The present study aimed to assess the contribution of energy intake to positive energy balance and weight gain with insulin therapy. Changes in energy intake (self-report and weighed food intake), dietary behavior (auto-questionnaires), resting energy expenditure (REE) (indirect calorimetry), physical activity (accelerometry), and glucosuria were monitored over the first 6 months of insulin therapy in 46 diabetic adults. No change in REE, activity, or glucosuria could explain weight gain in the type 1 (4.1 +/- 0.6 kg, P < 0.0001) or type 2 (1.8 +/- 0.8 kg, P = 0.02) diabetic groups. An increase in energy intake provides the most likely explanation for weight gain with insulin. However, it is not being recognized because of significant underestimation of self-reported food intake, which appears to be associated with increased dietary restraint.