of thromboembolic events (0∑4 3∑0% per patient-year) and bleeding (0∑3 0∑8% per patient-year). A third of patients developed atrial fibrillation during the first 3 months. In the absence of high-quality data, use of the same antithrombotic treatment in patients with valvuloplasty bands and rings as in patients receiving bioprosthetic valves is recommended. The ACC/AHA and ACCP guidelines do not provide recommendations for the antithrombotic management of valvuloplasty rings and bands. The ESC guidelines41 recommend the use of vitamin K antagonists for 3 months (target international normalised ratio 2∑5, range 2∑0 3∑0), and the European Association for Cardiothoracic Surgery (EACTS) recom- mends either vitamin K antagonists for 3 months or antiplatelet therapy.61 Anticoagulation for dental procedures and surgery Patients with bileaflet mechanical valves in the aortic position are at low risk (9% per patient-year28) of thromboembolic complications during temporary dis- continuation of anticoagulation for procedures. Those with mechanical valves in the mitral position, or tilting-disk and caged-ball valves in any position, are at high risk (~1∑5 2 times the thromboembolic rates of the low-risk group, or 13 18% per patient-year28) for thrombo- embolic events. Moderate-risk patients include those with mechanical aortic valves and at least one additional risk factor (eg, atrial fibrillation of previous thrombo- embolic event) for thromboembolism (9 13% per patient-year).62 However, no randomised studies have been done to compare bridging anticoagulation to no anticoagulation, and thus management recommendations are based on observational studies or expert opinion. In patients with low-risk mechanical heart valves who require discontinuation of warfarin for a procedure, treatment should be interrupted 4 5 days before surgery and restarted within 24 h after surgery, if haemostasis is secure. 14 prospective cohort studies involving 1367 patients with a mechanical heart valve have shown that those in the moderate-risk and high-risk groups bridged with unfractionated or low-molecular-weight heparin, started 2 3 days before the procedure and recommenced the day after surgery in combination with vitamin K antagonists, have low rates of thromboembolism (0∑8%) and major bleeding (0∑1%).62 Management of thrombotic complications The reported rate of prosthetic valve thrombosis from a large randomised trial and two large series done in developed countries with modern valves and routine oral therapy with vitamin K antagonists is 0∑03 0∑13% per patient-year (10 15-year follow-up).63 65 The risk is highest during the first year after valve implantation and in patients with a mechanical tricuspid valve replacement.66,67 The long-term rate of prosthetic valve thrombosis is similar in patients with bioprosthetic valves (with or without antiplatelet therapy) and in properly anti- coagulated patients with mechanical valves, but most cases occur in patients with mechanical valves who are inadequately anticoagulated or have additional risk factors such as atrial fibrillation.63 Prosthetic valve thrombosis can be either obstructive or non-obstructive. Patients with obstructive thrombosis present with dyspnoea or acute pulmonary oedema, arrhythmia, cardiogenic shock, or systemic embolism. Heart sounds might be muffled or absent, especially in patients with mechanical valves, and a regurgitant murmur might be present. Patients with non-obstructive thrombosis are more likely to present with embolic events but almost 50% are asymptomatic. The clinical history can be helpful to distinguish prosthetic valve thrombosis from pannus (the two often occur con- currently) and from vegetation. Valve thrombosis is more likely in patients with a history of subtherapeutic anticoagulation, thromboembolic risk factors (prosthetic valve in the mitral or tricuspid position, atrial fibrillation, or hypercoagulable states), or soon after surgery. Fever can occur with both prosthetic valve thrombosis and endocarditis, and thus blood cultures should be done in febrile patients with suspected thrombosis to rule out endocarditis.68 The diagnosis of prosthetic valve thrombosis can be confirmed 85% of the time by the combination of transthoracic echocardiography and fluoroscopy, but transoesophageal echocardiography remains the gold standard test.69 Infective and other non-thrombotic causes (eg, pannus, tumour) should be excluded before making a diagnosis.68 Current recommendations for the management of prosthetic valve thrombosis, based on observational data from case series and cohort studies, favour the use of thrombolytic therapy as first-line treatment.68,70 The reported success rate for thrombolytic therapy is 71 88%, with complication rates (thromboembolism or bleeding) of 15 25% and mortality rates of 3 12%.71 74 Thrombolytic therapy is effective irrespective of valve type and position,72,73,75 but success rates seem to be higher in patients with New York Heart Association (NYHA) class I or II symptoms than in those with more severe symptoms.74,75 Bleeding and thromboembolic events after thrombolytic therapy are common in patients with NYHA Review Prosthetic valve thrombosis confirmed on TTE and fluoroscopy or TEE Obstructive Thrombus >5 mm Non-obstructive Contraindication to thrombolysis No contraindication to thrombolysis Thrombus size ì! or ë! leaflet mobility Unfractionated heparin+vitamin K antagonist (INR 2∑5 3∑5) ±low-dose aspirin Thrombus d"5 mm TT+anticoagulation +TEE monitoring Anticoagulation +TEE after 48 h Surgery Thrombus size same or ë! or leaflet mobility not improved class III or IV symptoms and in those with haemodynamic instability, previous history of stroke, a thrombus area of 0∑8 cm2 or more, and rapid infusion of thrombolysis.72,74 Observational data suggest that streptokinase is more effective than urokinase or tissue plasminogen activator.74 Patients with non-obstructive prosthetic valve thrombosis are frequently stable clinically but are at high risk of thromboembolic complications. Success of thrombolytic therapy depends on the size of the thrombus; in small case series, the reported success rate is 82% with thrombi smaller than 5 mm, with thromboembolism and death rates of 4% and 8%, respectively. For thrombi larger than 5 mm, the success rate drops to 61%, with thromboembolic and death rates of 23% and 38%, respectively.68 Surgical mortality for prosthetic valve thrombosis ranges from 12% to 46%.70,75,76 Figure 3 shows a proposed algorithm for the management of patients with prosthetic valve thrombosis. Prosthetic valves carry a long-term thromboembolic risk of 0∑5 1∑7% per patient-year despite appropriate antithrombotic therapy. No good evidence exists on how to manage patients with prosthetic heart valves who experience cerebral embolisation during anticoagulant therapy. Our practice is to use imaging to confirm the absence or presence of intracerebral haemorrhage. In the absence of intracerebral hemorrhage, we continue treatment with vitamin K antagonists. In the presence of haemorrhage, we reverse anticoagulation with a combination of low-dose vitamin K and fresh frozen plasma. We generally restart treatment after 7 10 days if there is no recurrent bleeding. This approach is associated with a 5% rate of thromboembolism and 1% risk of rebleeding.77 All patients with a prosthetic heart valve who have thromboembolic events must be examined with echocardiographic scans to rule out valve thrombosis or endocarditis. Expert consensus guidelines22 recommend that patients with a bioprosthetic valve who have a thromboembolic event while not receiving any antithrombotic therapy should commence therapy with vitamin K antagonists or low-dose aspirin (75 100 mg per day) if they are within 3 months of valve implantation, or low-dose aspirin if they are more than 3 months after implantation. A vitamin K antagonist can be added in a patient already receiving low-dose aspirin. In patients with a prosthetic heart valve already receiving oral anticoagulation, adequacy of therapy should be assessed at the time of the thromboembolic event. In patients not adequately treated at the time of the event, efforts should focus on improving anticoagulant control. Aspirin can be added to vitamin K antagonist therapy in a patient adequately treated with oral anticoagulation at the time of the event or, if the patient is already taking aspirin, the target international normalised ratio can be increased by 0∑5. Future developments Vitamin K antagonist therapy is highly effective in Figure 3: Algorithm for the management of patients with prosthetic valve thrombosis* SHVD=Society of Heart Valve Disease. INR=international normalised ratio. TEE=transoesophageal echography. TT=thrombolytic therapy. TTE= transthoracic echography. *Based on the 2005 SHVD guidelines.68 Contraindications to thrombolysis: standard (active bleeding, history of haemorrhagic stroke, recent cranial trauma or neoplasm, or uncontrolled hypertension) and specific (left-sided thrombus e"10 mm or 0∑8 cm2, large left atrial [non-appendage] thrombus, recent ischaemic stroke [6 weeks], or recent major surgery [<4 days]). Metabolism *1/*1 Extensive, rapid, ultra-metaboliser *1/*3, *2/*3, *2/*2 Poor, slow Adapted from McClain and colleagues.79 Table 2: Cytochrome P450 2C9 variants and their association with warfarin metabolism *1/*2 Intermediate *3/*3 Extremely slow Enzyme production AA High (lower warfarin dose needed) AB Medium BB Low (higher warfarin dose needed) *Adapted from McClain and colleagues.79 Table 3: Vitamin K epoxide reductase complex 1 variants and their association with warfarin enzyme production* www.thelancet.com Vol374 August15,2009 571 reducing thromboembolic events in patients with prosthetic heart valves but increases the risk of bleeding even when the treatment is carefully monitored. The rate of bleeding is greatest during the initial weeks or months of starting warfarin therapy.78 About 40% of a patient s variability to warfarin dose can be explained by cytochrome P450 2C9 and vitamin K epoxide reductase complex 1 genotypes.79 Tables 2 and 3 show the effect of cytochrome P450 2C9 and vitamin K epoxide reductase complex 1 genotypes on warfarin metabolism and dose. Review Thromboembolism Prosthetic valve thrombosis Bleeding Developed countries (44 918 patient-years follow-up) 1∑8% per patient-year 0∑1% per patient-year 1∑0% per patient-year Developing countries (12 642 patient-years follow-up) 2∑6% per patient-year 1∑3% per patient-year 1∑9% per patient-year Table 4: Rates of thromboembolic and bleeding events for patients with similar mechanical heart valves in developed and developing countries114 130 572 www.thelancet.com Vol374 August15,2009 The risk of major bleeding could be increased by two-fold to four-fold in patients with poor-metabolising cytochrome P450 2C9 genotypes.78 82 Genotype-based warfarin therapy has shown potential to reduce complications. A report from the American Enterprise Institute-Brookings Joint Center, with input from the US Food and Drug Administration, concluded that routine genotyping could prevent 85000 serious bleeding events and 17 000 strokes every year in the USA. If these targets were achieved, corresponding health-care savings would be between US$100 million and $2 billion per year.83 Small, randomised trials comparing standard with genotype-based dosing84 86 have shown that genotype predicts the initial warfarin dose but with no evidence of improved clinical outcomes. The potential benefit of genotype-based therapy is substantial, but much research is needed before routine genotyping can be recommended to guide therapy. Patients who self-monitor therapy with vitamin K antagonists at home rather than in a laboratory are more often in the therapeutic range and have a lower incidence of complications and hospital admissions than those who do not.87 90 Meta-analyses of randomised trials have recently found that patient self-monitoring was associated with a 33% reduction of risk of death, a 55% reduction of risk of thromboembolism, and a slight decrease in major haemorrhage.91,92 Self-monitoring was also associated with improved quality of life and satisfaction. The main obstacle to widespread use of patient self-monitoring is cost. In the UK National Health Service, the estimated cost of patient self- monitoring is £122000 per quality-adjusted life year (QALY) over 5 years and £63000 over 10 years.91 This is not cost-effective considering the commonly accepted threshold of £30000 per QALY. Costs are related to the portable international-normalised-ratio-monitoring device, test strips, and patient education programmes. The US Medicare programme has recently decided to cover the cost of patient self-monitoring.93 Patients who can lead an independent and self-supporting life are candidates for its use.94 The oral direct thrombin inhibitor dabigatran etexilate and two oral direct factor Xa inhibitors rivaroxaban and apixaban are in advanced stages of clinical development and are expected to replace oral vitamin K antagonists for many indications.95 Dabigatran etexilate and rivaroxaban have been approved in Europe and Canada for prevention of venous thromboembolism, and trials of these agents and of apixaban in patients with atrial fibrillation are almost completed.96 The main advantages of these drugs compared with those of vitamin K antagonists include their predictable pharmacokinetics and pharmacodynamics, and reduced interactions with foods and drugs, which allows them to be administered in fixed doses without monitoring coagulation. None of these new agents have been studied in patients with prosthetic heart valves. Another device ThromboCheck (Cardosignal GmbH, Hamburg, Germany) that enables patient self- monitoring of mechanical heart valve function has been tested in observational studies.97 This device records the individual frequency spectrum of sounds created by a patient s valve. With each subsequent check, the recorded frequency is compared with the original one, and results are automatically sent to a medical centre. Changes in the measured sound frequency of the valve might indicate early onset of valve thrombosis, paravalvular leak, pannus formation, or endocarditis. An observational study97 of more than 500 patients with a mechanical heart valve who used the device has recently shown that an alarm signal had positive predictive values and specificities of 97% and 100%, respectively, for valvular pathology subsequently diagnosed by echocardiography and fluoroscopy. In 2000, Bonhoeffer and colleagues98 did the first successful human transcatheter valve implantation in the pulmonic valve position, which was followed by Cribier and colleagues99 in 2002 who implanted a valve in the aortic position. Inoperable patients with severe, symptomatic aortic stenosis can be given valves (Edwards SAPIEN and CoreValve) via a percutaneous (retrograde femoral artery) or transapical (mini thoracotomy and insertion through the left ventricular apex) approach. The percutaneous approach has given the following outcomes in clinical trials: successful implantation 78 86%, 30-day or in-hospital death 11 25%, stroke 0 10%, and major vascular complications 8 17%.100 107 With the transapical approach, the data are: successful implantation 90 100%, 30-day or in-hospital death 8 22%, stroke 3 5%, and major vascular complications 2%.108 110 Results are expected to improve as operators become increasingly more experienced and technology improves. The Placement of AoRTic TraNscathetER Valve Trial (PARTNER) is an international, multicentre trial that will randomly assign more than 1000 patients to open-heart aortic valve replacement, transcatheter aortic valve replacement (Edwards SAPIEN valve), or medical therapy (ClinicalTrials.gov identifier NCT00530894). Results are expected in 2014. Although no standardised recommendations for antithrombotic therapy in patients with transcatheter aortic valves currently exist, commonly used approaches include life-long low-dose aspirin (Walther T, Leipzig University, Germany, personal communication), life- long low-dose aspirin in combination with clopidogrel Review for 1 month (Webb JG, St Pauls Hospital, Vancouver, BC, Canada, personal communication), or life-long dual antiplatelet therapy with low-dose aspirin and clopidogrel (Svensson LG, Cleveland Clinic, Cleveland, OH, USA, personal communication). Prosthetic valves: a global perspective and research priorities Since prosthetic heart valves were first introduced in the 1950s and 1960s, great advances have been made to increase their durability, reduce their thrombogenicity, and improve the long-term care of patients who receive them. However, a huge unmet need remains in developing countries where the burden of valvular heart disease is greatest. An international population-based study2 estimated that there are currently 15 20 million people worldwide living with rheumatic heart disease, of whom more than three-quarters are in developing countries. There are an estimated 282 000 new cases of rheumatic heart disease every year worldwide, with 95% of these occurring in developing countries. Almost 500 000 people die every year from rheumatic heart disease, with most deaths occurring during childhood or early adulthood.2 Rheumatic valve disease is usually not amenable to surgical repair and requires replacement with a prosthetic heart valve.111 Mechanical valves are preferred over bioprosthetic valves for the management of rheumatic heart disease because they are more durable and the disease most commonly affects children and young adults. However, most patients in developing countries do not have access to cardiac surgery.112 For those who receive a mechanical heart valve, the need for long-term anticoagulation poses unique challenges because of illiteracy, poverty, remote distances, under- funded and underequipped medical facilities, and lack of drugs. A study done in Africa113 found that patients with mechanical valves attended their anticoagulation clinic once every 59 days and could maintain therapeutic international normalised ratio levels only 18% of the time. Consequently, rates of thromboembolic events, prosthetic valve thrombosis, and bleeding in patients with mechanical valves are much higher in developing countries than in developed countries (table 4).114 130 The development of durable valves that do not require anticoagulation and can be delivered safely with a catheter has the greatest potential to benefit developing countries, but patients are also least able to afford them.131 The PROACT trial is randomly assigning low-risk patients undergoing a mechanical On-X bileaflet aortic valve replacement to receive either combination of aspirin and clopidogrel or vitamin K antagonist therapy (ClinicalTrials.gov identifier NCT00291525). This valve is composed of pure pyrolytic carbon (without silicon), which might reduce its thrombogenicity compared with that of other prosthetic valves.8 Completion of this study is expected in 2015. Asian people are more sensitive and African people less sensitive to warfarin than white populations.132,133 Asia, being the most populous continent on earth, is where genotype-based warfarin dosing could prevent the greatest number of bleeding events by avoiding hyperanticoagulation when commencing therapy. Conversely, genotype-based dosing for patients of African descent could avoid delays in reaching therapeutic levels of anticoagulation, and thus prevent thromboembolic events. Conclusions Prostheticheartvalveshavegreatlyimprovedsurvivalof patients with severe valvular disease, and the number of valve implantations worldwide has steadily increased. Despite limited randomised studies, vitamin K antagonists are widely used and are highly effective for prevention of thromboembolic complications in patients with mechanical heart valves, if they are appropriately monitored. New developments and technologies have increased access to valve replacement and have improved the antithrombotic management of patients with prosthetic heart valves in developed countries, but the greatest unmet need is in developing countries, which have the greatest burden of valvular heart disease. Our future aims should ensure that improvements in valve technologies and antithrombotic therapies are applied globally to truly reduce the morbidity and mortality of patients with valvular heart disease. Contributors JCJS and JWE contributed to the design, literature review, data interpretation, and writing of the report. AL and MJD contributed to the literature review and writing of the report. Conflicts of interest JCJS has received research funding from Bristol-Myers Squibb and GlaxoSmithKline. JWE has received honoraria and/or research support from Astra-Zeneca, Bayer, Bristol-Myers-Squibb, Eisai, Eli Lilly, GlaxoSmithKline, Johnson & Johnson, McNeil, and Sanofi-Aventis. MJD is a consultant for Edwards Life Sciences and Medtronic Inc. AL declares that he has no conflicts of interest. ˇ˛Anxiety disorders david s Baldwin Khalil i ajel Matthew J Garner Abstract anxiety symptoms and disorders are common in community settings and in primary and secondary care. symptoms can be mild and transient, but many people are troubled by severe symptoms that cause great personal distress, and which impair social and occupational function. the societal burden associated with anxiety disorders is considerable, but many of those who might benefit from treatment are not recognized or treated by healthcare professionals. By contrast, some patients receive unneces- sary or inappropriate treatment. recognition relies on keen awareness of the psychological and physical symptoms which are seen in all anxiety disorders, and diagnosis rests on the identification of the particular anxi- ety disorders. the need for treatment is determined by the severity and persistence of symptoms, the level of associated disability and impact on everyday life, the presence of coexisting depressive symptoms and other features such as a good response to or poor tolerability of previ- ous treatment approaches. choice of treatment is influenced by patient characteristics, patient and doctor preferences, and the local availability of potential interventions. there is much overlap between the different anxiety disorders for evidence-based and effective therapies (such as the prescription of a selective serotonin reuptake inhibitor or undergoing individual cognitive behavioural therapy) but there are also important differences, and for this reason it is helpful to become familiar with the characteristic features and evidence-base for each anxiety disorder. Keywords anxiety disorder; cognitive behavioural therapy; recognition; selective serotonin reuptake inhibitor; treatment David S Baldwin MBBS DM FRCPsych is Reader in Psychiatry at the University of Southampton and Honorary Consultant Psychiatrist at the RSH Hospital, Southampton, UK. Competing interests: DSB has acted as a consultant to and holds or has held research grants (on behalf of his employer) from a number of companies with an interest in anxiety and depressive disorders (Asahi, AstraZeneca, Cephalon, Eli Lilly, GSK, Lundbeck, Organon, Pharmacia, Pierre Fabre, Pfizer, Roche, Servier, Sumitomo, and Wyeth). Khalil I Ajel MBChB MRCPsych is a Specialist Registrar in General Adult Psychiatry on the Wessex Deanery Rotational Scheme in Southampton, UK. Competing interests: Dr Ajel has received sponsorship to attend scientific meetings hosted by Eli Lilly and Lundbeck. Matthew J Garner BSc PhD is Lecturer in Psychology and Medicine at the School of Psychology in the University of Southampton. Competing interests: Dr Garner has received sponsorship from Eli Lilly to attend a scientific meeting. Anxiety symptoms and anxiety disorders Anxiety is a normal response to threat or stress, and is usually transient and controllable. It probably represents an  alarm , allowing someone to prepare for physical response to a per- ceived danger (the  fight-or-flight response). Anxiety is common among patients undergoing examination, investigation or treat- ment. Anxiety symptoms are clinically important when they are abnormally severe, unduly prolonged, occur in the absence of stress, and are associated with impairment of physical, social or occupational functioning. Anxiety symptoms can lead to presen- tations to Emergency Departments and to repeated consultations in primary and secondary care settings (Table 1). By convention, a distinction is made between physical symp- toms that are mainly due to autonomic arousal (for example, tremor, shortness of breath, and palpitations) and psychologi- cal symptoms, including apprehension, irritability and restless- ness. Anxiety disorders are diagnosed when a patient has the required number of symptoms for more than a minimum speci- fied period, and these symptoms cause significant distress and are associated with impairment in everyday function. All disor- ders share common psychological and physical symptoms, but each differs from the next in having characteristic features that aid diagnosis (Table 2)  for example, behavioural symptoms such as recurrent unexpected panic attacks and secondary ago- raphobia in panic disorder, or the cognitive symptom of fear of embarrassment and humiliation in social phobia  and dif- ferential diagnosis of a suspected anxiety disorder is aided by a simple algorithm (Figure 1). Community studies in the general population aged 18 65 years show that anxiety disorders have a lifetime prevalence of approximately 21%, with an overall female-to-male ratio of approximately 2:1 across this age range1 (Figure 2). Anxiety dis- orders (especially generalized anxiety disorder) are also common in older adults. Psychiatric disorders What s new? " GAD: refinement of cognitive models to explain origin of symptoms in Gad; use of ssris in acute treatment and in prevention of relapse; use of antipsychotic drugs to augment response after limited response to initial ssri treatment. " Panic disorder: recognition that panic attack frequency is not the best measure of illness severity or treatment response; demonstration that psychotropic drugs can enhance efficacy of cBt; emphasis given to cBt in Nice guidelines. " Social phobia: recognition that social phobia is more common than was previously thought. " PTSD: Nice recommendations for sequential treatment in primary and secondary care settings for Ptsd. " OCD: Neuropsychological studies in ocd showing abnormalities of executive function; suggestion that some cases may have their origin in autoimmune processes. MediciNe 36:8 415 © 2008 elsevier Ltd. all rights reserved. Acute presentations of anxiety disorders " Panic disorder. Panic disorder is characterized by recurrent unexpected surges of severe anxiety ( panic attacks ) which typically reach their peak within ten minutes and last around 30 45 minutes. Many patients believe they are in imminent danger of death or collapse and seek urgent medical attention. " Generalized anxiety disorder (GAD) is characterized by prolonged and excessive worrying that is not restricted to particular circumstances. Worries often centre on possible physical ill health, affecting themselves or family members, and patients can repeatedly present with medically unexplained physical symptoms, craving reassurance or requesting inappropriate medical investigations. " Social phobia (social anxiety disorder)  people with social phobia have a marked and persistent fear of being observed or evaluated negatively by other people, in social or performance situations. Many avoid consulting doctors, but some present with physical symptoms (such as excessive perspiration) or psychological symptoms (such as the fear of vomiting whilst in public). " Post-traumatic stress disorder (PTSD) typically develops some months after a traumatic event in which the individual felt intense helplessness or horror: characteristic symptoms include intrusive recollections, disturbed sleep and hyper- arousal. Patients can also present with symptoms of associated conditions or behaviours, such as alcohol abuse or after non-fatal self-harm. " Obsessive compulsive disorder (OCD) is characterized by recurrent obsessional ruminations, images or impulses, and/ or recurrent physical or mental rituals. common symptoms include ruminations about possible accidents, and counting or checking rituals. obsessive compulsive symptoms can be a feature of some neurological conditions (for example, tourette syndrome). Table 1 Mixed anxiety and depression  Depressive symptoms often accompany anxiety disorders  and approximately one-third of people with anxiety disorders also fulfil diagnostic criteria for major depression, this pattern being named  comorbidity . Treat- ment of depression will usually relieve anxiety symptoms when depression is the primary diagnosis, but if depression is comor- bid or follows an anxiety disorder, each condition will require separate consideration and often separate treatment.2 General considerations in the treatment of anxiety disorders Anxiety symptoms exist on a continuum and many people with mild symptoms of recent onset and associated with stressful events or situations will improve without the need for treatment. However, the chronic nature and associated disability of anxiety disorders means that most patients who fulfil diagnostic criteria are likely to benefit from treatment, whether this is psychological or pharmacological. The need for treatment is determined by the intensity and duration of symptoms, the degree of disability and its impact on everyday life, the presence of coexisting depressive symptoms, and other features such as a good response to, or poor toler- ability of, previous treatment approaches. The choice of a par- ticular treatment should be influenced by patient characteristics (for example, comorbid physical ill health and treatment contra- indications), by patient and doctor preferences, and by the local availability of potential interventions.2 In general, the efficacy of psychological and pharmacologi- cal approaches is similar in the acute treatment of anxiety dis- orders, with best evidence for judicious prescription of selective serotonin reuptake inhibitors (SSRIs)2 or manualized cognitive behaviour therapy (CBT) delivered by trained and supervised staff.3 It remains uncertain whether combining these approaches is associated with greater improvement than with either treat- ment given alone, and for this reason it is best to plan sequential steps in patient management.2,4,5 Many patients worry about starting pharmacological treat- ment, fearing problems like unwanted sedation and the risk of developing a dependence on prescribed medication. Conversely, others are reluctant to engage in a psychological treatment that can be limited in its availability, time-consuming and costly. Regardless of treatment modality, patients need to understand that transient worsening of symptoms can occur, and that pro- longed efforts are needed to consolidate an initial response to treatment. Stigma can influence how a psychiatric diagnosis or its treat- ment is perceived. Patients with anxiety disorders may fear being dismissed as having minor or  lifestyle complaints and therefore can be discouraged from seeking help and undergoing treatment. Antidepressant drugs have been the subject of media attention for many years, and patients may receive conflicting messages about the benefits and risks of pharmacological treatment. Simi- larly, access to evidence-based psychological treatments may be reserved more for those with  severe and enduring mental ill- nesses such as schizophrenia, reflecting a failure to recognize the impairment and chronicity of many anxiety disorders. Patients should be referred by general practitioners to second- ary care services when there is uncertainty regarding possible underlying diagnosis; after non-response to two evidence-based acute treatment approaches; when there has been a recurrence of symptoms despite continuing treatment; when anxiety disorders are comorbid with other disorders (e.g. depression or substance misuse); and when there is a need for a specialist intervention (e.g. antipsychotic augmentation after partial response to an SSRI). Generalized anxiety disorder Clinical features and differential diagnosis Generalized anxiety disorder (GAD) is characterized by excessive worrying lasting more than six months and not restricted to par- ticular circumstances (for example only when attending a social gathering) Common features include apprehension, tension, difficulty in concentrating and autonomic anxiety, with symp- toms such as dry mouth and abdominal discomfort. It is one of the most frequent mental disorders in primary care but is often not recognized, possibly because only a minority present with MediciNe 36:8 416 © 2008 elsevier Ltd. all rights reserved. Psychiatric disorders Psychiatric disorders Important common differential diagnoses in anxiety disorders Condition depressive illness Psychotic illness Psychostimulant abuse drug withdrawal Physical ill health Differentiation from anxiety disorder early morning waking, feeling worse in the morning, loss of capacity for pleasure, constipation, guilty thoughts, and suicidal thoughts all suggest depression rather than anxiety. depressive symptoms in anxiety disorders tend to develop after the psychological and somatic symptoms that characterize the anxiety disorder (e.g. anticipation of embarrassment, anxiety and avoidance in social phobia) delusions, hallucinations and thought disorder are not seen in patients with primary anxiety disorders. some patients with chronic ocd begin to think their obsessional ruminations may be reasonable but the long-standing nature of the illness supports the diagnosis of  obsessive compulsive disorder with poor insight Use of amphetamines, ecstasy, cocaine, and hallucinogens can all result in agitation and severe anxiety, including panic attacks. Primacy of drug-seeking behaviour and physical signs of intoxication (such as stereotypic movements with amphetamine abuse) support the diagnosis of drug abuse. excess consumption of caffeine- containing drugs can result in physical and psychological symptoms of anxiety abrupt withdrawal of opiates, alcohol, barbiturates, benzodiazepines or antidepressants can result in agitation, tremor, dizziness, gastrointestinal upset and insomnia. anxiety disorders are not associated with acute confusional states, or with marked autonomic instability. characteristic physical signs are seen after withdrawal from certain drug classes, such as pupillary dilatation in people withdrawing from opiates anxiety symptoms are common in many physical health problems, and can be the presenting feature  for example, in thyrotoxicosis, hypoglycaemia, complex partial seizures, paroxysmal tachycardia, and phaeochromocytoma Table 2 the characteristic psychological symptoms. The most important differential diagnosis is depressive illness (although GAD and major depression often occur together), and patients should be asked about key depressive symptoms such as reduced interest, loss of weight, and suicidal thoughts, but distinction between the two conditions can be difficult.6 Aetiology Genetic studies suggest that GAD and major depression have a common basis, their manifestation being influenced by environmental factors.7 As in depression, disturbances of sero- tonergic and noradrenergic neurotransmission may be present, and particular GABA receptor sub-types may be abnormal in Algorithm to facilitate diagnosis of anxiety disorders Significant anxiety-related symptoms and impaired function Yes Also moderate/severe depression? No Predominant symptom focus Treat depression? Trauma history and flashbacks Check for PTSD Obsession ± compulsions Check for OCD Uncontrollable worry about several areas Check for GAD Intermittant panic/anxiety attacks and avoidance Fear of social scrutiny Check for social phobia Discrete object/situation Check for specific phobia Some uncued/ spontaneous Check for panic disorder PTSD, post-traumatic stress disorder; OCD, obsessive compulsive disorder; GAD, generalized anxiety disorder. Modified from Baldwin DS, Anderson IM, Nutt DJ, et al. J Psychopharmacol 2005; 19: 567 96 with permission from Sage publications.2 Figure 1 MediciNe 36:8 417 © 2008 elsevier Ltd. all rights reserved.