Background Stevens-Johnson Syndrome (SJS) and Toxic Epidermal Necrolysis (TEN) are rare but potentially life-threatening skin reactions, commonly associated with drugs. They are part of the same disease spectrum and characterised by painful blistering and shedding of the skin. There is currently no systematic review assessing non-genetic risk factors for the development of drug-induced SJS/TEN. This is important because SJS/TEN has high mortality and morbidity.

The primary objective of this systematic review was to pool the relative risk of non-genetic risk factors for the development of drug-induced SJS/TEN. The secondary objective was to provide separate pooled estimates for males and females.

Methods The protocol was prospectively registered on PROSPERO on 2^nd September 2022. Medline, EMBASE, and grey literature databases were searched, with no language or date restrictions. Cohort, cross-sectional and case-control studies were included. A control group, such as people with a different skin condition, was part of the inclusion criteria. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were adhered to. Title and abstract screening, full-text screening and quality assessment were conducted by two researchers independently. Critical appraisal was conducted using the Joanna Briggs Institute criteria. Three domains, indicative of exposure validity, outcome validity and confounding were used to identify high-quality studies. Random-effects meta-analysis was conducted for each risk factor with two or more sufficiently homogeneous studies. Heterogeneity was assessed with I².

Results This systematic review included 15 papers, and these all had a high risk of bias. Of the 15, five were included in meta-analysis. The control groups were patients with Erythema Multiforme (EM) or Drug Reaction with Eosinophilia and Systemic Symptoms (DRESS). In the five papers, the number of cases of SJS/TEN ranged between 48 and 384. We found male gender and positive HIV status to significantly increase the risk of development of SJS/TEN, compared to EM, with pooled risk ratios of 1.24 (95% CI 1.01–1.54) and 1.70 (95% CI 1.32–2.20), respectively. I² was low for both meta-analyses. Gender, diabetes, liver disease, chronic kidney disease, infection, malignancy, and hypertension were not significant risk factors (p>0.05) when the control group was patients with DRESS. There was insufficient data to estimate separate risk ratios for males and females.

Conclusion Males and people with HIV have a statistically significant, increased risk of developing SJS/TEN following drug administration, compared to developing EM. Further research is required at a population-level with large sample size, healthy controls, and appropriate study design, to minimise bias.