Emulation of a randomized controlled trial in ulcerative colitis with US and French claims data: Infliximab with thiopurines compared to infliximab monotherapy. Issue 2 (26th September 2021)

• Record Type: Journal Article
• Authors: Kirchgesner, Julien; Desai, Rishi J.; Schneeweiss, Maria C.; Beaugerie, Laurent; Kim, Seoyoung C.; Schneeweiss, Sebastian
• Is Part Of: Pharmacoepidemiology and drug safety. Volume 31:Issue 2(2022)
• Page Start: 167
• Page End: 175
• Publication Date: 2021-09-26
• Subjects: anti‐TNF -- RCT emulation -- real‐world evidence -- ulcerative colitis; Pharmacoepidemiology -- Periodicals; Chemotherapy -- Periodicals; Epidemiology -- Periodicals; 615.705
• DOI: 10.1002/pds.5356
• Languages: English
• ISSNs: 1053-8569

Abstract

Abstract: Purpose: To understand the validity of real‐world evidence (RWE) studies in ulcerative colitis (UC), we emulated the SUCCESS randomized controlled trial (RCT) on the effectiveness of infliximab plus thiopurines, using US and French healthcare insurance claims data. Methods: The SUCCESS trial showed improved remission with infliximab plus thiopurines combined compared to infliximab monotherapy in patients with UC. Based on two US commercial claims databases (IBM MarketScan and Optum) and the French nationwide health insurance database (SNDS) from 2004 through 2019, all patients with UC who initiated combination therapy or infliximab alone were identified. The primary outcome of treatment failure was emulated by: Hospitalization related to UC or colectomy, treatment switch to another biologic or immunosuppressant, or use of corticosteroids 16 weeks after infliximab initiation. We estimated risk ratios with 95% confidence intervals after 1:1 propensity score (PS) matching. Results: Among 620 PS‐matched pairs of combination therapy and infliximab monotherapy users, treatment failure occurred in 124 (20%) of patients initiating combination therapy and 170 (27%) during monotherapy. Like in SUCCESS, the risk of treatment failure was decreased with combination therapy in the overall cohort (RR = 0.73; 95% CI: 0.60–0.90). Findings were consistent across MarketScan, Optum, and SNDS: RR = 0.76 (0.57–1.02), 0.82 (0.54–1.24), and 0.61 (0.41–0.90). Similar results were observed for each component endpoint. Conclusions: RWE results across three large claims databases were consistent with RCT findings. These findings provide support for the use of RWE to assess treatment effectiveness in UC.