Findings in 3 clinical trials challenge the accuracy of the Institute of Medicine's estimated average requirements for vitamin A in children and women

Abstract

Background: Vitamin A (VA) estimated average requirements (EARs) for women and children are extrapolated from rats and adult males. The retinol isotope dilution (RID) test can sensitively characterize VA status and intake requirements.

Objectives: These studies evaluated current EARs for children 4-8 y and women 19-30 y old.

Methods: Zambian children (n = 133, ages 5-7 y), US women (n = 51, ages 19-27 y), and Indonesian women (n = 29, ages 19-30 y) were provided diets or supplements containing 30%-155% of VA EARs for 42-90 d. RID was performed before and after the intervention to quantify changes in total body VA stores (TBSs) and total liver VA reserves (TLRs). Linear regression was performed between VA intake and change in TBSs or TLRs.

Results: Baseline mean ± SD TLRs were hypervitaminotic in Zambian children (1.13 ± 0.41 μmol VA/g liver), optimal in US women (0.46 ± 0.32 μmol/g VA/g liver), and deficient to marginal in Indonesian women (0.10 ± 0.08 μmol VA/g liver). VA intakes, resulting in no change in TBSs or TLRs, were 185 (95% CI: 18, 288) or 257 (95% CI: 124, 411) and 285 or 330 (CIs undefined) μg retinol activity equivalents (RAE)/d in the Zambian and US trials, respectively, but inconclusive in Indonesian women. The regression was not significant in either group of women.

Conclusions: Point estimates of VA intakes to maintain stores were below the current EARs of 275 (children) and 500 (women) μg RAE/d despite the TLRs being higher than the EARs were formulated to maintain (i.e., 0.07 μmol VA/g liver). Interventions based on these EARs may need to be scaled back. Lack of change in VA stores in women taking lower doses may result from physiological adaptation resulting in lower VA utilization. Longer, larger, and controlled studies are needed to accurately define EARs for VA.These trials were registered at Clinicaltrials.gov as NCT04123210 and NCT01814891.

Keywords: EAR; dietary requirements; retinol; stable isotope; vitamin A.

FIGURE 1

Study design and ¹³C₂-RID protocol. (A) Timeline for each study. The 14-d ¹³C₂-RID test was performed at baseline and after a washout after a period of VA supplementation. Zambian children consumed a study menu and a supplement or placebo 6 d/wk for 90 d. Diet was recorded on the seventh day to account for home VA intake. US women consumed a study menu and supplement or placebo daily for 42 d. Indonesian women were given a VA supplement or placebo daily for 60 d. (B) ¹³C₂-RID test protocol as performed in all 3 studies. RID, retinol isotope dilution; VA, vitamin A.

FIGURE 2

CONSORT flowcharts for studies to estimate the dietary vitamin A intake to maintain stores in US women (A) and Indonesian women (B). ^†Specific causes for exclusion are not known. RAE, retinol activity equivalents; RID, retinol isotope dilution.

FIGURE 3

Linear regression (and 95% CI) of TBSs (A, C) or TLRs (B, D) on daily intake of VA from food and supplements in hypervitaminotic Zambian children [using the placebo (Z₁) and supplement (Z₂) groups to avoid error from variable conversion of carotenoids, group Z_βc is plotted for reference only (triangles)] (A, B) and VA-sufficient US women (C, D). The estimated average requirement is calculated at x = 0 (the daily intake that would result in no change in TBSs or TLRs). Outliers were excluded from analysis but are plotted for reference (open symbols). RAE, retinol activity equivalents; TBS, total body vitamin A store; TLR, total liver vitamin A reserve; VA, vitamin A.