Combining vitamin C and carotenoid biomarkers better predicts fruit and vegetable intake than individual biomarkers in dietary intervention studies

Purpose The aim of this study was to determine whether combining potential biomarkers of fruit and vegetables is better at predicting FV intake within FV intervention studies than single biomarkers. Design Data from a tightly controlled randomised FV intervention study (BIOFAV; all food provided and two meals/day on weekdays consumed under supervision) were used. A total of 30 participants were randomised to either 2, 5 or 8 portions FV/day for 4 weeks, and blood samples were collected at baseline and 4 weeks for plasma vitamin C and serum carotenoid analysis. The combined bio-marker approach was also tested in three further FV intervention studies conducted by the same research team, with less strict dietary control (FV provided and no supervised meals). Results The combined model containing all carotenoids and vitamin C was a better fit than either the vitamin C only (P < 0.001) model or the lutein only (P = 0.006) model in the BIOFAV study. The C-statistic was slightly lower in the lutein only model (0.85) and in the model based upon factor analysis (0.88), and much lower in the vitamin C model (0.68) compared with the full model (0.95). Results for the other studies were similar, although the differences between the models were less marked. Conclusions Although there was some variation between studies, which may relate to the level of dietary control or participant characteristics, a combined biomarker approach to assess overall FV consumption may more accurately predict FV intake within intervention studies than the use of a single biomarker. The generalisability of these findings to other populations and study designs remains to be tested. Clinical trial Registration Number NCT01591057 (www.clinicaltrials.gov).