Genetic determinants of response to biologic disease-modifying antirheumatic drugs (bDMARDs) in rheumatoid arthritis (RA) remain insufficiently characterized. We developed a targeted sequencing panel of candidate immunogenetic variants identified through a literature review and evaluated associations with response to TNF-α inhibitors, response to IL-6 inhibitors, and difficult-to-treat RA (D2TRA). Whole-blood DNA was sequenced using custom hybridization capture on a DNBSEQ-G400 platform (2 × 150 bp). Reads were aligned to hg38 using BWA-MEM, and variants were called with GATK HaplotypeCaller. After quality control, the D2TRA analysis included 154 patients (59 D2TRA and 95 non-D2TRA); drug-response analyses included 66 anti-TNF patients (26 responders, 40 non-responders) and 69 anti-IL-6 patients (29 responders, 40 non-responders). Treatment-response associations were assessed using unadjusted allelic chi-square tests with multiple-testing correction; D2TRA was analyzed using logistic regression adjusted for age, sex, and body mass index. For IL-6 inhibitor response, rs11656130/MAP2K6 (p = 0.0163; FDR = 0.3704) and rs4910008/GALNT18 (p = 0.0212; FDR = 0.3704) were identified, while rs7767069/LINC02549 was associated with TNF-α inhibitor response (p = 0.0171; FDR = 0.5968). For D2TRA, rs1813443 in CNTN5 (OR = 1.85, p = 0.017) and rs12081765, an intergenic variant near LMX1A/RXRG (OR = 0.59, p = 0.028), were nominally associated; neither association remained significant after multiple-testing correction. With the current sample sizes, the study could detect only large genetic effects, and was underpowered for these after Bonferroni correction (α = 0.00143). The findings highlight the limited reproducibility of candidate pharmacogenetic markers in RA and provide a targeted panel and analytical framework for larger replication studies.

Background/Objectives: Difficult-to-treat rheumatoid arthritis (D2T-RA) affects approximately 11.7% of patients. Real-world data on treatment trajectories, biologic and targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs), and adverse events (AEs) remain limited. This study evaluated treatment patterns, line survival, and AEs in D2T-RA. Methods: This retrospective study included 97 D2T-RA patients. Treatment-line survival was assessed using Kaplan–Meier analysis with log-rank tests and a cluster-robust Cox proportional hazards model. Generalized estimating equation models (GEE) identified factors associated with AEs. Cluster analysis (K-modes, MCA followed by K-means, and hierarchical clustering) was performed to identify treatment patterns. Results: A median time to first prescription of b/tsDMARDs was 1369 days (IQR 852–2557). Inefficacy was the most common reason for switching (65.8%); csDMARD/GCs were discontinued for inefficacy more frequently than b/tsDMARD (70.1% vs. 61.3%). Among 470 b/tsDMARD-containing treatment lines, interleukin-6 (IL-6) inhibitors were most common (127 lines, 27.0%), followed by tumor necrosis factor-α (TNF-α) inhibitors (109 lines, 23.2%), Janus kinase (JAK) inhibitors (89 lines, 18.9%), rituximab (82 lines, 17.4%), and abatacept (61 lines, 13.0%). The agents used included tocilizumab, levilimab, sarilumab, and olokizumab; adalimumab, etanercept, infliximab, golimumab, and certolizumab pegol; and tofacitinib, baricitinib, and upadacitinib. In adjusted GEE models, treatment type was not a significant predictor of AE occurrence (OR 0.93, 95% CI 0.63–1.36); line number and female sex were independently associated with AEs, and a previous AE increased the risk of AE on the current line (OR 1.72, 95% CI 1.13–2.60; p = 0.011). TNF-α inhibitors were prescribed without concomitant methotrexate in 59 of 109 treatment lines (54.1%). At the final treatment line, lower disease activity was observed with abatacept, levilimab, and olokizumab, whereas higher values were observed with rituximab and etanercept; between-treatment differences were significant for DAS28-ESR (p = 0.0269), but not for CDAI, SDAI, or DAS28-CRP. Cluster analysis identified three exploratory treatment patterns, differing primarily in the use of tsDMARDs and hydroxychloroquine. Conclusions: D2T-RA is characterized by multiple treatment lines and frequent switching due to inefficacy, particularly for csDMARD/GC therapy. Among the targeted therapies evaluated, IL-6 inhibitors and abatacept were associated with lower disease activity at the final treatment line, whereas rituximab and etanercept were associated with higher values; these exploratory observations do not establish comparative effectiveness. The frequent use of TNF-α inhibitors without concomitant methotrexate highlights a potential target for treatment optimization. A history of AEs may help identify patients requiring closer safety monitoring.