Development and Validation of a Comprehensive Semi-Quantitative Food Frequency Questionnaire that Includes FODMAP Intake and Glycemic Index

Background Fermentable, short chain carbohydrates (FODMAPs) have been identified as triggers for functional gastrointestinal symptoms. In addition, excess FODMAP consumption has been implicated in the onset of Crohn's disease, and animal studies suggest that a low glycemic index diet can impair absorption of fructose, a major dietary FODMAP. Such hypotheses cannot be tested without the ability to quantify FODMAP ingestion with a validated dietary assessment tool. Objective To assess the validity and reproducibility of a 297-item comprehensive, semi-quantitative food frequency questionnaire (FFQ) in estimating intake of macro- and micronutrients, FODMAPs, and glycemic index/load. Subjects/setting One hundred healthy participants were recruited to complete the FFQ on two occasions, plus four 1-week food diaries kept during a 12-month period. Participants exhibiting major dietary change during the study period or low energy reporting on the FFQ were excluded. Main outcome measures Validation and reproducibility of the semi-quantitative FFQ by comparison with the mean of four 1-week food diaries. Statistical analyses performed Validation was assessed using Wilcoxon signed rank test, Spearman's correlation, Bland-Altman, and weighted K statistics. Reproducibility was examined using Shrout-Fleiss intraclass correlation coefficient. Results Seventy-two participants fulfilled inclusion and exclusion criteria. Demographics of the participants were comparable with 2006 Australian Census data. Consistent with other reported FFQs, the FFQ overestimated nutrient intake by a mean 140% (range=95% to 249%). However, based on the other analyses performed, it demonstrated validity for intake of sugars, fiber, alcohol, glycemic index, glucose, FODMAPs, calcium, folate, phosphate, potassium, iron, and magnesium; moderate validation for energy, total fat, saturated fat, carbohydrates, sodium, thiamin, sucrose, and retinol; poor validation for protein, mono/polyunsaturated fat, starch, glycemic load, niacin, and zinc. Riboflavin intake was not validated. Intraclass correlation coefficients for reproducibility ranged from 0.352 to 0.928. Conclusions The FFQ was validated for assessment of a wide range of nutrients, including the new class of carbohydrates, FODMAPs, and glycemic index. This provides a useful tool for dietary research, particularly in the area of gastroenterological disorders. J Am Diet Assoc. 2010;110:1469-1476.