点击显示 收起
1 From the Department of Epidemiology, Institute of Public Health, University of Copenhagen (HF); the Blair Research Laboratory, Ministry of Health, Harare, Zimbabwe (EG); the Department of Immunology, University of Zimbabwe, Harare, Zimbabwe (EG); the Department of Clinical Pharmacology, University of Zimbabwe, Harare, Zimbabwe (NN); the Department of Medical Laboratory Sciences, University of Zimbabwe, Harare, Zimbabwe (PN); the Department of Clinical Biochemistry, Aalborg University Hospital, Aalborg, Denmark (HK); and the Department of Human Nutrition, Royal Veterinary and Agricultural University, Frederiksberg, Denmark (PK and KFM)
2 Supported by the Council for Development Research, Danish International Development Assistance (to HF); the Danish Council for Medical Research (to HF); Southampton Insurance, Zimbabwe; the Foundation of 1870; BASF Health and Nutrition; the Hørslev Foundation; the Dagmar Marshall Foundation; the Sophus Jacobsens Foundation; and the Lily Benthine Lunds Foundation. 3 Address reprint requests to H Friis, Department of Epidemiology, Institute of Public Health, University of Copenhagen, Panum Institute, Blegdamsvej 3, DK-2200 Copenhagen, Denmark. E-mail: h.friis{at}pubhealth.ku.dk.
ABSTRACT
Background: Multiple micronutrient deficiencies may contribute to low birth weight, which is a major global determinant of mortality.
Objective: We assessed the effect of prenatal multimicronutrient supplementation on gestational length and birth size.
Design: We conducted a randomized, placebo-controlled, double-blind effectiveness trial among antenatal care attendees in Harare, Zimbabwe. Pregnant women (22–35 wk of gestation) were randomly allocated to receive a multimicronutrient or placebo supplement daily until delivery. Supplementation with iron and folic acid was part of antenatal care.
Results: Of 1669 women, birth data were available from 1106 (66%), of whom 360 (33%) had HIV infection. The mean gestational length was 39.1 wk, and 16.6% of the women had a gestational length < 37 wk. The mean birth weight was 3030 g, and 10.5% of the infants had a birth weight < 2500 g. Multimicronutrient supplementation was associated with tendencies for increased gestational length (0.3 wk; 95% CI: –0.04, 0.6 wk; P = 0.06), birth weight (49 g; –6, 104 g; P = 0.08), and head circumference (0.2 cm; –0.02, 0.4 cm; P = 0.07) but was not associated with low birth weight (birth weight < 2500 g) (relative risk: 0.84; 0.59, 1.18; P = 0.31). The effect of multimicronutrient supplementation on birth weight was not significantly different between HIV-uninfected (26 g; –38, 91 g) and HIV-infected (101 g; –3, 205 g) subjects (interaction, P > 0.10).
Conclusion: Antenatal multimicronutrient supplementation may be one strategy to increase birth size.
Key Words: Pregnancy • micronutrient supplementation • HIV • gestational length • birth size • low birth weight • Zimbabwe
INTRODUCTION
In developing countries, low birth weight is common, mainly because of impaired intrauterine growth. Low birth weight is a major determinant of child morbidity and mortality (1) and may also increase the risk of infectious and chronic diseases in adolescence and adulthood (2–4). Consequently, low birth weight is one of the major global health problems, particularly because persons born with low weight are at increased risk of being stunted and malnourished when reaching reproductive age and thus are at increased risk of maternal morbidity and mortality (5) and of delivering low-birth-weight babies (6).
Inadequate energy and macronutrient intakes contribute to low birth weight in some settings (7, 8), and efforts to improve the diets of women of reproductive age are of fundamental importance. Nevertheless, multiple micronutrient deficiencies may also contribute to low birth weight, and multimicronutrient supplementation is a cheap and feasible public health intervention, in particular if delivered as part of antenatal care (9).
The United Nations Children’s Fund (UNICEF) has for many years supported developing countries in providing iron and folic acid supplements to pregnant women (10). However, pregnant women are often deficient in a wider range of vitamins and minerals, which can be provided in the same supplement at a marginal extra cost (11). Therefore, a new supplement containing the recommended dietary allowance (RDA) of 15 different micronutrients, including iron and folic acid, has been proposed (12). Two trials from Nepal and Mexico, however, found no additional benefits of micronutrients in addition to iron and folic acid (13, 14), whereas effects on birth weight were found in 2 studies from Africa: a study among HIV-infected Tanzanian women (15) and a recent study among women from Guinea-Bissau, where the HIV prevalence is relatively low (P Kæstel, unpublished observations, 2004).
We assessed the effects of a prenatal multimicronutrient supplement on gestational length and birth size in a randomized, controlled effectiveness trial in Harare, Zimbabwe, where HIV infection is common but malaria is not endemic. Recruitment was up to 36 wk of gestation, because women in developing countries usually attend antenatal care late in pregnancy and because both micronutrient requirements and fetal growth increase toward the end of pregnancy. The supplement contained 3000 µg vitamin A and the RDA of 11 additional micronutrients, excluding iron and folic acid, which were provided through antenatal care. Data on micronutrient status (16, 17), body composition (18), and HIV viral load (19) was reported previously.
SUBJECTS AND METHODS
Study area and population
The study was conducted in the Mbare residential area of Harare, Zimbabwe, in 1996–1997. Harare is 1500 m above sea level and has a tropical climate with mean daily maximum and minimum temperatures of 25 °C and 12 °C, respectively. There are 4 main seasons: the early (June through August) and late (September through November) dry seasons and the early (December through February) and late (March through May) rainy seasons. The common diet is maize meal porridge eaten with vegetables and occasionally meat. Unemployment is very high in Mbare, especially among women, and most of those employed are general laborers. Malaria is not endemic in the study area because of the high altitude, but most women often return to their rural homes outside Harare, where malaria transmission may occur.
Women registering for antenatal care at the Edith Oppermann Maternity Hospital (EOMH) were informed about the study, and those who were between 22 and 36 wk of gestation were offered inclusion. The women were informed about HIV testing and counseled before they gave their written consent to participate. Counseling was provided by a nongovernmental organization, AIDS Counseling Trust. HIV results were made available within 2 wk, when those who wished to know their results received posttest counseling. A questionnaire was administered, and clinical examination and blood sampling were done. A research nurse obtained demographic data and medical and obstetric histories. Gestational age at recruitment and gestational length of pregnancy were calculated from the first day of the last menstruation or estimated by using fundus height. Antiretroviral drugs for prevention of mother-to-child transmission were not available. The initial sample size of 1700 was chosen to have 80% power to detect an 80-g difference in birth weight, with the assumption of a mean (± SD) of 3000 ± 500 g, with a significance level of 5%, and allowance for 20% loss to follow-up.
Permission was obtained from the ethical and scientific committees of the Medical Research Council of Zimbabwe, the Harare City Health Department, and the Ministry of Health. The Danish Central Medical Ethics Committee also approved the study. The national Ministry of Health recommendations were adhered to when giving women information on breastfeeding. Women who were found to be sick during clinical examination were referred for treatment at the EOMH or the Harare Central Hospital.
Examinations of the mothers
Clinical examinations of the mothers were done as part of antenatal care. Heights were measured to the nearest 0.1 cm and weights to the nearest 0.1 kg. Venous blood was collected, and serum was separated and stored in liquid nitrogen at –196 °C until tested for HIV. Samples found to be negative with the Genelavia Mixt HIV-1/2 enzyme immunoassay kit (Sanofi, Paris, France) were reported as negative. Positive and indeterminate samples were further tested by using the Recombigen HIV-1/2 enzyme immunoassay kit (Cambridge Biotech, Dublin) with different antigens from those in the Genelavia kit. Samples that were positive or indeterminate with the Recombigen kit were tested at the Department of Clinical Chemistry, Aalborg University Hospital, Aalborg, Denmark, by using an in-house reverse transcriptase-polymerase chain reaction method (20).
Study intervention
The study intervention was a daily tablet containing approximately the RDA for pregnant or lactating women of 13 micronutrients (21), except for vitamin A, for which a higher amount was given. The composition of the multimicronutrient supplement is shown in Table 1. An iron and folic acid supplement was given separately as part of routine antenatal care and was therefore not included in the multimicronutrient tablet. Similarly, iodine was not included because salt is iodized in Zimbabwe.
View this table:
TABLE 1. Composition of multimicronutrient supplement1
Allocation to daily supplementation with multimicronutrient or identical-looking placebo tablets was based on simple, blocked randomization. The digits 0–5 in a computer-generated random sequence were replaced by 6 preassigned permuted blocks of 4: AABB, ABAB, ABBA, BABA, BBAA, and BAAB; the digits 6–9 were deleted. Containers with 110 multimicronutrient or placebo tablets, which were coded A or B, respectively, were delivered by the manufacturer (Almega, Ringsted, Denmark) together with the code in 2 sealed envelopes. Duplicate containers, which corresponded to the random sequence, were consecutively numbered from 1 to 1800.
The study participants were numbered consecutively at recruitment. After baseline examinations, each participant was given a container with the same number and instructed to take one tablet with a meal every morning until delivery. For the women who delivered their infants at the EOMH, the remaining tablets were counted, and the number of tablets taken was estimated and used as a measure of compliance.
Birth records
Infants delivered at the EOMH were weighed, and length and head circumference were measured within 72 h by a trained research nurse. Birth weight was determined to the nearest 0.01 kg by using a Sohnle infant scale (Sohnle, Murrhardt, Germany). Length and head circumference were measured with a portable measuring mat and a measuring tape, respectively, to the nearest 1 cm. When study participants were not seen by the research nurse at the EOMH after delivery, the place of delivery was established through home visits. Birth records were later retrieved, and the necessary routine data were extracted. As standard procedure, all health centers keep antenatal records. The delivery records contain demographic information (name, address, and date of birth), obstetric history, plurality (ie, twins or triplets), vaginal or cesarean delivery, episiotomy, etc. In addition, birth weight, head circumference, and length of the infant were recorded. All data were recorded within 72 h.
Statistical analysis
All singletons and the firstborn of multiple pregnancies were included in the analysis, which was based on intention to treat. The outcome variables were measurements of gestational length and birth size (ie, birth weight, length, and head circumference). Based on these measurements, the ponderal index [birth weight [in kg)/length3 (in m)], percentage of infants delivered preterm (gestational length < 37 wk), percentage of infants with a birth weight < 2500 g, and percentage of infants with intrauterine growth retardation-low birth weight (IUGR-LBW, defined as gestational age > 37 wk and birth weight < 2500 g) were computed.
Data were analyzed by using SPSS version 11 (SPSS Inc, Chicago). The distribution of the continuous outcome variables conformed to normality as assessed by normal probability plots. The two-sample t test and the chi-square test were used to test for differences in means and proportions, respectively. The effects of the intervention were presented as differences in means for continuous variables and as relative risks for binary variables. All effect measures were given with 95% CIs. Effect estimates were presented after stratification by maternal HIV infection status, despite the lack of statistical interaction between the intervention and HIV infection status.
Multiple linear and logistic regression analysis was used to assess for effect modification and confounding by maternal HIV infection status, gravidity, season of delivery, anemia, and infant sex. If the P for interaction was <0.10, then interaction terms were computed to allow estimation separately for each level of the effect-modifying variable.
RESULTS
Baseline equivalence between the 837 women allocated to receive multimicronutrients and the 832 women allocated to receive placebo was achieved. However, the proportion of primigravidae was slightly higher in the placebo group, and the proportion of women with HIV infection was lower in the placebo group (Table 2).
View this table:
TABLE 2. Baseline comparison between multimicronutrient and placebo groups
Birth records were available for 1106 (66.3%) of the 1669 women recruited. For 235 (21.2%) of the 1106 women, routine birth records were used. There was no significant difference in the proportion of routine birth records in the micronutrient and placebo groups (22.6% compared with 19.9%, respectively; P = 0.43). As seen in Table 3, loss to follow-up differed with season of enrollment. Furthermore, mean gestational age tended to be higher in the 1106 women who were followed up than in the 563 women who were not followed up, and the proportion of primigravidae was significantly lower in the women who were followed up. However, there were no significant differences in age, anthropometric measures, or HIV prevalence. Importantly, there was no significant difference in treatment allocation between those who were or were not followed up. The analysis included 1093 (98.8%) singletons and the firstborn of 12 pairs of twins and 1 set of triplets. Of these 1106 neonates, 11 (1.0%) were stillborn. Only 22 (2.0%) of the 1106 deliveries were by cesarean delivery, and these were mainly emergency deliveries because of preeclampsia or fetal distress. The mean gestational length was 39.1 wk, and 184 (16.6%) of the women had a gestational length < 37 wk. The mean birth weight was 3030 g, and 116 (10.5%) of the infants had a birth weight < 2500 g. Sixty-six (6.0%) of the infants had a birth weight < 2500 g but a gestational length > 37 wk (IUGR-LBW).
View this table:
TABLE 3. Comparison of the baseline characteristics of the 1106 women for whose infants birth weight data were available with the baseline characteristics of the 563 women who were not included in the analyses
There were no significant differences between the multimicronutrient and placebo groups in the distribution of cesarean sections [11 (2.0%) compared with 11 (2.0%), respectively; P = 0.92] and stillbirths [4 (0.7%) compared with 7 (1.3%), [respectively; P = 0.39]. As seen in Table 4, gestational length was 0.3 (95% CI: –0.04, 0.6) wk longer in the women allocated to receive daily multimicronutrient supplementation than in those allocated to receive placebo, although the difference was not significant (P = 0.06). Similarly, compared with placebo supplementation, daily multimicronutrient supplementation tended to be associated with a 49-g (95% CI: –6, 104 g) higher birth weight and a 0.2-cm (95% CI: –0.02, 0.4 cm) higher head circumference. The relative risks of preterm delivery, birth weight < 2500 g, and IUGR-LBW were in the range of 0.71–0.84 but were not significant (P 0.15) (Table 5). Although there were no interactions between multimicronutrient supplementation and HIV infection status with respect to any of the outcomes (interaction, P > 0.10), the effects of supplementation are shown separately for the HIV-uninfected and the HIV-infected subjects in Tables 6 and 7 as decided in advance.
View this table:
TABLE 4. Effects of maternal supplementation with multimicronutrients on gestational length, birth weight and length, ponderal index, and head circumference in 1106 infants1
View this table:
TABLE 5. Effects of maternal supplementation with multimicronutrients on percentages and relative risks of preterm delivery, birth weight < 2500 g, and intrauterine growth retardation-low birth weight (IUGR-LBW) in 1106 infants
View this table:
TABLE 6. Effects of maternal supplementation with multimicronutrients on gestational length, birth weight and length, ponderal index, and head circumference in 1106 infants by maternal HIV infection status1
View this table:
TABLE 7. Effects of maternal supplementation with multimicronutrients on percentages and relative risks of preterm delivery, birth weight < 2500 g, and intrauterine growth retardation-low birth weight (IUGR-LBW) in 1106 infants by maternal HIV infection status1
Effect modification and confounding by gravidity, season of delivery, anemia, and infant sex were assessed. A marginally significant interaction between multimicronutrient supplementation and infant sex was found for birth length (P = 0.07) because of the stronger effect of multimicronutrient supplementation on length in girls (0.5 cm; 95% CI: 0.1, 1.0 cm; P = 0.017) than in boys (0.02 cm; 95% CI: –0.4, 0.4 cm; P = 0.92). There were no interactions with gravidity, season of delivery, or anemia (P > 0.10). Adjustment for gravidity, HIV infection status, season of delivery, anemia, and infant sex did not change the estimated effect of multimicronutrient supplementation either among all the women or among the women stratified by HIV infection status.
The mean number of days from recruitment until delivery was 69 (95% CI: 67, 71). Among the 558 (52.8%) women for whom tablet intake could be assessed, the mean number of tablets taken was 62 (95% CI: 59, 64), and the mean compliance, expressed as the percentage of days from recruitment to delivery when a tablet was taken, was 80%. The mean tablet intake did not differ significantly between the multimicronutrient and placebo groups (62 compared with 62, respectively; P = 0.96).
DISCUSSION
The percentage of low birth weight (<2500 g) found in the present study (ie, 10.5%) was similar to that previously reported from Zimbabwe (22, 23). Although this percentage is low compared with that in other developing countries (24) and is close to the global goal of <10% set by the World Summit for Children in 1990 (25), even birth weights between 2500 and 3000 g have been estimated to be associated with mortality that is several times higher than that associated with birth weights between 3000 and 3500 g (1).
Interestingly, the almost 50-g mean increase in birth weight that was seen in the present study was not confined to the infants in the lower range of the birth weight distribution. An effect of micronutrient supplementation on birth weight may be mediated by prevention of growth-inhibiting maternal or fetal micronutrient deficiencies or by reduction of the effect of infectious disease morbidity on fetal growth. A larger effect of micronutrient supplementation in HIV-infected women was therefore expected because even early asymptomatic HIV infection seems to increase micronutrient requirements (26, 27). However, the point estimates were not significantly different between the HIV-infected and the HIV-uninfected women in the present study.
The loss to follow-up was high because many women went to rural areas to deliver, and this was temporarily exacerbated by civil unrest. In addition to reducing the statistical power of a study, loss to follow-up may introduce bias if it is associated with both the intervention and the outcome. However, in the present study, there was little difference in baseline variables between those who were followed-up and those who were lost to follow-up. More importantly, the loss to follow-up was similar in the micronutrient and placebo groups. The study was an effectiveness trial because the participants were given all tablets at recruitment and were advised to take one tablet daily up to delivery. Because the women were not seen again before delivery, if at all, there were no opportunities to supervise or encourage adherence to the tablet-taking regimen. Consequently, adherence was not optimal, with only one-half of the women taking >85% of the tablets. Because data on compliance were obtained from only one-half of the women, the data do not permit assessment of a dose-response relation. It is likely that compliance was lower among those from whom data were not available. The effects reported here may reflect what can be expected during programmatic implementation. Similarly, the routine iron and folate intakes were not assessed.
Although our study predates the initiative by UNICEF to replace the current iron and folic acid prenatal supplement with a new supplement containing 13 additional micronutrients, the composition and doses are similar, albeit not identical (12). As for differences in composition, iron and folic acid were not included in the supplement tested but were given separately as part of routine antenatal care. Similarly, iodine was not included because salt is iodized in Zimbabwe. Our supplement contained the 12 remaining micronutrients contained in the UNICEF supplement. The main difference in doses between our supplement and the UNICEF supplement was that our supplement contained 3000 µg retinol equivalents preformed vitamin A and 3.5 mg ß-carotene; the doses of the remaining 11 micronutrients in our supplement were similar to those in the UNICEF supplement (ie, approximately the RDA).
Few such studies on the effect of multimicronutrients have been conducted (28). However, several studies to assess the effect of the new UNICEF prenatal supplement and supplements with different compositions or higher doses are currently in progress. Some of these studies not only have birth size as an outcome, but are also powered to study effects on perinatal and neonatal mortality. These studies were recently described in detail (29). A trial among Mexican women compared the effect of a supplement with 13 micronutrients (1–1.5 times the RDA) and iron with the effect of iron alone (14). Despite the facts that the women were recruited before week 13 of gestation and that the supplements were consumed under supervision, no effects were detected. In Nepal, 4 different micronutrient supplements were compared with placebo (13). Iron and folic acid increased birth weight by 37 g (95% CI: –16, 90 g), whereas neither folic acid alone nor iron, folic acid, and zinc combined had any significant effect. The effect of the multimicronutrient, which included iron, folic acid, and zinc, was of the same magnitude as that of iron and folic acid. Thus, adding more micronutrients to the iron and folic acid supplement produced no additional benefit in terms of improving birth weights. In contrast, a recent trial in Guinea-Bissau found that the UNICEF tablet increased birth weight but that a similar tablet containing 2 times the RDA had a larger effect (P Kæstel, unpublished observations, 2004).
Among HIV-infected pregnant women in South Africa, vitamin A supplementation was shown to reduce the risk of preterm deliveries (30). This result was not found in a similar study in Malawi, in which vitamin A increased birth weight by 95 g (31). In contrast to these studies, a study among pregnant Tanzanian women with HIV did not find an effect of vitamin A on either preterm delivery or birth weight (15). Nevertheless, multivitamin supplementation, which was assessed simultaneously by using a two-by-two factorial design, halved the risks of fetal loss, preterm delivery, and low birth weight and increased mean birth weight by 150 g. The dose of vitamin A in the Malawi trial was similar to ours, whereas the vitamin A intervention in both the South African and the Tanzanian trials consisted of 1500 µg retinol equivalents preformed vitamin A and 30 mg ß-carotene. The multivitamin intervention in the Tanzanian trial contained large doses of B vitamins, vitamin C, and vitamin E. In contrast to the Tanzanian trial, the present study did not assess the effects of the intervention on mother-to-child HIV transmission. However, we were able to assess interactions between multimicronutrient supplementation and HIV infection for various birth size outcomes. Although the effect on birth weight appeared to be larger among the HIV-infected women, this could not be substantiated by statistical tests of interaction.
In a randomized, placebo-controlled, double-blind effectiveness trial, we found that daily multimicronutrient supplementation of pregnant Zimbabwean women was associated with marginally significant effects on several birth-size variables. Given differences in micronutrient status, the coexistence of multiple deficiencies, and micronutrient-micronutrient interactions, the effects of similar interventions might differ across populations. Although multimicronutrient supplementation may be beneficial in increasing birth weight and in HIV-afflicted populations, its effects on a broader range of maternal and infant health outcomes need to be assessed, and effect modifiers need to be identified.
ACKNOWLEDGMENTS
We thank the women and staff at the EOMH for their participation and cooperation; the City Health Council, Harare, and the staff at the Blair Research Laboratory for assistance in conducting the study; and the Secretary for Health, Zimbabwe, for permission to publish.
The study was conceived and planned by HF, EG, NN, and KFM. Field and laboratory work was done by HF, EG, NN, PN, and HK. Data processing was done by HF and PK. The first draft of the manuscript was written by HF, and all authors contributed to the final version. None of the authors had any conflicts of interest.
REFERENCES