Objective: To examine the association of energy and % energy from fat with body mass index (BMI) and determine if self-reported dieting altered observed associations.
Design: Dietary intake data based on dietary recalls from four nonconsecutive days over a 1 year period were examined relative to BMI. The relation between energy intake and % energy from fat and BMI was examined by linear regression analysis.
Subjects: The sample included 1854 free-living women aged 19-50 years who participated in the 1985-6 Continuing Surveys of Food Intakes by Individuals conducted by the United States Department of Agriculture.
Results: Reported energy intake was inversely associated with BMI (regression coefficient (beta) = -0.001 24, standard error (s.e.) = 0.000 31). Controlling for low energy dieting alone reduced the inverse energy intake-BMI association by approximately 20% (beta = -0.001 00, s.e. = 0.000 31), compared to reductions of 16%, 13% and 10%, respectively, when health status, age and education were added individually to the energy intake-BMI linear regression. Physical activity, smoking status, % energy from fat and report of low fat dieting did not reduce the energy intake-BMI association. Controlling for nondietary factors related to BMI and potentially influencing energy intake reduced the inverse energy intake-BMI association by approximately 22% (beta = -0.000 97, s.e. = 0.00025). Further adjustment for low energy dieting on day 1 reduced the inverse energy intake-BMI association by 40% (beta = -0.000 74, s.e. = 0.000 26), suggesting that intermittent energy restriction was a significant factor in the reduced energy intake reported among overweight women. Percent energy from fat was not associated with BMI (beta = 0.049, s.e. = 0.025, P = 0.055). Exclusion of 37 women reporting poor health status further attenuated the inverse association between energy intake and BMI (beta = -0.000 64), s.e. = 0.000 26), while it strengthened the previously non-significant positive association between % energy from fat and BMI (beta = 0.062; s.e. = 0.024).
Conclusion: Intermittent energy restriction appeared to be a significant factor in the reduced energy intake reported among overweight women in this sample. Adequate assessment of energy expenditure is required to correctly interpret the association of energy intake to body weight.