Low-dose aspirin therapy to prevent preeclampsia

John C. Hauth, MD, Robert L. Goldenberg, MD, C. Richard Parker, Jr., PhD,
Joseph B. Philips III, MD, Rachel L. Copper, MSN, CRNP, Mary B. DuBard, MA, and
Gary R. Cutter, PhD
Birmingham, Alabama

OBJECTIVE: Our aim was to test the hypothesis that acetylsalicylate (aspirin) treatment reduces the
incidence or severity of pregnancy-associated hypertension.
STUDY DESIGN: Patients were nulliparous, healthy, and with a singleton gestation at between 20 and 22
weeks' gestation. A sample size of 600 patients was calculated on the basis of p ~ 0.05 and 90% power
of observation. A 2-week placebo-controlled "run-in" was used to select compliant patients.
Randomization occurred at 24 weeks, with 60 mg of aspirin or placebo treatment from randomization to
delivery.
RESULTS: Follow-up was maintained on 99% of the patients. The randomized patients had a 94% pill
compliance index. At randomization, serum thromboxane medians were similar in both groups.
Thromboxane 8 2 levels in the aspirin group decreased significantly from baseline at 29 to 31 weeks, 34
to 36 weeks, and at delivery as compared with an overall increase in the placebo group. Preeclampsia
developed in five of 302 women (1.7%) who received aspirin versus 17 of 302 (5.6%) who received the
placebo (p = 0.009). Preeclamspia was severe in one aspirin and in six placebo recipients (p = 0.06).
CONCLUSION: Daily ingestion of 60 mg of aspirin beginning at 24 weeks' gestation significantly reduced
the occurrence of preeclampsia. (AM J OBSTET GYNECOL 1993;168:1083-93.)

Key words: Preeclampsia prevention, aspirin, serum thromboxane

Pregnancy-associated hypertensive disorders, partic-
ularly preeclampsia or eclampsia, are common compli-
cations that are responsible for many fetal, neonatal,
and maternal deaths in the United States." 2 High-risk
populations for preeclampsia or eclampsia encompass
nulliparous women or those with chronic hypertension,
diabetes mellitus, multifetal gestations, or chronic vas-
cular disease.
Preeclampsia is manifested by increased peripheral
vascular resistance with arteriolar vasospasm and endo-
thelial damage. The characteristic resistance to the
pressor effects of the infusion of the potent vasopressor
angiotensin II, which occurs in normal pregnancy, is not
present in women with preeclampsia. As early as 24 to
28 weeks' gestation, 90% of women who subsequently
become hypertensive are significantly more sensitive to
an infusion of angiotensin II than are women who never
become hypertensive.' Except for a history or presence
From the Division of Maternal-Fetal Medicine, Department of Ob-
stetrics and Gynecology, The University of Alabama at Birmingham.
Supported by National Institute of Child Health and Human Devel-
opment grant No.1 ROI HD24496-01.
Presented by invitation at the Eleventh Annual Meeting of the
American Gynecological and Obstetrical Society, Hot Springs, Vir-
ginia, September 10-12, 1992.
Reprint requests: John C. Hauth, MD, University of Alabama at
Birmingham, Department of Obstetrics and Gynecology, UAB Station,
Birmingham, AL 35233-7333.
Copyright 1993 by Mosby-Year Book, Inc.
0002-9378/93 $1.00 + .20 6/6/44821
of the maternal medical and obstetric complications
detailed above, a positive angiotensin II test (marked
sensitivity) is the only reliable method to predict which
nulliparous women are destined to develop preeclamp-
sia later in pregnancy.
Maternal sensitivity to angiotensin II infusion can be
increased by cyclooxygenase inhibitors (such as 500 mg
of acetylsalicylate [aspirin] or 25 mg of indomethacin
given twice 2 hours apart4 ). Conversely, a low dose of
aspirin (80 mg) causes refractoriness to the pressor
effects of infused angiotensin 11.5.7 A proposed mecha-
nism of low-dose aspirin-induced maternal refractori-
ness to angiotensin II is an alteration in the maternal
thromboxane A2 (TxA2)/prostacyclin (vasoconstric-
tor/vasodilator) ratio in favor of the vasodilator prosta-
cyclin (as determined by measurement of their metab-
olites thromboxane B2 (TxB 2) and 6-keto-prostaglan-
din F 1.. (PGF 1..6 . 9 ). Moreover, Walsh et al.1O showed that
placental tissue of women with preeclampsia produced
more TxB2 and less 6-keto-PGF 1.. than did placentas
from women with normal pregnanl=ies. Thus the con-
cept has evolved that eicosanoids mediate vascular
reactivity during pregnancy and that altered production
of these substances may lead to the development of
preeclampsia. To test this relationship, Wallenburg et
al." administered 60 mg of aspirin a day to prevent
preeclampsia in primigravid women who were identi-
fied to be at high risk because of increased sensitivity to
angiotensin II at 28 weeks' gestation. He reported a

1083
1084 Hauth et al.
significant reduction in the incidence of preeclampsia
and eclampsia that did not occur in 21 women treated
with aspirin as compared with eight of 23 (35%) women
who received a placebo. Schiff et at." used a rollover test
to screen 791 pregnant women who were at risk for
preeclampsia. Women with an abnormal rollover test
were randomized to either 100 mg of aspirin a day
(N = 34) or a placebo (N = 31). The women who
received aspirin had a significantly lower incidence of
preeclampsia. The mean ratio of serum TxB2 to serum
6-keto-PGF 1a decreased significantly in the aspirin-
treated group and increased significantly in the placebo
group (suggesting that 6-keto-PGF la remains stable
during pregnancy).
However, because of their complexity and or lack of
specificity, angiotensin II sensitivity or rollover tests
have not been generally used for screening a popula-
tion of nulliparous women. Thus our purpose was to
substantiate the above reports in a prospective, double-
blinded, randomized clinial trial in a general popula-
tion of ostensibly normal nulliparous women. Our hy-
pothesis was that daily maternal prophylaxis with 60 mg
of aspirin would significantly reduce the incidence of or
ameliorate the severity of pregnancy-associated hyper-
tensive disorders (encompassing pregnancy-induced
hypertension, preeclampsia, and eclampsia). Finally, it
was apparent that the fetal and newborn risks related to
daily maternal aspirin therapy from 24 weeks' gestation
until delivery, although seemingly very low, needed to
be assessed in a larger population than was studied by
Wallenburg et al. II or Schiff et at."
Methods
Study subjects. The study population consisted of
nulliparous women. Patients with illnesses or conditions
known to increase the incidence of preeclampsia or
pregnancy-induced hypertension, such as renal disease,
collagen vascular disease, diabetes mellitus, multifetal
gestation, and chronic hypertension, were excluded.
Age below 28 years was chosen as an inclusion criterion
because women below this age account for 58% of all
deliveries in our population and have an incidence of
pregnancy-induced hypertension or preeclampsia of
9%, with < 1% having chronic hypertensive disease.
Older women have more chronic hypertension and less
preeclampsia and pregnancy-induced hypertension.
Potential candidates for our study were selected (by
means of our obstetric automated records system,
OBARI2) from among the 4500 women who yearly
attend antepartum public health clinics in Birmingham.
Potential subjects were contacted at no later than 22
weeks of gestation. At the patient's initial visit to the
research clinic, the gestational age was derived from the
maternal last menstrual period and physical examina-
tion and was confirmed with fetal ultrasonographic
April 1993
Am J Obstet Gynecol
measurements. The study was explained to each patient
and informed consent obtained. The 962 eligible pa-
tients who consented to the study then underwent a
compliance assessment during a 2 to 3 week "run in"
placebo trial, during which they were instructed to not
take any aspirin-containing medications. The 606 pa-
tients who returned with their pills for all scheduled
visits, took ;e: 75% of their pills during the "run in," and
had a serum TxB2 level > 1000 pglml on their second
visit were considered to have passed the "run in" and
were randomized to the double-masked medication. Of
the other 356 women who began the "run in" trial, 28
were not eligible because of subsequent refusal and 13
were delivered before 24 weeks' gestation. Thus 315
women participated in but failed the "run in." The
majority were subsequently delivered at our institution.
These 315 women had study inclusion criteria similar to
those randomized to aspirin or placebo. They returned
to public health department prenatal clinics for the
remainder of their pregnancies. Outcome data on this
group, including the determination for preeclampsia or
pregnancy-induced hypertension, were collected in a
manner identical to that for women randomized to
aspirin or placebo.
Treatment protocol. The patients were randomized
to receive a daily 60-mg capsule of aspirin or an
identical-appearing placebo containing a lactose filler.
At randomization all participants had baseline tests
performed as shown in the study design (Fig. 1). They
were then given a random but known amount of either
aspirin or placebo capsules and a list of common over-
the-counter products that contain aspirin. The patients
were asked to avoid these products and were given a
supply of acetaminophen to take for minor discomforts.
The random number (known by the pharmacy and the
safety officers) of excessive aspirin or placebo capsules
was used to discourage the patients from attempting to
appear compliant with our directions by emptying their
bottles before each visit. Moreover, each patient was
informed that her capsule bottle contained a random
number of excess capSUles. At each clinic visit the
importance of clinic attendance and treatment adher-
ence were emphasized. All study patients were seen at
2-week intervals by the same antepartum team. To
enhance patient compliance, all follow-up visits were
scheduled on the same day of the week; however, the
clinic functioned every day with the same obstetric
research nurses in attendance. Therefore patients who
came at an unscheduled time still received care by the
same nursing team. Patients who did not come for their
biweekly clinic visits were called by a research nurse and
seen at the next earliest convenient time.
Sample size calculations. To test our primary hy-
pothesis, it was determined that 297 patients would be
required in both the aspirin and placebo groups to
Volume 168, Number 4
Am J Obstet Gynecol
achieve significance of p s; 0.05 for a one-tailed test
and a 90% power of observation. This calculation was
based on the following assumption: the incidence of
pregnancy associated hypertensive disorders (pregnan-
cy-induced hypertension, preeclampsia, eclampsia) in
our target population is 9%; as reported by Wallen burg
et al.,1\ we anticipated that the placebo group would
have a sixfold greater incidence of pregnancy-induced
hypertension, preeclampsia, or eclampsia if aspirin
therapy was successful. Therefore the treatment group
incidence would be 1.5%. Finally, we assumed that only
80% of the aspirin group would complete the protocol
and be compliant to the aspirin regimen. Thus 80% of
the patients in the aspirin group would experience an
incidence of hypertensive disorders of 1.5%, but 20%
(dropouts and those who did not comply) would remain
at a 9% incidence. Thus the effective treatment group
incidence rate of hypertensive disorders was estimated
to reach 3%. Similarly, we estimated that 5% of the
placebo group might take aspirin (crossover), thus
yielding a reduced incidence of hypertension in the
control group of 8.6% (95% experience the 9% inci-
dence and 5% experience the aspirin rate of 1.5%). It
should be noted that for correlations or associations
with continuous outcomes, the power was substantially
higher, permitting some subgroup hypotheses to be
tested.
Diagnoses. Before any project deliveries, hyperten-
sive disorders were defined as follows: (1) Pregnancy-
induced hypertension: mild, a diastolic blood pressure
of ~ 90 but < 110 mm Hg on at least two occasions at
least 1 hour apart and before or during labor or within
12 hours post partum; severe, a diastolic blood pressure
of ~ 110 mm Hg on at least two occasions at least 1
hour apart and before or during labor or within 12
hours post partum. (If a study patient had a platelet
count of < 100,000/mm' or an aspartate aminotransfer-
ase ~ 100 U/L, she was considered to have severe preg-
nancy-induced hypertension even if the blood pressure
criteria for severe pregnancy-induced hypertension were
not met.) (2) Preeclamspia: mild, a diastolic blood pres-
sure ~ 90 but < 110 mm Hg on at least two occasions at
least 1 hour apart and before or during labor or within
12 hours post partum and proteinuria of ~ 1 + on two
or more occasions at least 1 hour apart in the absence
of a urinary tract infection or gross hematuria or
~ 0.5 gm per 24 hours; severe, a diastolic blood pres-
sure ~ 110 mm Hg on at least two occasions at least 1
hour apart before or during labor or within 12 hours of
delivery and proteinuria as defined for mild preeclamp-
sia. One of the following was also present: headaches or
visual disturbance, epigastric pain, oligura, thrombocy-
topenia, increased bilirubin, increased aspartate ami-
notransferase, fetal growth retardation, or proteinuria
~ 3 to 4( +) or 4 gm per 24 hours. Fetal growth retar-
Hauth et al. 1085
Identification ...... 1
Nulliparous Women (Gestation 16-20 weeks)
..... Informed Consent
~ Two week "run in" trial
Randomization
Baseline Tests (23 1: 1 weeks gestation) to include:
1) Serum thromboxane B,
2) Plasma thromboxane B,
3) Plasma 6-Keto prostaglandin F,.
4) Umbilical doppler flow assessment
5) Ultrasound
~
Begin Double Blinded Capsules (week 24)
~
Repeat above Baseline Tests at 28-30 and 34-36 weeks' gestation
Newborn ..... At Delivery ...... Mother
- Clinical Evaluation for:
+ Abnormal coagulation - Central placental bed biopsy
+ Persistent fetal circulation (if cesarean delivery)
+ Fetal growth retardation - Blood samples for prostaglandin
- Umbilical Arterylvenous: metabolites
+ pH and blood gases - Assessment for hypertension
+ Platelet count - Assessment for hemorrhagic
+ Blood for prostaglandin metabolites complications
- Pre- and post ductus arteriosus
Transcutaneous oxygen monitoring
for 3 hours
Fig. 1. Experimental design summary.
dation was defined as a birth weight s; 10th percentile
according to the growth standard of Brenner et al. 13
In the research clinic blood pressure was determined
with the patient seated and the cuff at heart level after
the patient had rested for 5 minutes. A standard air or
mercury sphygmomanometer was used with a standard
(25 to 35 cm) or large adult cuff (33 to 47 cm) selected
as determined by the patient's upper arm circumfer-
ence. Korotkoff 1 and 5 were recorded unless Korotkoff
5 = 0; then Korotkoff 4 was used. If, at an antepartum
clinic visit, the Korotkoff 5 was ~ 90, the blood pressure
was repeated after the patient had rested for 15 to 30
minutes. During initial evaluation at labor and delivery,
similar equipment, patient position, and criteria were
used; however, these aspects were not standardized
throughout a subsequent labor, delivery, and post-
partum period that encompassed up to 24 to 48
hours.
Patient compliance. Clinic attendance and treatment
adherence were managed as detailed above. Primary
analysis was by intent to treat; however, to allow subse-
quent analysis of outcomes and patient adherence to
treatment, along with the motivating effect of monitor-
ing, we assessed patient compliance by a capsule count
at each 2-week visit. The compliance index was calcu-
lated by the formula (number of pills taken)l(number of
days between visits) x 100. The compliance index was
1086 Hauth et al.
25,000 "T'""----------------,
Placebo
+305%
Aspirin
20,000
~
Q.
&15,000
~ standard curve ranged from 0 to 1000 pgltube, and the
~ lowest detectable amount of TxB2 generally ranged
.8
eE
10,000
.s::
l-
E
::>
~ respectively; within-assay coefficients of variation were
5,000
O~-"T'""---~---r_--~._~
24 29-31 34-36 Delivery
Gestation (weeks)
Fig. 2. Median serum TxB2 levels at randomization, at 29 to
31 and 34 to 36 weeks' gestation, and at delivery in women
who received aspirin or placebo. asterisks, Significant decrease
from randomization in aspirin group (jJ < 0.001) at all sam-
pling times; plus and minus percent, changes relative to baseline
values for both groups.
calculated for each 2-week period between visits. If
compliance was < 80% the patients were counseled
regarding the importance of taking their pills. We also
determined serum TxB2 levels in each patient at ran-
domization, at 28 to 30 weeks, at 34 to 36 weeks, and at
delivery for subsequent secondary analysis of compli-
ance, response, and potential crossover (women who
unintentionally ingest aspirin or other nonsteroidal
antiinflammatory drugs that can also reduce serum
TxB2 levels).
Randomization and treatment. The University of
Alabama at Birmingham Investigational Drug Serivce of
the Drug Resource Center generated a blocked ran-
domization scheme with randomly chosen block sizes.
With this scheme each of the 604 patients in the
double-blind trial was randomly assigned to take either
60 mg of aspirin or placebo daily. Confirmation of the
60 mg capsules of aspirin was provided by National
Medical Services (Willow Grove, Pa.). Ten separate
samples from two bottled formulations averaged 65.8
and 63.5 mg of aspirin per capsule.
Serum TxB 2 Serum (0.4 ml) was acidified to pH 3 to
4 by the addition of 1N hydrochloric acid and then
extracted with 3 ml of ethyl acetate. Two milliliters of
the ethyl acetate extract was removed and evaporated;
the dried residue was suspended in 1.2 ml of assay
buffer (0.01 moUL phosphate containing 0.1 % bovine
serum albumin, pH 7.0). Aliquots (20 and 100 f.LI) of the
April 1993
Am J Obstet Cynecol
extract were quantified by a specific TxB2 radioimmu-
noassay. Anti-TxB 2, TxB2 standard, and iodine 125-
TxB2 were purchased from Advanced Magnetics, Cam-
bridge, Mass., whose assay protocol was used through-
out with the exception that separation of antibody-
bound from nonbound '25I_TxB2 was achieved by use of
donkey antirabbit "{-globulin attached to cellulose beads
(Sac Cell, Boldon, Tyne, and Ware, U.K.). The assay
from 3 to 7 pg/tube. Between-assay coefficients ofvari-
ation for sera with high TxB2 (> 10,000 pg/ml) and for
low TxB2 (200 to 2000 pg/ml) were l.6% and 4.9%,
comparable. All samples obtained during the clinical
trial from a given patient were quantified in the same
assay procedure. Laboratory personnel were blinded as
to the treatment of all patients.
Data analysis. Outcome data collection occurred af-
ter delivery but before the assignment code was broken.
All patients (including those who failed the "run-in"
trial) having any hypertensive disease were further re-
viewed by the principal investigator, masked to treat-
ment group, to determine project-specific hypertension
diagnoses. All outcome data were accumulated and
entered into a relational data base containing project-
specific data. Data "lock" occurred before breaking the
double-blind code. Our reported analysis is by "intent
to treat" for the efficacy of daily 60 mg of aspirin in
preventing hypertensive disorders. Statistical analyses
were conducted by means of the Statistical Analysis
System release 6.04 for personal computers. Student
t test, X2 tests of proportion, and one-tailed Fisher's
exact tests were used where appropriate, with a
p :5 0.05 considered significant.
Results
Of the 2475 women screened from July 1, 1988,
through Aug. 1, 1991, 606 were randomized to either
the aspirin or placebo treatment groups. One patient
from each group was lost to follow-up, leaving 302
patients in each group for analysis. The patients in each
of these groups were of similar age (aspirin mean
20.3 2.6 years vs placebo mean 20.4 2.7 years)
and of similar race (aspirin group 30% white, 70% black
vs placebo group 27% white, 73% black, p = 0.416).
Both groups had a similar number of antepartum visits
(1l.6 2.2 in the aspirin group and 11.4 2.3 in the
placebo group).
Compliance was similar in the aspirin and placebo
treatment groups. The mean compliance index, with a
maximum of 100, was 94.5 (SD 5.4) in the aspirin
group and 94.4 (SD 7.53) in the placebo group
(p = 0.43, not significant). Compliance in the overall
aspirin and placebo treatment groups was also con-
Volume 168, Number 4 Hauth et al. 1087
Am J Obstet Gynecol

Table I. Obstetric complications in 604 Table II. Perinatal outcomes in 604

women who received either aspirin or women who received either aspirin or
placebo daily placebo daily
Aspirin Placebo Aspirin Placebo
(N = 302) (N = 302) (N = 302) (N = 302)

Indicated pre term delivery 12 (4%) 13 (4.3%) Fetal death 1 1

Spontaneous preterm de- 8 (2.7%) 11 (3.7%) Fetal growth retarda- 17 (5.6) 19 (6.3)
livery tion (No. and %)
Preterm premature rupture 9(3%) 8 (2.7%) Fetal growth retarda- 16 (5.8) 15 (5.6)
of membranes tion (2: 37 weeks)
Labor induced 66 (22%) 84 (28%) (No. and %)
Cesarean delivery 72 (24%) 66 (22%) Birth weight (mean 3249 577 3169 553*
SD, gm)
Differences between groups were not significant. Birth weight at 2: 37 3342 485 3280 425
weeks' gestation
(mean SD, gm)
Apgar score at 1 min 7.93 1.47 7.70 1.67
(mean SD)
firmed by serum TxB2 determinations. Median serum Apgar score at 5 min 8.86 0.54 8.81 0.65
TxB2 levels were similar at randomization in both (mean SD)
groups, but the women who received aspirin had sig- Umbilical artery pH 7.26 0.08 7.26 0.07
(mean SD)
nificantly lower serum levels at 29 to 31 (jJ = 0.0001) Umbilical artery pH at 7.26 0.08 7.26 0.07
and 34 to 36 (jJ = 0.0005) weeks' gestation and at 2: 37 weeks' gesta-
delivery (jJ = 0.0001) than the placebo group and from tion (mean SD)
Umbilical artery pH 39/241 (16) 37/233 (16)
their TxB2 level at randomization (Fig. 2). There was an <7.20 (No. and %)
overall increase in serum TxB2 from randomization to
delivery in the women who received the placebo. *P = 0.085; all others p > 0.1.
Selected obstetric complications were similar in the
604 women who received either aspirin or placebo
(Table I). There was no significant difference in the rate The effect of maternal aspirin treatment on the inci-
of indicated or spontaneous preterm delivery, preterm dence of hypertensive disorders is depicted in Table Ill.
amnion rupture, induction of labor, or incidence of The incidence of pregnancy-induced hypertension was
cesarean delivery. Selected perinatal outcomes were similar in women who received aspirin or placebo. How-
also similar in the infants of the women who received ever, the incidence of preeclampsia was significantly less
aspirin or placebo. The incidence of fetal growth retar- in women who received aspirin. Only five (1.7%) women
dation overall and at term, the mean birth weight at who received aspirin versus 17 (5.7%) who received pla-
term, the mean umbilical arterial pH levels, the percent cebo developed preeclampsia (p = 0.009). Preeclamp-
of pH values < 7.2, and the mean pH values at term sia was severe in six women who received placebo but in
gestation were all similar (Table II). There was a trend only one who received aspirin (p = 0.06). The mean
toward a higher overall mean birth weight in the gestational age at delivery was 36.0 4.58 weeks in the
women who received aspirin (jJ = 0.085). One fetal five women who developed preeclampsia and who re-
death occurred in each of the aspirin and placebo ceived aspirin and 36.4 3.67 weeks in the 17 women
treatment groups. One of these women, who was taking who received the placebo (jJ = 0.85). The combined in-
the placebo, had an unexplained death of a 2430-gm cidences of pregnancy-induced hypertension and pre-
fetus at 38 weeks' gestation. Her urine drug screen and eclampsia were not significantly different between the
Kleihauer-Betke test were negative, and pathologic ex- two groups. An elevated aspartate aminotransferase
amination of the fetus and placenta revealed no appar- (~ 100 U/L) occurred in three women who received aspi-
ent cause for the death. The patient did not have rin and eight women who received placebo (p = 0.128).
preeclampsia or pregnancy-induced hypertension, Thrombocytopenia 100,000 mm3 ) was documented
abruptio placentae, or evidence of fetal hydrops. An- in five women who received aspirin and three women
other fetal death occurred in a women who was receiv- who received placebo (jJ = 0.477). As expected by some,
ing aspirin; At 31 weeks' gestation, while lying in bed, a uric acid level of > 5.5 mg/dl at delivery was more
she had acute onset of severe abdominal pain and frequent in women with preeclampsia (22/27, 81.5%)
severe headache. The patient was delivered vaginally of than in women with pregnancy-induced hypertension
a 2000-gm fetus and required transfusion of eight units (24/39, 61%),P < 0.08.
of packed red blood cells. Fetal death was confirmed Finally, the incidence of pregnancy-associated hyper-
and felt to be caused by preeclampsia and abruptio tensive diseases were compared in those women who
placentae. The urine drug screen was negative. received aspirin or placebo, as compared with the 316
1088 Hauth et al. April 1993
Am J Obstet Gynecol

15.0-r---------------------,
Aspirin
II Placebo
o Failed "Run In"

10.0

5.0

0.0

Fig. 3. Pregnancy-associated hypertensive disorders in women who received aspirin or placebo or

who failed "run in" trial. asterisks, p < 0.009 and p < 0.05 for aspirin group compared with placebo
and "run in" failure groups, respectively. PAH, Pregnancy-associated hypertension; PIH, pregnancy-
induced hypertension.

Table III. Effect of maternal aspirin treatment (60 mg/day) on the incidence of pregnancy-associated
hypertensive disorders
Aspirin group Placebo group 95% Confidence
(N = 302) (N = 302) Odds ratio interval

Pregnancy-induced 19 (6.3%) 17 (5.6%) 1.19 0.62-2.27

hypertension
Preeclampsia
Mild 4 II
Severe I 6
Total 5 (1.7%) 17 (5.6%) 0.29 0.11-0.79*
Pregnancy-induced 24 (8%) 34 (11.3%) 0.71 0.43-1.16
hypetension or
preeclampsia

*P = 0.009.

women who failed the "run in" trial (pregnancy-in-
duced hypertension 5.8%, preeclampsia 4.1%). The
incidence of any hypertension or pregnancy-induced
hypertension was similar in all three groups. However,
preeclampsia was increased in those who failed the "run
in" or who received the placebo as compared with those
who received aspirin (Fig. 3). The birth weight at
delivery, gestational age at delivery, and perinatal mor-
tality are depicted in Fig. 4 for the three groups. There
were six pregnancy losses after 24 weeks' gestation in
the 301 women who failed the "run in" and in whom all
outcome data was available. These exceeded the num-
ber occurring in those who received aspirin or placebo
(p = 0.06). Of the six perinatal losses in the women who
failed the "run in," four were fetal deaths and two were
neonatal deaths. Among these deaths, two of the moth-
ers had hypertensive disorders. Three of the six losses
were in newborns who weighed < 750 gm at delivery.
Comment
TxA2, whether of placental or platelet ongm, is a
powerful vasoconstrictor, whereas prostacyclin, of endo-
thelial cell origin, is a potent vasodilator. In the placen-
tas of women with preeclampsia, the production of
TxA2 is increased and that of prostacyclin is decreased
relative to their production in placentas from normal
pregnancies. 10 Whether this alteration in the placental
production of these eicosanoids occurs after the onset
of the clinical syndrome of hypertension, edema, and
proteinuria or precedes and mediates the development
of preeclampsia has not been determined.
If a model for the development of preeclampsia
presumes that there is an early and increasingly (with
advancing gestational age) disparate ratio between ma-
ternal TxA2 and prostacyclin levels, then many clinical
and physiologic observations regarding preeclampsia
can be explained. Women with increased sensitivity to
angiotensin II as early as 24 weeks' gestation have an
increased incidence of preeclampsia. 3 Both these phe-
nomena would be expected if these women had an
excessive level of an endogenous vasoconstrictor or a
deficit of an endogenous vasodilator. Also, most preg-
nant women have a lowered total peripheral vascular
resistance. This is probably because of an increase of
Volume 168, Number 4
Am J Obstet Gynecol
10
Perinatal Mortality
iii
D 5
E
:::>
c: N=1'
0
40
Gestational Age
""'"
Q)
39
~
38
3,300
Birthweight
3,250
'"~ 3,200
C> 3,150
3,100
Aspirin
Fig. 4. Mean birth weight, gestational age, and perinatal mortality in women who received aspirin or
placebo or who failed "run in" trial. asterisks, Increased perinatal mortality in 301 women who failed
"run in" as compared with either aspirin or placebo group, p = 0.06.
various smooth muscle relaxants, including progester-
one, prostacyclin, and endothelial-derived relaxing fac-
tor. Thus normal pregnant women tend to have an
"underlilled" intravascular capacity, as viewed, so to
speak, by their renal and adrenal systems. However,
women with excessive vasoconstriction and increased
peripheral resistance from early gestation would be
expected to have a diminution of the normal increase of
plasma and red blood cell volume. This lack of intra-
vascular volume expansion commonly occurs in women
with severe preeclampsia or eclampsia. 14 A continued
maternal imbalance of TxB2 and prostacyclin, favoring
vasoconstriction, would explain the increased occur-
rence of uteroplacental insufficiency and the resultant
fetal growth impairment that is .common in women with
severe preeclampsia or eclampsia. Finally, with worsen-
ing of the vasoconstrictor/vasodilator imbalance, the
clinical syndrome of hypertension, edema, and protein-
uria would become manifest. Sequelae of this syndrome,
such as hepatic dysfunction, major motor seizures, en-
dothelial disruption with microangiopathic hemolytic
anemia, and thrombocytopenia can also be explained
by excessive arteriolar spasm produced by an increased
eicosanoid vasoconstrictor. If an imbalance in vasocon-
strictor and vasodilator substances occurs before the
overt clinical syndrome of preeclampsia and it either
causes or mediates this illness, then pharmacologic
prophylactic intervention before the clinical syndrome,
aimed at reversing this imbalance, should ameliorate or
prevent subsequent preeclampsia.
The dosage of aspirin necessary to inhibit TxA2
synthesis without affecting prostacyclin synthesis is very
low. Masotti et a1.15 reported platelet cyclooxygenase
inhibition with a single aspirin dose of 3.5 mg/kg of
Hauth et al. 1089
N=6
N=1'
3,127
Placebo Failed 'Run In'
body weight (or about 300 mg) with minimal inhibition
of prostacyclin production. Ylikorkala et a1. 16 gave ei-
ther a placebo or 100 or 500 mg of aspirin to 27 healthy
women at term and found that at delivery (0.5 to 11
hours after treatment) umbilical artery prostacyclin me-
tabolite levels were similar in both the placebo and 100
mg group and lower only in the 500 mg group. In
contrast, TxB2 levels were significantly higher in the
placebo group than in women who ingested either 100
or 500 mg of aspirin. Hoogendijk and ten Cate l7
reported that in women treated with "low-dose aspirin"
the platelet production of malondialdehyde progres-
sively declined to 95% of baseline production within 10
days, which is similar to the platelet life span. Spitz et
al. 6 and Brown et a1." found that a maternal dosage of
81 mg of aspirin significantly decreased the TxB2/6-
keto-PGF ln ratio. Sibai et al." tested 20, 60, and 80 mg
aspirin dosages and reported that only the 60 and 80
mg dosages significantly altered the TxB 2/prostacyclin
ratio to favor the vasodilator. Recently, Schiff et al. 9
reported that women who received 100 mg of aspirin
daily had a 35% decrease in their serum TxB 2/6-keto-
PGF In ratio, whereas the ratio in those who received the
placebo increased by 51 % with advancing gestation.
Similarly, in our study patients taking the placebo
median serum TxB2 levels increased by 28% from 24
weeks' gestation to delivery (mean 38.9 weeks) but
decreased by 60% from 24 weeks' gestation to delivery
(mean 39.1 weeks) in the women who received 60 mg of
aspirin daily.
In 1979 Crandon and Isherwood ls reported a lower
incidence of preeclampsia in nulliparous women who
took aspirin more than once each 2 weeks compared
with those who had taken no aspirin. Beaufils et al. 19
1090 Hauth et al.
accomplished the first trial of early (13 to 26 weeks'
gestation) maternal aspirin prophylaxis to ameliorate
or prevent preeclampsia. They randomized 102 women
at high risk for the development of preeclampsia or
fetal growth retardation to either 150 mg of aspirin and
300 mg of dipyridamole daily or to a control group.
Women who received aspirin had a significantly lower
incidence of preeclampsia and increased birth weights
and gestational ages at delivery. No significant mater-
nal or newborn hematologic abnormalities occurred
with the maternal aspirin treatment. Fetal and neonatal
losses occurred only in the control group (5 of 45) and
none of the newborns whose mothers took aspirin were
severely growth retarded compared with four newborns
in the control group. Significantly more infants weighed
less than the 10th and 3rd percentiles for gestational
age at delivery in the control group. Thus maternal
treatment with low-dose aspirin plus dipyridamole re-
duced the incidence of hypertensive disorders and fetal
growth retardation. Wallenburg et al. 11 also found a
significant reduction in preeclampsia in 44 primigravid
women selected to be at high risk on the basis of a
positive angiotensin sensitivity test at 28 weeks' gesta-
tion who were then randomized to either a placebo or
to 60 mg of aspirin daily until delivery. Schiff et a1. 9
used a rollover test to screen 791 pregnant women who
had various risk factors for preeclampsia. The 69
women with an abnormal rollover test were randomized
to either 100 mg of aspirin a day (n = 36) or a placebo
(n = 31). These authors also observed less preeclamp-
sia in the women who received aspirin (seven of 31 vs
one of 34), P = 0.023. However, like Wallenburg et
al.,11 they found no difference in the incidence of
pregnancy-induced hypertension (three of 34 vs four of
31).
Our data confirm that pharmacologic prophylaxis
significantly reduced the incidence of preeclampsia in
the aspirin group (five of 302 vs 17 of 302 in the
placebo group, p = 0.009). The severity of preeclamp-
sia was also ameliorated, because severe preeclampsia
developed in only one of 302 women who received
aspirin compared with six of 302 who received placebo
<p = 0.029). Our study population was different from
those studied previously, in that all women known to
have any medical or historic risk factor associated with
an increased risk of hypertensive disorders were ex-
cluded. The women selected had only to be nulliparous,
be available to participate at < 23 weeks' gestation, and
to have successfully completed a 2- to 3-week "run in"
trial.
Other than a reduction in preeclampsia in 12 of 302
women who received aspirin, we did not find additional
evidence of significant maternal, fetal, or neonatal ben-
efit. Because our sample size was based on the end
points of pregnancy-induced hypertension and pre-
eclampsia, there was not sufficient power to find im-
April 1993
Am J Obstet Gynecol
provement, if present, in perinatal mortality or morbid-
ity. Others have previously reported evidence for a
positive effect of low-dose aspirin on fetal growth. 11. 19
Most recently, Uzan et al. 20 reported significantly less
fetal growth retardation and increased birth weight in
women who received 150 mg of aspirin a day compared
with placebo. Our data were suggestive only of a posi-
tive effect of aspirin on fetal growth in that the mean
birth weight of infants of aspirin-treated women was
3249 gm compared with 3169 gm in those who received
the placebo <p = 0.08).
Should all nulliparous women receive daily prophy-
lactic low-dose aspirin therapy? Before any such recom-
mendation can be made, several factors must be con-
sidered. First, although all the women in this study were
medically at low risk, the fact that many were black and
all were poor undoubtedly accounted for the relatively
high background rate of preeclampsia. A middle-class
white population would be expected to have a lower
background risk of preeclampsia, and the effect of
aspirin would probably be less dramatic. Second, uni-
versal aspirin prophylaxis in nulliparous women would
be more attractive if an improvement in fetal or neo-
natal morbidity or mortality could be demonstrated.
However, even if an improvement in the perinatal
outcome cannot be demonstrated, universal aspirin
prophylaxis might still be valuable. Treating 600
women with daily low-dose aspirin might well cost less
than providing more intensive medical care for the 12
additional women with preeclampsia. Data for that
analysis are being collected.
Finally, the question of universal aspirin prophylaxis
in nulliparous women would not arise if there were
better predictors of preeclampsia in nulliparous women
with no other risk factors. In spite of the apparent
efficacy of aspirin for preventing preeclampsia in this
population, a good predictive test for preeclampsia
would be extremely useful. Nevertheless, in populations
at high risk for preeclampsia because of proverty, race,
and age, daily low-dose aspirin holds promise of achiev-
ing a substantial reduction in the incidence and severity
of preeclampsia and perhaps an increase in birth
weight.
REFERENCES
1. Cunningham FG, MacDonald PC, Gant NF. Hypertensive
disorders in pregnancy. In: Cunningham FG, MacDonald
PC, Gant NF, eds. Williams' obstetrics. 18th ed. Norwalk,
Connecticut: Appleton & Lange, 1989:653.
2. Kaunitz AM, Hughes jM, Grimes DA, Smith jC, Rochat
RW, Kafrissen ME. Causes of maternal mortality in the
United States. Obstet Gynecol 1985;65:605-12.
3. Gant NF, Daley GL, Chand S, Whalley Pj, MacDonald PC.
A study of angiotensin II pressor response throughout
primigravid pregnancy. j Clin Invest 1973;52:2682-9.
4. Everett RB, Worley Rj, MacDonald PC, Gant NF. Effect of
prostaglandin synthetase inhibitors on pressor response to
angiotensin II in human pregnancy. j Clin Endocrinol
Metab 1978;46:1007-10.
Volume 168, Number 4
Am J Obstet Gynecol
5. Sanchez-Ramos L, O'Sullivan M], Garrido-Calderon ].
Effect of low-dose aspirin on angiotensin II pressor re-
sponse in human pregnancy. AM] OBSTET GYNECOL 1987;
156:193-4.
6. Spitz B, Magness RR, Cox SM, Brown CEL, Rosenfeld CR,
Gant NF. Low-dose aspirin. I. Effect on angiotensin II
pressor responses and blood prostaglandin concentrations
in pregnant women sensitive to angiotensin II. AM ]
OBSTET GYNECOL 1988;159:1035-43.
7. Brown CEL, Gant NF, Cox K, Spitz B, Rosenfeld CR,
Magness RR. Low-dose aspirin. II. Relationship of angio-
tensin II pressor responses, circulating eicosanoids, and
pregnancy outcome. AM] OBSTET GYNECOL 1990;163:1853-
61.
8. Sibai BM, Mirro R, Chesney CM, Leffler C. Low-dose
aspirin in pregnancy. Obstet Gynecol 1989;74:551-7.
9. Schiff E, Peleg E, Goldenberg M, et al. The use of aspirin
to prevent pregnancy-induced hypertension and lower the
ratio of thromboxane A2 to prostacyclin in relatively high
risk pregnancies. N Engl] Med 1989;321:351-6.
10. Walsh SW. Preeclampsia: an imbalance in placental pros-
tacyclin and thromboxane production. AM] OBSTET Gy-
NECOL 1985;152:335-40.
11. Wallenburg HCS, Makovitz ]W, Dekker GA, Rotmans P.
Low-dose aspirin prevents pregnancy-induced hyperten-
sion and preeclampsia in angiotensin-sensitive primi-
gravidae. Lancet 1986;1:1-3.
12. Wirtschafter DD, Blackwell WC, Goldenberg RL, et al. A
county wide obstetrical automated medical record system.
] Med Syst 1982;6;277-90.
13. Brenner WE, Edelman DA, Hendricks CH. A standard of
fetal growth for the United States of America. AM] OBSTET
GYNECOL 1976;126:555-64.
14. Pritchard]A, Cunningham FG, Pritchard SA. The Park-
land Memorial Hospital protocol for treatment of
eclampsia: evaluation of 245 cases. AM ] OBSTET GYNECOL
1984;148:951-63.
15. Masotti G, Galanti G, Poggesi I, Abbate R, Serneri GG.
Differential inhibition of prostacyclin production and
platelet aggregation by aspirin. Lancet 1979;2: 1213-7.
16. Ylikorkala 0, Makila UM, Kaapa P, Viinikka L. Maternal
ingestion of acetylsalicylic acid inhibits fetal and neonatal
prostacyclin and thromboxane in humans. AM ] OBSTET
GYNECOL 1986; 155:345-9.
17. Hoogendijk EM, ten Cate ]W. Aspirin and platelets. Lan-
cet 1980; 1:93-4.
18. Crandon A], Isherwood DM. Effect of aspirin on incidence
of preeclampsia. Lancet 1979; 1: 1356.
19. Beaufils M, Uzan S, Donsimoni R, Colau]C. Prevention of
pre-eclampsia by early anti platelet therapy. Lancet 1985;
1:840-2.
20. Uzan S, Beaufils M, Breart G, Bazin B, Capitant C, Paris].
Prevention of fetal growth retardation with low-dose
aspirin: findings of the EPREDA trial. Lancet 1991;337;
1427-31.
Discussion
DR. FRANK C. MILLER, Lexington, Kentucky. The an-
algesic property of willow bark, which contains salicy-
lates, was known by Hypocrites and was used by Native
Americans, among others. Aspirin was developed and
became available in 1899 as a result of research by Felix
Hoffman at Bayer Industries. lOver the next 60 years
aspirin became the most widely used pharmaceutical
product in the world. Yet the use of aspirin to alter
prostaglandin production was not contemplated until
the 1970s, when it was discovered that aspirin affects
platelet aggregation through interference with platelet
synthesis of prostaglandin, presumably TxA2, a potent
vasoconstrictor and promoter of platelet aggregation
Hauth et al. 1091
produced mainly by platelets. Aspirin irreversibly inhib-
its cyclooxygenase and nonnucleated platelets cannot
produce cyclooxygenase. Thromboxane synthesis by
platelets can recover only as new platelets enter the
circulation. The life span of platelets is about 8 to 11
days. The endothelial cells that produce prostacyclin
can recover from aspirin suppression more rapidly and
thus change the thromboxane/prostacyclin ratio in fa-
vor of prostacyclin. It seems reasonable therefore that
aspirin may have a therapeutic role in the prevention of
preeclampsia, a disease characterized by vasoconstric-
tion and increased platelet aggregation.
The first reported prospective clinical trial using
low-dose aspirin to prevent preeclampsia was by Beau-
fils in 1985, followed closely by Wallenberg and others
(see references 11 and 19 of article). In these early trials
with small numbers of patients, there was a suggestion
that aspirin in doses of 60 to 150 mg daily during the
second and third trimesters reduced the instance of
pregnancy-induced hypertension. Subsequently, larger
trials showed significant reduction only in true pre-
eclampsia (hypertension plus proteinuria) and not
pregnancy-induced hypertension. It is not uncommon
in the medical literature for several clinical trials with
low statistical power to demonstrate no difference be-
tween compared groups. In situations such as this,
meta-analysis may be useful. In 1991 Imperiale and
Petralis 2 reported a meta-analysis of clinical trials of
low-dose aspirin for the prevention of pregnancy-in-
duced hypertension. Six clinical trials meeting their
inclusion criteria were evaluated by means of the fol-
lowing nine selected standards and principles for clin-
ical trials: (l) randomization, (2) blinded ness (single,
double, or nonblinded), (3) assessment of compliance,
(4) explicit inclusion criteria, (5) explicit exclusion cri-
teria, (6) baseline similarity between groups of impor-
tant features (these included maternal age, parity,
diastolic blood pressure, weight, and smoking), (7) ad-
ministration of principal maneuver (aspirin) in an
equivalent and comparable fashion within and among
trials, (8) equivalent and comparable use of ancillary
treatments, and (9) equal detection of outcomes be-
tween groups. The overall results of these six clinical
trials suggest that low-dose aspirin significantly reduces
the risk of pregnancy-induced hypertension and severe
low birth weight among newborns with no adverse
effects associated with taking aspirin.
The current study is the first to address the benefits
of low-dose aspirin in preventing preeclampsia in the
general patient population of normal nulliparous
women without pre screening for angiotensin-II sensi-
tivity or the "rollover" test. This prospective, double-
blind randomized study was designed to test the hy-
pothesis that low-dose (60 mg) aspirin treatment re-
duces the incidence or severity of pregnancy-associated
hypertension in a healthy group of nulliparous women.
The study design meets all of the standards and prin-
ciples of clinical trials outlined above. Patient selection
occurred at 20 to 22 weeks' gestation. A 2-week "run in"
was conducted to determine compliance primarily by
means of serum TxB2 levels > 1000 pg/ml on the
1092 Hauth et al.
second visit. Three hundred sixteen of 962 women
screened for the study failed the "run in" and were
followed in the routine clinic. Their outcomes were
evaluated separately. Fig. 1 outlines the study protocol,
which includes plasma 6 keto-PGF In and umbilical
Doppler flow assessment at randomization and 28 to 30
and 34 to 36 weeks' gestation.
Each group had one fetal death. The fetal death in
the woman receiving aspirin was secondary to abruptio
placentae associated with preeclampsia. No significant
differences were observed in induced or spontaneous
delivery rates, premature rupture of membranes, induc-
tion of labor, or cesarean delivery. All newborn param-
eters were stated to be similar, with only a trend toward
higher birth weights in the aspirin-treated group. The
umbilical Doppler flow results were not reported. Fig. 2
depicts the serum TxB2 levels in each group. Women
who received aspirin had significantly lower serum lev-
els at all three sample points. SEM bars should be
included to demonstrate variation.
The combined incidence of pregnancy-induced hy-
pertension and preeclampsia was not significantly dif-
ferent between groups, but preeclampsia and in partic-
ular severe preeclampsia was significantly less frequent
in the aspirin-treated group.
The women who failed the "run in" trial had essen-
tially the same incidence of any hypertension or preg-
nancy-induced hypertension as did the aspirin and
placebo group but more preeclampsia and more preg-
nancy losses after 24 weeks.
I would like to ask the following questions of Dr.
Hauth: (1) Why did the authors choose to define pre-
eclampsia as two elevated blood pressure recordings
only 1 hour apart and not at 6 hours apart as in most
standard texts? (2) Why did the authors report only the
levels of serum TxB 2, a metabolite of thromboxane,
and not 6-keto-PGF In' the metabolite of the vasodilator
prostacyclin, because the proposed mechanism of low-
dose aspirin is an alteration in the TxA2/prostacyclin
ratio in favor of prostacyclin? (3) Did the authors
perform umbilical Doppler flow assessments as stated in
the study design? If so, what were the results, and were
they clinically significant? (4) The incidence of labor
induction (nonindicated) was reported as 26% in the
aspirin group and 28% in the placebo group. Why did
so many normal, healthy primiparous women have
induced labor? and (5) How can one translate the
observation that the incidence of pregnancy-associated
and pregnancy-induced hypertension is not altered
with low-dose aspirin, whereas preeclampsia is?
REFERENCES
1. Mills JA. Aspirin, the ageless remedy? N Engl J Med
1991;325:1303-4.
2. Imperiale TF, Petralis AS. A meta-analysis of low-dose
aspirin for the prevention of pregnancy induced hyperten-
sive disease. JAMA 1991;266:260-4.
DR. ROBERT C. CEFALO, Chapel Hill, North Carolina.
Dr. Hauth, you do recognize that spiral arteriole prob-
lems associated with preeclampsia may begin very early
in pregnancy. My question is why did you select 22 to
April 1993
Am J Obstet Gynecol
24 weeks rather than 10 to 12 weeks to have your
patients start the aspirin?
DR. RONALD S. GIBBS, Denver, Colorado. Dr. Hauth, I
compliment you on the design, conduct, and analysis of
your study. I would like some help on the ability to
generalize your results to the population of nulliparous
women as a whole. In particular, I would be interested in
the concordance of your diagnosis of pregnancy-induced
hypertension with the diagnosis of clinicians in general.
DR. JlAROLD SCHULMAN, Mineola, New York. We have
heard about the state of our knowledge, and I person-
ally found it disappointing that we would begin a study
of preeclampsia by selecting entry criteria of nullipa-
rous women. I would hope we know more about the
disease than to select that as an entry criterion.
There are three studies, two in Britain and one in
France, in which the study entry point was 24 weeks,
and an abnormal uteroplacental Doppler test result was
used as the entry criterion. The study was then random-
ized, aspirin versus placebo, and showed a significant
reduction in both pregnancy-induced hypertension and
proteinuric hypertension, much more impressive than
your results.
So I think your study design is severely diluted with
normal pregnancies. There are so many normal preg-
nancies involved in this series that showing significant
differences would be a problem.
You mentioned that you used umbilical Doppler
ultrasonography in the evaluation of these women. I
hope a break point of ~ 3 SD was used to evaluate its
significance. I am curious about what your data showed.
DR. ROBERT F. CAsPER, Toronto, Ontario, Canada. I
wonder if it would be sensible to use aspirin every
second or third day rather than every day. Aspirin is an
irreversible inhibitor of cyclooxygenase. Blockade of
thromboxane and platelets is long lasting because the
platelet doesn't have the mechanism to regenerate
thromboxane. However, endothelium does have the
ability to regenerate cyclooxygenase and produce more
prostacyclin.
Dr. Hauth has demonstrated that thromboxane was
blocked very well, and still there wasn't much difference
in the two groups. I wonder whether there would be a
better chance of getting a prostacyclinlthromboxane
balance if there were an interruption in the aspirin.
DR. FREDERICK ZUSPAN, Columbus, Ohio. The cause of
preeclampsia is not known, but perhaps it's related to
an implantation defect at the time of conception. Ad-
ditionally, it's probably also related to an alteration in
the adrenergic mechanism around the base of the spiral
arteries, which reinforces what Dr. Gibbs said. Why not
start the aspirin at 10 or 12 weeks of gestation?
Second, there was no difference in perinatal outcome,
so it would seem to me that your next entry is going to
have to constitute more ill patients to define whether
this regimen is of value. I congratulate you on a care-
fully done study.
DR. HAUTH (Closing). Our design that was formulated
in 1987 is similar to that recommended in a recent
description of metaanalysis criteria regarding low-dose
aspirin for prevention of preeclampsia.
Let me answer first that we anticipated difficulty with
Volume 168, Number 4
Am J Obstet Gynecol
a consensus diagnosis. Currently we are part of a
National Institute of Child Health and Human Devel-
opment calcium trial and a National Institute of Child
Health and Human Development trial of high-risk
patients with low-dose aspirin. One difficulty was con-
currence of the National Institute of Child Health and
Human Development Maternal-Fetal Medicine Units
Network Centers on the diagnosis of preeclampsia. The
American College of Obstetricians and Gynecologists
(ACOG) Hughes definition uses 6 hours and the Inter-
national Society for the Study of Hypertension in Preg-
nancy (ISSHP) definition is 4 hours. Moreover, the
method of blood pressure determination is difficult
because few large centers have total control of labor and
delivery procedures. Internists do not understand that
the labor and delivery environment may last many
hours. During that time patients are not sitting quietly
before a blood pressure determination. They have con-
tractions, receive regional anesthetics, and are left in
the left lateral position very frequently.
As summarized, before the double-blind code was
broken, in addition to diagnosing preeclampsia with
our described criteria, we also categorized these pa-
tients by the ISSHP and the ACOG (Hughes) criteria.
The ISSHP uses only the diastolic pressure with an
open-ended postpartum duration. The Hughes method
uses 15 mm Hg diastolic and 30 mm Hg systolic from
20 weeks' gestation on and up to 24 hours post partum.
Our definition was most similar to that of the ISSHP
since we just used the diastolic blood pressure. Our
definition resulted in identical numbers of women with
mild and severe preeclampsia as compared with those
of the ISSHP and similar to those of the Hughes
method.
We investigated the origin of the 6-hour rule when
designing the National Institute of Child Health and
Human Development calcium and low-dose ASA trial in
high-risk patients. It likely was codified by the Commit-
tee on Maternal Welfare that was active in the 1940s
and 1950s. Considering the practice profiles of the
1940s and 1950s, my sense is that women with hyper-
tension were sent home from clinics and told to come
back in the morning, hence a 6-hour rule. The 4- and
6-hour definitions were not based on a prospective trial
as related to outcomes.
As stated, our definition of pregnancy-induced hy-
pertension and preeclampsia resulted in an identical
number of these diagnoses as with the ISSHP defini-
tion. Our numbers were similar to those with the ACOG
definition except for more mild pregnancy-induced
hypertension with the ACOG definition, which was
because of the intrapartum 30 or 15 mm Hg rule as
compared with blood pressures early in pregnancy.
Dr. Miller asked about prostacyclin in the plasma. We
used serum thromboxane because of the large values
with means of 8000 to 10,000. We reported these values
to assess population compliance (ASA versus placebo).
We will evaluate individual patient compliance in the
future.
We do have plasma for prostacyclin and thrombox-
ane analysis, but we did not measure serum prosta-
cyclin.
Hauth et al. 1093
As for labor inductions, most were for postterm ges-
tations and oligohydramnios. Our concern was that we
were causing oligohydramnios, but after the code was
broken, we found that oligohydramnios was similar in
women who received placebo and in those who received
aspirin (11 % and 13%, respectively). The 316 women
with failure who were followed up in routine clinics only
had a 6% incidence of oligohydramnios, but in that
group there were six fetal deaths, and all were associ-
ated with fetal growth restriction. It is possible that, with
the research clinic care and three or four additional
sonograms, we detected more oligohydramnios even in
low-risk women and that was the reason for the induc-
tions.
Why did aspirin only reduce preeclampsia? I don't
know. My opinion is that pregnancy-induced hyperten-
sion is probably a poor term. A better term may be
transient gestational hypertension. Also, if systolic blood
pressure is used, the pain and anxiety of labor may be
the cause of the increase in systolic pressure. Chesley'
has taught us over many years and most recently in his
1985 article that preeclampsia is a more precise diag-
nOSIS.
Why select 20 to 24 weeks instead of 12? I don't think
we would have gotten 10 to 12 weeks through our
research committee. Also, pragmatically, in our obstet-
ric population we don't see most of the patients at 10
weeks. We care for an indigent Medicaid population
that may not be generalized to private patients, which
was Dr. Gibbs' question.
In terms of general recommendations, we are not
prepared to recommend that all nulliparous women
take low-dose aspirin. We are performing various sub-
analyses related to changes in serum thromboxane in
individual patients in our population. The results may
allow us to select out the highest-risk patients in our
population.
I didn't understand Dr. Schulman's question. Our
data did show significance in favor of a positive low-
dose aspirin effect. I am aware of his interest in Doppler
technology and his belief that it predicts many adverse
outcomes. I remember him saying once about 8 years
ago, at ACOG's OB Committee, that the Doppler tech-
nology would revamp how we diagnose preeclampsia.
That has not happened and our Doppler data on these
patients do not support his premise.
No, there wasn't a different perinatal outcome, but
the power of our population wasn't designed to address
that issue. We designed the trial in two ways. One was to
evaluate intent to treat (i.e., can we get a population to
take a medicine and within that population show a
benefit?). In our subanalysis, to be reported later, we
will evaluate individual patients' serum thromboxane
levels to reflect who actually complied and who were
crossovers. This "compliance analysis" may answer your
perinatal benefit question.
REFERENCE
1. Chesley LC. Diagnosis of preeclampsia. Obstet Gynecol
1985;65:423.