HEMODYNAMIC MONITORING
MINERVA ANESTESIOL 2002;68:219-25
Hemodynamic monitoring
The goal of hemodynamic monitoring is to
maintain adequate tissue perfusion. Classical
hemodynamic monitoring is based on the
invasive measurement of systemic, pulmo-
nary arterial and venous pressures, and of
cardiac output. Since organ blood flow can-
not be directly measured in clinical practice,
arterial blood pressure is used, despite limita-
tions, as estimate of adequacy of tissue perfu-
sion. A mean arterial pressure (MAP) of 70
mm Hg may be considered a reasonable tar-
get, associated with sign of adequate organ
perfusion, in most patients. In the approach
to hypotension, which is the most common
cause of hemodynamic instability in critical
ill patients, increasing levels of monitoring
may be used. Assuming that central venous
pressure (CVP) and pulmonary artery occlu-
sion pressure (PAOP) are adequate estimates
of the volume of the systemic and pulmonary
circulation respectively, the following deci-
sion tree is suggested: 1) make a working dia-
gnosis based on the relationship between
pressure (CVP and PAOP) and cardiac output
or stroke volume (CO or SV); 2) consider con-
ditions that may alter reliability of CVP and
PAOP in estimate adequately circulating volu-
mes such as abnormal pressure/volume rela-
tionship (compliance) of the RV or LV,
increased intrathoracic pressure (PEEP, auto-
PEEP, intra-abdominal pressure), valvular
heart disease (mitral stenosis); 3) look at the
history; 4) separating RV and LV by reciprocal
variations of CVP, PAOP and SV. CVP is often
used as sole parameter to monitor hemody-
Address reprint requests to: L. Bigatello - Department of
Anesthesia and Critical Care, Massachusetts General Hospi-
tal - 55 Fruit Street - Boston MA 02114.
E-mail: lbigatello@partners.org
Vol. 68, N. 4 MINERVA ANESTESIOLOGICA
L. M. BIGATELLO, E. GEORGE
From the Department of Anesthesia
and Critical Care, Massachusetts General Hospital
Harvard Medical School, Boston, Massachusetts
namic. However CVP alone may not differen-
tiate between changes in volume (different
venous return curve) or changes in contracti-
lity (different starling curve). Finally, other
techniques such as echcardiography, transe-
sophageal Doppler and volume-based moni-
toring system are now available.
Key words: Hemodynamics - Blood pressure -
Blood pressure determination.
T he goal of hemodynamic monitoring is
to maintain adequate tissue perfusion.
In critically ill patients, hypoperfusion of
vital organs may lead to multiple organ
systems dysfunction and death. A great deal
of controversy exists over the indications
and safety of invasive monitoring. Classical
hemodynamic monitoring is based on the
invasive measurement of systemic and pul-
monary arterial and venous pressures and
of cardiac output. Although burdened with
potential flaws and complications, classical
pressure monitoring remains the most fre-
quently employed means of monitoring
patient hemodynamics in the operating
room and in Intensive Care Unit (ICU).
Newer monitoring techniques are promi-
sing but, for various reasons, have not yet
219
BIGATELLO
Hypovolemia
Decreased
venous return Mechanical
obstruction
Hypotension Myocardial
dysfunction
Arterial
vasodilatation
Fig. 1.—A physiological approach to hypotension.
reached widespread acceptance. The aim of
this lecture is to guide clinicians through
the interpretation of hemodynamic data
based on the application of classic circula-
tory physiology.
Why measure arterial blood pressure
Regrettably, tissue perfusion (i.e., organ
blood flow) cannot be directly measured in
clinical practice:
Organ Blood Flow = (arterial pressure-
venous pressure) / resistance
In the absence of a measurement of
actual blood flow to individual organs, and
assuming constant venous pressure and
constant resistance, measurement of arterial
blood pressure gives a reasonable estimate
of the adequacy of tissue perfusion. One
can easily see how crude this measurement
is: by measuring the blood pressure at the
brachial or radial artery we hope to estima-
te the adequacy of blood flow to the kid-
neys, brain, coronary circulation, etc.
However, physiology helps our limited
capacity. Under normal circumstances,
organ blood flow is strictly maintained
within normal range by autoregulation, i.e.,
during wide changes of arterial blood pres-
sure, blood flow remains constant through
constriction or dilation of the afferent ves-
sels. Unfortunately, in pathological condi-
tions, (hypertension, trauma, sepsis, etc.)
autoregulation can be significantly impaired
and flow may become directly dependent
on perfusion pressure. For the above rea-
sons, we normally do not measure our own
220 MINERVA ANESTESIOLOGICA
HEMODYNAMIC MONITORING
blood pressure as it changes widely with
our daily activities, but we resolve to invasi-
ve means to measure arterial blood pressu-
re in critically ill patients.
Despite the limitations of peripheral
blood pressure measurement, maintaining a
reasonable value of arterial pressure is asso-
ciated with signs of adequate organ func-
tion in most critically ill patients. The fol-
lowing suggestions may enhance the effec-
tiveness of arterial blood pressure monito-
ring.
a) The mean arterial pressure (MAP) is
the best physiological estimate of perfusion
pressure and is less subject to measurement
variability than the systolic pressure.
b) A MAP ≥70 mmHg is a reasonable tar-
get for most patients. At times (chronic
hypertension, cerebral edema, spinal cord
ischemia, etc.), higher values are necessary.
Controversy exists on the accuracy of clini-
cal parameters of vital organ function, such
as urine output and acid-base status, as
early indicators of tissue hypoperfusion.
However, no other proposed parameter,
such as serum lactate and gastric mucosal
pH, has yet been shown consistently to be
superior.1
c) Optimal blood flow through vital
organs is first achieved by maintaining an
adequate circulating volume. An increase in
blood pressure achieved using vasocon-
strictor agents in hypovolemic patients does
not provide adequate organ perfusion and
can be deleterious.
How to measure arterial blood
pressure
Non invasive (generally automated) oscil-
lometric blood pressure measurement is
widely available. Its accuracy is decreased
by the presence of rapidly changing blood
pressure, arrhythmias, hypotension and
hypertension. It should not be used in
hemodynamically unstable patients.
Intra-arterial blood pressure measure-
ment via a catheter-transducer system is
extremely reliable if the system is properly
Aprile 2002
HEMODYNAMIC MONITORING
TABLE I.—Example of a stepwise approach to the management of hypotensive patients.
Patient with severe hypotension/shock
Young, previously healthy patient with lower
extremity injury
65 y.o. patient with lower extremity injury and
a history of heart disease
Same patient, who did not respond to a limi-
ted volume challenge
Same patient, hypotensive a week later in the
ICU
set up.2 It should be used whenever possi-
ble in hemodynamically unstable patients.
Physiological approach to hypotension
Hypotension is the most common cause
of hemodynamic instability in critically ill
patients, particularly surgical patients. We
will limit our discussion of hemodynamics
to the interpretation of hypotension, but the
general principles illustrated here apply to
hemodynamic monitoring in general. We
suggest a simple approach to the diagnosis
of hypotension, summarized in Figure 1.
This schema can be applied to clinical
practice using increasing levels of monito-
ring. Table I shows a suggested stepwise
approach to the hemodynamic monitoring
of a hypotensive patient.
In many cases, the diagnosis can be sug-
gested on clinical grounds. For example,
hypotension in a young patient bleeding
from a duodenal ulcer should be first attri-
buted to hypovolemia and invasive monito-
ring should not be required to guide fluid
resuscitation. In less obvious cases, it is rea-
sonable to “try” an intervention, and con-
firm or reject the diagnosis post-hoc (“trial
and error”). For example, if our above
patient had also a history of coronary artery
disease, one should think of myocardial
dysfunction as a contributor to the hypoten-
sion. Given the clinical scenario, volume
resuscitation could still be a reasonable ini-
tial step, with a plan to escalate the level of
monitoring if the desired response is not
observed. As patients develop complex
problems during a prolonged ICU course,
the etiology of hypotension becomes be
Vol. 68, N. 4
BIGATELLO
Monitoring Treatment
Arterial line Volume
Arterial line “Trial & error”: volume, in a limited
amount
Arterial line + central monitoring According to central monitoring
Arterial line + central monitoring According to central monitoring
TABLE II.—Central pressures and cardiac output
changes in hypotension.
Hypotension CVO/PAOP CO/SV
Decreased venous return:
— hypovolemia ⇓ ⇓
— obstruction ⇑ ⇓
Myocardial dysfunction ⇑ ⇓
Arterial vasodilation ⇔ ⇑
more and more difficult to sort out, thus
requiring invasive monitoring.
Central venous pressure (CVP) monito-
ring provides a useful estimate of the volu-
me status of the systemic circulation and
(see below the discussion of interpretation
of CVP). The main limitations of CVP moni-
toring are that (a) it does not allow to mea-
sure cardiac output; and (b) it does not pro-
vide reliable information on the status of
the pulmonary circulation in the presence
of left ventricular dysfunction.
Pulmonary artery (PA) pressure monito-
ring with a PA catheter allows to measure
(CO) and stroke volume (SV), PA pressure
and PA occlusion pressure (PAOP) and hen-
ce to separately assess the performance of
the right and the left ventricle (RV and LV).
Central pressure measurements are inten-
ded to estimate volumes, i.e., the CVP esti-
mates the volume of the systemic circula-
tion and the PAOP or “wedge pressure”
estimates the volume of the pulmonary cir-
culation. As long as the postulate that cen-
tral pressures accurately reflect volumes
holds true, the characteristic hemodynamic
findings of hypotensive patients are strai-
ghtforward, as summarized in Table II.
The following decision tree may guide in
MINERVA ANESTESIOLOGICA 221
BIGATELLO
Pressure (cm H2O)
End-systolic
Systolic
120
Art. BP
End-diastolic
16
PAOP 10
60 90 140 170 Volume (mls)
Fig. 2.—Hemodynamic effect of a hypothetical acute epi-
sode of LV dysfunction (dotted lines). See text for expla-
nation.
the interpretation of hemodynamic data
obtained with invasive monitoring in a
hypotensive patient.
1) Make a working diagnosis based on
the relationship between pressures (CVP
and PAOP) and cardiac output (CO or SV)
as summarized in Table II. We assume at
this point that the CVP and the PAOP are
adequate estimates of the RV and LV end-
diastolic volumes respectively and that the
right (CVP) and left (PAOP) side of the cir-
culation are equally affected by the cause of
hypotension.
2) Revise our basic assumption that CVP ≈
volume of the right side of the circulation
and that PAOP ≈ volume of the left side of
the circulation. Unfortunately, this assump-
tion is often flawed. Knowledge of the basic
physiology underlying the pressure/volume
relationship in the central circulation will
greatly increase the accuracy of hemodyna-
mic monitoring. Our basic assumption can
be altered under three main circumstances:
a) When the volume/pressure relation-
ship (compliance) of the RV or LV is abnor-
mal, as it may happen with concentric LV
hypertrophy (LVH) from hypertensive car-
diomyopathy and aortic stenosis. In this
case, the measured PAOP overestimates the
LV end-diastolic volume.
b) When the pressure measurement does
not estimate the actual transmural pressure
across a cardiac chamber. An increase in
the intrathoracic pressure may be transmit-
222 MINERVA ANESTESIOLOGICA
HEMODYNAMIC MONITORING
ted to the blood vessel where the tip of our
catheter is lodged and increase the measu-
red pressure without any actual increase in
circulating volume. Common causes of
increased intrathoracic pressure that
mislead the interpretation of CVP and PAOP
include PEEP, autoPEEP and increased
intra-abdominal pressure. The fraction of
pressure transmitted through the blood ves-
sel depends on a number of factors, inclu-
ding the compliance of the anatomical
structures involved and the tension of the
blood vessel wall. While an exact estimate
of the fraction of pressure transmitted is
often impossible, a reasonable idea can be
derived from the compliance of the lung
and chest wall. For example, transmission
of auto-PEEP in a patient with COPD (com-
pliant lungs) may be substantial, while tran-
smission of applied PEEP in a patient with
ARDS (stiff lungs) may be minimal.
c) Mitral stenosis. Valvular heart disease
may affect the interpretation of hemodyna-
mic monitoring in many ways, and yet inva-
sive monitoring may be crucial in the inter-
pretation of hypotension in patients with
valvular defects. With significant mitral ste-
nosis, the PAOP may not correctly estimate
the LV end diastolic pressure because of
inadequate LV filling time. Hence, a high
PAOP may be recorded when the LV is still
underfilled.
It is very important to note that in all the
above situations, the pressures measured
are indeed correct, rather than measure-
ment errors. A high PAOP in the presence
of severe concentric LVH from aortic steno-
sis is indeed an accurate reflection of the
high pressure in the pulmonary veins and
left atrium and, as such, can result in acute
pulmonary edema. However, the LV may
still be underfilled. This example undersco-
res both the difficulty and the possible
benefit of the correct interpretation of inva-
sive hemodynamic monitoring in complex
circumstances such as valvular heart disea-
se.
3) Look at the history. The current
hemodynamic values have to be put in the
patient’s context. Although any properly
obtained hemodynamic profile should be
Aprile 2002
HEMODYNAMIC MONITORING
interpreted as the reflection of a specific
moment in time, looking at previous num-
bers as well as at all other relevant clinical
variables improves our understanding.
4) Separating RV and LV. The CVP and
PAOP may change independently, because
they measure two separate entities. The CVP
reflects the volume of the systemic circula-
tion and the PAOP the volume of the pulmo-
nary circulation. To better understand this
point, let’s examine the effect of a hypotheti-
cal episode of acute, isolated LV dysfunction
from, e.g., ischemia. The decreased LV con-
tractility causes a decrease of SV ejected by
the LV, and a few less mls of blood enter the
systemic circulation. This decrease of SV is
so minimal in respect to the size of the
venous reservoir that has no discernible
effect on the venous return to the RV. Thus,
the RV continues to eject an approximately
normal SV, which will be “accommodated”
by the dysfunctional LV by increasing the LV
end-diastolic volume and pressure (≈PAOP).
At steady state, the SV of the two ventricles
will be again equal and close to normal: the
RV “drives the circulation”! Figure 2 summa-
rizes graphically these events.
The two thick lines represent the LV end-
systolic V/P relationship (an estimate of LV
“contractility”) and the LV end-diastolic V/P
relationship (an estimate of LV complian-
ce). The quadrangle encased by the conti-
nuous line is the cardiac cycle, starting from
opening of the mitral valve (left lower cor-
ner) and going counterclockwise. The area
of this quadrangle estimates the “work”
during that cycle. In this hypothetical set-
ting, the systolic arterial pressure is 120 mm
Hg, the LV end-diastolic pressure as estima-
ted by the PAOP is 10 mm Hg, the SV is 80
mls (see below).
The ischemic event causes an acute
decrease of the slope of the end-systolic
curve (dotted line), i.e., a decrease in con-
tractility and in systemic blood pressure.
The events described above (i.e., decreased
SV of the LV and maintained SV of the RV)
move the new cardiac cycle to the right
and, once steady state is reached, the SV of
the LV is the same as prior to the event, but
at a higher PAOP, as described. This series
Vol. 68, N. 4 MINERVA ANESTESIOLOGICA
BIGATELLO
TABLE III.—Hemodynamics of RV and LV failure.
Parameters CVP PAOP SV
Isolated LV failure ⇔ ⇑ ⇔
LV failure + compensated ⇑ ⇑ ⇔
RV failure
Biventricular failure ⇑ ⇑ ⇓
Isolated RV failure ⇑ ⇔ ⇓
RV failure + volume infusion ⇑ ⇔ ⇔
of events can be further explained in terms
of LV ejection fraction (EF), a volumetric
measurement that can be obtained by echo-
cardiography:
EF = (EDV - ESV) / EDV
Where EDV = end-diastolic volume, ESV
= end-systolic volume and (EDV - ESV) =
SV. If we apply to this equation the num-
bers posted in Figure 2:
1) starting EF = (140 - 60) / 140 mls = 57%
2) post-event EF = (170 - 90) / 170 mls =
47%
Note that the SV before and after the
event is the same by virtue of an increase of
the EDV. Calculation of the EF confirms that
the LV has suffered an acute event. The dia-
gram of the type shown in Figure 2 can be
extremely helpful because it can be used to
visualize the effects of other possible
responses to the ischemic event. For exam-
ple, if the systolic blood pressure is not
changed, the same SV can be ejected by the
dysfunctional LV with less “work”, at a
much lower systemic pressure. Such a com-
pensation is what can be achieved, if indi-
cated, by “afterload reduction”. Or, one can
describe the possible compensation of a
different acute event characterized by dia-
stolic dysfunction instead of a decrease in
contractility.
In the occurrence of a longstanding injury
to the LV, this acute picture may evolve over
time in biventricular failure through the
development of pulmonary hypertension
and RV overload (Table III). As the RV fails,
CVP increases and eventually the SV decrea-
ses. Note that when the SV of the RV falls,
the SV of the LV falls as well, due to the
small size of the pulmonary venous reser-
voir. Again: the RV “drives the circulation”!
223
BIGATELLO HEMODYNAMIC MONITORING

Venous return (L/min) CO (L/min) with different physiological implications.

The two possible extremes are that the CVP
has increased only due to an increase in
6
volume (new venous return curve) or that it
4 has increased only due to a decrease in
contractility (new Starling curve). Clearly, a
combination of both phenomena is possi-
ble. The same thinking process can be illu-
strated for a decrease in CVP. Hence, an iso-
lated CVP value can represent very different
RA pressure RV diastolic pressure hemodynamic conditions, and without a
Fig. 3.—Curves of venous return and cardiac output at CO measurement, we have to use clinical
different “driving” pressures. equivalents to interpret the change in CVP.
In a reasonably stable patient, changes in
MAP should parallel changes in CO. An
Venous return and CO (L/min)

increase in CVP will be likely due to an

increased circulating volume if the MAP
also increases. An increase in CVP will be
likely due to a decreased contractility if the
MAP decreases. In an unstable patient,
measurement of the CO may be necessary.
RA and
CVP RV diastolic
pressure (cm H2O)
The role of echocardiography
4 8
Fig. 4.—Interpretation of an increase in CVP (see text).
in the ICU

We have described how we use the mea-

Interpretation of the CVP surement of pressures (CVP and PAOP) to
estimate RV and LV volumes. We also
Most times the CVP is monitored with a reviewed how at times these pressures do
simple central venous catheter without the not correctly estimate volumes. The possi-
capability of measuring CO. The following bility of actually measuring RV and LV volu-
discussion is designed to assist in the inter- mes through echocardiography is an exci-
pretation of the CVP when the measure- ting addition to the monitoring armamenta-
ment of the CO is not available. rium in the ICU. At present, its use is still
Venous return and CO are described by erratic in most ICUs. Obstacles to its wide-
these two curves (Fig. 3):3, 4 spread use include the sporadic availability
In the venous return graph., the upper of cardiologists, the lack of training of
curve represents a normal circulating volu- intensivists, and the technical difficulties
me and the lower represents hypovolemia. often encountered in critically ill patients.
The CO graph. shows two contractility Given its potential diagnostic benefits, it is
(“Starling”) curves, the lower one represen- likely that its use in the ICU will rapidly
ting a decreased contractility. At steady sta- increase.5
te, venous return and CO are equal, and the
two curves can be constructed on the same
graph. The point at which the two curves Other techniques of hemodynamic
intersect is the CVP (Fig. 4). monitoring
In this example, the starting CVP is 4
cmH2O. If the CVP increases to 8 cmH2O, Other means of hemodynamic monito-
the new value can occur at a variety of CO ring have been recently developed, aimed
values, as shown by the dotted vertical line, either at a less invasive approach or at a

224 MINERVA ANESTESIOLOGICA Aprile 2002

HEMODYNAMIC MONITORING
volume-based monitoring. These include
transesophageal 6 and transtracheal 7 Dop-
pler flow measurement and volume moni-
toring through the estimate of the intratho-
racic blood volume. These techniques
require more consistent clinical validation.8
Pulmonary artery catheters
and outcome
It seems reasonable to believe that, if
hemodynamic data are interpreted correc-
tly, they could benefit the care of complex
unstable patients. However, there is no
clear evidence of such benefit. Rather, in a
recent retrospective study, the use of PA
catheters was associated with an increased
mortality of critically ill patients.9 The mor-
tality was not related to any complication
from the PA catheters. Although these
results do not seem sufficiently strong to
discourage the use of invasive hemodyna-
mic monitoring, they underscore the impor-
tance of using appropriate indications for
invasive monitoring and of correctly inter-
preting the data obtained.
Acknowledgments.—We are grateful to the house
staff of the Surgical Intensive Care Unit of the Massachu-
setts General Hospital: fellows, residents and nurses,
who make our daily care of critically ill patient an
always new learning experience.
Riassunto
Monitoraggio emodinamico
Lo scopo del monitoraggio emodinamico è quel-
lo di mantenere un’adeguata perfusione tissutale. Il
monitoraggio emodinamico classico si basa sulla
misurazione invasiva delle pressioni arteriosa siste-
mica, pressione arteriosa polmonare e venosa cen-
trale, e della portata cardiaca. Dato che il flusso di
sangue di un organo non può essere direttamente
misurato nella pratica clinica, si usa la pressione
arteriosa, malgrado i suoi limiti, per stimare l’ade-
guatezza della perfusione tissutale. Una pressione
arteriosa media (MAP) di 70 mmHg, associata a
segni di adeguata perfusione d’organo, può consi-
derarsi in un target ragionevole nella maggioranza
dei pazienti. Nell’approccio alla ipotensione, causa
più frequente di instabilità emodinamica in terapia
intensiva, si possono utilizzare livelli crescenti di
Vol. 68, N. 4 MINERVA ANESTESIOLOGICA
BIGATELLO
monitoraggio. Assumendo che la pressione venosa
centrale (CVP) e la pressione polmonare arteriosa
d’incuneamento (PAOP) siano delle stime adeguate
rispettivamente del volume circolante sistemico e
polmonare, si può suggerire il seguente albero deci-
sionale: 1) fare una diagnosi operativa basata sulla
relazione tra pressioni (CVP, e PAOP) e portata o
gettata cardiaca (CO or SV); 2) prendere in conside-
razione le condizioni che possono alterare l’affida-
bilità della CVP e della PAOP nello stimare adegua-
tamente i volumi circolanti, come alterazioni della
curva pressione/volume (compliance) del RV e del
LV, aumenti della pressione intratoracica (PEEP,
autoPEEP, pressione intra-addominale), malattie
valvolari cardiache (stenosi mitralica); 3) osservare
la storia del paziente; 4)separare la il RV dal LV
osservando le variazioni reciproche di CVP, PAOP e
SV. La CVP è spesso utilizzata come unico parame-
tro nel monitoraggio emodinamico. Tuttavia sulla
base della sola CVP, non è possibile differenziare tra
variazioni in volume (differente curva di ritorno
venoso), e variazioni in contrattilità (differente cur-
va di starling). Infine, altre tecniche, quali l’ecocar-
diografia, il Doppler transesofageo, e i sistemi basa-
ti sul monitoraggio volumetrico, sono attualmente
disponibili.
Parole chiave: Emodinamica - Pressione arteriosa -
Monitoraggio - Pressione venosa centrale.
References
1. Consensus. Tissue hypoxia. How to detect, how to
correct, how to prevent. Am J Respir Crit Care Med
1996;154:1573-8.
2. Kleinman B, Powell S, Kumar P, Gardner R. The fast
flush test measures the dynamic response of the enti-
re blood pressure monitoring system. Anesthesiology
1992;77:1215-20.
3. Guyton A. Venous Return, Handbook of Physiology.
Washington, DC: American Physiological Society,
1963.
4. Jacobsohn E, Chorn R, O’Connor M. The role of the
vasculature in regulating venous return and cardiac
output: historical and graphical approach. Can J
Anaesth 1997;44:849-67.
5. Vitarelli A, Gheorgiade M. Transthoracic and transe-
sophageal echocardiography in the hemodynamic
assessment of patients with congestive heart failure.
Am J Cardiol 2000;86:36G-40.
6. Leone D, Servillo G, De Robertis E, Rossano F, Tufa-
no R. Monitoring cardiac output: esophageal Doppler
vs thermodilution. Minerva Anestesiol 1998;64:351-6.
7. Wallace A, Salahieh A, Lawrence A, Spector K,
Owens C, Alonso D. Endotracheal cardiac output
monitor. Anesthesiology 2000;92:178-89.
8. Neumann P. Extravascular lung water and intrathora-
cic blood volume: double vs single indicator dilution
technique. Intensive Care Med 1999;25:216-9.
9. Connors A, Speroff T, Dawson N, Thomas C, Harrell
F, Wagner D et al. The effectiveness of right heart
catheterization in the initial care of critically ill
patients. JAMA 1996;276:889-97.
225