OBJECTIVES: To estimate the association between abatacept use and the incidence of malignancy excluding non-melanomatous skin cancers (NMSCs). METHODS: Systematic database searches were performed, to April 2024, to identify phase II/III/IV randomised clinical trials (RCTs), long-term extension (LTE) and observational cohort studies of abatacept in people with rheumatoid arthritis and psoriatic arthritis. Network and pairwise meta-analyses were performed to estimate incidence rate ratios (IRRs) for malignancy excluding NMSC, comparing abatacept with placebo and tumour necrosis factor inhibitors (TNFi) in RCT/LTE studies. Pairwise meta-analyses evaluated the same outcome in observational studies, comparing abatacept with conventional synthetic DMARDs (csDMARDs) and biologic/targeted synthetic disease modifying antirheumatic drugs (b/tsDMARDs). RESULTS: In 18 eligible RCTs and 10 LTE studies, there were 15 535 person-years of exposure to abatacept, 1495 to placebo, and 733 to TNFi. In network meta-analyses of combined RCT/LTE data, the incidence of all malignancies excluding NMSCs was not significantly different between abatacept and placebo (IRR 0.58
95% CI 0.32-1.09) or TNFi (IRR 0.72
95% 0.27-1.87). In observational data, the incidence of malignancy was higher with abatacept, relative to other b/tsDMARDs (IRR 1.21
95% CI 1.15-1.28), but not significantly different compared with csDMARDs (IRR 0.97
95% CI 0.90-1.06). CONCLUSIONS: Abatacept was associated with a higher incidence of malignancy compared with other b/tsDMARDs in observational studies, but not when compared with placebo or TNFi in RCT/LTE data. Further pharmacovigilance data is essential to help elucidate whether abatacept modifies cancer risk. PROSPERO REGISTRATION NUMBER: CRD42023382314.