Estimates of the components of nutrient intake variation are needed for modelling distributions of usual intake or predicting the usual intake of individuals. Season is a potential source of variation in nutrient intakes in addition to within- and between-person variation, particularly in low- or middle-income countries. We aimed to describe seasonal variation in nutrient intakes and estimate within-person, between-person and other major components of intake variance among Zambian children. Children from rural villages and peri-urban towns in Mkushi District, Zambia aged 4-8 years were enrolled in the non-intervened arm of a randomised controlled trial of pro-vitamin A carotenoid biofortified maize (n 200). Up to seven 24-h dietary recalls per child were obtained at monthly intervals over a 6-month period covering the late post-harvest (August-October), early lean (November-January) and late lean (February-April) seasons (2012-2013). Nutrient intakes varied significantly by season. For energy and most nutrients, intakes were highest in the early lean season and lower in the late post-harvest and late lean seasons. Season and recall on a market day had the strongest effects on nutrient intakes among covariates examined. Unadjusted within- to between-person variance ratios ranged from 4·5 to 31·3. In components of variance models, season accounted for 3-20 % of nutrient intake variance. Particularly in rural settings in low- and middle-income countries, where availability of locally grown, nutrient-rich foods may vary seasonally, studies should include replicates across seasons to more precisely estimate long-term usual intakes.
Keywords: 24-h recall; Dietary assessment; Nutrient intake; Preschool-aged children; School-aged children; Season; Sub-Saharan Africa.