Weight change after myocardial infarction: statistical perspectives for future study

OBJECTIVES: Survivors of acute myocardial infarction (MI) often experience weight changes and weight management is often appropriate. Estimates of the sample size required in intervention weight loss studies are essential to the success of future evaluations. The aims of this study were therefore two-fold: (1) to describe pilot data on the effectiveness of advice for weight loss; (2) to provide information on the sample size required for future research to assess weight management in similar patients. METHODS: Further analysis of data from a randomized controlled study carried out in 84 post-MI patients attending cardiac rehabilitation. Forty-three intervention patients received dietary advice in line with current UK guidelines. Additionally, overweight intervention subjects were given individualized weight management advice. Forty-two control patients were recruited and received usual care. Anthropometric measurements were made at baseline and followed up at 52 weeks post-rehabilitation. Power calculations were performed using these data to determine the required sample size to adequately power a study examining the effectiveness of weight management. RESULTS: Seventy patients completed the study. At 52 weeks anthropometric measurements were unchanged in the 25 overweight patients provided with weight management advice, and also for all those (n = 20) with body mass index <25 kg m-2. In contrast, anthropometric measurements increased significantly (body weight +2.4 kg, P = 0.02; waist circumference +2.6 cm, P = 0.008) in overweight control patients. A minimum sample size of 71 patients is required for a future study of weight change in overweight subjects, and 58 for a study of healthy weight subjects. CONCLUSIONS: Pilot data suggested that significant weight changes occur in patients not given nutritional or weight management advice after MI. Power calculations carried out on these data indicate that a sample size at least three times as large as the present study is required to accurately evaluate weight management in this group.