Substance abuse is associated with high risk for tuberculosis (TB) and poor adherence to medication regimens. This study compared completion rates for isoniazid (INH) preventive therapy for injection drug users (IDUs) randomly assigned to methadone treatment combined with directly observed preventive treatment (DOPT) versus those assigned to routine TB clinic referral without methadone treatment. One hundred and eleven opioid-dependent patients with latent TB were assigned to one of three 6-month treatment conditions: standard methadone treatment including substance abuse counseling combined with daily INH DOPT (n=37); minimal methadone treatment without counseling, also combined with daily INH DOPT (n=35); or routine care referral to TB clinic for monthly INH supplies without DOPT and without methadone treatment (n=39). INH completion rates were 77.1% for minimal methadone and 59.5% for standard methadone, as compared with only 13.5% for routine care (P<0.0001). Mean duration of INH treatment retention was 5.7, 5.0 and 1.6 months, respectively (P<0.0001). TB incidence at 4-year follow-up was 0 of 54 subjects who completed preventive therapy versus 2 of 57 who failed to complete. One of these two had been assigned to routine care, and the other to minimal methadone. In conclusion, INH retention time and completion rates were significantly improved by methadone treatment combined with observed INH, whether or not substance abuse counseling was provided. The results of this study indicate that methadone treatment offers clear public health benefits when it is used to deliver preventive medical services.