PO:06:081 | Safety of JAK inhibitors compared to non-TNF-targeted biologic therapies in rheumatoid arthritis patients who are inadequate responders to TNF inhibitors

Lorenzo Di Luozzo¹, Alessandro Giollo¹, Mariangela Salvato¹, Francesca Frizzera¹, Kiren Khalid¹, Margherita Zen¹, Andrea Doria¹ | ¹Azienda ospedaliera universitaria di Padova, unità di Reumatologia, Padova, Italy

Background. Rheumatoid arthritis (RA) patients with inadequate response to tumor necrosis factor inhibitors (TNFi) are commonly switched to Janus kinase inhibitors (JAKis) or biologic DMARDs with other mechanisms of action (OMA). Comparative safety data between these options are limited, as most studies focus on placebo or TNFi comparisons. This study aimed to generate real-world evidence on the safety of JAKis versus OMA in TNFi-experienced RA patients.

Methods. In this retrospective cohort study, we included RA patients who had previously received at least one TNFi and were subsequently treated with JAKis or OMA between 2014 and 2024, with a follow-up period of 24 months. We calculated and compared the incidence rate (IR) [95% CI] of treatment-emergent adverse events (TEAEs) per 100 person-years (100 PY) exposure of RA patients initiating JAKis compared to OMA. The following TEAEs occurring during follow-up were investigated: serious AEs (SAEs), infections (including COVID-19), cardiovascular disease (CVD), venous thromboembolic events (VTE), gastrointestinal disorders, major adverse cardiac events (MACE), malignancy (excluding non-melanoma skin cancer), and biochemical abnormalities. We performed Poisson regression to estimate Incidence Rate Ratios (IRRs) with 95% confidence intervals (CIs) for TEAEs, adjusting for disease activity (DAS28CRP) and accounting for PY of follow-up as an offset.

Results. We enrolled 305 RA patients initiating JAKis (n=163) or OMA (n=142). Baseline CVD risk and comorbidities were balanced; OMA patients had slightly higher disease activity and erosions, but fewer seropositive and multi-bDMARD refractory patients (Table 1). No significant differences in TEAE frequency (n=119) were found based on IRRs or IRs (Figure 2). Most TEAEs occurred during M0–M6 and were numerically higher with OMA (n=65; IR 87 [65,112]) than JAKis (n=51; IR 67 [50,88]). Infections (n=90) were the most common TEAEs, more frequent in OMA (IR 31 [19,47]) than JAKis (IR 22 [13,34]) during early follow-up. VTEs were numerically more frequent with JAKis (n=3; IR 0.82 [0.17,2.4]) than OMA (n=1; IR 0.33 [0.01,1.86]) over the entire study. SAEs (n=40) mainly occurred in M0–M6 and were more frequent with OMA (IR 12 [5,24]) than JAKis (IR 6 [2,15]), as were TEAEs leading to discontinuation (OMA: IR 43 [29,62]; JAKis: IR 16 [8,28]). Three malignancies were reported in M12–M24 (2 with JAKis, 1 with OMA). GI AEs and lab abnormalities were uncommon in both groups.

Conclusions. The safety profile of JAKis and OMA in RA patients who were TNFi-IR was comparable. IRs of SAEs were in line with previous data from clinical trials and observational studies.