The effect of a comprehensive rehabilitation programme, including diet therapy, was studied in overweight male myocardial infarction (MI) patients randomly assigned to a rehabilitation group (n = 69) and a control group (n = 64). The results show a mean weight decrease of 3.8 kg in the rehabilitation group and a mean weight increase of 1.1 kg in the control group (t = -5.78, p less than 0.001). The proportion of patients reducing their weight more than 5 kg was 33% in the rehabilitation group and 5% in the control group (chi 2 = 15.6, p less than 0.001). Patient characteristics which significantly affected the weight change by univariate analysis were age, level of overweight, smoking habits, employment situation and social problems. Multivariate analyses yielded a two-variable model with moderate explanatory power in the rehabilitation group. The model includes the variables Body Mass Index and cessation of smoking before MI.