Dexfenfluramine (dF) was compared to placebo as adjuvant to a very energy-restricted diet (1.6-4.2 MJ/24 h). The diet was continued as long as possible or until a satisfactory weight loss had been achieved, and dietary efforts were continued throughout the study. Of the 37 females and five males included, 71 per cent could be followed up for 12 months. Initial BMI ranged from 28 to 54 kg/m2. The lowest body weight was reached 1 month earlier in the dF group (P = 0.037). Throughout the study, the reduction of excess weight (REW) was greater in dF patients (P less than 0.05 only at 4 and 6 months). At 6 months, excess weight had declined by 15 per cent more in the dF group than in the placebo group (95 per cent confidence limits of the median being 1-31 per cent). Between 6 and 12 months, both groups regained weight significantly, the rates of regain differing only insignificantly. At 12 months, excess weight showed a net decrease of only 8 per cent more in the dF group than in the placebo group (95 per cent confidence limits being -7 to +24 per cent). Therefore, REW showed no significant group difference after 12 months. Type of obesity (android or gynoid) as determined by waist to hip ratio had no significant impact on either weight loss, REW, reductions of waist and hip circumferences, or on waist to hip ratio changes. S-alkaline phosphatases and s-uric acid declined significantly in the dF group only. Side-effects were all mild and their prevalence showed no group difference.(ABSTRACT TRUNCATED AT 250 WORDS)