Background: There are discordant results regarding a potential increased risk of gastro-intestinal perforation (GIP) in rheumatoid arthritis (RA) patients treated with tocilizumab (TCZ) compared to conventional synthetic disease modifying anti rheumatic drugs (csDMARDs) or TNF inhibitors (TNFis) (1–3).

Objectives: The aim of our study was to compare the risk of diverticulitis and GIP in RA patients treated with TCZ compared to rituximab (RTX) and abatacept (ABA).

Methods: We conducted a multicentric study of patients with RA, prospectively followed in 3 observational French registries evaluating the effectiveness and safety of RTX (Autoimmunity and Rituximab (AIR)), ABA (Orencia and Rheumatoid Arthritis (ORA)), and TCZ (REGistry–RoAcTEmra (REGATE)). Using a propensity score approach, we compared the risk of diverticulitis or GIP during treatment with TCZ vs RTX and ABA. The following covariates were included in the propensity score: age, sex, history of diabetes and neoplasia, Charlson Comorbidity Index, number of previous csDMARDs and TNFi, history of TNFi, daily dose of glucocorticoids (GCs) at baseline, co-treatment with a csDMARDs, average DAS28 during follow-up, duration of RA, and exposure time to the considered bDMARDs.

Results: 4501 patients (1496 treated by TCZ, 1986 by RTX and 1019 by ABA) were included. 21 and 9 GIP occurred in the TCZ treated patients, compared to 10 and 8 in the RTX treated patients and 10 and 2 in the ABA treated patients (corresponding incidence rate (IR) are shown in table 1). Two deaths occurred in patients experiencing GIP: 1 (12.5%) due to undetermined rectal perforation among a RTX treated patient, and 1 (11.1%) due to a perforated ulcer among a TCZ treated patient. Based on inverse probability weighting (IPW), there was an increased risk of diverticulitis and GIP in the TCZ treated patients compared with RTX or ABA (table 1). In a subgroup analysis, we confirmed an increased risk of GIP due to diverticulitis but not to any other etiology. Older age (p=0.05), GCs at baseline (p=0.10) and average daily dose of GCs during follow-up (p=0.08) seemed associated with GIP only in univariate analysis. Compared to RTX and ABA, diverticulitis and GIP among TCZ patients occurred earlier after the last perfusion (p=0.01), with atypical clinical presentation (slow transit in 30%, p=0.04) and lower acute phase reactants when the event occurred (C-reactive protein: 31.2±58.4 vs 88.2±89.6 mg/L, p=0.005). Perforated diverticulitis seemed to have higher dose of GCs at the time of the event compared to diverticulitis without perforation in univariate analysis (p=0.06).

Conclusion: TCZ was associated with an increased risk of diverticulitis, and GIP due to diverticulitis, compared to RTX and ABA. Our study confirms an increased risk of GIP in RA patients treated with TCZ, which might be explained by an increased risk of diverticulitis with misleading clinical presentation.

References: [1]Strangfeld A et al. Ann Rheum Dis. 2016 Jul 12

[2]Xie F, Yun H et al. Arthritis Rheumatol. 2016 May 1

[3]Barbulescu A et al. OP0231, Ann Rheum Dis. 2018 Jun 1;77(Suppl 2):164–5.

Disclosure of Interests: Claire Rempenault: None declared, Cédric Lukas: None declared, Bernard Combe Grant/research support from: Novartis, Pfizer, Roche-Chugai, Consultant of: AbbVie; Gilead Sciences, Inc.; Janssen; Eli Lilly and Company; Pfizer; Roche-Chugai; Sanofi, Speakers bureau: Bristol-Myers Squibb; Gilead Sciences, Inc.; Eli Lilly and Company; Merck Sharp & Dohme; Pfizer; Roche-Chugai; UCB, Thierry Schaeverbeke: None declared, Daniel Wendling: None declared, Thao Pham Speakers bureau: Novartis, Janssen, Lilly, Xavier Mariette Consultant of: BMS, Gilead, Medimmune, Novartis, Pfizer, Servier, UCB, Jacques-Eric Gottenberg Grant/research support from: BMS, Pfizer, Consultant of: BMS, Sanofi-Genzyme, UCB, Speakers bureau: Abbvie, Eli Lilly and Co., Roche, Sanofi-Genzyme, UCB, Jacques Morel: None declared