Semaglutide and Gallbladder Events

A 2026 pooled analysis of four parallel-group STEP trials found about twice the risk of gallbladder-related events with once-weekly semaglutide 2.4 mg as with placebo. The pooled risk ratio was 2.08, but the absolute rates were 1.92 versus 0.82 events per 100 patient-years^[1]. Put plainly, these events were uncommon in both groups. The analysis estimates roughly one additional gallbladder event for every 90 to 100 people treated for a year. It supports a known safety signal rather than revealing a new, frequent complication.

The distinction between relative and absolute risk matters here. A doubling sounds large because the placebo rate starts low. The current US Wegovy label already warns about acute gallbladder disease and reports cholelithiasis in 1.6% of adults receiving the injection versus 0.7% receiving placebo. Cholecystitis was reported in 0.6% versus 0.2%^[4]. The pooled paper helps place those label figures across several trials, while its design limits how confidently we can separate drug effects from weight-loss effects.

What the pooled analysis included

This was not a new randomized trial. It was a pooled analysis of adverse-event counts from STEP 1, STEP 2, STEP 3 and STEP 5, all placebo-controlled trials of subcutaneous semaglutide 2.4 mg for weight management. Together, the primary analysis included 3,683 participants: 2,269 assigned to semaglutide and 1,414 to placebo^[2]. STEP 1, 3 and 5 enrolled adults without diabetes. STEP 2 enrolled adults with type 2 diabetes. Follow-up lasted 68 weeks in STEP 1 through STEP 3 and 104 weeks in STEP 5.

The underlying STEP trials were designed mainly to assess weight loss, not gallbladder disease. STEP 1, for example, randomly assigned 1,961 adults without diabetes to semaglutide 2.4 mg or placebo plus lifestyle intervention for 68 weeks. Mean body weight changed by 14.9% with semaglutide and 2.4% with placebo^[3]. Gallbladder diagnoses were collected through clinical adverse-event reporting. The trials did not use scheduled ultrasound to look for silent gallstones.

The pooled outcome combined cholelithiasis, cholecystitis and gallbladder disease not otherwise specified. Cholelithiasis accounted for most events where the original reports separated diagnoses. The authors calculated risk ratios for each trial, applied a continuity correction where an arm had no events, and combined the results with a random-effects model^[2]. That is a reasonable method for rare events, though estimates remain sensitive to a small number of diagnoses.

The event counts behind the headline

Across STEP 1, 2, 3 and 5, researchers counted 59 gallbladder-related events in the semaglutide groups and 16 in the placebo groups. After accounting for follow-up, the rates were 1.92 and 0.82 per 100 patient-years. The pooled risk ratio of 2.08 had a 95% confidence interval from 1.16 to 3.72, and statistical heterogeneity was low at an I-squared value of 4.3%^[2]. The wide confidence interval still allows a range of plausible effect sizes. This is a signal with uncertainty, not a precise forecast for one person.

Individual trials were less tidy. STEP 3 recorded 20 events among 407 semaglutide participants and three among 204 placebo participants. STEP 2 recorded one event with semaglutide 2.4 mg and three with placebo. STEP 5 recorded four and two, respectively^[2]. None of those trial-level estimates alone gives a stable answer. Pooling gains statistical power, but it does not erase differences in participant populations or lifestyle programmes.

Why STEP 4 changes the result

STEP 4 used a withdrawal design. Everyone first received semaglutide during a 20-week open-label run-in. Only participants who reached the maintenance dose were then randomized either to continue semaglutide or switch to placebo. Both randomized groups therefore had prior exposure to the drug, and gallstones that formed during the run-in could appear later in either arm. That makes STEP 4 unsuitable for the cleanest drug-versus-no-drug comparison.

When the authors added STEP 4 in a sensitivity analysis, the combined dataset grew to 4,486 participants and the risk ratio fell to 1.47. Its 95% confidence interval ran from 0.74 to 2.95, so the result was no longer statistically significant. Heterogeneity rose to 45%^[2]. This does not prove the primary analysis wrong. It shows how one structurally different trial can alter a rare-event estimate, which is exactly why the authors prespecified the four parallel-group trials as primary.

What may be driving the risk

Rapid or substantial weight loss can raise the risk of gallstones. The Wegovy label states that acute gallbladder disease occurred more often with treatment even after accounting for the degree of weight loss^[4]. The pooled paper also discusses a possible direct GLP-1 effect on gallbladder emptying. These mechanisms can coexist, but this analysis was not designed to measure either one. It contains event counts, not serial imaging or tests of gallbladder contraction.

That limitation also means the analysis probably misses asymptomatic stones. A person counted only if a clinical event was reported. Scheduled ultrasound might have found more gallstones in both groups, but a silent imaging finding is not the same outcome as painful biliary disease or surgery. Readers looking at the wider balance of effects can review our semaglutide evidence guide and the separate semaglutide safety section.

What this changes in practice

The pooled result reinforces existing product information. It does not establish a screening schedule, prove that preventive medicine is useful, or identify a subgroup that should receive different treatment. The authors note that preventive ursodeoxycholic acid has evidence after bariatric surgery, but direct evidence for that approach in people starting a GLP-1 receptor agonist is lacking^[2]. That unanswered question should not be converted into a treatment recommendation.

The label advises gallbladder studies and clinical follow-up when cholelithiasis is suspected^[4]. For a reader, the useful takeaway is narrower: gallbladder events are a recognized risk, the relative increase in this pooled analysis was meaningful, and the absolute frequency was low. Individual risk depends on medical history and the course of weight loss, so symptom assessment belongs with a qualified clinician rather than an online rule.

What we still do not know

The analysis lacked patient-level time-to-event data, standardized ultrasound and enough events to pool each diagnosis separately. It cannot tell us when risk peaks, how much comes from weight loss versus a direct drug effect, or whether people with pre-existing silent gallstones face a different absolute risk. The STEP programme also cannot answer whether the same estimate applies to every semaglutide formulation or dose.

Medical disclaimer: This article is for educational and journalistic purposes only and does not constitute medical advice. Peptides discussed may be classified as prescription medicines or research chemicals depending on your jurisdiction. Always consult a qualified healthcare professional before using any peptide product. PeptideMethods.com does not sell, distribute, or facilitate the sale of any peptide product.