1. Objectives

To evaluate the effectiveness of supplementation with vitamin A or one of its derivatives during pregnancy, alone or in combination with other micronutrients, on maternal and newborn clinical and laboratory outcomes

2. How studies were identified

The Cochrane Pregnancy and Childbirth Group’s Trials Register was searched to March 2015, which includes records identified from the following databases:

• CENTRAL (The Cochrane Library)
• MEDLINE
• EMBASE
• CINAHL

Relevant journals and conference proceedings were included in the search, and reference lists from identified studies were also handsearched

3. Criteria for including studies in the review

3.1 Study type

Randomized or quasi-randomized trials, including cluster-randomized trials

3.2 Study participants

Pregnant women in areas with inadequate or adequate vitamin A intake, as defined by the WHO global database on vitamin A deficiency

(Studies in which women were supplemented exclusively in the postnatal period were excluded)

(Studies aimed at preventing vertical transmission of HIV were included if they met inclusion criteria and reported relevant outcomes)

3.3 Interventions

Vitamin A supplementation alone or in combination with other micronutrients compared to placebo, no treatment, or micronutrients without vitamin A

3.4 Primary outcomes

• Maternal mortality (defined as death during pregnancy or within 42 days of cessation of pregnancy, for any cause related to or aggravated by pregnancy except accidental or incidental causes)
• Perinatal mortality (defined as number of stillbirths and deaths in the first week of life per 1000 live births)

Secondary outcomes included neonatal mortality (death during the first 28 days of life), stillbirths, maternal anaemia (haemoglobin <11.0 g/dL), maternal clinical infection, maternal night blindness, preterm birth (<37 weeks’ gestation), neonatal anaemia, neonatal clinical infection, congenital malformations, low birth weight (<2.5 kg)

4. Main results

4.1 Included studies

Nineteen trials involving over 310,000 women were included in this review:

• Three studies were cluster-randomized, and 16 were individually randomized. Two of the 19 trials were quasi-randomized
• Ten trials assessed vitamin A (or one of its derivatives) alone in comparison with a control group; five trials evaluated vitamin A (or one of its derivatives) in combination with other micronutrients in comparison to a control group; and four trials were multi-armed assessing both the use of vitamin A (or one of its derivatives) alone and vitamin A in combination with other supplements compared with a control group
• Control groups were exposed to a placebo, no treatment, or another intervention (for example, iron)
• Three trials were conducted in women with HIV infection
• The number of enrolled women ranged from 44 to over 207,000, and interventions ranged from a single day to over 18 weeks in duration
• Vitamin A doses ranged from 3000 IU per day up to 44,4000 IU per day, with some trials dosing on a weekly basis; single doses given at delivery were approximately 200,000 IU; and one trial used single doses of 600,000 IU given intra-muscularly or orally at the time of delivery

4.2 Study settings

• Bangladesh, China, Ghana (2 trials), India, Indonesia (6 trials), Malawi (3 trials), Nepal, South Africa, the United Kingdom of Great Britain and Northern Ireland, the United Republic of Tanzania, and the United States of America
• Sixteen studies were conducted in populations with moderate vitamin A deficiency, two were conducted in populations without vitamin A deficiency (the United Kingdom of Great Britain and Northern Ireland and the United States of America), and one study was conducted in a population with severe vitamin A deficiency (Nepal)
• Studies were predominantly conducted in rural settings

4.3 Study settings

How the data were analysed
Three comparisons were made: i) vitamin A alone versus placebo or no treatment; ii) vitamin A alone versus micronutrient supplements without vitamin A; and iii) vitamin A with other micronutrients versus micronutrients without vitamin A. Data from cluster-randomized trials were adjusted for clustering using standard methods. Fixed effects meta-analysis was used to produce pooled risk ratios (RR) with 95% confidence intervals (CI) where it was reasonable to assume that the treatment effects measured in the included trials were sufficiently similar. Where substantial heterogeneity was identified, random effects meta-analysis was used. The following subgroup analyses were planned for the primary outcomes maternal and perinatal mortality:

• By infant mortality rate in the country: high (≥30 per 1000 live births), low
• By maternal mortality rate in the country: high (≥100 per 100000 live births), low
• By prevalence of vitamin A deficiency in the country: high, low
• By prevalence of HIV infection in the country: high (>3%), low
• By dose of vitamin A: 10,000 IU/day, other doses
• By supplementation regimen: daily, weekly
• By duration of supplementation: ≤ one month, > one month
• By trimester of pregnancy at initiation of supplementation: pre-pregnancy; first, second, and third trimester

Results
Vitamin A alone versus placebo or no treatment
Maternal and infant mortality
The risk of maternal mortality was not significantly reduced with vitamin A supplementation in comparison to placebo or no treatment (RR 0.88, 95% CI [0.65 to 1.20], p=0.42; 4 studies/101,574 women). The risk of perinatal mortality was not affected by maternal vitamin A treatment (RR 1.01, 95% CI [0.95 to 1.07], p=0.74; 1 study/76,176 women), and neither was neonatal mortality (RR 0.97, 95% CI [0.90 to 1.05], p=0.50; 3 studies/89,556 women) or stillbirth (RR 1.04, 95% CI [0.98 to 1.10], p=0.19; 2 studies/122,850 women). Subgroup analyses did not alter results meaningfully.

Additional outcomes
Among women treated with vitamin A, the risk of anaemia was reduced by a statistically significant 36% (RR 0.64, 95% CI [0.43 to 0.94], p=0.025; 3 studies/3818 women). In addition, vitamin A supplementation significantly reduced the risk of maternal clinical infection (RR 0.45, 95% CI [0.20 to 0.99], p=0.047; 5 studies/1918 women), and maternal night blindness (RR 0.79, 95% CI [0.64 to 0.98], p=0.029; 2 studies/10,608 women). Vitamin A supplementation did not affect the risk of neonatal anaemia (RR 0.99, 95% CI [0.92 to 1.08], p=0.87; 1 study/409 women), preterm birth (RR 0.98, 95% CI [0.94 to 1.01], p=0.19; 5 studies/40,137 women), or the risk of having a low birth weight baby (RR 1.02, 95% CI [0.89 to 1.16], p=0.81; 4 studies/14,599).

Adverse events
There were no reports of side effects or adverse events in the included trials.

Vitamin A alone versus micronutrient supplements without vitamin A
In two studies including 591 women, vitamin A alone in comparison to micronutrients without vitamin A did not reduce the risk of maternal clinical infection (RR 0.99, 95% CI [0.83 to 1.18], p=0.91).

Vitamin A with other micronutrients versus micronutrient supplements without vitamin A
Supplementation with vitamin A and other micronutrients did not have a statistically significant effect on the risk perinatal mortality in comparison to treatment with micronutrients without vitamin A (RR 0.51, 95% CI [0.10 to 2.69], p=0.42; 1 study/179 women). There was also no evidence of an effect on the risk of stillbirth (RR 1.41, 95% CI [0.57 to 3.47], p=0.46; 2 studies/866 women), maternal anaemia (RR 0.86, 95% CI [0.68 to 1.09], p=0.22; 3 studies/706 women), maternal clinical infection (RR 0.95, 95% CI [0.80 to 1.13], 2 studies/597 women), preterm birth (RR 0.39, 95% CI [0.08 to 1.93], p=0.25; 1 study/136 women), congenital malformations (RR 0.34, 95% CI [0.04 to 3.18], 1 study/179 women), neonatal mortality (RR 0.65, 95% CI [0.32 to 1.31], p=0.23; 1 study/594 women), or neonatal anaemia (RR 0.75, 95% CI [0.38 to 1.51], p=0.43; 2 studies/1052 women). In one study of 594 women, the risk of low birth weight was reduced with vitamin A supplementation (RR 0.67, CI [0.47 to 0.96], p=0.027).

5. Additional author observations*

Overall, the risk of bias in the included trials was low to unclear, with all but three studies reporting adequate allocation concealment. For the main comparison (vitamin A alone versus placebo or no treatment), the quality of the evidence was assessed as being high for maternal mortality, perinatal mortality, and preterm birth; moderate for maternal anaemia; and low for maternal clinical infection. Populations studied differed with respect to baseline vitamin A status, which may have affected the results, and there were problems with follow-up in several large trials.

Current evidence does not support the supplementation of pregnant women with vitamin A to reduce maternal or perinatal mortality. However, antenatal vitamin A supplementation was demonstrated to reduce maternal anaemia in vitamin A-deficient populations and in HIV-infected pregnant women, and maternal night blindness and infection were also reduced, although data contributing to the latter finding were not of high quality.

Future trials need to assess baseline vitamin A status, as the effect of supplemental vitamin A is likely to depend on whether the participating population are vitamin A deficient or not. Further research on whether vitamin A supplementation reduces maternal infection is warranted, as are trials designed to determine the optimal dose of vitamin A and duration of supplementation for the prevention of maternal anaemia. Standardised international criteria for maternal and neonatal outcomes should also be employed in future research.