Background/Purpose: Cardiovascular disease is a major comorbidity in patients with psoriasis (PsO) and psoriatic arthritis (PsA). Increasing evidence suggests a potential reduction in the risk of cardiovascular (CV) events with biological disease-modifying antirheumatic drugs (bDMARDs) treatment, mainly including tumor necrosis factor inhibitors (TNFi) with the more recent focus on atrial fibrillation (AF). However, the comparative CV risk with different bDMARDs has not been adequately studied in this patient population.

Methods: Using two large US commercial insurance databases, Optum Clinformatics and MarketScan, we identified patients with PsO or PsA diagnosis initiating ustekinumab or TNFi (i.e., adalimumab, etanercept, infliximab, certolizumab, and golimumab) from September 25, 2009 to September 30, 2015. The primary outcome was incident AF and the secondary outcome was a composite CV event including myocardial infarction, stroke, and coronary revascularization. Patients were followed until the first occurrence of the following events: 1) outcomes, 2) death, 3) plan disenrollment, 4) switching or 5) discontinuing treatment. To account for potential confounding, we estimated propensity score (PS) as the predicted probability of receiving ustekinumab conditioning on 62 covariates including demographic and clinical factors in each database. We used weighting based on PS fine stratification with 50 strata to control for confounding. The adjusted hazard ratio (HR) of each outcome was estimated using a weighted Cox proportional hazards regression model. We performed all the analyses separately in each of the databases and then combined the HR by a random-effects meta-analysis.  

Results: We identified 60,028 (15,470 in Optum and 44,558 in MarketScan) adult patients with PsO or PsA and no prior AF. The mean age of the cohort was 47.2 ± 12.7, and 51% were female.  81% had diagnosis for PsO while 46% had PsA, and 27% had both PsO and PsA.  After PS stratification, covariates were well balanced between the two groups. As presented in Table 1, 60 incident AF occurred in 9,071 ustekinumab initiators while 323 incident AF did in 50,957 TNF inhibitor initiators (IR=4.7 and 5.0 cases per 1,000 person-year, respectively). MACE occurred in 74 patients among ustekinumab initiators, and 421 of patients initiated TNF inhibitors (IR=6.2 and 6.1 cases per 1,000 person-year, respectively).  The combined HR for AF among the ustekinumab initiators compared to those initiated TNFi was 1.15 (95% CI 0.70-1.89), and for MACE, the HR was 1.14 (95% CI 0.89-1.46).

Conclusion: In this large cohort study of 60,028 patients with PsO or PsA, we found no substantially different risk of AF or CVD after initiation of ustekinumab or TNFi. However, a possible small effect cannot be ruled out warranting further studies with a longer followup.