In a study of more than 142,000 patients, those on Ozempic and similar drugs in its class had lower rates of alcohol, opioid, nicotine, and cocaine use disorder, but the authors call the finding an association and warn against prescribing for addiction.
The same prescription that millions of people take for diabetes and weight loss is showing up in their medical records next to a quieter pattern: fewer diagnoses of alcohol, opioid, nicotine, and cocaine use disorder. A new study of more than 142,000 patients in the NIH All of Us research program found that the roughly 20,000 who had been prescribed a GLP-1 medication such as Ozempic were far less likely to carry those four addiction diagnoses than matched patients who had not. The authors are explicit about what their data cannot say.
The study is observational, not a randomized trial, and the effect sizes are large enough to invite other explanations. People who start a weight-loss drug often change how they eat, drink, and move, and they see clinicians more often, all of which can shift substance use independently of any drug effect. Doctors may also be more cautious about prescribing controlled substances to patients already on a high-profile injectable. The lead author, Tadesse M. Melaku Abegaz at the University of Texas at El Paso, writing with colleagues in Frontiers in Psychiatry, describes the result as an association, not a treatment, and says the team does not support prescribing GLP-1s for addiction today.
What makes the signal hard to dismiss is its shape. It appears across four different substances with different mechanisms of action: alcohol, opioids, nicotine, and cocaine. The reported drops in odds, summarized by SciTechDaily from the paper, range from 68% for nicotine use disorder to 75% for cocaine use disorder. A drug that suppressed withdrawal or craving in one class would be interesting; the same pattern showing up across all four points at something more general, which is also what makes the absence of a clear mechanism more conspicuous.
GLP-1 drugs act on receptors in the gut and the brain that help regulate appetite and blood sugar, and there is a plausible reason the brain connection might matter for addiction. The same reward circuitry that drives hunger overlaps with the circuitry that makes alcohol, nicotine, opioids, and cocaine feel compelling. If a metabolic drug were nudging that circuitry downward across the board, it would help explain why a single prescription appears to track with changes in substances as different as wine and fentanyl. That story is, for now, a hypothesis the authors point to, not a finding.
The next step the team calls for is the kind of study their data cannot be: a randomized trial that gives some patients a GLP-1 and tracks their substance use prospectively, alongside mental health and quality-of-life measures. Smaller trials of semaglutide in alcohol use disorder are already underway at other centers, and observational signals on opioids have circulated for years. What the All of Us analysis adds is scale and the same pattern across four substances, which is exactly the kind of signal that justifies the larger and slower trial the authors are asking for. The population already on these drugs is in the tens of millions and growing, which is why the answer matters even before the trials come in.