Not restricted and only human studies were included. Bibliographies of the included articles were examined to identify additional studies.Study selectionThe overall odds ratios (ORs) with corresponding 95 confidence intervals (CIs) were calculated to determine the association AZD-8835 biological activity between the presence of HLA-B*5801 in at least one allele and allopurinol-induced SJS/TEN. In some studies [11,12], allele frequency data were presented. Based on a study of Tanaka et al [20] and database of Major Histocompatibility Complex [8], we converted allele frequency data into the number of patients with HLA-B*5801 present in at least one allele. All analyses were performed using DerSimonian and Laird method [21] under a random-effects model. Sensitivity analyses were also performed to determine the robustness of the findings by ethnicity. The analyses were also performed separately for those using different types of control groups (e.g. controls obtained from the study, controls obtained from the population database). Statistical heterogeneity was assessed via the Q-statistic and I-squared tests [22]. A Q-value of 0.10 indicated statistically significant between-study heterogeneity and I-squared values < 25 denoted no/minimal heterogeneity across studies. Begg's test and Egger's test were used to evaluate the publication bias [23,24].ResultsStudy selectionTwo reviewers independently evaluated titles and abstracts retrieved from the comprehensive searches forA total of 94 articles were identified. After exclusion of duplication (6 articles), review articles or case reports (59 articles), non-human studies (1 article), studies in which patients did not receive allopurinol (5 articles), studies which did not examine the association betweenSomkrua et al. BMC Medical Genetics 2011, 12:118 http://www.biomedcentral.com/1471-2350/12/Page 3 ofHLA-B*5801 genotype and SJS or TEN outcomes (14 articles) and studies which did not have comparator (3 articles), 6 remaining studies included in the meta-analysis [10-15] (Figure 1). No additional articles were identified via review of the bibliographies of included studies.Study CharacteristicsCharacteristics of included studies are summarized in Table 1 and Table 2. All studies [10-15] included 96 SJS/TEN cases, 678 matched-controls and 3378 population-controls. The reported average age was 57.4 years old (range from 50.0-70.9 years) [10-15] and 51.6 years old (range from 35.9-63.5 years) [10,13-15] for cases and matched-control, respectively. Most patients were men (58.3 for cases and 77.7 for matched-controls). Kaniwa et al [11] and Lonjou et al [12] did not report gender percentage and mean age in the control groups. All SJS/TEN cases in all 6 studies were diagnosed according to the consensus definition [16-18]. All cases required confirmation of diagnosis by dermatologist [10,13], allergy specialist [15], Japan Severe Adverse Reaction (JSCAR) research group [11], or RegiSCAR expert committee [12]. All studies, except Kaniwa et al [11], required specific criteria for allopurinol exposure for the case definition. They included the duration of exposure to allopurinol no longer than 42 days [12], PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/27362935 60 days [15], or 3 months [10,13], and improvement of symptoms after drug discontinuation [10,13]. Hung et al [7] was the only study specifying that patients without symptoms upon re-exposure must be excluded. Despitea requirement of being exposed to allopurinol of cases, three studies [11-13] reported no information on dose o.