World Health Organisation multicentre randomised trial of supplementation with vitamins C and E among pregnant women at high risk for pre-eclampsia in populations of low nutritional status from developing countries


  • The views expressed in this document are solely the responsibility of the authors and do not necessarily represent the views of the World Health Organization or its Member States.

  • Trial registration number: ISRCTN86677348
    Link to trial:

Dr J Villar, Nuffield Department of Obstetrics and Gynaecology, University of Oxford, John Radcliffe Hospital, Oxford, OX3 9DU, UK. Email


Objective  To determine if vitamin C and E supplementation in high-risk pregnant women with low nutritional status reduces pre-eclampsia.

Design  Multicentred, randomised, controlled, double-blinded trial.

Setting  Antenatal care clinics and Hospitals in four countries.

Population  Pregnant women between 14 and 22 weeks’ gestation.

Method  Randomised women received 1000 mg vitamin C and 400 iu of vitamin E or placebo daily until delivery.

Main outcome measures  Pre-eclampsia, low birthweight, small for gestational age and perinatal death.

Results  Six hundred and eighty-seven women were randomised to the vitamin group and 678 to the placebo group. Groups had similar gestational ages (18.1; SD 2.4 weeks), socio-economic, clinical and demographical characteristics and blood pressure at trial entry. Risk factors for eligibility were similar, except for multiple pregnancies: placebo group (14.7%), vitamins group (11.8%). Previous pre-eclampsia, or its complications, was the most common risk factor at entry (vitamins 41.6%, placebo 41.3%). Treatment compliance was 87% in the two groups and loss to follow-up was low (vitamins 2.0%, placebo 1.3%). Supplementation was not associated with a reduction of pre-eclampsia (RR: 1.0; 95% CI: 0.9–1.3), eclampsia (RR: 1.5; 95% CI: 0.3–8.9), gestational hypertension (RR: 1.2; 95% CI: 0.9–1.7), nor any other maternal outcome. Low birthweight (RR: 0.9; 95% CI: 0.8–1.1), small for gestational age (RR: 0.9; 95% CI: 0.8–1.1) and perinatal deaths (RR: 0.8; 95% CI: 0.6–1.2) were also unaffected.

Conclusion  Vitamins C and E at the doses used did not prevent pre-eclampsia in these high-risk women.


Pre-eclampsia/eclampsia is a leading cause of maternal and perinatal morbidity and mortality worldwide.1 Oxidative stress has been suggested as a pathogenic mechanism that leads to the disease.2,3 It was therefore logical to hypothesise that supplementation of women at risk for pre-eclampsia with antioxidants might prevent the disorder or blunt its severity. Unfortunately, and following a promising preliminary report,4 the results of three double-blinded randomised controlled trials failed to corroborate the earlier observation.5–7 It should be noted, however, that two of these trials were performed in well-nourished populations.5,6

Nutritional supplementation during pregnancy may theoretically prevent complications by two mechanisms.8 The supplemented nutrient can have a pharmacological effect in well-nourished women, a concept explored in recent trials4,6,9 or may achieve benefit through correction of a deficiency secondary to low dietary intake.10 The present trial evaluates the latter concept: whether or not supplementation with vitamins C and E prevents pre-eclampsia and low birthweight (LBW) among women at high risk for pre-eclampsia from communities at risk of poor nutritional status in developing countries.

Materials and methods

A multicentre, randomised, placebo-controlled, double-blind trial was performed between October 2004 and December 2006, at antenatal clinics that served populations with low socio-economic status and had evidence of overall low nutritional status from a previous WHO survey.11 These clinics, located in Nagpur, India; Lima and Trujillo, Peru; Cape Town, South Africa; and Ho Chi Minh City, Viet Nam form part of the WHO Maternal and Perinatal Research Network––all having extensive experience in conducting large multicentre randomised trials. The trial followed the research protocol used in the recently completed United Kingdom based multicentre trial of vitamins C and E (the VIP trial)5 with only minor adaptations to accommodate local resources. We designed and analysed the trial in coordination with the United Kingdom-VIP trial that included use of their Data Safety Monitoring Committee.

Pregnant women considered high risk for pre-eclampsia (chronic hypertension, renal disease, pre-eclampsia-eclampsia in the pregnancy preceding the index pregnancy requiring delivery before 37 weeks’ gestation, HELLP syndrome in any previous pregnancy, pregestational diabetes, primiparous with a body mass index ≥30 kg/m2, history of medically indicated preterm delivery, abnormal uterine artery Doppler waveforms and women with antiphospholipid syndrome) were considered eligible, and randomised between 14 and 22 weeks’ gestation to receive supplementation or placebo. Women were not eligible if ingesting vitamin supplements that contained ≥200 mg of vitamin C and/or ≥50 iu of vitamin E and women receiving warfarin or unable to give informed consent. We did not evaluate Doppler waveforms routinely during the trial in contrast with the United Kingdom trial. Women ingesting medications with aspirin-like compounds were not excluded.

Women were assigned randomly to receive vitamin C and E tablets or identical placebos from enrolment to delivery, the assigned prescription continued even after pre-eclampsia or hypertension was diagnosed.5 Custom Pharmaceuticals (Hove, East Sussex, UK) prepared vitamin C (1000 mg) and identical placebo tablets (microcrystalline cellulose) with addition of tartaric and citric acid to provide similar acidic taste and Banner Pharmacaps (Europe BV, Tilburg, the Netherlands) prepared identical gelatin capsules containing 400 iu natural source vitamin E (RRR-α-tocopherol) or a placebo (sunflower seed oil). DHP Investigational Medicinal Products Clinical Trials Supplies (Crickhowell, Powys, Wales and UK) packaged the tablets and capsules sealed in blister strips, each with a 1-week supply.5 The active and placebo tablets for each vitamin were identical in form, colour and taste and were provided in boxes containing four blister packs, each marked Monday to Sunday. The women were instructed to take one tablet and one capsule daily and to leave unused tablets or capsules in the blister and to return the blisters at the subsequent trial visit, regardless of whether all tablets and capsules had been taken. Women who failed to return their blisters packs or missed scheduled visits were contacted by phone or home visits.

Randomisation was performed by the statisticians of the British VIP Trial. The randomisation sequence was blocked by centre in groups of two to ten individuals (mean size 6·7). Copies of the randomisation sequence were provided to the company in charge of data management (MedSciNet AB, Stockholm, Sweden) and DHP.

Data were recorded on specifically designed forms, before transfer to the internet-based data management system. All blood pressure equipment was serviced locally, prior to the study, and calibrated periodically. Blood pressure recorded with a mercury sphygmomanometer was determined as the average of two measurements recorded at 3-minute interval with the subject seated for ≥5 minutes with the cuff at the level of the heart on the right arm. Principal investigators and nurse-coordinators, most of whom had been certificated for measuring blood pressure during our previous similar trial10 by WHO, Geneva, in a course conducted by Shared Care, Inc. (CA, USA), were retrained and reevaluated for this trial. Midstream urines, obtained at all antenatal visits, on admission in labour, or before elective caesarean section, were screened for proteinuria, the latter verified through urethral catheterisation if the membranes had ruptured. The WHO document, How to measure blood pressure and proteinuria,12 previously used by the clinics’ staff was available to all personnel participating in the trial and used as part of the manual of operation.

Routine antenatal care protocols, standardised by the WHO/RHR Maternal and Perinatal Health Unit during previous randomised trials, were followed and supported by local treatment practices for complications. Standardised methodologies, maintenance of equipment and quality control procedures were similar to those in our previously described trials.10 Once pre-eclampsia or eclampsia occurred, the patient was treated per standard protocols at the participating hospitals, but data collection continued until patient and/or newborn was discharged. Weight and height were recorded at booking and questions were posed at each visit about treatment compliance. Gestational age and birthweight were recorded at delivery by local staff also blinded to supplementation.

Data management was based on a secure internet-based trial management system, which allowed data entry at the local site and quality control. We have recently used this internet-based system in a large multicentre study in developing countries,13 and it is identical to that of the vitamins C and E trial conducted in the UK.5 The database was held on a server with backup at the central unit in Stockholm, while the identifying characteristics and randomisation code were kept in a secure and separate database in that server unavailable to the research team, until all data analyses were completed. All data were collected and used within the context of the UK Data Protection Act. No details of any particular woman were released to any other organisation.


The maternal primary outcome was occurrence of pre-eclampsia, defined as de-novo hypertension plus new-onset proteinuria, both after gestational week 20, and up to 72 hours postpartum, excluding labour. Gestational hypertension was defined as two or more readings of diastolic blood pressure ≥90 mmHg (Korotkoff V) taken ≥4 hours, but <168 hours apart and occurring after 20 weeks of pregnancy or in the early postnatal period and excluding labour. Proteinuria was defined as excretion of ≥300 mg in 24 hours or two readings of >1+ on dipstick of MSU/CSU if a 24 h collection was not available. Severe gestational hypertension was defined as two diastolic blood pressure measurements ≥110 mmHg recorded at least 4 hours, but <168 hours apart or a single diastolic pressure level ≥120 mmHg or if the first measurement was immediately followed by treatment with an antihypertensive.

Severe pre-eclampsia was diagnosed when blood pressures were ≥160 mmHg systolic and/or ≥110 mmHg diastolic on two occasions, at least 4 hours but not more than 168 hours apart, or if the first measurement was immediately followed by treatment with an antihypertensive, either of these scenarios being associated with the presence of proteinuria. Early onset pre-eclampsia was defined as the disorder appearing before 32 weeks’ gestation. Eclampsia was defined as a seizure in a woman with pre-eclampsia in the absence of a known or subsequently diagnosed convulsive disorder. Eclampsia was considered jointly with pre-eclampsia as a primary outcome and independently as a secondary outcome. Placental abruption was also a secondary outcome, with the data for this extracted from records and/or pathological reports. For women previously hypertensive, pre-eclampsia was defined by the de-novo appearance of proteinuria as above. For women with pre-existing proteinuria, pre-eclampsia was defined as de-novo hypertension or by the identification of clinical or biochemical markers of pre-eclampsia, for example, HELLP syndrome, eclampsia.

Primary neonatal outcomes were low birthweight (LBW) (<2500 g), small for gestational age (<10th centile of the WHO recommended standard14) and intrauterine or neonatal death before hospital discharge. Secondary neonatal outcomes were preterm delivery (<37 weeks), early preterm delivery (<34 weeks), very LBW (<1500 g), ≥7 days in the neonatal intensive care unit and congenital malformations.

The expected sample size of 1740 pregnancies (500 for each centre except South Africa: 240), was determined by protocol to have 75% power to detect a 30% reduction in pre-eclampsia from 15% in the placebo to 10.5% in the vitamins group at α equal to 0.05. Upon learning of safety concerns reported in other trials,5,6 the Data Safety and Monitoring Committee requested and performed an interim analysis during 2006 and recommended continuing the trial. Nevertheless, one centre stopped recruitment earlier, precluding attainment of the planned sample size. However, the sample size of 1365 pregnancies reached had an overall pre-eclampsia rate of 24%, substantially greater than the 15% anticipated and used in the original power calculations. Our final sample size has statistical power of 88% to detect a reduction of 30% (as dictated by the protocol) from the 24% pre-eclampsia rate to a rate of 17%.

Data from all subjects were included, irrespective of compliance or follow-up failures (intention-to-treat principle). Baseline characteristics were used to compare groups at trial entry. Comparisons of outcomes were performed using relative risk and 95% CI. Only subgroup analyses specified in the protocol were undertaken for the primary outcomes based on the risk factors considered for inclusion criteria. All analyses conducted (overall and stratified) followed the approved trial protocol and all end-points studied are reported in the study. No additional exploratory analyses were conducted. Only in the event of a significant overall effect of vitamin E and C supplementation on the primary outcomes, and according to the protocol, did we plan to investigate the effect of compliance levels on the treatment effect, that is, a ‘Dose Effect’ relationship. All analyses were performed using stata 8/se (

This trial was approved by the Scientific and Ethical Review Group of the UNDP/UNFPA/WHO/World Bank Special Programme for Research, Development and Research Training in Human Reproduction, the WHO Secretariat Committee on Research involving Human Subjects, and the Institutional Review Boards of participating centres. Government approval was obtained as required for centres in India and Peru. Good Clinical Practice procedures15 were followed in all centres and monitored by the coordinating unit on a regular basis.


Of 1445 women screened, 1381 were eligible of whom 1365 agreed to participate and were randomised (687 vitamins group; 678 placebo group). Pre-eclampsia information was unavailable for 14 women (2%) in the vitamins and 9 (1.3%) in the placebo group, but the remainder of the data from these women was included in the analyses (Figure 1). There were data from 79 supplemented and 94 placebo-treated women with multiple pregnancies, for whom newborn outcomes were considered separately. We also conducted a sensitivity analysis for primary outcomes excluding these multiple births.

Figure 1.

 Trial profile: all numbers in the figure represent women (multiple births are counted as one pregnancy).

Overall, women were short (mean height 1.53 m in both groups) and had low early pregnancy weight (mean 62.4 kg) supporting the concept of a moderately undernourished population. The mean BMI among women that were enrolled because risk factors other than BMI >30 kg/m2 was 25.1 (6.8) kg/m2 and 24.9 (6.2) kg/m2 in the vitamins and placebo groups, respectively. The treatment groups were well balanced for baseline characteristics at entry, there being no biologically relevant differences in age, smoking, maternal weight and height, socio-demographic and obstetrics characteristics, systolic and diastolic pressure and gestational age at randomisation (Table 1). Risk factors for trial eligibility were also similar in the two groups, with the most common risk factor being a history of pre-eclampsia followed by chronic hypertension and primiparity with BMI ≥ 30 kg/m2. Women carrying multiple fetuses comprised 11.8% of the supplemented group, which was lower than in the placebo group (14.7%) (Table 2). The groups were also well balanced within each study site. Treatment compliance was calculated as the proportion of tablets not returned in the blisters over the total number of tablets given to each woman. Compliance was similar in both groups (median 87%). Follow-up losses were low and also similar between groups (vitamins 2.0%, placebo 1.3%).

Table 1.   Baseline characteristics at trial entry according to supplementation group
(n = 687)
(n = 678)
  1. Presented are means (SD) or percentages.

Maternal age (years)27.6 (5.6)27.6 (5.0)
Maternal age <20 years (%)4.86.0
Marital status (single) (%)10.08.5
Ever smoked (%)16.316.5
Years of schooling (<10 years) (%)17.017.0
Working outside home (%)20.121.1
Primigravida (%)16.417.0
Maternal weight (kg)62.4 (18.1)61.8 (17.3)
Maternal height (cm)153.4 (6.4)153.0 (6.4)
Systolic blood pressure at randomisation (mmHg)113.4 (15.7)113.2 (16.0)
Diastolic blood pressure at randomisation (mmHg)71.5 (12.1)71.4 (12.3)
Proteinuria at randomisation (≥++) (%)1.01.0
Treatments or supplements
Aspirin (%)34
Heparin (%)0.10.1
Folate (%)16.317.8
Routine antenatal multivitamins (%)35.536.1
Gestational age at entry (weeks)18.1 (2.4)18.2 (2.5)
Table 2.   Distribution of risk groups at trial entry according to supplementation group*
 Vitamins group
(n = 687) N (%)
Placebo group
(n = 678) N (%)
  1. *Some women had more than one risk factor.

Pre-eclampsia in the previous pregnancy217 (31.6)205 (30.2)
HELLP syndrome or eclampsia in any previous pregnancy67 (9.7)73 (10.8)
Chronic hypertension163 (23.7)170 (25.1)
Diabetes40 (5.8)32 (4.7)
Chronic renal disease16 (2.3)12 (1.8)
BMI > 30 kg/m2 in primiparous women128 (18.6)121 (17.8)
Any abnormal uterine artery Doppler waveform37 (5.4)40 (6.0)
Multiple pregnancy81 (11.8)100 (14.7)

The incidence of pre-eclampsia was 24.1% in the vitamins and 23.3% in the placebo group (RR: 1.0, 95% CI: 0.9–1.3). Severe pre-eclampsia, gestational hypertension and severe gestational hypertension were also similar in the two groups. Eclampsia and HELLP syndrome appeared to be more common in the vitamins group, but placental abruption was lower in the vitamins group but these differences did not achieve statistical significance. There were very few maternal deaths or admission to intensive care unit events, precluding meaningful evaluation (Table 3). Adjusting the results for maternal age using logistic regression analysis and excluding women carrying multiple fetuses did not modify these patterns.

Table 3.   Maternal outcomes according to supplementation group
 Vitamins group
n/N (%)
Placebo group
n/N (%)
RR95% CI
Pre-eclampsia164/681 (24.1)157/674 (23.3)1.00.9–1.3
Severe pre-eclampsia22/681 (3.2)29/674 (4.3)0.80.4–1.3
Eclampsia3/681 (0.4)2/674 (0.3)1.50.2–8.9
HELLP syndrome10/681 (1.5)8/674 (1.2)1.20.5–3.1
Placental abruption6/681 (0.9)9/674 (1.3)0.70.2–1.8
Gestational hypertension64/681 (9.4)52/674 (7.7)1.20.9–1.7
Severe gestational hypertension18/681 (2.6)19/674 (2.8)0.90.5–1.8
Any maternal admission to intensive care1/681 (0.01)5/674 (0.07)0.20.02–1.7
Maternal death  0  1

Stratified analyses by risk categories at trial entry were conducted as per protocol for the primary maternal outcome: pre-eclampsia (Table 4). Risk was almost identical in the supplemented and placebo-ingesting women enrolled because they had chronic hypertension, previous pre-eclampsia or carried multiple pregnancies. There was a marginally statistically significant reduction of pre-eclampsia among overweight women receiving vitamins (RR: 0.6; 95% CI: 0.3–1.0) (Table 4). Overall, among women with any previous hypertensive condition the risk of pre-eclampsia was similar in both groups (RR: 1.1.; 95% CI: 0.8–1.3).

Table 4.   Risk of pre-eclampsia according to supplementation group and trial entry criteria*
 Vitamins group n/N (%)Placebo group n/N (%)RR95% CI
  1. *Some women had more than one risk factor.

Pre-eclampsia in the previous pregnancy160/216 (74.1)157/204 (76.9)1.00.9–1.1
HELLP syndrome or eclampsia in any previous pregnancy 13/67 (19.4) 21/73 (28.8)0.70.4–1.2
Chronic hypertension 53/163 (32.5) 52/170 (30.6)1.10.8–1.5
Any previous hypertensive condition117/415 (28.2)105/396 (26.5)1.10.8–1.3
Diabetes  0/40  0/32 
Chronic renal disease  0/16  0/12 
BMI ≥ 30 kg/m2 in primiparous 16/128 (12.5) 27/121 (22.3)0.60.3–1.0
Any abnormal uterine artery Doppler waveform  0/37  0/40 
Multiple pregnancy 23/81 (28.4) 23/100 (23.0)1.20.7–2.0

The primary neonatal outcomes studied were LBW, small for gestational age and perinatal death. When all newborns, including those from multiple-fetus gestations were considered, the rate of LBW was 33.2% in the vitamins group and 36.4% in the placebo group (RR: 0.9; 95% CI: 0.8–1.1); the rate of small for gestational age (<10th percentile) was 23.8% in the vitamins group and 26% in the placebo group (RR: 0.9; 95% CI: 0.8–1.0). Perinatal mortality was 74.4 per 1000 live births in the vitamins group and 89.2 per 1000 live births in the placebo group (RR: 0.8; 95% CI: 0.6–1.2). None of these comparisons reached statistically significant levels (Table 5).

Table 5.   Neonatal outcomes according to supplementation group
 Vitamins group
n/N (%)
Placebo group
n/N (%)
RR95% CI
  1. *According to 14.

  2. **Neonatal intensive care unit.

Small for gestational age (<10th centile)*141/592 (23.8)149/573 (26.0)0.90.8–1.1
Preterm delivery <37 weeks188/674 (27.9)213/669 (31.8)0.90.7–1.0
Delivery for pre-eclampsia <37 weeks66/674 (9.8)76/669 (11.4)0.90.6–1.2
Preterm delivery <34 weeks73/674 (10.8)93/669 (13.9)0.80.6–1.0
Delivery for pre-eclampsia <34 weeks29/674 (4.3)31/669 (4.6)0.90.6–1.5
Low birthweight <2500 g250/753 (33.2)277/762 (36.4)0.90.8–1.0
Low birthweight <1500 g62/753 (8.2)76/762 (10.0)0.80.6–1.1
Any admission to NICU**64/753 (8.5)80/762 (10.5)0.80.6–1.1
>7 days in NICU**31/753 (4.1)36/762 (4.7)0.90.5–1.4
Perinatal death56/753 (7.4)68/762 (8.9)0.80.6–1.2
Any congenital malformation19/753 (2.5)12/762 (1.6)1.60.8–3.3

Among secondary outcomes, preterm delivery (<37 weeks) and very preterm delivery (<34 weeks) and preterm delivery related to pre-eclampsia tended to be lower in the intervention group (RRs were between 0.8 to 0.9), while there were more clinically evident congenital malformations in the offspring of supplemented women (2.5% in the vitamins group versus 1.6% for the placebo group; RR: 1.6; 95% CI: 0.8–3.3) but, again, none of the comparisons was statistically significant (Table 5).

Protocol-directed stratified analyses by maternal risk factors focused on the three neonatal primary outcomes (low birthweight, small for gestational age and perinatal death). Risks were virtually identical in the previous pre-eclampsia, eclampsia, HELLP syndrome or chronic hypertension subgroups and in women with multiple fetuses. Supplementation appeared to reduce LBW in women with chronic hypertension (RR: 0.8; 95% CI: 0.61.0) and very low birthweight (<1500 g) among women with BMI > 30 kg/m2 (RR: 0.4; 95% CI: 0.11.0).

As we observed lower numbers of multiple births in the supplementation group, we conducted a sensitivity analysis including only singletons with more than 22 weeks of gestation. The relative risks were almost identical to those in the overall population.

Any adverse events requiring a formal report from the clinics to the trial coordinating unit (Adverse Event Report Form) were similar. Rates were 4.9 and 4.3% in the vitamins and placebo groups respectively.


We performed a double-blinded randomised trial to evaluate the effect of vitamin C and E supplementation in pregnant women at high risk for pre-eclampsia attending antenatal care clinics serving a pregnant population with documented low nutritional status. We hypothesised that this population should benefit most if vitamins C and E supplementation were to prevent pre-eclampsia. The results, however, were similar to the two previous trials in populations with adequate nutritional status conducted in Australia and in the UK5,6 and to another trial conducted in Brazil,7 all failing to demonstrate beneficial effects of the supplementation with respect to maternal or neonatal primary outcomes. On the other hand, and unlike the United Kingdom study5 in which the incidence of low birthweight was greater in the supplemented group, we found no evidence of harm attributable to supplementation with vitamins C and E.

Our trial followed a protocol virtually identical to that of the United Kingdom’s VIP Trial5 complementing that study by probing a different population, one at nutritional risk and drawn from populations of different ethnic origin. We followed standardised procedures, used electronic data entry and used well-trained staffs with considerable experience from participating in other large WHO multicentre trials. Similarly to our previous trials in developing countries,10 most women approached consented to participation and treatment compliance (87%) was high while follow-up loss was very low in both groups. Although we reached a sample size 20% smaller than planned, the higher than anticipated pre-eclampsia rate observed, provided sufficient power (88%) to detect the difference between groups projected in the protocol. The high rates of pre-eclampsia (24% compared with 16% in the United Kingdom trial also among high-risk women5), may reflect the low socio-economic and nutritional status including very low calcium intake10,11 of our study populations that were, in addition, being provided with a very intense antenatal care follow-up, seldom offered to such high-risk women. The incidence of pre-eclampsia (74%) among women who had pre-eclampsia in a previous pregnancy was also high, but similar to recently reported recurrence risk.16,17

We did not document circulating vitamin levels in the participants, although such data would have helped in evaluating compliance and exploring mechanisms of effect. However, the trial’s primary aim was the supplements’ ability to influence substantive clinical outcomes rather than serum levels. Also, had supplementation been proven effective, applying results to clinical practice would not require screening pregnant women for baseline vitamin levels as a requirement to initiate treatment.

The study groups’ low nutritional intake was based on the evaluation of the pregnant population served by these clinics and not by a direct evaluation of the intake of the actual women randomised. This assessment was based on data collected by us in these same clinics10,11,18 focusing on calories, protein and calcium intake. We assumed that vitamin C and E intake follows the pattern of the other major nutrients. However, we did document in the randomised population low maternal height (a very good indicator of chronic nutritional deficiency), low early pregnancy weight and high rates of small for gestational age (close to 25%). The mean maternal height and weight were considerably below those in the United Kingdom trial.5

With the above limitations in mind, we note that our results are in agreement with the two large trials recently published (5–7), demonstrating that independent of the study population, including its risk level and nutritional status, VITAMINS C AND E supplements are unlikely to prevent pre-eclampsia, low birthweight, small for gestational age or perinatal death. Therefore, such supplements should not be recommended for clinical practice.

A lack of effectiveness in this and previous trials makes the issue that we did not observe side-effects less relevant. No ineffective treatment should be provided to patients, even if there were no known evidence of harmful effects. However, previous trials did suggest that supplementing with vitamins C and E might increase risk.5,6 And while we did not observe adverse effects, such as earlier and more severe pre-eclampsia and greater fetal loss, the wide confidence intervals for several outcomes do not completely exclude such possibility. It has been suggested that adverse outcomes in trials performed in nutritionally replete populations5,6 might be because of ‘hypervitaminosis’, which would have been less likely in our undernourished population.

How can we explain the discrepancies between laboratory results suggesting a causal path and the results from randomised trials of vitamins C and E supplementation? It was postulated that supplementation with vitamins C and E would reduce maternal and placental oxidative activity and thus reduce oxidative stress. Perhaps previous data from small samples in studies evaluating oxidative stress were over-interpreted, influenced, in part, by publication bias of negative results. Alternatively, the observed biochemical alterations suggesting an association between oxidative stress and pre-eclampsia may reflect a role in the pathophysiology of the disease but not in its causal pathway. In the latter case, it would be unlikely that reversing the oxidative stress would decrease the incidence of pre-eclampsia and might be responsible for some of the harmful effects of supplementation observed in populations with good nutritional status.

Finally, high oxidative stress might be causal for pre-eclampsia, but vitamins C and E supplementation starting later in gestation, after placentation has occurred, may be ‘too little, too late’ to prevent the clinical appearance of the condition. Also we cannot discount an effect of inadequate dose of vitamins C and E nor that other antioxidants might be more effective. Administration of antioxidants which provide a ‘broad-brush’ to target oxidative stress during pregnancy may be inappropriate. Development of specific agents to target the cellular signalling pathways responsible for superoxide synthesis, for example NADPH oxidase may offer a more effective approach in the future.

To conclude, we believe that this and the other recently published trials5,6 are a good example of a rapid clinical evaluation of a promising preventive nutritional intervention. They provide solid evidence of its lack of effectiveness and discourage the use of these supplements during pregnancy for preventing pre-eclampsia and impaired fetal growth.

Conflict of interest

No conflicts of interest.

Contribution to authorship

José Villar, Mario Merialdi, Andrew Shennan and Lucilla Poston were responsible for the conception of the trial and prepared the protocol. José Villar and Mario Merialdi supervised and coordinated the trial’s execution. Mario Merialdi and José Villar with support from Marius Kublickas were responsible for data monitoring and management. Mario Merialdi and José Villar conducted the analysis. Manorama Purwar, Nelly Zavaleta, Nguyen Thi Nhu Ngoc, Annemarie De Greeff and John Anthony collaborated in the preparation of the protocol and coordinated the trial implementation in their respective countries; they actively contributed to the overall undertaking of the study. José Villar wrote the paper with input from all the investigators. All the investigators read this report and made substantive suggestions on its content.


Supported by UNDP/UNFPA/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction, Department of Reproductive Health and Research, World Health Organization. The Cape Town, South Africa, study site was supported by funds provided by the United Kingdom authors.

We are indebted to Annette Briley for her important contributions to the data management of the trial. Marshall Lindheimer made important contributions during the preparation of the paper. We would also like to thank Paul Seed for the statistical advice he provided.

The members of the Vitamin E and C Supplementation Trial Data Safety and Monitoring Committee were: Peter Brockelhurst (Chairman), André Van Asache, Douglas G. Altman. We thank Paul FA Van Look for his critical review of the paper in its final stages.

Participating institutions and staff

Nagpur, India

M Purwar, country coordinator, C Doifode, S Fusey, C Sarodey, J Lalani, R Waghmare, S Zodpey, S Ughade, R Patil, P Attal, S Salve, S Mundle, Residents OBGYN, and Nursing staff of indoor wards and antenatal clinics of the Government Medical College and Hospital Nagpur, India.

Lima, Peru

N Zavaleta, country coordinator, M Huatuco, L Haro, T Enco, R Wong, S Chávez, S Ricco, T García, (Instituto De Investigación Nutricional).

O Chumbe, K Espinoza, O Guzmán, I Ramírez, R Hinojosa, C Dávila, T Galarza J Guerola, G Yucra, S Palian, L Gutierrez, N Cordero, L Chamberga, L Aguirre, N Casas, R Motta, L Untiveros, A Pinedo, P Cardoso, C Coveñas, S ninaquispe, C Solano, R Ponce, H Vicuña, M Obeso, B Paredes, D Zuta, J Alva, J huatuco, S Carcía (Health centres, Ministry of Health, Peru).

Ho Chi Minh City, Viet Nam

Ngoc NTN, Hieu NT, Hanh Le T and staff at the antenatal clinic of Hungvuong Hospital, Ho Chi Minh City.

Cape Town, South Africa

J Anthony, A de Greeff, A de Groot.