Details
In this report, we aimed to assess the risk of incident renal cell carcinoma (RCC) among adults with evidence of type 2 diabetes mellitus had they been randomized to initiate canagliflozin, a sodium-glucose cotransporter-2 inhibitor, or a comparator drug sitagliptin, and to achieve at least 180 days of cumulative exposure in the first 360 days of follow-up (treatment strategy of interest). We used a clone-censor-weighting design for this analysis. We applied treatment propensity score-based (i.e., inverse probability of treatment [IPT]) weighting to balance baseline characteristics between the initiator groups, and inverse probability-of-censoring (IPC) weighting to account for potential selection bias due to artificial censoring associated with treatment strategy non-adherence. We estimated the RCC hazard ratio (HR) we would have observed had all canagliflozin initiators been randomized to canagliflozin versus sitagliptin and accrued at least 180 days of exposure in the first 360 days of follow-up in the Sentinel Distributed Database (SDD).
Follow-up analyses were conducted post hoc to assess the balance in distribution of post-index IPC-weighted and IPT/IPC-weighted covariates among individuals with at least 360 days of follow-up, and to examine IPT-weighted and IPC-weighted HRs separately (rather than using the combined IPT and IPC weights).
The study period included data from March 1, 2013 to March 31, 2025. We distributed the initial request to six Sentinel Data Partners on November 6, 2025 and the follow-up request to the same partners on March 20, 2026.
The analytic package associated with this analysis can be found externally in Sentinel's Git Repository located here. The Git Repository serves as Sentinel's version control tracking system for analytic packages and technical documentation.