The aim of the present study was to determine the accuracy of reported energy intake according to a food-frequency questionnaire (FFQ) and dietary records (DR) in obese subjects with metabolic syndrome risk factors. Subjects were twenty-three men and twenty-seven women with mean BMI of 35.7 (range 30.5-43.8) kg/m(2) who participated in a dietary interview based on a FFQ and completed weighed DR. Total energy expenditure was measured with the doubly labelled water method. Total energy expenditure, measured RMR and physical activity level did not differ between under-reporters (50 % of the sample) and non-under-reporters. Under-reporters had lower median intake of sweets, desserts and snacks than non-under-reporters (100 v. 161 g/d (P = 0.0008) and 61 v. 128 g/d (P = 0.0002) according to the FFQ and DR, respectively). The DR also showed lower energy density (6.7 (sd 1.3) v. 7.9 (SD 1.6) kJ/g; P = 0.0064), lower intake of sugary drinks (0 v. 167 g/d; P = 0.0063) and higher scores for dietary restraint (9.0 (sd 5.0) v. 6.1 (SD 3.5); P = 0.0285) in under-reporters. Energy density was associated with accuracy according to the FFQ (Spearman's rank correlation coefficient (RS) 0.406; P = 0.0034) and the DR (RS 0.537; P < 0.0001). In multivariate analysis, consumption of bread and sweets, desserts and snacks measured by the FFQ was positively associated with accuracy (R(2)adjusted 0.46 (95 % CI 0.32, 0.70)). According to the DR, consumption of sweets, desserts and snacks was also associated with accuracy, as was dietary restraint (inversely) (R(2)adjusted 0.67 (95 % CI 0.54, 0.83)). In obese subjects with metabolic risk factors, intake of sweets, desserts and snacks, bread and dietary restraint were determinants of reporting accuracy.