Substitution treatment with methadone has proven effective in reducing heroin use and providing an opportunity for improvement in health and social functioning for dependent drug users.2 However, methadone maintenance does not suit all people. LAAM, a full opioid agonist, and buprenorphine, a partial agonist, are promising alternatives to methadone maintenance. Buprenorphine has a lower risk of overdose than methadone, and both buprenorphine and LAAM can be given less frequently than daily, providing increased flexibility and reduced costs. Previous research supports the feasibility of substitution treatment with LAAM and buprenorphine, and each of these drugs has been compared with methadone in controlled studies.2 The rationale for this trial by Johnson and colleagues was to compare LAAM and buprenorphine with standard ("high dose") methadone maintenance and a control in a single study. Methadone at 20 mg daily was chosen as the control because it has the capacity to suppress opioid withdrawal, but is only minimally effective as a maintenance treatment. For ethical reasons, a rescue treatment was available for all study participants who responded poorly to experimental treatment.

Participants were stratified, then allocated by random number generation, with participants and clinic staff unaware of group assignments and doses. Each day participants received three solutions (two oral, one sublingual), only one of which contained active medication. Otherwise all groups received equivalent treatment. It seems likely that participants in the "low dose" methadone group might have been able to guess their group allocation, but the adequacy of blinding was not discussed by the authors. There were clear definitions for transfer to rescue treatment, drop-out, and calculation of retention in treatment, and adjustments for missing data. All analyses were based on intention to treat.

LAAM and buprenorphine are significantly more effective than "low dose" methadone in reducing illicit opioid use, and of similar effectiveness to "high dose" methadone. Illicit opioid use was somewhat lower for the LAAM group — the use of larger group sizes might have provided sufficient statistical power for a significant difference to be detected. Participants in the "high dose" methadone group were retained in treatment for more days, although the difference achieved statistical significance only for "high dose" methadone compared with LAAM. Most of the difference between "high dose" methadone and LAAM occurred in the first two weeks of treatment.