Allopurinol Hypersensitivity: A Systematic Review of All Published Cases, 1950–2012

Abstract

Allopurinol is the primary therapy for the management of chronic gout. Utilization of allopurinol has increased in tandem with the growing prevalence of gout globally. This exposes more patients to the risk of allopurinol hypersensitivity (AH), a rare adverse reaction characterised by a spectrum of cutaneous reactions and systemic manifestations. Severe forms of AH have been associated with high mortality. The pathophysiology underlying this reaction remains unknown, but several risk factors have been proposed. The aim of this study was to review all published cases of AH documented in the literature in order to better understand the constellation of factors predisposing to this reaction, building on previous reviews by Lupton and Odom, Singer and Wallace and Arellano and Sacristan. A literature search was conducted in MEDLINE and EMBASE to identify relevant articles published between January 1950 and December 2012, with no language restrictions imposed. Articles that were included reported either allopurinol-induced cutaneous manifestations alone or satisfied the diagnostic criteria for AH as defined by Singer and Wallace. Nine hundred and one patients (overall AH cohort) were identified from 320 publications. Of these patients, 802 satisfied the Singer and Wallace criteria ('Singer and Wallace' cohort) while 99 patients had only mild cutaneous manifestations ('non-Singer and Wallace' cohort). Data were often incomplete; hence the results reported reflect the fractions of the subsets of the cohort where the data in question were available. In the overall AH cohort, 58 % (416/722) were male. The majority (73 %; 430/590) of patients were Asian. Renal impairment (48 %; 182/376) and hypertension (42 %; 160/376) were the most common chronic conditions; accordingly, diuretics (45 %; 114/252) and antihypertensives (39 %; 99/252) were the most prevalent concomitant medications. Allopurinol was prescribed for approved indications (chronic gout and chemoprophylaxis) in only 40 % (186/464) of patients. The median allopurinol dose was 300 mg/day (range 10-1,000 mg/day) and was taken by 50 % (168/338). There was no significant association between a higher dose (>300 mg/day) and an increased risk of severe cutaneous manifestations [odds ratio (OR) 1.76; 95 % CI 0.73-4.22; p = 0.23]. Approximately 90 % (489/538) of patients developed AH within 60 days of initiating allopurinol therapy. Serum oxypurinol (the active metabolite of allopurinol) concentration was only recorded in six patients, four of whom had levels within the putative therapeutic range of 30-100 μmol/L. The HLA-B*5801 allele was present in 99 % (166/167) of patients tested, with the majority (147/166) being of Asian ancestry. The all-cause mortality rate was 14 % (109/788) with 94 AH-related deaths, all of which occurred in the cohort meeting the Singer and Wallace criteria. The publications included in this review utilized different laboratory reference ranges to classify the non-cutaneous manifestations of AH; this may have introduced some variation in the cases identified as AH. A majority of the articles included in this analysis consisted of case reports and series--publication types that are not recognized as best-quality evidence; this thus limited the conclusions we could draw about the many risk factors we were interested in evaluating. Risk factors associated with AH, such as concomitant diuretic use, pre-existing renal impairment and recent initiation of allopurinol, were commonly present in AH patients; however, their role in the mechanism of AH remains to be established. A clear risk factor was the HLA-B*5801 status; this was especially relevant in Asian populations where there is a higher carriage rate of the allele. High allopurinol dose, previously suggested to be a risk factor, was not confirmed as such. The paucity of well-documented case reports and studies of AH render it difficult to draw more concrete conclusions or construct a meticulous profile of patients at risk of AH. Future case reports of AH need to be better documented to contribute to understanding the risks for, and mechanisms of, AH.