REPORTS

Allopurinol is the most common cause of Stevens-Johnson syndrome and toxic epidermal necrolysis in Europe and IsraelSima Halevy, MD,a Pierre-Dominique Ghislain, MD,b Maja Mockenhaupt, MD, PhD,c Jean-Paul Fagot, PharmD,d Jan Nico Bouwes Bavinck, MD, PhD,e Alexis Sidoroff, MD,f Luigi Naldi, MD,g Ariane Dunant, MS,b,h Cecile Viboud, PhD,d and Jean-Claude Roujeau, MD,b for the EuroSCAR Study Group Beer-Sheva, Israel; Creteil, Paris, and Villejuif, France; Freiburg, Germany; Leiden, The Netherlands; Innsbruck, Austria; and Bergamo, ItalyBackground: Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are rare severe cutaneous adverse reactions. Objectives: We sought to update knowledge on the causes of SJS or TEN with a focus on the rate of allopurinol-associated cases and to identify risk factors for allopurinol-associated SJS or TEN. Methods: We conducted a multinational case-control study. Results: In all, 379 patients with severe cutaneous adverse reactions validated as SJS or TEN and 1505 matched hospitalized control subjects were enrolled. Allopurinol was the drug most frequently associated with SJS or TEN, with 66 exposed patients (17.4%) and 28 exposed control subjects (1.9%) (adjusted odds ratio = 18, 95% condence interval: 11-32). Allopurinol use was greater than in a previous case-control European study. Daily doses equal to or greater than 200 mg were associated with a higher risk (adjusted odds ratio = 36, 95% condence interval: 17-76) than lower doses (adjusted odds ratio = 3.0, 95% condence interval: 1.1-8.4). The risk was restricted to short-term use ( # 8 weeks). The use of comedications did not increase the risk. Limitations: Nonsystematic recording of the indications for allopurinol use was a limitation. Conclusions: Results of this multinational study (EuroSCAR) revealed that allopurinol is the drug most commonly associated with SJS or TEN. The incidence of allopurinol-associated SJS or TEN has increased possibly because of increased use and dosages of this drug. ( J Am Acad Dermatol 2008;58:25-32.)From the Department of Dermatology, Soroka University Medical Center, Ben-Gurion University of the Negev, Beer-Shevaa; Reference Center for Toxic and Autoimmune Blistering Diseases, Department of Dermatology, Hopital Henri Mondor, Creteil, Paris XII Universityb; Dokumentationszentrum schwerer Hautreaktionen (dZh), Department of Dermatology, University Medical Center, Freiburgc; Inserm U 444, Faculte de Medecine St Antoine, Parisd; Department of Dermatology, Leiden University Medical Centere; Department of Dermatology and Venereology, Medical University of Innsbruckf; GISED Study Center, Department of Dermatology, Azienda Ospedaliera Ospedali Riuniti di Bergamog; and Biostatistics and Epidemiology Unit, Institut Gustave-Roussy, Villejuif.h Dr Ghislain is currently affiliated with St-Luc University Hospital, UCL, Brussels, Belgium, and Dr Viboud is currently affiliated with Fogarty International Center, National Institutes of Health, Bethesda, Maryland. Supported by the following institutions/companies (unrestricted grants): ADIR and Cie, Bayer Pharma/AG/Vital, Boehringer Ingelheim, Cassenne, Ciba Geigy/Novartis, Cilag GmbH, Dr Willmar Schwabe, Goedecke Parke Davis, Glaxo Wellcome/Glaxo SmithKline, Hoechst AG/Hoechst Marion Roussel/Aventis, HoffmannLa-Roche, IRIS Servier, Jouveinal Lab, LEO, LILLY, MSD Sharp and Dohme, Pfizer, Rhone Poulenc Rorer, Sanofi Winthrop/Sanofi Synthelabo GmbH, and Schering AG. Funding from pharmaceutical companies in France was managed through contract with Institut National de la Sante et de la Recherche Medicale, French Ministry of Health (PHRC AOM 98027). Conflicts of interest: None declared. Previously presented as: Roujeau JC, Ghislain PD, Mockenhaupt M, Bouwes Bavinck JN, Naldi L, Sidoroff A, Halevy S, Paty C. Doseeffect relation in SJS or TEN related to allopurinol. 5th International Congress on Cutaneous Adverse Drug Reactions, Satellite Congress of the 32nd ESDR Annual Meeting, Palexpo Geneva, Switzerland, September 18-19, 2002 [abstract]. J Invest Dermatol 2002;119:725; and Halevy S, Ghislain PD, Mockenhaupt M, Fagot JP, Schlingmann J, Roujeau JC. Allopurinol is the medication most frequently associated with Stevens-Johnson syndrome or toxic epidermal necrolysis in Europe. International Investigative Dermatology (IID-03), Fontainebleau Hilton, Miami Beach, Florida, April 30-May 4, 2003 [abstract]. J Invest Dermatol 2003;121:72. Accepted for publication August 9, 2007. Reprint requests: Sima Halevy, MD, Department of Dermatology, Soroka University Medical Center, Beer-Sheva 84101, Israel. E-mail: halevy@bgu.ac.il. Published online October 8, 2007. 0190-9622/$34.00 2008 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2007.08.036

25

26 Halevy et al

J AM ACAD DERMATOLJANUARY 2008

Abbreviations used: CI: NSAID: OR: SCAR: SJS: TEN: confidence interval nonsteroidal anti-inflammatory drug odds ratio severe cutaneous adverse reactions Stevens-Johnson syndrome toxic epidermal necrolysis

case-control study (EuroSCAR) conducted from 1997 to 2001.9

METHODSDesign The EuroSCAR study, a European case-control surveillance of SCAR, was conducted in 6 countries (Austria, France, Germany, Israel, Italy, and the Netherlands) between April 1997 and December 2001. Patients were actively detected in a network of about 1800 hospitals covering about 100 million inhabitants. Included were patients who developed the adverse reaction in the community, outside the hospital, and who were admitted because of symptoms of SCAR. For each case 3 hospital control subjects were matched on age, sex, region, and date of interview. They were selected according to a predefined list of admission diagnoses relating to acute conditions including infections (eg, pneumonia), trauma (eg, fractures), and abdominal emergencies (eg, appendicitis, ruptured ovarian cyst, strangulated hernia) that did not occur as a complication of an underlying chronic disease. We verified that our control group adequately represented the population10 in terms of prevalence of chronic disease and exposure to medications in common use (eg, benzodiazepines and drugs for diabetes mellitus [data not shown]). After obtaining informed consent from the study participants, trained investigators conducted personal interviews of patients and control subjects (by direct contact) with a structured questionnaire that included clinical data and information on medical history, infection, and drug exposure in the 4 weeks before hospitalization. The internal review board (Helsinki committee) of each of the participating centers in each country approved the study. Validation of patients and control subjects An international expert committee, composed of the 6 national study coordinators (all dermatologists) who were blinded to information on drug exposure and other risk factors validated the patients by reviewing the clinical data, photographs (available for 93% of patients), and results of the pathologic slides (available for 75% of patients). The patients were validated by means of a predened scoring system, which consisted of clinical and histopathologic parameters (ie, presence of mucous membrane erosions, skin detachment, epidermal sheets, atypical target lesions or spots, a positive Nikolskys sign, and epidermal necrosis). The expert committee also determined the date of onset of the disease (probable index day) and checked the validity of control

Allopurinol, an inhibitor of xanthine oxidase, is an effective uric acidelowering drug.1 Most of allopurinols activity is a result of the metabolite oxypurinol, a noncompetitive inhibitor of xanthine oxidase that prevents oxidation of xanthine to uric acid.1 Allopurinol has been implicated repeatedly in lifethreatening severe adverse drug reactions including hypersensitivity syndrome, Stevens-Johnson syndrome (SJS), and toxic epidermal necrolysis (TEN).2,3 SJS and TEN are rare, life-threatening, bullous cutaneous diseases generally considered as immune-mediated reactions to drugs. These severe cutaneous adverse reactions (SCAR) are characterized by epidermal necrosis, extensive detachment of the epidermis, erosions of mucous membranes and severe constitutional symptoms. There is evidence that SJS and TEN are a single disease with common causes and mechanisms.4,5 They differ mainly in the extent of detachment, which is limited in SJS (\10% body surface area), more widespread in TEN ([30%), and in-between in SJS/TEN overlap (10%30%). Although rare (2 cases/million population/y), SJS and TEN have a significant impact on public health in view of their high mortality (20%-25%) and morbidity.6 In a previous case-control study (SCAR study),7 the risk of SJS or TEN in relation to the use of medications was assessed in 4 European countries from 1989 to 1993. Allopurinol treatment was recorded in 13 of 245 patients with SJS or TEN (5.3%) and in 11 of 1147 control subjects (1%).7 At that time, allopurinol was used less frequently by patients with SJS or TEN than other high-risk medications such as anti-infectious sulfonamides (especially cotrimoxazole), phenobarbital, oxicam-type nonsteroidal anti-inflammatory drug (NSAID), chlormezanone, and carbamazepine.7 Based on the study results, regulatory agencies decided on withdrawal of chlormezanone from the market and restricted indications for cotrimoxazole and phenobarbital. Yet, allopurinol is still widely used in spite of its well-known association with SCAR. Furthermore, new indications for the use of allopurinol in cardiovascular diseases have been approved over the course of recent years.8 Accordingly, the risk of allopurinol-associated SJS or TEN was re-evaluated in a new European

J AM ACAD DERMATOLVOLUME 58, NUMBER 1

Halevy et al 27

subjects, based on the admission diagnosis and the date of onset of their acute condition. Analyses Odds ratio (OR) and 95% condence interval (CI) were estimated by standard methods (SAS, Version 9.1, SAS Institute Inc, Cary, NC and StatXact-5, Cytel Software Corporation, Cambridge, Mass). When at least 3 patients and control subjects were exposed to the evaluated risk factor, multivariate unconditional logistic regression was used to adjust for potential confounding factors. The data were adjusted for country, sex, age, exposure to high-risk drugs for SJS or TEN (anti-infection sulfonamides, carbamazepine, phenobarbital, phenytoin, lamotrigine, oxicamtype NSAIDs, nevirapine), any other suspected drugs (aminopenicillins, tetracyclines, quinolones, cephalosporins, macrolides, diclofenac and related NSAIDs, corticosteroids, acetaminophen, pyrazolones, acetylsalicylic acid), any other drug, and suspected nondrug risk factors (HIV infection, recent cancer, recent radiotherapy, or collagen vascular disease). Stratified analyses were performed to assess the impact of comedication on the risk of allopurinolassociated SJS or TEN. For most drugs, patients and control subjects were considered to have been exposed to a medication if it was taken in a 7-day window before the probable onset of the disease (probable index day). This window of exposure period was extended for all drugs with long elimination half-lives, including allopurinol (because of the long elimination half-life of its main metabolite, oxypurinol). Duration of drug exposure was classified as either # 8 weeks or [ 8 weeks before the index day.

had a definite index day and adequate information on drug exposure. Allopurinol exposure in patients and control subjects The risk estimates for drugs that were associated with high risk for SJS or TEN are presented in Table I. Allopurinol was the high-risk drug most frequently associated with SJS or TEN, with 66 exposed patients (17.4%) and 28 exposed control subjects (1.9%) (adjusted OR = 18, 95% CI: 11-32). However, it was not the drug with the highest OR. The distribution of the 66 patients according to country was Austria 5, France 12, Germany 41, Israel 2, Italy 5, and the Netherlands 1. The distribution of the 28 exposed control subjects was Austria 0, France 12, Germany 15, Israel 0, Italy 0, and the Netherlands 1. The rate of exposure to allopurinol among patients and control subjects was highest in Germany (24% of patients and 3% of control subjects). The risk estimates (adjusted OR) were similar for Germany and France at 16 (7.6-34) and 12 (4.0-34), respectively. It might have been higher in other countries (57; 7.3444), but the small numbers did not enable us to conduct a multivariate analysis. A comparison of allopurinol exposure between the previous SCAR study (1989-1993) and the current EuroSCAR study (April 1997-December 2001) (Table II) showed a 3-fold increase in the exposure of patients from 5.3% to 17.4%, and a 2-fold increase in the exposure of control subjects from 1% to 1.9%. The OR also increased from 5.5 to 18 (both adjusted) for all users of allopurinol (short and long term), and from 52 to 261 (both unadjusted) for short-term users of allopurinol ( # 8 weeks). Doses and duration of use of allopurinol As shown in Table III, the doses of allopurinol were higher in patients with SJS or TEN (median 300 mg/d, mean 258 6 66 mg/d) compared with control subjects (median 150 mg/d, mean 189 6 95 mg/d; P = .001). The difference persisted when calculated per kilogram of body weight. A daily dose of 200 mg or more of allopurinol was associated with an increased risk for SJS or TEN (adjusted OR = 36, 1776) compared with lower daily doses (3.0, 1.1-8.4). There was no association between the severity of illness, as manifested by the maximal erythema, the maximal detachment of the epidermis, or death and the daily dose of allopurinol (data not shown). As previously observed with allopurinol and other drugs, the risk for SJS or TEN was restricted to recent users ( # 8-week interval between initiation of treatment and onset of reaction). For short-term use of allopurinol the unadjusted OR was 261 (36-), but

RESULTSDescriptive analysis A total of 1763 control subjects were interviewed, of whom 1505 were enrolled in the study. Of 513 potential patients with SJS or TEN who were interviewed, 379 community-acquired patients were analyzed (SJS 134; SJS/TEN overlap 136; TEN 109). Patients developing SCAR during hospitalization for another condition were not included because their basal medical condition and medications were different from those of communityacquired patients and because control subjects had been selected as being representative of the community population. The 134 patients who did not fulll the criteria for SJS or TEN, as determined by the international expert committee, were excluded from the study (Fig 1). All patients and control subjects enrolled in the study

28 Halevy et al

J AM ACAD DERMATOLJANUARY 2008

Fig 1. Flow chart of patients with Stevens-Johnson syndrome or toxic epidermal necrolysis and control subjects enrolled in EuroSCAR study.

there was no signicant risk for long-term use ([8 weeks) (adjusted OR 0.9, 0.3-2.4). Furthermore, only 4 of 56 recent users (7%) of allopurinol also took another high-risk drug compared with 5 of 10 longterm users of allopurinol (50%), who took another high-risk drug (carbamazepine 2, oxicam-type NSAID 2, sulfasalazine 1). These results suggested that the risk for SJS or TEN was limited to recent exposure to allopurinol, whereas long-term users probably had other causes of the disease. Therefore, further analyses were restricted to the subgroup of 56 patients, who developed SJS or TEN within 8 weeks of initiating treatment with allopurinol. Demographic and clinical parameters The majority of the 56 patients recently exposed to allopurinol (33 of 56, 59%) were female. The patients were signicantly older than those with SJS or TEN not exposed to allopurinol (64 6 16 and 44 6

25, respectively, P\.0001). Furthermore, there was a trend toward a lesser extent of the skin lesions (ie, more of these cases were classified as SJS or SJS/TEN overlap and the minority as TEN), but the difference did not reach statistical significance. However, the death rate was significantly higher in patients recently exposed to allopurinol (18 of 56, 32%) than among patients not exposed to allopurinol (59 of 313, 18.8%) (P = .024). Comedication with allopurinol As shown in Table IV, there was no significant difference between patients and control subjects in the total number of concomitant medications (mean 6.9 and 6.6, respectively; median 7.0 and 5.5, respectively). Exposure to furosemide in 56 patients recently exposed to allopurinol and 28 control subjects exposed to allopurinol (36% and 32%, respectively)

J AM ACAD DERMATOLVOLUME 58, NUMBER 1

Halevy et al 29

Table I. Estimates of odds ratio for the 7 high-risk drugs most often associated with Stevens-Johnson syndrome or toxic epidermal necrolysis in the EuroSCAR studyDrug* Patients (%) n = 379 Control subjects (%) n = 1505 Unadjusted OR (95% CI) Adjusted OR (95% CI)

Allopurinoly Carbamazepine Cotrimoxazole Nevirapine Phenobarbitalz Phenytoin Lamotrigine

66 31 24 21 20 19 14

(17.4) (8.2) (6.3) (5.5) (5.3) (5.0) (3.7)

28 (1.9) 4 (0.3) 1 (0.1) 0 5 (0.3) 3 (0.2) 0

11 33 102 17 26

(7.0-18) (12-95) (14-754) (22-) (6.2-45) (7.8-90) (14-)

18 (11-32) 72 (23-225) ND ND 16 (5.0-50) 17 (4.1-68) ND

CI, Confidence interval; ND, not done because of less than 3 patients or control subjects; OR, odds ratio. *Exposure window: 7 days before the probable index day. y Exposure window: 14 days before the probable index day. z Exposure window: 21 days before the probable index day.

was higher than exposure to furosemide in the entire study group (10.8% and 3.3%, respectively). Similarly, exposure to any diuretic, including furosemide, in 56 patients recently exposed to allopurinol and 28 control subjects exposed to allopurinol (59% and 64%, respectively) was higher than exposure to diuretics in the entire study group (19.5% and 9.8%, respectively). Thus, patients and control subjects exposed to allopurinol were overexposed to furosemide and other diuretics compared with patients and control subjects in the entire study group, but this exposure did not alter the risk for SJS or TEN. Comedication of allopurinol with angiotensin-converting enzyme inhibitors, aminopenicillins, acetylsalicylic acid, or diclofenac was also not associated with an increased risk for SJS or TEN. Comorbidity Comorbidity rates with renal dysfunction, any infection, malignancy, collagen vascular disease, diabetes mellitus, or atopic dermatitis did not differ signicantly between patients and control subjects.

Table II. Comparison of allopurinol exposure in patients with Stevens-Johnson syndrome or toxic epidermal necrolysis and control subjects between the current EuroSCAR study and the previous SCAR studyAllopurinol* Study Patients (%) Control subjects (%) Adjusted OR (95% CI)

EuroSCAR study, 1997-2001 SCAR study, 1989-1993

66 of 379 (17.4) 13 of 245 (5.3)

28 of 1505 (1.9) 11 of 1112 (1.0)

18 (11-32) 5.5 (2-15)

CI, Confidence interval; OR, odds ratio; SCAR, severe cutaneous adverse reactions. *14 Days before the probable index day.

DISCUSSIONData from the EuroSCAR study, a multinational case-control surveillance of SCAR conducted in 6 countries, suggest that allopurinol, recorded in 17.4% of 379 patients and 1.9% of 1505 control subjects (adjusted OR = 18, 95% CI: 11-32), is the most common drug associated with SJS or TEN in Europe and Israel. Similar results were recorded in a recent case-control study conducted in a hospital in Taiwan, which showed allopurinol exposure in 17% of 35 patients with SJS and TEN and 2% of control subjects (adjusted OR = 186.7, 95% CI: 9.2-9,239), suggesting that allopurinol also plays a major role in the induction of SJS or TEN in the Asian population.11 The results of this study show that the risk for allopurinol-associated SJS or TEN is restricted to

short-term use ( # 8 weeks) (unadjusted OR = 261, 95% CI: 36-). Similar results were reported in the previous SCAR study.7 Furthermore, the current study suggests that the risk for allopurinol-associated SJS or TEN could be dose dependent. The use of 200 mg or more of allopurinol daily was associated with a higher risk for SJS or TEN (adjusted OR = 36, 95% CI: 17-76) compared with lower daily doses (3.0, 1.1-8.4). The recommended adult dosage of allopurinol starts at 100 mg/d and is increased at 1-week intervals until the desired serum uric acid concentration is attained.12 Previous reports have already shown that allopurinol is frequently prescribed in inappropriate doses13-16 and that adverse reactions to allopurinol are dose dependent.17,18 The increased risk for SJS or TEN in the EuroSCAR study was associated with both recent initiation of allopurinol treatment and high allopurinol dose. This finding suggests that the high doses of allopurinol that were used at the beginning of treatment may have contributed to the risk for SJS or TEN.

30 Halevy et al

J AM ACAD DERMATOLJANUARY 2008

Table III. Risk for Stevens-Johnson syndrome or toxic epidermal necrolysis according to allopurinol daily dose and duration of useAllopurinol Patients (%) n = 379 Control subjects (%) n = 1505 Unadjusted OR (95% CI) Adjusted OR (95% CI)

Daily dose \200 mg/d $200 mg/d Unknown Duration of use #8 wk [8 wk

9 (2.4) 49 (12.9) 8 (2.1) 56 (14.8) 10 (2.6)

12 (0.8) 11 (0.7) 5 (0.3) 1 (0.07) 27 (1.8)

3.5 (1.5-8.5) 21 (11-41)

3.0 (1.1-8.4) 36 (17-76)

261 (36-) 1.4 (0.7-3)

ND 0.9 (0.3-2.4)

CI, Confidence interval; ND, not done because of less than 3 patients or control subjects; OR, odds ratio.

Table IV. Comedication in 56 patients with Stevens-Johnson syndrome or toxic epidermal necrolysis and recent allopurinol use compared with 28 control subjects with allopurinol useDrug Patients recent use (%) n = 56y

Patients long-term use (%) n = 10

Control subjects (%) n = 28

P*

Furosemide Any diureticsz ACE inhibitors Aminopenicillins Acetylsalicylic acid Diclofenac

20 33 16 1 14 4

of of of of of of

56 56 56 56 56 56

(36) (59) (29) (2) (25) (7)

5 of 10 (50) 7 of 10 (70) 5 of 10 (50) 0 of 10 (0) 2 of 10 (20) 0 of 10 (0)

9 of 28 (32) 18 of 28 (64) 8 of 28 (29) 0 of 28 (0) 6 of 28 (21) 3 of 28 (11)

.60 .74 .76 1.00 .77 .68

ACE, Angiotensin-converting enzyme. *Patients with recent allopurinol use compared with control subjects with allopurinol use. The chi-square or Fishers exact test was used for differences in proportions. y Furosemide exposure in the whole study group (379 patients, 1505 control subjects). Patients: 10.8%; control subjects: 3.3%. z Furosemide; spironolactone; thiazides, and other diuretics. Diuretics exposure in the whole study group (379 patients, 1505 control subjects). Patients 19.5%; control subjects: 9.8%.

A drug interaction between allopurinol and other medications, including aminopenicillins, angiotensin-converting enzyme inhibitors, and acetylsalicylic acid or other salicylates, has been suspected to play a role in allopurinol-induced eruptions.1 The literature has focused also on interactions between allopurinol and diuretic agents in view of diuretic-induced hyperuricemia19 and furosemide-induced decreased urinary excretion of oxipurinol.20 Based on the results of the EuroSCAR study, it appears that exposure to furosemide, or diuretics in general, increased the likelihood of taking allopurinol, but was not associated with an increased risk for SJS or TEN. Comedication of allopurinol with aminopenicillins, angiotensin-converting enzyme inhibitors, acetylsalicylic acid, or diclofenac was also not associated with an increased risk for SJS or TEN. The established indications for allopurinol are treatment of hyperuricemia associated with chronic gout, acute uric acid nephropathy, recurrent uric acid stone formation, enzyme disorders, and cancer or cancer chemotherapy. There is general agreement that allopurinol is not indicated in the majority of

patients with asymptomatic hyperuricemia.1,12,21 Published studies revealed inappropriate indications for allopurinol in up to 86% of patients.13,22-26 In the EuroSCAR study the indications for allopurinol use were recorded in a nonsystematic manner and the patients original charts were not reviewed. A comparison of allopurinol exposure between the previous SCAR study (1989-1993) and the current EuroSCAR study (1997-2001) showed a 2- to 3-fold increase in the exposure of control subjects and patients. The rate of exposure in control subjects (1.9%) and the stable prevalence of gout in Europe (# 0.3%)27 suggests a trend toward treating asymptomatic hyperuricemia (prevalence # 28% in some areas28), a practice that is not supported by evidencebased data. It may be hypothesized that the increased rate of exposure to allopurinol among patients and control subjects in the current EuroSCAR study, compared with the previous SCAR study, is associated with increased sales of the medication, especially highdose tablets. Data from France show that from 1991 (mid term of SCAR study) to 1999 (mid term of

J AM ACAD DERMATOLVOLUME 58, NUMBER 1

Halevy et al 31

EuroSCAR study) the total sales for all brands and dosages of allopurinol increased by 20% with a strong increase in the market share of generics.29 Increases in sales affected principally 200- and 300mg formulations. Adherence to accepted guidelines for the use of allopurinol14 could lead to a significant decrease in morbidity and mortality from allopurinol-associated SJS or TEN. Based on the literature, which shows inappropriate allopurinol prescription in up to 86% of treated patients,22,24 it may be assumed that up to 48 of 56 patients with SJS or TEN with recent intake of allopurinol in the EuroSCAR study could have been prevented. Based on the known incidence of SJS or TEN (2 cases/million population/y), the extrapolation of our results (14.8% of SJS or TEN cases induced by recent allopurinol use with a 32% death rate) to the whole European population (376 million) leads to an estimate that about 100 cases of SJS or TEN and as many as 30 deaths related to these reactions could be avoided each year in Europe by a more appropriate use of allopurinol.We are indebted to all the patients whose participation made this study possible. In addition, we thank all the collaborating hospitals and colleagues for their enthusiastic support. Those who also participated in the study are as follows. Austria: Elisabeth Kowald, Dennis Linder, ` Gudrun Ratzinger; France: Fabienne Bazin, Helene Bocquet, Laure Dangoumaud, Yves Dorleans, Alain Dupuy, Catherine Paoletti, Annie-Claude Paty, Sylvie Rodriguez, Bruno Sassolas, Patricia Thion, Loc Vaillant; Germany: Konrad Bork, Uwe-Frithjof Haustein, Olaf Hering, Klaus Kitta, Brigitte Loleit, Oliver Pfeiffer, Thekla Renker, Birgit Schneck, Werner Schroder, Esra Tas, Dieter Vieluf; Israel: Arnon D. Cohen, Sara Weltfriend; Italy: Laura Atzori, Grazia Manfredi, Davide Melandri, Annalisa Pinna; the Netherlands: Christianne Bearda Bakker-Wensveen, Joost Govaert, Vina Williams-Snijders. We thank Mrs Peggy Sekula, Dipl Math (FH), from the Institute of Medical Biometry and Medical Informatics, University Medical Center Freiburg, for her critical review of the article and contributory comments. We thank Juergen Schneck, Dipl Stat, from the Department of Dermatology, University Medical Center, Freiburg, for his contribution to the statistical analyses in this study.REFERENCES 1. Sweetman S, editor. Martindale: the complete drug reference [database online]. London: Pharmaceutical Press; 2006. Electronic version; Greenwood Village, Colo: Thomson Micromedex. Edition expires September 2006. 2. Aubock J, Fritsch P. Asymptomatic hyperuricemia and allopurinol induced toxic epidermal necrolysis. Br Med J 1985;290: 1969-70. 3. Arellano F, Sacristan JA. Allopurinol hypersensitivity syndrome: a review. Ann Pharmacother 1993;27:337-43. 4. Auquier-Dunant A, Mockenhaupt M, Naldi L, Correia O, Schroder W, Roujeau JC, et al. Correlations between clinical patterns and

5.

6.

7.

8.

9.

10.

11.

12.

13.

14. 15.

16. 17.

18.

19.

20.

21.

22.

23.

causes of erythema multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis: results of an international prospective study. Arch Dermatol 2002;138:1019-24. Bastuji-Garin S, Rzany B, Stern RS, Shear NH, Naldi L, Roujeau JC. Clinical classification of cases of toxic epidermal necrolysis, Stevens-Johnson syndrome, and erythema multiforme. Arch Dermatol 1993;129:92-6. Rzany B, Mockenhaupt M, Baur S, Schroder W, Stocker U, Mueller J, et al. Epidemiology of erythema exudativum multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis in Germany (1990-1992): structure and results of a population-based registry. J Clin Epidemiol 1996;49:769-73. Roujeau JC, Kelly JP, Naldi L, Rzany B, Stern RS, Anderson T, et al. Medication use and the risk of Stevens-Johnson syndrome or toxic epidermal necrolysis. N Engl J Med 1995;333:1600-7. Struthers AD, Donnan PT, Lindsay P, McNaughton D, Broomhall J, MacDonald TM. Effect of allopurinol on mortality and hospitalizations in chronic heart failure: a retrospective cohort study. Heart 2002;87:229-34. Mockenhaupt M, Viboud C, Dunant A, Naldi L, Halevy S, Bouwes Bavinck J, et al. Stevens-Johnson syndrome and toxic epidermal necrolysis: assessment of medication risks with emphasis on recently marketed drugs, the EuroSCAR study. J Invest Dermatol 2007. doi:10.1038/sj.jid.5701033. Published online September 6, 2007. Kelly JP, Auquier A, Rzany B, Naldi L, Bastuji-Garin S, Correia O, et al. An international collaborative case-control study of severe cutaneous adverse reactions (SCAR): design and methods. J Clin Epidemiol 1995;48:1099-108. Lin MS, Dai YS, Pwu RF, Chen YH, Chang NC. Risk estimates for drugs suspected of being associated with Stevens-Johnson syndrome and toxic epidermal necrolysis: a case-control study. Intern Med J 2005;35:188-90. Klasco RK. Drugdex system [database online]. Greenwood Village, Colo: Thomson Micormedex. Edition expires September 2006. Devlin JW, Bellamy N, Bayliff CD. Observations and effects of educational consults on allopurinol prescribing. Can J Hosp Pharm 1992;45:21-7. Smith P, Karlson N, Nair BR. Quality use of allopurinol in the elderly. J Qual Clin Pract 2000;20:42-3. Stamp L, Gow P, Sharples K, Raill B. The optimal use of allopurinol: an audit of allopurinol use in South Auckland. Aust N Z J Med 2000;30:567-72. McClintock AD, Egan AJ, Woods DJ, Pillans PI. A survey of allopurinol dosage prescribing. N Z Med J 1995;108:346-7. McInnes GT, Lawson DH, Jick H. Acute adverse reactions attributed to allopurinol in hospitalized patients. Ann Rheum Dis 1981;40:245-9. Fam AG, Dunne SM, Iazzetta J, Paton TW. Efficacy and safety of desensitization to allopurinol following cutaneous reactions. Arthritis Rheum 2001;44:231-8. Gurwitz JH, Kalish SC, Bohn RL, Glynn RJ, Monane M, Mogun H, et al. Thiazide diuretics and the initiation of anti-gout therapy. J Clin Epidemiol 1997;50:953-9. Yamamoto T, Moriwaki Y, Takahashi S, Tsutsumi Z, Hada T. Effect of furosemide on renal excretion of oxypurinol and purine bases. Metabolism 2001;50:241-5. Dincer HE, Dincer AP, Levinson DJ. Asymptomatic hyperuricemia: to treat or not to treat. Cleve Clin J Med 2002;69: 594, 597, 600-2. Singer JZ, Wallace SL. The allopurinol hypersensitivity syndrome: unnecessary morbidity and mortality. Arthritis Rheum 1986;29:82-7. Bellamy N, Brooks PM, Emmerson BT, Gilbert JR, Campbell J, McCredie M. A survey of current prescribing practices of anti-

32 Halevy et al

J AM ACAD DERMATOLJANUARY 2008

inflammatory and urate-lowering drugs in gouty arthritis in New South Wales and Queensland. Med J Aust 1989;151:531-2, 535-7. 24. Zell SC, Carmichael JM. Evaluation of allopurinol use in patients with gout. Am J Hosp Pharm 1989;46:1813-6. 25. Stuart RA, Gow PJ, Bellamy N, Campbell J, Grigor R. A survey of current prescribing practices of antiinflammatory and uratelowering drugs in gouty arthritis. N Z Med J 1991;104:115-7. 26. Khoo BP, Leow YH. A review of inpatients with adverse drug reactions to allopurinol. Singapore Med J 2000;41:156-60.

27. Harris CM, Lloyd DC, Lewis J. The prevalence and prophylaxis of gout in England. J Clin Epidemiol 1995;48:1153-8. 28. Gresser U, Gathof B, Zollner N. Uric acid levels in southern Germany in 1989: a comparison with studies from 1962, 1971, and 1984. Klin Wochenschr 1990;68:1222-8. 29. Agence Francaise de Securite Sanitaire des Produits de Sante. Analyse des ventes de medicaments aux officines et aux hopitaux en France: 1991-2001. AFSSAPS ed, Paris 2003. Available at: http://www.agmed.sante.gouv.fr. Accessed September 18, 2007.

Effective January 1, 2008, CME articles published in the Journal of the American Academy of Dermatology will be designated for a maximum of 1 AMA PRA Category 1 CME CreditTM as per American Medical Association policy governing journal CME.