No trial has been designed to test cannabis with Ozempic. A 2026 review of registered GLP-1 substance-use trials found none on cannabis [5]. The current Ozempic label does not mention cannabis or marijuana [6]. The largest study to touch the question followed US veterans with type 2 diabetes for up to three years [2]. Starting a GLP-1 drug was linked to 2.25 fewer cannabis use disorder diagnoses per 1,000 people than starting an SGLT-2 inhibitor. That was a hazard ratio of 0.86.
This page reports what studies measured. It does not advise on cannabis use, with or without a GLP-1. The evidence comes in three kinds, in this order of strength:
- Randomized trials. One exists, and cannabis was a side outcome in it.
- Cohort studies of health records. Two large ones exist.
- Reports from users online. One analysis exists.
The same kinds of evidence exist for alcohol, where a dedicated trial has been run. That result is in whether semaglutide reduces alcohol use.
What the labels say
The current Ozempic, Wegovy and Zepbound labels contain none of these words: cannabis, marijuana, THC or dronabinol [6] [8] [9]. The labels list no interaction with cannabis, and none says one was studied. An absent warning is not a finding of safety. It means the question was not addressed.
The cohort studies
Both cohorts counted cannabis use disorder diagnoses in health records. A diagnosis is not the same as use. The 2024 study puts US cannabis users at more than 45 million, with about one-third meeting the diagnosis [1].
- Semaglutide, 2024. The study drew on US electronic health records [1]. It matched 85,223 patients with obesity on semaglutide or another anti-obesity drug. Over 12 months, semaglutide was linked to fewer new diagnoses (HR 0.56, 95% CI 0.42 to 0.75). Among people with a prior diagnosis, it was linked to fewer repeat diagnoses (HR 0.62, 95% CI 0.46 to 0.84).
- Veterans, 2026. The trial emulation followed 524,817 veterans with type 2 diabetes [2]. It compared starting a GLP-1 drug with starting an SGLT-2 inhibitor. The cannabis use disorder hazard ratio was 0.86 (95% CI 0.81 to 0.90). The absolute difference was 2.25 fewer diagnoses per 1,000 people over three years.
The 2024 authors call their result preliminary evidence. They state that randomized trials are needed before semaglutide is used for cannabis use disorder [1]. The veterans study describes its data as observational [2]. Neither study measured how much cannabis anyone used. A pooled review that included cannabis outcomes is logged in the 82-study substance-use review.
The one randomized result
A 2026 trial randomized 50 adults with moderate to severe alcohol use disorder [3]. They took oral semaglutide, 3 mg then 7 mg a day, or placebo for 8 weeks. Cannabis use was an exploratory outcome. Cannabis use days fell significantly on semaglutide.
The trial was built for alcohol, not cannabis. Its primary outcome, craving in a laboratory test, did not fall significantly. An exploratory result in 50 people is a lead for a future trial. It is not an answer.
What users report
A 2024 study coded Reddit posts about GLP-1 drugs and substance use [4]. It drew on 5,859 threads and their comments. About 30% of alcohol-related comments described stopping alcohol. For cannabis, the results were mixed. Posts are self-selected, and none was verified.
Cannabis use disorder and mental health
One more cohort bears on safety. A 2025 study of people with type 2 diabetes looked at suicide attempts [7]. GLP-1 use was linked to lower risk overall (adjusted HR 0.63). Cannabis use disorder was linked to much higher risk (adjusted HR 5.50). Among people with cannabis use disorder, GLP-1 use showed no difference (adjusted HR 1.00, 95% CI 0.37 to 2.69).
That is an association in records, not a drug effect. The broader cohort evidence on these drugs and suicidal thoughts is in semaglutide and suicidal ideation. The nicotine side of the substance-use question is in the exenatide smoking-cessation trial.