FODMAP project: Development, validation and reproducibility of a short food frequency questionnaire to estimate consumption of fermentable carbohydrates

Background & aims: Irritable bowel syndrome (IBS) is a functional disorder that is characterized by gastrointestinal symptoms and that has a major impact on quality of life, resulting in direct and indirect health care costs. The majority of patients with IBS suffer from food intolerances, most commonly related to the consumption of fermentable oligosaccharides, disaccharides, monosaccharides and polyols (FODMAPs). This study aimed to develop and verify the validity and reproducibility of a short food frequency questionnaire (FFQ) to assess typical FODMAP consumption in adults with IBS. Methods: The primary FFQ list consisted of source foods of FODMAPs that contributed at least 10% to the frequency of consumption among 855 adults from a population-based study in the municipality of Campinas in 2014/2015. In addition, source foods of FODMAPs (according to the Monash University Low FODMAP Diet application) and foods commonly consumed by the Brazilian population (according to the FFQ for adults validated in the city of Sao Paulo) were included. One hundred and five (n = 105) healthy subjects were recruited to respond to the FFQ twice and to respond the 24-h dietary recall (24HR) three times during a 3-month period. The relative validity of the proposed instrument was compared with the average of the three 24HRs, and the reproducibility of the instrument was assessed by comparing both FFQ applications. The following statistical analyses were used for validation and reproducibility: Wilcoxon's test, Spearman's correlation analysis, weighted kappa, Bland Altman's plot and index, and interclass correlation coefficient. Results: The final list of items for the short FFQ included 54 different foods. The foods were organized by FODMAP groups: free fructose, lactose, total oligosaccharides and total polyols, with variations of categories of responses for consumption frequency between 0 and 10 times and the unit of time in days, weeks or months. In the validity analyses, the correlation coefficients ranged from 0.209 (polyols) to 0.652 (lactose) (p < 0.05). There was no correlation between the methods in the fructose and oligosaccharide groups. The lactose group presented good agreement, and the remaining groups had a lack of agreement, with a mean of 15.7%. The Bland-Altman index values were 4.7% (fructose), 3.8% (lactose), 5.7% (oligosaccharides) and 6.6% (polyols). Regarding reproducibility, the interclass and Spearman's correlation coefficients varied from ICC = 0.781 and r = 0.725 (oligosaccharides) to ICC = 0.913 and r = 0.807 (lactose) (p < 0.05), showing strongly reproducible results for lactose and polyols and good results for fructose and oligosaccharides. Accurate agreement between FFQ applications had a mean of 67.3%, and 3.0% showed disagreement between FFQ1 and FFQ2. The weighted kappa coefficient ranged from 0.576 (polyols) to 0.645 (lactose). Conclusion: The semi-quantitative short FFQ was developed to evaluate the consumption of FODMAPs in adults in Sao Paulo. The instrument presents good reproducibility for all groups of FODMAPs, good validity for lactose and weaker validity for fructose, polyols and oligosaccharides. As the short FFQ was carefully designed for the study population, its estimates are relatively reliable at the population group level. A future reanalysis of this questionnaire would be useful when the chemical composition data of FODMAPs are available. (C) 2020 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.