GLP-1 constipation: prevalence vs incidence

Constipation is one of the most commonly reported side effects of tirzepatide and semaglutide, but a June 2026 letter in Clinical Gastroenterology and Hepatology argues that the headline numbers are misleading. The study, led by Kyle Staller at Massachusetts General Hospital, found that many GLP-1 receptor agonist users already had constipation before starting treatment. Prevalence (how many users have constipation at any given time) was high, but incidence (how many developed it new after starting the drug) was low^[1].

Why the distinction matters

Prevalence counts everyone with constipation, whether the drug caused it or not. Incidence counts only new cases that appear after treatment begins. The two numbers answer different questions, and confusing them leads to different clinical conversations.

People prescribed GLP-1 receptor agonists tend to have obesity, type-2 diabetes, or both. Those conditions carry higher baseline rates of gastrointestinal symptoms, including constipation, than the general population. When a study surveys GLP-1 users at a single point in time and reports constipation rates, it is measuring a mix of cases that existed before treatment and cases the drug caused. The Chang, Wang, Berschback and Staller letter makes exactly this point: the GLP-1 RA user population carries pre-existing GI burden that inflates the headline prevalence figure.

What the trial labels actually report

The Mounjaro prescribing information lists constipation at 6% to 7% across the 5, 10 and 15 mg doses versus 1% on placebo in the pooled SURPASS diabetes trials^[2]. In the obesity setting, rates run higher: SURMOUNT-3 reported constipation in 23.0% of tirzepatide-treated participants versus 6.8% on placebo^[3]. Semaglutide labels show a similar dose-dependent pattern, with constipation at 3.1% to 5.0% on Ozempic (diabetes doses) versus 1.5% on placebo^[4].

These trial figures capture new-onset events during the study period, which is closer to incidence. Randomised trials typically exclude patients with uncontrolled GI conditions at baseline, so the enrolled populations are healthier than the average clinic patient starting a GLP-1 RA.

Real-world cross-sectional surveys paint a different picture. The NIH All of Us cohort analysis by Aldhaleei and colleagues surveyed 10,328 adults with type-2 diabetes or obesity who were new GLP-1 RA users and found constipation in 30.4% of them. That number blends pre-existing and treatment-emergent cases without separating them, which is exactly the conflation the Staller letter warns against.

The mechanism behind GLP-1 constipation

GLP-1 receptor agonists slow gastric emptying. That delayed motility is part of the satiety signal (food stays in the stomach longer, so you feel full sooner), and it is also what causes nausea, bloating and constipation during dose escalation. The effect is dose-dependent: higher doses slow transit more, and the constipation signal is strongest in the first weeks after each dose step.

A 2025 review in US Pharmacist noted that the constipating effect is driven by reduced GI motility and recommended smaller, more frequent meals rather than high-fibre loading in the early weeks of therapy^[5]. The same review cautioned against using ondansetron for nausea alongside GLP-1 drugs because it can worsen constipation further. Most trial protocols mitigate GI side effects through slow titration over several weeks, which is why the labelled rates are lower than what clinicians see in real-world practice where dose escalation sometimes moves faster.

What we do not yet know

The Staller letter is brief. It does not publish the full dataset, break down results by specific GLP-1 RA (semaglutide vs tirzepatide vs liraglutide), or track how long constipation persists in those who do develop it on treatment. Whether the prevalence-incidence gap holds equally across all agents and dose levels remains open.

It also does not address whether patients with pre-existing constipation experience worsening on GLP-1 therapy, a different and clinically relevant question. Someone who already has slow-transit constipation and then adds a drug that further slows motility may have a harder time than the trial averages suggest. That subgroup has not been studied separately in any of the major GLP-1 obesity or diabetes programmes.

Practical takeaway

If you are reading about constipation rates on GLP-1 RA drugs, check whether the number cited is prevalence or incidence. They answer different questions. A 30% prevalence figure does not mean 30% of people develop constipation because of the drug. The trial-reported incidence figures, in the range of 6% to 23% depending on indication and dose, are a more accurate reflection of drug-attributable risk, and most of those cases are mild and concentrated during dose escalation. For country-specific prescribing details, see the tirzepatide regulation pages and the semaglutide regulation pages.