Effects of prenatal and postnatal maternal multiple micronutrient supplementation on child growth and morbidity in Tanzania: a double-blind, randomized–controlled trial

Abstract

Background

Maternal micronutrient status is critical for child growth and nutrition. It is unclear whether maternal multiple micronutrient supplementation (MMS) during pregnancy and lactation improves child growth and prevents child morbidity.

Methods

This study aimed to determine the effects of prenatal and postnatal maternal MMS on child growth and morbidity. In this double-blind, randomized–controlled trial, 8428 HIV-negative pregnant women were enrolled from Dar es Salaam, Tanzania, between 2001 and 2004. From pregnancy (12–27 weeks of gestation) through to 6 weeks postpartum, participants were randomized to receive daily oral MMS or placebo. All women received daily iron and folic acid during pregnancy. From 6 weeks postpartum through to 18 months postpartum, 3100 women were re-randomized to MMS or placebo. Child-growth measures, haemoglobin concentrations and infectious morbidities were assessed longitudinally from birth to ≤18 months.

Results

Prenatal MMS led to modest increases in weight-for-age z-scores (mean difference: 0.050; 95% confidence interval: 0.002, 0.099; p = 0.04) and length-for-age z-score (mean difference: 0.062; 95% confidence interval: 0.013, 0.111; p = 0.01) during the first 6 months of life but not thereafter. Prenatal or postnatal MMS did not have benefits for other child outcomes.

Conclusions

Whereas maternal MMS is a proven strategy to prevent adverse birth outcomes, other approaches may also need to be considered to curb the high burdens of child morbidity and growth faltering.

Key Messages

• It is unclear whether maternal multiple micronutrient supplementation (MMS) during pregnancy and lactation improves child growth and prevents child morbidity.

• We did not find substantial evidence of benefits of prenatal or postnatal maternal MMS on a wide range of child outcomes.

• Alternative strategies other than supplementing mothers with micronutrients may need to be considered to curb the high burdens of child morbidity and growth faltering.

Introduction

The period from conception to the second birthday (i.e. the first 1000 days of life) is commonly referred to as the ‘window of opportunity’ for targeting undernutrition. Adequate nutrition and growth during this period are vital to the development and health of children and their ability to achieve full potential as adults. There is a high burden of child malnutrition in low- and middle-income countries (LMICs). In Tanzania, ∼3 million children under 5 years of age were stunted and 90 000 had severe acute malnutrition in 2018.¹ Adequate maternal intake of micronutrients is critical for a healthy pregnancy and the optimal nutrition and growth of the child. With the increased nutritional demands during pregnancy and breastfeeding, maternal micronutrient deficiencies are prevalent in LMICs.^2,3 These deficiencies adversely impact not only maternal health, but also child growth and development.

Iron and folic acid (IFA) supplementation has been the standard of antenatal care for preventing maternal anaemia and adverse pregnancy outcomes⁴ and is also recommended for ≥3 months during the postpartum period.⁵ Malnourished pregnant and lactating women are often deficient in more than one micronutrient. Therefore, multiple micronutrient supplementation (MMS), with three or more micronutrients (including iron and folic acid) provided simultaneously, may confer greater benefits for mothers and children, and is increasingly recognized as a better alternative to IFA supplements.^6,7 A consistent body of literature suggests that, compared with IFA, prenatal MMS reduces the risks of small for gestational age and low birthweight, and possibly prevents preterm births and stillbirths.^7,8 The World Health Organization (WHO) recently recommended the use of prenatal MMS in the context of rigorous research.⁹ In addition to the short-term benefits for pregnancy outcomes, prenatal MMS may also improve child growth and nutrition through early fetal programming.¹⁰ There is, however, conflicting evidence on the impacts of prenatal MMS on child growth^11–16 and few studies have evaluated the impacts of prenatal MMS on infectious morbidities during early childhood.^17–19

Breast milk is the sole source of micronutrients during the first 6 months when the infants are recommended to be exclusively breastfed and continues to provide important nutritional support thereafter. For mothers with low micronutrient stores, breast milk may not provide adequate levels of micronutrients to the child and maternal micronutrient supplementation may be beneficial. Limited evidence exists on the impacts of postnatal maternal MMS on the child.^6,20,21 No randomized–controlled trials (RCTs) have examined the impacts of postnatal MMS on child growth or morbidity outcomes.

Eliminating maternal micronutrient deficiencies is critical for the health of women and their offspring. A better understanding of the impacts of maternal micronutrient supplementation during pregnancy and breastfeeding is critical for the design of nutritional interventions in resource-limited settings. Therefore, we aimed to determine the effects of prenatal and postnatal maternal MMS on child growth and morbidity in Tanzania.

Methods

Study design and study population

This study used data from a randomized, double-blind, placebo-controlled trial of prenatal and postnatal maternal MMS in Tanzania. The primary aim of the parent study was to evaluate the effects of prenatal MMS on the primary outcomes of fetal loss, low birthweight and preterm birth. The findings of the prenatal phase of the study with the primary outcome measures have been described elsewhere.²²

Pregnant women who attended antenatal clinics in Dar es Salaam, Tanzania, were recruited between 2001 and 2004. All women were confirmed to be seronegative for HIV infection based on antibody tests. The study enrolled 8428 eligible women, of whom 49 died or were lost to follow-up by the time of delivery. Among the remaining 8379 women with data on birth outcomes, 156 were pregnant with twins or triplets, resulting in 8223 women pregnant with singletons. From recruitment during the second trimester (12–27 weeks of gestation) through 6 weeks postpartum, women were randomly assigned to receive daily oral MMS or placebo. The MMS supplements included 20 mg of vitamin B-1 (thiamine), 20 mg of vitamin B-2 (riboflavin), 100 mg of vitamin B-3 (niacin), 25 mg of vitamin B-6 (pyridoxine), 0.8 mg of folic acid, 50 μg of vitamin B-12 (cobalamin), 500 mg of vitamin C and 30 mg of vitamin E. The doses were twice the recommended dietary allowance for vitamin E and 6–10 times for vitamin C and B vitamins. The active tablets and placebo were packaged in identical coded bottles and were similar in shape, size and colour. A randomization list was prepared based on a randomization sequence in blocks of 20. At enrolment, each eligible woman was assigned to the next numbered bottle. Participants and study staff who collected the data were unaware of the intervention assignments. At every monthly visit, a new bottle was given to each woman and the pills remaining in the used bottles were counted. Compliance with the assigned regimen was calculated as the number of tablets absent from the returned bottles divided by the total number of tablets that the participant should have taken. The median compliance was 96% with no discernible difference in compliance between the treatment arms.²² Per Tanzania’s standard of antenatal care at the time, all women were given daily doses of 60 mg of elemental iron and 0.25 mg of folic acid during pregnancy, irrespective of the assigned regimen. All women were also provided with malaria prophylaxis in the form of sulfadoxine–pyrimethamine tablets at 20 and 30 weeks of gestation.

Women who returned to the clinic for a postnatal visit between 4 and 10 weeks postpartum, had a live child between 4 and 10 weeks of age at that visit and provided additional informed consent were re-randomized to receive postnatal MMS (with the same composition as in the prenatal phase) or placebo through to 18 months postpartum. Therefore, among the 3100 participants included in the postpartum phase, the study effectively employed a two-by-two factorial design with prenatal MMS as one treatment factor and postnatal MMS as the other. As the study was conducted prior to the WHO recommendation of postpartum IFA, the postnatal control group received placebo without active ingredients per Tanzania’s standard of postnatal care at the time.

Data collection

Women (before birth) and mother–child dyads (after birth) had monthly study visits during which nutritional and health outcomes were assessed until ≤18 months of age. Child-anthropometry measures, including weight, length, head circumference and mid-upper arm circumference (MUAC), were collected by trained study nurses using standardized instruments and based on a standard operating procedure (Supplementary Box S1, available as Supplementary data at IJE online). Study nurses also assessed symptoms of child morbidities present on the day of the visit and during the past month. At birth and intermittently during the postnatal follow-up visits, child blood samples were collected and haemoglobin concentrations were measured. Every 3 months, study physicians conducted a scheduled physical examination and provided diagnoses and necessary medical interventions. Study physicians also examined and treated children when nurses or parents noted acute illnesses.

Definitions of child outcomes

The child-growth and nutrition outcomes were (i) weight-for-age z-score (WAZ), (ii) length-for-age z-score (LAZ), (iii) weight-for-length z-score (WLZ), (iv) head circumference-for-age z-score (HCZ), (v) arm-circumference-for-age z-score (ACZ) and (vi) haemoglobin concentrations. We computed all z-scores based on the WHO 2006 Child Growth Standards²³ using the WHO’s SAS macro.²⁴ Child-anthropometry measures that were outliers were set to missing based on the WHO flagging system, defined as WAZ < –6 or > 5, LAZ < –6 or > 6, WLZ < –5 or > 5, HCZ < –5 or > 5 and ACZ < –5 or > 5.²⁵ We treated the z-scores and haemoglobin concentrations as continuous variables. The continuous z-scores were also dichotomized as underweight (WAZ < –2), stunting (LAZ < –2), wasting (WLZ < –2), small head circumference (HCZ < –2) and small arm circumference (ACZ < –2). We defined severe acute malnutrition (SAM) as WLZ < –3 or MUAC < 115 mm,²⁶ moderate acute malnutrition (MAM) as WLZ between –2 and –3 or MUAC between 115 and <125 mm²⁷ and global acute malnutrition (GAM) as either SAM or MAM.²⁸

Based on the WHO definitions of anaemia for children aged 6–59 months,²⁹ any anaemia was defined as haemoglobin <11 g/dL, mild anaemia as haemoglobin between of 10–10.9 g/dL, moderate anaemia as haemoglobin of 7–9.9 g/dL and severe anaemia as haemoglobin of <7 g/dL.

The nurse-assessed child-morbidity outcomes include diarrhoea, cough, fever and common cold, assessed through monthly nurse visits. The physician-diagnosed child-morbidity outcomes include acute diarrhoea, acute upper respiratory infection (AURI), acute lower respiratory infection (ALRI) and acute respiratory infection (ARI) assessed through scheduled (every 3 months) or unscheduled (during acute illness episodes) physician visits and persistent diarrhoea assessed through scheduled physician visits (Supplementary Box S1, available as Supplementary data at IJE online).

Statistical analysis

We performed intention-to-treat analyses using assigned intervention and child age as the explanatory variables. For all analyses, we excluded children without any outcome data. We analysed the effects of prenatal and postnatal MMS in separate models, using prenatal and postnatal controls as the reference groups, respectively. For analyses on the effects of postnatal MMS, we additionally excluded children whose mothers were not included in the postnatal phase of the study and restricted to child outcome measures collected after the mother had started the postnatal regimen.

We used linear mixed-effects models to evaluate the effects of prenatal and postnatal MMS on continuous outcomes, including haemoglobin concentrations and child-growth z-scores. The models included a fixed effect for the intervention, fixed effects for indicator variables for time periods of child age (>6 to 12 and >12 to 18 months, with $\le$6 months as the reference), interaction terms between intervention and time periods of child age, random intercepts and random effects for time periods. An independent (variance component) covariance matrix was used. Although using the discrete time periods is possibly less accurate than using continuous time with potential nonlinear components, we believe that the superior interpretability of the current approach outweighs the potential inaccuracy. We reported the mean differences comparing supplementation to control arms during the three time periods ($\le$6, >6 to 12 and >12 to 18 months).

For the undernutrition outcomes (underweight, stunting, wasting, small head circumference, small arm circumference, SAM, MAM and GAM), we used Cox proportional-hazards models with the time to the first incidence (calculated as the date of incidence minus the birth date) as the outcome and used the exact method for ties. Children who did not develop the outcome due to either child death or loss to follow-up were censored at the last nurse visit. We assessed the assumption of proportional hazards by comparing Kaplan–Meier curves across treatment arms. We reported the hazard ratios (HRs) comparing supplementation to control arms.

For child-morbidity outcomes assessed by monthly nurse visits (diarrhoea, cough, fever and common cold), relative risks (RRs) were computed using generalized estimating equations with log link and binomial distribution.³⁰ For anaemia (any, mild, moderate and severe) and physician-diagnosed child-morbidity outcomes (acute diarrhoea, persistent diarrhoea, AURI, ALRI and ARI), we used generalized estimating equations with log link and Poisson distribution to compute incidence rate ratios (IRRs). All generalized estimating equations models included the intervention, indicator variables for time periods of child age and interaction terms between the intervention and time periods. We reported the RRs or IRRs comparing supplementation to control arms during the three time periods ($\le$6, >6 to 12 and >12 to 18 months). We used an exchangeable working covariance matrix and empirical (robust sandwich) standard errors to compute 95% confidence intervals (CIs) and p-values.³¹ We conducted subgroup analyses to evaluate whether the effects of postnatal MMS were modified by maternal and child characteristics (Supplementary Box S2, available as Supplementary data at IJE online). All analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary, North Carolina). We did not adjust for multiple testing.

Results

Maternal characteristics at study baseline and child characteristics at birth are presented in Table 1 by prenatal and postnatal regimens, and Figure 1 shows the study profile. The analyses for the prenatal regimen included 7783 mother–child pairs (3872 in the control group and 3911 in the MMS group). The analyses for the postnatal regimen included 3100 mother–child pairs (1542 in the control group and 1558 in the MMS group). The women were ∼25 years old and 21 weeks pregnant on average when enrolled in the study. Approximately 45% and 28% of the women were primigravida and secundigravida, respectively.

Table 1.

Maternal characteristics at baseline and child characteristics at birth, by prenatal and postnatal treatment arm, in a randomized–controlled trial of maternal multiple micronutrient supplementation among HIV-negative women in Dar es Salaam, Tanzania (2001–2006)^a

	Prenatal intervention		Postnatal intervention
	Control	MMS	Control	MMS
Maternal characteristics
Number of women (N)	3872	3911	1542	1558
Age at enrolment (years)	25.1 (5.1)	25.3 (5.1)	25.5 (5.1)	25.3 (5.0)
Gestational age at enrolment (weeks)	21.3 (3.5)	21.3 (3.4)	21.3 (3.3)	21.2 (3.3)
Parityb (%)
0 (primigravida)	1702 (44.3)	1748 (44.9)	691 (45.1)	676 (43.7)
1	1083 (28.2)	1086 (27.9)	414 (27.0)	459 (29.7)
2	589 (15.3)	567 (14.6)	244 (15.9)	231 (14.9)
$\ge$3	472 (12.3)	493 (12.7)	185 (12.1)	181 (11.7)
Maternal educationc (%)
0–4 years	422 (11.0)	455 (11.7)	168 (10.9)	172 (11.1)
5–7 years	2615 (67.9)	2562 (65.8)	1001 (65.2)	1018 (65.9)
8–11 years	626 (16.3)	673 (17.3)	283 (18.4)	276 (17.9)
$\ge$12 years	187 (4.9)	206 (5.3)	83 (5.4)	80 (5.2)
Marital statusd (%)
Married or cohabiting	3397 (88.5)	3449 (88.8)	1374 (90.0)	1382 (89.8)
Single	440 (11.5)	437 (11.3)	153 (10.0)	157 (10.2)
Maternal occupatione (%)
Housewife	2836 (75.8)	2768 (73.4)	1093 (76.0)	1154 (78.0)
Employed	904 (24.2)	1004 (26.6)	346 (24.0)	325 (22.0)
BMI at enrolmentf	24.6 (4.0)	24.6 (3.8)	24.8 (4.2)	24.5 (3.9)
BMI category at enrolmentf (%)
BMI <18.5	64 (1.9)	63 (1.8)	23 (1.6)	36 (2.5)
BMI: 18.5 to <25	2050 (60.2)	2020 (59.1)	857 (59.0)	851 (59.4)
BMI: 25 to <30	952 (27.9)	1031 (30.1)	429 (29.6)	416 (29.1)
BMI $\ge$30	341 (10.0)	307 (9.0)	143 (9.9)	129 (9.0)
Haemoglobin concentrations at enrolmentg (g/dL)	10.3 (1.6)	10.2 (1.5)	10.3 (1.6)	10.3 (1.5)
Total energy intakeh (kcal/d)	2317.2 (824.7)	2319.0 (808.8)	2413.5 (735.9)	2389.1 (706.2)
Child characteristics
Number of children (N)	3872	3911	1542	1558
Child age at the start of re-randomized regimen (weeks)	NA	NA	5.7 (0.14)	5.7 (0.14)
Girlsi (%)	1890 (49.1)	1835 (47.4)	781 (50.7)	769 (49.4)
Birthweightj (g)	3101.8 (504.7)	3166.3 (499.8)	3162.8 (485.9)	3169.7 (469.4)
Birth lengthk (cm)	47.4 (5.6)	47.5 (5.5)	47.4 (5.1)	47.3 (5.2)
Birth head circumferencel (cm)	34.4 (2.8)	34.4 (2.6)	34.3 (2.6)	34.4 (3.0)
Preterm birthm (%)	642 (18.2)	655 (18.6)	224 (15.9)	203 (14.7)
Small for gestational agen (%)	641 (19.1)	519 (15.4)	246 (17.8)	237 (17.5)
Large for gestational ageo (%)	480 (14.3)	493 (14.7)	184 (13.3)	180 (13.3)
Low birthweightp (%)	271 (7.3)	221 (5.9)	77 (5.1)	72 (4.7)
Haemoglobin concentrations at birth (g/dL)	14.0 (2.4)	13.9 (2.3)	14.1 (2.2)	14.1 (2.2)
Haemoglobin concentrations at 6 months of age (g/dL)	10.2 (1.6)	10.1 (1.7)	10.5 (1.8)	10.5 (2.0)
Haemoglobin concentrations at 12 months of age (g/dL)	10.2 (2.2)	10.3 (1.9)	11.4 (2.2)	11.3 (2.1)

^aValues are means (SDs) for continuous variables and counts and percentages for categorical variables. MMS, multiple micronutrient supplementation; NA, not applicable.

^bParity was missing for 26 women in the prenatal control group and 17 women in the prenatal MMS group.

^cMaternal education was missing for 22 women in the prenatal control group and 15 women in the prenatal MMS group.

^dMarital status was missing for 35 women in the prenatal control group and 25 women in the prenatal MMS group.

^eMaternal occupation was missing for 132 women in the prenatal control group and 139 women in the prenatal MMS group.

^fBMI at enrolment was missing for 465 women in the prenatal control group and 490 women in the prenatal MMS group.

^gHaemoglobin concentrations at enrolment were missing for 533 women in the prenatal control group and 543 women in the prenatal MMS group.

^hTotal energy intake was calculated as the average intake during pregnancy based on multiple 24-hour recalls; missing for 182 women in the prenatal control group and 189 women in the prenatal MMS group.

ⁱChild sex was missing for 26 children in the prenatal control group and 42 children in the prenatal MMS group.

^jBirthweight was missing for 172 children in the prenatal control group and 189 children in the prenatal MMS group.

^kBirth length was missing for 1179 children in the prenatal control group and 1195 children in the prenatal MMS group.

^lBirth head circumference was missing for 645 children in the prenatal control group and 685 children in the prenatal MMS group.

^mPreterm birth was defined as gestational age at birth at <37 weeks; missing for 343 children in the prenatal control group and 388 children in the prenatal MMS group.

ⁿBirthweight less than the 10th percentile for gestational age based on the INTERGROWTH-21st newborn size standards; missing for 507 children in the prenatal control group and 550 children in the prenatal MMS group.

^oBirthweight larger than the 90th percentile for gestational age based on the INTERGROWTH-21st newborn size standards; missing for 507 children in the prenatal control group and 550 children in the prenatal MMS group.

^pBirthweight <2500 g; missing for 172 children in the prenatal control group and 189 children in the prenatal MMS group.

Figure 1.

Study profile of a randomized–controlled trial of prenatal and postnatal maternal multiple micronutrient supplementation on child growth and morbidity among HIV-negative women in Dar es Salaam, Tanzania (2001–2006). MMS, multiple micronutrient supplementation.

Prenatal MMS increased child WAZ during the first 6 months of life; children whose mothers received MMS during pregnancy had 0.050 (95% CI: 0.002, 0.099; p = 0.04) SDs higher WAZ compared with children whose mothers received prenatal IFA alone (Table 2). However, this effect was attenuated after 6 months. Also, prenatal MMS did not have any effects on child underweight as a dichotomous outcome (HR: 1.01; 95% CI: 0.92, 1.11; p = 0.84). Similarly, prenatal MMS led to a higher child LAZ during the first 6 months [mean difference: 0.062 standard deviations (SDs); 95% CI: 0.013, 0.111; p = 0.01] but not thereafter, nor did prenatal MMS have any effect on stunting as a dichotomous outcome (HR: 0.98; 95% CI: 0.92, 1.05; p = 0.63). Prenatal MMS did not have any effect on child WLZ, HCZ, ACZ or haemoglobin concentrations. Prenatal MMS had no benefits for other undernutrition outcomes (Table 3), nurse-assessed morbidities (Table 4), physician-diagnosed morbidities (Table 5) or child-anaemia outcomes (Table 6).

Table 2.

Effects of prenatal and postnatal maternal multiple micronutrient supplementation on child-growth z-scores and haemoglobin concentrations in a randomized–controlled trial of maternal multiple micronutrient supplementation among HIV-negative women in Dar es Salaam, Tanzania (2001–2006)^a

	Prenatal intervention				Postnatal intervention
	Control	MMS	Mean difference (MMS—control) (95% CI)b	P b	Control	MMS	Mean difference (MMS—control) (95% CI)b	P b
	Children/visits (n)	Children/visits (n)			Children/visits (n)	Children/visits (n)
Weight-for-age z-score	3539/43 018	3587/43 240			1535/22 777	1549/23 161
$\mathrm{ }\le$ 6 months	3443/17 476	3449/17 618	0.050 (0.002, 0.099)	0.04	1533/7633	1542/7696	–0.042 (–0.113, 0.030)	0.25
>6 to 12 months	3267/17 558	3306/17 690	0.010 (–0.045, 0.066)	0.72	1456/7960	1470/8133	–0.044 (–0.125, 0.037)	0.29
>12 to 18 months	1675/7984	1675/7932	0.037 (–0.031, 0.105)	0.29	1370/7184	1405/7332	–0.039 (–0.126, 0.047)	0.37
Length-for-age z-score	3523/42 277	3570/42 585			1535/22 751	1549/23 135
$\mathrm{ }\le$ 6 months	3375/16 869	3391/17 065	0.062 (0.013, 0.111)	0.01	1533/7629	1542/7698	0.008 (–0.064, 0.080)	0.83
>6 to 12 months	3264/17 445	3301/17 610	0.005 (–0.048, 0.058)	0.86	1455/7951	1470/8120	–0.005 (–0.083, 0.072)	0.90
>12 to 18 months	1666/7963	1666/7910	0.027 (–0.038, 0.092)	0.42	1368/7171	1401/7317	–0.044 (–0.127, 0.039)	0.30
Weight-for-length z-score	3518/42 125	3566/42 415			1535/22 665	1548/23 055
$\mathrm{ }\le$ 6 months	3370/16 765	3386/16 965	–0.003 (–0.050, 0.044)	0.90	1533/7579	1541/7645	–0.080 (–0.152, –0.008)	0.03
>6 to 12 months	3262/17 415	3300/17 559	0.010 (–0.044, 0.065)	0.71	1455/7930	1470/8102	–0.053 (–0.136, 0.030)	0.21
>12 to 18 months	1660/7945	1658/7891	0.036 (–0.035, 0.107)	0.32	1367/7156	1400/7308	–0.015 (–0.105, 0.074)	0.74
Head circumference-for-age z-score	3535/42 759	3583/42 960			1535/22 702	1549/23 044
$\mathrm{ }\le$ 6 months	3440/17 363	3449/17 494	0.025 (–0.018, 0.069)	0.25	1533/7607	1542/7662	–0.034 (–0.097, 0.030)	0.30
>6 to 12 months	3275/17 450	3308/17 571	0.003 (–0.044, 0.049)	0.91	1457/7945	1469/8088	–0.041 (–0.108, 0.027)	0.24
>12 to 18 months	1671/7946	1685/7895	0.012 (–0.044, 0.068)	0.67	1376/7150	1404/7294	–0.055 (–0.126, 0.016)	0.13
Arm-circumference-for-age z-score	2758/27 027	2778/27 031			1519/20 418	1527/20 747
$\mathrm{ }\le$ 6 months	1951/6314	1968/6310	–0.015 (–0.067, 0.038)	0.58	1489/5275	1491/5305	–0.039 (–0.105, 0.027)	0.25
>6 to 12 months	2628/12 828	2635/12 869	0.001 (–0.051, 0.053)	0.98	1457/7958	1469/8110	–0.041 (–0.111, 0.029)	0.26
>12 to 18 months	1585/7885	1594/7852	0.016 (–0.047, 0.080)	0.61	1376/7185	1405/7332	–0.038 (–0.115, 0.038)	0.33
Haemoglobin, g/dL	3766/10 659	3773/10 729			1350/3283	1366/3338
$\mathrm{ }\le$ 6 months	3687/7410	3675/7463	0.024 (–0.051, 0.098)	0.53	1079/1386	1107/1395	–0.043 (–0.187, 0.101)	0.56
>6 to 12 months	1952/2403	1947/2378	0.056 (–0.075, 0.187)	0.40	918/1211	931/1248	0.018 (–0.137, 0.174)	0.82
>12 to 18 months	691/846	732/888	–0.044 (–0.262, 0.174)	0.69	536/686	544/695	–0.109 (–0.313, 0.096)	0.30

^aAll child-growth z-scores were based on the WHO 2006 Child Growth Standards. MMS, multiple micronutrient supplementation.

^bObtained from linear mixed-effects models comparing supplementation to control arms (control arm as the reference group). All models included MMS, indicator variables for time periods and interaction terms between MMS and time periods.

Table 3.

Effects of prenatal and postnatal maternal multiple micronutrient supplementation on undernutrition in a randomized–controlled trial of maternal multiple micronutrient supplementation among HIV-negative women in Dar es Salaam, Tanzania (2001–2006)

	Prenatal intervention				Postnatal intervention
	Control	MMSa	HR (95% CI)b	P b	Control	MMSa	HR (95% CI)b	P b
	Events/child-years (n)	Events/child-years (n)			Events/child-years (n)	Events/child-years (n)
Underweight	869/3370	887/3412	1.01 (0.92, 1.11)	0.84	431/1722	464/1723	1.07 (0.94, 1.22)	0.30
Stunting	1603/2823	1611/2884	0.98 (0.92, 1.05)	0.63	835/1375	859/1372	1.03 (0.94, 1.13)	0.54
Wasting	824/3386	856/3363	1.05 (0.95, 1.15)	0.36	405/1746	438/1725	1.09 (0.95, 1.25)	0.21
Small head circumferencec	564/3556	574/3587	1.01 (0.90, 1.13)	0.90	244/1836	257/1850	1.05 (0.88, 1.25)	0.61
Small arm circumferenced	379/3036	398/3039	1.05 (0.91, 1.21)	0.50	246/1849	283/1848	1.15 (0.97, 1.37)	0.11
Severe acute malnutritione	263/3119	282/3099	1.08 (0.91, 1.28)	0.36	130/1924	150/1936	1.14 (0.90, 1.44)	0.29
Moderate acute malnutritionf	737/2937	755/2882	1.04 (0.94, 1.16)	0.40	357/1785	403/1767	1.14 (0.99, 1.31)	0.07
Global acute malnutritiong	833/2875	867/2812	1.06 (0.97, 1.17)	0.21	406/1745	447/1721	1.11 (0.97, 1.27)	0.12

^aMultiple micronutrient supplementation.

^bObtained from Cox proportional-hazards models comparing supplementation to control arms (control arm as the reference group).

^cHead circumference-for-age z-score < –2.

^dArm circumference-for-age z-score < –2.

^eWeight-for-length z-score < –3 or mid-upper arm circumference <115 mm.

^fWeight-for-length z-score between –2 and –3 or mid-upper arm circumference between 115 and <125 mm.

^gSevere acute malnutrition or moderate acute malnutrition.

Table 4.

Effects of prenatal and postnatal maternal multiple micronutrient supplementation on child infectious morbidities by nurse evaluations in a randomized–controlled trial of maternal multiple micronutrient supplementation among HIV-negative women in Dar es Salaam, Tanzania (2001–2006)

	Prenatal intervention				Postnatal intervention
	Control	MMSa	RR (95% CI)b	P b	Control	MMSa	RR (95% CI)b	P b
	Events/visits (n)	Events/visits (n)			Events/visits (n)	Events/visits (n)
Diarrhoea	1686/43 477	1714/43 710			879/22 997	887/23 312
$\mathrm{ }\le$ 6 months	457/17 575	485/17 665	1.06 (0.93, 1.20)	0.42	190/7658	177/7741	0.92 (0.75, 1.14)	0.44
>6 to 12 months	856/17 670	871/17 790	1.01 (0.92, 1.12)	0.82	375/7984	364/8124	0.95 (0.82, 1.11)	0.52
>12 to 18 months	373/8232	358/8255	0.96 (0.83, 1.11)	0.56	314/7355	346/7447	1.09 (0.93, 1.27)	0.28
Cough	9906/42 597	10215/42 829			5531/22 650	5564/23 012
$\mathrm{ }\le$ 6 months	3580/17 211	3598/17 315	1.00 (0.95, 1.05)	0.99	1603/7532	1578/7613	0.98 (0.91, 1.05)	0.54
>6 to 12 months	4296/17 258	4476/17 347	1.03 (0.99, 1.08)	0.14	2058/7853	2039/8026	0.97 (0.91, 1.03)	0.32
>12 to 18 months	2030/8128	2141/8167	1.05 (0.98, 1.12)	0.14	1870/7265	1947/7373	1.03 (0.96, 1.10)	0.47
Fever	4297/43 183	4384/43 411			2372/22 873	2347/23 203
$\mathrm{ }\le$ 6 months	1257/17 486	1229/17 590	0.97 (0.90, 1.06)	0.52	553/7640	510/7723	0.92 (0.81, 1.04)	0.17
>6 to 12 months	2011/17 526	2128/17 634	1.05 (0.98, 1.12)	0.15	876/7936	904/8091	1.01 (0.92, 1.12)	0.80
>12 to 18 months	1029/8171	1027/8187	0.98 (0.90, 1.08)	0.72	943/7297	933/7389	0.98 (0.89, 1.08)	0.68
Common cold	8972/42 744	9325/43 003			4858/22 747	4796/23 099
$\mathrm{ }\le$ 6 months	3691/17 207	3782/17 327	1.02 (0.97, 1.06)	0.50	1577/7556	1474/7633	0.93 (0.87, 1.00)	0.05
>6 to 12 months	3575/17 373	3807/17 479	1.05 (1.00, 1.10)	0.04	1723/7893	1723/8067	0.98 (0.91, 1.05)	0.50
>12 to 18 months	1706/8164	1736/8197	1.00 (0.93, 1.07)	0.98	1558/7298	1599/7399	1.01 (0.94, 1.09)	0.73

^aMultiple micronutrient supplementation.

^bObtained from generalized estimating equations with log link and binomial distribution comparing supplementation to control arms (control arm as the reference group). All models included MMS, indicator variables for time periods and interaction terms between MMS and time periods.

Table 5.

Effects of prenatal and postnatal maternal multiple micronutrient supplementation on child infectious morbidities by physician diagnosis in a randomized–controlled trial of maternal multiple micronutrient supplementation among HIV-negative women in Dar es Salaam, Tanzania (2001–2006)^a

	Prenatal intervention				Postnatal intervention
	Control	MMS	IRR (95% CI)b	P b	Control	MMS	IRR (95% CI)b	P b
	Events/visits (n)	Events/visits (n)			Events/visits (n)	Events/visits (n)
Acute diarrhoea	424/19 390	432/19 546			162/7948	175/8070
$\mathrm{ }\le$ 6 months	123/9497	143/9585	1.15 (0.90, 1.48)	0.27	42/2771	53/2832	1.22 (0.81, 1.83)	0.35
>6 to 12 months	249/7114	235/7201	0.93 (0.77, 1.11)	0.41	71/2737	79/2777	1.10 (0.79, 1.52)	0.58
>12 to 18 months	52/2779	54/2760	1.07 (0.73, 1.55)	0.74	49/2440	43/2461	0.86 (0.58, 1.29)	0.47
Persistent diarrhoea	27/18 021	23/18 189			11/7485	19/7607
$\mathrm{ }\le$ 6 months	8/8882	4/8989	0.49 (0.15, 1.64)	0.25	1/2591	4/2622	3.98 (0.44, 35.61)	0.22
>6 to 12 months	13/6458	13/6535	0.99 (0.46, 2.13)	0.98	4/2546	9/2609	2.19 (0.68, 7.09)	0.19
>12 to 18 months	6/2681	6/2665	1.01 (0.33, 3.11)	0.99	6/2348	6/2376	0.99 (0.32, 3.06)	0.98
Acute upper respiratory infection	1872/19 390	1871/19 548			716/7944	670/8074
$\mathrm{ }\le$ 6 months	837/9494	841/9588	0.99 (0.90, 1.10)	0.88	219/2767	223/2831	0.99 (0.82, 1.19)	0.88
>6 to 12 months	770/7115	775/7201	0.99 (0.89, 1.09)	0.80	261/2739	226/2780	0.86 (0.72, 1.02)	0.08
>12 to 18 months	265/2781	255/2759	0.96 (0.82, 1.14)	0.67	236/2438	221/2463	0.93 (0.77, 1.12)	0.42
Acute lower respiratory infection	1174/19 399	1129/19 549			442/7948	456/8077
$\mathrm{ }\le$ 6 months	497/9502	483/9587	0.97 (0.85, 1.10)	0.60	167/2771	191/2835	1.12 (0.90, 1.38)	0.31
>6 to 12 months	563/7116	527/7202	0.93 (0.82, 1.04)	0.20	165/2739	176/2778	1.06 (0.86, 1.32)	0.59
>12 to 18 months	114/2781	119/2760	1.07 (0.82, 1.39)	0.61	110/2438	89/2464	0.81 (0.60, 1.08)	0.15
Acute respiratory infection	2931/19 390	2898/19 544			1135/7943	1096/8071
$\mathrm{ }\le$ 6 months	1277/9497	1279/9587	0.99 (0.92, 1.08)	0.86	375/2769	402/2831	1.04 (0.90, 1.19)	0.59
>6 to 12 months	1284/7114	1250/7199	0.96 (0.89, 1.03)	0.27	418/2738	390/2778	0.92 (0.81, 1.06)	0.25
>12 to 18 months	370/2779	369/2758	1.01 (0.88, 1.16)	0.89	342/2436	304/2462	0.88 (0.76, 1.03)	0.10

^aIRR, incidence rate ratio; MMS, multiple micronutrient supplementation.

^bObtained from generalized estimating equations with log link and Poisson distribution comparing supplementation to control arms (control arm as the reference group). All models included MMS, indicator variables for time periods and interaction terms between MMS and time periods.

Table 6.

Effects of prenatal and postnatal maternal multiple micronutrient supplementation on child anaemia in a randomized–controlled trial of maternal multiple micronutrient supplementation among HIV-negative women in Dar es Salaam, Tanzania (2001–2006)^a

	Prenatal intervention				Postnatal intervention
	Control	MMS	IRR (95% CI)b	P b	Control	MMS	IRR (95% CI)b	P b
	Events/visits (n)	Events/visits (n)			Events/visits (n)	Events/visits (n)
Any anaemiac	1872/3249	1847/3266			684/1897	763/1943
>6 to 12 months	1581/2403	1524/2378	0.97 (0.94, 1.01)	0.19	511/1211	564/1248	1.07 (0.99, 1.16)	0.10
>12 to 18 months	291/846	323/888	1.05 (0.93, 1.19)	0.42	173/686	199/695	1.14 (0.95, 1.36)	0.16
Mild anaemiad	621/3249	667/3266			274/1897	312/1943
>6 to 12 months	511/2403	522/2378	1.03 (0.93, 1.15)	0.54	192/1211	217/1248	1.09 (0.92, 1.30)	0.32
>12 to 18 months	110/846	145/888	1.26 (1.00, 1.58)	0.05	82/686	95/695	1.15 (0.87, 1.52)	0.33
Moderate anaemiae	1189/3249	1128/3266			400/1897	439/1943
>6 to 12 months	1012/2403	962/2378	0.96 (0.90, 1.02)	0.21	311/1211	338/1248	1.06 (0.93, 1.20)	0.39
>12 to 18 months	177/846	166/888	0.88 (0.73, 1.06)	0.19	89/686	101/695	1.12 (0.86, 1.47)	0.40
Severe anaemiaf	62/3249	52/3266			10/1897	12/1943
>6 to 12 months	58/2403	40/2378	0.70 (0.47, 1.05)	0.08	8/1211	9/1248	1.09 (0.42, 2.82)	0.86
>12 to 18 months	4/846	12/888	2.84 (0.95, 8.47)	0.06	2/686	3/695	1.49 (0.25, 8.92)	0.66

^aIRR, incidence rate ratio; MMS, multiple micronutrient supplementation.

^bObtained from generalized estimating equations with log link and Poisson distribution comparing supplementation to control arms (control arm as the reference group). All models included MMS, indicator variables for time periods and interaction terms between MMS and time periods.

^cHaemoglobin concentration <11 g/dL for children aged 6–18 months.

^dHaemoglobin concentration between 10 and 10.9 g/dL for children aged 6–18 months.

^eHaemoglobin concentration between 7 and 9.9 g/dL for children aged 6–18 months.

^fHaemoglobin concentration <7 g/dL for children aged 6–18 months.

We did not observe any effects of postnatal MMS on haemoglobin concentrations or child-growth z-scores (Table 2), except for WLZ, which was slightly reduced by postnatal MMS during the first 6 months of life (mean difference: –0.080 SDs; 95% CI: –0.152, –0.008; p = 0.03). Postnatal MMS did not have effects on any of the undernutrition outcomes (Table 3), nurse-assessed morbidities (Table 4), physician-diagnosed morbidities (Table 5) or child-anaemia outcomes (Table 6). Results from subgroup analyses by potential effect modifiers are similar to the primary results (Supplementary Box S3 and Supplementary Tables S1–S5, available as Supplementary data at IJE online).

Discussion

In this analysis of a double-blind, randomized–controlled trial in Tanzania, we found that prenatal MMS resulted in a slight increase in child WAZ and LAZ during the first 6 months of life and not thereafter. Further, this specific combination of prenatal micronutrients did not have benefits for preventing undernutrition, infectious morbidities or anaemia outcomes. We observed no benefits of supplementing lactating women with this particular regimen on any child outcomes examined.

Substantial evidence indicates that prenatal MMS reduces the risk of various adverse pregnancy outcomes. Compared with IFA or iron alone, prenatal MMS prevents small for gestational age and low birthweight, and may also be effective in reducing the risks of stillbirths and preterm births.⁷ The WHO recently updated their recommendations on antenatal care from IFA to MMS and has judged that prenatal MMS that includes iron and folic acid is recommended in the context of rigorous research.⁹ It has been suggested that micronutrient status during pregnancy may affect fetal growth and postnatal growth.^10,12 Nevertheless, the current body of literature is inconsistent regarding the impacts of prenatal MMS on postnatal child growth and nutrition.^11–16 In the MISAME trial in Burkina Faso, MMS increased LAZ, WLZ and HCZ, and reduced stunting until 30 months of age.¹² In the JiVitA-3 trial in rural Bangladesh, MMS provided from pregnancy through to 3 months postpartum improved LAZ and reduced stunting through to 3 months of age but not thereafter.¹⁵ In a trial in Vietnam, MMS provided from the third trimester until 3 months postpartum increased infant WAZ, LAZ and HCZ from birth to 3 months.¹⁶ Trials in Vietnam¹⁴ and China,¹³ however, found no differences in child-growth measures (at 6 months in Vietnam and from birth to 30 months in China) comparing prenatal MMS to IFA. In the MINIMat trial in Bangladesh, prenatal MMS resulted in more stunting from birth to 54 months compared with IFA, especially among boys.¹¹

The inconsistent findings from previous studies may be attributable to differences in the composition and dosage of the supplements. Some previous work^11–16 has used the United Nations International Multiple Micronutrient Antenatal Preparation or similar formulations,³² which included some micronutrients not present in our regimen such as vitamin A and minerals (copper, selenium, iodine and zinc). Different durations of supplementation may also have led to mixed results. Some studies provided prenatal MMS from pregnancy until 3 months postpartum, as opposed to 6 weeks in our study,^15,16 making it challenging to separate the effect of postnatal supplementation from that of prenatal supplementation. The discrepancy among previous studies may also be explained by different study settings and baseline nutritional levels. Given the national differences in maternal and child nutrition, it is vital to conduct MMS trials across diverse settings.³³ Our study is the first to assess the effects of prenatal MMS on child growth and morbidity in Tanzania—a country with decreasing yet still high burdens of maternal and child malnutrition.^1,34 Our findings on child-morbidity outcomes are consistent with the small but relatively consistent body of evidence suggesting that, among children of HIV-negative mothers, prenatal MMS has no effect on child infectious morbidities such as fever,^18,19 diarrhoea,¹⁹ cough,¹⁸ acute lower respiratory tract infection¹⁹ or early neonatal morbidity symptoms.¹⁷

The demands for micronutrients are greatly increased during breastfeeding to levels that are much higher than those during pregnancy.³⁵ Under the recommendation of exclusive breastfeeding,³⁶ breast milk is expected to provide all the essential micronutrients for infants during the first 6 months of life. Thus, it is plausible that supplementing lactating women with micronutrients may increase breast-milk micronutrient concentrations, which will then confer benefits for the child.⁶ However, there is scarce evidence on the impacts of postnatal maternal MMS from randomized trials. A recent Cochrane systematic review⁶ identified only two small RCTs investigating the effects of postnatal maternal MMS among HIV-negative women^20,21 and neither included child outcomes. The iLiNS-DYAD-M trial in Malawi³⁷ and the iLiNS-DYAD-G trial in Ghana³⁸ provided IFA or MMS to lactating women for ≤6 months postpartum. However, as the postnatal regimens in these studies were the same as the prenatal regimens received during pregnancy, it was challenging to disentangle the independent effects of prenatal vs postnatal supplementation. The unique factorial design of our study with a postpartum re-randomization provides an opportunity to address this knowledge gap on postnatal supplementation. The lack of benefits of postnatal MMS for any child outcome is in line with previous work that provided supplementations directly to children.^33,39–42 In a two-by-two factorial RCT among HIV-unexposed Tanzanian infants and young children, we found a small increase in WAZ when daily MMS was provided together with zinc; however, when MMS was provided by itself without zinc, no effects were observed on the incidences of underweight, stunting or wasting.⁴² We also reported that daily zinc supplementation to infants reduced the risk of physician-diagnosed diarrhoea and AURI. However, providing MMS conferred no additional benefits for any physician-diagnosed or nurse-assessed morbidities.³³ The slight decrease in WLZ during the first 6 months of life in the postnatal MMS group compared with the control group in our study is difficult to explain and may be due to random error. It should be noted that we find no effect in either direction of postnatal MMS on wasting, which is perhaps a more clinically meaningful outcome. Future studies should keep evaluating the effect of postnatal MMS on short- and long-term WLZ and wasting.

The small and transient effects of prenatal MMS on WAZ and LAZ during the first 6 months in this study may be partially explained by the effect of prenatal supplementation on increased birthweight and birth length. The lack of long-term benefits may be due to the failure of prenatal or postnatal supplementation to result in sustained improvements in the concentrations of micronutrients in the mother or the child. In another analysis of the same study, we reported that prenatal supplementation resulted in a modest increase in breast-milk vitamin B-12 concentrations at 6 weeks postpartum.⁴³ A recent study based on the JiVitA-3 trial also reported that prenatal MMS had modest direct effects on newborn micronutrient status.⁴⁴ However, a trial in India supplementing mothers with 50 μg/d of vitamin B-12 (the same dosage of B-12 as in the current study) from the first trimester through to 6 weeks postpartum resulted in increases in maternal plasma B-12 during pregnancy, breast-milk B-12 at 6 weeks postpartum and infant plasma B-12 at 6 weeks of age.⁴⁵ The benefits of micronutrient supplementation during pregnancy or lactation for maternal and child nutrient status remain uncertain and further examination of the potential modifiers of its benefits is warranted. The feasibility, acceptability, equity and cost-effectiveness aspects of postnatal MMS also need to be further evaluated.

The strengths of this study include a large number of participants, a long period of follow-up and the inclusion of a wide range of child-growth, anaemia and morbidity outcomes. This study also had some limitations. First, we did not examine the impacts of prenatal or postnatal MMS on micronutrient status. Maternal and child plasma and maternal breast-milk samples were collected in our study. We plan to measure the stored samples to assess the impacts of MMS on micronutrient biomarkers in future efforts. Second, as all women were recruited from an urban setting of Tanzania, the results may not be fully generalizable to rural contexts with higher burdens of maternal and child malnutrition and morbidity.⁴⁶

In conclusion, this study contributes to the evidence base for future nutritional interventions of antenatal and postnatal micronutrient supplementation. We did not find substantial evidence of benefits of prenatal or postnatal maternal MMS on a wide range of child outcomes. Whereas maternal MMS has demonstrated efficiency in preventing adverse birth outcomes, other approaches may need to be considered in conjunction with maternal MMS to curb the high burdens of child morbidity and growth faltering.

Supplementary data

Supplementary data are available at IJE online.

Ethics approval

This study was approved by the institutional review boards at Muhimbili University of Health and Allied Sciences and Harvard T.H. Chan School of Public Health. The ClinicalTrials.gov registration number of the study is NCT00197548; child growth and morbidity evaluated in this analysis were registered as the secondary outcomes of the study. All women enrolled in the study provided written, informed consent to participate. The study was carried out in accordance with the guidelines and regulations in the Declaration of Helsinki.

Funding

This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health [NICHD R01 37701].

Data availability

The data underlying this article cannot be shared publicly for the privacy of the individuals who participated in the study. The data will be shared on reasonable request to the corresponding author.