点击显示 收起
1 From the Kintampo Health Research Centre, Health Research Unit, Ghana Health Service, Accra, Ghana (SKT, SN, and SO-A); the Department of Nutritional Sciences, University of Wisconsin, Madison, WI (SAT); the Department of Nutrition for Health and Development, World Health Organization, Geneva, Switzerland (BdB); and the Department of Biochemistry, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana (FKNA and AT)
2 Presented in part in abstract form (abstr W10) at the 2004 International Vitamin A Consultative Group meeting, Lima, Peru, 15-17 November 2004. 3 Supported by an International Atomic Energy Agency fellowship (SKT); the National Research Initiative of the USDA Cooperative State Research, Education and Extension Service, grant 2003-35200-13754 (SAT); and the World Health Organization (BdB). 4 Reprints not available. Address correspondence to SA Tanumihardjo, Department of Nutritional Sciences, 1415 Linden Drive, Madison, WI 53706. E-mail: sherry{at}nutrisci.wisc.edu.
ABSTRACT
Background: Vitamin A deficiency is an important public health problem in many developing countries. Women of childbearing age and children are documented as the most affected groups.
Objective: The objective was to determine the length of time mothers are protected postpartum against vitamin A depletion after receiving either 400 000 IU vitamin A in 2 divided doses or 200 000 IU as a single dose plus a placebo 24 h apart.
Design: Mothers (n = 168) were recruited by trained fieldworkers 7–10 d after delivery. Modified-relative-dose-response (MRDR) tests were performed at baseline in 167 women, and vitamin A was administered within 6 wk after delivery. The women were randomly assigned to 2 main treatment groups, and each treatment group was divided into 3 follow-up subgroups. Each subgroup was invited back once at month 1, 3, or 5 for a second MRDR test.
Results: The serum retinol concentration and the MRDR value were 1.4 ± 0.5 µmol/L and 0.048 ± 0.037, respectively, at baseline. A significant improvement in vitamin A status occurred after vitamin A treatment as assessed by the MRDR test (P < 0.0001). Serum retinol concentrations were not different after vitamin A treatment (P = 0.87).
Conclusions: The mothers had marginally depleted liver reserves of vitamin A at baseline on the basis of MRDR test results. Liver reserves of vitamin A significantly improved in both treatment groups, and the improvement was maintained for 5 mo.
Key Words: Modified relative dose response • vitamin A • 3,4-didehydroretinol • lactation • vitamin A status
INTRODUCTION
The vitamin A status of underprivileged women and children in much of the world is compromised for many reasons (1-4). For women, low dietary intakes and increased requirements during pregnancy and lactation cause vitamin A inadequacy. For children, inadequate dietary intake and multiple infections can result in depleted liver reserves of vitamin A, and an increased risk of morbidity is associated with a mild deficiency of vitamin A (5). During acute infection, retinol is excreted in significant amounts in the urine (6), which increases the total requirements of vitamin A. These factors are exacerbated by low socioeconomic status and a poor living environment. Because vitamin A status in developing countries is tenuous, high-dose supplements were introduced as an effective way to ward off the deleterious effects of vitamin A deficiency (7). Typically, the efficacy of these programs was evaluated by using serum retinol concentrations. However, serum retinol concentrations are only useful when liver reserves are severely depleted, and many groups have a marginal vitamin A status (8).
Assessing vitamin A status is important in determining which population groups are at risk of deficiency, depletion, and excess (9). Dietary surveys are useful for determining dietary diversity but inappropriate for determining vitamin A status because multiple factors (eg, the plant matrix and fat content of the meal) affect the absorption and use of provitamin A carotenoids (10, 11), which is the main source of vitamin A for much of the world. Biochemical assessment of vitamin A status is not straightforward. Serum concentrations of retinol are homeostatically controlled and do not decline until liver reserves are dangerously low (12). Also, in times of infection, serum concentrations are transiently reduced because of the acute-phase response (13). In healthy persons, the liver contains 80–90% of the total body reserves for vitamin A mostly in the form of retinyl esters. A liver reserve of 0.070 µmol retinol/g liver is defined as adequate for humans (4). The reported range of liver reserves in well-nourished healthy American adults is 0.44–0.74 µmol retinol/g liver (12). Because liver biopsies of humans are not practical, indirect methods to determine liver reserves are used.
Biochemical techniques to indirectly determine liver reserves of vitamin A, which are commonly used to assess vitamin A status in response to intervention trials aimed at improving the health of women and children, include the modified-relative-dose-response (MRDR) test (4, 8, 14) and the deuterated retinol dilution assay (9, 15-20). As liver reserves become depleted, apo-retinol binding protein accumulates in the liver. For the MRDR test, a challenge dose of 3,4-didehydroretinyl acetate is administered, and the response of dehydroretinol-holo-retinol binding protein complex is measured in the serum 5 h after dosing. An abnormal MRDR value is 0.060. The MRDR test is a categorical indicator of vitamin A status and lacks utility in defining the actual liver reserve of vitamin A (4). Nonetheless, the MRDR test offers more information about the population than serum retinol alone, is relatively easy to implement in field surveys (21), and has been used with several diverse groups of lactating women (1, 22-24). This study evaluated the efficacy of the newly proposed high-dose supplement regimen (2 x 200 000 IU vitamin A) in lactating women in Ghana by using the MRDR test over time.
SUBJECTS AND METHODS
Subjects
This was an individually randomized controlled trial of postpartum mothers (n = 168) selected from rural villages in Kintampo District of the Brong-Ahafo Region in Ghana. The district has a population of 140 000 with a relatively poor socioeconomic status. Mothers were recruited by trained fieldworkers 7–10 d after delivery. After baseline MRDR tests were successfully done for the mothers (n = 167) as described below, they were randomly assigned into 2 treatment groups and dosed with vitamin A within 6 wk postpartum. Computer-generated randomization was performed with the use of the weekly alternate cohort system. In brief, mothers who delivered in week 1 of the trial were put in the 200 000 IU vitamin A group, those delivering in week 2 were put in the 400 000 IU vitamin A group, and those delivering in each subsequent week were alternatively assigned to the 2 treatment groups until recruitment aims were met. Mothers in the 200 000 IU vitamin A group (control, n = 82) received a single oral dose of 200 000 IU vitamin A on their registration day and a placebo 24 h later. Mothers in the 400 000 IU vitamin A group (intervention, n = 85) received two 200 000 IU vitamin A oral doses, given on their registration day and 24 h later. Each of the treatment groups was randomly divided into 3 monthly follow-up groups with projected sample sizes of 25 women per subgroup for the MRDR test in either month 1, 3, or 5.
The Ghana Health Service ethics committee at Kintampo Health Centre approved the consent form and study protocol. The same consent form was re-read to the mothers at the second MRDR test.
MRDR test
The MRDR test involved giving a single oral dose of 8.8 µmol 3,4-didehydroretinyl acetate dissolved in 290 µL corn oil in the morning with the use of a 0.3-mL insulin syringe. This is the standard amount given to women. The mothers were dosed at their homes, and 5 h later a single finger prick blood sample (500 µL) was taken. The blood samples were stored on ice away from light in a cooler until transported to the laboratory. Clotted blood samples were centrifuged at 600 x g for 10 min at ambient temperature, and the serum samples were stored at –20 °C until shipped. After completion of the trial, samples were shipped frozen to the Vitamin A Assessment Laboratory at the University of Wisconsin, Madison. All samples arrived frozen and were immediately stored at –80 °C until analyzed. After 1 mo of intensive training, SKT measured the samples for 3,4-didehydroretinol and retinol with the use of a standardized method developed specifically for small serum volumes (25).
Extraction and HPLC procedures
Serum was thawed, and 200-µL aliquots were treated with 250 µL ethanol to denature proteins and extracted 3 times with 300 µL hexanes. All extractions were done under gold fluorescent lighting to minimize vitamin A destruction and isomerization. HPLC-purified retinyl acetate dissolved in ethanol was used as an internal standard to determine extraction efficiencies. The hexane layers were pooled and dried under argon. The samples were redissolved in 40 µL methanol:ethylene dichloride (3:1; by vol) to dissolve the lipid phase, of which 35 µL was injected into a 5-µm Waters Sunfire 15-cm, C18 reversed-phase column (Milford, MA). The HPLC system was run in isocratic mode [injector: Rheodyne (Cotati, CA); detector: Waters 2487 dual channel (Milford, MA); pump: Waters 600 Multisolvent Delivery System (Milford, MA); and data processor: Shimadzu C-R7A (Kyoto, Japan)]. The wavelength of detection was set at 350 nm to optimize 3,4-didehydroretinol detection. The flow rate was 1 mL/min and the mobile phase was 74:20:6 methanol:acetonitrile:water (by vol) with 0.05% triethylamine added as a modifier. External standards of 3,4-didehydroretinol and retinol were prepared by saponifying the respective acetate forms with alcoholic potassium hydroxide and purifying by HPLC. Standard curves were constructed to quantify the 3,4-didehydroretinol and retinol in the serum.
Statistical analysis
SAS STATISTICAL COMPUTER SOFTWARE (version 8.2; SAS Institute, Cary, NC) was used in the analysis of data. A 3-factor repeated-measures analysis of variance (ANOVA) was used to determine the effects of treatment group assignment, month of follow-up, and the pretest and posttest values for the serum MRDR values and retinol concentrations. P values were those generated by the type III tests of fixed effects for main effects and interactions. If the results of the ANOVA were significant, subgroup analysis was done with differences of least squares means. Fisher’s exact and chi-square tests were used to compute differences in proportions. P < 0.05 was considered significant.
RESULTS
Baseline characteristics of the postpartum mothers were similar (Table 1). The infants were within 6 wk of age at baseline measurements. The mothers were of low socioeconomic status as assessed by an assets list database kept by the Kintampo Health Centre. The baseline serum concentration of retinol and MRDR value were 1.4 ± 0.5 µmol/L and 0.048 ± 0.037, respectively. Of the 168 women originally enrolled in the study, complete data sets were available for 156 mothers (93%) for analysis. The goal was to maintain 25 women in each subgroup by the end of the trial, and only one subgroup (200 000 IU, month 1) fell short of this goal (Figure 1). This was due in part to dropouts (n = 3) and by the computer randomization by week of delivery. The number of women in each subgroup thus varied from 21 to 30. Twenty-two percent of the women (n = 36) at baseline were categorized as having a depleted vitamin A status with the use of the MRDR value cutoff of 0.060. This magnitude is considered a public health problem (26). A significant improvement in vitamin A status was measured by the MRDR test after vitamin A treatment (P < 0.0001) and within the treatment subgroups at all times (P value range <0.0001 to 0.038) with the use of least squares means differences (Figure 2). However, the MRDR values were not different by treatment group with the use of ANOVA (P = 0.73). After vitamin A supplementation, the grouped MRDR values as compared with baseline in these women were 0.026 ± 0.015, 0.031 ± 0.020, and 0.023 ± 0.012 for months 1, 3, and 5, respectively. These mean values indicate an adequate vitamin A status after treatment for the group as a whole, which is currently defined as a MRDR value 0.030.
View this table:
TABLE 1. Baseline characteristics collected from Ghanaian lactating mothers enrolled in a vitamin A supplementation trial
FIGURE 1.. Distribution of the follow-up testing times for women who received either 400 000 IU in 2 equally divided doses or 200 000 IU as a single dose plus a placebo 24 h apart after a baseline vitamin A assessment with the modified-relative-dose-response (MRDR) test within 6 wk postpartum. Ninety-three percent of the women returned for their follow-up assessment.
FIGURE 2.. Effect of 2 different vitamin A supplementation regimens [400 000 IU in 2 equally divided doses (n = 30, 25, and 25 for months 1, 3, and 5, respectively) or 200 000 IU as a single dose plus a placebo 24 h apart (n = 21, 27, and 28 for months 1, 3, and 5, respectively)] on vitamin A status as measured by the modified-relative-dose-response (MRDR) test. The posttreatment MRDR values were significantly different from baseline (B) (P < 0.0001; repeated-measures ANOVA with fixed effects), but no 2- or 3-factor interactions and no differences by treatment group (P = 0.73) or follow-up time (P = 0.51) were observed. Values are means ± SDs.
The proportion of women with MRDR values >0.060 (inadequate storage) decreased from baseline in all subgroups but was not different by level of treatment with the use of Fisher’s exact test (Table 2). Although the proportion of positive MRDR values was higher at 3 mo than at 5 mo, the same women were not tested at these time points, and the mean values were both different from baseline (Figure 2).
View this table:
TABLE 2. Modified-relative-dose-response (MRDR) values in Ghanaian women enrolled in a postpartum supplementation trial at baseline and during follow-up at 1, 3, and 5 mo after treatment with either 200 000 or 400 000 IU vitamin A1
Serum retinol concentration did not vary by group (P = 0.52) before and after treatment (P = 0.87) or by time of follow-up (P = 0.15), and no interactions were observed (P < 0.29). The mean value (1.44 ± 0.50 and 1.44 ± 0.47 µmol/L before and after vitamin A treatment, respectively) was well above that which is considered to indicate deficient (0.7 µmol/L) or mar-ginal (1.05 µmol/L) vitamin A status in women (27). A similar percentage of women had serum retinol concentrations <1.05 µmol/L before (21.8%) and after (19.2%) treatment, and these frequencies did not differ with the use of a chi-square test.
DISCUSSION
Supplemental vitamin A as retinyl ester (200 000 IU, 210 µmol) is routinely provided to children 12 mo and to postpartum women, and 400 000 IU (420 µmol) has been provided in limited situations to postpartum women (28, 29). When equivalent doses were administered to lactating sows as a single bolus, it was theorized (30) and then shown that the average nursing infant may not benefit from the higher dose (31). Moreover, detoxifying metabolites are increased in lactating sows given the higher dose (32). Thus, the current recommendation of the International Vitamin A Consultative Group to split the dose by 1 d seems appropriate (7). This trial evaluated the change in vitamin A status of Ghanaian lactating women who received either 200 000 or 400 000 IU supplemental vitamin A within 6 wk of delivery. This study served as an evaluation of a much larger World Health Organization multicenter clinical trial in Ghana and Tanzania which was run in parallel to the current study (unpublished data, 2004).
The postpartum mothers in this study had a marginal liver reserve of vitamin A at baseline as assessed by the MRDR test. Vitamin A status improved in both the intervention and the control groups and was maintained for 5 mo after dosing. Serum retinol concentrations did not change either within or between the treatment groups, further supporting the assessment of the women as having a marginal and not deficient vitamin A status at the outset. That is, serum retinol was still under homeostatic control and therefore did not change in response to treatment. We originally hypothesized that the 400 000 IU dose would be more protective than the 200 000 IU dose. However, in this group of women, both doses resulted in improvement of liver reserves for the duration of the study. These women were from a rural setting; therefore, generalizing to the entire population is not possible in this study.
The MRDR test is a categorical, qualitative indicator of vitamin A status. Therefore, persons and groups are either abnormal or normal. It is not a quantitative measure of actual liver reserves which can be estimated with the use of stable-isotope methods (4). Although we can conclude from this study that both dosing regimens improved the vitamin A status of the women for 5 mo, we cannot conclude that the 400 000 IU vitamin A dose offered longer protection in these women. Had the women been vitamin A deficient and not marginally depleted to begin with, this may have been the case. We have observed a dose-dependent increase in liver reserves of lactating sows after large vitamin A doses (30). Therefore, the 400 000 IU dose probably increased liver reserves more than the 200 000 IU dose, but the MRDR test was not able to show this because it indicated normal reserves of vitamin A at all posttreatment assessments for each group as a whole.
Both dose amounts were beneficial to the subjects, and no adverse effects were reported during the 5 mo. This 5-mo period is when the infant depends on the mother for its vitamin A from breast milk, and introduction of weaning foods to nursing infants starts at the latter part of the fourth month in this population group. In this area of Ghana, breastfeeding is universal with >98% initiating breastfeeding within 3 d of birth (33). Most infants are weaned in their fifth month and receive vitamin A from foods other than their mother’s breast milk. This study shows that the mothers were able to maintain normal liver reserves of vitamin A to the weaning period. By boosting reserves of vitamin A after birth, perhaps the mothers were more able to maintain their own vitamin A status during lactation, even while shunting newly ingested vitamin A to the mammary gland to meet the needs of the infant.
This group of women from rural areas in Ghana have a similar vitamin A status as that observed in Bangladeshi women (23) and a better status than that observed in 2 groups of Indonesian women (22, 24). The vitamin A statuses of lactating Indonesian women (n = 23) were determined at 3 monthly intervals (times 1, 2, and 3) during lactation and then again (time 4) after they had ingested vitamin A capsules (8.4 µmol, 8000 IU) daily for 35 d. After supplementation mean MRDR values and serum concentrations of retinol were 60% lower and 38% higher, respectively, than the measurements before supplementation (22). Even though the MRDR value is a categorical indicator, the fact that the Ghanaian baseline measure (0.048 ± 0.037) is half that of the Indonesian group before vitamin A supplementation (0.094 ± 0.01) coupled with an increase in serum retinol concentrations in the Indonesian group and no increase in the Ghanaian women supports the observation of differences between population groups. Therefore, determination of vitamin A status within different population groups is crucial for determining who will benefit the most from high-dose supplements.
The current recommendation of the International Vitamin A Consultative Group is 2 divided doses of 200 000 IU 1 d apart as close to birth as possible and not beyond 6 wk (7). This study shows that both dosing regimens (200 000 and 400 000 IU) were protective in rural Ghanaian women. This may not have been the case if the testing were applied in the urban areas of Ghana where dietary diversification is limited. We have observed that the mean MRDR value in a group of suburban Ghanaian women was 0.09 ± 0.05 (unpublished observation, 2005) even though their serum retinol concentration (1.5 ± 0.6 µmol/L) was identical to the rural population group studied here. This MRDR value is almost double the value observed in the current study. These women may have further benefited from the second dose of vitamin A to improve or prolong adequate vitamin A status. The argument that differences in MRDR values between groups are due to differences in serum retinol concentrations (34) does not fit these data. We observed the opposite finding (ie, significant differences in MRDR values with almost identical serum retinol concentrations in groups of women from the same country but different environments). Both were from low socioeconomic status, but one group was rural, whereas the other was urban with generally less purchasing power and resources.
Therefore, on the basis of the results of this study, we can conclude that these women from a rural environment benefited from treatment with vitamin A and that this benefit lasted for 5 mo. Future studies to determine the quantitative improvement in liver reserves with the use of the stable-isotope method in response to 1 or 2 high doses of vitamin A are needed to evaluate the current recommendations of the International Vitamin A Consultative Group (4). Studies looking at divided doses of vitamin A over time on the benefit to infants are also needed before global policy can be enforced. Because recent dietary intake is delivered to the mammary gland in chylomicra (35, 36), it is logical that divided doses would benefit the infant more. Moreover, in population groups in which the vitamin A status can be defined as marginal and not deficient, social marketing efforts to improve dietary quality may be more effective, less costly, and sustainable in the long-term than public health programs to continuously administer vitamin A capsules.
ACKNOWLEDGMENTS
We thank Rebecca Surles, graduate student, for her unselfish time spent in the training of SKT in the appropriate analysis of the MRDR test in SAT’s laboratory. Peter Crump, University of Wisconsin, Madison, College of Agriculture and Life Sciences Statistical Consulting Service, provided statistical consultation. Ashley Valentine assisted in constructing Figure 2. We thank all staff members of Kintampo Health Research Centre who assisted in field activities; special thanks to George Adjei, who assisted with data management.
SKT was responsible for orchestrating the study, sample and statistical analyses, and manuscript preparation. SAT, along with the valuable input of the late Paul Arthur, was responsible for the study design, dose preparation, data analysis, and manuscript preparation and revision. BdB helped to secure the funding. SO-A and SN contributed to the study design, field data collection, analysis, and interpretation of the results. FKNA and AT supervised and assisted with the thesis. None of the authors had any financial interest in the work or a conflict of interest with the sponsors of this study.
REFERENCES