What role does measuring medication compliance play in evaluating the efficacy of naltrexone?

Abstract

Background: Compliance with medication in pharmacotherapy trials of alcoholism has been shown to be equal to, or more, important than in other areas of medicine. Research has suggested that naltrexone's effectiveness can be greatly influenced by the compliance of participants in clinical trials. Presently, we compare 2 compliance measurement methods [urine riboflavin and medication event monitoring system (MEMS)] used simultaneously to evaluate naltrexone's efficacy and the impact of compliance on the size of observable treatment effects.

Methods: One hundred and thirty-seven of 160 randomized alcoholic patients completed 12-weeks (84 days) of naltrexone or placebo and cognitive behavioral therapy (CBT) or motivational enhancement therapy (MET). Urine riboflavin was determined during study weeks 2, 6, and 12. The MEMS provided a detailed computerized record of when a participant opened their medication bottle throughout the trial. Baseline predictors of MEMS (80% openings) and urine riboflavin (>or=1,500 ng/mL by fluorimetry) compliance were examined. The effects of the treatments in the compliant participants defined by one, the other, or both methods were compared and contrasted with a previously reported intent-to-treat analysis where compliance was not taken into account.

Results: Age was predictive of compliance. 105 participants were deemed compliant via urine riboflavin criteria, 87 via MEMS, and 77 when both criteria were met, with no significant differences between treatment groups. The most compliant participants showed a significant medication by therapy interaction. Those treated with naltrexone/CBT showed more abstinence days (p<0.03), less heavy drinking days (p<0.03) and less total drinks (p<0.03) than the other groups. The effect size of this interaction increased from about 0.2 in the intent-to-treat analysis, to about 0.4 to 0.5 in the compliant group analyses, with little difference between compliance measurement methods.

Conclusions: Compliance measurement does appear to influence the evaluation of the efficacy of naltrexone within the context of CBT. Treatment effect sizes approximately doubled in the most compliant individuals. Measuring compliance by either of 2 distinct methods provides approximately similar results. As compliance with naltrexone within the context of CBT has such a large impact of treatment outcome, methods of enhancing compliance during treatment should be given the utmost attention.