A 2026 umbrella review of 16 systematic reviews and 371 trials offers a broad look at where medications have the strongest evidence in addiction care—and where major gaps remain.
The review found strong evidence for medications used in tobacco and opioid use disorders, more modest and less consistent benefits for alcohol use disorder, and no approved medications within its scope for stimulant or cannabis use disorders as of January 1, 2026.
Where the evidence is strongest
The review counted 41 approved medications, representing 25 distinct molecules, across alcohol, opioid, tobacco and sedative use disorders. The evidence summary included 223 tobacco trials, 86 alcohol trials and 49 opioid trials.
Researchers searched for systematic reviews published from 2016 through 2025 in PubMed, MEDLINE and the Cochrane Library. Of 24 reviews assessed in full, eight were excluded because of high risk of bias or low quality. The researchers did not combine the results into a new pooled analysis; instead, they summarized each medication using the most recent, highest-quality review available.
Smoking cessation medications
Varenicline produced quit rates 2.32 times those of placebo across 41 trials involving 17,395 people. Bupropion had a risk ratio of 1.64 across 45 trials, while nicotine replacement therapy had a risk ratio of 1.55 across 133 trials involving 64,640 people.
In the EAGLES trial, varenicline outperformed placebo, a nicotine patch and bupropion. It did not significantly increase moderate or severe psychiatric side effects, including among participants with existing psychiatric disorders.
These comparisons should not be read as a definitive ranking across every medication. Many smoking-cessation studies compared one treatment with placebo, making some comparisons indirect.
Opioid treatment: retention and safety context
Across 16 trials, six-month retention was 24% lower with buprenorphine than with methadone. The review reported a risk ratio of 0.76, with a 95% confidence interval of 0.67 to 0.85. A separate analysis of 21 observational studies involving 155,111 people found a similar difference, with a risk ratio of 0.77.
Retention measures whether people remain in treatment; they do not by themselves establish that one medication is better for every patient. In a smaller Cochrane analysis, urine-confirmed opioid use did not differ significantly between buprenorphine and methadone.
A 2017 meta-analysis of 19 cohorts reported lower all-cause death rates during treatment than after leaving treatment. For methadone, the rates were 11.3 per 1,000 person-years during treatment versus 36.1 after leaving it. For buprenorphine, the corresponding figures were 4.3 and 9.5. These were observational estimates without an untreated comparison group, so they do not provide a precise estimate of lives saved.
Alcohol medications show smaller effects
Acamprosate lowered the risk of any return to drinking by 12% across 20 trials involving 6,380 people. It did not lower the risk of returning to heavy drinking.
Oral naltrexone at 50 milligrams lowered the risk of any drinking by 7% and heavy drinking by 19%. Baclofen was associated with a 13% to 17% lower relapse risk across two reviews, but had little effect on heavy drinking.
Disulfiram showed no pooled benefit across three trials. Benefits appeared only in supervised settings. Overall, the review described alcohol medication benefits as modest and less consistent than the evidence for tobacco or opioid medications.
The gap in stimulant and cannabis treatment
No medication was approved within the review’s scope for cannabis, cocaine, methamphetamine or other stimulant use disorders, hallucinogen use, inhalant use or caffeine use.
One important research signal came from the phase 3 ADAPT-2 trial. It tested injectable naltrexone combined with daily oral bupropion in 403 adults with methamphetamine use disorder. In that trial, 13.6% of participants receiving the combination met the response definition, compared with 2.5% receiving placebo—a difference of 11.1 percentage points.
However, the naltrexone-bupropion combination is not approved for methamphetamine use disorder. The result came from one positive phase 3 trial, and more than eight in 10 people receiving the combination did not meet the response definition.
GLP-1 receptor agonists, including semaglutide, were being studied in phase 2 and phase 3 trials across several substances. None was approved for addiction as of 2026.
What the findings can—and cannot—show
The review provides a map of the current evidence rather than a universal treatment ranking. Its outcomes differed by disorder: quitting was the main measure for tobacco, retention was prominent in opioid studies, and relapse or drinking days were used in alcohol research. Those differences prevent a valid numerical comparison between medications such as varenicline and methadone.
The review also noted that the absence of a qualifying systematic review is not the same as evidence that a medication failed. Six approved drugs, including heroin-assisted treatment and levomethadone, had no qualifying review in the analysis.
For people seeking care, the practical message is that medication options and evidence vary substantially by substance. Treatment decisions require consideration of the specific disorder, the available evidence and a person’s circumstances; results from one condition should not be assumed to apply to another.
AI tools were used to assist with the preparation of this article.
Leave a Reply