Two Spectroscopy Devices Can Approximate Fruit and Vegetable Intake in a Racially and Ethnically Diverse Sample

Background: Assessing fruit and vegetable intake is important to determine the effectiveness of nutrition intervention studies. Few objective measures of fruit and vegetable intake have been directly compared within the same study participants. Objectives: This study aims to compare the relative validity of 2 different skin spectrophotometry devices to serve as biomarkers of plasma carotenoid concentrations, self-reported fruit and vegetable intake (FVI), and carotenoid intake, to examine if both devices can be used to approximate FVI. Methods: Human skin carotenoid scores (SCS, Veggie Meter-assessed) and color space measurement data from 2 prior studies were used. Participants’ skin color space parameters (lightness, redness, and yellowness) were measured using the Konica Minolta CM 700D spectrophotometer. Self-reported diet was assessed using food frequency questionnaires. We analyzed data from the 2 prior studies to examine correlations of skin lightness, redness, and yellowness parameters with SCS, plasma carotenoids, and FVI and carotenoid intake. Results: Skin lightness and redness demonstrated weak correlations with skin carotenoids (r = –0.15 to 0.14), plasma carotenoids (r = –0.19 to 0.21), and fruit and vegetable (FV) and carotenoid intake (r = –0.14 to 0.13). Conversely, skin yellowness demonstrated statistically significant (P < 0.05/63) correlations with SCS ranging from r = 0.60 to r = 0.67. The correlations between skin yellowness and total plasma carotenoids (r = 0.46–0.57), plasma α-carotene (r = 0.38–0.54), β-carotene (r = 0.50–0.57), α- and β-cryptoxanthin (r = 0.38–0.43), and lutein and zeaxanthin (r = 0.30–0.48) were all statistically significant (P < 0.00079), but the correlation between skin yellowness and plasma lycopene was not statistically significant (r = 0.06–0.17). There were positive, statistically significant correlations between skin yellowness and FVI in study 1 (r = 0.27), as well as positive, statistically significant correlations between changes in skin yellowness with changes in SCS (r = 0.51), changes in total plasma carotenoids (r = 0.45), and changes in plasma α-carotene (r = 0.51) and β-carotene (r = 0.45). Conclusions: Both devices can be used to approximate FVI. More research is needed to understand the reasons the devices may not detect plasma lycopene well. This study was registered at clinicaltrials.gov as NCT04056624 (https://clinicaltrials.gov/c t2/show/NCT04056624).