Baricitinib (Olumiant) did not demonstrate noninferiority to tumor necrosis factor inhibitors (TNFi) for venous thromboembolism (VTE) risk in patients with rheumatoid arthritis (RA) enriched for VTE risk factors, according to pooled long-term data from the RA-BRIDGE and RA-BRANCH post-marketing safety trials presented at the European Congress of Rheumatology (EULAR) 2026 Congress.1
The hazard ratio for time to first adjudicated VTE was 1.61 (95% CI, 0.97-2.66; P = .066) for combined baricitinib doses versus TNFi, with the upper confidence bound exceeding the prespecified noninferiority margin of 1.8.1 Serious infections, including new-onset COVID-19, were significantly more frequent with baricitinib.1
Two FDA-mandated post-approval safety studies were designed specifically to generate comparative long-term data in an RA population enriched for VTE, cardiovascular, malignancy, and infection risk. RA-BRIDGE and RA-BRANCH enrolled older patients and those with obesity to maximize the chance of detecting a safety signal where one might exist.
What did RA-BRIDGE and RA-BRANCH Show?
RA-BRIDGE (NCT03915964, global) and RA-BRANCH (NCT04086745, US-based) are randomized, open-label, active-controlled safety trials in adults with moderate to severe RA and at least one predefined VTE risk factor, including prior VTE, age 60 years or older, body mass index (BMI) of 30 kg/m² or higher, or age 50 to under 60 years with BMI 25 to under 30 kg/m².1 Patients were randomized 1:1:1 to baricitinib 2 mg once daily, baricitinib 4 mg once daily, or a TNFi (adalimumab or etanercept, per local label).1 Data from both trials were pooled per a prespecified analysis plan, yielding 3,640 patients (baricitinib 2 mg, n = 1,219; baricitinib 4 mg, n = 1,214; TNFi, n = 1,207).1
Combined baricitinib exposure totaled 7,694.4 patient-years versus 3,830.1 patient-years for TNFi, with median on-treatment exposure of approximately 3.7 years in both groups.1 For the primary endpoint, baricitinib produced a VTE incidence rate of 0.79 per 100 patient-years (95% CI, 0.60-1.01; 62 events) versus 0.51 per 100 patient-years for TNFi (95% CI, 0.31-0.79; 20 events), for a hazard ratio of 1.61 (95% CI, 0.97-2.66; P = .066).1 The upper confidence bound exceeded the prespecified margin of 1.8, so noninferiority was not demonstrated.1
Results were consistent across doses, with a hazard ratio of 1.70 (95% CI, 0.94-3.08) for baricitinib 2 mg and 1.69 (95% CI, 0.97-2.94) for baricitinib 4 mg, without a dose-dependent pattern.1
Key Facts
Did baricitinib show noninferiority to TNF inhibitors for VTE risk in RA-BRIDGE and RA-BRANCH?
No. The hazard ratio was 1.61 (95% CI, 0.97-2.66), and the upper confidence bound exceeded the prespecified noninferiority margin of 1.8.
Were serious infections more common with baricitinib than TNF inhibitors?
Yes. Serious infection incidence rates were 3.17 versus 2.46 per 100 patient-years (hazard ratio, 1.32; 95% CI, 1.04-1.68), partly driven by new-onset COVID-19.
Was MACE risk increased with baricitinib in this enriched population?
No. MACE risk was similar between baricitinib and TNF inhibitors (hazard ratio, 1.06; 95% CI, 0.64-1.76).
What Were the Secondary Safety End Points?
Major adverse cardiovascular event (MACE) risk was similar between combined baricitinib and TNFi groups (hazard ratio, 1.06; 95% CI, 0.64-1.76).1 Malignancy risk excluding nonmelanoma skin cancer was also similar (hazard ratio, 1.27; 95% CI, 0.85-1.89), though incidence rates were numerically higher with baricitinib 2 mg (1.28 per 100 patient-years; 95% CI, 0.90-1.75) than baricitinib 4 mg (0.94; 95% CI, 0.67-1.29) or TNFi (0.83; 95% CI, 0.57-1.15).1 Arterial thromboembolic events, opportunistic infections, and mortality were similar across groups.1
Serious infections were significantly more frequent with baricitinib than TNFi, with incidence rates of 3.17 versus 2.46 per 100 patient-years (hazard ratio, 1.32; 95% CI, 1.04-1.68).1 COVID-19 accounted for approximately 20% of serious infection events in the baricitinib group versus approximately 13% in the TNFi group.1 Treatment discontinuation rates were balanced between arms, at 55.2% for baricitinib and 57.3% for TNFi, with similar rates attributable to adverse events (about 14%) and patient withdrawal (14%).1
The investigators noted the enriched study population, selected specifically for VTE risk factors and elevated cardiovascular, malignancy, and infection risk, may limit direct extrapolation to broader RA populations with lower baseline risk.1 They characterized the findings as informing the ongoing benefit-risk profile of baricitinib and supporting individualized treatment decisions based on patient-specific VTE and infection risk factors.1
References
Taylor PC, Curtis J, Giles JT, et al. Long-Term Safety Evaluation of Baricitinib vs. TNF Inhibitors in Rheumatoid Arthritis Patients Selected for Venous Thrombo-Embolism Risk. Abstract LB0009. Presented at: European Congress of Rheumatology (EULAR) 2026; June 3-6, 2026; London, England. doi:10.1136/annrheumdis-2026-eular.LBA_B.23
Molander V, Bower H, Frisell T, Delcoigne B, Di Giuseppe D, Askling J. Venous thromboembolism with JAK inhibitors and other immune-modulatory drugs: a Swedish comparative safety study among patients with rheumatoid arthritis. Ann Rheum Dis. 2023;82(2):189-197. doi:10.1136/ard-2022-223050
