(P20-41-26) Polyphenol Estimator: A New Tool to Estimate Dietary Polyphenol Intake From ASA24 and NHANES Dietary Data

Objectives: Recent food-based recommendations for flavan-3-ols highlight a growing need to understand the breadth of our dietary polyphenol exposure. However, the estimation of dietary polyphenol intake remains challenging, requiring custom computational tools that are often difficult to implement or not fully reproducible. Thus, we developed an automated and user-friendly tool to estimate polyphenol intake from diet recalls and records

Methods: Polyphenol Estimator can process diet data which utilize the Food and Nutrient Database for Dietary Studies (FNDDS) infrastructure, such as the Automated Self-Administered 24-Hour Dietary Assessment Tool (ASA24) and Automated Multiple-Pass Method from the National Health and Examination Survey (NHANES). In brief, Polyphenol Estimator disaggregates foods using the FDA Food Disaggregation Database (up to 986 unique ingredients). Ingredients are matched to FooDB to derive food polyphenol content (up to 356 unique quantified polyphenols and 189 unique foods). Polyphenol intake is then summarized at various levels (total, class, compound). Users may use their polyphenol estimates to calculate the Dietary Inflammatory Index (DII). Polyphenol Estimator is freely available on GitHub (https://bit.ly/polyphenol-estimator) with a start guide for users with limited programming experience.

Results: As an illustration of Polyphenol Estimator, we applied it to two days of diet recalls from adults at least 20 years old in NHANES 2021 - 2023 (n = 2278). For 97.7% of participants, less than 2.5% of reported foods went unmapped, with 75.7% of participants having complete mappings. Total polyphenol intake was 517 ± 439 (mean ± SD) mg/1000 kcal, stemming largely from green tea, coffee, black tea, apples, wine, oranges, and blueberries. At the class level, polyphenols classified as organooxygen compounds, flavonoids, and cinnamic acids and derivatives were the top contributors to intake. At the compound level, cyptochlorogenic acid, neocholorogenic acid, and caffeic acid were top contributors. Lastly, the DII was 1.2 ± 1.9, indicating the average diet had proinflammatory potential.

Conclusions: Polyphenol Estimator offers a way to obtain polyphenol estimates from diet data. This tool can assist future efforts to understand polyphenol intake exposures and their biological impact on human health.

Funding Sources: USDA-ARS