Abstract

Background There have been few reports of differential responses to nutrition interventions in women and children from poor households. Women from poor households have greater nutritional risks and are potentially a target group for programmes. We assessed the modifying effects of household wealth on responses to micronutrient supplements in pregnancy on newborn anthropometry and perinatal mortality.

Methods A cluster randomized double-blind controlled trial conducted in two rural counties in northwestern China. All pregnant women in villages were randomly allocated from enrolment until delivery to daily supplementation with folic acid (control), iron/folic acid or multiple micronutrients (MMNs) with a recommended allowance of 15 vitamins and minerals. Wealth was based on a score from an inventory of household assets.

Results In the pregnant women from the poorest one-third of the households, MMN supplements significantly increased birth weight by 68 g [95% confidence interval (CI) 4–131 g], reduced low birth weight by 60% [relative risks (RRs) 0.40, 95% CI 0.21–0.78] and tended to reduce early neonatal mortality by 52% (RR 0.48, 95% CI 0.17–1.36) compared with folic acid. Iron/folic acid significantly increased the duration of gestation by 0.41 weeks (95% CI 0.18–0.65), reduced pre-term birth by 45% (RR 0.55, 95% CI 0.32–0.93) and significantly reduced early neonatal mortality by 90% (RR 0.10, 95% CI 0.01–0.79) compared with folic acid. Iron/folic acid and MMN supplements had no significant effects in women from wealthier households.

Conclusions In rural China, women from the poorest households had the largest perinatal outcome responses to micronutrient supplementation. In these women, standard iron/folic acid provided more protection for neonatal survival than MMN supplements.

Introduction

The Millennium Development Goals (MDGs) and the 2005 World Health Report are helping to focus attention on maternal and child health and nutrition.1 Almost 40% of deaths of children aged <5 years occur in the neonatal period and this proportion is increasing as the number of post-neonatal deaths falls more rapidly.1 To meet the fourth MDG, which calls for mortality in children under the age of 5 years to be reduced by two-thirds, more rapid reductions will be required, particularly in early neonatal deaths.2 More than 9 million deaths occur before or just after birth each year, 98% of them in developing countries. Pre-term birth complicates ∼5–10% of all deliveries worldwide and is the major cause of perinatal mortality and long-term physical and neurological morbidity both in developing and industrialized countries.3 Birth weight is closely associated with the health and survival of infants in the developing world, where 96% of the 20 million low birth weight (LBW) babies are born each year. LBW is associated with perinatal and infant mortality.4 Imbalances in maternal nutrition can adversely affect normal fetal growth and development. Impaired fetal growth has been associated with negative short- and long-term outcomes such as increased perinatal morbidity and mortality, and infant mortality.5 Iron deficiency is the most commonly recognized nutritional deficit in women of reproductive age in both the developed and the developing world.6 Both low and high haemoglobin concentrations during pregnancy are associated with adverse outcomes, and as yet it is not known whether iron supplementation decreases or increases the risk of adverse outcomes.7,8 Interventions for maternal nutrition [supplements of iron folic acid, multiple micronutrients (MMNs)] can improve outcomes for maternal health and births, but few have been assessed on a sufficient scale.9

China is the most populous developing country in the world, and the prevalence of LBW varies across different socio-economic groups and geographical areas in the country. In China, the Information and Statistics Centre of the Ministry of Health has classified rural counties into one of four groups based on the score from a basket of socio-economic indicators, with Type 4 the poorest and Type 1 the wealthiest.10 A sampling survey on LBW in China conducted in 1999 showed that the prevalence of LBW was 4.5% in the ‘Type 1’ counties, 5.6% in the ‘Type 2’ counties, 7.6% in ‘Type 3’ counties and 11.8% in the ‘Type 4’ counties.11 Similar trends were found in the same survey for the prevalence of anaemia during pregnancy, which was 31.1% in the ‘Type 1’ counties, 40.3% in the ‘Type 2’ counties, 55.2% in ‘Type 3’ counties and 50.6% in the ‘Type 4’ counties.12 Despite these findings, there are no specific policies or programmes in China to support the distribution of micronutrient supplements during pregnancy, including iron/folic acid, even to disadvantaged women.

We recently reported from Shaanxi Province, China, a cluster randomized controlled trial (CRCT) of the impact of iron/folic acid and MMN supplements during pregnancy, compared with folic acid alone, on birth anthropometry, duration of gestation, third-trimester maternal haemoglobin and perinatal mortality.13 Antenatal iron–folic acid increased the duration of gestation, reduced early pre-term delivery (<34 weeks of gestation) and was associated with a significant reduction in early neonatal mortality compared with folic acid. Antenatal MMN supplements modestly increased birth weight compared with folic acid, but this weight gain was not associated with reduced early neonatal mortality.

For this community-based trial, three subgroup analyses were planned to identify factors that might potentially modify the responses to antenatal micronutrient supplementation. The factors examined included timing of the start of supplementation, baseline maternal body mass index and household wealth, each of which is being reported separately.

Women from poor households have greater nutritional risks, including iron deficiency anaemia that in part is related to poor diets with limited bioavailable iron. In a dietary assessment of the women participating in the Shaanxi trial, we found significantly lower intakes of iron (as well as other micronutrients) among women from the poorest households.14 This group of women are potentially a specific target group for programmes that aim to reduce iron deficiency anaemia. In this secondary analysis of the Shaanxi trial data, we examine the modifying effects of household wealth on the impact of the micronutrient supplements on birth anthropometry, duration of gestation and perinatal mortality.

Methods

Experimental design and population

The details of the study design and methodology have previously been described.13 Briefly, a double-blind CRCT was conducted in Changwu and Bin rural counties, in Shaanxi Province, northwest China, from August 2002 to January 2006. There were 34 townships that were further divided into 561 villages as the unit of randomization in these two counties. The randomization of the villages was stratified by county with a fixed ratio of treatments (1 : 1 : 1) and blocking of 15. The same treatment was allocated to all pregnant women detected in a given village. A treatment colour code was assigned to each village based on the treatment allocation schedule. The treatment codes were only opened once all data had been collected and blinded analysis of the primary hypothesis was completed. A total of 5828 eligible pregnant women were recruited and randomized. After taking account of the cluster randomization, this sample was estimated to provide 80% power to detect a 50-g difference in birth weight between either iron/folic acid or MMNs and folic acid (control) groups, assuming a two-tailed test, with α = 0.05 and birth weight standard deviation (SD) of 436 g. This sample size had lower power for the other outcomes; for example, it had 80% power to detect a 75% reduction in early neonatal mortality between treatment groups, assuming a two-tailed test with α = 0.05 and an early neonatal mortality rate of 15 per 1000 live births.

Enrolment and pregnancy surveillance procedures

All the women of reproductive age who were likely to become pregnant including newly married couples, couples not using contraception or couples expressing the desire for another child were visited by village doctors every month and asked about the date of their last menstrual period (LMP). Informed verbal consent for pregnancy monitoring was obtained, and consenting women were visited every month and women with periods delayed by >5 days had a urine pregnancy test, and confirmed pregnancies were reported to the township Maternal and Child Health care (MCH) worker. Pregnant women who were township or county citizens were passively detected at antenatal clinics in township or county health facilities. Newly identified, consenting pregnant women were enrolled into the trial and a baseline interview conducted to record their socio-demographic status, and their menstrual, reproductive, medical and family histories. The recruited pregnant women received three, free antenatal care checks, at which they were asked about pregnancy complications and had a physical examination including blood pressure and weight measurement. All the information collected during the pregnancy until the 6-week follow-up visit after delivery was recorded in a Pregnancy Care Record Book that served both as a clinical record and data capture instrument.

Interventions

All the women in the trial received one of the following three daily antenatal supplements: folic acid (400 µg) alone as control; iron (60 mg)/folic acid (400 µg) or MMN (30 mg iron, 400 µg folic acid, 15.0 mg zinc, 2.0 mg copper, 65.0 µg selenium, 150.0 µg iodine, 800.0 µg vitamin A, 1.4 mg vitamin B1, 1.4 mg vitamin B2, 1.9 mg vitamin B6, 2.6 µg vitamin B12, 5.0 µg vitamin D, 70.0 mg vitamin C, 10.0 mg vitamin E and 18.0 mg niacin). The MMN supplements were formulated to contain approximately one recommended dietary allowance for the micronutrients included. The supplements were packaged in blister packs with identical appearance.

At enrolment, each pregnant woman received 15 capsules with instructions to take one capsule daily. The women were visited every 2 weeks by the village doctor who replenished supplements, retrieved the used blister strips and recorded the number of remaining capsules. The number of supplements consumed throughout the trial was summed to estimate compliance, which is defined as the percentage of expected days that supplements were consumed

Measurement of perinatal outcomes

Birth anthropometry was measured by the nursing staff within 1 h of delivery at hospital or measured by township MCH staff within 72 h for home deliveries. Birth weight was measured using an electronic scale (Type BD-585 TANITA Corporation, Dongguan, Guangdong Province, PR China) with precision to the nearest 50 g in hospital delivery or by a baby scale (Type RTZ-10A-RT Wuxi Weigher Factory, Wuxi, PR China) with precision to the nearest 50 g in home delivery. Recumbent length was measured to the nearest 1 mm using a portable measuring board with fixed head piece. LBW was defined as a birth weight <2500 g.

Gestational age at birth was calculated as completed days based on the first day of the LMP, obtained at the baseline interview. Pre-term delivery and early pre-term delivery were defined as delivery before 37 or 34 weeks of gestation. Small for gestational age babies were defined as those whose weight was below the 10th centile of the gestational age–sex-specific US reference for fetal growth.15

Fetal losses during pregnancy, birth outcome and delivery information, and neonatal and maternal deaths were reported by the village doctors or hospital staff and recorded by MCH staff using pre-coded structured forms. Neonatal survival was assessed at the 6-week follow-up visit. Abortions were terminations of a confirmed pregnancy before 28 weeks of gestation and recorded as spontaneous or induced; stillbirths were fetuses delivered at 28 weeks of gestation or later with no signs of life and recorded as occurring during or prior to the onset of labour; early neonatal deaths were deaths among live-born infants occurring 0–6 days after delivery; late neonatal deaths were infant deaths occurring from 7 to 27 days after delivery and perinatal deaths were stillbirths plus early neonatal deaths.

All women who had an abortion, stillbirth or neonatal death were re-interviewed by project staff to check the reported information and, for hospital or clinic deliveries, to cross check with medical records.

Statistical analysis

To assess the economic status of the household, the ownership of 16 different household assets or facilities (bicycle, motorbike, orchard, radio/cassette player/stereo system, television/VCD, refrigerator/freezer, washing machine, poultry, goat and/or sheep and/or pigs, buffalo and/or cow, tractor, main material used for house, main material used in the floors of the house, electricity available in the house, potable water available in the house, the kind of toilet) were used to construct a household wealth index using a principal components method to assign a weight for each asset.16 In the analyses, the index was categorized as tertiles indicating the poorest, middle and the wealthiest households. This wealth index has been validated in the study counties.17 The primary analyses were for singleton births alone, but additional analyses for combined singleton and multiple births were also conducted.

The baseline characteristics of the pregnant women and compliance with supplementation were examined across treatment groups by household wealth categories. Mean differences for anthropometry and gestation at birth by household wealth categories, with the folic acid group as the reference, and the 95% confidence intervals (CIs), were estimated using generalized estimating equation (GEE) linear models, adjusted for the cluster randomization design with an independent correlation structure.

To adjust for cluster randomization, GEE binomial regression models with log link and exchangeable correlation structures were applied to estimate the relative risks (RR) and 95% CIs for LBW and pre-term delivery by household wealth categories with the folic acid group as the reference. Similar analysis strategies were applied to estimate the RRs and 95% CIs for stillbirth, neonatal and perinatal death rates, with the folic acid group as the reference, by household wealth. Analyses of combined singleton and multiple births included a factor to adjust for multiple births. Wald chi-squared tests assessed the overall effects of the three treatments for each outcome.

We followed the existing guidelines for reporting subgroup analyses from randomized controlled trials.18,19 Formal statistical tests for interaction were undertaken by adding an interaction term between treatment assignment and household wealth in the models for each of the outcomes, and the P-value for this interaction term has been reported in the results tables. In the subgroup analyses for the trial, one baseline factor (household wealth) was examined for each of 11 outcomes. We estimate that less than one statistically significant interaction test (P < 0.05) could be expected by chance alone. All analyses were done with STATA version 9.2 (Stata/SE 9.2 Stata Corporation, College Station, TX, USA; 2006).

Trial registration

The study was registered as an International Standard Randomised Controlled Trial (ISRCTN08850194).

Ethical approval

The research protocol was approved by the Human Research Ethics Committee of the College of Medicine, Xi’an Jiaotong University (No. 2002001). The project was also approved by the Ministry of Health, China.

Role of the funding source

The study sponsor, UNICEF, invited the investigators to compare their recommended MMN supplements with the standard recommendation in China of folic acid alone, and with iron/folic acid which is commonly recommended. The sponsor played no role in the design or implementation of the collection, analysis or interpretation of data, or in the writing of this report, or in the decision to submit this paper for publication.

Results

Figure 1 shows the flow chart of participants in the trial. In total, 5828 pregnant women were enrolled in the trial, but of these 133 pregnant women were lost to follow-up, 279 pregnant women stopped taking supplements and refused to continue to participate and there were 601 fetal losses. There were a total of 4650 pregnancies that resulted in at least one live birth. There were 167 stillbirths for whom birth weight was usually not measured. There were a total of 4604 live singleton births, among which 4330 (94%) had birth weight available for analysis. There were 48 multiple births with 47 pairs of twins and 1 set of triplets. These multiple births were not balanced across the treatment groups with 19 pairs of twins (of 1580 births) in the iron/folic acid group, 12 twins and 1 triplets (of 1551 births) in the MMN group and 17 pairs of twins (of 1711 births) in the folic acid group (P = 0.031).

Figure 1

Participant flow chart

Figure 1

Participant flow chart

Table 1 presents the baseline characteristics among the three treatment groups in the different household wealth groups. The timing of initiation of supplementation and the average consumption were balanced among the three treatment groups in the poorest and wealthier households. The baseline characteristics were different among pregnant women from the poorest and wealthier households. The women from the poorest households were slightly older (P < 0.001), had lower education (P < 0.001), higher parity (P < 0.001) and shorter body stature (P < 0.001) than the women from the wealthier households. The overwhelming majority of women from the poorest households (96%) were farmers, but this declined to ∼78% for women from the wealthier households (P < 0.001).

Table 1

Baseline characteristics of pregnant women by household wealth index and treatment groups

 Household wealth index
 
 Poorest
 
Wealthier
 
 Treatment groups
 
Treatment groups
 
 Folic acid Iron/folic acid MMN Folic acid Iron/folic acid MMN 
Number of clusters 143 149 148 269 285 258 
    Population in clusters in 2001, mean (SD) 866 (514) 793 (425) 929 (530) 910 (519) 829 (443) 984 (480) 
    Births in clusters in 2001, mean (SD) 7.3 (5.5) 6.7 (4.5) 7.4 (5.0) 7.7 (5.6) 7.0 (4.6) 8.7 (9.5) 
    Pregnancies during trial/cluster/years, mean (SD) 3.9 (3.8) 3.5 (2.8) 3.9 (3.8) 4.0 (4.4) 3.8 (4.8) 3.9 (5.1) 
    Number of family members/cluster, mean (SD) 4.1 (0.9) 4.2 (0.8) 4.1 (0.9) 4.3 (0.9) 4.3 (1.0) 4.4 (1.0) 
Number of pregnant women 553 488 531 1150 1075 1011 
Maternal age (years)       
    Mean (SD) 25.4 (4.5) 25.1 (4.4) 25.6 (4.6) 24.4 (4.3) 24.3 (4.2) 24.3 (4.3) 
    15–19 44 (8.0) 45 (9.2) 47 (8.9) 90 (7.8) 84 (7.8) 84 (8.3) 
    20–24 213 (38.5) 186 (38.1) 194 (36.5) 592 (51.5) 557 (51.8) 528 (52.2) 
    25–29 186 (33.6) 168 (34.4) 176 (33.2) 286 (24.9) 277 (25.8) 260 (25.7) 
    30–34 98 (17.7) 78 (16.0) 101 (19.0) 164 (14.3) 142 (13.2) 119 (11.8) 
    ≥35 12 (2.2) 11 (2.3) 13 (2.5) 18 (1.6) 15 (1.4) 20 (2.0) 
Women’s education       
    <3 years 80 (14.5) 46 (9.5) 68 (12.8) 32 (2.8) 31 (2.9) 29 (2.9) 
    Primary 216 (39.1) 204 (41.9) 213 (40.2) 232 (20.3) 202 (18.9) 216 (21.4) 
    Secondary 239 (43.3) 218 (44.8) 225 (42.5) 677 (59.2) 628 (58.7) 561 (55.7) 
    High school plus 17 (3.1) 19 (3.9) 24 (4.5) 202 (17.7) 209 (19.5) 202 (20.0) 
Women’s occupation       
    Farmer 526 (95.5) 468 (96.5) 511 (96.8) 901 (79.2) 830 (77.6) 788 (78.4) 
    Others 25 (4.5) 17 (3.5) 17 (3.2) 237 (20.8) 239 (22.4) 217 (21.6) 
Gestation at enrolment (weeks)       
    Mean (SD) 14.2 (5.8) 14.3 (5.4) 14.5 (5.6) 14.2 (5.8) 13.8 (5.8) 14.3 (5.6) 
    <12 228 (41.2) 175 (35.9) 201 (37.9) 466 (40.5) 462 (43.0) 400 (39.6) 
    12–15 127 (23.0) 118 (24.2) 122 (23.0) 249 (21.7) 240 (22.3) 242 (23.9) 
    ≥16 198 (35.8) 195 (40.0) 208 (39.2) 435 (37.8) 373 (34.7) 369 (36.5) 
Number of previous pregnancies, mean (SD) 1.8 (0.9) 1.8 (0.8) 1.8 (0.8) 1.6 (0.8) 1.6 (0.8) 1.6 (0.8) 
Parity       
    0 301 (54.4) 272 (55.7) 295 (55.6) 777 (67.6) 741 (68.9) 710 (70.2) 
    1 224 (40.5) 188 (38.5) 208 (39.2) 327 (28.4) 303 (28.2) 267 (26.4) 
    ≥2 28 (5.1) 28 (5.7) 28 (5.3) 46 (4.0) 31 (2.9) 34 (3.4) 
Height (cm)       
    n 549 488 528 1140 1067 1001 
    Mean (SD) 157.8 (5.3) 157.6 (5.4) 157.5 (5.6) 159.5 (4.9) 159.4 (4.9) 159.4 (5.1) 
Weight (kg)       
    n 545 483 528 1121 1046 983 
    Mean (SD) 52.4 (6.1) 52.1 (6.8) 52.8 (6.7) 52.8 (6.0) 52.8 (6.1) 52.7 (5.7) 
BMI (kg/m2)       
    n 543 482 523 1121 1045 980 
    Mean (SD) 21.0 (2.3) 21.0 (2.4) 21.3 (2.5) 20.7 (2.1) 20.8 (2.2) 20.8 (2.0) 
Mid-upper arm circumfrance (cm)       
    n 544 483 528 1134 1060 1003 
    Mean (SD) 23.2 (2.2) 23.2 (2.2) 23.4 (2.1) 23.1 (2.0) 23.2 (2.1) 23.1 (2.1) 
 Household wealth index
 
 Poorest
 
Wealthier
 
 Treatment groups
 
Treatment groups
 
 Folic acid Iron/folic acid MMN Folic acid Iron/folic acid MMN 
Number of clusters 143 149 148 269 285 258 
    Population in clusters in 2001, mean (SD) 866 (514) 793 (425) 929 (530) 910 (519) 829 (443) 984 (480) 
    Births in clusters in 2001, mean (SD) 7.3 (5.5) 6.7 (4.5) 7.4 (5.0) 7.7 (5.6) 7.0 (4.6) 8.7 (9.5) 
    Pregnancies during trial/cluster/years, mean (SD) 3.9 (3.8) 3.5 (2.8) 3.9 (3.8) 4.0 (4.4) 3.8 (4.8) 3.9 (5.1) 
    Number of family members/cluster, mean (SD) 4.1 (0.9) 4.2 (0.8) 4.1 (0.9) 4.3 (0.9) 4.3 (1.0) 4.4 (1.0) 
Number of pregnant women 553 488 531 1150 1075 1011 
Maternal age (years)       
    Mean (SD) 25.4 (4.5) 25.1 (4.4) 25.6 (4.6) 24.4 (4.3) 24.3 (4.2) 24.3 (4.3) 
    15–19 44 (8.0) 45 (9.2) 47 (8.9) 90 (7.8) 84 (7.8) 84 (8.3) 
    20–24 213 (38.5) 186 (38.1) 194 (36.5) 592 (51.5) 557 (51.8) 528 (52.2) 
    25–29 186 (33.6) 168 (34.4) 176 (33.2) 286 (24.9) 277 (25.8) 260 (25.7) 
    30–34 98 (17.7) 78 (16.0) 101 (19.0) 164 (14.3) 142 (13.2) 119 (11.8) 
    ≥35 12 (2.2) 11 (2.3) 13 (2.5) 18 (1.6) 15 (1.4) 20 (2.0) 
Women’s education       
    <3 years 80 (14.5) 46 (9.5) 68 (12.8) 32 (2.8) 31 (2.9) 29 (2.9) 
    Primary 216 (39.1) 204 (41.9) 213 (40.2) 232 (20.3) 202 (18.9) 216 (21.4) 
    Secondary 239 (43.3) 218 (44.8) 225 (42.5) 677 (59.2) 628 (58.7) 561 (55.7) 
    High school plus 17 (3.1) 19 (3.9) 24 (4.5) 202 (17.7) 209 (19.5) 202 (20.0) 
Women’s occupation       
    Farmer 526 (95.5) 468 (96.5) 511 (96.8) 901 (79.2) 830 (77.6) 788 (78.4) 
    Others 25 (4.5) 17 (3.5) 17 (3.2) 237 (20.8) 239 (22.4) 217 (21.6) 
Gestation at enrolment (weeks)       
    Mean (SD) 14.2 (5.8) 14.3 (5.4) 14.5 (5.6) 14.2 (5.8) 13.8 (5.8) 14.3 (5.6) 
    <12 228 (41.2) 175 (35.9) 201 (37.9) 466 (40.5) 462 (43.0) 400 (39.6) 
    12–15 127 (23.0) 118 (24.2) 122 (23.0) 249 (21.7) 240 (22.3) 242 (23.9) 
    ≥16 198 (35.8) 195 (40.0) 208 (39.2) 435 (37.8) 373 (34.7) 369 (36.5) 
Number of previous pregnancies, mean (SD) 1.8 (0.9) 1.8 (0.8) 1.8 (0.8) 1.6 (0.8) 1.6 (0.8) 1.6 (0.8) 
Parity       
    0 301 (54.4) 272 (55.7) 295 (55.6) 777 (67.6) 741 (68.9) 710 (70.2) 
    1 224 (40.5) 188 (38.5) 208 (39.2) 327 (28.4) 303 (28.2) 267 (26.4) 
    ≥2 28 (5.1) 28 (5.7) 28 (5.3) 46 (4.0) 31 (2.9) 34 (3.4) 
Height (cm)       
    n 549 488 528 1140 1067 1001 
    Mean (SD) 157.8 (5.3) 157.6 (5.4) 157.5 (5.6) 159.5 (4.9) 159.4 (4.9) 159.4 (5.1) 
Weight (kg)       
    n 545 483 528 1121 1046 983 
    Mean (SD) 52.4 (6.1) 52.1 (6.8) 52.8 (6.7) 52.8 (6.0) 52.8 (6.1) 52.7 (5.7) 
BMI (kg/m2)       
    n 543 482 523 1121 1045 980 
    Mean (SD) 21.0 (2.3) 21.0 (2.4) 21.3 (2.5) 20.7 (2.1) 20.8 (2.2) 20.8 (2.0) 
Mid-upper arm circumfrance (cm)       
    n 544 483 528 1134 1060 1003 
    Mean (SD) 23.2 (2.2) 23.2 (2.2) 23.4 (2.1) 23.1 (2.0) 23.2 (2.1) 23.1 (2.1) 

Comparison of baseline characteristics of the women enrolled in the multi-micronutrient supplementation in pregnancy trial by household wealth index and treatment groups and all combined, Shaanxi Province, China 2002–05.

Values are number (%) of women (include multiple births) unless otherwise mentioned.

Impact of antennal nutrient supplementations on birth anthropometry

In the women from the poorest households, MMN supplements significantly increased the birth weight by 68 g (95% CI 4–131, P = 0.037) compared with the folic acid group. Similar results were found for MMN supplements when multiple births were included in the analysis although only with borderline significance (63 g, 95% CI −0.4 to −126, P = 0.051). Iron/folic acid increased birth weight by 55 g (95% CI −9 to −119, P = 0.092), but this difference with the folic acid group was not significant. In contrast, among women from the wealthier households neither the MMN supplements nor the iron/folic acid supplements had any effect on birth weight. However, as seen in Table 2, the interaction for birth weight was not statistically significant. There were no significant effects on birth length for iron/folic acid or MMN supplements during pregnancy compared with folic acid in the women from the poorest or wealthier households (Table 2).

Table 2

Mean birth weight, birth length, gestation weeks at birth by household wealth index and treatment group (just include singleton births)

   Comparing MMN or Iron/folic acid with folic acid alone
 
 
Household wealth index n Mean (SD) Adjusted difference (95% CI)a P values for interaction terms 
Birth weight (g)b    0.642 
    Poorest     
        Folic acid 465 3142 (464)   
        Iron/folic acid 426 3197 (419) 54.9 (−8.9 to 118.8)  
        MMN 446 3209 (414) 67.5 (4.2 to 130.7)  
    Wealthier     
        Folic acid 1050 3178 (422)   
        Iron/folic acid 1006 3186 (413) 7.2 (−32.9 to 47.3)  
        MMN 937 3208 (434) 29.6 (−9.9 to 69.2)  
Birth length (cm)c    0.776 
    Poorest     
        Folic acid 416 48.7 (3.0)   
        Iron/folic acid 363 49.0 (2.8) 0.34 (−0.12 to 0.79)  
        MMN 381 49.0 (2.6) 0.33 (−0.12 to 0.78)  
    Wealthier 
        Folic acid 927 49.0 (2.8)   
        Iron/folic acid 902 49.2 (2.6) 0.19 (−0.04 to 0.42)  
        MMN 847 49.1 (2.9) 0.16 (−0.14 to 0.45)  
Gestation at birth (weeks)    0.048 
    Poorest     
        Folic acid 524 39.5 (1.9)   
        Iron/folic acid 458 40.0 (1.6) 0.41 (0.18 to 0.65)  
        MMN 495 39.9 (1.7) 0.34 (0.11 to 0.56)  
    Wealthier     
        Folic acid 1112 39.7 (1.9)   
        Iron/folic acid 1041 39.8 (1.7) 0.11 (−0.04 to 0.26)  
        MMN 974 39.9 (1.7) 0.15 (−0.001 to 0.29)  
   Comparing MMN or Iron/folic acid with folic acid alone
 
 
Household wealth index n Mean (SD) Adjusted difference (95% CI)a P values for interaction terms 
Birth weight (g)b    0.642 
    Poorest     
        Folic acid 465 3142 (464)   
        Iron/folic acid 426 3197 (419) 54.9 (−8.9 to 118.8)  
        MMN 446 3209 (414) 67.5 (4.2 to 130.7)  
    Wealthier     
        Folic acid 1050 3178 (422)   
        Iron/folic acid 1006 3186 (413) 7.2 (−32.9 to 47.3)  
        MMN 937 3208 (434) 29.6 (−9.9 to 69.2)  
Birth length (cm)c    0.776 
    Poorest     
        Folic acid 416 48.7 (3.0)   
        Iron/folic acid 363 49.0 (2.8) 0.34 (−0.12 to 0.79)  
        MMN 381 49.0 (2.6) 0.33 (−0.12 to 0.78)  
    Wealthier 
        Folic acid 927 49.0 (2.8)   
        Iron/folic acid 902 49.2 (2.6) 0.19 (−0.04 to 0.42)  
        MMN 847 49.1 (2.9) 0.16 (−0.14 to 0.45)  
Gestation at birth (weeks)    0.048 
    Poorest     
        Folic acid 524 39.5 (1.9)   
        Iron/folic acid 458 40.0 (1.6) 0.41 (0.18 to 0.65)  
        MMN 495 39.9 (1.7) 0.34 (0.11 to 0.56)  
    Wealthier     
        Folic acid 1112 39.7 (1.9)   
        Iron/folic acid 1041 39.8 (1.7) 0.11 (−0.04 to 0.26)  
        MMN 974 39.9 (1.7) 0.15 (−0.001 to 0.29)  

Comparison of mean, SD and 95% CIs of birth weight (g), birth length (cm), gestation at births (weeks) of infants born to mothers enrolled in the nutrient supplementation trial by household wealth index and treatment groups, Shaanxi Province, China 2002–05.

aTaking into account cluster randomization with GEE linear models.

bMissing birth weight: 121 in folic acid treatment, 67 in folic acid/iron treatment, 86 in MMN treatment.

cMissing birth length: 293 in folic acid treatment, 234 in folic acid/iron treatment, 241 in MMN treatment.

The interaction for treatment group and household wealth was statistically significant (P = 0.045) for LBW (Table 3). Among the newborn infants from the poorest households randomized to the MMN supplements, there was a significant reduction in LBW of 60% (RR 0.40, 95% CI 0.21–0.78, P = 0.007) compared with the folic acid group. In the iron/folic acid group, there was a non-significant 43% reduction (RR 0.57, 95% CI 0.30–1.09, P = 0.087) in the rate of LBW compared with the folic acid group. When multiple births were included in the analysis, the results were similar with a significant reduction in the LBW (RR 0.57, 95% CI 0.33–0.98, P = 0.043) for MMN supplements compared with folic acid. As seen in Table 3, there were no effects on LBW across the MMN supplements or iron/folic acid groups among the newborns from the wealthier households.

Table 3

LBW, small for gestational age, pre-term and early pre-term delivery by household wealth index and treatment group (just include singleton births)

Household wealth index n (%) Comparing MMN or Iron/folic acid with folic acid alone
 
P-value for interaction terms 
  RR (95% CI)a  
LBW (<2500 g)   0.045 
    Poorest    
        Folic acid 29 (6.2)   
        Iron/folic acid 15 (3.5) 0.57 (0.30–1.09)  
        MMN 11 (2.5) 0.40 (0.21–0.78)  
    Wealthier    
        Folic acid 42 (4.0)   
        Iron/folic acid 38 (3.8) 0.94 (0.62–1.43)  
        MMN 35 (3.7) 0.91 (0.59–1.41)  
Small for gestational ageb   0.585 
    Poorest    
        Folic acid 92 (19.8)   
        Iron/folic acid 79 (18.5) 0.95 (0.71–1.26)  
        MMN 74 (16.6) 0.85 (0.64–1.14)  
    Wealthier    
        Folic acid 173 (16.5)   
        Iron/folic acid 179 (17.8) 1.09 (0.90–1.33)  
        MMN 156 (16.7) 1.03 (0.84–1.26)  
Pre-term (<37 weeks)   0.032 
    Poorest    
        Folic acid 42 (8.0)   
        Iron/folic acid 20 (4.4) 0.55 (0.32–0.93)  
        MMN 22 (4.4) 0.56 (0.36–0.86)  
    Wealthier    
        Folic acid 53 (4.8)   
        Iron/folic acid 50 (4.8) 1.01 (0.70–1.45)  
        MMN 45 (4.6) 0.97 (0.66–1.42)  
Early pre-term (<34 weeks)   0.589 
    Poorest    
        Folic acid 10 (1.91)   
        Iron/folic acid 3 (0.66) 0.36 (0.11–1.22)  
        MMN 4 (0.81) 0.46 (0.15–1.36)  
    Wealthier    
        Folic acid 15 (1.35)   
        Iron/folic acid 10 (0.96) 0.71 (0.33–1.52)  
        MMN 8 (0.82) 0.60 (0.25–1.46)  
Household wealth index n (%) Comparing MMN or Iron/folic acid with folic acid alone
 
P-value for interaction terms 
  RR (95% CI)a  
LBW (<2500 g)   0.045 
    Poorest    
        Folic acid 29 (6.2)   
        Iron/folic acid 15 (3.5) 0.57 (0.30–1.09)  
        MMN 11 (2.5) 0.40 (0.21–0.78)  
    Wealthier    
        Folic acid 42 (4.0)   
        Iron/folic acid 38 (3.8) 0.94 (0.62–1.43)  
        MMN 35 (3.7) 0.91 (0.59–1.41)  
Small for gestational ageb   0.585 
    Poorest    
        Folic acid 92 (19.8)   
        Iron/folic acid 79 (18.5) 0.95 (0.71–1.26)  
        MMN 74 (16.6) 0.85 (0.64–1.14)  
    Wealthier    
        Folic acid 173 (16.5)   
        Iron/folic acid 179 (17.8) 1.09 (0.90–1.33)  
        MMN 156 (16.7) 1.03 (0.84–1.26)  
Pre-term (<37 weeks)   0.032 
    Poorest    
        Folic acid 42 (8.0)   
        Iron/folic acid 20 (4.4) 0.55 (0.32–0.93)  
        MMN 22 (4.4) 0.56 (0.36–0.86)  
    Wealthier    
        Folic acid 53 (4.8)   
        Iron/folic acid 50 (4.8) 1.01 (0.70–1.45)  
        MMN 45 (4.6) 0.97 (0.66–1.42)  
Early pre-term (<34 weeks)   0.589 
    Poorest    
        Folic acid 10 (1.91)   
        Iron/folic acid 3 (0.66) 0.36 (0.11–1.22)  
        MMN 4 (0.81) 0.46 (0.15–1.36)  
    Wealthier    
        Folic acid 15 (1.35)   
        Iron/folic acid 10 (0.96) 0.71 (0.33–1.52)  
        MMN 8 (0.82) 0.60 (0.25–1.46)  

Percentage of Infants born to mothers enrolled in the nutrient supplementation trial with LBW (<2500 g), small for gestational age (<10th percentile), pre-term delivery (<37 weeks gestation), early pre-term delivery (<34 weeks gestation) and the 95% CIs, by household wealth index and treatment group.

aAdjusted for cluster randomization in GEE binomial models.

bDefined as newborns with weight below 10th centile of gestational age-sex-specific US reference for fetal growth.14

Impact of antenatal micronutrient supplements on duration of gestation at birth

The interaction for treatment group and household wealth was statistically significant (P = 0.048) for duration of gestation. In the women from the poorest households, there were significant increases in the duration of gestation at birth with iron/folic acid (0.41 weeks, 95% CI 0.18–0.65, P = 0.001) and MMN (0.34 weeks, 95% CI 0.11–0.56, P = 0.004) supplements compared with folic acid (Table 2). When multiple births were included in the analysis, the results were similar with significant increases in the duration of gestation at birth with iron/folic acid (0.45 weeks, 95% CI 0.21–0.69, P < 0.001) and MMN (0.29 weeks, 95% CI 0.04–0.54, P = 0.022) supplements compared with folic acid. There was no effect of either MMN or iron folic acid supplements on duration of gestation in women from the wealthier households (Table 2).

The interaction for treatment group and household wealth was statistically significant (P = 0.032) for pre-term delivery. There were significant reductions in the rate of pre-term delivery for women from the poorest households for both iron/folic acid (RR 0.55, 95% CI 0.32–0.93, P = 0.027) and MMN supplements (RR 0.56, 95% CI 0.36–0.86, P = 0.009) compared with folic acid (Table 3). Rates of early pre-term delivery were also reduced for iron/folic acid and MMN supplements compared with folic acid in women from the poorest households, but these differences were not statistically significant. When multiple births were included in the analysis, there was a significant reduction in the rate of pre-term delivery for women from the poorest households for iron/folic acid (RR 0.56, 95% CI 0.33–0.96, P = 0.035) but not for MMN supplements (RR 0.72, 95% CI 0.47–1.11). As seen in Table 3, there were no differences in the rate of pre-term births in women from the wealthier households across the treatment groups.

Both iron/folic acid and MMN supplements had no effect on the rate of small for gestational age (SGA) births in both the poorest and wealthier household compared with folic acid alone.

Impact of micronutrient supplementation on perinatal mortality

There was no impact on the risk of perinatal mortality either for the iron/folic acid or MMN supplements compared with folic acid among infants from the poorest or wealthier household (Table 4). In the poorest households, there were no significant increases in the stillbirth rate for iron /folic acid or MMN supplements, although the RRs were increased relative to the folic acid group.

Table 4

Mortality outcomes (including only singleton babies) by household wealth index and treatment group

 Household wealth index Folic acid
 
Iron/folic acid
 
MMN
 
Iron/folic acid vs folic acid MMN vs folic acid P values for interaction term 
n Rate/ n Rate/ n Rate/ RR (95% CI) RR (95% CI) 
1000 1000 1000 
Pregnancies with single live or still birth           
Stillbirths (≥28 weeks)a          0.009 
 Poorest 22 40.3 23 47.8 32 60.7 1.24 (0.69–2.23) 1.56 (0.87–2.80)  
 Wealthier 30 26.3 24 22.5 31 30.8 0.86 (0.50–1.49) 1.21 (0.74–1.97)  
Live births           
 Poorest 524  458  495     
 Wealthier 1112  1041  974     
Neonatal deathsa          0.385 
 Poorest 14 26.7 8.7 14.1 0.33 (0.11–1.04) 0.53 (0.21–1.34)  
 Wealthier 19 17.1 12 11.5 11 11.3 0.69 (0.34–1.38) 0.65 (0.32–1.35)  
Early neonatal deathsa          0.147 
 Poorest 11 21.0 2.2 10.1 0.10 (0.01–0.79) 0.48 (0.17–1.36)  
 Wealthier 13 11.7 8.6 10 10.3 0.75 (0.30–1.84) 0.89 (0.39–2.02)  
Perinatal deathsa          0.021 
 Poorest 33 60.4 24 49.9 37 70.2 0.84 (0.50–1.40) 1.19 (0.73–1.93)  
 Wealthier 43 37.7 33 31.0 41 40.8 0.83 (0.53–1.32) 1.12 (0.72–1.73)  
 Household wealth index Folic acid
 
Iron/folic acid
 
MMN
 
Iron/folic acid vs folic acid MMN vs folic acid P values for interaction term 
n Rate/ n Rate/ n Rate/ RR (95% CI) RR (95% CI) 
1000 1000 1000 
Pregnancies with single live or still birth           
Stillbirths (≥28 weeks)a          0.009 
 Poorest 22 40.3 23 47.8 32 60.7 1.24 (0.69–2.23) 1.56 (0.87–2.80)  
 Wealthier 30 26.3 24 22.5 31 30.8 0.86 (0.50–1.49) 1.21 (0.74–1.97)  
Live births           
 Poorest 524  458  495     
 Wealthier 1112  1041  974     
Neonatal deathsa          0.385 
 Poorest 14 26.7 8.7 14.1 0.33 (0.11–1.04) 0.53 (0.21–1.34)  
 Wealthier 19 17.1 12 11.5 11 11.3 0.69 (0.34–1.38) 0.65 (0.32–1.35)  
Early neonatal deathsa          0.147 
 Poorest 11 21.0 2.2 10.1 0.10 (0.01–0.79) 0.48 (0.17–1.36)  
 Wealthier 13 11.7 8.6 10 10.3 0.75 (0.30–1.84) 0.89 (0.39–2.02)  
Perinatal deathsa          0.021 
 Poorest 33 60.4 24 49.9 37 70.2 0.84 (0.50–1.40) 1.19 (0.73–1.93)  
 Wealthier 43 37.7 33 31.0 41 40.8 0.83 (0.53–1.32) 1.12 (0.72–1.73)  

aStillbirths, neonatal deaths, early neonatal deaths, perinatal deaths: RR (95% CIs) adjusted for cluster randomization with GEE binomial model.

However, in women from the poorest households, there was a large reduction in the risk of early neonatal mortality in the iron/folic acid group (90% reduction, RR 0.10, 95% CI 0.01–0.79) and a non-significant reduction in MMN supplements (52% reduction, RR 0.48, 95% CI 0.17-1.36) compared with folic acid, although the interaction for treatment group and household wealth was not statistically significant (P = 0.147) for early neonatal deaths (Table 4). When tertiles for household wealth were used, there was a progressive increase in the protective effect of iron/folic acid supplements for early neonatal mortality from no protection for women from wealthiest households (RR 0.96, 95% CI 0.22–4.15), to some protection for women from middle-level households (RR 0.55, 95% CI 0.18–1.73) to substantial protection for infants of women from the poorest households (RR 0.10, 95% CI 0.01–0.79). When multiple births were included in the analysis, the risk of early neonatal mortality in the iron/folic acid group was reduced (75% reduction, RR 0.25, 95% CI 0.06–1.14).

In the wealthier households, there were no significant differences in the stillbirth rate or the risk of early neonatal mortality for iron/folic acid or MMN supplements compared with folic acid (Table 4).

Discussion

Summary of key findings

In this trial, the responses to antenatal micronutrient supplementation were strongly modified by the level of wealth of the trial participants’ household with the largest impact of the supplements on the trial outcomes in women from the poorest households, but with no effect in women from the wealthier households. In women from the poorest households, MMN supplements compared with folic acid significantly increased birth weight (Table 2) and reduced LBW (Table 3). Although there were similar effects for iron/folic acid, they were not significant. Also in the poorest women, there was a significant increase in the mean duration of gestation of pregnancy for both iron/folic acid and MMN supplements (Table 2) compared with folic acid, and a significant reduction in pre-term delivery for iron/folic acid and MMN supplements (Table 3). Household wealth also strongly modified the effects on neonatal mortality. In women from the poorest households, there was a statistically significant 90% reduction in early neonatal mortality with iron/folic acid supplementation, and a 52% non-significant reduction for MMN supplements, but no mortality-sparing effects for the women from the wealthier households. All these key findings were not altered by the other factors examined in subgroup analyses, such as the timing of initiation of supplementation.

Significance of study findings

This trial reported important differences in the response to micronutrient supplementation in pregnancy by the economic status of the women’s household. It demonstrates the important role micronutrient supplementation in pregnancy can play to increase birth weight, duration of gestation and to reduce neonatal mortality in women from disadvantaged households. Poorer women have less adequate diets;20 greater micronutrient deficiencies;20 less access to information about how to stay healthy in pregnancy;21 and less access to healthcare services.21 These factors combined make these women and their babies more vulnerable to adverse outcomes. Antenatal micronutrient supplementation is a simple intervention that can be delivered through existing healthcare structures and is likely to be highly cost effective. Our findings highlight the need to target this intervention to poorer women to maximize reductions in neonatal mortality and help accelerate reaching the child survival MDG.

Strengths and limitations of study

Although the study was located in a socio-economically disadvantaged area of western China, there was a surprising amount of heterogeneity in household wealth as illustrated in the range of wealth index values reported across the groups examined (Table 1). The trial subjects included women living in county towns located on a major highway that links these communities to the rapidly expanding Chinese economy, through to women living in households without electricity and in villages with limited road access. The large range of household wealth in the study population has helped highlight the differences in the effects of the interventions across groups of women of widely differing economic status. Finally, the wealth index used in the analysis was based on an internationally recommended method.16 This asset-based index has been validated using income and expenditure data for a sample household from the study counties. In this validation study, the median household annual income for the poorest quintile of households classified by the wealth index was RMB 1111 and this steadily increased with each quintile to RMB 2400 for the wealthiest quintile and this trend was statistically significant (P < 0.001).17

The unique situation in China in which current antenatal care programmes do not distribute iron/folic acid to women in pregnancy allowed a design with a control group with folic acid alone. This design has led to more valid comparisons of the effects of different combinations of micronutrients on a range of pregnancy and birth outcomes, and contributes to more effective observation of the different responses in the household wealth groups.

The double-blind cluster randomized controlled design of the trial is another strength of this analysis. Although the findings presented are from post hoc, subgroup analyses by household wealth, the randomization procedure has produced balanced baseline characteristics across each household wealth stratum (Table 1). The design of the randomization, with stratification by county and blocking to ensure balance of treatments by geographical area within counties, accounts for this balanced distribution of confounders across treatment groups in each household wealth stratum. The large number of village clusters randomized has also contributed to ensuring the balance of treatment groups by household wealth. Our analyses have used appropriate statistical methods to adjust for the cluster randomization.

A limitation of the study was the low power to detect changes in the prevalence of LBW, pre-term delivery and perinatal mortality events, especially in the subgroup analyses. Although important statistically significant effects were found in many subgroups, the small sample size often meant the CIs were wide and there was uncertainty about the precision of the point estimates. For example, we found a 90% reduction in early neonatal mortality in poorest women given iron/folic acid compared with folic acid alone, but the CIs ranged from a 21% to 99% reduction. Also the rate of neonatal mortality in this group of women (2.2/1000 live births) was lower than in the women from the wealthiest households (9.4/1000 live births) whose rate was similar to that in high-income countries (1–11/1000 live births).1 The neonatal mortality rate for the iron/folic acid-treated poor women was lower than expected and suggested the magnitude of effect might have been overestimated. The low power also limits the examination of the effects of the different supplements on mortality in the household wealth subgroups. For example, there was a 52% reduction in early neonatal mortality with MMN supplementation in poorest women, but this effect was not significant, although it might have been significant with a larger sample. Despite these limitations, there are clear patterns of effect across the household wealth groups, for example, with the steadily increasing size of the protective effect of iron/folic acid for early neonatal mortality for women from progressively poorer households. There are similar patterns evident for MMN supplementation, although none of the effects was statistically significant.

The lack of baseline micronutrient status of the trial participants also restricts our interpretation of the reasons for the greater effects in women from poorest households. Were the differences in response by household wealth due to different levels of iron and other micronutrient deficiencies in the women from the poorest compared with the wealthiest households?

Comparisons with other studies

There are no reports of trials of nutrition interventions, including trials of responses to micronutrient supplementation in pregnancy, in which the modifying effects of household economic status of the participants have been examined. There is some evidence of differential responses to antenatal micronutrient supplementation in studies conducted in different countries that might be related to variations in economic status of the study populations. Larger effects on birth weight have been reported for multi-micronutrient vs iron/folic acid supplementation in pregnancy in poorer countries such as Nepal (75 g difference),22 Burkina Faso (31 g difference)23 and Niger (54 g difference)24 than in wealthier countries such as China (17 g difference),13 Indonesia (15 g difference)25 and Mexico (16 g difference).26 These trials, apart from our study in China, have not reported subgroup analyses based on the household economic status of the participants, and they have no economic measurements that allow direct comparisons of the studies stratified by economic status of the population.

Possible mechanisms and implications for policymakers

Our study is unable to determine the exact mechanisms that account for the larger impact of the antenatal micronutrient supplements in women from poorest households. It is likely that these women have more severe micronutrient deficiencies, but the lack of baseline biochemical assessment prevents confirmation of this as the likely mechanism. We did assess dietary intake in a subsample of the trial population and found that overall 64% of the pregnant women had an inadequate iron intake.14 The trial participants also had high phytate/iron ratios indicating poor bioavailability of the iron. Furthermore, the women from the poorest households had significantly lower intakes of calcium, iron, zinc, riboflavin and vitamin C than women from wealthier households.14 These dietary intake data provide evidence of lower micronutrient intakes in women from poorer households and suggest why their responses to supplementation may have been greater. We have also reported that women in the trial from poorer households had higher rates of home delivery, with 29% of women from the lowest quintile of household wealth delivering at home without trained birth attendants to only 6% for women from the highest quintile of household wealth.17 Thus, by preventing pre-term delivery and improving birth weight in the group of women with the least access to trained birth attendants, the micronutrient supplements especially iron/folic acid led to a substantial reduction in neonatal mortality in women from the poorest households.

Unanswered questions and future research

Our results highlight the importance of collecting information about the economic status of the households of participants in future trials of nutrition interventions to prevent malnutrition in women and children. Although the need to prioritize nutrition and environmental interventions towards the poor to improve child survival has been reported,27 the differential response to nutrition interventions in poor households has not been carefully considered in part because of the lack of data. A better understanding of the extent to which nutrition interventions have a greater impact among the poor is vital to provide evidence to policymakers on how best to maximize the impact of these interventions.

Conclusions

We have demonstrated that household wealth modifies the perinatal responses to micronutrient supplements in pregnant women, in a rural population in China. The women from the poorest households had larger responses of perinatal outcomes to antenatal micronutrient supplementations than women from the wealthiest households. Micronutrient supplementation in pregnancy in rural China, especially among the poorest women, has an important role to play in improving perinatal health outcomes. The standard recommended iron/folic acid supplements appear to provide more protection for neonatal survival than MMN supplements.

Funding

The project was supported by two grants: one from the United Nations Children’s Fund (grant no. YH101-H12/03) through a cooperative agreement between UNICEF and the Centers for Disease Control and Prevention, Atlanta, USA; the other from the National Natural Science of Foundation of China (grant no. 30271131), Beijing, China. The nutrient supplement premix was provided by Sight and Life and DSM Nutritional Products, China, and manufactured by the Beijing Vita Nutritious Products, Beijing, China. The investigators remained independent of the organizations funding the project including all decisions related to this publication.

Acknowledgements

We extend our thanks to the following members of the Shaanxi study team who contributed to the successful implementation of the trial: Dang Shaonong, Kang Yijun, Fang Bo and Liang Weifeng (field supervisors); and the area coordinators of the Maternal and Child Healthcare Station and Health Bureau in Changwu and Bin counties. Ray Yip, Mary Cogswell, Zuoguo Mei, Suying Chang and Lingzhi Kong provided useful comments about the study design and field procedures. Li Qiang, Wang Quanli, Xiao Shengbin, Xing Yuan, Zhou Xiaoyan, Wang Bei and Xie Hong helped with data management. M.J.D. was a paid consultant for UNICEF China during the conduct of the trial, and during the same period, as a consultant for UNICEF Pyongyang.

Conflict of interest: None declared.

KEY MESSAGES

  • Household wealth modifies the perinatal responses to micronutrient supplements in pregnant women in a rural population in China.

  • The women from the poorest households had larger responses of perinatal outcomes to antenatal micronutrient supplementations than women from the wealthier households.

  • The standard recommended iron/folic acid supplements appear to provide more protection for neonatal survival than MMN supplements in rural China.

References

1
 
WHO. The World Health Report 2005: Making Every Mother and Child Count. 2005
2
Darmstadt
GL
Lawn
JE
Costello
A
Advancing the state of world's newborn
Bull WHO
 , 
2003
, vol. 
81
 (pg. 
224
-
25
)
3
Villar
J
Keirse
MJNC
Pre-term delivery syndrome: the unmet need
New Perspectives for the Effective Treatment of Pre-term Labour. An International Consensus. Research & Clinical Forum
 , 
1994
UK
Wells Medical
(pg. 
9
-
33
)
4
UNICEF and WHO
Low Birthweight: Country, Regional and Global Estimates
 , 
2004
New York
UNICEF
5
Fall
CHD
Yajnik
CS
Rao
S
Davies
AA
Brown
N
Farrant
HJW
Micronutrients and fetal growth
J Nutr
 , 
2003
, vol. 
133
 (pg. 
1747S
-
56S
)
6
Yip
R
Iron
In: Present Knowledge in Nutrition
 , 
2001
8th edn
Washington, DC
International Life Science Institute-Nutrition Foundation
7
Yip
R
Significance of an abnormally low or high hemoglobin concentration during pregnancy: special consideration of iron nutrition
Am J Clin Nutr
 , 
2000
, vol. 
72
 (pg. 
272S
-
79S
)
8
Rasmussen
K
Is there a causal relationship between iron deficiency or iron-deficiency anemia and weight at birth, length of gestation and perinatal mortality?
J Nutr
 , 
2001
, vol. 
131
 (pg. 
590S
-
601S
)
9
Zulfi qar
A
Bhutta
TA
Black
RE
, et al.  . 
for the Maternal and Child Undernutrition Study Group
What works? Interventions for maternal and child undernutrition and survival
Lancet
 , 
2008
, vol. 
371
 (pg. 
417
-
40
)
10
Rudan
I
Chan
KY
Zhang
JS
, et al.  . 
Causes of deaths in children younger than 5 years in China in 2008
Lancet
 , vol. 
375
 (pg. 
1083
-
89
)
11
Lin
L
Liu
Y
Zhang
X
Mi
J
Cao
L
Sampling survey on low-birth weight in China in 1998
Zhonghua Yu Fang Yi Xue Za Zhi
 , 
2002
, vol. 
36
 (pg. 
149
-
53
)
12
Institute of Pediatrics Capital
Survey of women anaemia status at child-bearing age in China in 1998
Zhong Guo Sheng Yu Jian Kang Za Zhi
 , 
2002
, vol. 
13
 (pg. 
102
-
8
)
13
Zeng
LX
Cheng
Y
Dang
SN
, et al.  . 
Impact of micronutrient supplementation during pregnancy on birth weight, duration of gestation, and perinatal mortality in rural western China: double blind cluster randomised controlled trial
BMJ
 , 
2008
, vol. 
337
 (pg. 
1211
-
15
)
14
Cheng
Y
Dibley
MJ
Zhang
X
Zeng
L
Yan
H
Assessment of dietary intake among pregnant women in a rural area of western China
BMC Public Health
 , 
2009
, vol. 
9
 pg. 
222
 
15
Alexander
GR
Himes
JH
Kaufman
RB
Mor
J
Kogan
M
A United States national reference for fetal growth
Obstet Gynecol
 , 
1996
, vol. 
87
 (pg. 
163
-
68
)
16
Filmer
D
Pritchett
LH
Estimating wealth effects without expenditure data—or tears: an application to educational enrollments in states of India
Demography
 , 
2001
, vol. 
38
 (pg. 
115
-
32
)
17
Zeng
L
Yan
H
Chen
Z
Measurement of the living standards of family in rural area and relationship between index and perinatal care status
J Hyg Res
 , 
2008
, vol. 
37
 (pg. 
714
-
17
)
18
Wang
R
Lagakos
SW
Ware
JH
Hunter
DJ
Drazen
JM
Statistics in medicine—reporting of subgroup analyses in clinical trials
N Engl J Med
 , 
2007
, vol. 
357
 (pg. 
2189
-
94
)
19
Lagakos
SW
The challenge of subgroup analyses—reporting without distorting
N Engl J Med
 , 
2006
, vol. 
354
 (pg. 
1667
-
69
)
20
Black
RE
Allen
LH
Bhutta
ZA
, et al.  . 
Maternal and child undernutrition: global and regional exposures and health consequences
Lancet
 , 
2008
, vol. 
371
 (pg. 
243
-
60
)
21
Tang
S
Meng
Q
Chen
L
Bekedam
H
Evans
T
Whitehead
M
Tackling the challenges to health equity in China
Lancet
 , 
2008
, vol. 
372
 (pg. 
1493
-
501
)
22
Christian
P
Khatry
SK
Katz
J
, et al.  . 
Effects of alternative maternal micronutrient supplements on low birth weight in rural Nepal: double blind randomised community trial
BMJ
 , 
2003
, vol. 
326
 pg. 
571
 
23
Roberfroid
D
Huybregts
L
Lanou
H
, et al.  . 
Effects of maternal multiple micronutrient supplementation on fetal growth: a double-blind randomized controlled trial in rural Burkina Faso
Am J Clin Nutr
 , 
2008
, vol. 
88
 (pg. 
1330
-
40
)
24
Zagre
NM
Desplats
G
Adou
P
Mamadoultaibou
A
Aguayo
VM
Prenatal multiple micronutrient supplementation has greater impact on birthweight than supplementation with iron and folic acid: a cluster-randomized, double-blind, controlled programmatic study in rural Niger
Food Nutr Bull
 , 
2007
, vol. 
28
 (pg. 
317
-
27
)
25
Shankar
AH
Jahari
AB
Sebayang
SK
, et al.  . 
Effect of maternal multiple micronutrient supplementation on fetal loss and infant death in Indonesia: a double-blind cluster-randomised trial
Lancet
 , 
2008
, vol. 
371
 (pg. 
215
-
27
)
26
Ramakrishnan
U
Gonzâalez-Cossâio
T
Neufeld
LM
Rivera
J
Martorell
R
Multiple micronutrient supplementation during pregnancy does not lead to greater infant birth size than does iron-only supplementation: a randomized controlled trial in a semirural community in Mexico
Am J Clin Nutr
 , 
2003
, vol. 
77
 (pg. 
720
-
25
)
27
Gakidou
E
Oza
S
Vidal Fuertes
C
, et al.  . 
Improving child survival through environmental and nutritional interventions: the importance of targeting interventions toward the poor
JAMA
 , 
2007
, vol. 
298
 (pg. 
1876
-
87
)