Between a quarter and a third of people who have bariatric surgery regain a significant amount of weight within two to five years, according to the figures cited by the authors of a new review. Their meta-analysis, published in Frontiers in Endocrinology, pulls together what is known about treating that regain with a GLP-1 drug instead of a second operation. Liraglutide and semaglutide each took off about 9% of body weight over a year, and tirzepatide was ahead of semaglutide at six months. The evidence behind those numbers is thinner than the headline figures suggest.
What the review pooled
The team, at the University of Porto, searched three databases in June 2025 and included 27 studies of adults who started liraglutide, semaglutide or tirzepatide at least a year after surgery, most of them after a gastric bypass or a sleeve gastrectomy. Only three were randomised trials; 22 were cohort studies and two were case series. Nineteen fed the meta-analysis. Weight loss was measured from the day the drug was started, so every figure below is loss on top of whatever the surgery had already achieved.
At 12 months, liraglutide produced a pooled total weight loss of 9.22% and semaglutide 9.02%. For someone who weighs 110 kg when they start, 9% is roughly 10 kg. The 12-month intervals are very wide, though: 0.83% to 17.60% for liraglutide and 3.43% to 14.62% for semaglutide, which means the true average could plausibly sit anywhere from almost nothing to close to a fifth of body weight. The six-month estimates are much tighter, at 8.34% (6.00 to 10.67) and 9.34% (7.72 to 10.97).
Only two studies compared tirzepatide with semaglutide. At three months the gap, 1.69 percentage points, had an interval that crossed zero (−1.99 to 5.37), so no difference was shown. At six months tirzepatide was 4.23 points ahead (1.42 to 7.05). The single study of tirzepatide alone followed 21 patients for six months and recorded a 12% loss.
No serious adverse events attributed to the drugs were reported. In the studies that counted them, 39% of liraglutide patients and 24% of semaglutide patients had at least one side effect, mostly gastrointestinal; among liraglutide users nausea was the most common. The authors warn that those two figures come from different studies and populations and should not be read as a comparison.
The pooled results sit below the roughly 15% and 21% average losses the authors cite from the main semaglutide and tirzepatide trials in people without surgery, although the authors read them as broadly consistent with that trial range. Dose may explain part of the gap: several of the included studies used doses below the standard 2.4 mg weekly maintenance dose on the Wegovy label, or below the 15 mg top dose on the Zepbound label. That is our reading of the study tables, not a conclusion the authors draw.
The direction of the tirzepatide result matches SURMOUNT-5, the head-to-head trial in people who had not had surgery. The more important point is that people with a previous bariatric operation were excluded from the trials that got these drugs approved, so pooled observational evidence like this is most of what exists for them. The authors propose, as expert opinion awaiting validation, lifestyle measures alone when the goal is under 5% loss, a drug when the goal is 5% to 20%, and revision surgery only when more than 20% is needed. Our GLP-1 research hub covers how the drugs themselves compare.
What it cannot tell you
Heterogeneity was high in most of the pooled analyses, reaching an I² of 98.6% for liraglutide at 12 months, because the studies differed in the type of surgery, how long after surgery the drug began, the dose, and even what counted as regain: no two liraglutide studies used the same definition. The 12-month semaglutide estimate rests on three studies and the tirzepatide comparison on two. The authors did not test for publication bias because too few studies fed each analysis, and they left the three randomised trials out of the pooling because those differed too much from the rest. Most of the studies did not report what happened to weight after patients stopped the drug, so whether the extra loss holds without continued treatment is unknown. The review includes no cost analysis either; current self-pay prices are on our cost without insurance page.
Sources: Santos-Pereira M, Frutuoso JF, Pereira SS, Guimarães M, Monteiro MP. “Use of incretin receptor agonists in patients submitted to metabolic bariatric surgery – a systematic review and meta-analysis.” Frontiers in Endocrinology, 26 August 2026, doi:10.3389/fendo.2026.1870008; full text on PubMed Central. PROSPERO registration CRD42022371697. The authors declare no commercial or financial conflicts of interest.
