Serum carotenoids are strongly associated with dermal carotenoids but not self-reported fruit and vegetable intake among overweight and obese women.

Background: Accurate assessment of fruit and vegetable intake (FVI) is essential for public health nutrition research and surveillance. Blood carotenoid concentrations are robust biomarkers of FVI, but collecting blood samples typically is not feasible in population-based studies. Understanding how well non-invasive measures compare to blood estimates is important for advancing surveillance and evaluation. The objective of this study was to examine the associations between serum carotenoids and four non-invasive measures of FVI in overweight and obese women.

Methods: This study utilized baseline data from 157 overweight or obese women (95.5% white, mean age 58.56 years ± 9.49 years) enrolled in the Strong Hearts, Healthy Communities randomized trial, including two direct measures of carotenoids and three self-reported measures of FVI. Participants completed a fasting blood draw, dermal carotenoid scans using resonance Raman spectroscopy (RRS), a two-item FVI screener modeled after the American Heart Association's Life's Simple 7 "My Life Check" tool (modified AHA tool), the National Cancer Institute's All-Day Fruit and Vegetable Screener (FVS), multiple 24-h dietary recalls, physical measurements, and demographic and health behavior questions. We analyzed blood for total carotenoids and derived total FVI estimates from self-report tools. We used multivariate linear regression models to examine associations between each non-invasive tool and serum carotenoids under four scenarios analogous to different research contexts in which varying breadths of participant data are available. We also calculated adjusted Pearson's correlations between serum carotenoids, dermal carotenoids, and the self-reported measures.

Results: Dermal carotenoids were strongly correlated with serum carotenoids (0.71, P < 0.00067) and associated with serum carotenoids in all regression models (0.42-0.43, P < 0.002). None of the self-reported FVI measures were significantly associated with serum or dermal carotenoids in adjusted regression models or correlation analyses.

Conclusions: Compared to self-reported FVI, we found dermal carotenoids measured by RRS to be a superior method to approximate serum carotenoids among overweight and obese women. More research is needed to investigate these assessment methods in diverse populations.

Trial Registration: ClinicalTrials.gov Identifier: NCT02499731, registered July 16, 2015.