1. Objectives

To assess the effects of vitamin A supplementation, with or without other micronutrients, for postpartum women on maternal and infant health outcomes

2. How studies were identified

The following databases were searched in February 2016:

• Cochrane Pregnancy and Childbirth Group’s Trials Register
• CENTRAL (The Cochrane Library 2016)
• MEDLINE
• EMBASE
• LILACS
• CINAHL
• Web of Science

Reference lists were also searched

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials, including cluster-randomized trials

3.2 Study participants

Postpartum women, breastfeeding or otherwise, in settings of vitamin A deficiency

(Studies that exclusively enrolled women with conditions such as HIV were excluded and maternal data from studies conducted in areas with a high prevalence of HIV were also excluded if no diagnostic testing was performed at baseline)

3.3 Interventions

Vitamin A supplementation (beta-carotene or retinyl palmitate in oil or water-miscible formulation), with or without other micronutrients, compared with placebo, no intervention, other micronutrients without vitamin A, or a lower dose of vitamin A, and commenced up to six weeks postpartum

3.4 Primary outcomes

Maternal outcomes

• Mortality
• Morbidity (febrile illness, respiratory tract infection, diarrhoea, anaemia, and others)
• Adverse effects within three days of receiving vitamin A

Infant outcomes

• Mortality
• Morbidity (febrile illness, respiratory tract infection, diarrhoea, anaemia, and others)
• Adverse effects within three days of receiving vitamin A

Secondary outcomes included serum retinol concentration (maternal and infant), vitamin A hepatic reserves (maternal and infant), breast milk retinol concentration, maternal vitamin A deficiency (clinical: impaired visual adaptation to darkness, night blindness, xerophthalmia; subclinical: abnormal conjunctival impression cytology), and infant vitamin A deficiency (clinical: signs of xerophthalmia)

4. Main results

4.1 Included studies

Fourteen randomized controlled trials, enrolling 25,758 women and infant dyads, were included in this review

• Nine trials compared vitamin A with placebo, one trial compared vitamin A with no intervention, one trial compared vitamin A plus iron to iron alone, three trials compared a lower dose of vitamin A (200,000 IU) with a higher dose (400,000 IU)
• Ten studies administered a single dose of vitamin A (200,000, 300,000, or 400,000 IU) in the form of retinyl palmitate or water-miscible formulation within the first days or weeks postpartum
• One trial compared 7.8 mg of daily beta-carotene to placebo
• Ten of the 14 trials reported that women breastfed their infants, with seven of these trials reporting women at least partially breastfed until six months postpartum. All four studies that did not report breastfeeding directly indicated that infants were at least partially breastfed

4.2 Study settings

• Bangladesh (2 trials), Brazil (2 trials), Gambia, Ghana, India (3 trials), Indonesia, Kenya, the United Republic of Tanzania, Zimbabwe, and a multinational study (Ghana, India, Peru)
• All studies were conducted in countries with a high risk of inadequate vitamin A intakes, and all studies were conducted in low- and middle-income settings

4.3 Study settings

How the data were analysed
Two comparisons were made: i) vitamin A (as retinyl, water-miscible or beta-carotene) versus placebo or no vitamin A treatment; and ii) 400,000 IU of vitamin A (as retinyl) versus 200,000 IU of vitamin A (as retinyl). Fixed effects meta-analysis was used to generate risk ratios (RR) and 95% confidence intervals (CI) for dichotomous outcomes and mean differences (MD) and 95% CI for continuous outcomes. Random effects meta-analysis was used when substantial clinical or statistical (I²>30%) heterogeneity was detected. To explore potential sources of heterogeneity, the following subgroup analyses were planned:

• By type of supplement: vitamin A (retinyl palmitate or water-miscible formulation) or beta-carotene
• By duration of supplementation: daily, single or double bolus dose
• By dose: 200,000 to 300,000 IU versus 400,000 IU of vitamin A
• By duration of breastfeeding

Results
Vitamin A (as retinyl, water-miscible or beta-carotene) versus placebo or no vitamin A treatment
Primary maternal outcomes
Maternal mortality was not significantly affected by vitamin A supplementation in either of two trials reporting on this outcome (RR 0.50, 95% CI [0.09 to 2.71]; 1 trial/564 participants; and hazard ratio 1.01, 95% CI [0.44 to 2.21]; 1 trial/8577 participants; data not pooled). Maternal morbidity (fever, respiratory tract infection, diarrhoea) was assessed in one small trial of 50 women in which no differences between vitamin A and control groups were reported (analyses not provided). In a subset of one trial including 786 women, no significant difference in the risk of the following adverse effects was found within 30 hours of vitamin A supplementation: headache (RR 1.21, 95% CI [0.74 to 1.99]), blurred vision (RR 1.64, 95% CI [0.39 to 6.82]), drowsiness (RR 2.09, 95% [0.91 to 4.79]), nausea (RR 1.38, 95% CI [0.44 to 4.31]), vomiting (RR 0.33, 95% CI [0.03 to 3.14]), poor appetite (RR 2.22, 95% CI [0.69 to 7.14]), and abdominal pain (RR 1.28, 95% CI [0.95 to 1.73]).

Additional maternal outcomes
Serum retinol at three to 3.5 months postpartum was improved with 200,000 IU to 400,000 IU of vitamin A (MD 0.11 μmol/L, 95% CI [0.03 to 0.19]; 5 trials/704 women). When subgrouped by dose, 200,000 IU to 300,000 IU remained statistically significant (MD 0.17 μmol/L, 95% CI [0.07 to 0.19]; 4 trials/302 women) while 400,000 IU did not (MD 0.04 μmol/L, 95% CI [-0.01 to 0.09]; 1 trial/402 women). Daily beta-carotene did not significantly improve serum retinol at three to 3.5 months postpartum (MD 0.10 μmol/L, 95% CI [-0.14 to 0.34]; 1 trial/54 women). Serum retinol at six or nine months postpartum was not improved by vitamin A supplementation overall, or by beta-carotene, although a dose 200,000 IU to 300,000 IU of vitamin A improved serum retinol at six months (MD 0.13 μmol/L, 95% CI [0.03 to 0.23]; 3 trials/242 women). The risk of low hepatic vitamin A reserves was not different between treatment and control groups at three, six, or nine months postpartum regardless of dose or type of supplement. Breast milk retinol concentrations were improved with 200,000 IU to 400,000 IU of vitamin A supplementation at three to 3.5 months postpartum (MD 0.20 μmol/L, 95% CI [0.08 to 0.31]; 6 trials/837 women), but not with 7.8 mg beta-carotene/day (MD 0.02 μmol/L, 95% CI [-0.17 to 0.21]; 1 trial 109 women). At six months postpartum, only 400,000 IU of vitamin A had a significant effect on breast milk retinol (MD 0.06 μmol/L, 95% CI [0.01 to 0.11]; 1 trial/354 women). At eight to nine months postpartum, 7.8 mg beta-carotene/day had a significant effect on breast milk retinol (MD 0.21 μmol/L, 95% CI [0.01 to 0.41]; 1 trial/103 women), while 200,000 IU to 300,000 IU of vitamin A did not. The risk of breast milk retinol concentrations <1.05 μmol/L was reduced at three months postpartum with 200,000 IU to 300,000 IU of vitamin A (RR 0.56, 95% CI [0.37 to 0.84]; 3 trials/304 women). The risk of breast milk retinol concentrations <0.28 μmol/g of fat at six months postpartum was reduced with daily beta-carotene supplementation (RR 0.76, 95% CI [0.60 to 0.97]; 1 trial/102 women). Maternal abnormal conjunctival impression cytology was not different by treatment group.

Infant primary outcomes
No statistically significant effect of maternal vitamin A supplementation on infant mortality was found (400,000 IU, to 14 weeks of age: RR 0.71, 95% CI [0.33 to 1.54], 1 trial/279 infants; 400,000 IU, to 12 months of age: RR 1.28, 95% CI [0.83 to 1.98], 1 trial/4601 infants; 300, 000 IU, to 12 months of age: RR 0.88, 95% CI [0.34 to 2.24], 1 trial/598 infants; 200, 000 IU, to 2 months of age: RR 0.33, 95% CI [0.01 to 7.90], 1 trial/66 infants; 200, 000 IU, to 6 months of age: RR 1.54, 95% CI [0.26 to 9.17], 1 trial/546 infants). The risk of infant diarrhoea was not significantly affected by 300,000 IU of maternal vitamin A supplementation (RR 1.02, 95% CI [0.98 to 1.06]; 1 trial/456 infants) or by 200,000 IU of maternal vitamin A supplementation (200,000 IU 0.12 episodes versus 0.11 episodes in the control group; 1 trial/50 infants; p=0.59), and nor was the duration of diarrhoea (0.74 days versus 0.71 days in the control group; p=0.78; 1 trial/50 infants). Infant gastroenteritis (RR 6.03, 95% CI [0.30 to 121.82]; 1 trial/84 infants), acute respiratory tract infection (RR 1.00, 95% CI [0.96 to 1.03]; 1 trial/456 infants; and 0.42 episodes versus 0.45 episodes in the control group; 1 trial/50 infants; p=0.51), and upper respiratory tract infection (RR 0.91, 95% CI [0.22 to 3.81]; 1 trial/84 infants) did not differ between treatment and control groups. The duration of acute respiratory tract infection was reduced with vitamin A supplementation by 0.6 days in one trial of 50 infants (p=0.03) and the mean number of febrile illness episodes was also lower in the same trial (0.1 versus 0.3 episodes in the control group; p<0.002). The adverse effect of bulging fontanelle was not significantly increased with 400,000 IU of maternal vitamin A supplementation in one trial including 444 mother-infant pairs (RR 2.00, 95% CI [0.61 to 6.55]).

Additional infant outcomes
Infant serum retinol at three to 3.5 months of age was not improved overall (200, 000 IU to 400, 000 IU), but was marginally improved by maternal supplementation with 200, 000 IU of vitamin A (MD 0.20 μmol/L, 95% CI [0.00 to 0.39]; 3 trials/215 infants; p=0.047). Overall, maternal supplementation with 200, 000 IU to 400, 000 IU had no statistically significant effect on infant serum retinol at six months (MD 0.04 μmol/L, 95% CI [0.00 to 0.09]; 4 trials/330 infants; p=0.08). No differences were found between treatment groups for the outcome low infant hepatic vitamin A stores.

400,000 IU of vitamin A (as retinyl) versus 200,000 IU of vitamin A (as retinyl)
Maternal outcomes
No trials of 400,000 IU versus 200,000 IU of vitamin A reported on maternal primary outcomes. No statistically significant difference in serum retinol concentrations was found at two months (2 trials/429 women), four months (1 trial/213 women), or six months (1 trial/200 women) postpartum. In one trial including 377 women, no difference between treatment groups was found in breast milk retinol concentrations between two and six months postpartum, or in the risk of breast milk retinol concentrations <1.05 μmol/L at three or six months postpartum.

Infant outcomes
No trials comparing 400,000 IU to 200,000 IU of vitamin A reported on infant primary outcomes. Infant serum retinol concentrations at two months (2 trials/362 infants), four months (1 trial/292 infants), and six months (1 trial/173 infants) of age did not differ by treatment group.

5. Additional author observations*

The overall risk of bias in the included trials was judged to be low to unclear. Evidence for the outcome maternal mortality was assessed to be of moderate quality using GRADE criteria, while evidence for the outcome infant mortality was rated as low quality. Evidence for both maternal and infant morbidity outcomes was rated as very low quality, and evidence for the outcomes adverse effects of supplementation and breast milk retinol concentrations was judged to be low quality.

In regions with a high risk of inadequate vitamin A intake, vitamin A supplementation in the postpartum period provided limited benefit to women and their infants and did not significantly increase the risk of adverse effects.

Further research on the effect of postpartum vitamin A supplementation on maternal and infant mortality and morbidity, and the interaction between vitamin A and other micronutrients, such as iron and zinc, is warranted. Sustainable interventions to improve vitamin A-rich food intake is also an important avenue of investigation to pursue.