OP0022 COMPARATIVE EFFECTIVENESS OF BIOLOGIC AND TARGETED SYNTHETIC DMARDS IN PSORIATIC ARTHRITIS ACCORDING TO COMORBIDITIES AT TREATMENT START: A NORDIC COLLABORATIVE STUDY BASED ON ENRICHED CLINICAL REGISTERS

Background
Considering recent safety concerns for JAK-inhibitors (JAKi), clinicians need to make informed treatment decisions balancing expected effectiveness against risk for adverse events (AE). Current guidelines specifically caution that the use of JAKi in patients with pre-existing comorbidities increases risk of AE, but more data are needed regarding expected drug performance in these patients.
Objectives
To compare 1-year treatment retention and 6-month effectiveness in PsA patients with or without comorbidities at treatment start for A) IL17i versus TNFi, and B) JAKi versus TNFi.
Methods
Patients with PsA starting TNFi, IL17i or JAKi in routine care (years 2012-2020), were included from 5 Nordic rheumatology registries. Due to preferred use of IL17i and JAKi in bDMARD-experienced patients, b/tsDMARD naïve patients were excluded. Baseline patient characteristics were identified, as well as 6- and 12-month follow-up data. Linkage to national patient registries was used to identify comorbidities at treatment start (0-5 years for congestive heart disease, ischemic heart disease, thrombosis, depression/anxiety, chronic obstructive or interstitial pulmonary disease, diabetes, hospitalized infection, and chronic kidney disease; ever for malignancies). Hazard ratios (HR) for discontinuation at 1 year were estimated through Cox regression, while odds ratios (OR) at 6 months for low disease activity (LDA, DAPSA28≤14, LUNDEX adjusted) were estimated using logistic regression (adjusted for baseline variables, see Figure 1 text). Multiple imputation was performed to account for missing baseline data. For sensitivity, analyses limiting the TNFi group to adalimumab-treated patients were performed.
Results
A total of 10 276 treatment starts among 6118 patients with PsA were included. Patients initiating a TNFi were more often men and more often used concomitant csDMARDs, and they were less likely to have any (vs. none) of the above-mentioned comorbidities at treatment start (43%) compared to IL17i (46%) and JAKi (55%). IL17i and JAKi were frequently used in patients having failed >3 b/tsDMARDs (Table 1). Patients with a comorbidity were less likely to discontinue IL17i vs. TNFi (HR=0.87, 95%CI 0.77-0.98), no difference was seen for JAKi vs TNFi. Patients without comorbidities were less likely to discontinue IL17i than TNFi but more likely to discontinue JAKi (HR=1.42, 1.15-1.75) (Figure 1). For LDA, among patients with at least one co-morbidity, there was no difference in the chance of reaching LDA between TNFi (33%), IL17i (30%) and JAKi (29%). Among patients without comorbidities, the odds of reaching LDA were lower for JAKi and for IL17i vs. TNFi (Figure 1). Sensitivity analyses replacing TNFi with adalimumab showed similar estimates for both discontinuation and effectiveness (not shown).