Comparison of energy intakes determined by food records and doubly labeled water in women participating in a dietary-intervention trial

The accuracy of estimates of usual energy intake derived from food records in participants of a long-term dietary-intervention trial was studied in a subset of 29 women aged 48.7 +/- 5.0 y and weighing 61.9 +/- 6.5 kg. This sample was similar to the population in the whole trial (n = 715), from which it was selected in terms of age, weight, body mass index (BMI), and reported energy and fat intakes. During the validation study, reported energy intake was derived from 7 consecutive days of food records, and total energy expenditure was measured by the doubly labeled water method over 13 d. Reported energy intake (6.98 +/- 1.58 MJ/d) was significantly lower than energy expenditure (9.00 +/- 2.08 MJ/d) and represented 79.8 +/- 17.6% of expenditure. The correlation between reported energy intake and expenditure was 0.46 (P = 0.01, 95% CI: 0.15, 0.71). Body weight, BMI, height, length of time in the dietary trial, and percentage of energy from fat and carbohydrate were not significantly associated with the accuracy of reporting. These results indicate that energy intake derived from food records is an imprecise measure that substantially underestimates energy intake in middle-aged women participating in a long-term dietary-intervention trial.