Summary: In this prespecified secondary analysis of the SELECT trial (17,604 adults with overweight or obesity and established cardiovascular disease but without diabetes), semaglutide 2.4 mg weekly raised the proportion who were normoglycaemic at week 156 (69.5% vs 35.8%) and lowered progression to biochemical diabetes (1.5% vs 6.9%; NNT 18.5). The mean HbA1c difference was modest at -0.32 percentage points, and after an early nadir HbA1c rose in parallel in both arms, so the drug did not slow glycaemic progression over time.
PICO Summary
| Element | Detail |
|---|---|
| Population | 17,604 adults aged ≥45 years with BMI ≥27 kg/m², preexisting cardiovascular disease, and no diabetes (HbA1c <6.5%); multinational, multicentre, double-blind randomised controlled trial (secondary analysis of SELECT). Mean intervention exposure 152 ± 56 weeks. |
| Intervention | Subcutaneous semaglutide 2.4 mg once weekly (n = 8,803). |
| Comparison | Matched placebo injection once weekly (n = 8,801). |
| Outcome | Mean HbA1c difference -0.32 percentage points (95% CI -0.33 to -0.30; -3.49 mmol/mol, 95% CI -3.66 to -3.32), favouring semaglutide throughout (P < 0.0001). At week 156, normoglycaemia (HbA1c <5.7%) in 69.5% vs 35.8% and biochemical diabetes (HbA1c ≥6.5%) in 1.5% vs 6.9% (both P < 0.0001); number needed to treat 18.5 to prevent one case of diabetes. Body weight plateaued at 65 weeks and was 8.9% lower with semaglutide. After a shared HbA1c nadir at 20 weeks, HbA1c rose similarly in both arms, so semaglutide did not slow glycaemic progression over time. Weight reduction mediated 24.5% of progression and 27.1% of regression. |
Semaglutide and glycaemic regression in SELECT
Secondary analysis of RCT · overweight/obesity, CVD, no diabetes · 156 weeks
Semaglutide nearly doubled the proportion who were normoglycaemic at week 156 and cut progression to biochemical diabetes to about a fifth, but HbA1c rose in parallel in both arms after an early nadir, so it shifted the starting point without slowing glycaemic progression.
Expert Commentary
This prespecified secondary analysis is best read as hypothesis-generating rather than as a glycaemic efficacy trial. The headline proportions are striking, with nearly twice as many semaglutide-treated participants normoglycaemic at week 156 and progression to biochemical diabetes cut to roughly a fifth, but the interpretation hinges on a nuance the original framing obscured. After a shared nadir at 20 weeks, HbA1c rose in parallel in both arms, so semaglutide shifted the starting point without altering the slope and did not slow glycaemic progression over time. The mean HbA1c separation of -0.32 percentage points is small, and the categorical outcomes rest on threshold crossings near that shifted baseline rather than on a divergent trajectory. The principal limitation is that diabetes was defined biochemically by a single HbA1c rather than by a clinical diagnosis with confirmatory testing, which inflates apparent regression and progression and limits transfer to practice. The trial was funded by the manufacturer and several authors are employees, and these were not the primary cardiovascular endpoints. Can I use this with my patients? Cautiously yes for an adult with obesity and established cardiovascular disease already prescribed semaglutide 2.4 mg, where lower diabetes risk is a welcome secondary benefit, but not as a standalone reason to start the drug. I would welcome a trial powered for incident clinically diagnosed diabetes with durable post-treatment follow-up.
References
Kahn SE, Deanfield JE, Jeppesen OK, Emerson SS, Boesgaard TW, Colhoun HM, et al. Effect of semaglutide on regression and progression of glycemia in people with overweight or obesity but without diabetes in the SELECT trial. Diabetes Care. 2024;47(8):1350-1359. doi:10.2337/dc24-0491
