Aims To compare the effects of levo-alpha-acetylmethadol (LAAM) and methadone maintenance (MM) on treatment retention and abstinence from opiate use. Design A two-group experimental design with patients randomly assigned (2 : 1 LAAM : MM) to receive LAAM (three doses per week) or methadone (daily dosing). Setting A community clinic in Los Angeles, California. Participants A total of 315 patients seeking LAAM or methadone maintenance. Interventions LAAM or methadone maintenance, plus ancillary services available to all patients. LAAM and methadone dose levels varied according to clinical judgement. Electrocardiograms were administered to LAAM patients monthly. Measurements Treatment status at 26-week follow-up and number of days retained in treatment, weekly clinical urine tests and 26-week research urine test. Findings LAAM and methadone patients did not differ on treatment retention. LAAM patients were less likely to test positive for opiate use during treatment (40% versus 60%) and at 26-week follow up (39.8% versus 60.2%). Benefits of LAAM were confined to patients (n = 204) still in treatment at 26 weeks (33% positive in patients receiving LAAM and 61% in patients receiving methadone). No adverse events, cardiological or otherwise, were observed with LAAM administration. Conclusions LAAM is an effective medication for the treatment of opiate dependence with clinical advantages due not only to the reduction of opiate use but also to the alternate-day dosing schedule. LAAM may be more effective than methadone in promoting abstinence from opiate use among patients for whom LAAM is an acceptable alternative to methadone.