Abstract
Introduction Recent publications have described drug reaction with eosinophilia and systemic symptoms (DRESS) with topiramate. Topiramate has been associated with other severe cutaneous adverse reactions, including Stevens-Johnson syndrome, but a relationship to DRESS has not been established. To determine if there is a causal association between topiramate and DRESS, we conducted a comprehensive review of the data in the Janssen Research & Development Global Safety Database (GSD), signaling databases, and the literature. Methods The primary data were post-marketing reports of DRESS in the Janssen topiramate GSD (cumulative through 1 July 2022), representing >14,000,000 patient-years (PY) exposure. Cases were reviewed, assigned a Registry of Severe Cutaneous Adverse Reaction (RegiSCAR) score, and assessed for overall contribution of topiramate to DRESS based on temporality, concomitant medications, dechallenge/rechallenge, and baseline patient factors. Statistical disproportionality was evaluated in European Medicines Agency's EudraVigilance (EV) safety database and the United States Food and Drug Administration Adverse Event Reporting System (FAERS). For EV, the overall disproportionality threshold was the lower limit of the 95% confidence interval (CI) for the reporting odds ratio (ROR025) >1 and N ≥5. The overall threshold for FAERS was the Empirical Bayesian Geometric Mean (EBGM) ≥2, lower bound of the 90% CI (EB05) of >1, and N ≥3. To account for the role of concomitant drugs, Empirical Bayes regression-adjusted arithmetic mean (ERAM) scores were calculated, with a threshold ≥2, a lower bound of the 90% CI (ER05) of >1, and N ≥3. An integrated search of major biomedical literature was performed for reports of topiramate and DRESS. Results There were 17 reports of DRESS in the GSD (reporting rate 0.12/100,000 PY). RegiSCAR scores ranged from -3 to 7 (average -0.4). No cases met full diagnostic criteria and were highly confounded by the presence of other suspect drugs. Disproportionality scores exceeded thresholds for statistical significance in FAERS (N=72, EBGM=2.06, EB05=1.69), but not in EV (N=33, ROR025=0.79). When accounting for co-administered drugs, ERAM was statistically significant for carbamazepine (4.53), lamotrigine (ERAM=6.54), phenytoin (ERAM=2.91), and zonisamide (ERAM=2.25) exceeding disproportionality thresholds, but the score of topiramate was no longer significant (0.25). Conclusion A comprehensive review of all available evidence does not support a causal association between topiramate and DRESS.