Opioid choice in GLP-1RA users: oxycodone was associated with more severe constipation and obstruction

Design

Population-based new-user cohort, US claims 2016–2025, propensity-score weighted

Population

411,188 adults with type 2 diabetes on a GLP-1RA starting oxycodone, hydrocodone or tramadol

Primary outcome

30-day severe constipation, bowel obstruction or gastroparesis

Effect

0.51% oxycodone vs 0.33% tramadol, RR 1.55 (95% CI 1.33 to 1.79); vs hydrocodone RR 1.48 (1.30 to 1.69)

A new-user cohort from US insurance claims (2016 to 2025) followed 411,188 adults with type 2 diabetes already on a GLP-1 receptor agonist who started oxycodone, hydrocodone or tramadol. Propensity-score weighting balanced the groups, and the outcome was a 30-day composite of severe constipation, bowel obstruction and gastroparesis.

The weighted 30-day risk was 0.51% with oxycodone, 0.35% with hydrocodone and 0.33% with tramadol. Oxycodone was associated with a 48% relative increase against hydrocodone and 55% against tramadol. Hydrocodone and tramadol were indistinguishable. The excess sat in constipation and obstruction; gastroparesis did not differ.

Both drug classes slow the gut, and the question of which opioid adds least to a GLP-1RA has had almost no data. This gives a first ranking, though the absolute gap is under two events per thousand prescriptions.

In clinic, it is a reason to think about bowel risk when a GLP-1RA user needs short-term opioid analgesia, not a reason to prefer tramadol, which carries its own hypoglycaemia, serotonergic and seizure concerns. Hydrocodone is not marketed in India, so the practical comparison here is oxycodone or tapentadol against tramadol.

• Ask about bowel habit before prescribing an opioid to a GLP-1RA user
• Co-prescribe an osmotic laxative with any opioid course
• Tell patients that vomiting, distension or no bowel movement for several days needs review
• Weigh tramadol's own risks (hypoglycaemia, serotonin syndrome, seizures) before switching to it

Why it matters

Opioid choice is usually made without thinking about the GLP-1RA already slowing the patient's gut.

Don't overread it

This was observational claims data; it cannot show that switching opioid prevents obstruction, and the absolute difference is small.

The statistics, in plain English

A risk ratio of 1.55 sounds large, but it is 0.51% against 0.33%: a risk difference of 0.18 percentage points, or about one extra severe GI event for every 550 people started on oxycodone instead of tramadol. The confidence interval (1.33 to 1.79) excludes 1, so chance is an unlikely explanation, but residual confounding (why a clinician chose oxycodone, such as more severe pain after surgery) can produce differences of this size.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.