Association between biological immunotherapy for psoriasis and time to incident inflammatory arthritis: a retrospective cohort study

Background Targeted biological immunotherapies have been highly effective in controlling skin disease in patients with psoriasis, but whether therapy delays progression to inflammatory arthritis is unclear. The aim of this study was to compare the time to incident inflammatory arthritis among patients newly receiving biological therapies for psoriasis. Methods In this retrospective cohort study, we obtained data on a national sample of patients in the USA from the electronic health records database of the US-based TriNetX network. We included adult patients (aged ≥18 years) with two diagnostic codes for psoriasis (>30 days apart; International Classification of Diseases [ICD] codes) who had been newly prescribed a biologic (inhibitors of tumour necrosis factor [TNF], interleukin [IL]-17, IL-23, or IL-12/23, first prescribed on or after the date of receiving a first psoriasis diagnosis code). The time to incident inflammatory arthritis, defined by first occurrence of a diagnostic code for psoriatic arthritis or other inflammatory arthritis after initiation of biological therapy, was graphed with use of the Kaplan-Meier estimate. Time-dependent risk for inflammatory arthritis was calculated with weighted Cox proportional hazards regression with anti-TNF exposure as the reference, adjusted for demographic and clinical covariables. Sensitivity analyses were used to evaluate incident cases of psoriasis, increased exclusion periods for prevalent cases of inflammatory arthritis, drug switching, and more stringent disease and outcome definitions. Findings Between Jan 1, 2014, and June 1, 2022, we identified 15 501 patients with psoriasis (mean age 50·2 years [SD 15·0]; 8399 [54·2%] women and 7102 [45·8%] men; 11 175 [72·1%] White). 976 (6·3%) of the 15 501 patients developed inflammatory arthritis, with a cumulative incidence of 2·6 cases per 100 person-years. In multivariable regression analyses, the risk of developing inflammatory arthritis was significantly lower in patients prescribed IL-12/23 inhibitors (adjusted HR 0·58, 95% CI 0·43–0·76) or IL-23 inhibitors (0·41, 0·17–0·95) than in patients prescribed TNF inhibitors. We found no significant difference for IL-17 inhibitors (0·86, 0·54–1·38) compared with TNF inhibitors. For IL-12/23 inhibitors, the results persisted in all sensitivity analyses. For IL-23 inhibitors, the results persisted in three of six sensitivity analyses, when a higher diagnostic threshold for incident arthritis was used and when excluding patients who developed arthritis within 3 or 6 months after first biologic prescription. Interpretation In this large cohort study of patients with psoriasis, treatment with IL-12/23 inhibitors or IL-23 inhibitors was associated with reduced risk of progression to inflammatory arthritis compared with TNF inhibitors. Prospective observational cohorts with disease activity measures and pooled analyses of previous randomised trials are required to confirm these findings. Funding None.