Eighteen months, 14 new trials: the obesity-drug evidence base has doubled in size

In January 2025, Annals of Internal Medicine published a systematic review of GLP-1 receptor agonists and co-agonists for weight loss in adults without diabetes. It covered 26 randomised trials and 15,491 participants across 12 agents, three of them commercially available: liraglutide, semaglutide and tirzepatide [s2]. An updated version published on 1 September covers 38 trials and 25,816 participants — 14 new trials adding 11,000 people [s1].

Read side by side, the two reviews are a compressed record of how fast this field is moving and of what has and has not changed.

What the update reports

The updated review searched MEDLINE, Embase and the Cochrane Central Register of Controlled Trials from 5 October 2024 through 25 March 2026, including randomised controlled trials of at least 16 weeks' duration [s1]. Its placebo-subtracted weight-loss figures for the commercially available therapies are:

• liraglutide, up to −5.8% (95% CI, −8.0% to −3.6%)
• subcutaneous semaglutide, −14.8% (CI, −16.2% to −13.4%)
• oral semaglutide, −14.3% (CI, −17.2% to −11.4%)
• orforglipron, −12.4% (CI, −15.1% to −9.7%)
• tirzepatide, −19.0% (CI, −21.6% to −16.4%) [s1]

Numerically greater reductions were seen with emerging multiagonists: −23.9% (CI, −29.3% to −18.5%) with amycretin and −22.1% (CI, −24.9% to −19.3%) with retatrutide [s1].

Head-to-head data — absent from the earlier review, which noted no head-to-head trials were available [s2] — now show greater weight loss with semaglutide and JNJ-64565111 than liraglutide, and greater weight loss with tirzepatide and cagrilintide-semaglutide than with semaglutide [s1].

What the comparison shows

Two shifts are visible.

The first is that the ceiling for the marketed agents has risen. The 2025 review put tirzepatide at up to 17.8% (CI, 16.3% to 19.3%) after 72 weeks and semaglutide 2.4 mg at up to 13.9% (CI, 11.0% to 16.7%) after 68 weeks [s2]; the update reports 19.0% and 14.8% [s1]. Those are not directly comparable estimates — the earlier figures are tied to specific doses and durations, the newer ones summarise a larger and differently composed evidence base — but the direction is consistent.

The second is the arrival of oral agents at parity. Oral semaglutide at −14.3% and orforglipron at −12.4% [s1] sit close to injectable semaglutide's −14.8% [s1]. Neither had comparable representation in the earlier review, which covered nine premarket agents alongside the three marketed ones [s2].

What has not changed is retatrutide's position at the top of the experimental field. The 2025 review reported it at up to 22.1% (CI, 19.3% to 24.9%) after 48 weeks [s2]; the update reports −22.1% (CI, −24.9% to −19.3%) [s1].

Safety: more data, same shape

Gastrointestinal adverse events remained common in the update — 76.0% on a GLP-1 drug versus 40.1% on placebo [s1]. Discontinuation because of adverse events was generally low, 10.7% versus 3.4%, but numerically higher with some oral agents [s1]. Serious adverse events (6.5% vs 5.2%) and deaths (0.1% vs 0.0%) were rare, and the authors identified no new safety signals [s1].

The earlier review found the same shape with wider ranges, reflecting a smaller evidence base: adverse events in 80% to 97% of GLP-1 arms versus 63% to 100% of placebo arms, mostly gastrointestinal (47% to 84% vs 13% to 63%), with discontinuation for adverse events at 0% to 26% versus 0% to 9% and serious adverse events at 0% to 10% versus 0% to 12% [s2].

The narrowing of those ranges is itself informative. It reflects more trials with more consistent reporting, not necessarily safer drugs.

The limits both reviews name

Neither review could pool its results. Heterogeneity precluded quantitative synthesis in the update [s1] and prevented meta-analysis in the original [s2]. Every figure above is a trial-derived estimate, not a meta-analytic one, which means the confidence intervals describe individual trials rather than the field.

The update also states that safety outcomes were inconsistently reported [s1]. For a drug class now being prescribed at population scale to people without diabetes, that is the more consequential of the two limitations. Efficacy is measured with a scale; harms are measured with whatever each trial protocol chose to collect.

Both reviews are registered on PROSPERO under the same identifier, CRD42024505558, and both report no primary funding source [s1][s2].

The conclusion the update draws is narrow and correct: GLP-1 receptor agonists demonstrate substantial weight loss in adults without diabetes, with an expanding range of therapeutic options including oral and multiagonist therapies [s1]. What it cannot speak to — because the trials mostly do not run long enough — is what happens across the years these drugs are now expected to be taken.

Sources

• [s1] "Efficacy and Safety of Glucagon-like Peptide-1 Receptor Agonists and Co-agonists for Weight Loss Among Adults Without Diabetes: An Updated Systematic Review," Annals of Internal Medicine, 1 September 2026. https://doi.org/10.7326/ANNALS-25-05519
• [s2] "Efficacy and Safety of Glucagon-Like Peptide-1 Receptor Agonists for Weight Loss Among Adults Without Diabetes: A Systematic Review of Randomized Controlled Trials," Annals of Internal Medicine, 7 January 2025. https://doi.org/10.7326/ANNALS-24-01590