Public Health Nutrition

Assessment and methodology

Strength of the relationships between three self-reported dietary intake instruments and serum carotenoids: the Observing Energy and Protein Nutrition (OPEN) Study

Stephanie M Georgea1 c1, Frances E Thompsona1, Douglas Midthunea2, Amy F Subara1, David Berrigana1, Arthur Schatzkina3 and Nancy Potischmana1

a1 Applied Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, 6130 Executive Blvd, EPN 4017A, Bethesda, MD 20892, USA

a2 Biometry Research Group, Division of Cancer Prevention, National Cancer Institute, Bethesda, MD, USA

a3 Nutritional Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD, USA


Objective To assess the strength of the relationships between serum carotenoids and three self-reported dietary intake instruments often used to characterize carotenoid intake in studies of diet and disease.

Design Participants completed a Diet History Questionnaire (DHQ), two 24 h dietary recalls (24HR), a fruit and vegetable screener and a fasting blood draw. We derived dietary intake estimates of α-carotene, β-carotene, cryptoxanthin, lutein, zeaxanthin and lycopene from each diet instrument and calculated sex-specific multivariate correlations between dietary intake estimates and their corresponding serum values.

Setting Montgomery County, Maryland, USA.

Subjects Four hundred and seventy women and men aged 40–69 years in the National Cancer Institute's Observing Protein and Energy Nutrition (OPEN) Study.

Results Serum carotenoids correlated more strongly with the DHQ (r = 0·34–0·54 for women; r = 0·38–0·56 for men) than with the average of two recalls (r = 0·26–0·47 for women; r = 0·26–0·40 for men) with the exception of zeaxanthin, for which the correlations using recalls were higher. With adjustment for within-person variation, correlations between serum carotenoids and recalls were greatly improved (r = 0·38–0·83 for women; r = 0·42–0·74 for men). In most cases, correlations between serum carotenoids and the fruit and vegetable screener resembled serum–DHQ correlations.

Conclusions Evidence from the study provides support for the use of the DHQ, a fruit and vegetable screener and deattenuated recalls for estimating carotenoid status in studies without serum measures, and draws attention to the importance of adjusting for intra-individual variability when using recalls to estimate carotenoid values.

(Received July 05 2011)

(Accepted November 09 2011)

(Online publication January 10 2012)


c1 Corresponding author: Email