GLP-1 Drugs and Opioid Use Disorder: What the Records Data Shows

No randomized trial has tested a GLP‑1 for opioid use disorder. What exists is records data, and it points one way. Across 503,747 people with a history of opioid use disorder, those with a GLP‑1 or GIP prescription had an adjusted opioid-overdose rate ratio of 0.60 (95% CI 0.43–0.83).^[1] In a smaller, matched study of people already on methadone, a GLP‑1 prescription was associated with a higher chance of entering remission (HR 1.75).^[2] Both are associations. Neither is a reason to change an opioid treatment plan.

The large cohort

The bigger study drew de-identified electronic health records from about 136 US health systems covering more than 100 million patients, from January 2014 to September 2022. It looked at 503,747 adults with a history of opioid use disorder and, separately, 817,309 with alcohol use disorder, comparing those with at least one GIP or GLP‑1 receptor agonist prescription against those with none.^[1]

The association held when the analysis was restricted to people who also had type 2 diabetes and obesity — which matters, because the obvious objection is that a GLP‑1 prescription is a marker for having diabetes rather than a cause of anything. That subgroup consistency does not eliminate confounding, but it removes the simplest version of it.

The methadone study

The second study is narrower and, for that reason, more interesting: everyone in it was already receiving methadone. It took adults on methadone with both opioid use disorder and type 2 diabetes from the TriNetX database between 2015 and 2024, propensity-matched them on baseline characteristics, and followed 1,157 people prescribed a GLP‑1 for a year.^[2]

The design is the point. This is not a GLP‑1 replacing opioid agonist therapy; it is a GLP‑1 sitting alongside methadone in people who were already receiving the treatment with the strongest evidence base there is. Nothing here suggests substituting one for the other, and nobody has studied that.

What would settle it, and what to do meanwhile

• There is no randomized trial of a GLP-1 for opioid use disorder. Two large observational studies pointing the same way is a reason to run one, not a reason to act as though it has been run.
• Buprenorphine and methadone are the treatments with randomized evidence for opioid use disorder. Nothing in this literature displaces them.
• The larger cohort found the same direction of effect for alcohol intoxication in a separate 817,309-person group, which is consistent with what has been seen for other substances.
• ⚠ If you take an opioid agonist therapy, changes to it belong with the clinician managing it. A GLP-1 is not a substitute and was not studied as one.

For the same class of drugs across every other substance measured — and for the randomized behavioral trial that missed its endpoints — see GLP-1s and substance use. For the one substance where a dedicated randomized cessation trial exists, see GLP-1s and quitting smoking.