RESUMO
BACKGROUND: Vitamin A (VA) deficiency (VAD) affects â¼19 million pregnant women worldwide. The extent of VAD in Zambian women of reproductive age is unknown owing to lack of survey inclusion or the use of static serum retinol concentrations, a low-sensitivity biomarker. OBJECTIVES: This cross-sectional study employed isotopic techniques to determine VA status with serum and milk among women aged 18-49 y (n = 197) either lactating with infants aged 0-24 mo or nonlactating with or without infants. METHODS: Assistants were trained and piloted data collection. Demographic data, anthropometry, and relevant histories were obtained including malaria and anemia. For retinol isotope dilution (RID), baseline fasting blood and casual breast milk samples were collected before administration of 2.0 µmol 13C2-retinyl acetate and 24-h dietary recalls. On day 14, blood (n = 144) and milk (n = 66) were collected. Prevalence of total liver VA reserves (TLR) ≤0.10 µmol/g was defined as VAD with comparison to the DRI assumption of 0.07 µmol/g as minimally acceptable for North Americans. RESULTS: When a 20% adjustment for dose lost to milk was made in the RID equation for lactation, mean total body VA stores (TBS) for lactating women were 25% lower than for nonlactating women (P < 0.01), which was not the case without adjustment (P = 0.3). Mean ± SD TLR for all women were 0.15 ± 0.11 µmol/g liver. Using retinol purified from breast milk instead of serum for RID analysis yielded similar TBS and TLR, which were highly correlated between methods (P < 0.0001). Serum retinol ≤0.70 µmol/L had 0% sensitivity using either VAD liver cutoff and milk retinol ≤1.0 µmol/L had 42% sensitivity for VAD at 0.10 µmol/g. CONCLUSIONS: Determining accurate VA status among women of reproductive age, especially lactating women, forms a basis for extrapolation to the general population and informing policy development and program implementation.
