Energy Intake in Socially Vulnerable Women Living in Brazil: Assessment of the Accuracy of Two Methods of Dietary Intake Recording Using Doubly Labeled Water.

Background: Diet is related to the prevention of chronic diseases, but assessing dietary intake is a complex task, especially in socially vulnerable populations.

Objectives: To assess the accuracy of the 24-hour food recall (24HFR) and the food frequency questionnaire (FFQ) methods in socially vulnerable women in Brazil and compare these methods against doubly labeled water (DLW).

Design: Cross-sectional study. Energy intake (EI) was measured using 3 24HFRs and 1 FFQ. Total energy expenditure (TEE) was measured using DLW. Cutoff points were calculated to assess underreporting and overreporting using both the difference and the ratio between the EI and TEE.

Participants/setting: Sixty-seven socially vulnerable, nondieting adult women with stable body weight and mothers of malnourished children.

Main Outcome Measures: EI, in kilocalories, derived from 24HFR and FFQ.

Statistical Analyses: Repeated measures analysis of variance was used to compare EI and TEE and Bland-Altman graphs were plotted to assess the agreement between these variables. Spearman correlation coefficient between the ratio of EI to TEE and socioeconomic or anthropometric variables was calculated.

Results: Mean TEE was 2,186 kcal (95% CI 2,063 to 2,309 kcal). EI obtained through 24HFR (mean=1,848.6 kcal [95% CI 1,737.5 to 1,959.7 kcal]) was significantly lower than TEE (P=0.01), and significantly lower than the EI obtained through FFQ (mean=2,084.5 [95% CI 1,929.0 to 2,240.0 kcal]; P<0.01). There were no statistically significant differences between the EI values of FFQ and TEE (P=0.89). The distribution of underreporters, nonunderreporters, and overreporters were more homogenous in the FFQ method. There was a positive correlation between body fat percentage and underreporting in the FFQ method (r=0.245; P=0.046).

Conclusions: The methods of assessing dietary intake in our study showed poor agreement with TEE obtained by DLW. These may, therefore, not be the most suitable methods for assessing EI in this population.