Original article 295
Continuous mixed venous and central venous oxygen
saturation in cardiac surgery with cardiopulmonary bypass
Pierre-Yves Lequeux, Yves Bouckaert, Hicham Sekkat, Philippe Van der Linden, Constantin Stefanidis,
Chi-Hoang Huynh, Gilbert Bejjani and Philippe Bredas
Background and objective Replacing mixed venous oxygen
saturation (SvO2) monitoring by central venous oxygen
saturation (ScvO2) monitoring in order to avoid the use of a
pulmonary artery catheter and its related complications is
still controversial in the setting of cardiac surgery. The influence
of surgery, cardiopulmonary bypass and anaesthesia drugs on
the relationship between SvO2 and ScvO2 has never been
studied.
Methods Fifteen patients scheduled for cardiac surgery with
cardiopulmonary bypass were included in the study. SvO2 (from
the pulmonary artery) and ScvO2 (from the superior vena cava)
were continuously measured with fibre-optic catheters from
induction of anaesthesia to 24 h postoperatively.
Results A total of 9267 pairs of measurements were recorded.
Mean bias between SvO2 and ScvO2 was 4.4% with limits of
Introduction
Mixed venous oxygen saturation (SvO2) is routinely
used to assess the oxygen supply/uptake balance in
patients undergoing cardiac surgery with cardiopulmon-
ary bypass (CPB).1 However, SvO2 measurements can
only be performed with a pulmonary artery catheter
which also measures pulmonary pressures and cardiac
output.2 These catheters are associated with potential
severe complications and their usefulness in the manage-
ment of haemodynamically unstable patients is still
debated.3–5 Cardiac output measurements can be per-
formed without pulmonary artery catheters with less
invasive monitoring.6 Central venous oxygen saturation
(ScvO2), however, can be obtained from a central venous
catheter that is easier to place and has fewer compli-
cations than a pulmonary artery catheter. The difference
between both measurements is that ScvO2 analyses blood
only from the superior vena cava, whereas SvO2 also takes
account of blood from the inferior vena cava and the
coronary sinus.
Several studies investigated the correlation between SvO2
and ScvO27–19 with contradictory results. Some of these
studies were performed in cardiac surgery patients7–11 but
only in the preoperative (at initial catheter placement)11 or
in the postoperative period.7–10 Surprisingly, the intra-
From the Department of Anesthesiology (P-YL, PB), Intensive Care Unit, CHU-
Tivoli, La Louviere (YB), Department of Anesthesiology, CHU-Brugmann, Brussels
(HS, PVdL), Department of Cardiac Surgery, CHU-Tivoli, La Louviere (CS, C-HH)
and Department of Anesthesiology, Erasme Hospital, ULB, Brussels (GB),
Belgium
Correspondence to Dr Pierre-Yves Lequeux, MD, Department of Anesthesiology,
CHU-Tivoli, La Louviere, Belgium
Tel: +32 64276111; fax: +32 64276529; e-mail: pilequeu@ulb.ac.be
0265-0215 ß 2010 Copyright European Society of Anaesthesiology
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
agreement of
13.6 and R22.5%, respectively. Trends of SvO2
and ScvO2 values followed very different patterns for some
patients. Surgery, cardiopulmonary bypass and anaesthesia
drugs did not influence the relationship between the two
methods.
Conclusion Because of the large interindividual variability in the
difference between SvO2 and ScvO2, the measure of ScvO2
should not replace the measure of SvO2 with a pulmonary artery
catheter for the management of patients undergoing cardiac
surgery with cardiopulmonary bypass.
Eur J Anaesthesiol 2010;27:295–299
Keywords: cardiac, central venous oxygen saturation, mixed venous oxygen
saturation, oximetry, surgery
Received 12 May 2009 Revised 27 July 2009
Accepted 29 July 2009
operative period, which is characterized by rapid and
profound haemodynamic changes, has never been studied.
In addition, the effect of CPB on the SvO2 –ScvO2 corre-
lation remains largely unexplored.
Most of the published studies compared paired measure-
ments of SvO2 and ScvO2 obtained from blood samples,
but only three of them used fibre-optic measure-
ments.14,15,18 Using continuous fibre-optic measure-
ments14,18 is more similar to clinical practice and allows
for a large number of comparisons.
Factors that influence the correlation between the two
measurements remain largely undefined. The haemato-
crit level,14 the blood pH,14 the patient’s temperature14
and even the cardiac index17 do not seem to influence this
relationship. During surgical procedures, other factors
may influence the correlation between SvO2 and ScvO2,
such as the use of anaesthetic drugs. Indeed, anaesthesia
induces a decrease in cerebral oxygen consumption and
to a lesser extent in cerebral blood flow, resulting in a
positive balance between oxygen supply and oxygen
demand from the brain compared with the awake state,20
which may influence the correlation between SvO2 and
ScvO2. This effect of anaesthetic drugs on the SvO2 –
ScvO2 relationship has never been studied.
The goal of this study was to evaluate the correlation
and agreement between continuous fibre-optic SvO2
and ScvO2 measurements in order to evaluate their
interchangeability during cardiac surgery with CPB
and to assess the still unexplored effects of anaesthesia,
surgery and CPB on the relationship between the two
methods.
DOI:10.1097/EJA.0b013e3283315ad0
 296 Lequeux et al.
Methods
Patient population
After approval from institutional review board and
patients’ written and informed consent, 15 patients
scheduled for cardiac surgery with CPB were enrolled
in this experiment.
Exclusion criteria included contraindication to pulmon-
ary artery catheter placement or age less than 18 years.
Protocol
Midazolam 5 mg was given orally 30 min before induction
of anaesthesia. Upon arrival in the operating room, rou-
tine monitoring for cardiac anaesthesia was placed includ-
ing five electrodes – electrocardiography, pulse oximetry,
noninvasive blood pressure cuff, capnography, bispectral
index (BIS) monitoring, 18-G femoral artery catheter for
invasive blood pressure monitoring (Arrow International
Inc., Reading, Pennsylvania, USA), urine output monitor-
ing, rectal temperature and transoesophageal echocardio-
graphy. A 16-G catheter (BD Venflon Pro, Becton Dick-
inson, Helsinborg, Sweden) was inserted in a left forearm
peripheral vein. A spinal anaesthesia was performed with
an intrathecal injection of 20 mg sufentanil and 0.5 mg
morphine. Anaesthesia was induced and maintained with
a target-controlled infusion (TCI) of propofol and remi-
fentanil in order to keep BIS values between 40 and
60. Neuromuscular block was achieved with cisatra-
curium. The trachea was then intubated and the lungs
were ventilated with a mixture of air and oxygen. A 7F,
20 cm long three-lumen central venous catheter (Arrow
International Inc.) and a 7.5F pulmonary artery catheter
(Swan-Ganz CCO/SvO2; Edwards Lifesciences, Irvine,
California, USA) through a 8.5F introducer (Intro-Flex;
Edwards Lifesciences) were then inserted through the
right internal jugular vein. Correct tip placement of both
central venous (1–2 cm above the right atrium) and
pulmonary artery catheters was confirmed with the ultra-
sonography microbubbles test and postoperative chest
radiography. The ultrasonography microbubbles test con-
sisted of injecting a solution of saline mixed with air
(obtained by rapidly transferring saline from a syringe to
another syringe filled with air) through the catheter in
order to visualize its tip. The pulmonary artery catheter
was connected to the Vigilance monitor (Edwards Life-
sciences) and a CeVOX optic fibre (Pulsion Medical
Systems, Munich, Germany) was inserted through the
distal lumen of the central venous catheter and connected
to the CeVOX monitor (Pulsion Medical Systems). SvO2
and ScvO2 values were recorded every minute for 24 h
except for the CPB period with the Multi Data Logger
software (Edwards Lifesciences) for the SvO2 values and
with the CeVOX Win software (Pulsion Medical Sys-
tems) for the ScvO2 values, both running on Windows
XP Professional (Microsoft Corporation, Redmond,
Washington, USA). The Vigilance (Edwards Life-
sciences) and CeVOX (Pulsion Medical Systems)
European Journal of Anaesthesiology 2010, Vol 27 No 3
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
monitors were calibrated after initial catheter placement,
after CPB and upon arrival in the ICU with concomitant
blood samples (1 ml) from the pulmonary artery and the
central venous catheters after withdrawal of dead space
flushing fluid. Blood samples were analysed with the
Bayer Blood Gas Analyzer Rapidlab 865 (Bayer Health-
care LLC, East Walpole, Massachusetts, USA) allowing
direct measurements of blood oxygen saturation. CPB
was performed under normothermia with 1800 ml mixed
crystalloid/colloid solution and myocardial preservation
with 800 ml cold blood cardioplegia (Kalium 30 mEq/l).
Transfusion triggers were classically haematocrit of
21 during surgery, 18 during CPB and 27 in the post-
operative period. These were not the only triggers used as
the decision to carry out transfusion was also based on
other parameters such as SvO2 or lactate level. Postopera-
tive sedation was performed with propofol and remifenta-
nil. Remifentanil infusion was gradually decreased and
shifted towards a morphine infusion titrated according to
patients’ needs. Propofol infusion was stopped when
patients were deemed ready to recover from anaesthesia.
Time at spontaneous eyes opening was recorded in
the ICU.
The intraoperative period was defined as the period from
induction of anaesthesia to the end of surgery and the
postoperative period as that from arrival to discharge from
the ICU. The effect of surgery was analysed by compar-
ing intraoperative and postoperative measurements; the
effect of CPB was analysed by comparing prebypass and
postbypass measurements; and the effect of the anaes-
thetic drugs was analysed by comparing measurements
obtained before patients opened their eyes with those
obtained after patients opened their eyes in the ICU.
Statistics
Correlation and agreement between SvO2 and ScvO2
measurements were calculated with the method
described by Bland and Altman,21 with limits of agree-
ment as the mean difference  1.96 SD.
Mean bias and SD between SvO2 and ScvO2 values for
each period (intraoperative and postoperative, pre-CPB
and post-CPB, anaesthetized and awake) were calculated
for each patient from the Bland and Altman results
(SD ¼ limit of agreement
bias/1.96). Data obtained
during the different periods were compared with a paired
Student’s t-test.
Results
The central venous catheter was misplaced for one patient
as confirmed by the intraoperative ultrasonography micro-
bubbles test and postoperative chest radiography. Results
from this patient were not analysed. Results for the pre-
CPB period from one patient and from the ICU period for
another could not be recorded for technical reasons. One
patient died intraoperatively; therefore, only intraopera-
tive measurements could be recorded.

Table 1 Patients characteristics
Weight (kg)
Height (cm)
Age (years)
Sex (male/female)
CPB duration (min)
Aortic clamping duration (min)
Surgery duration (min)
Data are expressed as mean  SD or ratio (sex). CPB, cardiopulmonary bypass.
Patient characteristics are presented in Table 1.
A total of 9267 pairs of measurements were recorded
(354 for the intraoperative pre-CPB period, 768 for the
intraoperative post-CPB period, 2834 for the postopera-
tive anaesthetized period and 5311 for the postoperative
awake period). The mean bias between SvO2 and ScvO2
(for the 9267 pairs of measurement) was 4.4% and the
limits of agreement were
13.6 and þ22.5%, respectively
(Fig. 1). The results of the Bland and Altman tests for
each patient are displayed in Fig. 2. The mean bias
between the two methods was >3% in 7284 out of
9267 measurements (79%). There was no statistical
difference for SvO2 –ScvO2 mean bias and SD between
the intraoperative and postoperative period, the pre-CPB
and post-CPB period or the anaesthetized and awake
state period (Table 2).
Discussion
Continuous SvO2 measurement is a classical method of
monitoring the oxygen supply/uptake balance during
cardiac surgery with CPB.1 Such measurements are
performed with a fibre-optic catheter inserted in the
pulmonary artery.2 These catheters may also measure
Fig. 1
Total Bland and Altman SvO2 –ScvO2 comparison. Bland and Altman
comparison between continuous ScvO2 from a central venous fibre-
optic catheter and continuous SvO2 from a pulmonary artery fibre-optic
catheter. The solid line represents the difference between both
measurements (mean bias) and the dotted line represents precision
(95% limits of agreement).
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
SvO2 vs. ScvO2 during cardiac surgery 297
Fig. 2
83  22
168  8
66  16
10/5
120  46
80  34
245  71
Per patient SvO2 –ScvO2 mean bias and 95% limits of agreement.
SvO2 –ScvO2 mean bias and 95% limits of agreement were calculated
by the methods described by Bland and Altman.
thermodilution cardiac output. But the placement of
these catheters is associated with potential severe com-
plications,3 accounting for the popularity of less invasive
methods of cardiac output monitoring which measure
cardiac output without a pulmonary artery catheter.6
Because of this, it has been proposed to replace SvO2
measurements by central venous oxygen saturation
(ScvO2), which can be obtained from a central venous
catheter that is easier to place and has fewer compli-
cations. Animal studies suggest that ScvO2 correlates well
with SvO2 and could reliably replace it.22,23
Human studies, however, are more contradictory. ScvO2
seems to correlate with postoperative complications24,25
and its use for early goal-directed therapy has been shown
to reduce mortality in septic shock26 or to reduce post-
operative complications and length of hospital stay when
used perioperatively.27 These studies, however, did not
assess interchangeability of both methods but only the
usefulness of ScvO2. Studies comparing ScvO2 and SvO2
are quite contradictory; some report an acceptable bias
between the two methods,15,17,19 whereas others do
not.8–13,16,18 An acceptable bias should be a change in
SvO2 values less than that which would induce a thera-
peutic change. Bendjelid et al.9 considered 3% to be the
maximal acceptable bias. But other authors consider the
bias between the two methods of minor importance,
arguing that ScvO2 could replace SvO2, provided that
Table 2 Effect of surgery, cardiopulmonary bypass and
anaesthesia drugs on SvO2 –ScvO2 bias and SD
Mean bias  SD (%) P
Intraoperative period 3.6  6.0
Postoperative period 4.6  6.5 0.40
Pre-CPB period 3.6  3.1
Post-CPB period 2.0  4.9 0.45
Anaesthetized state period 8,2  3,7
Awake state period 3.7  4,6 0.18
Awake state period was defined as beginning at spontaneous eye opening in the
ICU. CPB, cardiopulmonary bypass.
European Journal of Anaesthesiology 2010, Vol 27 No 3
 298 Lequeux et al.
Fig. 3
SvO2 –ScvO2 24 h time course: example of two patients. (A) Example of trends of SvO2 and ScvO2 values showing a similar pattern with a large
bias (11.2%). (B) Example of trends of SvO2 and ScvO2 values showing different patterns with a small bias (
2.8%). (a) Intraoperative
precardiopulmonary bypass period. (b) Intraoperative postcardiopulmonary bypass period. (c) Postoperative anaesthetized period. (d) Postoperative
awake period.
the trends of both methods show a similar pattern what-
ever the bias.13,14
Our results show a mean bias of 4.4 and 95% limits of
agreement of
13.6 and þ22.5%, and a mean bias of over
3% for 79% of the time, thus far above the maximal
tolerable difference described by Bendjelid et al.9 In
addition, our results reveal that some patients show,
although sometimes with a very large mean bias, a
remarkably comparable pattern in the trends of the
two methods (Fig. 3a), which suggests that ScvO2 can
be used to adequately replace SvO2, whereas other
patients show trends of SvO2 and ScvO2 of very different
patterns. In the example presented in Fig. 3b, ScvO2
remains very stable, around 70% during 24 h, whereas
SvO2 varies widely (down to 45%) over the same period,
which calls into question the reliability of ScvO2 in
detecting changes in tissue oxygenation in the setting
of cardiac surgery at least for some patients.
The present study also demonstrates a large interindivi-
dual variability in the bias between SvO2 and ScvO2 that
has never been emphasized in previous studies. Indeed,
mean bias between SvO2 and ScvO2 for each patient
varies from
7.1 to þ16.8%, with large 95% limits of
agreement for each patient (Table 2). Because the bias
varies widely from one patient to another and even for a
single patient over the course of time, it is almost impos-
sible to predict a SvO2 value from a ScvO2 value even if
the mean difference between the two methods is known.
However, the usefulness of ScvO2 monitoring in the
setting of cardiac surgery still remains to be investigated.
European Journal of Anaesthesiology 2010, Vol 27 No 3
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
Indeed, our study was designed to compare ScvO2 and
SvO2 values but not to evaluate the efficacy of ScvO2
monitoring to guide haemodynamic therapeutics along
with other parameters (preload parameters, cardiac output,
blood pH, haemoglobin, lactate levels, etc.). Our results
suggest that SvO2 and ScvO2 values are not equivalent, but
a randomized trial comparing cardiac surgery patients
monitored with either SvO2 or ScvO2 measurements is
required to evaluate the impact of both types of monitoring
on clinical outcomes and to better establish their respect-
ive roles in the management of cardiac surgery patients.
Several other studies have been performed in cardiac
surgery patients. Most of them compared SvO2 and
ScvO2 during the postoperative period7–10 and one during
the preoperative period only (at catheter placement).11 All
concluded that both methods are not interchangeable.
However, none of them studied the intraoperative period,
which is characterized by wide haemodynamic changes
that could influence the relationship between SvO2 and
ScvO2. Our results, however, suggest that neither surgery
nor CPB significantly influences the relationship between
SvO2 and ScvO2.
Anaesthesia is another factor that could influence this
SvO2 –ScvO2 relationship. Indeed, most anaesthetic drugs
induce a greater decrease in cerebral oxygen consump-
tion than in cerebral blood flow, resulting in a positive
balance between brain oxygen supply and oxygen uptake
compared with the awake state.20 The higher oxygen
saturation in the blood from the superior vena cava in
anaesthetized patients than in awake patients could have
an effect on the difference between ScvO2 and SvO2, but

our results do not demonstrate it. Factors influencing the
correlation between SvO2 and ScvO2 thus remain largely
unknown, as haematocrit, blood pH, patient temperature14
and cardiac index17 did not seem to affect this relation-
ship either.
A limit of the present study is the small number of patients
included. However, continuous fibre-optic measurements
allowed for the collection of large amounts of data per
patient and a large number of paired SvO2 –ScvO2
measurements (9267). In addition, the sample was suffi-
cient to demonstrate a large interindividual variability in
the bias between SvO2 and ScvO2.
In conclusion, our study demonstrates that the bias
between SvO2 and ScvO2 varies widely from one patient
to another and for a single patient over the course of time
and also that the trends in the two methods follow very
different patterns for some patients. In addition, as factors
such as anaesthesia, surgery and CPB as well as haema-
tocrit, blood pH, patient temperature and cardiac index
do not seem to influence the relationship between SvO2
and ScvO2, the difference observed between the two
methods still remains unexplained. Therefore, in patients
undergoing cardiac surgery with CPB, the value of ScvO2
obtained from a central venous catheter should not be used
to assess the value of SvO2 obtained from a pulmonary
artery catheter. Further investigations are required to
evaluate the usefulness of ScvO2 monitoring to guide
the management of cardiac surgery patients.
Acknowledgements
The authors thank the medical and nursing team from the Intensive
Care Unit (CHU-Tivoli, La Louviere, Belgium) for their collabora-
tion in collecting data and Pr C. Melot (Intensive Care Unit, Erasme
Hospital, Brussels, Belgium) for assistance in statistical analysis.
References
1 O’Connor JP, Townsend GE. Pro: perioperative continuous monitoring of
mixed venous oxygen saturation should be routine during high-risk cardiac
surgery. J Cardiothorac Anesth 1990; 4:647–650.
2 Swan HJ, Ganz W, Forrester J, et al. Catheterization of the heart in man with
use of a flow-directed balloon-tipped catheter. N Engl J Med 1970;
283:447–451.
3 Abreu AR, Campos MA, Krieger BP. Pulmonary artery rupture induced by a
pulmonary artery catheter: a case report and review of the literature.
J Intensive Care Med 2004; 19:291–296.
4 Djaiani G, Karski J, Yudin M, et al. Clinical outcomes in patients undergoing
elective coronary artery bypass graft surgery with and without utilization of
pulmonary artery catheter-generated data. J Cardiothorac Vasc Anesth
2006; 20:307–310.
5 Vincent JL. A reappraisal for the use of pulmonary artery catheters. Crit
Care 2006; 10 (Suppl 3):S1.
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
SvO2 vs. ScvO2 during cardiac surgery 299
6 De Wilde RB, Schreuder JJ, van den Berg PC, et al. An evaluation of
cardiac output by five arterial pulse contour techniques during cardiac
surgery. Anaesthesia 2007; 62:760–768.
7 Yazigi A, El Khouri C, Jebara S, et al. Comparison of central venous to mixed
venous oxygen saturation in patients with low cardiac index and filling
pressures after coronary artery surgery. J Cardiothorac Vasc Anesth 2008;
22:77–83.
8 Chawla LS, Zia H, Gutierrez G, et al. Lack of equivalence between central
and mixed venous oxygen saturation. Chest 2004; 126:1891–1896.
9 Bendjelid K, Treggiari MM, Suter PM, et al. Continuous SvO2
measurements and co-oximetry are not interchangeable immediately after
cardiopulmonary bypass. Can J Anaesth 2004; 51:610–615.
10 Pieri M, Brandi LS, Berolini R, et al. Comparison of bench central and mixed
pulmonary venous oxygen saturation in critically ill postsurgical patients.
Minerva Anestesiol 1995; 61:285–291.
11 Turnaoglu S, Tugrul M, Camci E, et al. Clinical applicability of the
substitution of mixed venous oxygen saturation with central venous oxygen
saturation. J Cardiothorac Vasc Anesth 2001; 15:574–579.
12 Varpula M, Karlsson S, Ruokonen E, et al. Mixed venous oxygen saturation
cannot be estimated by central venous oxygen saturation in septic shock.
Intensive Care Med 2006; 32:1336–1343.
13 Dueck MH, Klimek M, Appenrodt S, et al. Trends but not individual values of
central venous oxygen saturation agree with mixed venous oxygen
saturation during varying hemodynamic conditions. Anesthesiology 2005;
103:249–257.
14 Reinhart K, Kuhn HJ, Hartog C, et al. Continuous central venous and
pulmonary artery oxygen saturation monitoring in the critically ill. Intensive
Care Med 2004; 30:1572–1578.
15 Ladakis C, Myrianthefs P, Karabinis A, et al. Central venous and mixed
venous oxygen saturation in critically ill patients. Respiration 2001;
68:279–285.
16 Edwards JD, Mayall RM. Importance of the sampling site for measurement
of mixed venous oxygen saturation in shock. Crit Care Med 1998;
26:1356–1360.
17 Berridge JC. Influence of cardiac output on the correlation between mixed
venous and central venous oxygen saturation. Br J Anaesth 1992;
69:409–410.
18 Martin C, Auffray JP, Badetti C, et al. Monitoring of central venous oxygen
saturation versus mixed venous oxygen saturation in critically ill patients.
Intensive Care Med 1992; 18:101–104.
19 Tahvanainen J, Meretoja O, Nikki P. Can central venous blood replace
mixed venous blood samples? Crit Care Med 1982; 10:758–761.
20 Oshima T, Karasawa F, Satoh T. Effects of propofol on cerebral blood flow
and the metabolic rate of oxygen in humans. Acta Anaesthesiol Scand
2002; 46:831–835.
21 Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986; 1:307–310.
22 Reinhart K, Rudolph T, Bredle DL, et al. Comparison of central-venous to
mixed-venous oxygen saturation during changes in oxygen supply/demand.
Chest 1989; 95:1216–1221.
23 Shou H, Perez de Sa V, Larsson A. Central and mixed venous blood oxygen
correlate well during acute normovolemic hemodilution in anesthetized
pigs. Acta Anaesthesiol Scand 1998; 42:172–177.
24 Collaborative Study Group on Perioperative ScvO2 Monitoring.
Multicentre study on peri- and postoperative central venous oxygen
saturation in high-risk surgical patients. Crit Care 2006; 10:R158.
25 Kusano C, Baba M, Takao S, et al. Oxygen delivery as a factor in the
development of fatal postoperative complications after oesophagectomy.
Br J Surg 1997; 84:252–257.
26 Rivers E, Nguyen B, Havstad S, et al., Early Goal-Directed Therapy
Collaborative Group. Early goal-directed therapy in the treatment of severe
sepsis and septic shock. N Engl J Med 2001; 345:1368–1377.
27 Pearse R, Dawson D, Fawcett J, et al. Early goal-directed therapy after
major surgery reduces complications and duration of hospital stay.
A randomised, controlled trial. Crit Care 2005; 9:R687–R693.
European Journal of Anaesthesiology 2010, Vol 27 No 3