(A-312) Study design and comparator choices in evaluating the risk of suicidal ideation and suicidality with semaglutide: A real-world study

Background: Semaglutide, a GLP-1 analog, is prescribed for both type 2 diabetes and weight management. After reports of suicidal ideation (SI) prompted regulatory review, a real-world study estimated a HR of 0.27 for SI comparing new users of semaglutide to prevalent users of weight management ingredient (WMI) drugs (bupropion, naltrexone, orlistat, phentermine, or topiramate). Rigorous pharmacoepidemiologic designs with active comparators are warranted.

Objectives: To investigate the effect of semaglutide on SI and suicidality and the effect of study design and comparators choices on the association with real-world data.

Methods: We conducted two different real-world studies in the Merative® MarketScan™ Databases. We identified overweight or obese patients who initiated semaglutide, WMI drugs, or active comparators (combination bupropion/naltrexone, orlistat, phentermine, or combination phentermine/topiramate) and followed them for SI and suicidality (SI, suicide attempt, or intentional self-harm). We compared semaglutide to WMI drugs (Study 1) and active comparators (Study 2) using two distinct study designs: New user vs. Prevalent-user (NUPU) and New user (NU) to evaluate the influence of study design and comparator choice. WMI and active comparators patients were weighted to resemble semaglutide initiators to reduce confounding. We estimated risk ratios (RR) for SI and suicidality within the 183 days of drug initiation using Kaplan-Meier methods with 95% confidence intervals from 1,000 bootstrap replicates.

Results: In Study 1, semaglutide was associated with a reduced SI risk (RR 0.49 [0.32–0.73]) versus WMI drugs (NUPU), but this association was attenuated with the NU design (SI: RR 0.74 [0.39–1.13]). In Study 2, semaglutide was associated with a higher SI risk (NUPU: RR 1.87 [1.00–4.22]; NU: RR 1.36 [0.62–6.14]) compared to active comparators. These results were similar when examining suicidality.

Conclusions: Semaglutide was not associated with increased SI or suicidality when using an appropriate study design and compared to active comparators. Careful consideration of study design and appropriate comparator choice are essential in drug safety studies.