1. Objectives

To evaluate the effectiveness and safety of vitamin supplementation in women prior to conception, periconceptionally and in early pregnancy on the risk of spontaneous miscarriage

2. How studies were identified

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

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

In addition, reference lists of identified studies, relevant journals, and conference proceedings were handsearched

3. Criteria for including studies in the review

3.1 Study type

Randomized and quasi-randomized controlled trials, including cluster-randomized trials

3.2 Study participants

Pregnant women <20 weeks’ gestation or women planning to become pregnant in the near future, irrespective of whether they were at high or low risk of miscarriage

(Trials in which women were enrolled exclusively ≥20 weeks’ gestation or where the majority of women were enrolled after this time point were excluded)

3.3 Interventions

Any vitamin supplementation, either alone or in combination with other vitamins or interventions, in comparison to no treatment, placebo, other vitamins, or other interventions for the prevention of miscarriage, in areas where dietary intakes of that vitamin(s) are inadequate

3.4 Primary outcomes

• Total fetal loss, defined as combined total of early miscarriage (<12 weeks’ gestation), late miscarriage (≥12 to <24 weeks’ gestation) and stillbirth (pregnancy loss at ≥24 weeks’ gestation)
• Early or late miscarriage

Secondary outcomes included stillbirth, congenital malformations, and adverse effects of vitamin supplementation sufficient to discontinue supplementation (e.g., hypervitaminosis, headache, nausea, vomiting, diarrhoea)

4. Main results

4.1 Included studies

Forty randomized controlled trials, involving 276,820 women, were included in this review:

• Data from a total of 278,413 pregnancies were included in the review as two trials included women who were pregnant more than once during the study period
• Thirty-two studies were individually randomized (59,094 women with 51,146 pregnancies) and eight trials were cluster-randomized (217,726 women with 219,267 pregnancies)
• Timing of initiation and duration of supplementation varied widely, with some trials beginning preconceptionally and other trials enrolling women early to mid pregnancy, which in some cases included women ≥20 weeks’ gestation
• Interventions assessed included vitamin A alone or with iron, folic acid, multivitamins or β-carotene (7 trials); vitamin C with or without multivitamins or vitamin E (12 trials); folic acid with or without multivitamins and/or iron (8 trials); antioxidant vitamins; multivitamins with or without folic acid, vitamin A, vitamin E or iron and folic acid (17 trials); and multivitamins alone (2 trials)
• One trial was in HIV-infected women, one trial was in women with type 1 diabetes, one trial enrolled women at risk of pre-eclampsia, and one trial enrolled only nulliparous women

4.2 Study settings

• Australia, Bangladesh (3 trials), Brazil, Burkina Faso, Canada, China, Hungary, India, Indonesia (6 trials), Japan, Malawi (2 trials), Mexico, Nepal (2 trials), Niger, Nigeria (2 trials), Pakistan, Ireland, South Africa, Uganda, the United Kingdom of Great Britain and Northern Ireland (4 trials), the United Republic of Tanzania (2 trials), and the United States of America (3 trials). One trial involved 33 international centres and another trial was a multi-country study involving India, Peru, South Africa and Viet Nam
• Trials were conducted in both resource-rich and resource-constrained countries, and in both rural and urban settings

4.3 Study settings

How the data were analysed
Comparisons were separated on the basis of five main intervention categories: i) vitamin C supplementation (with or without other micronutrients) compared with placebo, no intervention, or other micronutrients without vitamin C; ii) vitamin A supplementation (with or without other micronutrients) compared with placebo, no intervention, or other micronutrients without vitamin A; iii) multivitamin supplementation (with or without other micronutrients) compared with placebo, no intervention, or another micronutrient; iv) folic acid supplementation (with or without other micronutrients) compared with placebo, no intervention, or other micronutrients without folic acid; and v) antioxidant vitamins supplementation (with or without other micronutrients) compared with placebo, no intervention, or other micronutrients without antioxidants. For trials recruiting women prior to pregnancy, the denominators were all women randomized; and where the number of women in each trial who became pregnant was known, the denominators were the number of women randomized with confirmed pregnancies. Fixed effects meta-analysis was used to generate pooled risk ratios (RR) with 95% confidence intervals (CI) where clinical and statistical heterogeneity were low, and random effects meta-analysis was used where heterogeneity was considered substantial. Sensitivity analyses were conducted by risk of bias for allocation concealment, and by excluding cluster-randomized trials. Data from studies with more than two treatment arms or studies that compared a vitamin intervention with another were only analysed in the subgroup analyses according to vitamin type. Where possible, subgroup analysis was performed as follows:

• By duration of supplementation, based on trial entry: <12 weeks’ gestation, 12 to 20 weeks’ gestation or both
• By dose of vitamin: below versus above the recommended dietary intake
• By risk of spontaneous miscarriage: high risk (presence of hyperhomocysteinemia, thrombophilia, antiphospholipid syndrome, systemic lupus erythematous) versus low risk
• By low or adequate dietary vitamin intake at trial entry

Results
Vitamin C supplementation
Primary outcomes
Total fetal loss did not differ between women receiving vitamin C supplementation and those that did not (vitamin C plus vitamin E: RR 1.14, 95% CI [0.92 to 1.40], 7 trials/18,949 women; vitamin C alone: RR 1.28, 95% CI [0.58 to 2.83], 2 trials/224 women; vitamin C plus multivitamins: RR 1.32, 95% CI [0.63 to 2.77], 1 trial/406 women). Likewise, there was no evidence of a difference between treatment groups in the risk of early or late miscarriage (vitamin C plus vitamin E: RR 0.90, 95% CI [0.65 to 1.26], 4 trials/13,346 women; vitamin C alone: RR 1.17, 95% CI [0.52 to 2.65], 2 trials/224 women; vitamin C plus multivitamins: RR 1.19, 95% CI [0.79 to 1.79], 2 trials/790 women).

Additional outcomes
There was no evidence of a difference in the risk of stillbirth among women receiving vitamin C plus vitamin E or vitamin C alone. Among women allocated to vitamin C plus vitamin E, the risk of congenital malformations or adverse events was not different to that of the placebo group.

Vitamin A supplementation
Primary outcomes
No difference in the risk of total fetal loss was found between women receiving vitamin A and control groups in analyses of vitamin A plus iron and folic acid (RR 1.01, 95% CI [0.61 to 1.66], 3 trials/1640 women), vitamin A alone (RR 1.05, 95% CI [0.90 to 1.23], 3 trials/52,480 women), β-carotene alone (RR 1.02, 95% CI [0.98 to 1.07], 2 trials/51,163 women), vitamin A or β-carotene (RR (RR 1.05, 95% CI [0.91 to 1.21], 1 trial/17,373 women), or vitamin A with or without multivitamins (RR 0.80, 95% CI [0.53 to 1.21], 1 trial/1074 women). Similarly, there was no evidence of a difference between treatment groups in the risk of early or late miscarriage (vitamin A plus iron and folic acid: RR 0.86, 95% CI [0.46 to 1.62], 2 trials/1397 women; vitamin A alone: RR 0.98, 95% CI [0.92 to 1.04], 1 trial/39,668 women; β-carotene alone: RR 1.00, 95% CI [0.94 to 1.06], 1 trial/39,860 women; vitamin A with or without multivitamins RR 0.76, 95% CI [0.37 to 1.55], 1 trial/1075 women).

Additional outcomes
No evidence of a difference between treatment groups was found for the risk of stillbirth, and no trials assessed congenital malformations or adverse effects of vitamin A supplementation.

Multivitamin supplementation
Primary outcomes
Multivitamins, given with or without vitamin A, reduced the risk of total fetal loss in comparison to treatment with vitamin A or placebo in one trial of 1074 women (RR 0.60, 95% CI [0.39 to 0.92], p=0.018). In addition, multivitamins without folic acid reduced the risk of total fetal loss in comparison to no multivitamins/folic acid (RR 0.49, 95% CI [0.34 to 0.70], p=0.000079; 1 trial/907 women). In all other analyses of the effect of multivitamin supplementation on total fetal loss or early or late miscarriage there were no differences between groups, including: multivitamins with folic acid versus no multivitamins or folic acid (total fetal loss: RR 0.91, [0.65 to 1.27]; early or late miscarriage: 0.95, 95% CI [0.67 to 1.34], 1 trial/1368 women); multivitamins with or without folic acid versus folic acid (total fetal loss: RR 0.96, 95% CI [0.70 to 1.33]; early or late miscarriage: RR 0.96, 95% CI [0.70 to 1.33], 2 trials/1644 women); multivitamins versus control (total fetal loss: RR 0.83, 95% CI [0.58 to 1.17]; early or late miscarriage: RR 0.83, 95% CI [0.58 to 1.17], 1 trial/5021 women); multivitamins with vitamin E versus multivitamins without vitamin E or control (total fetal loss: RR 0.92, 95% CI [0.46 to 1.83]; early or late miscarriage: RR 1.04, 95% CI [0.26 to 4.13], 1 trial/823 women); and multivitamins with iron and folic acid versus iron and folic acid (total fetal loss: RR 1.00, 95% CI [0.75 to 1.34]; early or late miscarriage: RR 0.99, 95% CI [0.72 to 1.38], 3 trials/6883 women).

Additional outcomes
The risk of stillbirth was reduced among women receiving multivitamins plus iron and folic acid compared to those receiving iron and folic acid only (RR 0.92, 95% CI [0.85 to 0.99], p=0.031; 10 trials/79,851 women). No other analyses of the outcomes stillbirth or congenital malformations revealed differences between treatment groups.

Folic acid supplementation
The risk of total fetal loss and early or late miscarriage were not reduced with folic acid supplementation in the following comparisons: folic acid with or without multivitamins versus no folic acid or multivitamins (total fetal loss: RR 0.97, 95% CI [0.69 to 1.35]; early or late miscarriage: RR 0.99, 95% CI [0.70 to 1.39], 1 trial/1364 women); folic acid with or without multivitamins versus multivitamins (total fetal loss: RR 1.14, 95% CI [0.82 to 1.57]; early or late miscarriage: RR 1.09, 95% CI [0.79 to 1.51], 2 trials/1644 women); folic acid and iron versus iron (total fetal loss: RR 0.23, 95% CI [0.01 to 4.59]; early or late miscarriage: RR 0.38, 95% CI [0.02 to 9.03], 1 trial/75 women); folic acid and iron versus no iron or folic acid (total fetal loss: RR 13.0, 95% CI [0.74 to 226.98]; early or late miscarriage: RR 13.0, 95% CI [0.74 to 226.98], 1 trial/160 women). There was also no difference detected between women receiving folic acid and those that did not for the outcomes congenital malformations and stillbirth.

Antioxidant supplementation
In one trial of 110 women comparing antioxidants with multivitamins versus multivitamins with a low antioxidant content, no evidence of a difference between treatment groups was noted in the risk early or late miscarriage (RR 1.12, 95% CI [0.24 to 5.29]), and no other outcomes of interest were reported on.

Subgroup and sensitivity analyses
Subgroup analyses could not be performed due to insufficient data. Results remained unchanged in sensitivity analyses by allocation concealment and by unit of randomization.

5. Additional author observations*

The methodological quality of some of the included trials was poor due to high rates of attrition or unclear allocation concealment. Definitions of the timing miscarriage varied across trials, which was attempted to be resolved using the outcome total fetal loss. In addition, supplementation in some trials was initiated in mid-pregnancy, which would have little effect on the risk of miscarriage. The quality of the evidence as assessed using GRADE criteria for vitamin C and vitamin E compared with control was high for total fetal loss, and moderate for early or late miscarriage, stillbirth, and adverse effects. For vitamin A plus iron and folic acid versus iron plus folic acid, the evidence was judged as low quality for total fetal loss, early or late miscarriage, and stillbirth. For multivitamins plus iron and folic acid versus iron plus folic acid, the evidence was judged as high quality for total fetal loss and stillbirth, and for early or late miscarriage was downgraded to moderate quality due to funnel plot asymmetry suggesting that smaller negative trials might have been missing from the published literature.

Overall, there was no evidence of an effect of supplementation with any single vitamin on the risk of early or late miscarriage in pregnant women. However, supplementation with multivitamins with or without iron and/or folic acid or vitamin A, may decrease the risk of total fetal loss and stillbirth.

Future trials should focus on women at high risk of miscarriage, should be of high methodological quality, should assess different doses and combinations of vitamins (i.e., individual vitamins or multivitamin preparations with or without vitamin A and folic acid), and should also assess potential harms, psychological effects, and long-term outcomes for both mother and child.