A minimum of six days of diet recording is needed to assess usual vitamin K intake among older adults

Abstract

There is a growing interest in the role of vitamin K in health, especially in aging populations. Knowledge of inter- and intra-individual variability of dietary vitamin K intake could be useful to accurately assess usual intake and rank participants in epidemiological studies. Our objectives were to: 1) estimate the variance components of vitamin K intake; 2) investigate whether day of the week, season, and energy intake are factors related to intra-individual variance; and 3) calculate the requisite number of days to achieve desired degrees of accuracy for estimating individual vitamin K intake, ranking individuals and estimating regression coefficient. Vitamin K intake was assessed in 939 older adults (67-84 y) enrolled in the Qué bec Longitudinal Study on Nutrition and Successful Aging study using 2 sets of 3 nonconsecutive multiple-pass 24-h dietary recalls (24HR) collected 6 mo apart. Each set included 2 weekdays and one weekend day. Intra- to interindividual variance ratios for vitamin K intake were 3.2 (95% CI = 2.6-3.9) overall, 2.6 (95% CI = 2.1-3.5) for men, and 3.7 (95% CI = 2.9-5.0) for women. Day of the week (weekdays) and season (May to October) were positively and significantly associated with vitamin K intake but explained a negligible part of intra-individual variation (<1%). Adjusting for energy intake explained <7% of variance and did not affect the variance ratio. Six to 13 24HR are required to properly rank individuals according to their usual vitamin K intake and limit attenuation of the regression coefficient. These results should be considered in studies planning to assess vitamin K intakes in older adults. © 2011 American Society for Nutrition.