The incretin evidence base now runs to 38 trials, and the numbers are no longer close

Design

updated systematic review of randomised controlled trials of ≥16 weeks, narrative synthesis, PROSPERO registered

Population

38 trials, 25,816 adults with overweight or obesity and without diabetes

Primary outcome

placebo-subtracted percentage weight loss, with adverse events and discontinuation

Effect

−14.8% subcutaneous semaglutide (95% CI −16.2 to −13.4), −19.0% tirzepatide (−21.6 to −16.4), −5.8% liraglutide (−8.0 to −3.6); gastrointestinal adverse events 76.0% vs 40.1%

This updated Annals systematic review pooled 38 randomised trials of at least 16 weeks in 25,816 adults with overweight or obesity and without diabetes, adding 14 trials and 11,000 participants to the previous version. It is a narrative synthesis — heterogeneity ruled out meta-analysis — so these are placebo-subtracted figures from individual trials, not pooled estimates.

Among commercially available agents, placebo-subtracted weight loss reached −5.8% for liraglutide (95% CI −8.0 to −3.6), −14.8% for subcutaneous semaglutide (−16.2 to −13.4), −14.3% for oral semaglutide (−17.2 to −11.4), −12.4% for orforglipron (−15.1 to −9.7) and −19.0% for tirzepatide (−21.6 to −16.4). Emerging multiagonists went further: −23.9% with amycretin and −22.1% with retatrutide. Head-to-head data favoured semaglutide over liraglutide, and tirzepatide and cagrilintide-semaglutide over semaglutide.

Safety is where a prescriber should slow down. Gastrointestinal adverse events occurred in 76.0% on active drug against 40.1% on placebo. Discontinuation for adverse events was 10.7% versus 3.4%, higher with some oral agents. Serious adverse events (6.5% vs 5.2%) and deaths (0.1% vs 0.0%) were rare and no new safety signals emerged — but 'no new signal' in trials of 16 weeks to two years is not the same as long-term safety in a drug people may take for decades.

• Liraglutide is now clearly the weakest of the available options — there is little reason to start it when semaglutide is accessible.
• Oral semaglutide performs close to the subcutaneous form, which matters where cold chain or injection acceptance is the barrier.
• Counsel on gastrointestinal effects as the expected course, not a complication: three in four patients get them.
• Amycretin and retatrutide are not licensed anywhere — do not let their numbers enter a conversation about what to start.
• In India, availability and price differ sharply between these agents and change every few months; check current status before naming one to a patient.

Why it matters

The question in clinic has moved from whether to offer these drugs to which one, and this is the first synthesis with enough head-to-head data to answer it.

Don't overread it

Heterogeneity precluded quantitative synthesis, so the between-drug differences here are mostly indirect comparisons, not a ranking the evidence establishes.

The statistics, in plain English

Heterogeneity prevented the authors pooling these trials, which means the figures are not directly comparable to each other — different populations, durations, and background lifestyle support sit behind each one. The confidence intervals shown are within-comparison, so a −19.0% and a −14.8% from separate trials cannot be subtracted to give a 4.2% advantage; only the head-to-head trials support ranking. Placebo-subtracted also means the absolute weight loss a patient sees is larger than these numbers, because the placebo arms lost weight too.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.