Chambers JC, Fusi L, Malik IS, Haskard DO, De Swiet M, Kooner JS. Association of Maternal Endothelial Dysfunction With Preeclampsia. JAMA. 2001;285(12):1607-1612. doi:10.1001/jama.285.12.1607
Author Affiliations: National Heart and Lung Institute (Drs Chambers, Malik, Haskard, and Kooner), Institute of Obstetrics and Gynaecology (Dr Fusi), Institute of Reproductive and Developmental Biology (Dr De Swiet), Imperial College School of Medicine, Hammersmith Hospital, London, England.
Context Preeclampsia is believed to result from release of placental factors
that damage maternal vascular endothelium. However, because most studies have
been conducted during pregnancy, it has not been possible to separate maternal
from placental mechanisms underlying endothelial dysfunction in preeclampsia.
Objective To determine whether endothelial function is impaired in nonpregnant
women with previous preeclampsia and whether endothelial dysfunction is mediated
by oxidative stress.
Design and Setting Case-control study conducted at 3 hospital maternity units in London,
England, between July 1997 and June 2000.
Participants A total of 113 women with previous preeclampsia (n = 35 with recurrent
episodes; n = 78 with a single episode) and 48 women with previous uncomplicated
pregnancies, all of whom were at least 3 months (median, 3 years) postpartum.
Main Outcome Measures Brachial artery flow-mediated (endothelium-dependent) and glyceryl trinitrate–induced
(endothelium-independent) dilatation were compared between previously preeclamptic
women and controls. To investigate oxidative stress, these measurements were
repeated after administration of ascorbic acid, 1 g intravenously, in 15 cases
and 15 controls.
Results Mean (SD) flow-mediated dilatation was lower in women with previous
preeclampsia compared with controls (recurrent group, 0.9% [4.1%]; single-episode
group, 2.7% [3.5%]; and control group, 4.7% [4.3%]; P<.001).
In contrast, glyceryl trinitrate–induced dilatation was similar in the
3 groups (recurrent, 19.5% [5.9%]; single-episode, 21.0% [8.0%]; and control,
21.0% [8.3%]; P = .65). Impaired flow-mediated dilatation
in previously preeclamptic women was not accounted for by recognized vascular
risk factors. Ascorbic acid administration increased flow-mediated dilatation
in previously preeclamptic women (baseline, 2.6% [3.3%]; after administration,
5.6% [3.0%]; P = .001) but not in controls (baseline,
6.2% [3.3%]; after administration, 6.7% [5.0%]; P
Conclusions Our results indicate that endothelial function is impaired in women
with previous preeclampsia and is not explained by established maternal risk
factors but is reversed by antioxidant ascorbic acid administration.
Preeclampsia remains a major cause of maternal and fetal morbidity and
mortality,1 complicating up to 10% of first
pregnancies and accounting for 40% of iatrogenic premature deliveries.2 The fetal and maternal mechanisms underlying preeclampsia
are not well understood. Endothelial dysfunction is considered to underlie
many of the manifestations of preeclampsia, including hypertension, proteinuria,
and edema.3 It is widely believed that inadequate
trophoblast invasion of the uterine spiral arteries leads to placental ischemia
and release of placental factors that damage the maternal vascular endothelium.3- 5 A role for placental
factors is further supported by findings of increased lipid peroxidation and
oxidation stress in the placentas of women with preeclampsia.6,7
Abnormal placentation may not be the sole basis for preeclampsia though, since
maternal factors such as hypertension, diabetes, and obesity are associated
with an increased risk of preeclampsia.8 However,
most studies of preeclampsia have been conducted during pregnancy, and it
has not been possible to separate maternal from placental mechanisms underlying
the development of preeclampsia.
In this study, we measured vascular responses of women with a history
of preeclampsia at a median interval of 3 years after delivery to examine
whether vascular endothelial function is impaired in preeclamptic women in
the absence of placenta. We additionally examined the role of oxidation stress
underlying vascular endothelial dysfunction in these women.
We conducted a case-control study to compare vascular endothelial function
between women with previous preeclampsia (35 with recurrent episodes; 78 with
single episodes), and women with uncomplicated pregnancies (n = 48) between
July 1997 and June 2000. Cases and controls were at least 3 months (median,
3 years) postpartum. Women with previous preeclampsia were identified from
the maternity units of Ealing, Hammersmith, and Queen Charlotte's Hospitals,
London. Criteria for preeclampsia were a blood pressure greater than 140/90
mm Hg after the 20th week of gestation, accompanied by a proteinuria of 2
+ on urinalysis or proteinuria greater than 300 mg in a 24-hour collection.9
Of a total of 485 women with preeclampsia who were sent invitations
to participate, at least 75 had moved, leaving 410 who received the invitation.
One hundred twenty-eight replied to the invitation, and 113 agreed to participate
(28% acceptance rate). There was no evidence for selection bias in recruitment:
in particular, responders and nonresponders were of similar age and parity,
and they had similar blood pressure at their initial prenatal visit. Controls
were women with uncomplicated deliveries at the same hospitals. Exclusion
criteria for both patients and controls included established atherosclerosis,
malignancy, major organ failure (including hepatic or renal failure), vasculitis,
systemic infection, recent major surgery or trauma, and known diabetes. The
study was approved by the local ethics committee, and all participants gave
written informed consent.
Clinical history, including past history of hypertension, diabetes,
habitual smoking, alcohol intake, and drug therapy, was recorded for all subjects.
The mean of 3 blood pressure readings was calculated for each participant
who sat for 10 minutes while the readings were measured by a mercury sphygmomanometer.
Height, weight, and waist-hip girth ratio were recorded according to standardized
protocols. Blood samples were collected in the fasting state (overnight) and
assayed for total cholesterol, high-density lipoprotein cholesterol (HDL-C),
and triglyceride and glucose levels (AU800 multi-channel analyzer; Olympus
Optical Ltd [UK]; Middlesex, England), and total plasma homocysteine,10 recognized determinants of vascular endothelial dysfunction.11,12 Measurements were also made of plasma
soluble E-selectin and intercellular adhesion molecule-1 (ICAM-1) by radioimmunoassay
(R & D Systems, Abingdon, England), as biochemical markers of endothelial
activation.13 Subjects were not investigated
at any specific time in relation to their menstrual cycle.
For each participant, brachial artery flow-mediated dilatation (endothelium
dependent), and glyceryl trinitrate-induced dilatation (endothelium independent),
were measured as described below. To investigate the role of oxidant stress
mechanisms in the observed vascular responses, brachial artery measurements
were repeated before and 1 hour after administration of ascorbic acid (1 g
in 100 mL of normal saline intravenously infused over 30 minutes) in 15 patients
and 15 controls selected as nonsmoking, nonobese (body mass index [BMI] <25
kg/m2), normotensive (blood pressure <140/90 mm Hg), with normal
total cholesterol (<250 mg/dL [<6.5 mmol/L]) and normal fasting plasma
glucose (108 mg/dL [<6.0 mmol/L]) levels.
Brachial artery flow-mediated dilatation was measured using a 7.0-MHz
linear array transducer (Acuson 128XP/10 system; Mountain View, Calif), and
high resolution ultrasonic vessel wall tracking system (Vadirec; Ingenious
Systems; Arnhem, the Netherlands) as previously described.14
In brief, the brachial artery was scanned longitudinally, and the brachial
artery diameter was measured at end diastole. After the baseline resting scan,
a pneumatic cuff placed at the level of the mid-forearm was inflated to 300
mm Hg for 4.5 minutes. The second scan was performed 55 to 65 seconds after
cuff deflation. Fifteen minutes was allowed for vessel recovery, after which
the second baseline scan was performed. Glyceryl trinitrate (400 µg)
was then administered and the fourth scan of the brachial artery undertaken.
Brachial artery velocity and blood flow, determinants of flow-mediated dilatation,
were measured using pulsed wave Doppler analysis 1 minute prior to inflation
and 10 seconds after deflation of the pneumatic cuff. Velocity data were recorded
with online angle correction on superVHS tape and analyzed offline. Peak systolic
velocity and velocity time integral were calculated as an average over 5 beats.
Resting and hyperemic brachial artery blood flow were derived from the velocity
time integral, by standard methods.11 Vessel
diameter and blood flow were measured by 2 independent observers unaware of
each participant's clinical details or the type and stage of the study. The
technique for measurement of brachial artery flow–mediated dilatation
is reproducible in our laboratory. The intraindividual, between-day coefficient
of variation for flow-mediated dilatation is 3%, which compares favorably
with other centers.15
Based on previous studies in our laboratory,14
we expected an SD for flow-mediated dilatation of 4% in healthy controls.
We planned to study 50 individuals in each of the 3 groups (recurrent preeclampsia,
single episode of preeclampsia, and controls), offering a 90% power to detect,
at a significance level of 5%, a true difference of 3% in flow-mediated dilatation
between groups, a difference comparable to that identified in other pathological
Data were analyzed using SPSS, Version 10.0 (SPSS Inc, Chicago, Ill) statistical
package. Continuous data are expressed as mean (SD) or as mean (95% confidence
interval [CI]). The differences between the 3 groups were investigated using
analysis of variance for continuous data, and the χ2 test for
categorical data. Post hoc t tests were used to localize
differences, with Bonferonni adjustment for multiple comparisons. Thereafter,
the contribution of possible confounding effects to the association of previous
preeclampsia with flow-mediated dilatation was examined: first, through linear
regression analysis and then separately through a subgroup analysis in which
cases and controls with identified, maternal risk factors for vascular disease
were excluded (BMI >25kg/m2, known hypertension, blood pressure
>140/90 mm Hg, fasting plasma glucose level >108 mg/dL [6.0 mmol/L], cholesterol
level >250 mg/dL [6.5 mmol/L], or cigarette smoking). For the purposes of
regression analysis and for the analysis of women without vascular risk factors,
cases with recurrent and single episodes of preeclampsia were combined into
1 group. The effects of ascorbic acid on brachial artery vascular responses
were determined using the paired samples t test.
Statistical significance was inferred at a P value
The characteristics of the subjects are summarized in Table 1. In comparison with controls, women with recurrent and single
episodes of preeclampsia had higher systolic and diastolic blood pressure,
BMI, waist-hip girth ratio, and total cholesterol to HDL-C ratio, and a higher
prevalence of hypertension and family history of hypertension. Concentrations
of soluble E-selectin, but not ICAM-1, were also higher in women with previous
preeclampsia compared with controls. There were no significant differences
in age, smoking rates, or concentrations of fasting glucose, triglycerides,
and homocysteine between the 3 groups.
Brachial artery flow-mediated, endothelium-dependent dilatation was
lower in women with previous preeclampsia compared with controls (recurrent
episode group, 0.9% [4.1%]; single episode group, 2.7% [3.5%]; control group,
4.7% [4.3 %]; P<.001). The defect was more severe
in women with recurrent preeclampsia compared with a single episode of preeclampsia
(P = .02).
In contrast, there were no significant differences in glyceryl trinitrate–induced,
endothelium-independent dilatation between the 3 groups (recurrent episode,
19.5% [5.9%]; single episode, 21.0% [8%]; control, 21.0% [8.3 %]; P = .65). Baseline brachial artery diameter was higher in women with
previous preeclampsia compared with controls but not when corrected for body
surface area (Table 2). There
were no significant differences between the groups in brachial artery velocity
or blood flow either at rest or during the reactive hyperemia after deflation
of the pneumatic cuff (Table 2).
In univariate analysis, flow-mediated dilatation was negatively associated
with BMI and baseline brachial artery diameter (Table 3). In multivariable regression analysis, the relationship
between previous preeclampsia and impaired flow-mediated dilatation was independent
of age, BMI, waist-hip girth ratio, blood pressure, family history of hypertension,
fasting plasma glucose levels, lipid profile, homocysteine concentration,
brachial artery diameter, and brachial artery blood flow (P = .008, Table 4).
In a separate analysis, we compared vascular responses of cases and
controls, among women who were nonobese and nonsmoking, who had normal blood
pressure and fasting glucose and cholesterol levels (Table 5). Flow-mediated, endothelium-dependent dilatation was lower
in women with previous preeclampsia compared with those in the control group
(previous preeclampsia, 2.5% [3.2%]; control, 4.6% [4.4%]; P = .03), confirming that the relationship between preeclampsia and
endothelial dysfunction was independent of risk factors generally associated
with vascular disease.
Administration of ascorbic acid increased flow-mediated dilatation in
preeclamptic women but not in controls. In contrast, glyceryl trinitrate–induced
dilatation was unchanged after ascorbic acid in both patients and controls
Vascular endothelial dysfunction is recognized to be a central disturbance
in preeclampsia. Evidence from previous studies suggests that endothelial
dysfunction occurs in response to abnormal placentation, which may lead to
placental ischemia and release of placental products that damage the maternal
vascular endothelium.3,19 To identify
whether maternal factors, independent of the placenta, contribute to endothelial
dysfunction in preeclampsia, we studied the vascular responses of preeclamptic
women remote from delivery.
We found that flow-mediated dilatation is reduced in women with previous
preeclampsia compared with women with uncomplicated pregnancies, at a median
interval of 3 years postpartum. Since flow-mediated dilatation is endothelium
dependent,20 our results demonstrate that vascular
endothelial function is impaired in women with previous preeclampsia. Impaired
endothelial function was more severe in women with recurrent preeclampsia
and was not accounted for by maternal obesity, hypertension, metabolic disturbances
associated with insulin resistance, dyslipidemia, elevated homocysteine concentrations,
or brachial artery flow characteristics, which are recognized as potential
determinants of vascular function.11,12
Endothelial dysfunction 3 years postpartum is also unlikely to be a consequence
of the preeclamptic episode since in other situations endothelial dysfunction
normalizes once the underlying cause has been removed. Previous studies have
shown recovery of endothelial function after lowering risk factors including
high cholesterol,21,22 triglyceride,23 and homocysteine concentration12,24;
high blood pressure25- 27;
insulin resistance28; physical inactivity29,30; and estrogen deficiency.31 Endothelial dysfunction is also only transiently
impaired by other stimuli such as cigarette smoking32
and systemic inflammation.33 Our observations
of impaired endothelial function in women with preeclampsia, remote from delivery
and independent of known vascular risk factors, suggest that novel factors
may contribute to material endothelial dysfunction in preeclampsia.
Recent studies show that endothelium-dependent dilatation can be inhibited
by NG-monomethyl-L-arginine (L-NMMA) infusion, an antagonist of
nitric oxide synthase.34 These observations
suggest that endothelium-dependent dilatation is largely mediated by the release
of nitric oxide although they do not exclude a separate role for prostacyclin
and other endothelium-derived relaxing factors. Although nitric oxide activity
was not directly measured in our patients, our findings of reduced endothelium-dependent
dilatation imply that the bioavailability of endothelial nitric oxide may
be reduced in preeclamptic women, even in the nonpregnant state. Reduced nitric
oxide, the major endothelium-derived vasodilator, promotes vasoconstriction,
platelet aggregation, and monocyte adhesion, all of which could contribute
to the vascular disturbances in preeclampsia.
Increasing evidence suggests that endothelial dysfunction in preeclampsia
is mediated by oxidative stress.36,37
Recent studies show that antioxidant vitamins improve biochemical markers
of endothelial activation and reduce the incidence of preeclampsia in high-risk
women.9 Observations that lipid peroxidation
and formation of peroxynitrite are increased in the placentas of women with
preeclampsia6,7 have led to the
hypothesis the placenta is the principal source of oxidative stress. In our
study, impaired endothelium-dependent dilatation in women with preeclampsia
was normalized by ascorbic acid. Ascorbic acid is a powerful water-soluble
antioxidant that scavenges oxygen-derived free radicals, including superoxide
anion, which would otherwise interact with nitric oxide and impair its vasoactive
functions.38 Our findings therefore support
the presence of oxidative stress, which may contribute to endothelial dysfunction
in women with previous preeclampsia.
A further important finding of this study is that blood pressure levels
are elevated following an episode of preeclampsia. We found that systolic
blood pressure was 14 mm Hg higher and diastolic blood pressure was 8 mm Hg
higher in previously preeclamptic women than in controls. These observations
indicate that in women with previous preeclampsia, blood pressure distribution
is shifted to the right compared with women with normal pregnancies and may
precede development of hypertension. This may be of clinical importance since
it is increasingly recognized that there is a graded relationship between
blood pressure and risk of vascular events, including myocardial infarction
In summary, we have found evidence of impaired vascular endothelial
function in preeclamptic women, at a median interval of 3 years after delivery.
Endothelial dysfunction is not explained by the presence of known vascular
risk factors but is reversed by ascorbic acid.