点击显示 收起
1 From the Division of Human Nutrition and Epidemiology, Wageningen University, Wageningen, Netherlands (MAD, FTW, and CEW); the Nutrition Research and Development Centre, Bogor, Indonesia (MAD, FTW, and M); the Departments of Internal Medicine (MAD and FTW) and Gastroenterology and Hepatology (CEW), University Medical Centre Nijmegen, Nijmegen, Netherlands
2 CE West is deceased. 3 Supported by the Dutch Foundation for the Advancement of Tropical Research (WOTRO) and Ter Meulen Fund (Royal Netherlands Academy of Arts and Sciences). 4 Address reprint requests to F Wieringa, Department of Internal Medicine, University Medical Centre Nijmegen, Bilderdijkkade AB, t/o 32A, 1053 AZ Amsterdam, Netherlands. E-mail: wieringa{at}tiscali.nl.
ABSTRACT
Background: Deficiencies of vitamin A, iron, and zinc are prevalent in women and infants in developing countries. Supplementation during pregnancy can benefit mother and infant.
Objective: We examined whether supplementation during pregnancy with iron and folic acid plus ß-carotene or zinc or both improves the micronutrient status of mothers and infants postpartum.
Design: Pregnant women (n = 170) were supplemented daily only during pregnancy with ß-carotene (4.5 mg), zinc (30 mg), or both or placebo plus iron (30 mg) and folic acid (0.4 mg) in a randomized, double-blind, placebo-controlled trial. Micronutrient status was assessed 1 and 6 mo postpartum.
Results: Six months postpartum, plasma retinol concentrations were higher in the women who received zinc during pregnancy than in women who did not. Infants born to mothers supplemented with ß-carotene + zinc had higher plasma retinol concentrations, with the frequency of vitamin A deficiency reduced by >30% compared with the other 3 groups. Breast-milk ß-carotene concentrations were higher in all women supplemented with ß-carotene, but breast-milk retinol concentrations were higher only in women who received ß-carotene + zinc. Zinc concentrations did not differ among groups in mothers and infants.
Conclusions: Zinc supplementation during pregnancy improved the vitamin A status of mothers and infants postpartum, which indicates a specific role of zinc in vitamin A metabolism. Addition of both ß-carotene and zinc to iron supplements during pregnancy could be effective in improving the vitamin A status of mothers and infants.
Key Words: Infants • breast milk • vitamin A deficiency • zinc • pregnancy • iron • growth • ß-carotene • folic acid • Indonesia
INTRODUCTION
Deficiencies of micronutrients during infancy can lead not only to poor growth but also to increased risk of morbidity and mortality from infectious diseases and to delayed psychomotor development (1-4). Maternal nutritional status is one of the most important factors determining the nutritional status of infants, especially with respect to vitamin A (5). Nutritional status during pregnancy determines to a large extent the nutritional stores with which infants are born, and nutritional status postpartum also affects breast-milk micronutrient content. In many developing countries, pregnant and lactating women and their infants are often deficient in various micronutrients. Supplementing women during pregnancy with micronutrients has the advantage of benefiting mothers and infants simultaneously.
Vitamin A deficiency is still a serious health problem worldwide, and supplementation with vitamin A was shown to reduce mortality in children aged <5 y by an estimated 23% (6). Several approaches to reduce vitamin A deficiency were implemented, including intermittent high-dose vitamin A supplementation of infants and children and providing a single high-dose of vitamin A to the mother directly postpartum. Because of the possible teratogenicity of vitamin A in humans (7), the World Health Organization recommends that only small doses (<10 000 IU/d or <25 000 IU/wk) should be given to mothers during pregnancy (8). Similarly, only small doses (3 doses of 50 000 IU with intervals of at least 1 mo) of vitamin A can be safely given to young infants (9, 10). ß-Carotene is a precursor of vitamin A and is not teratogenic. Therefore, it is considered a useful alternative to retinol for supplementation during pregnancy. ß-Carotene is also the main dietary source of vitamin A in developing countries, but both the uptake and the conversion of dietary ß-carotene to retinol are less efficient than previously thought (11).
Iron deficiency is the most prevalent micronutrient deficiency, and more than 50% of the pregnant women in developing countries are estimated to be iron deficient (12). Supplementation of iron, in combination with folic acid, during pregnancy was implemented as a standard program in many countries. High intake of phytate, which leads to iron deficiency, also leads to zinc deficiency, suggesting that zinc deficiency is also prevalent.
If there is deficiency of more than one micronutrient, supplementation with only one micronutrient will not adequately address all needs. Moreover, the benefit of supplementing with one micronutrient can be compromised by the presence of deficiency of other micronutrients, thus reducing effectiveness of the supplementation. For example, supplementation with vitamin A in addition to iron reduces the prevalence of anemia more than does iron supplementation alone (13). This study investigates whether supplementing women during pregnancy with ß-carotene and zinc, in addition to the standard supplementation with iron and folic acid, can improve vitamin A and zinc status of mothers and newborns 1 and 6 mo postpartum.
SUBJECTS AND METHODS
Study design and location
The study was designed as a double-blind, placebo-controlled study of mothers and their infants. During pregnancy, all mothers enrolled in a randomized, double-blind, controlled trial in which iron and folic acid were supplemented together with ß-carotene or zinc or both. On the basis of the factorial design, a required sample size of 47 subjects per group was calculated to detect a difference in plasma retinol concentrations (means ± SDs) of 0.20 ± 0.3 µmol/L, with a confidence level of 95% and a power of 0.9. To allow for 20% dropout, 230 subjects were recruited.
All women were recruited before 20 wk gestational age from 13 adjacent villages in a rural area in Bogor District, West Java, Indonesia. Age of the women was 25.1 ± 5.6 y, with a median parity of 2 [interquartile range (IQR): 1–4]. Each woman was supplemented daily during pregnancy until delivery. All women received iron and folic acid (30 mg iron as ferrous fumarate/d and 0.4 mg pteroylglutamic acid/d). In addition, one group of women received ß-carotene (4.5 mg as water-soluble granulate/d; ß-carotene group), one group received zinc (30 mg zinc as sulfate/d; zinc group), one group received ß-carotene plus zinc (4.5 mg ß-carotene and 30 mg zinc/d; ß-carotene + zinc group), and one group received only iron and folic acid (control group). Capsules were prepared containing the micronutrients specific for each group and were indistinguishable from each other. They were prepared by the pharmacy of the Gelderse Vallei Hospital (Ede, Netherlands) and given a letter code. The code allocation was not known to the researchers until all observations and biochemical analyses were completed. Each month, each subject received a bottle labeled with her name, containing 40 capsules. At the end of each month, the remaining capsules were counted as a measure of compliance. Compliance, expressed as a proportion of the intended supplements consumed during pregnancy, did not differ among the groups with a mean compliance of > 80% in all groups, and 90% of the women taking >50% of the intended dose. The effects of supplementation on pregnancy outcome will be published separately.
Subjects and procedures
Of the 230 women initially recruited, 179 remained in the study until delivery. Infants were born between November 1998 and July 1999. Exclusion criteria at enrollment were twin pregnancy and congenital abnormalities that interfered with growth, development, or metabolism. Nine women were not eligible for follow-up because of twin pregnancy (3), still birth (1), or neonatal death (5). Thus, 170 women were enrolled for follow-up of infant and mother until 6 mo postpartum (Figure 1). The protocol was approved by the Ethical Committees of the National Health Research and Development Institute of Indonesia and of the Royal Netherlands Academy of Arts and Sciences. The women were informed at the beginning of supplementation, and written informed consent was given. Mothers and infants were assessed anthropometrically each month, starting in the first month after delivery. Two breast-milk samples were collected from each mother; one in the first and one in the sixth month postpartum. Furthermore, a blood sample (5 mL) was collected at the end of the study of both the mother and the infant for assessment of micronutrient status.
FIGURE 1.. Trial profile.
Measurements
Anthropometry included measurement of weight and length by trained anthropometrists with use of standard methods. Nonfasting blood samples were collected between 0900 and 1100 by using a closed-tube vacuum system (Becton Dickinson, Leiden, Netherlands). Blood samples were immediately stored at 4 °C and separated within 5 h. Plasma and serum samples were divided into aliquots and stored at –30 °C until analysis. Breast-milk samples were obtained with use of standardized collection procedures (14). Breast milk was collected from the right breast 45–60 min after the last feeding from that breast. The breast was completely expressed with the use of a manual breast-milk pump. All milk was collected in one container and thoroughly mixed before dividing into aliquots, analyzed, and frozen at –30 °C.
Hemoglobin concentrations were measured by a standard cyanoblue method (Humalyzer, Tanusstein, Germany). Serum zinc concentrations were analyzed with flame atomic absorption spectrophotometry (Varian, Clayton South, Victoria, Australia) with use of trace element–free procedures (15). Plasma retinol and ß-carotene concentrations were measured by HPLC, with 2 separate extractions and separations as described by Ehrhardt et al (16). The CV for zinc, retinol, and carotenoid analyses (10% duplicate analysis and pooled control samples) was <5%. Vitamin A deficiency was defined as plasma retinol concentration < 0.70 µmol/L (3). Breast-milk retinol and carotenoid concentrations were measured as described by Jackson et al (17). In short, 200 µL breast milk was mixed with KOH-ethanol (12.5% wt:vol, 250 µL), and internal standard (9 µmol/L echinenone in ethyl alcohol, 100 µL) was added. The mixture was incubated at 45 °C for 2 h. Retinol and carotenoids were extracted twice with hexane, dried under nitrogen, and reconstituted in solvent for HPLC analysis. Creamatocrit was measured to estimate breast-milk fat content with use of a method analogous to hematocrit measurement, and breast-milk fat content was calculated according to Lucas et al (18). The CV for the measurement of the fat content and concentrations of retinol and ß-carotene in breast milk was <10%. Retinol and ß-carotene concentrations in breast milk were expressed in terms of fat to control for variations in the fat composition of breast milk. Ferritin was measured with use of a commercial enzyme-linked immunosorbent assay kit (IBL, Hamburg, Germany) according to the guidelines of the manufacturer. To control for the effects of the acute-phase response on indicators of micronutrient status, the acute-phase protein 1-acid glycoprotein (AGP) was measured with use of immunoturbidimetric techniques at the Northern Ireland Centre for Diet and Health, University of Ulster, Coleraine, Northern Ireland (Cobas Fara analyzer; Roche Products, Welwyn, United Kingdom). The CV for the ferritin and AGP assays was <10%.
Statistical analysis
Data were checked for normal distribution with use of the Kolmogorov-Smirnov test. To achieve normality, plasma concentrations of ferritin and ß-carotene and breast-milk concentrations of retinol, ß-carotene, and zinc were transformed to natural logarithms before statistical analyses. Differences in baseline characteristics among groups were tested with Pearson chi-square test or with analysis of variance (ANOVA). Correlations between birth weight and anthropometry at 6 mo of age were analyzed in a multiple linear regression model, controlling for sex. The effect of supplementation was analyzed with multivariate analysis (2-way ANOVA) for main effects and interactions of both zinc and ß-carotene supplementation. Significant interactions (P < 0.1) were further examined in group-by-group analysis with ANOVA or analysis of covariance (ANCOVA), with correction for the acute-phase response, using AGP concentrations. Differences in concentrations of hemoglobin and plasma ß-carotene were analyzed with ANOVA, and differences in plasma or serum concentrations of ferritin, retinol, and zinc were tested with ANCOVA with the acute-phase protein AGP as covariate to control for the effects of the acute-phase response on these indicators (19).
Furthermore, body mass index (BMI) of the mothers at recruitment (<20 wk gestational duration) as a general indicator of nutritional status at baseline was included as covariate in the analyses concerning the effect of supplementation on indicators of micronutrient status in the mothers 6 mo postpartum. Differences in breast-milk concentrations of retinol, ß-carotene, zinc, and fat among supplementation groups at 1 mo and 6 mo were tested with ANOVA. When the overall F test was significant (P < 0.05), groups were compared with the control group by using Tukey post hoc multicomparison test for ANOVA or with Bonferroni correction for ANCOVA. Correlations between micronutrient status of mothers and infants were investigated with multiple linear regression, controlling for AGP concentrations of mother and infant for retinol, ferritin, and zinc concentrations, as described earlier. Changes in breast-milk concentrations of fat, retinol, ß-carotene, and zinc from the first month to the sixth month of lactation were tested with use of independent t test, with all groups pooled. Cumulative frequency distribution curves were smoothed with use of a moving average, taking the 2 previous and 2 following values into account (GraphPad Prism 2.00, San Diego). Differences in prevalence in micronutrient deficiency among groups were investigated with logistic regression analysis, controlling for AGP in infants and for AGP and BMI at recruitment in mothers. Anthropometric z scores were calculated with EPI-Info, Version 6.04b (Centers for Disease Control and Prevention, Atlanta), and statistical analysis was carried out with the SPSS 10.0.1 (SPSS Inc, Chicago) software package.
RESULTS
During follow-up, 34 mother–infant pairs dropped out for various reasons (Figure 1). No difference was observed among the groups in baseline characteristics concerning socioeconomic status, village distribution, anthropometry, and gestational duration of the pregnant women at recruitment or in breastfeeding practices and complementary feeding patterns. The number of infants in all groups was similar (Table 1). There were 101 boy and only 69 girl infants, but both sexes were equally distributed over the groups. Seven infants were born preterm, of which 3 were in the control group, 2 in the zinc group, and 1 each in the ß-carotene group and the ß-carotene + zinc group. The birth weights of boy and girl infants combined did not differ among the groups (
View this table:
TABLE 1. General characteristics of the infants and of their mothers supplemented during pregnancy with ß-carotene alone, ß-carotene plus zinc, or zinc alone in addition to iron and folic acid1
Neither zinc nor ß-carotene supplementation significantly affected hemoglobin or ferritin concentrations in the women or their infants 6 mo postpartum (Table 2). Plasma retinol concentrations 6 mo postpartum were significantly higher in the women who received zinc during pregnancy than in the women who did not receive zinc (P < 0.01, ANCOVA controlling for BMI at recruitment and AGP; Table 2). Concomitantly, a tendency was observed toward higher plasma retinol concentrations 6 mo postpartum in the women who received ß-carotene during pregnancy than in the women who did not receive ß-carotene (P < 0.1, ANCOVA controlling for BMI at recruitment and AGP; Table 2). However, a significant interaction (P = 0.09) was observed between zinc and ß-carotene supplementation on plasma retinol concentrations in the mothers 6 mo postpartum.
View this table:
TABLE 2. Indicators of the micronutrient status of mothers and their infants 6 mo postpartum and main effects of ß-carotene and zinc supplementation during pregnancy
Similarly, both zinc and ß-carotene supplementation during pregnancy resulted in higher retinol concentrations in the infants 6 mo postpartum (P < 0.01 and P < 0.05, respectively, ANCOVA controlling for AGP; Table 2). However, a strong interaction was also observed between zinc and ß-carotene supplementation on plasma retinol concentrations in infants (P < 0.01).
Plasma concentrations of ß-carotene 6 mo postpartum were significantly related to ß-carotene supplementation during pregnancy in the mothers (P < 0.01, ANOVA) but not in the infants. Also, a significant interaction was observed between ß-carotene and zinc supplementation (P = 0.09) on plasma ß-carotene concentrations 6 mo postpartum in the mothers. Neither supplementation with zinc nor ß-carotene during pregnancy affected serum zinc concentrations in the mothers or the infants 6 mo postpartum.
Interactions were further explored in group-by-group analysis, but only for interactions with P < 0.1. Both plasma retinol and ß-carotene concentrations were significantly higher in the women who were supplemented with ß-carotene + zinc during pregnancy than the women in the control group (P < 0.01 for both, ANCOVA with Bonferroni correction; Table 2). The plasma retinol concentrations of infants of mothers supplemented during pregnancy with ß-carotene + zinc were significantly higher than those of infants from all other groups (P < 0.01, ANCOVA controlling for AGP, with Bonferroni correction).
The effect of supplementation during pregnancy on plasma retinol concentrations of mothers and infants 6 mo postpartum is clearly illustrated in the cumulative frequency distributions (Figures 2 and 3). The distribution curve of plasma retinol concentrations 6 mo postpartum of the mothers in the ß-carotene + zinc group is shifted to significantly higher concentrations than for the control and ß-carotene groups (P < 0.01, ANCOVA controlling for AGP and BMI at recruitment), with the lowest frequency of plasma retinol concentrations < 0.70 µmol/L (Figure 2). The distribution curves of women who received ß-carotene only or zinc only are not different from the control group. The distribution curve of plasma retinol concentrations of the infants born from mothers receiving ß-carotene + zinc during pregnancy is also clearly shifted toward higher retinol concentrations compared with all other groups (P < 0.01, ANCOVA controlling for AGP), whereas the distribution curves of the infants in the ß-carotene group and in the zinc group are almost similar to the control group (Figure 3).
FIGURE 2.. Cumulative frequency of plasma retinol concentrations 6 mo postpartum in women supplemented during pregnancy with ß-carotene or zinc or both in addition to iron and folic acid. Plasma retinol concentrations in the ß-carotene + zinc group were significantly higher than those in the control and ß-carotene groups (P < 0.01, analysis of covariance with control for BMI at recruitment and 1-acid glycoprotein with Bonferroni adjustment). The reference line indicates the cutoff for vitamin A deficiency (0.70 µmol/L). The prevalence of vitamin A deficiency was 9% in the ß-carotene + zinc group, 21% in the control group, 27% in the ß-carotene group, and 32% in the zinc group.
FIGURE 3.. Cumulative frequency of plasma retinol concentrations in infants aged 6 mo born to mothers supplemented during pregnancy with ß-carotene or zinc or both in addition to iron and folic acid. Plasma retinol concentrations in the ß-carotene + zinc group were significantly higher than those in all other groups (P < 0.01, analysis of covariance with control for 1-acid glycoprotein with Bonferroni adjustment). The reference line indicates the cutoff for vitamin A deficiency (0.70 µmol/L). The prevalence of vitamin A deficiency was 41% in the ß-carotene + zinc group, 73% in the control group, 73% in the ß-carotene group, and 74% in the zinc group.
Analysis of the prevalence of vitamin A deficiency 6 mo postpartum with logistic regression showed a significant interaction between ß-carotene and zinc supplementation on the prevalence of vitamin A deficiency for both mothers and infants (P < 0.05 for both). Infants born from mothers receiving ß-carotene and zinc during pregnancy had an odds ratio of 0.22 (95% CI: 0.07, 0.68; P < 0.01, logistic regression controlling for AGP) to be vitamin A deficient at 6 mo of age compared with the control group. This finding is also reflected in a >30% lower prevalence of vitamin A deficiency in the infants in the ß-carotene and zinc group than in the other groups (Figure 3). In mothers, the overall prevalence of vitamin A deficiency was lower; hence, the effect of supplementation on reducing the prevalence of vitamin A deficiency was less strong and failed to reach significance, although mothers who received ß-carotene and zinc during pregnancy had an odds ratio of 0.23 (95% CI: 0.04, 1.22; P = 0.08, logistic regression controlling for AGP and BMI at recruitment) of being vitamin A deficient 6 mo postpartum compared with the control group. Plasma concentrations of retinol and ß-carotene of the mothers were significantly correlated with those of their infants (R = 0.23, P < 0.01 multiple linear regression controlling for AGP concentrations of mothers and infants, and r = 0.45, P < 0.001 Pearson correlation, respectively; Table 2).
Breast-milk concentrations of retinol and zinc decreased significantly between the first and sixth month of lactation in all groups combined (P < 0.001, independent t test; Table 3). In the first month of lactation, the micronutrient concentrations and fat content of breast milk were not significantly affected by supplementation with either zinc or ß-carotene during pregnancy (ANOVA). As breast milk was collected from the women at some time during the first few weeks of lactation, women were possibly at various stages of milk production, resulting in a physiologic variation in the composition of the milk produced. This effect might have obscured differences in micronutrient concentrations of breast milk among the different supplementation groups in the samples taken during the first month of lactation. Supplementation during pregnancy with ß-carotene or zinc tended to increase breast-milk concentrations of retinol at 6 mo postpartum (P = 0.06 and P = 0.07 respectively, ANOVA). ß-Carotene supplementation during pregnancy resulted in higher concentrations of ß-carotene in breast milk 6 mo postpartum (P < 0.001, ANOVA). The zinc content of breast milk at 6 mo postpartum was not affected by either zinc or ß-carotene supplementation. Analysis per supplementation group showed that at 6-mo postpartum retinol concentrations in breast milk were significantly higher in the women supplemented during pregnancy with ß-carotene + zinc than in the women of the control group (P < 0.05, ANOVA with Tukey post hoc test; Table 3). Breast-milk ß-carotene concentrations 6 mo postpartum were significantly higher in both the women who were supplemented during pregnancy with ß-carotene + zinc as well as with ß-carotene alone than for women of the control group (P < 0.01, ANOVA with Tukey post hoc test; Table 3). Breast-milk lutein concentrations were not significantly different among the groups at 1 and 6 mo postpartum, with median (IQR) concentrations of 56.5 (25.2–107.6) nmol/g fat and 51.7 (31.7–82.8) nmol/g fat, respectively, indicating no differences in dietary intake of vegetables (20).
View this table:
TABLE 3. Micronutrient concentrations in breast-milk samples at 1 and 6 mo postpartum
DISCUSSION
This study shows that supplementation with ß-carotene is only effective in improving vitamin A status when given in combination with zinc. This effect of zinc on the conversion of ß-carotene to retinol has not been reported previously. Plasma retinol concentrations 6 mo postpartum were higher after supplementation during pregnancy with both ß-carotene and zinc in mothers as well as their infants, and the prevalence of vitamin A deficiency in infants was reduced by >30%. Intriguingly, ß-carotene supplementation alone, without zinc, did not significantly increase plasma retinol concentrations in either mothers or infants. Furthermore, retinol concentrations in breast milk 6 mo postpartum were higher for the women supplemented during pregnancy with ß-carotene and zinc but not for the women who received only ß-carotene.
These findings point to a specific role of zinc in the conversion of ß-carotene to retinol, distinct from the role of zinc in the mobilization of vitamin A as reported by others who showed that zinc deficiency can reduce plasma retinol binding protein (RBP) concentrations (21, 22). In zinc-deficient pregnant rhesus monkeys, plasma vitamin A and RBP concentrations were strongly correlated to plasma zinc concentrations but only below a certain threshold of plasma zinc concentrations (21). However, results were not equivocal and even conflicting (23). Differences in the degree of zinc and vitamin A deficiency and differences between animal models and humans further hamper interpretation.
In the present study, zinc supplementation could have improved the effectiveness of ß-carotene supplementation by enhancing the uptake of ß-carotene from the intestinal lumen, for instance by affecting the excretion of pancreatic enzymes or the composition of bile acids. Furthermore, zinc can affect transport, storage, and subsequent mobilization of ß-carotene and retinol by way of effects on chylomicron and lipoprotein metabolism and on RBP expression and synthesis. Much is still unclear about these processes and the role of zinc therein. However, if the main effect of zinc was improved uptake and transport of ß-carotene and retinol, improvement of vitamin A status is also expected in the zinc group. This improvement is expected especially because ß-carotene was the main dietary source of vitamin A, and deficiencies of vitamin A and zinc in this population were only marginal. Yet, there was no improvement of vitamin A status in the women who received only zinc.
Enhanced activity of the enzyme 15-15-dioxygenase, which splits ß-carotene centrally to retinol, could be the most likely mechanism to explain the findings. Activity of this enzyme is stimulated by iron (24), but it is possible that activity could also be stimulated by zinc. The primary site of enzyme activity is the intestinal mucosa, although it is expressed in many other tissues, including the liver. It was suggested that especially exogenous ß-carotene is primarily converted to retinol and not ß-carotene in endogenous pools (25, 26). However, recently Tang et al (27) showed that over an 8-wk period, 19% of the total conversion of ß-carotene to retinol after a test dose was by extraintestinal cleavage. Thus, endogenous conversion of ß-carotene to retinol, eg, in the liver, could well play a more important role in the long-term effects of ß-carotene supplementation than previously presumed.
In the present study, ß-carotene concentrations in breast milk were higher after supplementation with ß-carotene during pregnancy, with or without zinc. This finding implies that ß-carotene uptake, transport, and storage were not impaired in the women who did not receive zinc in addition to ß-carotene. Breast-milk retinol concentrations, however, are only higher after supplementation with zinc in addition to ß-carotene, probably directly reflecting improved vitamin A status for these women. In contrast to breast milk, ß-carotene concentrations in plasma are only higher after supplementation with zinc in addition to ß-carotene, suggesting that mobilization of ß-carotene to breast milk involves a different mechanism than to plasma. Apparently, mobilization of ß-carotene to plasma is affected by zinc. It is less likely that the higher ß-carotene concentrations in plasma reflect improved intestinal absorption of ß-carotene because of zinc, as the effect is not seen in the women receiving only zinc supplementation. In conclusion, the higher plasma retinol concentrations of the women who also received zinc in addition to ß-carotene are most likely due to an increased conversion of ß-carotene to retinol, although it is not clear whether this took place during the supplementation period itself (during pregnancy) or to what extent later conversion of endogenous ß-carotene to retinol contributed.
Increased neonatal vitamin A stores at birth could have made an important contribution to the higher vitamin A status in infants born from mothers who received zinc in addition to ß-carotene during pregnancy. Vitamin A stores in fetal liver accumulate during the last trimester of pregnancy, but stores are related to maternal plasma retinol concentrations (28). Colostrum and early milk contain high concentrations of retinol and ß-carotene, but concentrations fall after a few weeks of lactation. In the present study breast-milk concentrations of retinol and ß-carotene were higher at 6 mo of lactation after supplementation with ß-carotene and zinc during pregnancy. The contribution of breast-milk retinol concentrations toward maintaining vitamin A nutriture can be estimated. Assuming an average daily milk intake of 600 mL, and an average fat content of breast milk of 31 g/L, the median daily retinol intake from breast milk of the infants in the ß-carotene + zinc group 6 mo postpartum would be 216 RE (IQR: 133–303), whereas the median daily intake of retinol of the infants in the control group would be only 148 RE (IQR: 100–194). For comparison, the UK-estimated average requirement and lower reference nutrient intake for retinol in this age group are 250 RE and 150 RE, respectively (29). The contribution of breast-milk ß-carotene to the vitamin A status of the infants is small and estimated at only 1.4 RE (IQR: 1.1–2.3) in the ß-carotene and zinc group, assuming a conversion factor of ß-carotene-to-retinol of 1:6. Also, infants in the ß-carotene group do not have higher plasma retinol concentrations, although breast-milk ß-carotene concentrations were also high. Thus, an increase in breast-milk ß-carotene concentrations alone will not contribute substantially to improving vitamin A status in infants. Comparison of results of different supplementation strategies used to improve vitamin A status of young infants, including maternal high-dose vitamin A supplementation directly postpartum and ß-carotene supplementation of either the mother during lactation or the infant, suggests that supplementation during pregnancy is more effective than supplementation of either mother or infant postpartum (30, 31).
It is estimated that 2.6 µg ß-carotene in oil can be converted to 1 µg retinol (32). However, the bioavailability of ß-carotene from fruit and vegetables is much lower than previously thought. The US Institute of Medicine revised the conversion factor for a mixed diet from 6 µg to 12 µg ß-carotene to supply 1 µg retinol (33). However, estimates of the bioconversion of ß-carotene to retinol from a mixed diet have varied widely, ranging from 6 µg in European subjects (34) to 26 µg in some developing countries such as Indonesia (11, 35). Differences in zinc status might, in part, explain this wide range of estimates, and the low rate of conversion observed in developing countries could be due to zinc deficiency. Therefore, impaired bioconversion of provitamin A carotenoids to retinol because of underlying zinc deficiency could explain the paradox encountered in many developing countries of a high prevalence of vitamin A deficiency in the presence of apparently abundant dietary carotenoids. Further- more, concurrent zinc deficiency could reduce the effectiveness of ß-carotene supplementation. For instance, in Nepal, supplementation during pregnancy with ß-carotene was much less effective than with vitamin A in improving vitamin A status in infants (36), possibly because of underlying zinc deficiency.
Supplementation during pregnancy with zinc, either alone or combined with ß-carotene, had no effect on indicators of zinc status 6 mo postpartum. Plasma zinc concentration is a relative insensitive indicator for zinc status, because plasma zinc concentrations are homeostatically controlled and can remain in the normal range in marginal zinc deficiency. Furthermore, breast-milk zinc content is relatively independent of maternal zinc status. A physiologic response to zinc supplementation, eg, improved growth in children, is, therefore, considered a more reliable indicator of preexisting marginal zinc deficiency. The improvement in vitamin A status after supplementation of ß-carotene and zinc could be regarded as a physiologic response, indicating preexisting marginal zinc deficiency in the study population, with impaired conversion of ß-carotene to retinol because of zinc deficiency. This finding suggests preexistent zinc deficiency in these subjects, which is supported by the finding that mean plasma zinc concentrations were low in both mothers and infants 6 mo postpartum, with 29% of the mothers and 17% of the infants below the cutoff of marginal zinc deficiency (10.7 µmol/L).
The findings of this study clearly show that ß-carotene can only improve vitamin A status when zinc is also available. Supplementation during pregnancy with ß-carotene and zinc, but not with ß-carotene alone, is effective in improving the vitamin A status of both mothers and infants 6 mo postpartum. Increased breast-milk vitamin A content is an important contributor to the improved vitamin A status of the infants. In view of the high prevalence of vitamin A deficiency in infants and lactating mothers in developing countries and the direct relation between micronutrient deficiencies and increased morbidity and mortality, supplementing pregnant women with ß-carotene and zinc is clearly indicated. Because iron supplementation of pregnant women is already widely implemented, the addition of ß-carotene and zinc to these supplements could be an efficient and cost-effective measure for improving the health of both infants and lactating mothers.
ACKNOWLEDGMENTS
We thank all the mothers and the infants, the health volunteers, and the midwives who participated in this study, and we are grateful for the enthusiastic support we received from Hendra, Anni, and their staff. Furthermore, we thank the field team of NRDC (Bogor, Indonesia); D Dillon, Ibu Asih, and Ibu D’ari from the laboratory of SEAMEO, University of Indonesia, for their untiring efforts; DI Thurnham, CA Northrop-Clewes, J Coulter, and other staff from NICHE, University of Ulster, for their help with the analysis of retinol and ß-carotene in breast milk and AGP; and J Erhardt from Hohenheim University for advice on micromethods for the determination of retinol and carotenoids in plasma.
MAD and FTW were involved in the design, implementation, and analysis of the study and in the drafting of the manuscript. CEW was involved in the design and analysis of the study and in the drafting of the manuscript; Muhilal was involved in the design and implementation of the study. All authors contributed to the final version of the manuscript. There was no conflict of interest for any of the authors.
REFERENCES