Content uploaded by Eric E C de Waal
Author content
All content in this area was uploaded by Eric E C de Waal on Mar 30, 2020
Content may be subject to copyright.
Acta Anaesthesiol Scand 2007; 51: 1258–1267
Printed in Singapore. All rights reserved
#2007 The Authors
Journal compilation #2007 Acta Anaesthesiol Scand
ACTA ANAESTHESIOLOGICA SCANDINAVICA
doi: 10.1111/j.1399-6576.2007.01423.x
Limitations of arterial pulse pressure variation and left
ventricular stroke volume variation in estimating cardiac
pre-load during open heart surgery*
S. REX
1
,G.SCHA
¨LTE
1
,S.SCHROTH
1
,E.E.C.DE WAAL
2
,S.METZELDER
1
,Y.OVERBECK
1
,R.ROSSAINT
1
and W. BUHRE
3
1
Department of Anaesthesiology, University Hospital, Rheinisch-Westfa¨ lische Technische Hochschule Aachen, Aachen, Germany,
2
Division of
Perioperative and Emergency Care, University Medical Center, Utrecht, the Netherlands and
3
University of Witten-Herdecke,
Department of Anaesthesia and Intensive Care Medicine, Hospital Ko¨ln-Merheim, Cologne, Germany
Background: In addition to their well-known ability to predict
fluid responsiveness, functional pre-load parameters, such as
the left ventricular stroke volume variation (SVV) and pulse
pressure variation (PPV), have been proposed to allow real-
time monitoring of cardiac pre-load. SVV and PPV result from
complex heart–lung interactions during mechanical ventilation.
It was hypothesized that, under open-chest conditions, when
cyclic changes in pleural pressures during positive-pressure
ventilation are less pronounced, functional pre-load indicators
may be deceptive in the estimation of ventricular pre-load.
Methods: Forty-five patients undergoing coronary artery bypass
grafting participated in this prospective, observational study. PPV
and SVV were assessed by pulse contour analysis. The thermodilu-
tion technique was used to measure the stroke volume index and
global and right ventricular end-diastolic volume index. Trans-
oesophageal echocardiography was used to determine the left
ventricular end-diastolic area index. All parameters were assessed
before and after sternotomy, and, in addition, after weaning from
cardiopulmonary bypassbefore and afterchest closure (pericardium
left open). Patients were ventilated with constant tidal volumes
(8 2 ml/kg) throughout the study period using pressure control.
Results: SVV and PPV decreased after sternotomy and
increased after chest closure. However, these changes could
not be related to concomitant changes in the ventricular pre-
load. The stroke volume index was correlated with SVVand PPV
in closed-chest conditions only, whereas volumetric indices
reflected cardiac pre-load in both closed- and open-chest con-
ditions. SVV and PPV were correlated with left and right
ventricular pre-load in closed-chest–closed-pericardium condi-
tions only (with the best correlation found for the right ventric-
ular end-diastolic volume index).
Conclusions: SVV and PPV may be misleading when estimating
cardiac pre-load during open heart surgery.
Accepted for publication 31 May 2007
Key words: Cardiac pre-load; functional haemodynamic
monitoring; heart–lung interaction; pulse pressure variation;
sternotomy; stroke volume variation.
#2007 The Authors
Journal compilation #2007 Acta Anaesthesiol Scand
THE optimization of cardiac pre-load is an impor-
tant goal in the haemodynamic management
of cardiac surgical patients. In most cases, arterial
hypotension results from absolute or relative hypo-
volaemia, whereas vigorous fluid loading is asso-
ciated with the risk of volume overload and
pulmonary oedema (1), particularly in patients with
severely depressed ventricular function.
Although several static pre-load indices are rou-
tinely used in cardiac surgical patients, they all ex-
hibit significant shortcomings. Cardiac filling pressures
have been demonstrated to reflect ventricular pre-
load only inaccurately (2, 3), whereas volumetric
variables [e.g. left ventricular end-diastolic area in-
dex (LVEDAI) and global end-diastolic volume index
(GEDI)] cannot be measured continuously, and re-
quire sophisticated, additional monitoring equipment.
Since the pioneering study of Perel et al. in 1987 (4),
increasing amounts of data from numerous recent
reports have suggested that the assessment of dy-
namic pre-load parameters [such as the pulse pres-
sure variation (PPV) and stroke volume variation
(SVV)] may allow the real-time monitoring of cardiac
pre-load (5–9). However, in the majority of these
studies, functional haemodynamic monitoring was
validated by the investigation of the effects of one
*Presented in part at the Annual Meeting of the European
Association of Cardiothoracic Anaesthesiologists, 1–4 June 2005,
Montpellier, France.
1258
single intervention (i.e. volume withdrawal or load-
ing) on haemodynamic variables, making extrapola-
tion to the setting of cardiac surgery difficult. Cardiac
surgical patients are exposed to a variety of peri-
operative factors that, in addition to and independent
of changes in volume status, may affect the reliability
of functional haemodynamic monitoring. In particu-
lar, open-chest conditions and a loss of pericardial
constraint may abate the effects of cyclic changes in
intrathoracic pressure on heart–lung interactions.
Functional pre-load indicators are derived from the
respiratory variations of stroke volume and arterial
pressure typically occurring during mechanical ven-
tilation (1, 10, 11). Briefly, the changes in alveolar and
pleural pressure during each respiratory cycle result
in cyclic alterations of ventricular pre- and after-load
(12–14), affect the ventricular interdependence (15)
and are directly transmitted to the thoracic aorta (16).
For the majority of these mechanisms, it appears
fundamental that changes in intrathoracic pressure
during mechanical ventilation are transmitted to
cardiovascular structures within the thorax. There-
fore, it was hypothesized that, during open-chest
conditions, when changes in pleural pressures during
positive-pressure ventilation are less pronounced,
PPV and SVV may no longer accurately reflect the
cardiac pre-load.
Methods
After approval by the institutional review board
committee and written informed consent had been
obtained, 45 patients participated in the study. All
patients underwent elective coronary artery bypass
grafting with the use of cardiopulmonary bypass
(CPB). Patients with occlusive peripheral arterial
disease, intracardiac shunts, significant valvular
heart disease, arrhythmias, severely decreased left
ventricular function (ejection fraction, 0.30) or
emergency operations were excluded from the study.
Anaesthesia
The patients were pre-medicated with 10 mg of
oxazepam orally in the evening before surgery and
1–2 mg of flunitrazepam 1 h before arrival in the
operating room. Pre-operative medication was con-
tinued until the day of surgery. Anaesthesia was
induced with etomidate (0.1 mg/kg) and sufentanil
(0.5–2 mg/kg). Endotracheal intubation was facili-
tated with rocuronium (1 mg/kg). Anaesthesia was
maintained with a continuous infusion of sufentanil
(2 mg/kg/h) and isoflurane [0.5 minimum alveolar
concentration (MAC)].
Haemodynamic monitoring
Prior to the induction of anaesthesia, a 5-F thermistor-
tipped catheter (PV2015L20A, Pulsiocath, Pulsion
Medical Systems AG, Munich, Germany) was in-
serted into the femoral artery. After the induction of
anaesthesia, a 7.5-F central venous catheter (AG-
15854-E, Arrow International Inc., Reading, PA) and
an 8.5-F introducer sheath (SI-09880, Arrow Interna-
tional Inc.) were placed in the right internal jugular
vein. In a subset of 22 patients, a 7-F pulmonary
artery catheter (PV2047, VoLEF Catheter PACC 947,
Pulsion Medical Systems AG) was inserted into the
pulmonary artery under pressure guidance.
Routine haemodynamic variables (heart rate, mean
arterial and central venous pressures) were recorded
continuously (S/5, Datex-Ohmeda GmbH, Duisburg,
Germany). The arterial thermodilution catheter was
connected to a haemodynamic computer (PiCCO-
plus V 5.2.2, Pulsion Medical Systems AG) for the
assessment of transpulmonary thermodilution curves,
allowing the discontinuous measurement of the car-
diac index, stroke volume index (SVI) and global
GEDI (17). In addition, the arterial pressure, PPV,
pulse contour-derived stroke volume and SVV were
monitored continuously (18). The pulmonary artery
catheter was connected to a second haemodynamic
monitor (VoLEF V 1.0, Pulsion Medical Systems
AG) for the measurement of pulmonary artery pres-
sures and right ventricular end-diastolic volumes
(RVEDVs) (17, 19).
Indicator dilution measurements were performed
by triple bolus injections of 20 ml of ice-cooled saline
0.9% into the right atrium. Injections were randomly
spread over the respiratory cycle. Each value repre-
sents the average of three measurements. The results
were normalized to the body surface area. In ad-
dition, the transpulmonary thermodilution mea-
surements are a prerequisite for the calibration of
pulse contour analysis by the assessment of aortic
impedance (20).
Trans-oesophageal echocardiography (TOE)
A multiplane TOE probe (Omniplane II T6210, Philips
Medical Systems, Eindhoven, the Netherlands), con-
nected to an ultrasonograph (Sonos 5500, Philips
Medical Systems, Eindhoven, the Netherlands), was
positioned to visualize the transgastric short-axis view
of the left ventricle at the level of the mid-papillary
muscles. This position was maintained throughout the
whole study period. Simultaneously acquired TOE
images and electrocardiogram (ECG) signals were re-
corded on a magneto-optical disc, and analysed off-
line by an experienced investigator blind to the
1259
Pulse pressure variation and cardiac surgery
haemodynamic results. LVEDA was measured at the
peak of the electrocardiographic R-wave by manually
tracing the endocardial border including the papillary
muscles (21). The left ventricular end-systolic area was
defined as the smallest left ventricular cavity deter-
mined by visual inspection. For the estimation of
cardiac contractility, the left ventricular fractional area
change was calculated using the standard formula.
For each measurement, an average of at least four
consecutive cardiac beats throughout the respiratory
cycle was evaluated.
Study protocol
All patients were anaesthetized and underwent me-
chanical ventilation in a pressure-controlled mode.
The inspiratory pressure level was adjusted to
achieve a constant tidal volume of approximately
8 ml/kg throughout the procedure. The inspiratory-
to-expiratory time ratio was set at 1 : 1.
Haemodynamic measurements were performed dur-
ing two study periods (sternotomy and chest closure)
with two time points each: before and after sternot-
omy, and before and after sternal closure.
After each intervention, the patients were allowed
to stabilize for 5 min before haemodynamic data
were recorded. Opening of the chest was standard-
ized by retracting the sternum by 12 cm. Measure-
ments after sternotomy were performed whilst the
pericardium was still closed. During chest closure, by
contrast, the pericardium was left open according to
our surgeons’ clinical routine. Vasoactive medication
was not administered prior to CPB. During weaning
from CPB, all patients received vasoactive medica-
tion which was not changed throughout the second
study period. No fluids were administered during
the brief study periods.
Statistics
All data in the tables and figures are presented as
means standard deviations. The results were
analysed statistically using a commercially available
software package (Statistica
#
for Windows Version
6.0, Statsoft, Tulsa, OK). The effects of sternotomy
and chest closure on the haemodynamic variables
were tested using analysis of variance for repeated
measurements (RMANOVA) to take into account the
correlated observations (22). The factor time was
included as a fixed effect. Whenever the analysis
yielded a significant time effect, post hoc testing with
adjustment for multiple comparisons was performed
using Tukey’s honestly significant difference (HSD)
test. Linear regression analysis and Pearson’s prod-
uct moment correlation (r) were used to describe
correlations between different parameters. A level of
P<0.05 was considered to be statistically significant.
Results
The participating patients were aged 63 9 years,
with an average weight of 86 13 kg. The body
surface area was 2.02 0.17 m
2
. One patient suffered
from one-vessel disease, 13 patients from two-vessel
disease and 31 patients from three-vessel disease.
The median number of grafts performed was three
(range, 1–5).
Echocardiographic images of sufficient quality
(defined as more than 75% of the endocardial border
being clearly identifiable) were obtained in 35 pa-
tients. One patient was excluded from the second
study period because of acute right ventricular
failure after weaning from CPB.
Sternotomy was associated with a decrease in PPV
and SVV (Table 1). SVI increased significantly. Con-
comitantly, the left ventricular pre-load was signifi-
cantly increased after sternotomy, as indicated by
GEDI and LVEDAI. Before sternotomy, SVI was
significantly correlated with PPV (Fig. 1A) and all
volumetric indicators of cardiac pre-load (Table 2),
but the correlation with cardiac filling pressures was
not significant (Table 2). There was a weak, but
significant, correlation between PPV and GEDI, LVE-
DAI and the right ventricular end-diastolic volume
index (RVEDI), with the highest correlation found
for the right ventricular pre-load (Fig. 2A–C). After
sternotomy, the correlation between SVI and the
volumetric variables was still maintained (Table 2),
whereas SVI was only weakly related to PPV
(Fig. 1B). Similarly, PPV was no longer correlated
with the volumetric parameters (Fig. 2D–F). More-
over, changes in SVI induced by sternotomy were
significantly correlated with concomitant changes in
the volumetric pre-load indicators, but not to
changes in PPV (Table 3). No correlation could be
found between changes in PPV and changes in GEDI
(r¼0.26, P¼0.09), LVEDAI (r¼0.21, P¼0.21)
and RVEDI (r¼0.33, P¼0.13).
Closure of the chest with the pericardium left open
did not result in changes in either the left or right
ventricular pre-load (Table 1). By contrast, both PPV
and SVV increased significantly. To keep the tidal
volumes constant, inspiratory pressures had to be
increased according to the study protocol (Table 1).
Before and after chest closure, SVI was significantly
correlated with the volumetric parameters of cardiac
pre-load (Table 2), whereas SVI was significantly
1260
S. Rex et al.
correlated to PPV only after closing the sternum
(Fig. 1D). PPV was not correlated with GEDI, LVE-
DAI or RVEDI, either before or after chest closure
(Fig. 3). As during sternotomy, changes in SVI asso-
ciated with chest closure were correlated with
changes in the volumetric pre-load indices, but only
weakly with changes in PPV (Table 3).
In all conditions, PPV was significantly correlated
with SVV (before sternotomy: r¼0.71, P<0.01; after
sternotomy: r¼0.65, P<0.01; before chest closure:
r¼0.41, P¼0.01; after chest closure: r¼0.55,
P<0.01).
Accordingly, correlations between SVV and the
studied pre-load indices in the different conditions
were similar to the correlations found for PPV (data
not shown).
Discussion
Our data show that both PPV and SVV are correlated
with the left and right ventricular pre-load in closed-
chest–closed-pericardium conditions. Consequently,
PPV and SVV may serve as on-line parameters for the
beat-to-beat estimation of the ventricular pre-load in
these conditions. By contrast, PPV and SVV fail to
reflect the cardiac pre-load when used in cardiac
surgical patients with an open chest and/or pericar-
dium. However, there was an acceptable correlation
between SVI and the volumetric parameters of both
left and right ventricular pre-load, irrespective of
whether or not the chest was open.
Numerous studies have found that functional pre-
load indices show a high sensitivity for the detection
of changes in the circulating blood volume (4–9).
Respiratory variations in stroke volume and arterial
pressure are of greater magnitude in hypovolaemic
than in normovolaemic conditions (10). During hy-
povolaemia, both the vena cava (23) and the right
atrium (24) are more compliant and thus more
collapsible. Hence, pleural pressure changes are more
easily transmitted to these structures. Moreover,
underfilling of the pulmonary veins allows a more
pronounced effect of mechanical inspiration on the
right ventricular after-load (10). In addition, hypo-
volaemic ventricles operate on the steep (left) portion
of the Frank–Starling curve, so that a given pre-load
Table 1
Cardiorespiratory variables in the peri-operative time course.
Sternotomy Chest closure
Before After Before After P(RMANOVA)
HR (beats/min) 55 13 56 11 86 11†86 8†<0.01
CI (l/min/m
2
) 2.2 0.5 2.5 0.5 3.5 0.8†3.4 0.8†<0.01
SVI (ml/m
2
)4111 46 12* 41 12 40 10 <0.01
MAP (mmHg) 77 18 81 14 77 10 78 9 0.39
MPAP (mmHg) 22 4224235244 0.27
CVP (mmHg) 12 3123102†12 2‡<0.01
PAOP (mmHg) 14 3123123133 0.09
GEDI (ml/m
2
) 682 128 736 141†669 136 662 110 <0.01
RVEDI (ml/m
2
) 115 23 123 27 119 28 115 28 0.60
LVEDAI (cm
2
/m
2
) 8.31 2.47 9.46 2.60* 8.60 2.3 8.67 2.24 0.02
LV-FAC (%) 54 11 52 11 56 11 53 10 0.09
REF (%) 0.35 0.09 0.37 0.1 0.33 0.08 0.34 0.08 0.39
PPV (%) 12 595†93†11 3‡<0.01
SVV (%) 11 494125155†‡ <0.01
V
T
(ml) 8 2828271 0.26
PAW (mmHg) 18 4184163†19 4‡<0.01
NE (mg/kg/min) – – 0.03 0.03 0.04 0.03
E(mg/kg/min) – – 0.02 0.01 0.02 0.02
CI, cardiac index; CVP, central venous pressure; E, epinephrine; GEDI, global end-diastolic volume index; HR, heart rate; LVEDAI, left
ventricular end-diastolic volume index; LV-FAC, left ventricular fractional area change; M(P)AP, mean (pulmonary) arterial pressure; NE,
norepinephrine; PAOP, pulmonary artery occlusion pressure; PAW, inspiratory airway pressure; PPV, pulse pressure variation; REF, right
ventricular ejection fraction; RMANOVA, analysis of variance for repeated measurements; RVEDI, right ventricular end-diastolic volume
index; SVI, stroke volume index; SVV, stroke volume variation;V
T
, tidal volume.
Data are mean standard deviation.
*P<0.05 vs. before sternotomy.
†P<0.01 vs. before sternotomy.
‡P<0.01 vs. before chest closure.
1261
Pulse pressure variation and cardiac surgery
change (as imposed by ventilation) will induce
a significant change in stroke volume. In conclusion,
volume status is of crucial importance for the mag-
nitude of PPV and SVV. Consequently, arterial pres-
sure waveform analysis has been proposed (7, 9) as
a tool for the on-line monitoring of the ventricular
pre-load. In the present study, we demonstrated
significant correlations between PPV and indices of
the left and right ventricular pre-load, i.e. GEDI,
LVEDAI and RVEDI. Interestingly, the highest cor-
relation was found between PPV and the right ven-
tricular pre-load. This further emphasizes the crucial
importance of venous return, and hence right ventric-
ular pre-load, for the magnitude of PPV/SVV (25).
Accordingly, SVI was significantly correlated with
both PPV and SVV. Hence, PPV and SVV may serve
as beat-to-beat indicators of cardiac pre-load in
closed-chest–closed-pericardium conditions.
After opening the chest, however, both SVV and
PPV were no longer related to the ventricular pre-
load. In addition, SVI was no longer or only very
weakly related to PPVand SVV, both before and after
CPB. By contrast, SVI was still fairly correlated with
the volumetric pre-load indices, highlighting their
role as clinical gold standards for the estimation of
ventricular pre-load (2, 26). These findings suggest
that, in open-chest conditions, PPV and SVV (as
derived from the arterial waveform) may no longer
PPV (%) PPV (%)
0510 15 20 25 30 0 510 15 20 25
SVI (ml/m2)SVI (ml/m2)
0
20
40
60
80
100
0
20
40
60
80
100
PPV
(
%
)
05 10152025 05 10152025
SVI (ml/m2)SVI (ml/m2)
0
20
40
60
80
100
0
20
40
60
80
100
PPV
(
%
)
P < 0.01
r = -0.72
P = 0.04
r = -0.31
P = 0.06
r = -0.28
P < 0.01
r = -0.69
A. Before Sternotomy B. After Sternotomy
C. Before Chest Closure D. After Chest Closure
Fig. 1. Linear correlation analysis of the
relationship between the stroke volume
index (SVI) and arterial pulse pressure
variation (PPV), both before (A) and after
(B) sternotomy, and before (C) and after
(D) chest closure.
Table 2
Correlation coefficients (r) for linear regression analysis of the
relationship between the absolute values of the stroke volume
index and different static indices of ventricular pre-load.
Sternotomy Chest closure
Before After Before After
CVP r–0.27 –0.32 –0.2 0.01
P0.07 0.03 0.19 0.96
PAOP r–0.4 –0.15 0.05 0.15
P0.08 0.53 0.85 0.53
GEDI r0.57 0.47 0.4 0.43
P<0.01 <0.01 <0.01 <0.01
LVEDAI r0.53 0.56 0.77 0.4
P<0.01 <0.01 <0.01 0.01
RVEDI r0.42 0.43 0.51 0.62
P<0.05 <0.05 0.03 <0.01
CVP, central venous pressure; GEDI, global end-diastolic volume
index; LVEDAI, left ventricular end-diastolic area index; PAOP,
pulmonary artery occlusion pressure; RVEDI, right ventricular
end-diastolic volume index.
1262
S. Rex et al.
accurately reflect phasic changes in pre-load, and
hence stroke volume. The aortic impedance has been
found to be reduced in the presence of an increased
intrathoracic pressure (27). Hence, opening the chest
may increase aortic impedance and subsequently
alter the relationship between the stroke volume
and the pulse pressure. Similar observations have
been reported from severe hypovolaemia, where an
increase in aortic compliance results in an altered
relation between the stroke volume and the pulse
pressure (5). An alternative explanation for our
observations may be that factors other than cyclic
changes in the stroke volume could contribute to PPV
(11, 28). Indeed, Denault et al. (16) demonstrated that
ventilation-induced changes in systolic arterial pres-
sure reflected concomitant changes in intrathoracic
pressure rather than cyclic changes in left ventricular
volumes. Changes in intrathoracic pressures during
mechanical ventilation are less pronounced in open-
chest than in closed-chest conditions. Consequently,
and in agreement with other investigators (16, 18, 29,
30), we found that PPV was reduced after sternotomy.
GEDI (ml/m2)
4000 600 800 1000 1200
GEDI (ml/m2)
4000 600 800 1000 1200
PPV (%)
0
5
10
15
20
25
30
PPV (%)
0
5
10
15
20
25
30
PPV (%)
0
5
10
15
20
25
30
0
5
10
15
20
25
30
LVEDAI (cm2/m2)
0 5 10 15
LVEDAI (cm2/m2)
0 5 10 2015
RVEDI (ml/m2)
1000 50 150 200
RVEDI (ml/m2)
1000 50 150 200
PPV (%)
0
5
10
15
20
25
30
PPV (%)
0
5
10
15
20
25
30
PPV (%)
P = 0.01
r = -0.37
P = 0.40
r = -0.13
P = 0.02
r = -0.38
P < 0.01
r = -0.63
P = 0.48
r = -0.11
P = 0.22
r = -0.28
Before Sternotomy After Sternotomy
A
B
C
D
E
F
Fig. 2. Linear correlation analysis of the re-
lationship between the arterial pulse pressure
variation (PPV) and global end-diastolic vol-
ume index (GEDI) (A, D), left ventricular end-
diastolic area index (LVEDAI) (B, E) and right
ventricular end-diastolic volume index (RVEDI)
(C, F), both before and after sternotomy.
1263
Pulse pressure variation and cardiac surgery
However, Reuter et al. (18) attributed this decrease
to a concomitant increase in the left ventricular
pre-load induced by chest opening. The increase in
the left ventricular pre-load could have induced a
rightward shift of the Frank–Starling relation between
the ventricular pre-load and the stroke volume,
thereby decreasing the sensitivity of the heart to
cyclic fluid challenges imposed by positive-pressure
ventilation (10). Our data confirm that opening the
chest was associated with an increase in the left
ventricular pre-load, as indicated by two indepen-
dent volumetric indices, i.e. thermodilution-derived
GEDI and echocardiographically based LVEDAI.
However, the increase in the left ventricular pre-load
induced by sternotomy is not correlated with the
concomitant decrease in PPV or SVV, which may
indicate that the two changes occur independently of
each other. Accordingly, both PPVand SVV increased
after closing the chest, although the biventricular
pre-load remained essentially unaltered and the
ventricles still operated on the same portion of the
Frank–Starling relationship (i.e. no leftward shift).
These findings are in agreement with studies in
which the magnitude of the respiratory variations
in arterial pressure was found to be affected merely
by the decrease in chest wall compliance and inde-
pendent of changes in pre-load (4, 8).
The slope of the Frank–Starling relationship is
another major determinant of PPV and reflects ven-
tricular contractility (10). A flattening of the Frank–
Starling curves, as seen in failing ventricles, is
associated with a decrease in PPV. However, myo-
cardial contractility has been shown to be unaffected
by changes in intrathoracic pressures (31–33). Like-
wise, the parameters used in the present study for the
estimation of myocardial contractility (i.e. fractional
area change and right ventricular ejection fraction)
were not altered by opening or closing the sternum.
In addition, we demonstrated the significance of an
intact pericardium for the dependence of the func-
tional pre-load indicators on ventricular filling. Clos-
ing the sternum with the pericardium left open did
not restore the correlation between PPV/SVV and
GEDI, LVEDAI or RVEDI. Pericardiotomy has been
demonstrated to alter right heart filling (34). More-
over, the loss of pericardial constraint attenuates ven-
tricular interdependence (35) and, subsequently, the
contribution of right heart filling to cyclic changes in
the left ventricular stroke volume (36).
In summary, we speculate that the changes in PPV
and SVV, observed during sternotomy and chest
closure, do not reflect changes in ventricular pre-
load, but are most probably the result of altered in-
trathoracic pressures and chest wall compliance.
One limitation of the present study was that both
SVV and PPV were not tested with regard to their
well-known ability to predict fluid responsiveness
(37) in open-chest conditions. However, we particu-
larly wanted to draw attention to the fact that SVV
and PPV in cardiac surgical patients are affected by
additional factors other than the pre-load. Caution
should be warranted when basing fluid therapy
solely on functional haemodynamic monitoring. An
increase in SVV or PPV, as observed during chest
closure, may not necessarily indicate hypovolaemia
or justify fluid resuscitation.
Tidal volumes were kept constant during the study
period by adjusting the inspiratory pressures. How-
ever, we do not believe that changing the ventilatory
settings in the intra-operative time course signifi-
cantly confounded our results. Two recently pub-
lished clinical studies have demonstrated that both
the pulse pressure and SVV are mainly determined
by the tidal volume and not by inspiratory pressures
(38, 39), thus highlighting the importance of the un-
changed tidal volumes in this study.
Direct measurement of the intrathoracic pressures,
e.g. by assessing the oesophageal pressures, was not
performed. However, this method has been des-
cribed to have severe limitations for estimating
juxtacardiac pressures, complicating conclusions on
the mechanisms of PPV (40). Moreover, we did not
validate the individual respiratory-induced changes
in pulse pressure and stroke volume, as determined
by the PiCCO pulse contour algorithm, against
Table 3
Correlation coefficients (r) for linear regression analysis of the
relationship between changes in the stroke volume index and
corresponding changes (D) in different indices of the ventricular
pre-load.
Sternotomy Chest closure
DCVP r0.09 0.06
P0.58 0.68
DPAOP r0.19 0.4
P0.43 0.09
DGEDI r0.52 0.47
P<0.01 <0.01
DLVEDAI r0.58 0.55
P<0.01 <0.01
DRVEDI r0.57 0.58
P<0.01 <0.01
DPPV r0.28 0.34
P0.06 0.02
CVP, central venous pressure; GEDI, global end-diastolic volume
index; LVEDAI, left ventricular end-diastolic area index; PAOP,
pulmonary artery occlusion pressure; PPV, pulse pressure varia-
tion; RVEDI, right ventricular end-diastolic volume index.
1264
S. Rex et al.
another method that independently measures indi-
vidual stroke volume. Such a study is certainly
warranted for a final validation of the PiCCO algo-
rithm, but was beyond the scope of this investigation.
In conclusion, we have demonstrated that PPV and
SVV are indicative of ventricular pre-load in closed-
chest–closed-pericardium conditions only. In open-
chest conditions, PPV and SVV most probably do not
reflect the cyclic changes in stroke volume. These
limitations must be considered carefully in the hae-
modynamic monitoring and management of cardiac
surgical patients.
Acknowledgements
The Department of Anaesthesiology, University Hospital,
Rheinisch-Westfa¨lische Technische Hochschule Aachen, Aachen,
Germany was in receipt of a research grant from Pulsion Medical
Systems AG, Munich, Germany. W. Buhre is a member of the
advisory board and has received honoraria for lectures from
GEDI (ml/m2)
0 400 600 800 1000 2000
GEDI (ml/m2)
0 400 600 800 1000 2000
PPV (%)
0
5
10
15
20
25
PPV (%)
0
5
10
15
20
PPV (%)
0
5
10
15
20
PPV (%)
0
5
10
15
20
25
30
PPV (%)
0
5
10
15
20
25
30
LVEDAI (cm2/m2)
0 5 10 15
LVEDAI (cm2/m2)
0 5 10 15
RVEDI
(
ml/m2
)
050 100 150 200
RVEDI
(
ml/m2
)
050 100 150 200
PPV (%)
0
5
10
15
20
25
30
P = 0.25
r = -0.18
P = 0.08
r = -0.27
P = 0.52
r = -0.11
P = 0.19
r = 0.29
P = 0.06
r = -0.31
P = 1.00
r = -0.0006
Before Chest Closure After Chest Closure
A
B
C
D
E
F
Fig. 3. Linear correlation analysis of the re-
lationship between the arterial pulse pressure
variation (PPV) and global end-diastolic vol-
ume index (GEDI) (A, D), left ventricular
end-diastolic area index (LVEDAI) (B, E) and
right ventricular end-diastolic volume index
(RVEDI) (C, F), both before and after chest
closure.
1265
Pulse pressure variation and cardiac surgery
Pulsion Medical Systems AG. Financial support for the present
study was provided solely from institutional and departmental
sources. None of the authors has any financial interest in the
equipment used in the study.
References
1. Bendjelid K, Romand JA. Fluid responsiveness in mechan-
ically ventilated patients: a review of indices used in intensive
care. Intensive Care Med 2003; 29: 352–60.
2. Kumar A, Anel R, Bunnell E et al. Pulmonary artery occlu-
sion pressure and central venous pressure fail to predict
ventricular filling volume, cardiac performance, or the
response to volume infusion in normal subjects. Crit Care
Med 2004; 32:691–9.
3. Buhre W, Weyland A, Schorn B et al. Changes in central
venous pressure and pulmonary capillary wedge pressure do
not indicate changes in right and left heart volume in patients
undergoing coronary artery bypass surgery. Eur J Anaesthesiol
1999; 16: 11–7.
4. Perel A, Pizov R, Cotev S. Systolic blood pressure variation is
a sensitive indicator of hypovolemia in ventilated dogs
subjected to graded hemorrhage. Anesthesiology 1987; 67:
498–502.
5. Berkenstadt H, Friedman Z, Preisman S et al. Pulse pressure
and stroke volume variations during severe haemorrhage in
ventilated dogs. Br J Anaesth 2005; 94: 721–6.
6. Fujita Y, Yamamoto T, Sano I et al. A comparison of changes in
cardiac preload variables during graded hypovolemia and
hypervolemia in mechanically ventilated dogs. Anesth Analg
2004; 99: 1780–6.
7. Preisman S, DiSegni E, Vered Z, Perel A. Left ventricular
preload and function during graded haemorrhage and re-
transfusion in pigs: analysis of arterial pressure waveform
and correlation with echocardiography. Br J Anaesth 2002; 88:
716–8.
8. Tournadre JP, Allaouchiche B, Cayrel V et al. Estimation of
cardiac preload changes by systolic pressure variation in pigs
undergoing pneumoperitoneum. Acta Anaesthesiol Scand
2000; 44: 231–5.
9. Preisman S, Pfeiffer U, Lieberman N, Perel A. New monitors
of intravascular volume: a comparison of arterial pressure
waveform analysis and the intrathoracic blood volume.
Intensive Care Med 1997; 23: 651–7.
10. Michard F. Changes in arterial pressure during mechanical
ventilation. Anesthesiology 2005; 103: 419–28.
11. Magder S. Clinical usefulness of respiratory variations in
arterial pressure. Am J Respir Crit Care Med 2004; 169: 151–5.
12. Scharf SM, Caldini P, Ingram RH Jr. Cardiovascular effects of
increasing airway pressure in the dog. Am J Physiol 1977; 232:
H35–43.
13. Brower R, Wise RA, Hassapoyannes C et al. Effect of lung
inflation on lung blood volume and pulmonary venous flow.
J Appl Physiol 1985; 58: 954–63.
14. Fessler HE, Brower RG, Wise RA, Permutt S. Mechanism of
reduced LV afterload by systolic and diastolic positive pleural
pressure. J Appl Physiol 1988; 65: 1244–50.
15. Taylor RR, Covell JW, Sonnenblick EH, Ross J Jr. Dependence
of ventricular distensibility on filling of the opposite ventricle.
Am J Physiol 1967; 213: 711–8.
16. Denault AY, Gasior TA, Gorcsan J III et al. Determinants of
aortic pressure variation during positive-pressure ventilation
in man. Chest 1999; 116: 176–86.
17. Godje O, Peyerl M, Seebauer Tet al. Reproducibility of double
indicator dilution measurements of intrathoracic blood vol-
ume compartments, extravascular lung water, and liver
function. Chest 1998; 113: 1070–7.
18. Reuter DA, Goresch T, Goepfert MS et al. Effects of mid-line
thoracotomy on the interaction between mechanical ventila-
tion and cardiac filling during cardiac surgery. Br J Anaesth
2004; 92: 808–13.
19. Hofer CK, Furrer L, Matter-Ensner S et al. Volumetric preload
measurement by thermodilution: a comparison with trans-
oesophageal echocardiography. Br J Anaesth 2005; 94: 748–55.
20. Godje O, Hoke K, Goetz AE et al. Reliability of a new
algorithm for continuous cardiac output determination by
pulse-contour analysis during hemodynamic instability. Crit
Care Med 2002; 30: 52–8.
21. Liu N, Darmon PL, Saada M et al. Comparison between
radionuclide ejection fraction and fractional area changes
derived from transesophageal echocardiography us-
ing automated border detection. Anesthesiology 1996; 85:
468–74.
22. Ludbrook J. Repeated measurements and multiple com-
parisons in cardiovascular research. Cardiovasc Res 1994; 28:
303–11.
23. Barbier C, Loubieres Y, Schmit C et al. Respiratory changes in
inferior vena cava diameter are helpful in predicting fluid
responsiveness in ventilated septic patients. Intensive Care
Med 2004; 30: 1740–6.
24. Santamore WP, Amoore JN. Buffering of respiratory varia-
tions in venous return by right ventricle: a theoretical analy-
sis. Am J Physiol 1994; 267: H2163–70.
25. Vignon P. Evaluation of fluid responsiveness in ventilated
septic patients: back to venous return. Intensive Care Med 2004;
30: 1699–701.
26. Hoeft A, Schorn B, Weyland A et al. Bedside assessment of
intravascular volume status in patients undergoing coronary
bypass surgery. Anesthesiology 1994; 81: 76–86.
27. Pinsky MR, Matuschak GM, Bernardi L, Klain M. Hemody-
namic effects of cardiac cycle-specific increases in intratho-
racic pressure. J Appl Physiol 1986; 60: 604–12.
28. Pinsky MR. Probing the limits of arterial pulse contour
analysis to predict preload responsiveness. Anesth Analg
2003; 96: 1245–7.
29. Abel JG, Salerno TA, Panos A et al. Cardiovascular effects of
positive pressure ventilation in humans. Ann Thorac Surg
1987; 43: 198–206.
30. De Blasi RA, Palmisani S, Cigognetti L et al. Effects of
sternotomy on heart–lung interaction in patients undergoing
cardiac surgery receiving pressure-controlled mechanical
ventilation. Acta Anaesthesiol Scand 2007; 51: 441–6.
31. Katzenberg C, Olajos M, Morkin E, Goldman S. Effects of
changes in airway pressure on the left ventricle and left
atrium of dogs. Cardiovasc Res 1986; 20: 853–62.
32. Pizov R, Ya’ari Y, Perel A. The arterial pressure waveform
during acute ventricular failure and synchronized external
chest compression. Anesth Analg 1989; 68: 150–6.
33. Pizov R, Cohen M, Weiss Y et al. Positive end-expiratory
pressure-induced hemodynamic changes are reflected in the
arterial pressure waveform. Crit Care Med 1996; 24: 1381–7.
34. Wranne B, Pinto FJ, Hammarstrom E et al. Abnormal right
heart filling after cardiac surgery: time course and mecha-
nisms. Br Heart J 1991; 66: 435–42.
35. Baker AE, Dani R, Smith ER et al. Quantitative assessment
of independent contributions of pericardium and septum
to direct ventricular interaction. Am J Physiol 1998; 275:
H476–83.
1266
S. Rex et al.
36. Shaver JA, Reddy PS, Curtiss EI et al. Noninvasive/invasive
correlates of exaggerated ventricular interdependence in
cardiac tamponade. J Cardiol 2001; 37 (Suppl. 1): 71–6.
37. Rex S, Brose S, Metzelder S et al. Prediction of fluid respon-
siveness in patients during cardiac surgery. Br J Anaesth 2004;
93: 782–8.
38. De Backer D, Heenen S, Piagnerelli M et al. Pulse pressure
variations to predict fluid responsiveness: influence of tidal
volume. Intensive Care Med 2005; 31: 517–23.
39. Reuter DA, Bayerlein J, Goepfert MS et al. Influence of tidal
volume on left ventricular stroke volume variation measured
by pulse contour analysis in mechanically ventilated patients.
Intensive Care Med 2003; 29: 476–80.
40. Marini JJ, O’Quin R, Culver BH, Butler J. Estimation of
transmural cardiac pressures during ventilation with PEEP.
J Appl Physiol 1982; 53: 384–91.
Address:
Dr Steffen Rex
Klinik fu¨ r Ana¨ sthesiologie und Fachu¨ bergreifende
Klinik fu¨ r Operative Intensivmedizin Erwachsene
Universita¨ tsklinikum der RWTH Aachen
Pauwelsstr. 30
D-52074 Aachen
Germany
e-mail: srex@ukaachen.de
1267
Pulse pressure variation and cardiac surgery
