GLP-1 Drugs Beat SGLT2 Inhibitors for Weight Loss — But Only in Women and Non-Diabetics
A real-world study of nearly 4,000 patients reveals GLP-1 receptor agonists outperform SGLT2 inhibitors for weight loss, with striking sex and diabetes-status differences.
Summary
This real-world study compared 12-month weight loss between patients starting GLP-1 receptor agonists (like semaglutide) versus SGLT2 inhibitors (like empagliflozin) in a diverse US cohort of nearly 4,000 people. After rigorous statistical balancing, GLP-1 drugs produced about 1.2 percentage points greater body weight loss overall. The advantage was concentrated in women (−1.71 pp) and patients without type 2 diabetes (−1.96 pp), while men and those with type 2 diabetes showed no meaningful difference between the drug classes. GLP-1 initiators also achieved greater HbA1c reductions. With obesity and metabolic dysfunction central to accelerated aging, these findings support personalized drug selection for weight and cardiometabolic management rather than a one-size-fits-all approach.
Detailed Summary
Obesity and metabolic dysfunction are among the most powerful drivers of accelerated biological aging, cardiovascular disease, and reduced healthspan. Two drug classes — GLP-1 receptor agonists (GLP-1RAs) and SGLT2 inhibitors (SGLT2is) — are now widely prescribed for weight management and cardiometabolic risk reduction, but head-to-head real-world comparisons with robust methodology are scarce.
This retrospective active comparator new-user cohort study used the NIH All of Us Research Program (2018–2023), identifying 2,932 GLP-1RA initiators and 941 SGLT2i initiators with BMI ≥27 kg/m² and weight measurements at 12 months. Inverse probability of treatment weighting (IPTW) using a 17-covariate propensity score — including heart failure and chronic kidney disease — achieved excellent covariate balance across all 28 diagnostic checks.
After weighting, GLP-1RA users lost significantly more weight than SGLT2i users: a difference of −1.20 percentage points (95% CI −1.91 to −0.48; p = 0.001). The odds of achieving ≥10% weight loss were 75% higher with GLP-1RAs (OR 1.75), translating to a number needed to treat of 14. HbA1c fell more with GLP-1RAs (−0.261%; p = 0.032); blood pressure changes were similar between groups.
Critically, effect modification was substantial. The weight-loss advantage of GLP-1RAs was concentrated in patients without type 2 diabetes (−1.96 pp; p < 0.001) and essentially absent in those with type 2 diabetes (+0.08 pp; interaction p = 0.006). Sex was also a significant modifier: women gained −1.71 pp more weight loss with GLP-1RAs versus men, who showed virtually no difference (+0.12 pp; interaction p = 0.017).
These findings support individualized treatment selection when prescribing weight-loss medications. Clinicians managing obesity in women or patients without type 2 diabetes may see meaningfully greater weight-loss benefit from GLP-1RAs. Conversely, SGLT2is — with their established cardiovascular and renal protective benefits — may represent an equally effective or preferred option for men with type 2 diabetes. The authors appropriately flag the interaction findings as hypothesis-generating pending prospective confirmation.
Key Findings
- GLP-1RAs produced 1.20 percentage points greater 12-month weight loss than SGLT2 inhibitors overall (NNT = 14 for ≥10% weight loss).
- Women showed a −1.71 pp GLP-1RA weight-loss advantage; men showed virtually no difference (+0.12 pp) between drug classes.
- GLP-1RA weight-loss superiority was confined to obesity-only patients (−1.96 pp); patients with type 2 diabetes showed no meaningful difference.
- GLP-1RAs reduced HbA1c by an additional 0.261% compared to SGLT2 inhibitors; blood pressure changes did not differ significantly.
- Odds of achieving ≥10% body weight loss were 75% higher with GLP-1RAs (OR 1.75; 95% CI 1.40–2.19).
Methodology
Retrospective active comparator new-user cohort study using the NIH All of Us Research Program (2018–2023), comprising 3,873 patients with BMI ≥27 kg/m². IPTW with a 17-covariate propensity score achieved balance on all 28 diagnostic metrics (maximum SMD 0.077), with the primary outcome being percent body weight change at 12 months.
Study Limitations
Summary is based on the abstract only, as the full text is not open access. The study is retrospective and observational; despite robust propensity score weighting, unmeasured confounding cannot be excluded. Specific GLP-1RA agents (e.g., semaglutide vs. liraglutide) and SGLT2i agents were not individually analyzed, and the sex- and diabetes-status interaction findings are hypothesis-generating and require prospective validation.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: