(CSPOT17) A post-authorization safety study of golimumab in ulcerative colitis using the Spanish ENEIDA registry (MK8259-042)

Background: Simponi (golimumab [GLM), an anti–tumor necrosis factor α (anti-TNFα) agent, was approved to treat moderately to severely active ulcerative colitis (UC) in 2013. Given theoretical concerns about a greater risk of malignancy and colectomy with anti-TNFα therapy, a post-authorization study was conducted under real-world conditions.

Objectives: To evaluate the risk of colectomy due to intractable disease (hereafter “colectomy”) and advanced colonic neoplasia (ACN) (colorectal cancer or high-grade dysplasia) in UC among new GLM users compared with new users of (i) other anti-TNFα agents and (ii) thiopurines (TPs).

Methods: This observational study used a new-user cohort design with a secondary nested case-control analysis. It used data from the ENEIDA registry, a database maintained by the Spanish Working Group on Crohn's Disease and Ulcerative Colitis (GETECCU). Patients with UC were included who were aged ≥ 18 years, initiated GLM, an anti-TNFα agent other than GLM, or TP from September 2013 through December 2021; and had not had a study outcome before study entry. Crude risk analyses were performed for each outcome, and multivariable models were fit for outcomes with sufficient events.

Results: There were 474 patients in the GLM cohort, 1,737 patients in the other anti-TNFα cohort, and 1,380 patients in the TP cohort identified in 30 hospitals. Median age was 43 years. A total of 64 colectomy cases and 10 ACN cases occurred over the 8.5-year study.
Incidence rates (IRs) of colectomy were 4.4 (95% confidence interval [CI], 1.2-11.2) per 1,000 person-years (PY) with GLM use only, 12.4 (95% CI, 9.1-16.5) per 1,000 PY with other anti-TNFα agent use only, and 78.6 (95% CI, 16.2-229.7) per 1,000 PY with overlapping exposure to both GLM and other anti-TNFs around times of therapy switch. Compared with use of only other anti-TNFα agents, GLM-only use was not associated with an increased risk of colectomy in adjusted Poisson and Cox models (IR ratio = 0.40; 95% CI, 0.14-1.13; hazard ratio = 0.41; 95% CI, 0.15-1.15). (The TP comparator was evaluated only for ACN risk). The IR of ACN was 1.5 (95% CI, 0.2-5.4) per 1,000 PY in the GLM cohort, 1.3 (95% CI, 0.5-2.8) per 1,000 PY in the other anti-TNFα agent cohort, and 1.0 (95% CI, 0.3-2.6) per 1,000 PY in the TP cohort. Compared with use of other anti-TNFα agents, GLM use was not associated with an increased risk of ACN (adjusted hazard ratio = 1.09; 95% CI, 0.22-5.44). Similar results were found for the comparison with TPs. For both outcomes, results from age-adjusted nested case-control analyses were consistent with the main results.

Conclusions: Results from this study do not indicate an increased risk of colectomy due to intractable disease or of ACN among GLM users compared with users of other anti-TNFα agents in adult patients with UC in Spain.