GLP-1 drugs and severe GI motility risk

A new cohort study of more than 626,000 matched adults with type-2 diabetes found that people who started a GLP-1 based therapy (including tirzepatide) had a 37% higher relative risk of severe gastrointestinal motility events compared with those who started an SGLT-2 inhibitor^[1]. The absolute risk stayed below 1%, but the signal is large enough to inform clinical conversations about who should be monitored more closely.

What the study measured

Alkabbani and colleagues at Harvard and Brigham and Women's Hospital used two large deidentified US commercial healthcare databases to compare new users of GLP-1 receptor agonists or tirzepatide against new users of SGLT-2 inhibitors. All participants were adults with type-2 diabetes who had no prior major gastrointestinal conditions at baseline. The composite endpoint was the first occurrence of severe constipation, gastroparesis, or gastrointestinal obstruction. After propensity-score matching on 127 baseline covariates, the cohort included 313,342 individuals in each group, with a mean age of 60 years and 45% female. Median follow-up on treatment was 5.2 months^[1].

The incidence rate in the GLP-1 group was 1.02 per 100 person-years versus 0.75 in the SGLT-2 inhibitor group, giving a rate difference of 0.27 per 100 person-years and a hazard ratio of 1.37 (95% CI 1.30 to 1.45). The increased risk held across subgroups stratified by specific GLP-1 agent, age, sex, BMI, frailty level, diabetes severity, and opioid use. The consistency across subgroups is notable: it suggests the signal is not driven by a single molecule or a single patient profile.

Why absolute risk matters here

A 37% relative increase sounds alarming out of context. Relative risk figures often do. But the base rate is low: 0.75 per 100 person-years in the comparator group. In practical terms, the rate difference translates to roughly 3 additional severe GI motility events per 1,000 person-years of treatment. For most patients starting a GLP-1 therapy for type-2 diabetes, that figure does not change the benefit-risk calculus.

It does, however, flag specific populations where clinicians may want to monitor more carefully. Older adults, people already taking opioids (which independently slow gut motility), those with a history of slow-transit constipation, and patients who report persistent bloating or abdominal discomfort early in titration are the groups where this data is most clinically actionable.

How this fits the broader safety picture

Delayed gastric emptying is a known pharmacological effect of the GLP-1 class. The Mounjaro (tirzepatide) prescribing information states that the drug "delays gastric emptying" and is "not recommended in patients with severe gastroparesis"^[2]. A 2024 review in the Journal of Clinical Endocrinology and Metabolism by Jalleh and colleagues catalogued the downstream consequences: retained gastric contents at endoscopy (a four-fold increase in one 35,183-patient study), aspiration risk during anaesthesia, and inconsistent but non-zero signals for small bowel obstruction^[3].

The Alkabbani cohort study adds population-level scale to that picture. Because it uses an active comparator (SGLT-2 inhibitors) rather than placebo, it controls for the confounding that comes with simply being a person with type-2 diabetes who initiates a new drug. The active-comparator new-user design is widely regarded as the strongest observational approach for drug-safety questions of this kind. Residual confounding from unmeasured variables (diet, hydration, fibre intake) cannot be ruled out, but the 127-covariate propensity score is unusually thorough for a claims-based study.

What this does not tell us

The study did not separate individual GLP-1 agents at the primary analysis level, though subgroup analyses were consistent across agents. It did not include people using GLP-1 therapies for obesity without diabetes, so the findings do not directly extend to the Zepbound or Wegovy populations. And it did not capture over-the-counter laxative use or dietary changes, both of which could modify GI motility risk in ways the claims data cannot see.

There is also a duration question. Median follow-up was 5.2 months. Whether the elevated risk persists, attenuates, or increases over longer treatment courses is not addressed. Given that the Mounjaro label notes the gastric-emptying delay is "largest after the first dose" and "diminishes over time," the chronic-use risk profile may differ from what this cohort captured.

Where this lands for readers

If you are on a GLP-1 based therapy or considering one, the takeaway is not that these drugs are unsafe. The absolute event rate is low, and the benefits for glycaemic control and weight management are well established across the SURPASS and SURMOUNT programmes. The takeaway is that severe constipation, gastroparesis, and obstruction are real, if uncommon, downstream effects of a drug class that slows the gut by design. Mention any persistent GI symptoms to your prescriber. For background on how tirzepatide works and its broader safety profile, see our main page. For semaglutide the same class-level GI mechanism applies, and the safety sections on both peptide pages cover the gastrointestinal profile in detail.