Vous êtes sur la page 1sur 8

Antioxidant Therapy to Prevent Preeclampsia

A Randomized Controlled Trial

Joseph A. Spinnato II, MD, Salvio Freire, MD, Joao Luiz Pinto e Silva, MD,
Marilza Vieira Cunha Rudge, MD, Sergio Martins-Costa, MD, Matthew A. Koch, MD,
Norman Goco, Cleide de Barros Santos, MD, Jose Guilherme Cecatti, MD, Roberto Costa, MD,
Jose Geraldo Ramos, MD, Nancy Moss, PhD, and Baha M. Sibai, MD

OBJECTIVE: To study whether antioxidant supplementa- reduction in the risk of preeclampsia, assuming a placebo
tion will reduce the incidence of preeclampsia among group rate of 21% and !".05. The ! level for the final
patients at increased risk. analysis, adjusted for interim looks, was 0.0458.
METHODS: A randomized, placebo-controlled, double- RESULTS: Outcome data for 707 of 739 randomly as-
blind clinical trial was conducted at four Brazilian sites. signed patients revealed no significant reduction in the
Women between 12 0/7 weeks and 19 6/7 weeks of rate of preeclampsia (study drug, 13.8% [49 of 355]
gestation and diagnosed to have chronic hypertension or compared with placebo, 15.6% [55 of 352], adjusted risk
a prior history of preeclampsia were randomly assigned ratio 0.87 [95.42% confidence interval 0.611.25]). There
to daily treatment with both vitamin C (1,000 mg) and were no differences in mean gestational age at delivery
vitamin E (400 International Units) or placebo. Analyses or rates of perinatal mortality, abruptio placentae, pre-
were adjusted for clinical site and risk group (prior term delivery, and small for gestational age or low birth
preeclampsia, chronic hypertension, or both). A sample weight infants. Among patients without chronic hyper-
size of 734 would provide 80% power to detect a 40% tension, there was a slightly higher rate of severe pre-
eclampsia in the study group (study drug, 6.5% [11 of 170]
compared with placebo, 2.4% [4 of 168], exact P".11,
From the University of Cincinnati College of Medicine, Cincinnati, Ohio; odds ratio 2.78, 95% confidence interval 0.79 12.62).
Universidade Federal de Pernambuco, Hospital das Clnicas, Recife, Brazil;
Universidade Estadual de Campinas, Campinas, Brazil; Universidade Estadual
CONCLUSION: This trial failed to demonstrate a benefit
Paulista, Botucatu, Brazil; Universidade Federal Do Rio Grande Do Sul, of antioxidant supplementation in reducing the rate of
Hospital de Clnicas, Porto Alegre, Brazil; RTI International, Research Triangle preeclampsia among patients with chronic hypertension
Park, North Carolina; and National Institute of Child Health and Human and/or prior preeclampsia.
Development, Bethesda, Maryland.
Supported by Grant Number 1 U01 HD40565 cosponsored by the National
Institute of Child Health and Human Development and the Bill and Melinda www.ClinicalTrials.gov, NCT00097110
Gates Foundation. (Obstet Gynecol 2007;110:13118)
The authors thank Jutta Thornberry, Janet Bartz, Steve Litavecz, and Ty Hartwell, LEVEL OF EVIDENCE: I
RTI International; Susie Meikle, our initial National Institute of Child Health and
Human Development Project officer; and members of our site research teams:

Cincinnati: Les Myatt; Recife: Elias Ferreira de Melo, Antonio Carlos Barbosa
n response to evidence for the role of oxidative
Lima, Angelo Manoel Barreto, Jose Remgio Neto, Eduardo Costa Ramos; Botucatu:
Jose Carlos Peracoli, Joelcio Francisco Abbade, Anice Vieira de Camargo Martins, stress in the pathophysiology of the endothelial
Grasiela Bossolan, Kleber Campos, Tania Prevedel; Porto Alegre: Melissa Prade damage seen in preeclampsia,1 several clinical studies
Hemesath, Cristiano Dhil Zaffari; Campinas: Fernanda G Surita, Eliana Amaral,
Mary A. Parpinelli, Fabiana Krupa. Additionally, we thank Soubhi Kahhale and
have been reported that attempt to improve the
his research team at the University of Sao Paulo, Brasil, for their important initial antioxidant capability of women at risk for pre-
contributions to this research. eclampsia. In 1999 Chappell et al2 reported the results
Corresponding author: Joseph A. Spinnato II, MD, Department of Obstetrics and of their randomized study of 283 women who were
Gynecology, University of Cincinnati College of Medicine, PO Box 2670526, identified as being at an increased risk by abnormal
Cincinnati, OH 40267-0526; e-mail: spinnaja@ucmail.uc.edu.
two-stage uterine artery Doppler analysis or a previ-
Financial Disclosure
The authors have no potential conflicts of interest to disclose.
ous history of preeclampsia. Preeclampsia occurred in
24 (17%) of 142 women in the placebo group and 11
2007 by The American College of Obstetricians and Gynecologists. Published
by Lippincott Williams & Wilkins. (8%) of 141 in the group receiving daily vitamin C
ISSN: 0029-7844/07 (1,000 mg) and vitamin E (400 International Units)


(adjusted odds ratio (OR) 0.39, 95% confidence inter- incompatible with life, or prior participation in the
val [CI] 0.17 0.90, P!.02). study.
Subsequent to this report, several consecutive The protocol was approved by the National
studies have reported negative outcomes.37 In re- Institute of Child Health and Human Development
sponse to the Chappell study,2 we initiated a trial of and the institutional review boards at the University
antioxidant supplementation focusing upon pa- of Cincinnati, each participating site, and the data
tients with chronic hypertension and/or a history of coordinating center. Each woman gave written in-
preeclampsia. formed consent. Planned interim analyses were mon-
itored by the Global Networks independent data
monitoring committee.
MATERIALS AND METHODS Women were assigned randomly to receive daily
This clinical trial was conducted as a protocol within vitamin C 1,000 mg and vitamin E 400 International
the National Institute of Child Health and Human Units or placebo. The medications were manufac-
Development Global Network for Womens and tured as softgel capsules by J R Carlson Laboratories
Childrens Health Research. The primary clinical (Arlington Heights, IL). Each active treatment gel
center (Recife) and three additional clinical sites capsule contained 500 mg of ascorbic acid, 100
(Campinas, Botucatu, and Porto Alegre) are staffed by International Units of d-alpha tocopherol, 100 Inter-
the Senior Foreign Investigator (Recife) or Senior national Units of d-alpha tocopherol acetate, and
Collaborating Investigators (Campinas, Botucatu, and excipients (gelatin, soybean oil, glycerin, water leci-
Porto Alegre), a program coordinator, research phy- thin, and caramel color). The placebo gel capsules
sicians, and a data manager or staff nurse. Each sites contained excipients only and were externally identi-
major teaching hospital serves a primarily urban cal to the active drug. Participants were instructed to
low-income population. ingest two gel capsules daily from enrollment until
We enrolled women seeking prenatal care who delivery or until the diagnosis of preeclampsia. Cor-
were 12 0/7 to 19 6/7 weeks pregnant and diagnosed rect supplier randomization assignment was verified
with nonproteinuric chronic hypertension or a prior by the data coordinating center.
history of preeclampsia in their most recent preg- The randomization sequence was constructed by
nancy that progressed beyond 20 weeks of gestation. the data coordinating center as permuted blocks of
Patients were enrolled with the criterion of chronic random size, stratified by clinical center, and imple-
hypertension if they were receiving antihypertensive mented by a program residing on the clinical centers
medication or if the systolic blood pressure was 140 study computer. We calculated the sample size for
mm Hg or more or diastolic blood pressure was 90 80% power, by intent-to-treat analysis, to detect a
mm Hg or more on at least two occasions at least 4 relative risk of 0.60 for study drug compared with
hours apart or when the medical record was available placebo, assuming a placebo group preeclampsia rate
and supported the diagnosis or if the medical record of 21%, with (two-tailed) !!.05.
was not available, the patient answered affirmatively Study participants were discouraged from the use
to one of the two following questions: 1) When not of antioxidant vitamins, calcium supplements, and
pregnant, were you ever prescribed medication by a chronic use of aspirin. The women were followed at
health care provider to control your blood pressure? routine prenatal visits, typically every 4 weeks until 26
2) When not pregnant, were you ever told by a health to 28 weeks of gestation, every 2 to 3 weeks until 36
care provider that you had a problem with high blood weeks of gestation, and then weekly until delivery or
pressure? Patients were enrolled with the criterion of the onset of preeclampsia. At each visit, the womens
history of preeclampsia/eclampsia if the medical blood pressure, weight, and urinary protein excretion
record was available and supported the diagnosis of were measured. Urinary protein was measured with a
preeclampsia or, if the medical record was not avail- dipstick in a fresh, clean, midstream urine sample.
able, in the opinion of the enrolling physician, the Compliance with treatment was assessed by
patients history of the events of that pregnancy were counting residual pills at monthly return visits. A
highly consistent with the diagnosis of prior pre- computerized bottle cap, the MEMS V TrackCap
eclampsia. Exclusion criteria were planned delivery Child Resistant (APREX, a division of AARDEX,
elsewhere, multifetal gestation, allergy to vitamin C or Ltd., Union City, CA), which internally records the
vitamin E, requirement for aspirin or anticoagulant date and time of each opening of the pill bottle, was
medication, 24-hour urinary protein 300 mg or more, placed on the first pill bottle and then transferred to
prepregnancy diabetes mellitus, known fetal anomaly sequential bottles. Information from the TrackCap

1312 Spinnato et al Antioxidant Therapy to Prevent Preeclampsia OBSTETRICS & GYNECOLOGY

was shared with the patients at monthly intervals to analyses of subjects who had received at least 80% of
motivate optimal compliance through encouragement the intended doses, a threshold consistent with previ-
and constructive problem solving. ous studies4,6 as assessed by returned pill counts, were
The primary outcome was the development of performed, along with analyses by risk group (prior
preeclampsia, according to the following definitions: preeclampsia only, chronic hypertension only, or
hypertension (defined as a systolic blood pressure of both) and adjusted for covariates. Two interim anal-
140 mm Hg or a diastolic blood pressure of 90 mm yses that corresponded to 34% and 68% of the total
Hg) was designated as severe if two or more systolic planned results were performed according to the
values obtained four or more hours apart were 160 Lan-DeMets approach using an !-spending function
mm Hg or if two or more diastolic values were 110 analogous to the OBrien-Fleming procedure9 at an
mm Hg. Among patients without chronic hyperten- overall level of !!.05 (two-tailed); the significance
sion, gestational hypertension was defined as nonpro- level for the final analysis was !!.0458. Comparisons
teinuric hypertension after week 20 of gestation or between study drug and placebo groups were strati-
during the postpartum period. Preeclampsia was de- fied by site and risk group using Cochran-Mantel-
fined as hypertension plus proteinuria (either 300 mg Haenszel statistics10 for binary outcomes and analysis
per 24 hours or 2" or more by dipstick on two or of variance for continuous outcomes, with natural log
more occasions 4 hours apart). Severe preeclampsia transformation and supplemental Cochran-Mantel-
was defined as severe hypertension and proteinuria; Haenszel row mean tests with modified ridit scores (the
urinary protein excretion 5 g per day with any degree nonparametric van Elteren test)10 where distributions
of hypertension; hypertension complicated by pulmo- were skewed. Covariate adjustments analyses used lo-
nary edema or a low platelet count (less than 100,000/ gistic regression models.11 Data were analyzed using
mL); or hemolysis, an elevated serum aspartate ami- SAS/STAT 9.3 (SAS Institute, Inc., Cary, NC). Exact
notransferase concentration (more than 70 units/L), Cochran-Mantel-Haenszel procedures in StatXact 7
and a low platelet count (the hemolysis, elevated liver (Cytel, Inc., Cambridge, MA) were used when binary
enzymes, low platelets syndrome). Among chroni- outcomes were sparse, yielding odds ratios instead of
cally hypertensive patients, superimposed preeclamp- risk ratios. Lan-DeMets calculations were performed
sia was defined by hypertension plus proteinuria using East 4 (Cytel, Inc., Cambridge, MA).
(either 300 mg/24 h or 2" or more by dipstick).
Women were considered to have eclampsia if they RESULTS
met the criteria for pregnancy-induced hypertension Screening for enrollment began on July 2, 2003, at
and had convulsions. Women were followed through Recife; May 19, 2004, at Botucatu; July 5, 2004, at
the 14th day postpartum for the occurrence of pre- Porto Alegre, and February 2, 2005, at Campinas; and
eclampsia. All reports of suspected gestational hyper- concluded on May 15, 2006. Follow-up was com-
tension or preeclampsia underwent blinded review by pleted on November 23, 2006. From the general
the Senior Foreign Investigator (S.F.) and the Princi- population of obstetric clinic patients and of 835
pal Investigator (J.A.S.). women satisfying inclusion criteria, 739 of 753 (98%)
Secondary outcomes included the severity of eligible women were enrolled in the study (Recife,
preeclampsia, gestational hypertension, abruptio pla- 265; Campinas, 202; Botucatu, 152; Porto Alegre,
centae, premature rupture of membranes, preterm 120). Figure 1 describes the enrollment flow. In each
birth, small for gestational age and low birth weight treatment group, 4.3% of subjects were lost to fol-
infants. Premature rupture of membranes was defined low-up or withdrew consent. Thus, outcome data
as rupture of membranes before the onset of labor. were available for 707 women, of whom 355 were
Small for gestational age was defined as a birth weight assigned to the study drug and 352 to the placebo. Of
below the 10th percentile according to the growth these, 19 (5.4%) and 7 (2.0%) had early treatment
tables of Alexander et al.8 Abruptio placentae was termination but remained in follow-up. The demo-
diagnosed according to clinical findings or placental graphic and clinical characteristics of the enrolled
examination. subjects, including the criteria for inclusion in the
Data were analyzed using the intent-to-treat prin- study, were similar between the two groups (Table 1). Of
ciple: all randomly assigned subjects were included in patients enrolled with chronic hypertension, 52.8% were
the treatment group to which they were originally receiving antihypertensive therapy at onset of preg-
assigned. Participants with missing outcomes due to nancy. Gestational age at enrollment was determined by
withdrawal of consent or loss to follow-up were best obstetric estimate; ultrasound dating was performed
excluded from the analysis of outcomes. Secondary on 61.3% of the women and rejected the last menstrual

VOL. 110, NO. 6, DECEMBER 2007 Spinnato et al Antioxidant Therapy to Prevent Preeclampsia 1313
Fig. 1. Enrollment flow chart.
Spinnato. Antioxidant Therapy to
Prevent Preeclampsia. Obstet Gynecol

period dating in about half of those. Violations of Table 2 reports the rate of preeclampsia by
inclusion or exclusion criteria occurred in 25 subjects. treatment and risk group. No significant differences
Twenty-three women were enrolled outside the 12 week were noted for either the intent-to-treat cohort or the
to 19 week window for gestational age (711 weeks, compliant subgroup. In the intent-to-treat cohort, 49
n!13; 20 23 weeks, n!10). The majority were discov- of 355 patients (13.8%) in the vitamin group and 55 of
ered by ultrasound examinations performed after enroll- 352 patients (15.6%) in the placebo group developed
ment. There were two twin gestations, one lost to preeclampsia (P!.43, adjusted risk ratio (RR) 0.87
spontaneous abortion, and one that delivered liveborns, [95.42% CI 0.611.25]). No significant treatment dif-
resulting in one more infant analyzed than mothers in ferences in the frequency of preeclampsia were noted
the group receiving study drug. All 25 women remained within risk group. The results remained nonsignificant
on their assigned study treatment and continued in after additional adjustment by logistic regression for
follow-up. The percentage of patients judged by re- three baseline covariates specified a priori: prepreg-
turned pill counts as having received at least 80% of the nancy body mass, maternal cigarette smoking at
intended doses was substantial (606 of 707 patients enrollment, and mean arterial pressure at enrollment.
[85.7%]) and was similar between treatment groups Of these, only the mean arterial pressure at enroll-
(84.5%, 86.9%). ment was significantly associated with the subsequent

1314 Spinnato et al Antioxidant Therapy to Prevent Preeclampsia OBSTETRICS & GYNECOLOGY

Table 1. Baseline Characteristics of Enrolled Patients
All Patients Prior Preeclampsia Only Chronic Hypertension

Vitamins Vitamins Vitamins

Characteristic C and E Placebo C and E Placebo C and E Placebo
Total number enrolled 371 368 178 175 193 193
Study site
Botucatu 78 (21.0) 74 (20.1) 19 (10.6) 17 (9.7) 59 (30.6) 57 (29.5)
Campinas 100 (26.9) 102 (27.7) 44 (24.6) 42 (24.0) 56 (29.0) 60 (31.1)
Porto Alegre 60 (16.1) 60 (16.3) 35 (19.6) 27 (15.4) 25 (13.0) 33 (17.1)
Recife 133 (35.8) 132 (35.9) 80 (44.9) 89 (50.9) 53 (27.5) 43 (22.3)
Age (y) 28.9 (#6.3) 29.7 (#6.2) 27.2 (#5.6) 28.3 (#5.9) 30.4 (#6.5) 31.0 (#6.2)
Race (self-identified)
Brown 166 (44.7) 167 (45.4) 88 (49.4) 100 (57.1) 78 (40.4) 67 (34.7)
African American 40 (10.8) 45 (12.2) 16 (9.0) 22 (12.6) 24 (12.4) 23 (11.9)
White 165 (44.5) 156 (42.4) 74 (41.6) 53 (30.3) 91 (47.2) 103 (53.4)
Yellow 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Marital status
Married 157 (42.3) 168 (45.7) 76 (42.7) 69 (39.4) 81 (42.0) 99 (51.3)
Cohabiting 182 (49.1) 175 (47.6) 90 (50.6) 96 (54.9) 92 (47.7) 79 (40.9)
Other 32 (8.6) 25 (6.8) 12 (6.7) 10 (5.7) 20 (10.4) 15 (7.8)
Primigravida 37 (10.0) 34 (9.2) 0 (0.0) 0 (0.0) 37 (19.2) 34 (17.6)
Gestational age at enrollment (wk) 15.6 (#2.7) 15.7 (#2.5) 15.5 (#2.8) 15.6 (#2.5) 15.6 (#2.7) 15.7 (#2.6)
Source of EDD
Dated by LMP 137 (36.9) 149 (40.5) 71 (39.9) 71 (40.6) 66 (34.2) 78 (40.4)
LMP confirmed by ultrasonography 115 (31.0) 98 (26.6) 51 (28.7) 52 (29.7) 64 (33.2) 46 (23.8)
Ultrasonography rejects LMP 119 (32.1) 121 (32.9) 56 (31.5) 52 (29.7) 63 (32.6) 69 (35.8)
Prepregnancy BMI (kg/m2) 28.5 (#7.1) 28.8 (#7.0) 26.6 (#5.7) 26.8 (#6.3) 30.3 (#7.8) 30.8 (#7.1)
Urine protein by dipstick
Negative or trace 367 (98.9) 365 (99.5) 178 (100.0) 175 (100.0) 189 (97.9) 190 (99.0)
1" (less than 300 mg/24 h) 4 (1.1) 2 (0.5) 0 (0.0) 0 (0.0) 4 (2.1) 2 (1.0)
Blood pressure (mm Hg)
Mean arterial pressure* 106.7 (#14.6) 107.3 (#14.6) 100.7 (#11.5) 100.8 (#12.4) 112.3 (#14.9) 113.1 (#14.1)
Systolic 122.2 (#16.2) 122.7 (#16.3) 115.8 (#12.9) 115.5 (#13.9) 128.2 (#16.7) 129.2 (#15.6)
Diastolic 75.6 (#13.3) 76.4 (#13.1) 70.6 (#11.1) 71.4 (#11.5) 80.3 (#13.4) 81.0 (#12.7)
Smoking during gestation 52 (14.0) 50 (13.6) 26 (14.6) 22 (12.6) 26 (13.5) 28 (14.5)
Smoking at enrollment 27 (7.3) 31 (8.4) 12 (6.7) 14 (8.0) 15 (7.8) 17 (8.8)
Less than 1 pack per day 23 (85.2) 27 (87.1) 11 (91.7) 13 (92.9) 12 (80.0) 14 (82.4)
1 pack per day or more 4 (14.8) 4 (12.9) 1 (8.3) 1 (7.1) 3 (20.0) 3 (17.6)
EDD, estimated date of delivery; LMP, last menstrual period; BMI, body mass index.
Data are n, n (%), or mean (#standard deviation).
* Calculated as 2/3 systolic pressure"1/3 diastolic pressure.

development of preeclampsia (P!.0018, OR [for 10 (1.1% compared with 2.3%), induction of labor (14.6%
mm Hg difference] 1.31, 95% CI 1.10 1.56]). compared with 18.3%) or cesarean delivery (66.0%
Among patients without chronic hypertension, compared with 67.6%). Premature rupture of the
severe preeclampsia occurred more frequently in the membranes was more frequently observed in the
study group (6.5% [11 of 170]) compared with placebo study drug group (10.6% compared with 5.5%,
(2.4% [4 of 168]; P!.11, OR 2.78, 95% CI 0.79 P!.015, RR 1.89, 95% CI 1.113.23).
12.62), but not significantly so. There were no mater- Tables 3 and 4 detail perinatal outcomes. No
nal deaths. There was one occurrence of eclampsia in significant differences were noted in the rate of fetal
a patient with prior preeclampsia who was assigned to and neonatal deaths, preterm delivery, low birth
the placebo group. Four patients (two in each group) weight, small for gestational age, low 1- or 5-minute
developed hemolysis, elevated liver enzymes, low Apgar score, and neonatal morbidities.
platelets syndrome. The study treatment was well tolerated by the
There were no (study drug compared with pla- patients. Nine patients discontinued treatment perma-
cebo) differences in the frequency of gestational dia- nently because of difficulty taking the pills. Thirteen
betes (3.4% compared with 3.7%), abruptio placentae (seven study drug, six placebo) were discontinued

VOL. 110, NO. 6, DECEMBER 2007 Spinnato et al Antioxidant Therapy to Prevent Preeclampsia 1315
Table 2. Effect of Antioxidants on the Incidence of Preeclampsia* According to Risk Group
Incidence of Preeclampsia
Adjusted Risk Ratio
Group Vitamins C and E Placebo (95.42% Confidence Interval) P
Intent to treat
All groups 49/355 (13.8) 55/352 (15.6) 0.87 (0.611.25) .43
Prior preeclampsia only 20/170 (11.8) 19/168 (11.3) 1.03 (0.571.87) .93
Chronic hypertension 29/185 (15.7) 36/184 (19.6) 0.78 (0.501.23) .28
Prior preeclampsia 8/49 (16.3) 10/55 (18.2) 0.70 (0.281.79) .44
No prior preeclampsia 21/136 (15.4) 26/129 (20.2) 0.81 (0.481.37) .42
Compliant patients
All groups 43/300 (14.3) 47/306 (15.4) 0.94 (0.641.37) .73
Prior preeclampsia only 17/145 (11.7) 16/150 (10.7) 1.06 (0.552.05) .86
Chronic hypertension 26/155 (16.8) 31/156 (19.9) 0.87 (0.541.40) .55
Prior preeclampsia 6/39 (15.4) 8/49 (16.3) 0.87 (0.322.36) .78
No prior preeclampsia 20/116 (17.2) 23/107 (21.5) 0.87 (0.511.49) .60
Data are n/N (%) unless otherwise specified.
* Includes preeclampsia (in those without chronic hypertension at enrollment), superimposed preeclampsia (in those with chronic
hypertension at enrollment), eclampsia, or the hemolysis, elevated liver enzymes, low platelets syndrome.

Cochran-Mantel-Haenszel method, adjusting for study site and risk group (prior preeclampsia only, chronic hypertension only, chronic
hypertension and prior preeclampsia).

Patients whose returned pill counts indicated that they received at least 80% of the intended doses from enrollment to delivery or to
development of preeclampsia.

permanently because of perceived side effects, includ- women who were lost to follow-up. Reported adverse
ing itching (three), eczema (one), vomiting (two), events were almost exclusively related to expected
abdominal pain (two), abdominal pain and diarrhea pregnancy-related events. One mother attempted sui-
(one), headache (one), constipation (one), and malaise cide by ingesting 26 pills (placebo). Three fetuses had
and diminished vision (one). The remaining treat- developmental abnormalities, one with multiple mal-
ment discontinuations were related to other medical formations (study drug), one with neural tube defect
conditions or were for unknown reasons, including (placebo), and one with trisomy 13 (placebo).

Table 3. Fetal and Infant Outcomes*

Adjusted Risk Ratio
Outcome Vitamins C and E Placebo (95.42% Confidence Interval)* P*
Number of fetuses 356 352
Fetal and neonatal deaths 18 (5.1) 19 (5.4) 1.00 (0.531.87) 1.00
Spontaneous abortion 4 (1.1) 3 (0.9) 1.30 (0.219.44) 1.00
Therapeutic abortion 1 (0.3) 0 (0.0) $ (0.02$) .49
Stillbirth 7 (2.0) 10 (2.8) 0.73 (0.281.95) .53
Neonatal death 6 (1.7) 6 (1.7) 1.12 (0.363.47) .84
Number of completed infants 351 349
Preterm delivery
Less than 37 weeks of gestation 96 (27.4) 82 (23.5) 1.15 (0.891.50) .27
Less than 34 weeks of gestation 29 (8.3) 26 (7.4) 1.10 (0.651.84) .72
Birth weight (g) 3,019.7 (#779.3) 3,039.7 (#767.5) 15.8 (131.199.5) .78
Low birth weight (less than 2,500 g) 61 (17.4) 62 (17.8) 0.98 (0.711.36) .91
Very low birth weight (less than 1,500 g) 20 (5.7) 19 (5.4) 1.08 (0.582.00) .82
Small for gestational age 49 (14.0) 49 (14.0) 1.01 (0.701.46) .96
Apgar scores
Less than 4 at 1 min 15 (4.4) 21 (6.2) 0.72 (0.371.39) .32
Less than 7 at 5 min 8 (2.3) 12 (3.5) 0.72 (0.291.77) .46
Data are n (%) or mean (#standard deviation) unless otherwise specified.
* Cochran-Mantel-Haenszel method, adjusting for study site and risk group (prior preeclampsia only, chronic hypertension only, chronic
hypertension and prior preeclampsia).

Odds ratios, 95.42% confidence intervals, and P values are calculated from the exact Cochran-Mantel-Haenszel procedure.

Excludes infants of mothers who withdrew consent or were lost to follow-up (n!32) and spontaneous and therapeutic abortions (n!8);
denominators vary due to missing responses.

1316 Spinnato et al Antioxidant Therapy to Prevent Preeclampsia OBSTETRICS & GYNECOLOGY

Table 4. Neonatal Outcomes
Adjusted Risk Ratio
Outcome Vitamins C and E Placebo (95.42% Confidence Interval)* P*
Number of completed infants 351 349
Baby died before discharge, or received
care in NICU or intermediate nursery
(12 h or more) 38 (10.8) 39 (11.2) 0.93 (0.611.43) .75
Respiratory distress syndrome (definite
or suspected) 40 (11.4) 34 (9.7) 1.11 (0.721.71) .64
Ventilator support 16 (4.6) 13 (3.7) 1.29 (0.602.74) .50
Duration of ventilator support (d) 5.5 (3.0, 8.0) 3.0 (1.0, 4.0) .042
Bronchopulmonary dysplasia 4 (1.1) 0 (0.0) $ (0.56$) .13
Seizures 2 (0.6) 1 (0.3) 2.08 (0.10134.08) .62
Intraventricular hemorrhage (Grade 3 or 4) 1 (0.3) 0 (0.0) $ (0.04$) .38
Periventricular leukomalacia 0 (0.0) 0 (0.0) NA NA
Retinopathy of prematurity 3 (0.9) 0 (0.0) $ (0.34$) .25
Necrotizing enterocolitis 1 (0.3) 0 (0.0) $ (0.04$) .38
NICU, neonatal intensive care unit; NA, not applicable.
Data are n (%) or median (25th75th percentile) unless otherwise specified.
* Cochran-Mantel-Haenszel method, adjusting for study site and risk group (prior preeclampsia only, chronic hypertension only, chronic
hypertension and prior preeclampsia).

Excludes infants of mothers who withdrew consent or were lost to follow-up (n!32) and spontaneous and therapeutic abortions (n!8);
denominators vary due to missing responses.

P value from analysis of variance after natural log transformation, adjusting for study site and risk group (prior preeclampsia only, chronic
hypertension only, chronic hypertension and prior preeclampsia); P!.17 using the van Elteren nonparametric test.10

Odds ratios, 95.42% confidence intervals, and P values were calculated from the exact Cochran-Mantel-Haenszel procedure.

DISCUSSION hypertension, pregestational diabetes, or multifetal

This study failed to demonstrate a significant effect of gestation) to receive either vitamin C/E or placebo.
vitamins C and E on the rate of preeclampsia. These No difference in the rates of preeclampsia was ob-
findings were noted despite narrow inclusion criteria, served between groups (17.3% compared with 18.8%).
a population at risk, excellent patient compliance, and This incomplete study was stopped prematurely due
a limited number of clinical sites. Reassuringly, we to a loss of funding. Poston et al6 enrolled 2,410
failed to identify a deleterious influence of supple- patients between 14 and 21 weeks of gestation and
mentation on the frequency of low birthweight, small identified to be at increased risk of preeclampsia at 25
for gestational age, stillbirth, or measures of birth clinical sites in the United Kingdom. The population
asphyxia. The more frequent occurrence of prema- included a heterogeneous mix of patients with prior
ture rupture of membranes among supplemented preeclampsia before 37 weeks of gestation, chronic
patients was not expected, and it will be further hypertension, type 1 or 2 diabetes mellitus, antiphos-
analyzed and presented as a separate communication. pholipid syndrome, chronic renal disease, multiple
As determined by pill counts, 86% of the patients in pregnancy, abnormal uterine artery Doppler, and
this study were compliant with their study medication, primiparous women with body mass index greater
taking at least 80% of the expected number of pills. The than 30 kg/m2. The incidence of preeclampsia was
recently reported trials noted similarly defined compli- similar in treatment and placebo groups (15% com-
ance rates of 65 67%.4,6 Whether the improved compli- pared with 16%). Of note, there were significantly
ance noted in the current study represents population more low birth weight babies born to mothers in the
differences or was achieved in response to feedback and supplemented group, but small for gestational age
encouragement using information gained from the elec- frequency did not differ between groups. Among
tronic pill caps or other factors is uncertain. those with preeclampsia, the mean gestational age at
This study is the second completed trial among diagnosis of preeclampsia and the mean gestational
high-risk patients reported since the Chappell trial.2 age at delivery were significantly earlier in the anti-
The subsequently reported studies3,4,6 employed the oxidant group. Neither of these untoward findings
same dose of study drug as the Chappell trial. Beazley was observed in the current study.
et al3 randomly allocated 109 patients at increased Rumbold et al4 enrolled 1,877 normotensive,
risk for preeclampsia (previous preeclampsia, chronic nulliparous women at 14 to 22 weeks of gestation at

VOL. 110, NO. 6, DECEMBER 2007 Spinnato et al Antioxidant Therapy to Prevent Preeclampsia 1317
nine centers in Australia. Of the women enrolled in that might facilitate persistent proinflammatory reac-
the study, 935 were randomly assigned to the vitamin tions at the maternalfetal interface. When vitamins C
group and 942 to the placebo group. There were no and E are used in combination, either of these mech-
significant differences between the vitamin and pla- anisms could mask a potential beneficial effect of
cebo groups in the risk of preeclampsia (6.0% and vitamin C on the incidence of preeclampsia.
5.0%, respectively), death or serious outcomes in the
infant, or having an infant with a birth weight below
the 10th percentile for gestational age. But, as noted in
1. Spinnato JA, Livingston JC. Prevention of preeclampsia with
the editorial accompanying the article5; the therapeu- antioxidants: evidence from randomized trials. Clin Obstet
tic group was also noted to have increases in the risk Gynecol 2005;48:41629.
of hospitalization for hypertension (5.2% compared 2. Chappell LC, Seed PT, Briley AL, Kelly FJ, Lee R, Hunt BJ, et
with 3.4%) and for the use of antihypertensive therapy al. Effect of antioxidants on the occurrence of pre-eclampsia in
women at increased risk: a randomised trial. Lancet 1999;354:
(4.6% compared with 2.8%). Similar to the current 8106.
study, these findings raised the question whether 3. Beazley D, Ahokas R, Livingston J, Griggs M, Sibai BM.
supplementation was associated with more severe Vitamin C and E supplementation in women at high risk for
disease when hypertension occurred. preeclampsia: a double-blind, placebo-controlled trial. Am J
The current study has limitations. It was powered Obstet Gynecol 2005;192:5201.
for a RR of preeclampsia of 0.60, in hopes of confirm- 4. Rumbold AR, Crowther CA, Haslam RR, Dekker GA, Rob-
inson JS, ACTS Study Group. Vitamins C and E and the risks
ing the positive results of Chappell.2 Additionally, the of preeclampsia and perinatal complications. N Engl J Med
pathways to preeclampsia may differ for patients with 2006; 354:1796806.
prior preeclampsia compared with chronic hyperten- 5. Jeyabalan A, Caritis SN. Antioxidants and the prevention of
sion. Thus the confidence intervals for the primary preeclampsia unresolved issues. N Engl J Med 2006;354:
outcome and for enrollment subsets may fail to 18413.
exclude some differences that, if real, would be clin- 6. Poston L, Briley AL, Seed PT, Kelly FJ, Shennan AH, Vita-
mins in Pre-eclampsia (VIP) Trial Consortium. Vitamin C and
ically important. However, taken in the context of vitamin E in pregnant women at risk for pre-eclampsia (VIP
other recently published studies, the results suggesting trial): randomised placebo-controlled trial. Lancet 2006;367:
a lack of benefit are largely compelling. 114554.
It seems unlikely that further study of supplemen- 7. Lindheimer MD, Sibai BM. Antioxidant supplementation in
tation of vitamins C and E, initiated in the second pre-eclampsia. Lancet 2006;367:111920.
trimester, in combination at the doses used in these 8. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A
United States national reference for fetal growth. Obstet
studies is warranted. However, it would be overreach- Gynecol 1996;87:1638.
ing to suggest that these studies directly or indirectly 9. DeMets DL, Lan KK. Interim analysis: the alpha spending
challenge the validity of the suspected role of lipid function approach. Stat Med 1994;13:134156.
perioxidation in the pathogenesis of preeclampsia. 10. Kuritz SJ, Landis JR, Koch GG. A general overview of
Recently, Banerjee et al12 hypothesized that vitamin E Mantel-Haenszel methods: applications and recent develop-
pharmacotherapy could prevent an immunologic ments. Annu Rev Public Health 1988;9:12360.
switch from T-helper cell 1 to T-helper cell 2 that is 11. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd
ed. New York (NY): John Wiley & Sons; 2000.
thought to be vital for early-to-late transition in nor-
12. Banerjee S, Chambers AE, Campbell S. Is vitamin E a safe
mal pregnancies. Additionally, they suggest that vita- prophylaxis for preeclampsia? Am J Obstet Gynecol 2006;194:
min E could be a potential interferon-gamma mimic 122833.

1318 Spinnato et al Antioxidant Therapy to Prevent Preeclampsia OBSTETRICS & GYNECOLOGY