Key content
 Accurate blood pressure (BP) measurement is fundamental to early
diagnosis of hypertensive disorders in pregnancy.
 Poor auscultatory technique and lack of training leads to
inaccuracies in BP measurement using sphygmomanometry with
mercury and aneroid devices.
 Automated devices limit user error but require validation of
accuracy because they tend to underestimate BP
in pre-eclampsia.
 Systolic hypertension may better predict risk of adverse outcome
(such as haemorrhagic stroke) than diastolic hypertension.
 Ambulatory/self-monitoring increases the number of
representative readings available on which to base management,
limiting unnecessary intervention.
 Detection of hypotension in pregnancy is crucial to the diagnosis
of shock secondary to haemorrhage and sepsis.
Learning objectives
 To learn how to obtain accurate BP measurements in pregnancy.
 To understand the significance of hypertension in pregnancy.
Ethical issues
 A large proportion of maternal deaths are associated with
substandard care, often related to poor recognition of severity of
hypertension or shock and need for treatment.
 Lack of cheap, accurate, easy-to-use BP devices in low- and
middle-income countries, where risk of maternal and perinatal
mortality and morbidity secondary to pre-eclampsia and shock is
highest, continues to be a challenge.
Keywords: blood pressure measurement / diagnosis /
hypertension / pre-eclampsia

Introduction
Hypertensive disorders in pregnancy, which include preeclampsia,
gestational hypertension and chronic
hypertension, complicate 2–8% of pregnancies and confer
risk to the health of mother and fetus.1 Pre-eclampsia is one
of the three leading causes of maternal death in the UK and
can result in substantial maternal morbidity, including
intracranial haemorrhage, HELLP (haemolysis, elevated
liver enzymes and low platelet count) syndrome and
disseminated intravascular coagulation.2 In 2010 an
estimated 287 000 maternal deaths occurred globally, 99%
of which occurred in low- and middle-income countries
(LMICs). Approximately 14% of these deaths were thought
to be related to hypertensive disorders in pregnancy,3
although this figure may be higher when the contribution
of hypertensive disorders in pregnancy to other causes of
mortality (such as postpartum haemorrhage) is considered.
Pre-eclampsia also contributes to one-fifth of all preterm
births globally (and is the leading cause of iatrogenic preterm
birth) and one-quarter of stillbirths and neonatal deaths
in LMICs.1
Accurate measurement of blood pressure (BP) is crucial to
the diagnosis and management of hypertensive disorders in
pregnancy. BP monitoring is the most important, and
frequent, screening test in the antenatal period and is
undertaken by healthcare assistants, midwives, general
practitioners and obstetricians on a daily basis. Accuracy of
BP measurement impacts on maternal and perinatal clinical
outcomes, highlighted in the 2006–2008 UK Confidential

Enquiries into Maternal Deaths report,4 which found that the
most common reason for substandard care in deaths
secondary to pre-eclampsia/eclampsia was failure to
recognise and treat hypertension.
As well as detecting hypertension, BP devices are also
essential for the detection of acute haemodynamic
compromise and management of shock in pregnancy.
Obstetric haemorrhage, pregnancy-related sepsis and unsafe
termination of pregnancy are major contributors to
approximately 46% of maternal deaths worldwide and all
can present with signs and symptoms of shock.3 According to
the Confidential Enquiries report,4 the majority of maternal
deaths could have been avoided if early warning signs of
impending collapse had been recognised and acted on earlier.
Thus, the ability to measure BP accurately is an indispensible
skill for obstetricians, midwives and other healthcare
workers, regardless of setting, in order to prevent maternal
and perinatal morbidity and mortality, both in the UK
and worldwide.
This review considers the importance of accuracy of
BP measurement, including choice of device, technique
and common errors. It also discusses the evidence-based
hypertension BP thresholds at which there is increased
risk of morbidity and mortality in pregnancy, and the
importance of prompt medical intervention, as well as
the role of vital sign monitoring in women at risk
of shock.
Accuracy of BP measurement in pregnancy
BP measurement is a key part of the assessment of
hypertensive disorders in pregnancy, guiding diagnosis,
admission, antihypertensive treatment and timing of
delivery, as well as the assessment of haemodynamic shock
in pregnancy, secondary to obstetric haemorrhage or sepsis.
It is therefore important that all healthcare providers are
aware of the issues surrounding accuracy of BP measurement
in pregnancy.
Auscultatory technique
The National Institute for Health and Care Excellence
(NICE)5 Antenatal Care guidance recommends BP
measurement at every antenatal visit and outlines the steps
involved in BP measurement using the auscultatory
technique. This includes use of the correct-sized cuff, initial
inflation of the cuff 20–30 mmHg above the palpable systolic
BP, deflation at a rate of 2 mmHg per second, recording BP
to the nearest 2 mmHg and use of Korotkoff phase V to
indicate diastolic BP. For example, deflating the cuff too fast
will result in underestimation of the systolic BP and
overestimation of the diastolic BP. Despite these clear
recommendations, in clinical practice BP measurement is
often not performed correctly, leading to inaccurate readings,
because of inadequate training and equipment, time
constraints or lack of awareness of the importance of BP
monitoring as a screening and diagnostic test.
Korotkoff IV versus V
Up until the late 1990s there was debate as to whether
Korotkoff phase IV (muffling of sound, K4) or Korotkoff
phase V (disappearance of sound, K5) should be used to
classify diastolic BP in pregnancy. It was argued that K4 was
more appropriate considering the unique haemodynamics of
pregnancy and because it was thought that K5 could often
extend to or near zero (since shown to be very rare).6 A
randomised controlled trial published in The Lancet in 1998,7
comparing outcome in hypertensive disorders in pregnancy
managed according to either K4 or K5, demonstrated that an
episode of severe hypertension was more likely in women in
the K4 group, mainly because diastolic hypertension was
more likely to be recorded. However, the frequency of severe
systolic hypertension, simultaneous systolic and diastolic
hypertension, and maternal and fetal adverse clinical
outcome did not differ between the two groups.
Considering these findings and that K5 is better reflective
of intra-arterial pressure and is far more reproducible,8 the
use of K5 to classify diastolic BP was recommended and has
since been included in the NICE5 Antenatal Care guidelines.
Sources of error associated with auscultatory
technique
BP measurement using the auscultatory technique relies on
accurate transmission and interpretation of Korotkoff
sounds. Although auscultation using mercury
sphygmomanometry has been the gold standard of BP
measurement in the past, clinicians are gradually moving
away from such sphygmomanometers because of health and
environmental concerns regarding the use of mercury in
clinical settings. As of 2014, mercury sphygmomanometers
are banned and can no longer be purchased in Europe
because of environmental concerns.
Some non-automated BP devices require regular
calibration to ensure a leak rate (loss of air pressure)
within 4 mmHg/minute and a pressure scale accurate to
within 3 mmHg for any part of the pressure range.9
Observational studies assessing the calibration of BP devices
used in clinical practice demonstrated that 20–25% of devices
used in hospital and clinic settings had unacceptable
calibration errors.9,10 Despite recommendations to record
BP to the nearest 2 mmHg, a questionnaire-based study
reported that only 10% of midwives and obstetricians
recorded BP to the nearest 2 mmHg, with 23% recording
BP to the nearest 10 mmHg.11 A study assessing BP values in
women seen at antenatal clinic showed that 78% of readings
obtained by clinicians ended in a zero.12 This user preference
to round off BP values to a zero or five is referred to as terminal digit preference and is a source of error associated
with auscultation. Observer bias refers to the user adjusting
the BP reading to what is preferred or what it was
preconceived to be. This concept can extend to threshold
avoidance, where the observer adjusts the BP reading to avoid
thresholds that entail making a diagnosis or requiring
intervention. Again, this is a source of error more
commonly associated with auscultation.
Aneroid devices
Aneroid devices remove the need to use mercury in clinical
settings, but the inherent errors associated with
sphygmomanometry and the use of Korotkoff sounds with
auscultation remain. Furthermore, aneroid devices require more
frequent maintenance and calibration than mercury
sphygmomanometers. A survey of BP devices used by UK
general practitioners showed that only 50% of devices had been
serviced within 1 year and 24% had never been serviced,13
increasing the chance of error. Likewise, an observational study of
devices used in UK general practices demonstrated that 53% of
aneroid devices were reading in error by more than 3 mmHg,
far more than the mercury and automated devices.14 Although a
seemingly small difference from the true reading, a systematic
underestimation of BP by 3 mmHg would lead to one-quarter of
patients with hypertension being falsely classified as
normotensive.15 As long as aneroid devices are regularly
maintained and calibrated, their accuracy can be assumed to be
similar to mercury devices, as shown in an observational study
evaluating the accuracy of aneroid devices used clinically,
compared with a calibrated mercury sphygmomanometer.16
Oscillometry: an alternative to the auscultatory
technique
The auscultatory technique for measuring BP requires skill
and training, and is therefore prone to observer error. In
recent years, there has been a shift towards the use of
automated BP devices, which rely on detecting changes in the
amplitude of the intra-arterial oscillometric waveforms
produced during cuff deflation to determine BP. The Royal
College of Obstetricians and Gynaecologists (RCOG), the
British Hypertension Society and a number of international
organisations have recommended that automated devices are
independently validated according to a recognised protocol
to ensure accuracy (for example, the Association for the
Advancement of Medical Instrumentation criteria, British
Hypertension Society and International protocols).17,18
Despite this recommendation, of the hundreds of
commercially available automated devices, only a small
number have been evaluated and even fewer have
passed validation.
The issue of accuracy is even more important for those
devices used in pregnancy. The majority of devices validated
specifically in pregnant populations fail the protocol
requirements, likely because of the haemodynamic changes
of pregnancy. A particular concern is that automated devices
tend to underestimate BP in women with pre-eclampsia,
resulting in false classification of normotension in these
high-risk women.19 This is thought to be due to specific
pathological changes of pre-eclampsia, including decreased
arterial vascular compliance and increased interstitial
oedema, which may affect the amplitude and detection of
the oscillometric waveform.20 Separate validation in
pregnancy (including pre-eclampsia) is therefore
recommended but only a small number of devices have
passed validation for use in pregnancy, which includes
pre-eclampsia (Box 1).
For those automated devices that are validated for use in
pregnancy (including pre-eclampsia), the obvious benefit
over auscultation (mercury sphygmomanometry and
aneroid) is that inaccuracies secondary to observer error
are limited and BP measurement is simpler. In LMICs,
community healthcare providers caring for women
antenatally may have had limited training in BP
measurement. In these settings, automated devices may be
a more practical alternative. However, there are issues
regarding cost, powering and maintenance of these devices.
Appropriate BP cuff size
To ensure accuracy it is important to consider the size of BP
cuff used, particularly as obesity is a risk factor for
pre-eclampsia and therefore those with a large arm
circumference (33 cm or above) are at higher risk of
developing pre-eclampsia. In a clinical setting large cuffs
are often less readily available than standard cuffs. If the
appropriately sized cuff is not available, that is, if a standard
cuff is used on a woman with an arm circumference of more
than 32 cm, BP can be overestimated.21 Conversely, if a large
cuff is used on a woman with a normal arm circumference,
BP can be underestimated, although this error is much less.22
It is therefore important that a variety of cuffs are available in

the clinical setting and that arm circumference is correctly
estimated or measured. If in doubt, overcuffing (using a cuff
that is too large for the arm circumference) is better than
undercuffing (using a cuff that is too small for the
arm circumference).
Evidence-based hypertension thresholds
Obstetricians use BP values to guide management in women
with hypertensive disorders in pregnancy. If BP is
underestimated or overestimated through inaccurate
measurement, avoidable maternal and perinatal mortality
and morbidity can result. Furthermore, obstetricians use BP
thresholds recommended by national guidelines to aid in
management decision making. If these thresholds are not
supported by adequate evidence, maternal and perinatal
mortality and morbidity can again result. The NICE23
guidelines on hypertension in pregnancy define mild
hypertension as diastolic BP of 90–99 mmHg and/or
systolic BP 140–149 mmHg, moderate hypertension as
diastolic BP 100–109 mmHg and/or systolic BP
150–159 mmHg, and severe hypertension as diastolic BP of
110 mmHg or above and/or systolic BP of 160 mmHg or
above. Although these definitions are concise and widely
adopted, the BP thresholds that indicate the need for
treatment are less clearly defined.
BP thresholds for treating severe hypertension
For women with severe hypertension, there is consensus that
antihypertensive treatment should be given to reduce the risk
of maternal central nervous system complications, but the
specific thresholds for initiating treatment are based on
limited evidence.
A landmark (yet small) retrospective cohort study of 28
women who sustained strokes in association with severe
pre-eclampsia or eclampsia showed that all women had a
systolic BP of more than 155 mmHg immediately before the
stroke, whereas only 13% had a diastolic BP of at least
110 mmHg.24 A retrospective cohort study of women with
eclampsia showed that posterior reversible encephalopathy
syndrome, defined as the presence of neurological symptoms
and signs, together with radiological findings of vasogenic
cerebral oedema, occurred at lower systolic BP levels in
pregnancy (mean peak systolic BP of 173 mmHg), compared
with non-pregnant patients with hypertensive encephalopathy
(mean peak systolic BP of 191 mmHg).25 These studies
highlight the importance of prioritising the control of
systolic BP over diastolic BP, and support the 2010 NICE
guidelines recommending immediate treatment if systolic BP
is 150 mmHg or above.23 The 2010 NICE guidelines also
recommend immediate treatment if diastolic BP is at least
100 mmHg, although this is largely based on extrapolation of
risk, rather than direct evidence.23 However, a 2012 nested
case–control study investigating potential factors associated
with antenatal stroke demonstrated that after adjustment for
age, stroke increased by 3% (adjusted OR 1.03, 95% CI 1.00–
1.05) for every mmHg increase in highest recorded systolic BP,
compared with 8% (adjusted OR 1.08, 95% CI 1.03–1.13) for
every mmHg increase in diastolic BP.26
BP thresholds for treating mild-to-moderate
hypertension
There is a lack of consensus among the obstetric community
regarding the threshold of treatment for mild-to-moderate
hypertension in pregnancy. A 2014 update of the Cochrane
review assessing the effects of antihypertensive treatment of
mild-to-moderate hypertension during pregnancy (including
those with a diagnosis of pre-eclampsia)27 demonstrated a
reduction in the number of women developing severe
hypertension or requiring a second treatment agent.
However, well-powered studies demonstrating reductions in
adverse clinical outcome (maternal stroke, progressive renal
and other end-organ disease, and heart failure) following
antihypertensive treatment in these cases are limited.
Similarly, the review failed to demonstrate a significant
effect on preterm births or caesarean sections in those given
antihypertensives. The review concluded that it remains
unclear whether treatment of mild-to-moderate hypertension
is worthwhile. Current NICE/RCOG guidelines therefore do
not recommend treating hypertension of systolic BP of
140–149 mmHg or diastolic BP of 90–99 mmHg, except in
those women with target-organ damage secondary to
chronic hypertension.23
BP thresholds for treating postpartum hypertension
A 2005 Cochrane review on the management of postpartum
hypertension28 demonstrated insufficient evidence to form
robust recommendations on the thresholds for treatment.
The NICE guidance23 recommends initiating treatment at the
same thresholds as for during the antenatal and intrapartum
period; this is largely based on expert opinion. A 2013 clinical
review29 summarises the evidence for treatment options for
postpartum hypertension and provides a flow diagram with
management pathways.
No role for isolated incremental rise in BP
The use of an isolated incremental rise in BP to define
hypertension in pregnancy is now not recommended in the
guidelines.23 Research has demonstrated that women with an
incremental rise (for example, 30 mmHg systolic BP/
15 mmHg) from booking whose BP remained under the
threshold of 140/90 mmHg had normal pregnancy outcomes.30
Future research to improve evidence base
In response to a call for more robust evidence, an
international multicentre randomised controlled trial on Shock Index as a marker of haemodynamic
instability
Shock Index, the ratio of pulse to systolic BP, has been
proposed in a nonpregnant and, more recently, a pregnant
population, as an earlier marker of compromise than
conventional vital signs. A systematic review has
demonstrated that Shock Index was an accurate indicator
of decompensation in both non-pregnant individuals and
pregnant women (with AUROCs ranging 0.77–0.84,
compared with 0.56–0.74 for pulse and 0.56–0.79 for
systolic BP).46 However, to date, only two retrospective
studies have attempted to define the normal range of Shock
Index in women with obstetric haemorrhage according to
risk of adverse clinical outcome. Both studies concluded that
the upper limit of normal Shock Index was 0.9,49,50 and one
of the studies also suggested a second threshold of at least 1.7
indicating the need for urgent attention.50 The use of Shock
Index in the assessment of women with shock secondary to
obstetric haemorrhage or sepsis may help more promptly
identify those at risk of avoidable mortality and morbidity,
but more studies in this area are required.
Conclusion
Hypertensive disorders in pregnancy, obstetric haemorrhage,
sepsis and unsafe termination of pregnancy contribute to
more than half of all maternal deaths globally. The diagnosis
and management of each of these conditions is guided, in
part, by the measurement of BP. It is therefore important
that clinicians caring for women with these conditions are
able to measure BP correctly and to appreciate the
importance of accurate measurement, and the significance
of the measurement which is often very different from
non-pregnant patients. Healthcare professionals should be
aware of the advantages and disadvantages of the various BP
devices available and feel confident to raise concern
regarding devices inaccurate for use in pregnancy or poor
technique observed.
In the absence of high-grade evidence, the national
guidelines for hypertensive disorders in pregnancy on
intervention according to BP thresholds are largely based
on expert opinion, particularly regarding mild-to-moderate
hypertension. For severe hypertension, there are a small
number of studies suggesting that severe systolic
hypertension is a better indicator than diastolic
hypertension of the risk of adverse cerebrovascular events.
Further well-designed clinical studies and trials are required
to evaluate the optimal threshold for intervention in women
with different types of hypertension in pregnancy, with
assessment of the efficacy of varying antihypertensive drug
classes. Likewise, the evidence regarding the use of vital signs
in the diagnosis and management of maternal shock is
limited. The impact of recognition of haemodynamic
compromise on mortality and significant morbidity was
highlighted in the 2006–2008 Confidential Enquiries report.4
Research should now focus on determining the optimal vital
sign predictor(s) and thresholds of this predictor to guide
assessment in obstetric shock.