Indicators of vitamin A status were evaluated in nonpregnant breast-feeding (n = 265) and nonpregnant non-breast-feeding (n = 49) Indonesian women. The concentration of vitamin A (not including provitamin A carotenoids) and fat in breast milk was 30% and 20% higher, respectively, for women with a breast-fed child 7-18 mo old than for women with an infant 3-6 mo old. The vitamin A content of milk fat was constant throughout lactation. Breast-milk vitamin A was most sensitive to changes in vitamin A status when expressed per volume. Sensitivity and specificity for detecting serum retinol concentrations < 0.70 mumol/L were < 75% for the concentration of breast-milk vitamin A and serum retinol binding protein (RBP). The modified-relative-dose-response (MRDR) method suffers from a relatively large intraindividual variation in the ratio of dehydroretinol to retinol because of vulnerability of the dehydroretinol concentration to laboratory errors and to variation in dosing and absorption. Within categories of dehydroretinol:retinol, serum retinol concentration was lower in breast-feeding women than in non-breast-feeding women. Thus, it may be necessary to use different cutoff values for the ratio and for serum retinol concentration. Serum retinol concentration, which was just above marginal (0.85 mumol/L), had the smallest within-person variation and was also the most sensitive indicator for detecting a difference between groups in change in vitamin A status postintervention, requiring only 19 subjects per group. Serum RBP concentration, breast-milk vitamin A expressed per volume or per gram milk fat, and the MRDR method required groups of 35, 36, 139, and 53 subjects, respectively.
PIP: Although serum retinol concentration is the most widely used biochemical method for assessing vitamin A status, it is not sensitive to changes over a wide range of vitamin A status and is temporarily lowered during infections. Alternative indicators of vitamin A status--breast milk vitamin A concentration, serum retinol binding protein (RBP) concentration, and the modified relative dose response (MRDR) method--were evaluated in a cross-sectional and intervention study of 265 breast feeding women and 49 nonpregnant, non-breast-feeding women from West Java, Indonesia. Concentrations of vitamin A and fat in breast milk were 30% and 20% higher, respectively, for women with a breast-fed child 7-18 months of age than for mothers of a 3-6 month old infant. Sensitivity and specificity for detecting serum retinol concentrations under 0.70 mcmol/l were less than 75% for the concentration of breast milk vitamin A and serum RBP. The MRDR method yielded large intra-individual variation in the ratio of dehydroretinol to retinol because of vulnerability to laboratory errors and variation in dosing and absorption. Within categories of this ratio, serum retinol concentrations were 20-25% lower in breast-feeding women. This finding suggests a need for different cutoff values for the ratio and for serum retinol concentration. Serum retinol concentration had the smallest within-person variation and was the most sensitive indicator for detecting a difference between groups in change in vitamin A status post-nutritional intervention. However, more data are required on breast milk composition and its changes throughout lactation before conclusions can be drawn about how to use breast milk vitamin A as an indicator of vitamin A status.