Bellomo G, Narducci PL, Rondoni F, Pastorelli G, Stangoni G, Angeli G, Verdecchia P. Prognostic Value of 24-Hour Blood Pressure in Pregnancy. JAMA. 1999;282(15):1447-1452. doi:10.1001/jama.282.15.1447
Author Affiliations: Divisions of Medicine (Drs Bellomo and Rondoni), Obstetrics (Drs Narducci and Angeli), and Neonatology (Drs Pastorelli and Stangoni), Assisi Hospital, Assisi, and Cardiovascular Department, Hospital R. Silvestrini, Perugia (Dr Verdecchia), Italy.
Context Elevated blood pressure (BP) measured at the physician's office may
reflect true hypertension or white coat hypertension (WCH). The prognostic
value of WCH among pregnant women is unknown.
Objective To assess the prognostic value of WCH in pregnancy.
Design Prospective cohort study conducted between September 1994 and October
Setting Community hospital.
Patients Women without preexisting hypertension and not treated with antihypertensive
drugs and with high (n = 148) or normal (n = 106) office BP (high office BP
was defined as ≥140 mm Hg systolic and/or ≥90 mm Hg diastolic) matched
for gestational age during their third trimester of pregnancy. All women underwent
24-hour noninvasive BP monitoring, and women without hypertension on 24-hour
monitoring (125/74 mm Hg or less for average 24-hour BP) with office hypertension
were classified as having WCH. Women were followed up through the end of pregnancy.
Main Outcome Measures Duration of pregnancy, gestational hypertension, preeclampsia or eclampsia,
cesarean delivery, placental and neonatal weight, and length of maternal and
neonatal hospital stays for those with and without elevated office BP.
Results After application of exclusion criteria, data for 7 women were removed
from the analysis. For the remaining subjects, in the group with elevated
BP, prevalence of WCH was 29.2% (42/144). Duration of pregnancy was similar
in the normotensive and WCH groups (39.6 vs 39.8 weeks;P = .50), but shorter (38.3 weeks; P<.001)
in the true hypertension group. Incidence of preeclampsia was similar in the
normotensive and WCH groups (5.8% vs 7.1%; P = .86)
but higher in the true hypertension group (61.7%; P<.001).
Frequency of cesarean delivery was lower in the normotensive (12.4%) than
in the WCH (45.2%; P = .008) and true hypertension
(41.1%; P = .009) groups. Neonatal weight was lower
(P<.001) in the true hypertension (mean, 2911
g) than in the normotensive (3336 g) and WCH groups (3435 g), which did not
differ (P = .68). The duration of neonatal hospital
stay did not differ between the normotensive and the WCH group (5.3 vs 6.9
days; P = .13) but was longer in the true hypertension
group (12.3 days;P<.001).
Conclusions In women with elevated BP during their third trimester of pregnancy,
24-hour BP was superior to office BP (distinguishing true hypertension from
WCH) for prediction of the outcome of pregnancy. Outcomes in the normotensive
and WCH group were comparable, but the increased incidence of cesarean delivery
in the WCH group may reflect decision-making processes influenced by office
White coat hypertension (WCH) is generally defined by a persistently
elevated office blood pressure (BP) with normal BP outside the medical setting.1 Like any other category of hypertension, the definition
is arbitrary. A diagnosis of WCH is established by 24-hour BP monitoring,
and its prevalence has been estimated at 5% to 60%, depending on the clinical
characteristics of the populations studied and the diagnostic criteria used.2- 6
Whether WCH is a benign condition7- 9
or carries a definitely increased cardiovascular risk is not known.10- 13
White coat hypertension is common among young, nonobese women of childbearing
age.2 At present, there is little information
regarding prevalence and prognostic significance of WCH in pregnancy, although
WCH has been reported to be as frequent or even more frequent among pregnant
than nonpregnant women.14- 16
If this is the case, then a number of pregnant women may be subjected to expensive
diagnostic procedures and even drug treatment for a hypertensive condition
that might actually prove to be limited to the clinical visit.
We used 24-hour BP monitoring to evaluate a group of 148 clinically
hypertensive pregnant women in their third trimester of pregnancy and compared
the findings with those of a group of 106 clinically normotensive pregnant
women at the same gestational age. We followed up all subjects to completion
of pregnancy to establish prevalence and prognostic value of WCH in the third
trimester of pregnancy. The main evaluation criteria were duration of pregnancy,
incidence of transient gestational hypertension, preeclampsia and eclampsia,
incidence of cesarean delivery, placental and neonatal weight, Apgar score,
and length of hospital stay of mother and newborn.
The study began in September 1994 and ended in October 1997. The study
population consisted of 148 women with office hypertension in their third
trimester of pregnancy, between week 26 and week 38, and a control group composed
of 106 normotensive women at the same stage of pregnancy. At entry to the
study, BP was obtained on 2 separate occasions in a quiet environment with
the subject seated at an angle of approximately 60° to the horizontal.
A physician on the hospital staff measured BP with a standard mercury sphygmomanometer.
In our institution, all mercury sphygmomanometers are calibrated every 3 months
against an electronic pressure generator. Diastolic BP was taken at Korotkoff
phase 5, which has been shown to correspond more closely to intra-arterial
measurement and be less dependent on the examiner's skill than Korotkoff phase
4.17,18 Three measurements were
taken, 5 minutes apart, and the average was recorded: the mean of the averages
recorded on the 2 occasions was considered the actual BP. Blood pressure values
equal to or greater than 140 mm Hg systolic or 90 mm Hg diastolic were considered
abnormal and used for diagnosis of hypertension. Known heart disease, nephropathy,
or hypertension preceding pregnancy were reasons for exclusion from the study.
None of the subjects were treated with antihypertensive drugs on entry. All
the women participating in the study were informed about its purpose and the
procedures involved and gave their consent. The study was approved by the
hospital ethics committee.
All the participating women were admitted to the hospital on the day
preceding ambulatory BP monitoring, blood samples were taken, and a 24-hour
urine collection was started for determination of proteinuria.
A pressurometer validated for use in pregnant women (TM2420, A&D
Company, Tokyo, Japan) was used to measure 24-hour BP.19
Device accuracy was checked in every session against a mercury sphygmomanometer,
and a difference greater than 5 mm Hg was reason for exclusion from the study.
The pressurometer takes systolic BP at Korotkoff phase 1 and diastolic BP
at Korotkoff phase 5 with a microphone. Blood pressure was recorded at 15-minute
intervals both during the day and night for 25 hours, starting between 9 and
11 AM. Cuffs of appropriate size were used, always applied to the left arm.
The subjects were free to move around in the hospital and encouraged to engage
in social activities with friends and relatives. We chose to perform 24-hour
BP monitoring within a hospital setting to ensure a more controlled and homogeneous
environment for all subjects in terms of meal schedules, sleeping times, and
physical activity. Moreover, in our experience the number of artifacts and
missing values tends to be lower in a hospital setting than in an outpatient
setting when using an auscultatory method for BP recording (G.B., unpublished
All BP recordings from the first hour of monitoring were removed from
analysis because they might be influenced by an alarm reaction.1,20
All the readings with a pulse pressure lower than 20 mm Hg, a diastolic BP
greater than 140 mm Hg or lower than 40 mm Hg, or a systolic BP greater than
250 mm Hg or lower than 60 mm Hg were automatically rejected. The quality
of 24-hour BP monitoring was judged satisfactory when at least 70% of the
readings passed the editing criteria and at least 1 recording per hour was
obtained. Daytime was defined as 6 AM to 10 PM. Our upper normal limits for
24-hour BP monitoring were 125/74 mm Hg for average 24-hour BP, 128/78 mm
Hg for average daytime BP, and 121/70 mm Hg for average nighttime BP. These
reference values mark the 90th percentile of the distribution of 24-hour BP
monitoring in a population of 132 normotensive pregnant women (72 primipara
and 60 multipara) of mean age 29 years (range, 18-41 years) attending our
hospital between 1991 and 1994, whose office BP was persistently below 140
mm Hg systolic and 90 mm Hg diastolic during the entire pregnancy and puerperium.
In all these women, 24-hour BP monitoring was carried out during the third
trimester of pregnancy (26-40 weeks of gestation) in the hospital. Their mean
daytime 24-hour BP was 113 mm Hg (range, 92-136 mm Hg) systolic and 72 mm
Hg (range, 57-84 mm Hg) diastolic.
For the purpose of this study, WCH was defined as the presence of office
hypertension (office BP ≥140 mm Hg systolic and/or 90 mm Hg diastolic)
and mean 24-hour, daytime, and nighttime systolic and diastolic BP below the
reference values. The women with higher 24-hour BP were classified as having
Ambulatory BP reports were accessible to gynecologists and patients,
but these data were unlikely to influence therapeutic decisions because they
were considered experimental findings unsupported by specific guidelines.
All women were followed up at our hospital until the end of pregnancy
and regularly up to 1 month after delivery, unless their clinical conditions
dictated otherwise. According to standard criteria,21,22
preeclampsia was diagnosed as the presence of office BP greater than 140/90
mm Hg and proteinuria greater than 300 mg/24 h. Transient late gestational
hypertension was defined as the presence of normal office BP (<140 mm Hg
systolic and <90 mm Hg diastolic) 4 weeks after delivery in untreated women
who had hypertension without preeclampsia during the third trimester of pregnancy.
Statistical analysis was carried out using 1-way analysis of variance
and the Scheffé test or the χ2 test with Yates correction,
as appropriate. Results are expressed as mean (SD), unless otherwise indicated.
The 24-hour BP monitoring values from 4 women in the hypertensive group
and 3 in the normotensive group did not pass the inclusion criteria, and the
data relative to these subjects were not included in the analysis. On average,
91.8% of readings were valid (range, 75%-100%). Among the 144 remaining women
with clinic hypertension, 42 (29.2%) were classified as having WCH according
to the criteria defined in the "Methods" section. The relevant clinical, laboratory,
perinatal, and 24-hour BP monitoring data in the normotensive group, WCH group,
and true hypertensive group are summarized in Table 1 and Table 2.
The indications for cesarean delivery and the treatments received by the patients
are shown in Table 3 and Table 4. In particular, duration of pregnancy
(P=.50) and neonatal weight (P=.68) did not differ between the normotensive and WCH group.
Prevalence of pregnancy-induced hypertension (ie, transient late gestational
hypertension and preeclampsia combined) was greater in the WCH than in the
normotensive group (19.1% [8/42] vs 12.6% [13/103]; P=.02),
but lower than in the true hypertensive group (95% [97/102]; P<.001).
The incidence of preeclampsia alone did not differ significantly between
the WCH and the normotensive groups (7.1% [3/42] vs 5.8% [13/103]; P=.86), while it was considerably higher in the true hypertensive group
(61.7% [63/102]; P<.001).
At entry into the study, none of the women in the normotensive and WCH
group had 24-hour proteinuria greater than 300 mg; whereas in the true hypertensive
group, 34 women had 24-hour proteinuria greater than 300 mg. In the true hypertensive
group, among the 68 women who did not have significant proteinuria at entry,
29 developed preeclampsia later in the course of pregnancy. Three women in
the WCH group and 6 in the normotensive group whose proteinuria values were
less than 300 mg/24 h at entry later developed preeclampsia.
Sensitivity and specificity23 of office
hypertension, 24-hour BP monitoring, and 24-hour proteinuria values for diagnosis
of pregnancy-induced hypertension and preeclampsia alone are shown in Table 5.
At the follow-up visit 1 month after delivery, none of the women in
the normotensive group had office BP greater than 140 mm Hg systolic or 90
mm Hg diastolic, compared with 3 women in the WCH group (P=.04) and 26 women in the true hypertensive group (P<.001). Five women in the true hypertensive group still had 24-hour
urine protein values greater than 300 mg 1 month after delivery.
We evaluated separately the perinatal data for the women who gave birth
by cesarean delivery and those who delivered vaginally (Table 2). No significant differences were observed between the normotensive
group and the WCH group for Apgar score at 1 minute (cesarean delivery, 9.5
[1.2] vs 9.2 [2.4]; P=.24; vaginal delivery, 9.9
(0.5) vs 9.8 [0.8]; P=.76) and maternal hospital
stay (cesarean birth, 7.4 [1.4] vs 7.8 [1.1] days; P=.16;
vaginal delivery, 4.2 [0.9] vs 4.2 [0.8] days; P=.85).
One maternal and 1 neonatal death occurred in the hypertensive group,
1 stillbirth occurred in the WCH group (attributed to umbilical cord strangulation),
and 1 stillbirth in the normotensive group (due to placental insufficiency
caused by multiple infarcts). The maternal death occurred suddenly 36 days
after delivery, when the patient was at home, in apparent good health. An
autopsy did not show abnormalities. Death was attributed to "sudden death,
probably due to ventricular arrhythmia." The neonatal death in the hypertensive
group resulted from hyaline membrane disease and sepsis in a premature neonate,
who was born at 28 weeks of gestation and weighed 918 g.
Overall, 24-hour BP monitoring was well tolerated. Four women (1.6%)
were not able to complete the recording session because of excessive discomfort
caused by the procedure, and 74 (29.1%) complained of sleep disturbances caused
by cuff inflation. Five women (2.0%) presented with minor local allergic reactions
secondary to the contact of the skin with the microphone, and in 18 women
(7.1%) a slight swelling or bruising of the hand and/or forearm was observed.
Prevalence of WCH was 29.2% in our population of women with high BP
measured in an office during their third trimester of pregnancy. This number
is comparable, though slightly higher, with that observed in a nonpregnant
population of similar age.2 To the best of
our knowledge, few data exist on the clinical value of WCH in pregnancy,14,16 and none of these studies provided
a comprehensive assessment of the prognostic impact of this condition.
The only 2 available studies found a higher prevalence of WCH than ours.
However, only a subset of the pregnant women actually underwent 24-hour BP
monitoring in 1 study,14 while in the other
study,16 the dividing line between WCH and
true hypertension was set to 85 mm Hg diastolic for average 24-hour ambulatory
blood pressure, regardless of systolic BP, with a reported prevalence of WCH
greater than 60%. The criteria used in our study for diagnosis of WCH are
more restrictive, in line with the criteria currently used in nonpregnant
We studied hospitalized patients and used normal reference values derived
from hospitalized subjects: these values appear to be lower than those found
in other studies.26- 28
However, we derived the cutoff point for true hypertension from a population
evaluated in a hospital setting, and a 1996 study29
showed that 24-hour BP monitoring values in hospitalized subjects tend to
be slightly lower than in outpatients, probably due to a reduction of environmental
stress and physical activity. In the other studies,26- 28
BP was recorded every 30 minutes by means of an oscillometric method. We used
15-minute recording intervals and an auscultatory method. Moreover, in 2 of
the studies,27,28 the average
24-hour systolic and diastolic BP was higher than the office BP, and in 1
study,28 the pressurometer was shown to systematically
overestimate both systolic and diastolic BP. According to Clark et al,19 this is avoided when an auscultatory method is used.
Finally, 2 of the studies26,27
included women who developed pregnancy-induced hypertension, and 1 study26 included only primigravidas.
Our study shows that office systolic and diastolic BP were significantly
higher in the WCH group than in the normotensive group, although their 24-hour
BP values were almost identical. The women with true hypertension had 24-hour
BP significantly higher than the women in either the WCH or normotensive group.
Office BP in the true hypertensive group was considerably higher than in the
normotensive women, but only marginally increased in comparison with the women
in the WCH group.
Laboratory testing showed that the women in the true hypertensive group
had significantly higher 24-hour urine protein values, higher serum creatinine,
and uric acid values, and a lower platelet count compared with the other 2
groups. Such an observation is consistent with the greater proportion of women
who developed preeclampsia in the group with true hypertension. Of note, none
of these parameters differed between the normotensive and the WCH groups;
these 2 groups appeared to differ only in their office BP.
For outcome of pregnancy, the incidence of preeclampsia or eclampsia
was much greater in the women with true hypertension, whereas no differences
were noted between the WCH and normotensive groups.
Analysis of perinatal data (Table
2) showed that the duration of pregnancy and neonatal and placental
weights did not differ between the normotensive and WCH groups. However, the
women from the latter group had a higher incidence of cesarean delivery, with
a slightly lower Apgar score at 1 minute and a longer hospital stay for both
the mother and the newborn, although not as long as in the true hypertension
group. These observations indicate that WCH has an intermediate prognostic
impact on some variables, between that of normotension and true hypertension.
However, subgroup analysis (Table 2)
showed that the differences between the normotensive and WCH groups are mostly
explained by the different rate of cesarean deliveries. In fact, when the
data from the women who underwent cesarean delivery and those who delivered
vaginally are evaluated separately, the differences between the 2 groups disappear.
It is difficult to provide a definite explanation for the higher incidence
of cesarean delivery in the women with WCH compared with the normotensive
women. However, it is possible that the attitude of the physicians caring
for the women with increased office BP in their third trimester of pregnancy
may have played a role. Such speculation is supported by the lack of biochemical
or clinical abnormalities in the women in the WCH group. Moreover, many women
with WCH were treated with antihypertensive agents, including nifedipine,
which may depress uterine contractility and is also used as a tocolytic agent.30 It is worth noting that 6 women in this group underwent
cesarean delivery because of dystocia due to uterine inertia. The lower Apgar
score at 1 minute and the longer hospital stay of women in the WCH group may
have been direct consequences of an increased tendency to perform cesarean
delivery in these women, as suggested by the lack of differences between the
normotensive group and the WCH group when the 2 subsets with vaginal delivery
and cesarean delivery are analyzed separately.
Sensitivity of 24-hour BP for the diagnosis of transient late gestational
hypertension and preeclampsia was marginally lower than that of office hypertension,
but its specificity was considerably higher. Such a finding is in accordance
with the data provided by 2 recent studies,31,32
which show that 24-hour BP is superior to office BP in predicting occurrence
of severe hypertension in pregnancy. Furthermore, a recent study33
showed that fetal growth is inversely associated with 24-hour BP but not with
office BP. A study by Hermida and Ayala34 showed
sensitivity and specificity values lower than those we observed for diagnosis
of pregnancy-induced hypertension. However, in that study BP was measured
with an oscillometric method every 30 minutes during daytime and every 60
minutes during the night, 24-hour BP parameters were derived by mathematical
modeling rather than actual BP readings, and sensitivity and specificity were
computed separately for systolic and diastolic BP. The sensitivity of 24-hour
urine protein values for the diagnosis of pregnancy-induced hypertension was
low in our study (less than 50%), with, however, a 100% specificity. This
latter figure is probably overestimated, as we excluded women with known nephropathy
In summary, the results of this study show that 29.2% of women with
high office BP in the third trimester of pregnancy have WCH and that this
condition carries a prognostic impact that, as far as the outcome of pregnancy
is concerned, differs from the normotensive women only for a greater probability
of undergoing cesarean delivery, a slightly lower Apgar score at 1 minute,
and a higher incidence of transient late gestational hypertension. Taken together,
these findings indicate that 24-hour BP is superior to office BP for prediction
of the outcome of pregnancy. Twenty-four hour BP monitoring may be considered
a valuable diagnostic procedure in pregnant women with hypertension, particularly
in uncomplicated cases in which a therapeutic decision appears to be most
difficult. On the basis of our data, it is also possible to speculate that
a significant proportion of women diagnosed as having transient gestational
hypertension may actually have WCH. The higher incidence of cesarean delivery
in the group of women with WCH probably reflects decision-making processes
based on office BP. These data suggest the utility of a randomized study of
the effects of therapeutic decisions based on office BP vs 24-hour BP in women
with uncomplicated hypertension during their third trimester of pregnancy.