Stroke volume variation does not predict fluid responsiveness in ...

Acta Anaesthesiol Scand 2006; 50: 1068–1073
Printed in Singapore. All rights reserved
Stroke volume variation does not predict fluid
responsiveness in patients with septic shock
on pressure support ventilation
A. PERNER 1 and T. FABER 2
1 Department of Anaesthesia, Rigshospitalet and 2 Herlev Hospital, University of Copenhagen, Copenhagen, Denmark
Background: Stroke volume variation (SVV) – as measured by
the pulse contour cardiac output (PiCCO Ò ) system – predicts the
cardiac output response to a fluid challenge in patients on
controlled ventilation. Whether this applies to patients on pressure
support ventilation is unknown.
Methods: Thirty consecutive patients with septic shock were
included. All were on pressure support ventilation, monitored
using the PiCCO Ò system and receiving 500 ml of colloid on
clinical indications. Arterial pulse contour SVV and the transpulmonary
thermodilution cardiac index were measured before
and after fluid challenge.
Results: Forty-seven per cent of the patients were defined as fluid
responders by an observed increase of > 10% in the cardiac index
after fluid. Prior to fluid challenge, the cardiac index was lower in
responders compared with non-responders (mean SD, 3.0 0.6
vs. 4.0 1.2 l/min/m 2 , P < 0.01). In contrast, pre-infusion values
THERE are few safe and effective methods to guide
fluid resuscitation. The traditionally used cardiac
filling pressures do not predict fluid responsiveness
(i.e. an increase in cardiac output on fluid challenge) in
critically ill patients (1). Alternative methods, which
assess dynamic changes in cardiac stroke volume
during controlled ventilation, may predict fluid
responsiveness. The pulse contour cardiac output
system (PiCCO Ò ; Pulsion Medical Systems AG,
Munich, Germany) calculates the stroke volume variation
(SVV) by continuous analysis of the arterial
pulse contour. Variations in stroke volume may occur
in the case of cardiac preload reserve, if preload is
altered by the changes in pleural pressures during
ventilation (2). Several studies have shown that
increased SVV is a marker of fluid responsiveness in
critically ill patients (3–10), but all were performed in
patients on strictly controlled mechanical ventilation.
This study was presented in part at the 18th Annual Congress of
the European Society of Intensive Care in Amsterdam 2005.
1068
Copyright # Acta Anaesthesiol Scand 2006
ACTA ANAESTHESIOLOGICA SCANDINAVICA
doi: 10.1111/j.1399-6576.2006.01120.x
of SVV were similar between subsequent responders and nonresponders
(13 5vs.16 6%, P ¼ 0.26). The mean areas under the
ROC curves were 0.77 (95% confidence interval, 0.60–0.94) and
0.52 (0.30–0.73) for pre-fluid cardiac index and SVV, respectively,
indicating a predictive power of only the cardiac index.
Conclusions: SVV did not predict the response in cardiac output
to fluid challenge in patients with septic shock on pressure
support ventilation.
Accepted for publication 15 May 2006
Key words: blood volume; cardiac output; plasma volume
expanders; septic shock; stroke volume.
# Acta Anaesthesiologica Scandinavica (2006)
In patients on pressure support ventilation, heart–
lung interactions are more complex, because the
swings in pleural pressures may be insufficient to
change cardiac preload even in fluid responsive
patients. Moreover, the duration of respiratory cycle
and tidal volumes are not fixed, which potentially
hampers the use of any dynamic marker of fluid
responsiveness. On the other hand, there are no
clinical data to reject the concept of dynamic volume
monitoring in these patients.
The aim of the present study was therefore to
assess if SVV predicted an increase in cardiac output
in response to fluid challenge in patients with septic
shock on pressure support ventilation.
Materials and methods
The ethics committee of Copenhagen County
approved the study. Consent was waived because
all interventions and measurements were clinically
required.

Patients
Patients were enrolled consecutively when meeting
the following inclusion criteria: (i) persistent septic
shock (11) 12 h after the initial fluid resuscitation,
(ii) stable sinus rhythm, (iii) mechanical ventilation via
a cuffed endotracheal tube, (iv) monitored using the
PiCCO Ò system on clinical indication, and (v) due to
receive a fluid bolus as decided by the treating
clinician. The indications for fluid challenge included
signs of hypoperfusion (cold extremities and/or sustained
hyperlactatemia) or sustained need for norepinephrine.
There were no exclusion criteria. All patients
were ventilated with pressure support ventilation
with or without supplementary bilevel positive airway
pressure (BiPAP mode) (Evita XL; Dräger Medical AG,
Lübeck, Germany). If required, patients were sedated
by a continuous propofol infusion supplemented with
fentanyl if necessary. Source control had been obtained
by antibiotics and/or drainage, and all patients
received hydrocortisone (300 mg/day). Norepinephrine
was infused to maintain a mean arterial blood
pressure above 70 mmHg, and if the cardiac index was
below 2.5 l/min/m 2 , dobutamine was administered
to attain this level. Immediately prior to the study, the
motor activity assessment scale (MAAS) (12) and
sequential organ failure assessment (SOFA) scores
(13) were registered.
Protocol
A single bolus of 500 ml of 6% Dextran 70 in saline
was infused over 30 min with the patient positioned
supine. During this period, no changes in ventilation,
sedation or vasoactive treatments were performed,
and no other fluids were given except for nutrition.
Immediately prior to and after the fluid administration,
the cardiac index was measured by transpulmonary
thermodilution by repeated injections of 20 ml
of cold saline (4 8C) into the superior caval vein until
three measurements with a variation of less than 10%
had been obtained. The mean value of these three
measurements was used in the analyses and registered
with the concurrently displayed SVV. As the
SVV value given on the monitor is a running 30-s
average, both the electrocardiogram and the pulse
curve were observed and the SVV of the first 30-s
period of stable cardiac rhythm was registered. This
was done immediately after the thermodilution analysis
to avoid potential drift of the calibration of the
pulse contour analysis. In addition, heart rate, thermodilution
stroke volume and intrathoracic blood
volume indexes, mean arterial blood pressure, ventilator
modus, airway pressures, tidal volume and
the dose of the continuously infused drugs were
registered. Fluid responders were defined as the
patients who had an increase in cardiac index of
more than 10% after the fluid challenge, because the
reproducibility of cardiac index measurements has
been observed to be around 4% (14).
Statistics
The sample size was calculated (2a 0.05 and b 0.2) to
detect a difference in SVV of 5% (SD 5%) between
fluid responders and non-responders. This estimate
was based on previous data in septic patients (3). The
results are expressed as mean SD or 95% confidence
intervals (95% CI) and analyses were carried
out using the Student’s t-tests for unpaired or paired
variables or receiver operating characteristic (ROC)
curve or linear regression analyses where appropriate.
P-values less than 0.05 (two-tailed) were considered
statistically significant.
Results
Patient characteristics are given in Table 1. All patients
received norepinephrine, which was supplemented
with dobutamine in only two patients.
Thirteen patients were sedated with propofol and
seven received fentanyl, six of whom were treated
with both drugs. All patients were ventilated with
a positive end expiratory pressure (PEEP) of 5 cm
H2O or greater and negative airway pressures were
not observed during the study.
Table 1
Patient characteristics.
Gender (male/female) 18 (60)/12 (40)
Age (years) 67 11
SOFA score 13 3
ICU mortality
Source of sepsis
16 (53)
Pulmonary/abdominal/urinary 15 (50)/12 (40)/2 (7)/1 (3)
tract/uncertain
SVV in patients with septic shock
Norepinephrine dose (mg/kg/min) 0.17 0.17
MAAS 1.3 1.0
Propofol 13 (43)
Dosage, mg/h, n ¼ 13 98 97
Fentanyl 7 (23)
Dosage, mg/h, n ¼ 7 0.1 0.03
Ventilator settings
PSV/PSV þ BIPAP 16 (53)/14 (47)
Peak pressure/PEEP, cmH 2O 25 10/11 4
Tidal volume, ml/kg bodyweight 8.7 2.4
Data are numbers of patients (percentages) or means SD, n ¼
30. PSV, pressure support ventilation; BIPAP, bilevel positive
airway pressure; MAAS, motor activity assessment scale; PEEP,
positive end expiratory pressure; SOFA, sequential organ failure
assessment.
1069

A. Perner and T. Faber
Fourteen patients (47%) were fluid responders and
these had a lower baseline cardiac index compared
with the 16 (53%) non-responders (Table 2). In contrast,
SVV prior to fluid challenge was not different in
fluid responders and non-responders (P ¼ 0.26, see
Fig. 1 and Table 2). The area under the ROC curve
(AUC) for the cardiac index was different from
0.5 [mean (95% CI) 0.77 (0.60–0.94)] in contrast
to AUC for SVV, which was 0.52 (0.30–0.73)
(Fig. 2). SVV prior to fluid challenge was unrelated
to the changes in cardiac index after fluid challenge
(Fig. 3).
The administration of fluid increased the mean
arterial blood pressure in both groups whereas the
intrathoracic blood volume index (ITBVI) only
increased in the fluid responders (Table 2). In contrast,
SVV was unaltered by fluid (Table 2).
For SVV, additional ROC curves were constructed
for patients stratified for tidal volumes, PEEP or
motor activity assessment scale (MAAS) above or
below the overall mean values. The AUC’s were 0.52
(0.19–0.85) and 0.63 (0.33–0.92) for tidal volumes
below and above 8 ml/kg, respectively, 0.57 (0.30–
0.83) and 0.58 (0.18–0.97) for PEEP below and above
11 cm H2O, respectively, and 0.64 (0.34–0.94) and 0.58
(0.27–0.89) for MAAS below 2 and MAAS of 2 or
above, respectively.
Discussion
In patients with septic shock on pressure support
ventilation, we found that SVV did not discriminate
between those who would or would not increase the
cardiac index in response to a colloid challenge. In
the present study, neither the group comparisons,
ROC curve nor regression analysis showed any
predictiveness of SVV for the response of the cardiac
index to fluid. Moreover, similar results were obtained
if patients were stratified by sedation scores or
Table 2
ventilation with high tidal volumes or high PEEP.
The only predictor of fluid responsiveness in the
present study was cardiac index. However, the
sensitivity and specificity was low with values
around 70% for a cut-off value of 3.3 l/min/m 2 .
In contrast to our observations, others have found
that the SVV may predict fluid responsiveness. The
reasons for these discrepancies are most likely differences
in ventilation and sedation. The study by Marx
and coworkers (3) is the only one on septic shock. In
10 sedated patients on controlled ventilation, baseline
SVV correlated positively with the change in
cardiac output after fluid challenge. There are several
studies on the predictiveness of SVV in the perioperative
period. Patients have been assessed during
brain surgery (4) and before (5,6,9,10) and after
(7,8,10) cardiac surgery; all these patients were
anaesthetized and on controlled ventilation. The
majority of these studies found a predictive power
of SVV for fluid responsiveness either by group
comparison or correlation analysis. The study by
Wiesenack and colleagues (5) failed to show a relationship
between baseline SVV and the change in
cardiac index upon fluid loading, but a group comparison
was not made in that study. To our knowledge,
the present study is the first to assess the
validity of SVV as a marker of fluid responsiveness
in patients on pressure support ventilation. In addition,
our patients were less sedated, but stratifying
the data by sedation scores did not improve the
predictiveness of SVV. Although the secondary analyses
carried a high risk of a type 2 error, it is likely
that the mode of ventilation with triggering by
spontaneous breathing explains why SVV had no
predictive power in our patients. Other results may
be obtained with the use of higher tidal volumes
and airway pressures, which would induce larger
changes in pleural pressures and thereby cardiac
preload. But again, our secondary analyses of
Haemodynamic data and response to fluid challenge in subsequent fluid ‘responders’ and ‘non-responders’.
Fluid ‘responders’, n ¼ 14 Fluid ‘non-responders’, n ¼ 16
Baseline After fluid Paired t-test Baseline After fluid Paired t-test
Heart rate (beats/min) 92 16 91 17 P ¼ 0.64 98 14 97 13 P ¼ 0.16
Mean arterial blood pressure (mmHg) 70 10 78 12 P ¼ 0.02 69 10 74 10 P ¼ 0.02
Cardiac index (l/min/m 2 ) 3.0 0.6* 3.7 0.6 P < 0.01 4.0 1.2 3.9 1.2 P ¼ 0.10
Stroke volume index (ml/m 2 ) 35 11 42 12 P < 0.01 42 14 41 15 P ¼ 0.35
Intrathoracic blood volume index (ml/m 2 ) 1059 149 1125 204 P < 0.01 1120 152 1099 185 P ¼ 0.43
Stroke volume variation (%) 15 5 12 5 P ¼ 0.08 15 6 13 6 P ¼ 0.07
Data are means SD.
*P < 0.01 compared with the baseline value of ‘non-responders’ using the unpaired Student’s t-test.
1070

Fig. 1. Values of stroke volume variation (SVV) prior to fluid
challenge in patients with septic shock on pressure support
ventilation. Patients who increased more than 10% in the cardiac
index after 500 ml of colloid were categorized as fluid ‘responders’
and those who did not as ‘non-responders’. Dots and bars represent
values of single patients and means, respectively.
patients ventilated with high tidal volumes or PEEP
do not support this notion.
In the present study, all patients had undergone
initial blood volume resuscitation, and this may
explain why values of SVV were lower than previously
observed in patients with sepsis (3). Nevertheless,
half of our patients responded to fluid, and in
these patients the average increase in cardiac index
was 23%. The fact that half the patients had a very
SVV in patients with septic shock
significant increase in cardiac index illustrates the
difficulty of guiding fluid therapy in septic patients
who remain on vasopressors after the initial resuscitation.
A recent study of experimental hemorrhagic
shock showed that norepinephrine may blunt pulse
pressure variation (15), which is a marker of SVV. In
theory, this may not apply to SVV (16), but it is
currently unknown if norepinephrine alters SVV as
measured by pulse-contour analysis. If so, this could
have affected our results.
Intrathoracic blood volume increased in fluid responders,
but was unchanged in non-responders. This
may be as a result of low sensitivity of the single
indicator technique of intrathoracic blood volume
measurement for the detection of changes. Alternatively,
rapid extravasation of fluid may have
occurred in the non-responders. Also, mathematical
coupling may contribute, because intrathoracic blood
volume and cardiac index is calculated from the same
thermodilution curve. On the other hand, it cannot be
excluded that the non-responders had to be filled
more if they were to increase in cardiac output, which
questions the use of a fixed fluid bolus to all patients.
A fixed volume was chosen, because there are no firm
data to support the design of a titration protocol.
Moreover, studies of related methods in patients with
septic shock have used a fixed fluid bolus (17,18).
Interestingly, we found a small increase in blood
pressure on fluid challenge even in patients classified
Fig. 2. Receiver operating characteristic curves for the ability of cardiac index (CI, dashed line with open dots) and stroke volume variation
(SVV, solid line with solid dots) prior to fluid challenge to discriminate between responders and non-responders to fluid challenge. The areas
under the curves were 0.77 (P ¼ 0.01, compared with an area of 0.5) for the cardiac index and 0.51 (P ¼ 0.90) for stroke volume variation
(SVV).
1071

A. Perner and T. Faber
Fig. 3. Relationship in patients with septic shock on pressure support ventilation between stroke volume variation (SVV) prior to fluid
challenge and the subsequent change in cardiac index (Ä-CI) caused by 500 ml of colloid. Dots represent values of single patients. No
statistically significant relationship was found by linear regression analysis (P ¼ 0.86).
as non-responders as observed by others (4,19).
Vascular tone is probably not affected by fluid
loading, suggesting that errors in cardiac output
measurements or spontaneous changes may have
occurred in some patients. Also the stratification
using a fixed cut-off value for the increase in cardiac
index may have contributed. Although there is no
definite answer to explain this unexpected finding,
this may also happen in everyday clinical life. Thus
the addition of independent measures, such as central
venous oxygen saturation, may be of value, but
this should be confirmed in future studies.
Finally, the 30 s used by the PiCCO Ò system to
calculate SVV may include confounding variations in
tidal volumes or airway pressures. If so, assessment
within a single respiratory cycle or averaging of these
confounders by calculating SVV over an even longer
time period may improve its performance in patients
on pressure support ventilation. These matters
deserve further study, but until then we recommend
that clinicians do not use SVV to guide fluid resuscitation
in patients on pressure support ventilation.
References
1. Bendjelid K, Romand JA. Fluid responsiveness in mechanically
ventilated patients: a review of indices used in intensive
care. Intensive Care Med 2003; 29: 352–60.
2. Pinsky MR. Assessment of indices of preload and volume
responsiveness. Curr Opin Crit Care 2005; 11: 235–9.
1072
3. Marx G, Cope T, McCrossan L, Swaraj S, Cowan C, Mostafa
SM et al. Assessing fluid responsiveness by stroke volume
variation in mechanically ventilated patients with severe
sepsis. Eur J Anaesthesiol 2004; 21: 132–8.
4. 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: 984–9.
5. Wiesenack C, Prasser C, Rodig G, Keyl C. Stroke volume
variation as an indicator of fluid responsiveness using pulse
contour analysis in mechanically ventilated patients. Anesth
Analg 2003; 96: 1254–7.
6. Hofer CK, Müller SM, Furrer L, Klaghofer R, Genoni M,
Zollinger A. Stroke volume and pulse pressure variation for
prediction of fluid responsiveness in patients undergoing
off-pump coronary artery bypass grafting. Chest 2005; 128:
848–54.
7. Reuter DA, Felbinger TW, Kilger E, Schmidt C, Lamm P,
Goetz AE. Optimizing fluid therapy in mechanically ventilated
patients after cardiac surgery by on-line monitoring of
left ventricular stroke volume variations. Comparison with
aortic systolic pressure variations. Br J Anaesth 2002; 88: 124–6.
8. Reuter DA, Felbinger TW, Schmidt C, Kilger E, Goedje O,
Lamm P et al. Stroke volume variations for assessment of
cardiac responsiveness to volume loading in mechanically
ventilated patients after cardiac surgery. Intensive Care Med
2002; 28: 392–8.
9. Wiesenack C, Fiegl C, Keyser A, Prasser C, Keyl C. Assessment
of fluid responsiveness in mechanically ventilated
cardiac surgical patients. Eur J Anaesthesiol 2005; 22: 658–65.
10. Preisman S, Kogan S, Berkenstadt H, Perel A. Predicting fluid
responsiveness in patients undergoing cardiac surgery: functional
haemodynamic parameters including the Respiratory
Systolic Variation Test and static preload indicators. Br J
Anaesth 2005; 95: 746–55.
11. American College of Chest Physicians/Society of Critical
Care Medicine Consensus Conference. Definitions for sepsis

and organ failure and guidelines for the use of innovative
therapies in sepsis. Crit Care Med 1992; 20: 864–74.
12. Devlin JW, Boleski G, Mlynarek M, Nerenz DR, Peterson E,
Jankowski M et al. Motor Activity Assessment Scale: a valid
and reliable sedation scale for use with mechanically ventilated
patients in an adult surgical intensive care unit. Crit Care
Med 1999; 27: 1271–5.
13. Vincent JL, Moreno R, Takala J, Willatts S, De MA, Bruining H et
al. The SOFA (sepsis-related organ failure assessment) score to
describe organ dysfunction/failure. On behalf of the Working
Group on sepsis-related problems of the European Society of
Intensive Care Medicine. Intensive Care Med 1996; 22: 707–10.
14. Michard F, Alaya S, Zarka V, Bahloul M, Richard C, Teboul JL.
Global end-diastolic volume as an indicator of cardiac preload
in patients with septic shock. Chest 2003; 124: 1900–8.
15. Nouira S, Elatrous S, Dimassi S, Besbes L, Boukef R,
Mohamed B et al. Effects of norepinephrine on static and
dynamic preload indicators in experimental hemorrhagic
shock. Crit Care Med 2005; 33: 2339–43.
16. Pinsky MR. The dynamic interface between hemodynamic
variables and autonomic tone. Crit Care Med 2005; 33: 2437–8.
17. Tavernier B, Makhotine O, Lebuffe G, Dupont J, Scherpereel P.
Systolic pressure variation as a guide to fluid therapy in
patients with sepsis-induced hypotension. Anesthesiology
1998; 89: 1313–21.
18. 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: 134–8.
19. De Backer D, Heenen S, Piagnerelli M, Koch M, Vincent JL.
Pulse pressure variations to predict fluid responsiveness:
influence of tidal Volume. Intensive Care Med 2005; 31: 517–23.
Address:
Anders Perner
Department of Intensive Care 4131
Rigshospitalet
Blegdamsvej
DK-2100 Copenhagen
Denmark
e-mail: ap@dadlnet.dk
SVV in patients with septic shock
1073