Comparison of 24-Hour Recall and 3-Day Food Records during the Complementary Feeding Period in Thai Infants and Evaluation of Plasma Amino Acids as Markers of Protein Intake

Background: An accurate and reliable measurement of nutrient intake is the first and foremost step in order to optimise infant nutrition and evaluate its impact on health outcomes. However, research on the validity of dietary assessment tools used during the weaning period is limited, especially in lower-middle income countries. The primary aim of this study was to evaluate relative validity of a 24-h recall method (24-HR) using a 3-day food record (3-DFR). A secondary aim was to investigate association between protein intake from 3-DFR and plasma amino acids as a potential protein biomarker. Methods A multicentre, prospective cohort study was conducted in Chiang Mai, Thailand from June 2018 to May 2019. Food consumption data were collected in healthy infants using 24-HR and 3-DFR at 9 and 12 months of age. Blood samples were obtained at 12 months (M). Plasma amino acids were analysed using high performance liquid chromatography. Results Of 145 infants, 49% were female. At group level, paired t-tests/Wilcoxon signed rank tests did not show significant differences between average nutrient intakes from the 2 dietary assessment methods, except for vitamin A and vitamin C. Weighted kappa (Kw) was acceptable for all nutrients, except for vitamin A intake at 9 M (Kw = 0.15). The Bland–Altman analyses were unbiased for most nutrients with variable limits of agreement. At individual level, correlation coefficients (r) ranged from acceptable to excellent (r = 0.37–0.87) while cross-classifications showed acceptable outcomes, except for vitamin A. Multivariate analyses showed significant associations between protein intake at 12 M from the 3-DFR and plasma concentrations of branched-chain amino acids (BCAA) and essential amino acids (EAA), even after adjusting for gender, milk feeding type and energy intake. Conclusions For infants aged 9–12 M, a 24-HR can be used as a more practical alternative to a 3-DFR for most nutrients although caution is required for some micronutrients, especially vitamin A. A repeated interview might further improve the accuracy. Furthermore, protein intake, particularly animal-based protein, significantly predicted plasma BCAA and EAA concentrations regardless of gender, type of milk feeding and energy consumption.