To assess continued efficacy of anorexiants after 3 years of use, 52 participants (43% of those starting) entered a second double-blind trial to compare 60 mg sustained-release fenfluramine plus 15 mg phentermine resin versus placebo added to behavior modification, caloric restriction, and exercise. Although participants in both the active medication and placebo groups gained weight, participants receiving fenfluramine plus phentermine (n = 27) gained significantly (p less than 0.01) less (4.4 +/- 0.5 kg or 5.3% +/- 0.5% of initial weight) than participants receiving placebo (n = 24) (6.9 +/- 0.8 kg or 8.5% +/- 1.1% of initial weight). At week 190, both groups were still below their initial weight (fenfluramine plus phentermine group, 5.0 +/- 1.4 kg; placebo group, 2.1 +/- 1.2 kg; p less than 0.01). Overall, 12 participants (23.5% of those still in the study) were greater than or equal to 10% below initial weight. One participant dropped out during this phase because of personal reasons and loss of medication efficacy. During the 30 weeks, participants receiving fenfluramine plus phentermine had 26 moderate or severe complaints versus eight participants receiving placebo. Fenfluramine plus phentermine provided better appetite control and only slightly more bother. Analysis of participant response in this phase by treatment assignment in the first double-blind phase (weeks 6 to 34) indicated that initial receipt of medication did not have negative learning effects. Eleven participants receiving active medication between weeks 6 and 34 and receiving placebo between weeks 160 to 190 gained 5.1 +/- 1.0 kg. In contrast, 13 participants originally taking placebo gained 8.3 +/- 9 kg in this second double-blind phase.(ABSTRACT TRUNCATED AT 250 WORDS)