Influence of Using Different Sources of Carotenoid Data in Epidemiologic Studies

Abstract

Objective This study compared distributions of carotenoid intake and diet-serum correlations using two sources of carotenoid data: the US Department of Agriculture-National Cancer Institute (USDA-NCI) carotenoid food composition database and values accompanying the Block-NCI Health Habits and History Questionnaire (HHHQ). Design and subjects A 100-item food frequency questionnaire was used to collect dietary data from 2,152 adults, aged 43 to 85 years, who were participating in the Nutritional Factors in Eye Disease Study, a population-based study designed to evaluate nutritional factors associated with age-related eye disease. Blood samples were collected from a random sample of 400 nonfasting participants in the study. Results Median carotenoid intakes using HHHQ vs USDA-NCI data were alpha carotene (229 vs 223 μg/day), beta carotene (1,321 vs 1,325 μg/day), beta cryptoxanthin (72 vs 21 μg/day), lutein + zeaxanthin (653 vs 811 μg/day), and lycopene (593 vs 1,615 μg/day). All paired differences in carotenoid intake were significantly different from zero (Wilcoxon signed-rank, P<.0001). Despite these differences, the two databases similarly ranked individuals according to carotenoid intake: Spearman correlations ranged from .71 (lycopene) to .93 (alpha carotene). Differences between diet-serum correlations (adjusted for energy, body mass index, high density lipoprotein, and total cholesterol) using HHHQ vs USDA-NCI data were minor and not significant ( P>.05): alpha carotene ( r=.33 vs .32), beta carotene ( r=.27 vs .32), beta cryptoxanthin (r=48 vs .53), lutein + zeaxanthin (r=.28 vs .24), and lycopene (r=.29 vs .25). Conclusions Although estimates of carotenoid intake differed significantly, only minor differences in carotenoid rankings and diet-serum correlations were observed using either data source in this population. J Am Diet Assoc. 1996; 96:1271-1275.