Validation of a food frequency questionnaire in the assessment of dietary glycemic index, glycemic load, and protein intake in pregnant women with obesity

Objectives: Studies suggest that diets with a low glycemic index and high protein are favorable in aiding weight loss and improving weight maintenance; however, methods to measure dietary intake are comprehensive both for the participant and the study staff. We aimed to validate the accuracy of the dietary glycemic index and protein intake assessed through a food frequency questionnaire against a 4-d weighed food record in Danish pregnant women with obesity. Methods: A total of 31 pregnant women completed a 29-item food frequency questionnaire and a 4-d weighed food record with overlapping time periods. The women had a mean (± SD) age of 30.6 ± 3.9 y and a prepregnancy body mass index of 33.9 ± 3.5 kg/m². We evaluated the validity of the food frequency questionnaire by Bland-Altman plots and the Spearman correlation coefficient. Results: The results of the validation study found good acceptance of the 29-item food frequency questionnaire. The mean intake of glycemic index, glycemic load, and protein intake of the 29-item food frequency questionnaire and the weighed food record correlated well, although intake data of the 29-item food frequency questionnaire tended to be lower. Spearman correlation coefficients had moderate to high correlations for glycemic index (ρ = 0.73; P < 0.001) and protein intake (ρ = 0.70; P < 0.001). A moderate correlation was found for glycemic load (ρ = 0.55; P = 0.002). There was no correlation for carbohydrates (ρ = 0.21; P = 0.253). Conclusion: The results suggest no risk of bias between the two methods of assessment; hence, a 29-item food frequency questionnaire can be used to assess the mean glycemic index, glycemic load, and protein intake in pregnant women with obesity.