Background. A diet rich in fruits and vegetables is important for both children and adults, but dietary assessment of these foods for research and surveillance is difficult. Traditional self-report methods are subject to measurement error, and biomarkers of intake are increasingly used in dietary research. Carotenoids in blood or tissue are considered the best current biomarker of fruit and vegetable intake due to their widespread distribution in these foods, but biochemical analyses of carotenoids are invasive and impractical for use with children. Resonance Raman spectroscopy (RRS), a noninvasive method of measuring carotenoid status in skin using light, provides an alternative to biochemical analyses of carotenoid status. Research using RRS in adults was initiated by our group, but unanswered questions remained about the utility of this methodology. The overall aim of this dissertation was to evaluate RRS as a biomarker of fruit and vegetable intake in vivo for its eventual use in nutritional epidemiologic research.

Methods. Study I built on our previous work in healthy adults to identify time-dependent predictors of dermal carotenoid status, measured by RRS. Sun exposure and physical activity were select for examination as possible longitudinal determinants of carotenoid status. The association between time-fixed demographic covariates and dermal carotenoid status was also examined in detail. Studies II and III involved original data collection to broaden the RRS application for use in diverse populations. Specifically, Study II involved 381 economically-disadvantaged preschool children, who were enrolled into a baseline sample to estimate the variability and identify main determinants of carotenoid status in a population of children, using RRS. Study III used RRS to evaluate the effect of a 16-week fruit and vegetable intervention on change in dermal carotenoid status among economically-disadvantaged children by obtaining pre- and post-intervention RRS measures for 289 children.

Results. Study I: Self-reported recent sun exposure was associated with lower dermal carotenoid status, measured by RRS. We also observed seasonal variation in dermal carotenoid status over a six month period of time. Studies II and III demonstrated that RRS is a feasible, noninvasive method of nutritional status in the field, particularly among young children, which correlated with dietary intake. Study II: we observed variability in dermal carotenoid status in the preschool population similar to the variability we observed in adults. Carotenoid status was lower in younger children, those who had a higher BMI, and those whose family participated in the federal Supplemental Nutrition Assistance Program (SNAP). Study III: RRS did not detect a significant effect of the intervention, but lower baseline carotenoid status predicted a greater positive change in carotenoid status at post-intervention.

Conclusions. RRS was correlated with dietary intake, suggesting RRS is a feasible and valid biomarker of fruit and vegetable in children and adults. Further research is needed to continue the development of this technology for eventual use in epidemiologic research and surveillance.