Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254

Contents lists available at ScienceDirect

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier.com/locate/siny

Review

Hemodynamic monitoring of the critically ill neonate: An eye on the

future
Timur Azhibekov a, Sadaf Soleymani a, Ben H. Lee b, Shahab Noori a, Istvan Seri b, *
a
Division of Neonatology and the Center for Fetal and Neonatal Medicine, Department of Pediatrics, Children's Hospital Los Angeles and the LACþUSC
Medical Center, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
b
Sidra Medical and Research Center, Doha, Qatar

s u m m a r y
Keywords: By continuous assessment of dynamic changes in systemic and regional perfusion during transition to
Neonate extrauterine life and beyond, comprehensive neonatal hemodynamic monitoring creates numerous
Hemodynamics
opportunities for both clinical and research applications. In particular, it has the potential of providing
Cerebral autoregulation
Oxygen delivery
additional details about physiologic interactions among the key hemodynamic factors regulating sys-
Organ perfusion temic blood flow and blood flow distribution along with the subtle changes that are frequently transient
Genetic variability in nature and would not be detected without such systems in place. The data can then be applied for
predictive mathematical modeling and validation of physiologically realistic computer models aiming to
identify patient subgroups at higher risk for adverse outcomes and/or predicting the response to a
particular perturbation or therapeutic intervention. Another emerging application that opens an entirely
new era in hemodynamic research is the use of the physiometric data obtained by the monitoring and
data acquisition systems in conjunction with genomic information.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction criteria for initiating interventions [14] using various approaches

Neonatology continues to evolve rapidly along with advances in
other subspecialties, biomedical research and cutting edge
biomedical technology. However, our understanding of neonatal
hemodynamics and, as importantly, when and how to intervene
when deviations from the physiologic course occur remain quite
limited. In this context, the old story of six blind men and an
elephant comes to mind [1] (Fig. 1). Probably, each of us has come
across this parable in one form or another and the lesson it is meant
to teach can be applied to many areas of our lives. Accordingly, it
very well describes the limitations of our understanding of
neonatal cardiovascular physiology and pathophysiology, with vast
differences in opinions and ongoing debates about how to
approach the numerous challenges that the neonatologist faces
when providing care for neonates with hemodynamic compromise.
Severe cardiovascular compromise in the neonatal period is
associated with increased morbidity and mortality [2e6]. Despite
many attempts to define neonatal hypotension [7e13] and develop
* Corresponding author. Address: Sidra Medical and Research Center, PO Box
26999, Doha, Qatar. Tel.: þ974 4404 1995.
E-mail address: iseri@sidra.org (I. Seri).
and vasoactive medications [15e18], these and other related
questions remain mostly unanswered. Importantly, clinical trials on
treatment of neonatal hypotension have so far failed to demon-
strate improvement in clinically relevant long-term outcomes.
Some of the major underlying challenges include the hetero-
geneity of the neonatal population due to differences in gestational
age (GA) and postnatal age, variations in maturity for a given GA,
existing comorbidities (particularly lung disease and infection),
complex and multifactorial interactions between systemic and
regional perfusion [19] and underlying genetic heterogeneity. In
addition, conventional hemodynamic parameters [heart rate (HR),
arterial oxygen saturation (SpO2), and blood pressure (BP)], even if
monitored continuously, have significant limitations for accurate
assessment of both the hemodynamic status and the response to
interventions used for the treatment of the circulatory compromise.
Recognition of shock in its early, compensated phase along with
identification of the underlying pathophysiology is of paramount
importance in reversing shock before its progression to uncom-
pensated cardiovascular failure when the obvious clinical mani-
festations of shock become apparent. This requires, among others,
the use of comprehensive hemodynamic monitoring systems
capable of continuous and simultaneous acquisition of multiple

http://dx.doi.org/10.1016/j.siny.2015.03.003
1744-165X/© 2015 Elsevier Ltd. All rights reserved.
 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254
Fig. 1. “We have to remember that what we observe is not nature in itself, but nature exposed to our method of questioning” ‒ Werner Heisenberg, 1958 [1].
hemodynamic parameters that reflect both systemic and regional
blood flow and oxygen delivery [20e22].
2. Comprehensive hemodynamic monitoring
With advances in biomedical technology and computer science,
the ability of hemodynamic monitoring systems to collect, store,
and analyze the complex physiometric data provides a foundation
for advances in diagnosis and management of neonatal cardio-
vascular compromise. However, it must be emphasized that this
system also has a number of caveats concerning its accuracy,
reliability, feasibility, and the need for validation across different
subpopulations [21,22]. The utility power of such monitoring
systems lies in the comprehensiveness of the hemodynamic pa-
rameters monitored beyond the conventionally obtained param-
eters of HR, respiratory rate (RR), SpO2, BP, and transcutaneous
CO2 measurements. For blood pressure monitoring, invasive
continuous monitoring via umbilical arterial catheter (UAC) or
peripheral arterial line remains the preferred method [23,24]. The
macrocirculatory component of systemic perfusion can be
assessed by measurement of cardiac output (CO). Whereas func-
tional magnetic resonance imaging (fMRI) has become the “gold
standard,” its feasibility even for research purposes remains
limited. Bedside functional echocardiography (fECHO) and
impedance electrical cardiometry (IEC) have been used at the
bedside and these methods provide comparable CO measurements
[25,26]. Impedance electrical cardiometry has the additional
advantage of the capability of continuously assessing CO. Whereas
newer methods of non-invasive cardiac output monitoring are
emerging [27,28], further validation of IEC and validation of the
newer methods in the neonatal population are needed, particu-
larly in extremely preterm neonates. For assessment of peripheral
perfusion and microcirculation, a number of methods are
currently available, such as laser Doppler flowmetry, visible light
technology, orthogonal polarization (OPS) and side-stream dark-
field (SDF) imaging. Inclusion of perfusion index (PI) [29e32] as
additional parameter derived by pulse oximetry into the hemo-
dynamic monitoring system can potentially also improve the
247
assessment of peripheral perfusion, although such use in neonates
remains to be robustly validated. To evaluate regional organ blood
flow, near-infrared spectroscopy (NIRS) is a useful tool for non-
invasive and continuous assessment of tissue oxygen saturation.
Tissue oxygenation index (TOI), with caution, can be used as a
surrogate of organ blood flow [20,33]. For more in-depth overview
of the methods used for monitoring of the different cardiovascular
parameters and the corresponding references in the literature, we
encourage the reader to read previously published reviews on the
topic [20e22].
Beyond monitoring blood flow and oxygen delivery to the tis-
sues, simultaneous assessment of functional activity of target or-
gans is an additional important step to enhance the diagnostic
power of comprehensive hemodynamic monitoring systems.
Amplitude-integrated electroencephalography (aEEG) has been
studied primarily in term neonates with hypoxic‒ischemic en-
cephalopathy [34e36]. However, studies have recently also been
published on its use in preterm neonates [37,38] along with studies
on the correlation of brain activity patterns with NIRS data for
diagnostic and prognostic purposes [39,40]. These latter findings
lend support to the idea of integrating aEEG into hemodynamic
monitoring and data acquisition systems.
A hemodynamic monitoring “tower”, such as that designed and
described by the authors [21,22], allows for practical continuous
and simultaneous bedside monitoring and acquisition of the syn-
chronized physiologic data at high sampling rates in real time. It
also has the advantage of being a mobile, stand-alone unit that can
be utilized at the patient's bedside. Collected data are subsequently
analyzed to study minute-to-minute changes and interactions be-
tween multiple hemodynamic parameters.
For research applications on a larger scale, the output data from
bedside monitors and other devices (e.g. infusion pumps, ventila-
tors) from multiple patients can be acquired and routed to central
servers using hospital data networks. Preprogrammed software for
hospital-wide systems is available from third-party vendors [e.g.,
Bernoulli Enterprise (Cardiopulmonary Corporation, CT, USA) or
BedMaster (Excel Medical Electronics, FL, USA)]. Of note, their
implementation, in addition to financial cost, is a challenging and
248 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254

time-consuming process that requires close collaboration between
the vendor, biomedical engineering and information technology
departments, hospital administration, and healthcare providers
from different disciplines (e.g. physicians, respiratory therapists,
and nurses). It also requires additional infrastructure to store and
organize an enormous amount of continuously growing data and
the involvement of computer scientists and experts in bioinfor-
matics for large data handling and analysis. Once established,
however, accurately collected and appropriately stored data from
these multimodular monitoring systems comprise an invaluable
resource for multiple research ideas and hypotheses, quality
improvement activities that extend far beyond cardiovascular
physiology alone.
One such system is undergoing the final steps of testing and
validation at one of the author's institutions at Children's Hospital
Los Angeles, with the ability to capture data from bedside monitors,
ventilators, NIRS, IEC, and other compatible devices (Fig. 2). There
are certain challenges inherent in automated collection of physio-
logic datasets, not the least of which is the intermittent absence of
data and the presence of artifacts. Additionally, the crucial task of
feasible, reliable and accurate documentation of therapeutic in-
terventions and the adjustments in the dosages of vasoactive and
other medications synchronously time-stamped with the moni-
tored physiologic data remains a significant challenge. Newer
infusion pumps that have the capability of transmitting data
regarding medication and dose changes to a central data server
offer potential solutions to this problem.
Collected physiologic data can be retrieved either as a data
spreadsheet or as waveforms when applicable. Various software
packages then can be utilized for further processing of these data to
view, score, transform, or analyze parameters of interest. The
output format also offers flexibility allowing analysis of various
combinations of parameters, such as cerebral regional tissue oxy-
gen saturation (CrSO2) and BP to assess cerebral autoregulation,
rSO2 and SpO2 to estimate fractional tissue oxygen extraction
(FTOE).
It is important to remember that, to guard the patient's rights
and confidentiality, stored data represent protected health infor-
mation and should be accessed for research and quality improve-
ment purposes in accordance with federal, state and hospital-wide
regulations, policies and procedures.
3. Advantages of comprehensive hemodynamic monitoring
A growing number of investigators combine data from multiple
monitoring tools to increase the diagnostic and prognostic value of
these tools. In addition, such approaches can provide important
insights into the underlying pathophysiology that would not be
possible from the use of a single method [41e43]. Significant lim-
itations, however, lie in the fact that data acquisition is typically

Fig. 2. Real-time continuous recording of hemodynamic parameters during a 12 h period in a neonate weighing 825 g at 25 weeks' gestation. This 10-day-old extremely premature
neonate presented with respiratory failure, hemodynamically significant patent ductus arteriosus, hypotension, bilateral grade 2 intraventricular hemorrhage, stage 2A necrotizing
enterocolitis, anemia, and hyperglycemia. Treatment of hypotension was initiated with dopamine infusion at 5 mg/kg/min (start of infusion not shown). At 24 h the dose was
increased to 8 mg/kg/min (bottom channel) with gradual increase in arterial blood pressure. No apparent increase in CO immediately following dose increase was observed (no effect
on HR and/or stroke volume), suggesting that an increase in SVR via alpha-receptor activation was responsible for the gradual increase in BP. RrSO2 increased toward the normal
range along with the increase in BP, while CrSO2 remained unchanged. During titration of dopamine infusion over the following hours, BP continued to remain in the same range.
However, BP fluctuations up to 20 mmHg in magnitude were observed. These episodes did not coincide with the timing of changes in the dose of dopamine. Rather, they appear to
represent spontaneous significant fluctuations in peripheral vascular resistance to the point where SV and thus CO were negatively affected. Heart rate slowly increased over the
course of the monitoring period shown along with an increase in SpO2 and CrSO2. The latter finding suggests that no significant changes occurred in cerebral blood flow (CBF).
Finally, as the swings in BP were not accompanied by similar abrupt increases in CrSO2, and CrSO2, as expected, followed the changes in SpO2 instead, autoregulation of CBF was
likely intact in this 10-day-old patient. Hemodynamic data were collected using Bernoulli system at the sampling rate of 0.2 Hz, and preliminary processing was performed with
software package “MatLab R2013a” (The MathWorks, Inc.). For visual assessment of the trends, we utilized the average value of 60 consecutive samples, which is equivalent to a
5 min interval. The exact time of the adjustments of the dose of dopamine was entered manually after data collection was complete, based on the nursing documentation. CO,
cardiac output; SV, stroke volume; HR, heart rate; BP, blood pressure; TCOM, transcutaneous pCO2 measurements; SpO2, arterial oxygen saturation; CrSO2, cerebral tissue oxygen
saturation; RSO2, renal tissue oxygen saturation; SVR, systemic vascular resistance.
 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254
intermittent, being performed at specified time intervals. The time
intervals are determined by the investigator and the properties of
the monitoring devices. Although efforts may be made to base
them on the continuum of key physiologic events or their transi-
tional points, they may be defined simply on technical feasibility of
data collection. While dealing with known and frequently un-
known limitations in accuracy, precision and validity of such
monitoring tools, another layer of assumptions is added by the
possibility of physiologic phenomena that are missed due to their
subtle or transient nature. This increases the risk of data misin-
terpretation and, thus, might lead to misunderstanding or misin-
terpretation of the underlying physiologic mechanisms.
As an example, several studies have investigated the possible
association between the development of peri-intraventricular
hemorrhage (P/IVH) and CrSO2 in preterm neonates. CrSO2, as
well as FTOE, were used as surrogates of cerebral blood flow. Ver-
hagen et al. [44] compared these parameters collected intermit-
tently for periods of 2 h on postnatal days 1, 2, 3, 4, 5, 8, and 15. The
authors found that preterm neonates with both mild and severe
forms of P/IVH had significantly lower CrSO2 and higher FTOE than
controls during the first eight days. They concluded that infants
with P/IVH had persistently decreased cerebral perfusion at least
for the first eight days after birth, regardless of the severity of the
hemorrhage.
Conversely, Alderliesten et al. [45] reported findings of a nested
case‒control study based on a cohort of premature neonates 32
weeks gestation whose physiologic data was collected continuously
over the first 72 postnatal hours. They compared CrSO2 and FTOE
averaged over an approximately 24 h period before and after P/IVH
was detected by cranial ultrasound. They reported significantly
higher mean CrSO2 and lower mean FTOE values during the period
before P/IVH had been diagnosed in neonates with severe P/IVH as
compared to controls, suggesting cerebral hyperperfusion in the
patients before the bleeding occurred. Interestingly, the differences
in mean CrSO2 and FTOE values between the groups were no longer
significant after the severe P/IVH had been diagnosed.
Recently, our group demonstrated a pattern of systemic and
cerebral hemodynamic changes in extremely preterm neonates
affected by P/IVH that suggested a plausible pathophysiologic
explanation to such discrepancies in the findings of the above
studies [46]. Preterm neonates with GA of 27 weeks were fol-
lowed for ~72 h after birth. Evaluations included continuous CrSO2
monitoring and ultrasound assessment of cardiac function, cerebral
hemodynamics and absence or presence of P/IVH monitored every
12 h. Patients who later developed P/IVH had lower CO on the initial
echocardiogram at 4e6 h after birth. The lower CO coincided with
lower CrSO2 and higher cerebral FTOE indicating lower cerebral
blood flow as well. Interestingly, a subsequent increase in CO and
CrSO2 along with a decrease in cerebral FTOE and increases in
arterial carbon dioxide (PaCO2) and middle cerebral artery mean
velocity preceded the detection of P/IVH, indicating reperfusion of
the formally hypoperfused forebrain before P/IVH developed.
Finally, very few changes in the hemodynamic factors monitored
were noted in the group of patients not developing severe P/IVH.
Thus, the above findings suggest a dynamic continuum of he-
modynamic events in affected preterm neonates, starting with
initial systemic hypoperfusion that was, at least, partially related to
cardiac dysfunction and resulted in or at least contributed to ce-
rebral hypoperfusion. As systemic perfusion improved during the
ensuing 20e44 h of postnatal life, it was associated with increases
in cerebral perfusion as well, thus likely predisposing these patients
to the development of P/IVH due to the hypoperfusion‒reperfusion
cycle.
Therefore, despite the seemingly contradictory findings, the
results of the studies by Verhagen et al. [44] and Alderliesten et al.
249
[45] fit the overall hypothesis that the hypoperfusion‒reperfusion
cycle is one of the major pathogenetic factors in the development of
P/IVH along with other factors such as the incomplete high-priority
vascular bed assignment of the forebrain circulation in very pre-
term neonates [20]. It is likely that Verhagen et al. [44] did not
detect the period of increased cerebral perfusion due to the study
design using intermittent data collection. Conversely, Alderliesten
et al. [45] analyzed NIRS data collected over 24 h prior to diagnosis
of P/IVH, thus potentially not capturing some of the hemodynamic
changes that occurred earlier during transition.
Another important observation of the study [46] is that PaCO2
may play a substantial role in the pathogenesis of P/IVH: affected
neonates had increased PaCO2 compared to the unaffected group
prior to P/IVH being detected by cranial ultrasound. However,
PaCO2 levels were not continuously monitored in our study [46],
nor were they monitored in the two other studies [44,45] that
did not find an association between PaCO2 and development of
P/IVH using post-hoc analysis. Meticulously verified continuous
monitoring of PaCO2 initiated shortly after birth is likely to
provide valuable insights into the complex interactions between
the absolute values of and changes in PaCO2 with cerebral blood
flow and other hemodynamic parameters. Indeed, in hemody-
namically stable preterm neonates, it appears that the relation-
ship between PaCO2 levels and cerebral blood flow, assessed
by middle cerebral artery mean velocity changes, evolves and
undergoes “maturation” during the immediate postnatal transi-
tional period [47].
The overall hemodynamic status and the corresponding thera-
peutic interventions are also likely to be important potential con-
founding factors in these relationships. A number of studies
reported evidence of pressure-passive cerebral perfusion [48] and/
or increase in the incidence of P/IVH [45] in hypotensive neonates
that were treated with higher doses of vasopressor-inotropes.
Lastly, the importance of using multiple physiologic parameters
to identify patients at high risk for complications that may require
further evaluation and intervention is supported by the observation
that hypotension alone is not predictive of adverse neuro-
developmental outcomes [49]. Alderliesten et al. showed that
persistence of CrSO2 below 50% for more than 10% of the time
monitored, but not the hypotension itself, was associated with
lower neurodevelopmental scores at 18 months' corrected GA,
regardless of treatment for hypotension, adjustment for the
severity of P/IVH and that of respiratory status [49].
4. Computational modeling, from research to patient care
Decision-making in the neonatal intensive care unit needs to
take place quickly and requires synthesizing large amounts of pa-
tient data. Computational models are mathematical representa-
tions of human anatomy, physiology, and/or pathophysiology. The
computational models aim to assist the clinician and researcher in
improving understanding of the interrelationship between various
parameters or estimating unknown ones, aiding in the diagnosis,
and experimenting with potential interventions and procedures in
the model before introducing them to the patient [50].
Clear understanding of the research question by both the
investigator and the statistician is of critical importance for the
selection of appropriate study design, sample size calculations,
biostatistical analysis methods, and subsequent interpretation of
study findings. At present, it is not uncommon for biostatistical
analyses applied to longitudinal data to fail to adjust for both
within- and between-subject variability. Vigorous study design and
data analyses must also be more consistently driven by biological
plausibility rather than technical mathematical parameters in order
for the study results to be clinically meaningful and to contribute to
250 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254
better understanding of the disease or a pathway with diagnostic
and/or therapeutic implications.
With the implementation of comprehensive hemodynamic
monitoring systems, huge amounts of continuously growing
physiologic data require appropriate organization and storage that,
in turn, utilize a substantial amount of intellectual and technical
resources. These steps, along with subsequent data processing and
analysis, depend on the expertise and close collaboration between
basic science, translational and clinical researchers, scientists from
various specialties, biostatisticians, biomedical engineers, and
computer scientists. Only with such a multidisciplinary approach
can the development of sophisticated methods of data collection
and analysis in neonatal hemodynamic research be possible,
allowing for the identification of pathognomonic trends and pat-
terns that precede pathological changes in systemic and organ
blood flow and thus oxygen delivery. This will lay the foundation
for the development of algorithms that are able to predict both
impeding cardiovascular compromise and the responsiveness of a
given patient to a particular intervention.
As an encouraging example of such efforts, the use of the heart
rate characteristics monitoring system to identify neonates at risk
for sepsis has been validated to decrease mortality from late onset
neonatal sepsis [51,52] (see also B.A. Sullivan and K.D. Fairchild in
this issue). Similar findings have been reported in pediatric and
adult critical care literature of continuous monitoring of dynamic
changes in cardiovascular parameters in response to various factors
[53]. Clinical applications of such functional hemodynamic moni-
toring include prediction of the patient's responsiveness to fluid
administration [54e56], detection of compensated shock [57,58],
and loss of arterial tone determination [59,60].
Another emerging area of neonatal cardiovascular research
with important clinical implications is the use of the hemody-
namic monitoring data for validation of mathematical models
aiming at predicting cardiovascular responses to distinct stimuli.
Fig. 3. Simulation results of cardiovascular and autonomic changes following PDA ligation. Computational modeling of PDA ligation using the modified PNEUMA model. HR, heart
rate; LVO, left ventricular output; RVO, right ventricular output; PDA, patent ductus arteriosus; SBP, MBP and DBP, systolic, mean and diastolic blood pressure, respectively; TPR, total
peripheral vascular resistance.
Such mathematical, or parametric, models have already been
developed and described in the adult literature including the
comprehensive and physiologically realistic computer model
PNEUMA [61], originally developed to study cardiorespiratory
mechanisms of breathing in adults with sleep disorders. The
PNEUMA system incorporates three major systems: cardiovascu-
lar, respiratory and autonomic nervous systems. One of the au-
thors is currently working on adapting this system to neonatal
physiology to model and study the effects of patent ductus arte-
riosus (PDA) on systemic, pulmonary, and organ blood flow,
particularly with regard to its closure, medical or surgical. Fig. 3
demonstrates acute hemodynamic and autonomic changes
following PDA ligation predicted by this modified model [62] with
alteration of the resistive properties of the PDA. Immediately after
PDA closure, while total left ventricular output (LVO) rapidly de-
creases as expected, effective cardiac output (eCO) increases as the
entire LVO is directed to the systemic circulatory bed. This increase
in eCO results in a sudden increase in systemic blood pressure,
triggering a baroreflex cascade with ensuing vagal reduction in
heart rate and sympathetic drive. With the removal of the PDA
shunt, however, pulmonary blood flow also decreases. Following
surgical closure of the PDA, although eCO increases for a very brief
period of time, it then rapidly decreases due to a decrease in the
cardiac preload caused by the closure of the shunt. Although heart
rate returns to pre-ligation levels ~15 min after surgical PDA
closure, it does so with greater variability, associated with elevated
systemic blood pressure, eCO, and cerebral blood flow (data not
shown).
Hemodynamic and respiratory changes associated with PDA
closure have been studied in animal models by Clyman et al. [63].
Interestingly, even a small left-to-right PDA shunt was found to
lead to significant redistribution of regional blood flow, particu-
larly to abdominal organs, with reduction in regional blood flow
resulting from decreased perfusion pressure and/or regional
 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254
vasoconstriction. The decrease in perfusion pressure and the
contribution of vasoconstriction were dependent on the degree of
the ductal shunt and the characteristics of the given organ. For
example, decreased renal blood flow was found to be primarily
due to decreased perfusion pressure in the setting of a moderate
ductal shunt, whereas decreased flow to gastrointestinal tract was
due to localized vasoconstriction, even with small left-to-right
shunts.
Hemodynamic changes in premature infants following surgical
PDA closure have corroborated the findings in animal models.
Weisz et al. [64] studied postoperative CO changes in extremely
premature infants after PDA ligation using echocardiography and a
non-invasive CO-monitoring system based on thoracic bio-
reactance (NICOM) [27]. Based on high versus low LVO at 1 h after
ligation, the authors identified two groups of patients. After an
initial decrease in CO following PDA closure, a gradual increase in
LVO was observed. In addition, NICOM measurements were found
to consistently underestimate stroke volume (SV) and LVO
compared to echocardiography, with increasing disagreement in
the NICOM‒echocardiography congruity over the postoperative
course. Lien et al. reported similar findings in postoperative he-
modynamic changes in extremely premature neonates following
PDA ligation [65]. Using IEC measurements during 10 distinct
periods before and after PDA ligation, they reported an initial
decrease in CO and increase in SVR following ligation. Although
the observed reduction in CO was due almost entirely to a
decrease in SV, there was only a transient decrease in HR at 1 min
after ductal closure. Whereas CO gradually improved by 24 h post
ligation, the postoperative increase in SVR was sustained
throughout the study period. Although this study performed
intermittent rather than continuous hemodynamic assessments
due to the use of oscillometric blood pressure measurements, their
observations are consistent with the findings reported by Weisz
et al. [64]. Both of these studies also highlight the importance of
further efforts to validate non-invasive CO monitoring systems in
neonates and that caution should be used when interpreting the
measurements obtained by the use of the novel technologies
alone.
On a larger scale, hemodynamic physiologic data form an inte-
gral part of the so-called “complex medical dataset” that also in-
cludes physiologic data representing other organ systems, outcome
and healthcare economics data. Several professional networks have
been established [virtual PICU (vPICU), National Patient Centered
Clinical Research Network (PCORnet), various Clinical Research
Networks (CRNs), etc.] to co-ordinate research efforts on the use of
the complex medical data, including machine learning and artificial
intelligence platforms.
5. Genetic variability and phenotypic associations
Advances in molecular genetics and genetic epidemiology pro-
vide increasingly simplified access to whole genome data, affording
the opportunity to explore associations of genetic variability with
phenotypic presentations and opening a new era in hemodynamic
research and understanding of the multilayered nature of cardio-
vascular physiology and pathophysiology. It certainly adds another
level of complexity to this, already complex, and incompletely
understood topic. Advantages of studying genomic associations in
the neonatal population include the unique fact that neonates have
a largely limited and controlled exposure to the environment, per
se, reducing the potential for confounding gene‒environment in-
teractions on studied genomic and phenotypic associations. Addi-
tional environmental exposures, which are not under real-time
study control, such as intrauterine development and environment,
labor and delivery process, and the immediate neonatal period,
251
may be accounted for, at least in part, by using detailed maternal
health history and antenatal care records. Among these environ-
mental exposures, the epigenetic effects of a hostile intrauterine
environment have been studied most extensively.
An increasing number of genome-wide association studies
(GWAS) report identified genetic polymorphisms that are associ-
ated with certain hemodynamic parameters such as blood pressure
[66], resting heart rate [67,68], heart rate variability [69], and car-
diac function [70]. These findings provide new insights into un-
derlying biological mechanisms of inter-individual variability of the
hemodynamic parameters and their interaction with potentially
important diagnostic and therapeutic implications. For example,
Turner et al. recently identified single nucleotide polymorphisms
(SNPs) that are associated with the response to different classes of
antihypertensive medications (diuretics, angiotensin II receptor
blockers) [71,72]. Using both hypothesis-driven and agnostic ap-
proaches, further research in this direction may allow for the
identification of genetic variants associated with therapeutic effi-
cacy of pharmacologic agents used in the treatment of neonatal
cardiovascular compromise.
6. The next step: cardiopulmonary and neurocritical care
monitoring system
The use of a multimodular monitoring system has been histor-
ically constrained by the limitations in output data acquisition
management systems. However, with a sufficiently stable data
acquisition system, the number of monitoring modules that can
simultaneously obtain and store data can be increased in quantity
and quality, with the latter being dependent on real-time sampling
rates with clinically relevant time-intervals. This type of next-
generation cardiopulmonary and neurocritical care monitoring
system is being developed in one of the authors' institutions at
Sidra Medical and Research Center in Doha, Qatar. In addition to
hemodynamic parameters (HR and its characteristics, BP, CO, SpO2,
cerebral, renal and muscle rSO2 and transcutaneous CO2 mea-
surements), this system will be capable of simultaneously incor-
porating ventilator output parameters, “smart-pump” infusion
dynamics, and aEEG output. By focusing on cardiorespiratory and
neurologic physiometric data, the creation of predictive models for
short-term neonatal outcomes such as intracranial hemorrhage
and long-term outcomes such as neurodevelopmental scores will
become feasible. Furthermore, by incorporating solitary data such
as neuroimages, biomarkers, and single nucleotide polymorphisms,
such multimodular monitoring systems can contribute to the cre-
ation of a new generation of real-time clinical decision support
tools to improve survival and, as importantly, quality of life for
critically ill neonates.
As mentioned earlier in the chapter, acquisition of real-time
physiometric data using comprehensive multimodular moni-
toring systems creates a big-data warehouse of phenotypic data,
both descriptive and responsive, that will be comparable with
voluminous big-data warehouses of genomic data. In the expo-
nentially growing field of personalized whole genomic
sequencing, the potential for integrating genomic data for single
nucleotide polymorphisms with phenotypic data from multi-
modular monitoring systems has the potential to create significant
clinical analytic synergy. Furthermore, as the field of proteomics
continues to develop, the application of biomarker research with
genomic and phenotypic multimodular monitoring data may un-
lock the potential that these big-data fields hold for modern
medical science.
The impact of integrated multimodular monitoring systems can
be applied to research paradigms as well as clinical ones. With
such systems, not only can algorithms for complex predictive
252 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254
models be generated to assist with clinical decision-making, they
can also be used to identify discrete patient subpopulations for
refinement of personalized medical care plans and clinical trial
designs. In consideration of the latter point, one of the principal
values of a randomized clinical trial is the attempt to inherently
equalize the effects of confounding factors that are typically un-
measured, either by design or by nature. However, each clinical
trial is still limited by the generalizability of its findings due to the,
at times, significant differences between the study population and
the patient population they are applied to by a practitioner. This
results in persistent Type I and Type II errors at the population
level, typically due to differential misclassification biases in iden-
tified confounders. By utilizing risk and outcome stratification
protocols afforded by the use of real-time multimodular moni-
toring systems, it is possible that existing clinical trial designs can
be refined, although the caveat of small sample sizes will continue
to be present as long as the availability of biomedical informatics
hardware remains limited. Given this caveat, it is likely that data-
driven identification of patient subpopulations within the gesta-
tional- and postnatal-age-defined cohorts stratified based on the
multimodular monitoring data will allow more appropriate se-
lection of patients for interventional trials to optimize benefit-to-
risk ratios. The potential advantages of this approach include
decrease in the sample size required to detect clinically mean-
ingful differences as well as avoidance of unnecessary exposure of
each patient to interventions targeting clinical study outcomes
with an identified lower risk of occurrence (e.g., P/IVH) or higher
likelihood of physiologic resolution (e.g., hemodynamically insig-
nificant PDA).
7. Conclusion
Comprehensive hemodynamic monitoring of neonates at risk
for perinatal complications and cardiovascular compromise based
on antenatal factors and delivery events is a vital step for early
identification of patient subpopulations whose postnatal transition
will take an abnormal course leading to progression of patho-
physiologic changes before they become clinically apparent and/or
irreversible. The prompt and accurate recognition of impeding
failure of compensatory mechanisms along with implementation of
appropriate interventions may potentially improve both short- and
long-term outcomes. Examples of interventions in appropriately
identified patient populations include ‒ but clearly are not limited
to ‒ delayed cord clamping that increases intravascular volume in a
gradual manner and promotes stabilization of systemic and
regional perfusion in the most premature neonates immediately
after delivery, the judicious use of vasoactive medications with
careful, stepwise titration of their doses to avoid significant blood
pressure swings, and selective use of indomethacin in patients at
high risk of P/IVH due to reperfusion phenomena. Further stratifi-
cation of neonatal subpopulations based on common genetic var-
iants (SNPs) that are associated with increased susceptibility and
risk of cardiovascular and other organ system instability and cor-
responding complications, and/or largely determine the response
to particular therapies, is likely to provide additional benefits in
developing pathophysiology-targeted interventions applicable to
clinical practice.
We propose that the phenotypic characterization of hemody-
namic responses described in this chapter combined with geno-
typic data analysis is the crucial step toward better understanding
of the complex and multi-level nature of neonatal hemodynamics.
We also feel strongly that it is imperative to build upon this
knowledge by developing physiologically justified and evidence-
based approaches in management of neonates at high risk of car-
diovascular compromise.
Practice points
 Efforts to obtain indirect (clinical signs) and direct (mea-
surements or assessments) information on the status of
the circulation (cardiac output, pulmonary blood flow)
and organ blood flow distribution are essential for the
interpretation of the findings on blood pressure in
particular and the hemodynamic status in general.
 Emerging novel technological approaches to continu-
ously and non-invasively assess cardiac output and organ
blood flow distribution are promising but need further
validation.
 Correct and clinically relevant interpretation of the data
obtained by monitoring is not possible without the thor-
ough understanding of the principles of developmental
hemodynamics.
Research directions
 Further validation of the novel technologies used in the
continuous and non-invasive monitoring of cardiac
output and organ blood flow distribution by the gold
standard of the measurement of the given hemodynamic
parameter such as MRI for assessment of cardiac output.
 Use of predictive mathematical modeling, large data
handling and machine learning to enable us to better
predict occurrences of pathological events of clinical
significance during the next day, days, or beyond in the
course of the given patient.
 Using predictive modeling, subpopulations of patients
likely to develop preventable and/or treatable conditions
can be identified and be selected for interventional trials
to increase the benefit-to-risk ratio of the intervention
under investigation.
Conflict of interest statement
None declared.
Funding sources
None.
References
[1] Heisenberg W. Physics and philosophy. London: George Allen & Unwin; 1958.
p. 57. Available at: http://archive.org/details/PhysicsPhilosophy.
[2] Kluckow M. Low superior vena cava flow and intraventricular haemorrhage in
preterm infants. Archs Dis Childh Fetal Neonatal Ed 2000;82:188Fe94F.
[3] Hunt RW, Evans N, Rieger I, Kluckow M. Low superior vena cava flow and
neurodevelopment at 3 years in very preterm infants. J Pediatr 2004;145:
588e92.
[4] Fanaroff JM, Wilson-Costello DE, Newman NS, Montpetite MM, Fanaroff AA.
Treated hypotension is associated with neonatal morbidity and hearing loss in
extremely low birth weight infants. Pediatrics 2006;117:1131e5.
[5] Osborn DA, Evans N, Kluckow M. Hemodynamic and antecedent risk factors of
early and late periventricular/intraventricular hemorrhage in premature in-
fants. Pediatrics 2003;112:33e9.
[6] Pellicer A, Bravo MC, Madero R, Salas S, Quero J, Caban ~ as F. Early systemic
hypotension and vasopressor support in low birth weight infants: impact on
neurodevelopment. Pediatrics 2009;123:1369e76.
[7] Tan KL. Blood pressure in very low birth weight infants in the first 70 days of
life. J Pediatr 1988;112:266e70.
 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254
[8] Shortland DB, Evans DH, Levene MI. Blood pressure measurements in very low
birth weight infants over the first week of life. J Perinat Med 1988;16:93e7.
[9] Lee J, Rajadurai VS, Tan KW. Blood pressure standards for very low birth-
weight infants during the first day of life. Archs Dis Childh Fetal Neonatal Ed
1999;81:F168e70.
[10] Zubrow AB, Hulman S, Kushner H, Falkner B. Determinants of blood pressure
in infants admitted to neonatal intensive care units: a prospective multicenter
study. Philadelphia Neonatal Blood Pressure Study Group. J Perinatol 1995;15:
470e9.
[11] Cunningham S, Symon AG, Elton RA, Zhu C, McIntosh N. Intra-arterial blood
pressure reference ranges, death and morbidity in very low birthweight in-
fants during the first seven days of life. Early Hum Dev 1999;56:151e65.
[12] Bada HS, Korones SB, Perry EH, Arheart KL, Ray JD, Pourcyrous M, et al. Mean
arterial blood pressure changes in premature infants and those at risk for
intraventricular hemorrhage. J Pediatr 1990;117:607e14.
[13] Batton B, Li L, Newman NS, Das A, Watterberg KC, Yoder BA, et al., the Eunice
Kennedy Shriver National Institute of Child Health & Human Development
Neonatal Research Network.. Evolving blood pressure dynamics for extremely
preterm infants. J Perinatol 2014;34:301e5.
[14] Cordero L, Timan CJ, Waters HH, Sachs LA. Mean arterial pressures during the
first 24 hours of life in < or ¼ 600-gram birth weight infants. J Perinatol
2002;22:348e53.
[15] Roze JC, Tohier C, Maingueneau C, Lefevre M, Mouzard A. Response to
dobutamine and dopamine in the hypotensive very preterm infant. Archs Dis
Childh 1993;69:59e63.
[16] Bidegain M, Greenberg R, Simmons C, Dang C, Cotten CM, Smith PB. Vaso-
pressin for refractory hypotension in extremely low birth weight infants.
J Pediatr 2010;157:502e4.
[17] Paradisis M, Evans N, Kluckow M, Osborn D. Randomized trial of milrinone
versus placebo for prevention of low systemic blood flow in very preterm
infants. J Pediatr 2009;154:189e95.
[18] Osborn D, Evans N, Kluckow M. Randomized trial of dobutamine versus
dopamine in preterm infants with low systemic blood flow. J Pediatr
2002;140:183e91.
[19] Noori S, Stavroudis T, Seri I. Principles of developmental cardiovascular
physiology and pathophysiology. In: Kleinman C, Seri I, editors. Neonatology
questions and controversies: hemodynamics and cardiology. Philadelphia:
Saunders/Elsevier; 2012. p. 3e28.
[20] Azhibekov T, Noori S, Soleymani S, Seri I. Transitional cardiovascular physi-
ology and comprehensive hemodynamic monitoring in the neonate: rele-
vance to research and clinical care. Semin Fetal Neonatal Med 2014;19:45e53.
[21] Soleymani S, Borzage M, Seri I. Hemodynamic monitoring in neonates: ad-
vances and challenges. J Perinatol 2010;30:S38e45.
[22] Soleymani S, Borzage M, Noori S, Seri I. Neonatal hemodynamics: monitoring,
data acquisition and analysis. Expert Rev Med Devices 2012;9:501e11.
[23] Engle WD. Blood pressure in the very low birth weight neonate. Early Hum
Dev 2001;62:97e130.
[24] Nuntnarumit P, Yang W, Bada-Ellzey HS. Blood pressure measurements in the
newborn. Clin Perinatol 1999;26:981e96. x.
[25] Noori S, Drabu B, Soleymani S, Seri I. Continuous non-invasive cardiac output
measurements in the neonate by electrical velocimetry: a comparison with
echocardiography. Archs Dis Childh Fetal Neonatal Ed 2012;97:F340e3.
[26] Grollmuss O, Gonzalez P. Non-invasive cardiac output measurement in low
and very low birth weight infants: a method comparison. Front Pediatr
2014;2:16.
[27] Keren H, Burkhoff D, Squara P. Evaluation of a noninvasive continuous cardiac
output monitoring system based on thoracic bioreactance. Am J Physiol Heart
Circ Physiol 2007;293:H583e9.
[28] Weisz DE, Jain A, McNamara PJ, EL-Khuffash A. Non-invasive cardiac output
monitoring in neonates using bioreactance: a comparison with echocardiog-
raphy. Neonatology 2012;102:61e7.
[29] De Felice C, Del Vecchio A, Criscuolo M, Lozupone A, Parrini S, Latini G. Early
postnatal changes in the perfusion index in term newborns with subclinical
chorioamnionitis. Archs Dis Childh Fetal Neonatal Ed 2005;90:F411e4.
[30] Kinoshita M, Hawkes CP, Ryan CA, Dempsey EM. Perfusion index in the very
preterm infant. Acta Paediatr 2013;102:e398e401.
[31] Takahashi S, Kakiuchi S, Nanba Y, Tsukamoto K, Nakamura T, Ito Y. The
perfusion index derived from a pulse oximeter for predicting low superior
vena cava flow in very low birth weight infants. J Perinatol 2010;30:265e9.
[32] Vidal M, Ferragu F, Durand S, Baleine J, Batista-Novais AR, Cambonie G.
Perfusion index and its dynamic changes in preterm neonates with patent
ductus arteriosus. Acta Paediatr 2013;102:373e8.
[33] Van Bel F, Lemmers P, Naulaers G. Monitoring neonatal regional cerebral
oxygen saturation in clinical practice: value and pitfalls. Neonatology
2008;94:237e44.
[34] Ancora G, Maranella E, Grandi S, Sbravati F, Coccolini E, Savini S, et al. Early
predictors of short term neurodevelopmental outcome in asphyxiated cooled
infants. A combined brain amplitude integrated electroencephalography and
near infrared spectroscopy study. Brain Dev 2013;35:26e31.
[35] Ter Horst HJ, Sommer C, Bergman KA, Fock JM, van Weerden TW, Bos AF.
Prognostic significance of amplitude-integrated EEG during the first 72 hours
after birth in severely asphyxiated neonates. Pediatr Res 2004;55:1026e33.
[36] Rennie J, Chorley G, Boylan GB, Pressler R, Nguyen Y, Hooper R. Non-expert
use of the cerebral function monitor for neonatal seizure detection. Archs Dis
Childh Fetal Neonatal Ed 2004;89:F37e40.
253
[37] West CR, Groves AM, Williams CE, Harding JE, Skinner JR, Kuschel CA, et al.
Early low cardiac output is associated with compromised electroencephalo-
graphic activity in very preterm infants. Pediatr Res 2006;59:610e5.
[38] Shah D, Paradisis M, Bowen JR. Relationship between systemic blood flow,
blood pressure, inotropes, and aEEG in the first 48 h of life in extremely
preterm infants. Pediatr Res 2013;74:314e20.
[39] Toet MC, Lemmers PMA. Brain monitoring in neonates. Early Hum Dev
2009;85:77e84.
[40] Toet MC, Lemmers PMA, van Schelven L, van Bel F. Cerebral oxygenation and
electrical activity after birth asphyxia: their relation to outcome. Pediatrics
2006;117:333e9.
[41] Ter Horst HJ, Verhagen EA, Keating P, Bos AF. The relationship between
electrocerebral activity and cerebral fractional tissue oxygen extraction in
preterm infants. Pediatr Res 2011;70:384e8.
[42] Tsuji M, Saul JP, du Plessis A, Eichenwald E, Sobh J, Crocker R, et al. Cerebral
intravascular oxygenation correlates with mean arterial pressure in critically
ill premature infants. Pediatrics 2000;106:625e32.
[43] Soul JS, Hammer PE, Tsuji M, Saul JP, Bassan H, Limperopoulos C, et al. Fluctu-
ating pressure-passivity is common in the cerebral circulation of sick premature
infants. Pediatr Res 2007;61:467e73.
[44] Verhagen EA, Ter Horst HJ, Keating P, Martijn A, Van Braeckel KN, Bos AF.
Cerebral oxygenation in preterm infants with germinal matrix‒intraventric-
ular hemorrhages. Stroke J Cereb Circ 2010;41:2901e7.
[45] Alderliesten T, Lemmers PM, Smarius JJ, van de Vosse RE, Baerts W, van Bel F.
Cerebral oxygenation, extraction, and autoregulation in very preterm infants
who develop peri-intraventricular hemorrhage. J Pediatr 2013;162. 698e704.e2.
[46] Noori S, McCoy M, Anderson MP, Ramji F, Seri I. Changes in cardiac function
and cerebral blood flow in relation to peri/intraventricular hemorrhage in
extremely preterm infants. J Pediatr 2014;164. 264e270.e1e3.
[47] Noori S, Anderson M, Soleymani S, Seri I. Effect of carbon dioxide on cerebral
blood flow velocity in preterm infants during postnatal transition. Acta Pae-
diatr 2014;103:e334e9.
[48] Munro MJ, Walker AM, Barfield CP. Hypotensive extremely low birth weight
infants have reduced cerebral blood flow. Pediatrics 2004;114:1591e6.
[49] Alderliesten T, Lemmers PM, van Haastert IC, de Vries LS, Bonestroo HJC,
Baerts W. Hypotension in preterm neonates: low blood pressure alone does
not affect neurodevelopmental outcome. J Pediatr 2014;164:986e91.
[50] Neal M. Patient-specific modeling for critical care. In: Kerckhoffs RCP, editor.
Patient-specific modeling of the cardiovascular system. New York: Springer;
2010. p. 81e94.
[51] Fairchild KD, O’Shea TM. Heart rate characteristics: physiomarkers for
detection of late-onset neonatal sepsis. Clin Perinatol 2010;37:581e98.
[52] Fairchild KD, Schelonka RL, Kaufman DA, Carlo WA, Kattwinkel J, Porcelli PJ,
et al. Septicemia mortality reduction in neonates in a heart rate characteristics
monitoring trial. Pediatr Res 2013;74:570e5.
[53] Pinsky MR. Functional haemodynamic monitoring. Curr Opin Crit Care
2014;20:288e93.
[54] Michard F, Boussat S, Chemla D, Anguel N, Mercat A, Lecarpentier Y, et al.
Relation between respiratory changes in arterial pulse pressure and fluid
responsiveness in septic patients with acute circulatory failure. Am J Respir
Crit Care Med 2000;162:134e8.
[55] Lanspa MJ, Grissom CK, Hirshberg EL, Jones JP, Brown SM. Applying dynamic
parameters to predict hemodynamic response to volume expansion in
spontaneously breathing patients with septic shock. Shock 2013;39:155e60.
[56] Berkenstadt H, Margalit N, Hadani M, Friedman Z, Segal E, Villa Y, et al. Stroke
volume variation as a predictor of fluid responsiveness in patients undergoing
brain surgery. Anesth Analg 2001;92:984e9.
[57] Creteur J, Carollo T, Soldati G, Buchele G, De Backer D, Vincent JL. The prognostic
value of muscle StO2 in septic patients. Intensive Care Med 2007;33:1549e56.
[58] Mesquida J, Espinal C, Gruartmoner G, Masip J, Sabatier C, Baigorri F, et al.
Prognostic implications of tissue oxygen saturation in human septic shock.
Intensive Care Med 2012;38:592e7.
[59] García MIM, Cano AG, Romero MG. Dynamic arterial elastance to predict
arterial pressure response to volume loading in preload-dependent patients.
Crit Care 2011;15:R15.
[60] Hadian M, Severyn DA, Pinsky MR. The effects of vasoactive drugs on pulse
pressure and stroke volume variation in postoperative ventilated patients.
J Crit Care 2011;26. 328.e1e8.
[61] Ivanova O, K Khoo M. Simulation of spontaneous cardiovascular variability
using PNEUMA. IEEE Engng Med Biol Soc Annu Conf 2004;6:3901e4.
[62] Soleymani S, Khoo MC, Noori S, Seri I. Modeling of neonatal hemodynamics
during PDA closure. Fred S. Grodins Graduate Research Symposium 2013, Los
Angeles, CA, USA (Abstract).
[63] Clyman RI, Mauray F, Heymann MA, Roman C. Cardiovascular effects of patent
ductus arteriosus in preterm lambs with respiratory distress. J Pediatr
1987;111:579e87.
[64] Weisz DE, Jain A, Ting J, McNamara PJ, El-Khuffash A. Non-invasive cardiac
output monitoring in preterm infants undergoing patent ductus arteriosus
ligation: a comparison with echocardiography. Neonatology 2014;106:330e6.
[65] Lien R, Hsu K-H, Chu J-J, Chang Y-S. Hemodynamic alterations recorded by
electrical cardiometry during ligation of ductus arteriosus in preterm infants
[Epub ahead of print] Eur J Pediatr 2014 Oct 15. http://dx.doi.org/10.1007/
s00431-014-2437-9.
[66] International Consortium for Blood Pressure Genome-Wide Association
Studies, Ehret GB, Munroe PB, Rice KM, Bochud M, Johnson AD, Chasman DI,
254 T. Azhibekov et al. / Seminars in Fetal & Neonatal Medicine 20 (2015) 246e254
et al. Genetic variants in novel pathways influence blood pressure and car-
diovascular disease risk. Nature 2011;478:103e9.
[67] Eijgelsheim M, Newton-Cheh C, Sotoodehnia N, de Bakker PIW, Muller M,
Morrison AC, et al. Genome-wide association analysis identifies multiple loci
related to resting heart rate. Hum Mol Genet 2010;19:3885e94.
[68] Mezzavilla M, Iorio A, Bobbo M, D'Eustacchio A, Merlo M, Gasparini P, et al.
Insight into genetic determinants of resting heart rate. Gene 2014;545:170e4.
[69] Newton-Cheh C, Guo CY, Wang TJ, O’donnell CJ, Levy D, Larson MG. Genome-
wide association study of electrocardiographic and heart rate variability
traits: the Framingham Heart Study. BMC Med Genet 2007;8(Suppl. 1):S7.
[70] Vasan RS, Glazer NL, Felix JF, Lieb W, Wild PS, Felix SB, et al. Genetic variants
associated with cardiac structure and function: a meta-analysis and replica-
tion of genome-wide association data. JAMA 2009;302:168e78.
[71] Turner ST, Bailey KR, Schwartz GL, Chapman AB, Chai HS, Boerwinkle E.
Genomic association analysis identifies multiple loci influencing antihyper-
tensive response to an angiotensin II receptor blocker. Hypertension 2012;59:
1204e11.
[72] Turner ST, Boerwinkle E, O’Connell JR, Bailey KR, Gong Y, Chapman AB, et al.
Genomic association analysis of common variants influencing antihyperten-
sive response to hydrochlorothiazide. Hypertension 2013;62:391e7.