P672 Occurrence of malignancy among infliximab biosimilar and bio-originator initiators in Canada: a comparative population-based analysis

Abstract

Abstract Background Infliximab (INF) downregulates the immune system. Given the immune system's role in deleting pre-cancerous cells, malignancy is a potential concern. Still, relatively few real-world data analyses have compared biosimilar versus bio-originator INF in terms of malignancy incidence. We aimed to compare malignancy occurrence among initiators of infliximab biosimilar (INF-B) and originator (INF-O). Methods We used data from the National Prescription Drug Utilization Information System (NPDUIS), containing pan-Canadian (except Quebec) claims-level data on prescription dispensations paid from public drug programs linked to comprehensive day surgery and inpatient care (Discharge Abstract Database, DAD) and hospital-based and community-based ambulatory care (National Ambulatory Care Reporting System, NACRS). We studied adults (≥18 years) initiating INF between January 2015 and December 2019 with no history of malignancy, HIV or organ transplant (1-year pre-baseline). Follow-up began 365 days after treatment initiation (lag period) and continued until the first cancer diagnostic code (ICD-10 C00-C97), treatment discontinuation + 365 days, or end of study period. We compared INF-B and INF-O using Cox regression models, adjusting for potential confounders/ effect modifiers: sex, age at treatment initiation, prior use of other biologics or prednisone, province (Ontario versus others), and calendar year. Results New users of INF-B (2,019) and INF-O (5,183) were about half female (52%), with a median age (interquartile range) of 45 years (28-62). Overall, the malignancy incidence rate was 10.2 (95% confidence interval, CI, 8.7-11.8) events per 1,000 person-years. Comparing INF-B to INF-O, the adjusted hazard ratio for malignancy was 0.97 (95% CI 0.63-1.48). Conclusion In this real-world dataset, we were unable to identify clear differences in malignancy comparing INF-B and INF-O. Limitations include the inability to control for residual confounders (e.g., disease severity), potential outcome misclassification, and selection bias.