Ulcerative colitis reporting across medications: FAERS analysis and cross-species transcriptomic triangulation.

BACKGROUND

Ulcerative colitis (UC) is reported with medications from different therapeutic classes. We examined these reporting associations together with recorded clinical context and tested whether experimental cytokine-blockade signatures align with human UC mucosal expression.

METHODS

FAERS reports from 2004Q1-2024Q4 were restricted to health professionals and primary-suspect medications. Ten previously identified candidates were evaluated using unique report-generic counts, reporting odds ratios (RORs), proportional reporting ratios, and a supportive Bayesian information component. Sensitivity analysis excluded complete reports recording UC as an indication. Directional signatures from five antibody-versus-isotype contrasts in a 42-sample mouse colitis study were scored in two external human cohorts (87 UC/21 controls and 74 active UC/11 controls), using 12,109 common one-to-one orthologs. Ten primary score comparisons underwent joint Benjamini-Hochberg correction.

RESULTS

Among 7,300,798 eligible FAERS reports, 11,352 included UC. Eight candidates met the specified reporting criteria; seven retained support after UC-indication exclusion. Secukinumab and isotretinoin had 204 and 130 UC reports, with RORs of 4.49 (95% CI 3.90-5.16) and 3.79 (3.19-4.51). Valganciclovir decreased from nine to three reports after exclusion and became imprecise. In human mucosa, experiment-2 IL-17A and IL-17RA blockade signatures scored lower in UC in both cohorts; experiment-1 IL-17RA showed a positive difference in only one cohort. IL-23p19 and IL-12/23p40 blockade signatures scored lower in UC in both cohorts. These patterns persisted across the planned signature-size sensitivity analysis.

CONCLUSION

Event-centered analysis differentiated reporting associations by their indication context. External experimental and patient data did not support a uniform UC-like transcriptional pattern following IL-17 blockade, refining the mechanistic interpretation of the reporting signals. The findings identify priorities for drug-specific clinical investigation without estimating incidence or establishing causality.