A vaccine based on autoclaved Mycobacterium w was administered, in addition to standard multidrug therapy (MDT), to 156 bacteriologically positive, lepromin negative multibacillary leprosy patients compared to a well matched control group of 145 patients with a similar type of disease who received a placebo injection in addition to MDT. The MDT was given for a minimum period of 2 years and continued until skin smear negativity, while the vaccine was given at 3-month intervals up to a maximum of eight doses. The fall in clinical scores and bacteriological indices was significantly more rapid in vaccinated patients, from 6 months onward until years 2 or 3 of therapy. However, no difference was observed in the fall in bacteriological index in the two groups from year 4 onwards. The number of LL and BL patients released from therapy (RFT) following attainment of skin smear negativity, after 24-29 months of treatment was 84/133 (63.1%) in vaccinated and 30/120 (25.0%) in the placebo group; the difference was highly statistically significant (P < 0.0001). In all, 90.2% patients (146/162) converted from lepromin negativity to positivity in the vaccine group, as against 37.9% (56/148) in the placebo group. The average duration of lepromin positivity maintained following eight doses of vaccine administered over 2 years was 3.016 years in the vaccine and 0.920 years in the placebo group. Histological upgrading after 2 years of treatment in the LL type was observed in 34/84 (40.5%) cases in the vaccine and 5/85 (5.9%) cases in the placebo group, the difference being statistically significant (P < 0.001). The incidence of type 1 reactions was significantly higher (30.5%) in the vaccine group than (19.7%) in the placebo group (P = 0.0413); the difference was mainly observed in LL type (P = 0.009). The incidence of type 2 reactions was similar (31.8 and 34.6%) in vaccine and placebo groups. The vaccine did not precipitate neuritis or impairments over and above that encountered with MDT alone. After 5 years of follow-up following RFT, no incidence of bacteriological or clinical relapses was observed in both groups.