Because prisoners with preincarceration heroin dependence typically relapse following release, a pilot study examined a novel opioid agonist maintenance program whereby consenting males initiated levo-alpha-acetylmethadol (LAAM) treatment shortly before release from prison with opportunity to continue maintenance in the community. Treated prisoners (experimental group) were compared with controls who received community treatment referral information only and prisoners who withdrew from treatment prior to medication regarding treatment participation and community adjustment during nine months post-release. Nineteen of 20 (95%) prisoners who initiated maintenance in prison entered community treatment, compared with 3 of 31 (10%) controls, and 1 of 13 (8%) who withdrew. Moreover, 53% of experimental participants remained in community treatment at least six months, while no other participants did so. Differences in heroin use and criminal involvement between experimental participants and each of the other two groups, while not consistently statistically significant, uniformly favored the experimental group. Despite study limitations, robust findings regarding treatment attendance suggest that this intervention is highly promising.