CARDIOVASCULAR ANESTHESIA
SECTION EDITOR
KENNETH J. TUMAN
Stroke Volume Variation as an Indicator of Fluid
Responsiveness Using Pulse Contour Analysis in
Mechanically Ventilated Patients
Christoph Wiesenack, MD, Christopher Prasser,
and Cornelius Keyl, MD
Department of Anesthesia, University Hospital, Regensburg, Germany
Assessment of cardiac performance and adequate fluid
replacement of a critically ill patient are important goals
of a clinician. We designed this study to evaluate the
ability of stroke volume variation (SVV), derived from
pulse contour analysis, and frequently used preload
variables (central venous pressure and pulmonary cap-
illary wedge pressure) to predict the response of stroke
volume index and cardiac index to volume replacement
in normoventilated cardiac surgical patients. We stud-
ied 20 patients undergoing elective coronary artery by-
pass grafting. After the induction of anesthesia, hemo-
dynamic measurements were performed before (T1)
and subsequent to volume replacement by infusion of
6% hydroxyethyl starch 200/0.5 (7 mL/kg) with a rate
A
ssessment of cardiac performance and ade-
quate fluid replacement of a critically ill pa-
tient are important goals of a clinician. Usu-
ally, a pulmonary artery catheter is placed for
assessment of cardiac index (CI) and the estimation
of intravascular volume status. Although cardiac
filling pressures as central venous pressure (CVP)
and pulmonary capillary wedge pressure (PCWP)
are frequently used in clinical practice, they are
often misleading in the assessment of fluid respon-
siveness. Measuring left ventricular end-diastolic
area index by transesophageal echocardiography (1)
or assessment of intrathoracic blood volume index
(2) should provide an accurate estimation of existing
hypovolemia; however, these techniques cannot be
used routinely with most patients.
Systolic pressure variation (SPV), the difference be-
tween the maximal and minimal values of the systolic
Accepted for publication December 3, 2002.
Address correspondence and reprint requests to Christoph
Wiesenack, MD, University Hospital, Department of Anesthesia,
Franz-Josef-Strauss Allee 11, 93052 Regensburg, Germany. Address
e-mail to christoph.wiesenack@klinik.uni-regensburg.de.
DOI: 10.1213/01.ANE.0000053237.29264.01
1254 Anesth Analg 2003;96:12547
SOCIETY OF CARDIOVASCULAR ANESTHESIOLOGISTS
MD, Gabriele Rodig, MD,
of 1 mL kg1 min1. Except for heart rate, all hemo-
dynamic variables changed significantly (P 0.01) af-
ter volume loading. Linear regression analysis between
SVV at baseline (T1) and SVV after volume applica-
tion showed a significant correlation (r 0.97; P
0.01), whereas linear regression analysis between SVV
(T1) and percentage changes of stroke volume index (r
0.19) and cardiac index (r 0.17) did not reveal a
significant relationship between variables. The results
of our study suggest that SVV derived from pulse con-
tour analysis cannot serve as an indicator of fluid re-
sponsiveness in normoventilated cardiac surgical
patients.
(Anesth Analg 2003;96:1254 7)
blood pressure during one mechanical breath, has
been shown to be a valuable variable of cardiac pre-
load in several clinical studies (35). SPV was consid-
ered to be a sensitive indicator of hypovolemia and to
correlate with the response of CI to volume loading
(6 10).
Pulse pressure variation (PPV), defined as the
maximal pulse pressure less the minimum pulse
pressure divided by the average of these two pres-
sures, rather than SPV would more accurately re-
flect changes in left ventricular stroke volume
(LVSV) because it is not influenced by the intratho-
racic pressure-induced changes in arterial pulse
(6 8).
A similar method for continuous assessment of
fluid responsiveness is offered by the recently intro-
duced PiCCO system (Pulsion Medical System, Mu-
nich, Germany). Arterial pulse contour analysis con-
tinuously calculates stroke volume index (SVI) and
displays stroke volume variation (SVV), which is
the change in percentage of SV calculated over the
last 30 s. Comparable to SPV and PPV, SVV should
reflect the ventilation-induced changes in LVSV and
may serve as an indicator of fluid responsiveness,
2003 by the International Anesthesia Research Society
0003-2999/03
ANESTH ANALG CARDIOVASCULAR ANESTHESIA WIESENACK ET AL. 1255
2003;96:1254 7 STROKE VOLUME VARIATION AS AN INDICATOR OF FLUID RESPONSIVENESS

but until now, there has been only limited informa- CI, SVI, SVV, and hemodynamic variables were meas-
tion about the value of this new variable (1115). ured after the induction of anesthesia (T1). After volume
This study was designed to evaluate the ability of replacement by infusion of 6% hydroxyethyl starch
SVV and frequently used preload variables (CVP and (HES) 200/0.5 (7 mL/kg) with a rate of 1 mL
PCWP) to predict the response of SVI and CI to vol- kg1 min1 (mean, 552 mL), a second measurement
ume replacement in normoventilated cardiac surgical was performed. Measurements were achieved in a he-
patients. modynamic steady-state without any application of va-
soactive drugs. SPV and PPV were not calculated in our
investigation.
Methods After the assessment of normal distribution by the
Lilliefors modification of the Kolmogorov-Smirnov
After obtaining approval of the local Ethic Commit-
test, the Students t-test was used to compare vari-
tee and with written informed consent, we studied
ables. Linear regression analysis was performed be-
20 patients (16 men), aged 4373 yr (mean, 60 yr),
tween SVV at baseline (T1) and the percentage values
undergoing elective coronary artery bypass graft-
of changes in SVV (SVV), SVI (SVI), and CI (CI)
ing. Patients with valvular heart disease, intracar-
and between the changes in preload indicating vari-
diac shunts, peripheral vascular disease, preopera-
ables (SVV, CVP, and PCWP) and the changes in
tive dysrhythmias, and an ejection fraction of 50%
preload dependent variables SVI and CI. A P 0.05
were excluded from the study.
was regarded as significant.
Anesthesia was induced with fentanyl 5 g/kg
followed by etomidate until loss of consciousness
and pancuronium 100 g/kg and maintained using
0.5% isoflurane, supplemented with bolus doses of Results
fentanyl up to 20 g/kg and pancuronium 50 g/
Except for heart rate, all hemodynamic variables
kg. Mechanical ventilation without positive end-
changed significantly (P 0.01) after volume loading.
expiratory pressure with a constant tidal volume of
CI, SVI, MAP, CVP, and PCWP increased, whereas
10 mL/kg to an end-tidal Pco2 of 30 35 mm Hg was
SVV and systemic vascular resistance index decreased
maintained at a fraction of inspired oxygen of 0.5
(Table 1).
throughout the study. Peak airway pressures
Linear regression analysis between SVV at baseline
ranged from 14 to 22 mm Hg (mean, 17.6 mm Hg),
(T1) and SVV after volume application shows a lin-
whereas mean airway pressure ranged from 10 to
ear correlation (r 0.97; P 0.01), whereas linear
18 mm Hg (mean, 14.1 mm Hg).
regression analysis between SVV (T1) and percentage
A 4F arterial thermodilution catheter (PiCCO) was
changes of SVI (r 0.19) and CI (r 0.17) did not
inserted via the femoral artery for monitoring of arte-
reveal a significant correlation between variables, as
rial blood pressure (MAP), SVI and SVV assessment
presented in Figure 1.
derived from arterial pulse contour analysis, and for
Linear regression analysis between changes of pre-
intermittent transpulmonary thermodilution CI mea-
load indicating variables SVV, CVP, and PCWP and
surements. After the induction of anesthesia, a 7.5F
changes of the preload-dependent variables SVI and
pulmonary artery catheter (Baxter Healthcare Corpo-
CI did not show a significant correlation among
ration, Irvine, CA) was inserted via an 8.5F introducer
variables.
into the right internal jugular vein for monitoring of
CVP and PCWP (Siemens monitor SC 9000, Erlangen,
Germany).
The arterial catheter was connected to a computer Discussion
for pulse contour analysis (Pulsion Medical System).
Several studies have shown that SPV and PPV were
To calibrate the system, a reference thermodilution CI
valuable indicators of fluid responsiveness during me-
measurement was performed in triplicate to derive a
chanical ventilation (310), whereas cardiac filling
calibration factor and the individual aortic compliance
pressures (CVP and PCWP) were of little help when
(16).
deciding on adequate volume therapy. Accurate mea-
SVV is presented as the change in SV (in percent)
surement of SPV and PPV is difficult, and the tech-
calculated over the last 30 s according to:
nique has not yet been made commercially available.
SVV (SVmax SVmin)/SVmean The new PiCCO system offers continuous assess-
ment of SVV, which is simple to evaluate and may
where SVmax is the mean value of four maximum SVs indicate fluid responsiveness during mechanical ven-
of the last 30 s, SVmin is the mean value of four tilation. Positive-pressure ventilation induces cyclic
minimum SVs of the last 30 s, and SVmean is the mean changes in LVSV, which are mainly related to the
value SVs of the last 30 s. expiratory decrease in left ventricular preload because
1256 CARDIOVASCULAR ANESTHESIA WIESENACK ET AL. ANESTH ANALG
STROKE VOLUME VARIATION AS AN INDICATOR OF FLUID RESPONSIVENESS 2003;96:1254 7

Table 1. Hemodynamic Variables at Sample Points T1 of the inspiratory decrease in right ventricular filling
and T2 and ejection (7).
T1 T2 The results of our study suggest that SVV derived
from pulse contour analysis cannot serve as a variable
HR (bpm) 54.6 8.5 53.9 8.6
MAP (mm Hg) 66.3 9.2 75.3 10.8 *
of fluid responsiveness in mechanically ventilated car-
CVP (mm Hg) 6.9 1.7 9.0 1.9 * diac surgical patients as indicated by the lack of cor-
PCWP (mm Hg) 7.2 2.3 10.8 2.3 * relation between SVV at baseline and SVI after vol-
CI (L min1 m2) 1.88 0.3 2.60 0.5 * ume administration. The high correlation between
SVI (mL/m2) 33.9 7.0 48.9 8.6 * SVV at baseline and SVV only demonstrates a rela-
SVV (%) 13.6 8.1 4.76 1.8 * tionship between the ventilation-associated SVV dur-
SVRI (dyn s cm5 m2) 2855 682 2189 461 *
ing baseline conditions and the amount of decrease in
Values are mean sd. SVV subsequent to the infusion of a defined volume:
HR heart rate; MAP mean arterial blood pressure; CVP central
venous pressure; PCWP pulmonary capillary wedge pressure; CI cardiac the larger the amount of ventilation-associated SVV,
index; SVI stroke volume index; SVV stroke volume variation; SVRI the larger the decrease in SVV, which is of little clinical
systemic vascular resistance index.
* P 0.01. interest. Therefore, SVV can merely predict the
amount of the decrease in SVV after volume correction
considering the SVV value at baseline but not the
extent of the increase in SVI.
According to Berkenstadt et al. (11), an increase in
SVI by 5% or more cannot be expected in patients with
SVV values 9.5% subsequent to volume loading of
100 mL of 6% HES. This is of evident clinical interest
to reduce needless volume replacement, but these
findings cannot be supported by the results of our
study. In our investigation, the increase in SVI after
volume loading was 10% in all patients, although
five patients showed a SVV 9.5% at baseline. Larger
amounts of applied volume can obviously increase
SVI significantly despite SVV values 9.5%, but the
extent of this increase cannot be predicted. In contrast
to the study of Berkenstadt et al. (11), we examined
cardiac surgical patients with potential impaired ven-
tricular function caused by their coronary artery dis-
ease, which could be a possible explanation for the
differences. According to Michard and Teboul (7), the
increase in SV as a result of end-diastolic volume
increase depends on ventricular function because a
decrease in ventricular contractility decreases the
slope of the relationship between end-diastolic vol-
ume and SV.
A recent study of Reuter et al. (15) could demon-
strate the ability of SVV to predict myocardial respon-
siveness to volume administration, but the authors
used large tidal volumes up to 15 mL/kg, which can
possibly affect SVV assessment. SVV, as well as SPV
and PPV, depends on a positive-pressure breath and
therefore could be influenced by tidal volume
(6,10,17,18). As Szold et al. (10) and the respiratory
systolic variation test by Perel (18) demonstrate, in-
creased tidal volumes lead to progressively larger de-
creases in LVSV (6). However, Feissel et al. (12) found
Figure 1. (A) Linear regression analysis between stroke volume that analysis of respiratory changes in aortic blood
variation (SVV) at baseline (T1) and the absolute values of changes velocity and thus in LVSV is an accurate method for
in SVV (SVV) after volume replacement. (B) Linear regression predicting the hemodynamic effects of volume expan-
analysis for comparison between SVV at T1 and the percentage
changes in stroke volume index (SVI) after volume replacement. sion in septic shock patients receiving mechanical ven-
n.s. not significant. tilation using tidal volumes of 8 10 mL/kg. Thus,
ANESTH ANALG
2003;96:1254 7 STROKE VOLUME VARIATION AS AN INDICATOR OF FLUID RESPONSIVENESS
tidal volume issues may not be the main reasons for a
lack of agreement between studies.
Our results stand in contrast to the findings of other
authors (11,12,14,15) who demonstrated that SVV can
serve as a predictor of changes in SVI or CI caused by
volume replacement in mechanically ventilated pa-
tients, but confirms Denault et al. (19) who stated that
much of the power signal of the arterial pulse is al-
tered by ventilation, making the measure of SVV by
pulse contour technique questionable at best because
it uses the power spectral analysis to estimate volume
(13). Pulse contour-derived estimates of SVV calculate
SV from the impedance characteristics of the pulse-
pressure waveform, analyzing only the systolic part of
the arterial pressure wave using a complex algorithm
(16). A potential limitation of the variable constitutes
that SVV calculation has never been validated under
positive-pressure ventilation to reflect actual SVV (13),
just a variation in an impedance modeled variable that
itself may show variation independent of actual
changes in SV.
The possibility that some of the patients in our
study were vasoconstricted such that SVV would be
minimal even in preload-responsive patients (6) has to
be taken into account but could not be supported by
our systemic vascular resistance index data at
baseline.
A major limitation of the study is that other vari-
ables of fluid responsiveness such as SPV, PPV, or
transesophageal echocardiography-derived assess-
ment of aortic blood flow were not measured simul-
taneously with the SVV. Even with this methodolog-
ical limitation, our data suggest that in this setting of
preoperative cardiac surgery, the PiCCO-derived SVV
was not a predictor of preload responsiveness. Nev-
ertheless, further studies are required to confirm these
findings by simultaneously evaluating additional vari-
ables. The determinants of volume responsiveness
seem to be complex and not easily identified by meas-
ures of a single variable, such as arterial pulse
contour-derived estimates of SVV.
References
1. Buhre W, Buhre K, Kazmaier S, et al. Assessment of cardiac
preload by indicator dilution and transoesophageal echocardi-
ography. Eur J Anaesthesiol 2001;18:6627.
CARDIOVASCULAR ANESTHESIA WIESENACK ET AL. 1257
2. Sakka SG, Bredle DL, Reinhart K, Meier-Hellmann A. Compar-
ison between intrathoracic blood volume and cardiac filling
pressures in the early phase of hemodynamic instability of
patients with sepsis or septic shock. J Crit Care 1999;14:78 83.
3. Beaussier M, Coriat P, Perel A, et al. Determinants of systolic
pressure variation in patients ventilated after vascular surgery.
J Cardiothorac Vasc Anesth 1995;9:54751.
4. Rooke GA, Schwid HA, Shapira Y. The effect of graded hemor-
rhage and intravascular volume replacement on systolic pres-
sure variation in humans during mechanical and spontaneous
ventilation. Anesth Analg 1995;80:92532.
5. Tavernier B, Makhotine O, Lebuffe G, et al. Systolic pressure
variation as a guide to fluid therapy in patients with sepsis-
induced hypotension. Anesthesiology 1998;89:131321.
6. Gunn SR, Pinsky MR. Implications of arterial pressure variation
in patients in the intensive care unit. Curr Opin Crit Care
2001;7:2127.
7. Michard F, Teboul JK. Using heart-lung interactions to assess
fluid responsiveness during mechanical ventilation. Crit Care
2000;4:2829.
8. Michard F, Teboul JL. Predicting fluid responsiveness in ICU
patients. Chest 2002;121:2000 8.
9. Perel A. Assessing fluid responsiveness by the systolic pressure
variation in mechanically ventilated patients. Anesthesiology
1998;89:1309 10.
10. Szold A, Pizov R, Segal E, Perel A. The effect of tidal volume
and intravascular volume state on systolic pressure variation in
ventilated dogs. Intensive Care Med 1989;15:368 71.
11. Berkenstadt H, Margalit N, Hadani M, et al. Stroke volume
variation as a predictor of fluid responsiveness in patients un-
dergoing brain surgery. Anesth Analg 2001;92:984 9.
12. Feissel M, Michard F, Mangin I, et al. Respiratory changes in
aortic blood velocity as an indicator of fluid responsiveness in
ventilated patients with septic shock. Chest 2001;119:86773.
13. Pinsky RP. Functional hemodynamic monitoring. Intensive
Care Med 2002;28:386 8.
14. Reuter DA, Felbinger TW, Kilger E, et al. Optimizing fluid
therapy in mechanically ventilated patients after cardiac sur-
gery by on-line monitoring of left ventricular stroke volume
variations: comparison with aortic systolic pressure variations.
Br J Anaesth 2002;88:124 6.
15. Reuter DA, Felbinger TW, Schmidt C, et al. Stroke volume
variations for assessment of cardiac responsiveness to volume
loading in mechanically ventilated patients after cardiac sur-
gery. Intensive Care Med 2002;28:392 8.
16. Goedje O, Hoeke K, Lichtwarck-Aschoff M, et al. Continuous
cardiac output by femoral arterial thermodilution calibrated
pulse contour analysis: comparison with pulmonary artery ther-
modilution. Crit Care Med 1999;27:240712.
17. Michard F, Chemla D, Richard C, et al. Clinical use of respira-
tory changes in arterial pulse pressure to monitor the hemody-
namic effects of PEEP. Am J Respir Crit Care Med 1999;159:
9359.
18. Perel A. Analog values from invasive hemodynamic monitor-
ing. In: Vincent JL, ed. Updates in intensive care and emergency
medicine, applied cardiovascular physiology. Berlin: Springer,
1997:129 40.
19. Denault AY, Gasior TA, Gorscan J 3rd, et al. Determinants of
aortic pressure variation during positive pressure ventilation in
man. Chest 1999;68:1532 6.