Volpicelli 1997: naltrexone worked better in people who took it

What they tested

People with alcohol dependence were randomly given naltrexone (48 people) or a matching placebo (49 people) for 12 weeks. Everyone had individual counselling, twice a week for the first month and then weekly.

Earlier trials had excellent pill-taking. This one aimed to see how naltrexone did in a more everyday treatment setting, and how attendance and pill-taking affected results.

What they found

Over the whole sample, naltrexone showed only modest effects in reducing drinking during the 12 weeks.

Among people who completed treatment and were highly compliant with their medicine, naltrexone's effect improved across a variety of outcome measures.

The authors concluded that naltrexone is effective in people who stick with treatment and take the medicine, and that better pill-taking could improve its results.

Limits

• People weren't randomly assigned to take their pills well or badly, so the better results in reliable pill-takers may partly reflect differences between people who stick with treatment and those who don't.
• It was a small, 12-week trial.
• The abstract gives no numbers for the size of the effects.

What it means if you're considering naltrexone

This trial is part of why the site stresses daily pill-taking. Naltrexone's effect looked small when everyone was counted, and larger in those who took it as planned. It doesn't prove the pills caused all of that difference, but it does suggest that missing doses can make naltrexone look less helpful than it can be.