Objective:

To evaluate the association between initiating GLP-1 receptor agonists (GLP-1 RAs) and obesity-related cancer (ORC) incidence compared with other glucose-lowering therapies, across levels of baseline type 2 diabetes (T2D) severity.

Approach:

• Design and population: A cohort study using PCORnet electronic health record data emulated five active-comparator, new-user trials among adults with T2D and no prior ORC who initiated a GLP-1 RA or a comparator after a 365-day washout.
• Comparators and severity: Comparators were metformin, insulin, an SGLT2 inhibitor, a DPP4 inhibitor, and a sulfonylurea. Baseline T2D severity was categorized as low, moderate, or high based on documented complications.
• Analysis: Propensity-score weighting and adjusted regression models estimated 3- and 5-year ORC risks, risk differences (RDs), and hazard ratios (HRs). Severity comparisons were exploratory.

Key Findings:

• Across 3-year analyses, GLP-1 RA initiation was associated with lower ORC risk than insulin (RD, −1.25%; HR, 0.62) and sulfonylureas (RD, −1.14%; HR, 0.67).
• At 5 years, the RD versus insulin was −1.67%; HRs were 0.61 versus insulin and 0.75 versus sulfonylureas.
• Estimates versus metformin, SGLT2 inhibitors, and DPP4 inhibitors were closer to the null and were not consistently different.
• The clearest pattern was among adults with lower T2D severity. In the high-severity group, the 3-year RD favored SGLT2 inhibitors (1.61%; 95% CI, 0.17% to 2.89%).

Interpretation:

The authors reported lower ORC incidence with GLP-1 RAs than with insulin or sulfonylureas, particularly among adults with lower T2D severity. They stated that the findings do not justify selecting glucose-lowering therapy to prevent cancer and that established metabolic, cardiovascular, and kidney benefits should remain the basis for treatment decisions.

Limitations:

• Retrospective, single-system EHR data may limit generalizability and leave residual confounding or incomplete treatment-group overlap despite weighting.
• Exposure and outcome misclassification, broad severity categories, and sparse events in some strata may affect estimates.
• Treatment pathways differed: insulin was often initiated later in T2D, while SGLT2 inhibitor prescribing was influenced by cardiorenal indications.
• Many GLP-1 RA initiators entered the cohort recently, so 5-year estimates relied disproportionately on earlier initiators.

Conclusion:

The study found lower ORC incidence associated with GLP-1 RA initiation versus insulin or sulfonylureas, but did not establish treatment superiority for cancer outcomes.

Sources:

• Article context provided by the user