Oxycodone linked to a higher 30-day risk of severe gut problems than hydrocodone or tramadol in GLP-1 users

Adults taking a GLP-1 medicine who were then prescribed oxycodone for pain had more severe gut complications over the next month than similar patients given hydrocodone or tramadol. The finding comes from an analysis of 411,188 US insurance records published on September 2 in Diabetes Care.

GLP-1 medicines, the class that includes semaglutide and tirzepatide, work partly by slowing how fast the stomach empties. Opioid painkillers slow the gut as well. The study asked a narrow and practical question: when a patient is already on a GLP-1 and needs a painkiller, does the choice of opioid matter?

Within 30 days of the first opioid prescription, the researchers counted a combined measure of three problems. Severe constipation. Bowel obstruction, which is a blockage of the intestine and a surgical emergency. And gastroparesis, which means the stomach empties far more slowly than it should, causing nausea, vomiting and a full feeling after small meals.

Roughly one extra case for every 550 to 600 people given oxycodone

The rate of that combined measure was 0.51 percent for oxycodone, 0.35 percent for hydrocodone and 0.33 percent for tramadol.

Set against hydrocodone, oxycodone carried a risk ratio of 1.48 (95 percent confidence interval 1.30 to 1.69) and an absolute risk difference of 0.17 per 100 patients (0.11 to 0.22). Set against tramadol, the risk ratio was 1.55 (1.33 to 1.79) and the risk difference 0.18 per 100 (0.12 to 0.24).

A risk ratio near 1.5 sounds alarming on its own. The absolute numbers are small. A risk difference of 0.17 per 100 works out to about one extra case for every 550 to 600 people who receive oxycodone rather than one of the other two drugs. A confidence interval is the range the true value most likely sits in. None of the ranges above crosses 1.0 for the ratios, so chance alone is an unlikely explanation.

The gap came from severe constipation and bowel obstruction. The researchers did not find a difference in gastroparesis between the three opioids.

Hydrocodone and tramadol were indistinguishable from each other

Comparing hydrocodone with tramadol gave a risk ratio of 1.05 (0.91 to 1.21) and a risk difference of 0.02 per 100 (-0.03 to 0.06). Both ranges include no difference at all, so this study gives no reason to prefer one of those two over the other on gut safety.

Everyone studied had type 2 diabetes and averaged 62.8 years of age

The cohort came from US insurance claims filed between 2016 and 2025. All 411,188 patients had type 2 diabetes and were already taking a GLP-1. Mean age was 62.8 years and 53.8 percent were women. Of the group, 24.4 percent started oxycodone, 48.5 percent hydrocodone and 27.1 percent tramadol.

That shapes who the result speaks to. People taking a GLP-1 for obesity without diabetes, and younger patients generally, were not the population under study. The finding may well extend to them. This study does not show that it does.

Claims records cannot show why a clinician picked one opioid

This was a new-user cohort study, meaning it followed people from the moment they first filled one of these opioids rather than mixing in long-term users. The researchers applied propensity score matching weights, a statistical method that balances measured differences between groups, such as age, other conditions and previous medicines, so the comparison sits closer to like with like.

That method can only balance what was written down. Oxycodone is generally held back for more severe pain, and the reason a clinician reached for it does not appear in a claims database. Patients who receive oxycodone may differ from those who receive tramadol in ways the data cannot capture, and part of the gap could reflect those patients rather than the drug.

The authors flag their own caution, writing that variation across secondary analyses warrants cautious interpretation. Follow-up ran for 30 days, so the study says nothing about longer opioid use.

This is a prescribing decision, not a patient decision. Anyone on a GLP-1 who is offered a painkiller can raise the question with the prescriber. Nobody should stop or change a prescribed dose on their own.