| |
Vitamin C and vitamin E in pregnant women at risk for pre-eclampsia (VIP trial): randomised placebo-controlled trial
|
| |
| |
"....Concomitant supplementation with vitamin C and vitamin E does not prevent pre-eclampsia in women at risk, but does increase the rate of babies born with a low birthweight. As such, use of these high-dose antioxidants is not justified in pregnancy.....our findings of an increase in low birthweight and no benefit with respect to risk of pre-eclampsia suggest a contraindication of the studied doses of vitamin C and vitamin E in pregnancy..."
The Lancet 2006; 367:1145-1154
April 8, 2006
L Poston a , AL Briley a, PT Seed a, FJ Kelly a and AH Shennan a, for the Vitamins in Pre-eclampsia (VIP) Trial Consortium
a. Maternal and Fetal Research Unit, Division of Reproductive Health, Endocrinology and Development and Pharmaceutical Sciences Research Division, King's College London, St Thomas' Hospital, London SE1 7EH, UK
Summary
Background: Oxidative stress could play a part in pre-eclampsia, and there is some evidence to suggest that vitamin C and vitamin E supplements could reduce the risk of the disorder. Our aim was to investigate the potential benefit of these antioxidants in a cohort of women with a range of clinical risk factors.
Methods: We did a randomised, placebo-controlled trial to which we enrolled 2410 women identified as at increased risk of pre-eclampsia from 25 hospitals. We assigned the women 1000 mg vitamin C and 400 IU vitamin E (RRR ƒ¿ tocopherol; n=1199) or matched placebo (n=1205) daily from the second trimester of pregnancy until delivery. Our primary endpoint was pre-eclampsia, and our main secondary endpoints were low birthweight (<2-5 kg) and small size for gestational age (<5th customised birthweight centile). Analyses were by intention to treat. This study is registered as an International Standard Randomised Controlled Trial, number ISRCTN 62368611.
Findings: Of 2404 patients treated, we analysed 2395 (99-6%). The incidence of pre-eclampsia was similar in treatment and placebo groups (15% [n=181] vs 16% [n=187], RR 0-97 [95% CI 0-80-1-17]). More low birthweight babies were born to women who took antioxidants than to controls (28% [n=387] vs 24% [n=335], 1-15 [1-02-1-30]), but small size for gestational age did not differ between groups (21% [n=294] vs 19% [n=259], 1-12 [0-96-1-31]).
Interpretation: Concomitant supplementation with vitamin C and vitamin E does not prevent pre-eclampsia in women at risk, but does increase the rate of babies born with a low birthweight. As such, use of these high-dose antioxidants is not justified in pregnancy.
INTRODUCTION
Pre-eclampsia affects 2-3% of all pregnancies, and every year is responsible for about 60 000 deaths worldwide.1 In the UK, hypertensive disease in pregnancy is a major cause of maternal death.2 Pre-eclampsia is a syndrome in which the placenta is implicated in the evolution of a generalised maternal inflammatory response, characterised by activation of maternal vascular endothelial cells and leucocytes.3,4 Markers of oxidative stress are present in the placenta and maternal circulation of affected women, suggesting it as a cause of the disorder. As such, antioxidants might help to prevent pre-eclampsia, though this notion has yet to be tested in large randomised trials.4,5
In 1999, we published the results of a small, randomised controlled trial6 in which we assigned women at risk of pre-eclampsia daily vitamin C (1000 mg) and vitamin E (400 IU) from 16-22 weeks' gestation. The primary outcome was a reduction in maternal concentrations of biomarkers of pre-eclampsia; the disorder itself was a secondary outcome. The trial was stopped early after an interim analysis showed a significant improvement in the primary outcome (plasminogen-activator inhibitor [PAI]1-to-PAI-2 ratio); rate of pre-eclampsia was also reduced. Women entered the trial mainly on the basis of an abnormal uterine artery doppler waveform (a recognised risk factor for pre-eclampsia). The findings of another, smaller study7 of women judged at risk of pre-eclampsia on the basis of their clinical history indicated no evidence of benefit with the same antioxidants.
Neither of these previous trials was powered to assess clinical outcomes. Our aim, therefore, was to assess whether supplementation with vitamin C and vitamin E prevents pre-eclampsia in women at increased risk. We also looked at rate of low birthweight and babies born small for gestational age, since maternal and fetal disease can be affected independently.
DISCUSSION
Our findings do not lend support to the hypothesis that vitamin C (1000 mg) and vitamin E (400 IU) supplements given prophylactically from the second trimester of pregnancy lead to a reduction in the rate of pre-eclampsia in women at risk for the condition. Instead, they indicate an association between supplementation with antioxidants and low birthweight, and an umbilical artery cord pH of less than 7-0, which could be attributable to the earlier onset of pre-eclampsia in this group than in controls and could explain why some women who took supplements delivered at an earlier gestational age. Since early onset disease is usually severe, this explanation might also account for the mothers' increased need for treatment with magnesium sulphate, intravenous antihypertensive therapy, and antenatal steroids. This notion would not explain, however, the increased risk of gestational hypertension seen in the treated group.
Growth restriction, as assessed by birthweight centiles, was not increased in women who took antioxidants, but there was a significant increase in growth-restricted babies in the subgroup of women with diabetes who received antioxidants, and the trends in most other risk groups did not favour the supplements. The increase in unexplained stillbirths in women who received antioxidants is a concern, but might be due to chance, since this was a post-hoc analysis. There were fewer deaths from immaturity in the antioxidant group.
The rate of pre-eclampsia in our population was similar to that reported elsewhere;16 treatment with antioxidants had no effect, irrespective of nature of risk at enrolment, although this analysis could have been underpowered in some groups.
Our population had higher plasma concentrations of vitamin C at baseline than previously reported by us in a similar cohort of women.6,17 Improved ability to stabilise samples before assay, a higher dietary intake of vitamin C, or greater use of multivitamin supplements could explain this difference. The lower concentrations of vitamin C in plasma during gestation in those who did, than in those who did not, develop pre-eclampsia in the placebo group lends support to the findings of previous work,3,4 which suggest that oxidative stress is associated with the disorder. Our present data also concur with those derived from our previous longitudinal study17 of vitamin C in women taking placebo.
This study is different from our previous, smaller trial in women at high risk of pre-eclampsia, though the same intervention and dose were used in both. That study was powered on a biochemical endpoint (PAI-1-to-PAI-2 ratios) not pre-eclampsia and recruited women mainly on the basis of abnormal uterine artery doppler waveform. The two protocols were almost identical, except that in the previous investigation gestational age at trial entry was later (16-22 weeks' gestation compared with 14-22 weeks') and, previously, the women screened were a predominantly low-risk population. The reduction in pre-eclampsia that we noted in our earlier trial, however, is unlikely to be peculiar to the population chosen, since in the present study we noted no association between the condition and abnormal doppler waveforms. The source of medication in the two trials was different, but we confirmed adequate supplementation by biochemical analyses in both. We believe that our previous conclusion of reduced pre-eclampsia in the treatment group occurred through a type 1 statistical error based on only 35 events.
A Cochrane metanalysis18 of antioxidant supplements for prevention of pre-eclampsia suggests a modest benefit, but includes studies in which micronutrients other than antioxidants were used. Two previous relevant, but small studies in high-risk women used vitamin C and vitamin E in the same doses as in this trial. Neither was powered to detect differences in pre-eclampsia, although that by Chappell and colleagues6 showed a significant reduction. Both studies reported a slight, non-significant reduction in the weight of babies born to mothers taking antioxidant supplements. In another study,19 vitamin C supplements or placebo were given to women at risk of preterm labour to investigate the hypothesis that gestation would be prolonged. The study was stopped early because of an increase in spontaneous preterm labour in the treatment group. We noted no such association. Finally, findings of a non-randomised study20 indicate that high-dose vitamin E could lead to a reduction in the birthweight of babies.
Our trial will inevitably raise questions about the safety of use during pregnancy of multivitamins that contain vitamin C and vitamin E. It is noteworthy, therefore, that birthweight centiles after correction for potentially confounding variables were significantly higher in those women in the placebo group taking pregnancy-related multivitamin supplements than in those who were not. This finding suggests that such preparations are not associated with a detrimental effect, and does not exclude a benefit, but this observation should be treated with caution, since this investigation was an exploratory analysis of a subgroup and not a predefined outcome. These observational data from the placebo group concur with other data,21 notably from developing countries, that suggest that multivitamin supplementation increases birthweight. Furthermore, the association we noted between higher plasma concentrations of vitamin C at baseline and higher birthweight, has also been reported in a cohort of Korean women.22
We do not know why supplementation with vitamin C and vitamin E to above physiological doses does not reduce the risk of pre-eclampsia, but increases the rate of low birthweight babies. A detrimental effect on placental function is possible, given the lower birthweight. Direct effects on fetal growth also cannot be excluded. Studies are underway to investigate the effect of the intervention on markers of oxidative stress and placental function.
The daily doses of vitamin C and vitamin E that we gave were below the maximum recommended intake in pregnant women. In the USA, the Institute of Medicine's Food and Nutrition Board have set an upper tolerable limit of vitamin C ingestion in pregnancy at 2000 mg per day and of vitamin E ingestion at 1000 mg (1600 IU).23 Neither antioxidant is contraindicated in pregnancy, although there is limited evidence on which to assess their safety. Diarrhoea is the only consistent side-effect of vitamin C (120-6000 mg daily) in non-pregnant individuals.24 In controlled trials25 in preterm infants of vitamin E supplementation for the treatment of retinopathy of prematurity, the supplement was associated with increased risk of sepsis and necrotising enterocolitis. Cochrane reviews of vitamin E26 and vitamin C24 supplementation in pregnancy emphasise the need to establish the safety of their use in high doses in pregnant women. The absence of benefit and evidence of unfavourable outcomes in this study cannot be extrapolated to other antioxidants, including the same vitamins at lower doses. Additionally, our findings should not detract from the potential importance of oxidative stress in pre-eclampsia.
Our results provide another example of the lack of efficacy of high-dose antioxidants in prevention of disease despite consistent evidence for a state of oxidative stress.3,4 Pre-eclampsia shares many of the characteristics of a proatherogenic state, including oxidative stress, dyslipidaemia, and endothelial cell activation, yet most of the evidence from studies of antioxidant supplementation in cardiovascular disease indicates no positive affect.27 There is also little evidence of harm, although a recent meta-analysis28 claimed that persistently high intake of vitamin E (>400 IU daily) increased mortality. It is noteworthy, however, that the methods used for analysis have been questioned.29
Careful consideration of underlying pathophysiology should accompany further investigation and direct ongoing clinical trials in this area. In the mean time, our findings of an increase in low birthweight and no benefit with respect to risk of pre-eclampsia suggest a contraindication of the studied doses of vitamin C and vitamin E in pregnancy.
Results
Figure 1 shows the trial profile. Of 2404 women enrolled and treated, we analysed 2395 (99-6%). Based on counts of returned pills from 2070 women, 80% (n=1653) of women took at least 50% of their tablets, 65% (n=1345) took 80% or more, and 32% (n=661) took all of their tablets; 6% (n=125) did not take any tablets. Neither compliance nor baseline characteristics differed greatly between groups (table 1) or subgroups. More than a quarter of women were taking multivitamin supplements at enrolment. Table 1 also shows the proportion of women with each identified risk factor for pre-eclampsia at baseline; 21% of women had more than one risk factor. Although only 3% of women entered the study with an abnormal uterine artery doppler waveform, 274 (11%) had a recorded abnormal waveform during the course of the study. We did not do this test systematically in trial participants nor is it routinely undertaken in clinical practice. As such, the true prevalence of abnormal waveforms is not known in this population.
Table 2 shows the primary and secondary maternal outcomes and table 3 the primary outcome, pre-eclampsia, stratified by risk at enrolment. Overall, 15% (n=368) of women developed pre-eclampsia. Treatment with vitamin C and vitamin E did not reduce this risk (table 2), irrespective of risk at enrolment (table 3). There was, however, a variation in the rate of pre-eclampsia, ranging from 11% in primiparous obese women and women with a multiple pregnancy to 32% in women with chronic renal disease (table 3). Women with multiple risk factors had a rate of pre-eclampsia of 26%. There was no difference in the risk of severe pre-eclampsia (pre-eclampsia with severe hypertension) or early onset pre-eclampsia (delivery for pre-eclampsia <34 weeks' gestation) between groups, although significantly more women in the treatment group than in the control group developed gestational hypertension and received intravenous antihypertensive therapy (table 2), antenatal steroids, or magnesium sulphate. Six women taking supplements developed HELLP syndrome compared with one on placebo (not significant). Other indicators of serious morbidity were similar in antioxidant versus placebo groups: stroke (0 vs 1), cerebral haemorrhage (none), pulmonary oedema (3 vs 2), disseminated intravascular coagulation (1 vs 1), epigastric pain (16 vs 21), liver haematoma or rupture (none), and admission to intensive-care unit (6 vs 9). Neither of the women who died did so because of pre-eclampsia (table 2); one woman committed suicide 5 days postnatally and the second died 6 months postpartum after an infection.
Table 4 shows all neonatal outcomes, and figure 2 effect of treatment on number of children born with a low birthweight or small for gestational age. Babies of women who took supplements were significantly more likely to be born with a low birthweight than those of controls (table 4, figure 2); small size for gestational age did not differ between groups. Within the nine risk subgroups, there were significantly more small and growth-restricted singleton babies born to women with diabetes who were taking antioxidants (low birthweight 20% [n=19] vs 10% [n=6], risk ratio [RR] 3-26 [95% CI 1-36-7-84]; small for gestational age 12% [n=12] vs 3% [n=3], 4-12 [1-20-14-21]; figure 2).
18% (n=367) of singletons were delivered preterm, and 14% (n=284) were small for gestational age. There was no difference in rate of preterm birth between groups, but babies born to mothers taking antioxidants had a higher incidence of arterial cord pH of less than 7. Although more babies died in the antioxidant group than in the placebo group, this statistic was not significant, and the causes of death were balanced (table 4). The trial steering committee requested further analysis of the cause of death after allocation was revealed; the findings of this analysis showed significantly more unexplained stillbirths (antepartum deaths from 24 weeks' gestation), but fewer deaths due to immaturity, in women who took vitamins, even after correction for multiple births. The rate of congenital malformations was similar in both groups (table 4).
With respect to plasma analyses, 1355 women had their concentrations of vitamin C measured at recruitment. Neither baseline concentrations of vitamin C (mean 68 mmol/L, SD 26) nor vitamin E (corrected for plasma cholesterol 5-6 mmol/mmol, SD 1-2) differed between groups (figure 3). In the 208 women who provided longitudinal blood samples throughout pregnancy, however, there was a significant increase in concentration of both vitamins at all follow-up points in those assigned antioxidants. This finding was independent of development of pre-eclampsia (figure 3). In this subgroup, treatment compliance was better than in the group as a whole (85% [169 of 200] taking at least 80% of tablets, compared with 65% [1345 of 2070]). Plasma concentrations of vitamin C were lower throughout the gestational period studied in those women who took placebo and developed pre-eclampsia (n=20) than in women who did not (n=81; p=0-043). Similar analyses showed no statistical differences between these groups in corrected plasma concentrations of vitamin E.
In view of the results for the main outcomes we did some additional analyses that were not in the predefined analysis plan for additional insight. Similar rates of pre-eclampsia in the treatment (n=181) and placebo (n=187) groups meant we were able to meaningfully compare the women. The mean gestational age at diagnosis of pre-eclampsia was more than a week earlier (34+0vs 35+1, difference 8 days, 95% CI 2-14) in women taking antioxidants and the mean gestational age at delivery was significantly earlier (35+6vs 36+4, difference 5 days, 0-10). These women were also more likely than controls to have received magnesium sulphate (23% [n=41] vs 12% [n=22], RR 1-93, 1-2-3-1) and antenatal steroids (38% [n=69] vs 28% [n=53], 1-35, 1-0-1-8). The highest quartile of vitamin C at baseline after correction for risk group, degree of education, and housing status and smoking status was associated with fewer infants born small for gestational age (0-42, 0-26-0-67), a lower rate of low birthweight babies (0-39, 0-24-0-63), and a lower rate of pre-eclampsia (odds ratio [OR] 0-59, 0-38-0-93). Vitamin C concentrations were higher by 15 mmol/L (95% CI 11-18) in 313 women taking pregnancy-specific multivitamins at trial entry (mean 79 mmol/L, SD 25) than in 1005 women taking none (65 mmol/L, SD 26). We did not include women taking other multivitamins in this analysis. The use of pregnancy-specific multivitamins was associated with lower rates of low birthweight (RR 0-72, 95% CI 0-56-0-93) and better birthweight centiles, but after correction for risk group at baseline, degree of education, and housing and smoking status remained significant only for birthweight centiles (OR 0-68, 0-52-0-88).
Methods
Participants
Between Aug 6, 2003, and June 27, 2005, we did a randomised controlled trial (the Vitamins in Pre-eclampsia [VIP] trial) to which we enrolled women with clinical risk factors for pre-eclampsia from 25 UK hospitals in ten geographical areas. The last baby was delivered on Dec 3, 2005. Eligible women could be referred to trial centres from any location in the UK. 13 women were recruited in Amsterdam, Holland.
Our inclusion criteria were gestational age 14+0 - 21+6 weeks plus one or more of the following risk factors: pre-eclampsia in the pregnancy preceding the index pregnancy, requiring delivery before 37 completed weeks' gestation, diagnosis of HELLP syndrome (haemolysis, elevated liver enzymes, and low platelets) in any previous pregnancy at any stage of gestation, eclampsia in any previous pregnancy at any stage of gestation; essential hypertension requiring medication, currently or previously; maternal diastolic blood pressure of 90 mm Hg or more before 20 weeks' gestation in the current pregnancy; type 1 or type 2 diabetes, requiring insulin or oral hypoglycaemic therapy before the pregnancy; antiphospholipid syndrome;8 chronic renal disease (creatinine ≥125 mmol/L pre-pregnancy or ≥100 mmol/L during pregnancy, or significant proteinuria [≥500 mg per 24 h]); multiple pregnancy; abnormal uterine artery doppler waveform (18-22 weeks' gestation, mean resistance index >0-67 or pulsatility index >1-65 with or without the presence of unilateral or bilateral diastolic notches); primiparity with body-mass index (BMI) at first antenatal appointment of 30 kg/m2 or more.
We excluded eligible women if they were unable or unwilling to give written informed consent or were being treated with warfarin. We only excluded women who were taking vitamin supplements if they contained doses of vitamin C of 200 mg or more or of vitamin E of 40 IU or more daily.
We employed study-specific research midwives in each geographical area, and each midwife covered up to three maternity units. The midwives instigated and maintained strategies to identify potentially eligible women at each centre from referral letters, high-risk antenatal clinics, and ultrasound departments, as well as from referrals from midwifery and medical colleagues. Recruitment occurred in various settings, including ultrasound departments, and, most frequently, antenatal clinics of participating centres, but also occasionally by telephone. We used a secure internet-based trial management system, which permitted both data collection and quality assurance from any computer terminal in all geographical locations (MedSciNet, AB Stockholm, Sweden). This system facilitated contemporaneous data management by the central trial team. The database was held on a server with backup at a remote site, and fulfilled all criteria of the Data Protection Act (1998). We kept identifying characteristics of participants on a secure and separate database.
The South East Multi Ethics Research Committee provided ethics approval (number 00/01/027), and we acquired site-specific approval from each participating centre. All participants gave written informed consent. Whenever possible trial visits coincided with routine antenatal appointments.
|
|
| |
| |
|
|
|
