OBJECTIVE To compare levo-alpha-acetylmethadol (LAAM) and methadone maintenance (MM) on treatment retention, drug use during treatment and at follow-up, and abstinence. METHODS A two-group experimental design with patients assigned randomly (2:1) to receive fully subsidized LAAM or MM for 52 weeks. METHODS A community clinic providing maintenance treatment in Los Angeles, California. METHODS A total of 315 treatment-seeking patients willing to be assigned randomly to treatment condition; 289 (91.7%) were interviewed at 52 weeks. METHODS LAAM or MM, plus ancillary services available to all patients. Medication dose varied according to clinical judgement. METHODS Treatment retention and status at 52-week follow-up, weekly clinical urinalysis, self-reported drug use and research urinalysis on samples collected at follow-up. RESULTS LAAM participants were more likely to complete the planned 52 weeks (57.4%) than MM participants (46.2%) and were less likely to be discharged for arrest/incarceration. LAAM produced fewer during treatment clinic opiate-positive samples (M = 48.8) than MM (M = 62.3). Further, 24.4% on LAAM compared to 11.8% on MM were able to sustain at least 12 weeks of abstinence during the last 24 weeks of treatment. Opiate use at follow-up was lowest (50.9%) among LAAM participants in maintenance treatment. No adverse events, cardiological or otherwise, were observed with LAAM administration. CONCLUSIONS LAAM is an effective medication for the treatment of opiate dependence in community clinics with numerous behavioral and clinical advantages. LAAM is more effective than MM in promoting retention and extended reduction in and abstinence from opiate use while in treatment.