Hendershot 2017: naltrexone's effect on drinking and craving in the lab

What they tested

A meta-analysis pools the results of many studies. Here the researchers pooled human laboratory studies in which volunteers took naltrexone or a placebo for a short time and then had their drinking and craving measured under controlled conditions. In self-administration studies, people could choose how much alcohol to drink during a lab session.

They also checked whether the effect differed by who took part (dependent or non-dependent drinkers), the type of lab test, or how long people had been taking the medicine.

What they found

Compared with placebo, naltrexone reduced the amount people drank in the lab (Hedges' g = -.277, 95 percent CI -.421 to -.133, p < .001). Hedges' g is a standard measure of effect size, and a negative value here means less drinking on naltrexone.

It also reduced self-reported craving (g = -.286, 95 percent CI -.416 to -.156, p < .001).

The effects didn't differ by study population, lab method or how long people took the medicine. The authors say this supports the idea that naltrexone works partly by reducing craving and drinking in the moment.

Limits

• Lab studies are short and artificial. How much someone drinks in a lab session isn't the same as how they drink over months at home.
• The drinking analysis rested on only 9 studies.

What it means if you're considering naltrexone

Results like these help explain how naltrexone works for alcohol: it seems to blunt craving and how much people drink once alcohol is in front of them. The effect on drinking was small. The authors add that the results may support naltrexone's use in short or acute dosing schedules, the idea behind taking it before drinking.

The lab results don't by themselves show how much someone will cut back in daily life. That evidence comes from longer clinical trials.