Cochrane review delivers a cautious verdict on liraglutide for obesity

Cochrane, the international group that grades medical evidence, published a review on October 30 that takes a cautious view of liraglutide, one of the older GLP-1 receptor agonists used for obesity. Pooling 24 randomized controlled trials with 9,937 participants, the reviewers concluded the drug likely increases the share of people who lose at least 5% of their body weight, while calling the evidence on many other outcomes limited or very uncertain ^[1].

The review was funded by the World Health Organization and is one of three Cochrane reviews looking at GLP-1 drugs in adults living with obesity ^[1]. Authors searched five databases and two trial registries through December 17, 2024. Participants ranged in average age from 31.3 to 64.7 years and came mostly from middle- and high-income countries. Most trials compared liraglutide with placebo; two used semaglutide as an active comparator, one used orlistat, and two compared liraglutide with structured lifestyle programs. Seventeen trials focused on specific groups, such as people with diabetes, fatty liver disease, polycystic ovary syndrome, or obstructive sleep apnea ^[1].

What the numbers show

At medium-term follow-up, defined as 26 to 68 weeks, people taking liraglutide were about twice as likely as those on placebo to reach at least 5% weight loss (risk ratio 2.10, 95% confidence interval 1.80 to 2.45; 18 studies, 6,651 participants). Cochrane rated that finding moderate certainty, its second-highest grade ^[1].

Other results were graded lower. The average difference in percentage weight change was 4.72 percentage points more loss with liraglutide (95% CI -5.32 to -4.12), but the reviewers called that evidence very uncertain because results varied widely across trials ^[1]. Liraglutide may increase any adverse events (RR 1.07, 95% CI 1.04 to 1.11) and serious adverse events (RR 1.20, 95% CI 1.00 to 1.43), both rated low certainty. Evidence on side effects leading people to stop the drug (RR 1.98, 95% CI 1.30 to 3.02) was rated very uncertain ^[1].

On outcomes many patients care most about, the review found little signal. Liraglutide likely made little to no clinically meaningful difference in major adverse cardiovascular events (RR 0.86, 95% CI 0.66 to 1.10; 6 studies, 5,762 participants) or in quality of life measured by the IWQOL-Lite-CT physical function score (mean difference 2.90, 95% CI -0.08 to 5.89). Evidence on death was very uncertain (RR 0.44, 95% CI 0.08 to 2.25) ^[1].

Only three trials had long-term data, covering 104 to 162 weeks. There, the higher rate of at least 5% weight loss persisted (RR 1.78, 95% CI 1.51 to 2.09), but withdrawals due to adverse events roughly doubled (RR 2.16, 95% CI 1.59 to 2.93), and the percentage weight difference of 4.34 points was rated low certainty ^[1].

The funding question

The reviewers flagged that the drug manufacturer funded 22 of the 24 studies, with major involvement in the design, conduct, analysis or reporting of 13 of them, "raising concerns about potential conflicts of interest." They called for "further independent and long-term studies" ^[1]. Among the 16 listed review authors, one disclosed work as an independent contractor for Novo Nordisk; the rest declared none, and Cochrane stated that people who selected studies, extracted data and graded evidence were not involved in the included trials ^[1].

Why it matters for patients

Certainty grades are not the same as saying a drug does not work. Moderate certainty means further research could change the estimate; very low certainty means the true effect could be substantially different from what the pooled numbers suggest ^[1]. For someone weighing liraglutide, the practical takeaway is that the strongest evidence supports a specific, modest benchmark — reaching at least 5% weight loss — rather than a precise average percentage or a proven reduction in heart attacks, strokes or death ^[1].

The side-effect picture also matters for staying on treatment. More people on liraglutide stopped because of adverse events, and the review notes this "might further limit the sustainability of the initial effects" ^[1].

This review covers liraglutide only. The sources do not report head-to-head conclusions for semaglutide or tirzepatide, and the two companion Cochrane reviews on other GLP-1 drugs are not described here ^[1].

What happens next

Cochrane's search cutoff was December 17, 2024, so trials reported after that date are not included ^[1]. The updated protocol is registered as PROSPERO CRD420250654193 ^[1]. When the other two GLP-1 reviews will publish is not stated in the source.