Jonas 2014: what 122 trials show about alcohol medicines

What they tested

Researchers searched medical databases, the FDA website and trial registries from 1970 to March 2014 for randomized trials of at least 12 weeks testing medicines for adults with alcohol use disorders. They included 122 trials and 1 cohort study, with 22,803 participants in total.

Most studies looked at acamprosate (27 studies, 7,519 people), naltrexone (53 studies, 9,140 people) or both. Results were given as the number needed to treat (NNT), how many people must take a drug for one more to benefit, and the number needed to harm (NNH).

What they found

For oral naltrexone 50 mg a day, the NNT was 20 to prevent a return to any drinking (95% CI, 11 to 500) and 12 to prevent a return to heavy drinking. For acamprosate, the NNT to prevent any drinking was 12.

Trials comparing acamprosate directly with naltrexone found no statistically significant difference for return to any drinking or to heavy drinking.

For injectable naltrexone, pooled trials found no link with return to any drinking or heavy drinking, but did find fewer heavy drinking days (a weighted mean difference of -4.6%).

Among medicines used off-label, moderate evidence linked nalmefene and topiramate to improvements in some drinking measures. Nalmefene was linked to 2.0 fewer heavy drinking days a month, and topiramate to a lower percentage of heavy drinking days (a weighted mean difference of -9.0%).

Side effects led to more trial dropouts on naltrexone (NNH 48) and nalmefene (NNH 12). The risk was not significantly raised for acamprosate or topiramate.

Limits

• Only trials of at least 12 weeks were included, and the review covers outpatient treatment.
• The range around naltrexone's NNT for preventing any drinking was very wide (11 to 500), so that estimate is uncertain.

What it means if you're considering naltrexone

For someone weighing naltrexone, this review shows a real but modest benefit for the daily tablet, strongest against heavy drinking, and about the same results as acamprosate when the two were compared directly. Side effects led to a small number of extra dropouts. The authors suggest dosing frequency, side effects and availability can guide the choice between medicines.

The site's guides also cite a 2023 federal review that updated this one and put the number needed to treat to avoid heavy drinking at 11.