Abstract
BACKGROUND: Inaccurate energy intake assessments and dietary underreporting are important barriers to assess reliable health correlates of food consumption in nutritional epidemiology. Studies that do not account for this phenomenon may result in spurious diet-health associations. Whether underreporting occurs more frequently with special diets remains subject to investigation. OBJECTIVES: This study aimed to test the hypothesis whether low-calorie and carbohydrate-restrictive diets were associated with increased odds for energy intake underreporting and investigate whether a lower carbohydrate intake (in %/total energy intake) was associated with a higher discrepancy between self-reported energy intake and total energy expenditure. METHODS: This study used a predictive equation derived from 6497 doubly labeled water measurements to detect erroneous self-reported energy intake in the National Health and Nutrition Examination Surveys (NHANES, 2009-2018). Weighted underreporting prevalence was compared among 3 groups, namely, the United States general population without a special diet, individuals who reported low-calorie diets, and individuals who reported carbohydrate-restrictive diets. Crude and multivariate logistic binomial regression models were built to examine associations between diet and energy intake underreporter status. RESULTS: Data from 18,150 adult NHANES participants ≥20 y were analyzed. Underreporting occurred almost twice as often in participants reporting low-calorie diets [38.84% (CI: 34.87, 42.95)] and carbohydrate-restrictive diets [43.83% (CI: 33.02, 55.26)] as compared with the general population [22.89% (CI: 21.88, 23.93)]. Both diets were associated with significantly higher odds for underreporting even after an adjustment for sociodemographic factors [odds ratio (OR): 2.32; CI: 1.93, 2.79 and OR: 2.86; CI: 1.85, 4.42, respectively]. Subanalyses in participants denying any weight loss intention/with stable weight revealed a comparable picture. The lowest level of agreement between total energy expenditure and self-reported energy intake was found in carbohydrate-restrictive diets. CONCLUSIONS: Our findings have far-reaching implications, especially with regard to studies that associated carbohydrate restriction or low-calorie diets with favorable health outcomes while not accounting for the herein-suggested phenomena.