Unsuspected Adverse Drug Reactions to Dermatologic Medications: An Epidemiological Hypothesis-Free Screening Study of Real-World Data in Denmark.

backgroundMany adverse drug reactions (ADRs) to dermatological drugs may be underrecognized due to limitations in traditional surveillance. Systematic, hypothesis-free screening for such ADRs using real-world data is an underutilized approach in dermatology.

objectiveThe aim was to systematically screen commonly used dermatological drugs for potential unknown ADR signals using nationwide Danish health register data and to evaluate sequence symmetry analysis (SSA) as an epidemiological screening method in pharmacovigilance. This evaluation focuses on identifying key signals for dedicated follow-up studies, not on an exhaustive analysis of all potential associations.

methodsA nationwide, register-based, hypothesis-free screening study using Danish national health registers was conducted. The study cohort comprised 5,877,711 Danish residents prescribed dermatological drugs or relevant immunosuppressants linked to a dermatological indication between 1996 and 2022. Data were analyzed from 1995 to 2024. Exposures were the first dispensation of included dermatological drugs. The primary outcome measure was the trend-adjusted sequence ratio (SR), an estimate of the incidence rate ratio for rare events, with 95% confidence intervals (CIs), for incident drug-drug and drug-diagnosis pairs within ±12-month windows of exposure initiation. To assess directionality, supplementary case-crossover (CCO) analyses were performed on top-ranked signals, estimating odds ratios (ORs) with 95% CIs.

resultsThe screening of 22.5 million incident exposure prescriptions yielded 4010 drug-drug and 22,234 drug-diagnosis pairs. After filtering and review, several potential unknown ADR signals were identified and prioritized. Signals were manually reviewed and categorized by clinical experts to identify plausible novel ADRs. Notable associations (adjusted SR [95% CI]; CCO OR [95% CI]) included isotretinoin with systemic corticosteroids (1.44 [1.35-1.53]; 1.78 [1.61-1.98]) and hemorrhoid treatments (1.58 [1.49-1.68]; 1.83 [1.68-1.98]); azathioprine with bone-modifying drugs (1.42 [1.33-1.53]; 1.57 [1.36-1.81]); terbinafine with lipid-lowering therapy (1.04 [1.01-1.08]; 1.14 [1.07-1.20]) and vitamin B12/folate use (1.20 [1.14-1.26]; 1.12 [1.02-1.22]); and clobetasol with atrial fibrillation/flutter (1.11 [1.06-1.17]; 1.07 [1.00-1.14]).

conclusionsSystematic hypothesis-free screening using SSA on nationwide real-world data can effectively identify potential unknown ADRs for dermatological drugs. The specific signals found in this study potentially have clinical implications and warrant further targeted investigations to support causality. This approach supports the implementation of routine real-world data screening to supplement traditional pharmacovigilance.