Central venous saturation is a predictor of reintubation in

difficult-to-wean patients*
Cassiano Teixeira, MD; Nilton Brandão da Silva, PhD; Augusto Savi, RPT; Silvia Regina Rios Vieira, PhD;
Luis Antônio Nasi, MD; Gilberto Friedman, PhD; Roselaine Pinheiro Oliveira, MD;
Ricardo Viegas Cremonese, MD; Túlio Frederico Tonietto, MD; Mathias Azevedo Bastian Bressel, MSc;
Juçara Gasparetto Maccari, MD; Ricardo Wickert, RPT; Luis Guilherme Borges, RPT

Objective: To evaluate the predictive value of central venous
saturation to detect extubation failure in difficult-to-wean patients.
Design: Cohort, multicentric, clinical study.
Setting: Three medical-surgical intensive care units.
Patients: All difficult-to-wean patients (defined as failure to
tolerate the first 2-hr T-tube trial), mechanically ventilated for
>48 hrs, were extubated after undergoing a two-step weaning
protocol (measurements of predictors followed by a T-tube trial).
Extubation failure was defined as the need of reintubation within
48 hrs.
Interventions: The weaning protocol evaluated hemodynamic
and ventilation parameters, and arterial and venous gases during
mechanical ventilation (immediately before T-tube trial), and at
the 30th min of spontaneous breathing trial.
Measurements and Main Results: Seventy-three patients were
enrolled in the study over a 6-mo period. Reintubation rate was
42.5%. Analysis by logistic regression revealed that central ve-
nous saturation was the only variable able to discriminate out-
come of extubation. Reduction of central venous saturation by
>4.5% was an independent predictor of reintubation, with odds
ratio of 49.4 (95% confidence interval 12.1–201.5), a sensitivity of
88%, and a specificity of 95%. Reduction of central venous
saturation during spontaneous breathing trial was associated
with extubation failure and could reflect the increase of respira-
tory muscles oxygen consumption.
Conclusions: Central venous saturation was an early and in-
dependent predictor of extubation failure and may be a valuable
accurate parameter to be included in weaning protocols of diffi-
cult-to-wean patients. (Crit Care Med 2010; 38:491– 496)
KEY WORDS: central venous saturation; mechanical ventilation;
extubation; difficult-to-wean

T he weaning process comprises
progressive withdrawal from
the invasive ventilatory sup-
port until removal of the en-
dotracheal tube and it could represent
approximately 40% of the patient’s time
on mechanical ventilation (MV) (1). A
more prolonged time on MV increases the
risk of nosocomial pneumonia, traumatic
airway injury, sepsis, and bleeding of di-
gestive stress ulcers (2– 4). Furthermore,
*See also p. 708.
From the Intensive Care Unit (CT, NBdS, AS, RPO,
RVC, TFT, MABB, JGM, RW, LGB), Moinhos de Vento
Hospital, Porto Alegre, Brazil; Central-Intensive Care
Unit (CT, GF, RPO), Santa Casa Hospital, Porto Alegre,
Brazil; and the Intensive Care Unit (SRRV, LAN, GF,
MABB), Clínicas Hospital, Porto Alegre, Brazil.
Gaúcho Weaning Study Group: Kamile Borba Pinto,
RPT; Fernanda Callefe Moreira, RPT; Eubrando Silves-
tre Oliveira, MD; Flavio Cardona Alves, MD; Jorge
Amilton Höher, MD; Cristiane Trevisan, RPT; Sérgio
Fernando Monteiro Brodt, MD; José Hervê Diel Barth,
MD; André Santana Machado, MD; Fabiano Leichsen-
ring, RPT; Patrícia de Campos Balzano, MD; Luciano
Marques Furlanetto, MD; José Fernando Pires, MD;
Eduardo Monteiro da Silva, MD; Régis Bueno Albuquer-
que, MD; Daniele Munareto Dallegrave, MD; Marcelo
Jeffman, MD; Sandro Cadaval Gonçalves, MD; Eduardo
failure in this process increases the risk
of nosocomial pneumonia and mortality
(5, 6). Success of weaning depends on
improvement of the acute illness, ade-
quate oxygenation and ventilatory param-
eters, hemodynamic stability, adequate
mentation and cough, and normal acid
base and electrolytes values (1). Extuba-
tion is performed after the decision has
been made to disconnect the patient from
MV and after the patient has tolerated a
de Oliveira Fernandes, PhD; Marcelo Garcia da Rocha,
MD; Maicon Becker, MD; Sergio Pinto Ribeiro, PhD;
Alexandre Cordella da Costa, MD; Marcelo de Mello
Rieder, RPT; Marisa Helena Pilenghi Correa, MD; Cris-
tiane Magalhães Siqueira de Campos Morais, MD;
Paulo José Zimmermann Teixeira, PhD; Márcio Pereira
Hetzel, MD; André Petit Torelly, MD; Roger Weingart-
ner, MD; Patrícia Pickersgill de Leon, MD; Joyce
Michele Silva, RPT.
The authors have not disclosed any potential con-
flicts of interest.
For information regarding this article, E-mail:
cassiano.rush@terra.com.br
Copyright © 2010 by the Society of Critical Care
Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/CCM.0b013e3181bc81ec
spontaneous breathing trial (SBT). How-
ever, extubation failure (EF) occurs in
approximately 14% to 32% of the patients
meeting these criteria, indicating that
the traditional two-step weaning protocol
(evaluation of predictors followed by T-
tube trial) does not adequately detect fail-
ure in difficult-to-wean patients (failure
to tolerate their first SBT) patients (5–9).
Presence of cardiovascular dysfunc-
tion can contribute to weaning failure
by increasing loads and reducing neu-
romuscular capacity. Although respira-
tory muscles do not develop fatigue,
they perform a huge workload. Thus,
they rely on efficient oxygen transport
by the cardiovascular system (10). Jub-
ran et al (11) examined the hemody-
namics and mixed venous saturation
(SvO2) in patients during weaning tri-
als. Patients who failed weaning also
failed to increase oxygen delivery (ḊO2)
to the tissues, in part due to elevated
right and left ventricular afterloads.
Central venous oxygen saturation
(ScvO2), although less accurate than
SvO2, has been successfully used as an

Crit Care Med 2010 Vol. 38, No. 2 491

adequate resuscitation goal in critical
illness (10, 12). Earlier authors (13–15)
had demonstrated adequate correlation
between ScvO2 and SvO2, however, not
in critically ill patients. As such, in the
weaning process (recovery stage from
critical disease), measurement of ScvO2
could potentially be a reliable and con-
venient tool to warn rapidly about acute
changes in the oxygen supply and de-
mand of these patients.
Given the hypothesis that changes in
ScvO2 during SBT could predict EF, a
multicentric study was conducted to
evaluate the predictive value of mea-
surements of ScvO2 in difficult-to-wean
patients submitted to a standard two-
step weaning protocol and extubated af-
ter successful SBT.
METHODS
Design
This is a cohort, multicentric, clinical
study performed in three medical-surgical in-
tensive care units (ICUs), and approved by the
Ethic Research Committees of the three insti-
tutions. The informed consent was obtained
from all patients or next-of-kin.
Patients
Over a 6-mo period, we studied prospec-
tively all difficult-to-wean patients (defined
as failure to tolerate the first 2-hr T-tube
trial), mechanically ventilated for ⬎48 hrs,
in three medical-surgical ICUs (Moinhos de
Vento Hospital, Santa Casa Hospital, and
Clínicas Hospital, Porto Alegre, Brazil). All
patients were ventilated with Servo 900C,
and Servo 300 (Siemens-Elema AB, Solna,
Sweden) or Evita-4 (Dräger Medical AG, Lü-
beck, Germany). They were assessed daily
for presence of the following readiness-to-
wean criteria: a) improvement in the under-
lying condition that leads to acute respira-
tory failure; b) adequate oxygenation,
indicated by PaO2 ⬎60 torr (⬎8 kPa) on FIO2
⬍0.4 and positive end-expiratory pressure
⬍8 cm H 2 O; c) cardiovascular stability
(heart rate ⬍130 beats/min and no or min-
imal pressors); d) afebrile; e) adequate he-
moglobin (⬎8 g/dL); f) adequate mental sta-
tus (arousal, Glasgow Coma Scale score of
⬎13, and no continuous sedative infusions);
g) effective cough; and (h) normal acid base
and electrolytes (1). Exclusion criteria were:
a) tracheotomized patients; b) patients with
no central venous catheter; c) negative for
informed consent; d) weaned successfully in
the first weaning trial; and e) dead before
weaning trial (Fig. 1).
492
Mechanical ventilated
patients
(n= 768)
Eligible for
the study
(n= 304)
Success 1st
SBT
Failure 1st
(n= 231)
Daily SBT
Extubation
success
(n=42)
Figure 1. Enrollment. MV, mechanical ventilation; SBT, spontaneous breathing trial.
Weaning Protocol
Patients meeting these criteria were then
weaned in a semirecumbent position, using
a two-step weaning protocol (measurements
of predictors followed by a T-tube trial dur-
ing 30 mins). In spontaneous breathing, fre-
quency to tidal volume index (f/VT) was cal-
culated by the respiratory rate (RR) and VT
ratio measured, using an electronic respi-
rometer (295, Ainca, San Marcos, CA) spi-
rometer monitored during 1 min. Maximal
inspiratory pressure (MIP) and maximal ex-
piratory pressure were measured with a ma-
nometer (Support Famabra, São Paulo, Bra-
zil) and defined as the most negative and
positive values, respectively, produced by
three consecutive inspiratory and expiratory
trials against a unidirectional valve during
20 secs (16). Those with f/VT ⱕ105 were
submitted to spontaneous breathing on the
T piece for 30 mins (defined as SBT) with
supplementary humidified oxygen (4 –7
L/min) to achieve arterial oxygen saturation
ⱖ90% as measured by pulse oximetry. Pa-
tients with intolerance to SBT, defined by:
RR ⱖ35 breaths/min, oxygen saturation by
pulse oximetry (SpO2) ⱕ90%, heart rate
ⱖ130 beats/min or changes ⱖ20%, change
in mental status (drowsiness, coma, agita-
tion, anxiety), worsened discomfort, dia-
phoresis or signs of increased work of
breathing (use of accessory respiratory mus-
cles or thoracoabdominal paradox) were re-
turned to MV and underwent new evaluation
on the day after. At this point, patients were
selected for entry in the study.
Excluded by:
- MV < 48hs (n=102)
- Tracheostomy (n= 54)
- Death before SBT (n= 124)
- Patients with no VCL (n= 98)
- Negative for informed consent (n= 86)
SBT
(n= 73)
until
success
Extubation
failure
(n=31)
After successful completion of an SBT,
patients were extubated, and followed for
presence of postextubation respiratory dis-
tress during 48 hrs. EF was defined as need
of reintubation in 48 hrs. Noninvasive ven-
tilation was used to prevent respiratory dis-
tress after extubation in all patients with
chronic obstructive pulmonary disease
(COPD).
Measurements
Measurements of ventilatory parameters
were recorded at 1st min and at 30th min of
SBT. Respiratory compliance, RR, oxygen-
ation and pressure index (CROP) were mea-
sured immediately before SBT (during MV
support) and calculated by the formula:
[Cdyn ⫻ MIP ⫻ (Pa O 2/P AO 2)]/RR, where
Cdyn stands for dynamic compliance, PAO2
for alveolar oxygen pressure, and RR for
respiratory rate. Arterial and venous blood
samples were collected immediately before
SBT (during MV support) and at 30th min of
SBT. Hemodynamic variables (heart rate
and arterial blood pressure), demographic
data, Acute Physiology and Chronic Health
Evaluation (APACHE II) score (17) at first 24
hrs of ICU stay, ICU admission diagnosis,
comorbidities, Glasgow Coma Scale score,
days in ICU, MV days, drugs used (neuro-
muscular blocking agents, benzodiazepines,
opioids, adrenocortical steroids) were also
registered. The ScvO 2 and arterial blood
were sampled by central venous access
(placed in the internal jugular or subclavic
vein) and radial artery, respectively, and
were analyzed immediately, using a blood
Crit Care Med 2010 Vol. 38, No. 2
gas analyzer (ABL 520 Radiometer, Copen- Table 1. Demographic characteristics, clinical parameters, drug use, and ventilatory settings
hagen, Denmark). Oxygen extraction ratio
(O2ER) was calculated by the formula: (SaO2 ⫺ All Patients Extubation Success Extubation Failure
ScvO2)/SaO2. Variables (n ⫽ 73) (n ⫽ 42) (n ⫽ 31) pa

Statistical Analysis Age, yr 57 ⫾ 19 55 ⫾ 20 59 ⫾ 18 .34

Male, % 39 (53) 22 (56) 17 (44) .83
All data were expressed as mean ⫾ stan- Admission APACHE II 17 ⫾ 6 16 ⫾ 5 20 ⫾ 7 .006
dard deviation for continuous variables and GCS at extubation 14 ⫾ 2 14 ⫾ 2 13 ⫾ 3 .13
percentages for categorical variables. Differ- Hemoglobin concentration, g/dL 9.8 ⫾ 1.5 10.1 ⫾ 1.8 9.4 ⫾ 1.2 .81
ICU admission, %
ences between the two groups at baseline Sepsis 36 (49) 23 (55) 13 (42) .35
were analyzed with the use of Student’s t Stroke 8 (11) 3 (7) 5 (16) .23
test or Mann-Whitney U test for continuous Postoperative 5 (6) 4 (9) 1 (3) .38
variable and chi-square test for categorical Exacerbation of asthma 4 (5) 3 (7) 1 (3) .63
variable, including the Fisher test. Logistic Acute pulmonary edema 4 (5) 3 (7) 1 (3) .42
regression was performed for multivariable Exacerbation of COPD 1 (1) 0 1 (3) NA
Comorbidities, %
analysis for all univariate relevant variables Heart disease 30 (41) 17 (40) 13 (42) .54
that discriminate EF for extubation success COPD 12 (16) 3 (7) 9 (29) .02
(ES) patients. The Pearson test and Spear- Drug use, days
man test were used to determine correla- Neuromuscular antagonists 2.5 ⫾ 1.5 3⫾2 2⫾2 .51
tions of parametric and nonparametric vari- Opioids 7⫾9 4⫾3 10 ⫾ 13 .41
ables, respectively. Incremental analysis of Benzodiazepines 5⫾3 4⫾3 6⫾3 .08
Adrenocortical steroids 14 ⫾ 13 10 ⫾ 6 18 ⫾ 18 .69
the area under the receiver operating char- Ventilatory settings at weaning trial
acteristic curve was performed to quantify PEEP, cm H2O 5⫾1 5⫾1 5⫾1 .96
ScvO2 differences between MV and the 30th Inspiratory pressure, cm H2O 17 ⫾ 4 17 ⫾ 5 18 ⫾ 4 .71
min of SBT. Statistical analysis was per- VT (mL) 545 ⫾ 134 567 ⫾ 126 513 ⫾ 140 .08
formed by a statistician, using the commer- Cdyn (mL/cm H2O) 50 ⫾ 22 52 ⫾ 18 48 ⫾ 27 .45
cially available software (Statistical Package PaO2/FIO2 289 ⫾ 100 292 ⫾ 92 285 ⫾ 111 .77
for Social Science - SPSS 11.0, Chicago, IL).
APACHE II, Acute Physiology and Chronic Health Evaluation; ICU, intensive care unit; MV,
Statistical significance was set at p ⬍ .05.
mechanical ventilation; GCS, Glasgow Coma Scale; COPD, chronic obstructive pulmonary disease;
PEEP, positive end-expiratory pressure; VT, tidal volume; Cdyn, dynamic compliance; PaO2/FIO2,
RESULTS arterial pressure/oxygen fraction ratio; NA, not analyzed.
a
Comparing extubation failure to extubation success.
From of July 2005 to January 2006
(6 mos), 768 patients were submitted to
MV in the three study ICUs: 124 died than at MV measurements (92 ⫾ 24 torr The PaO2, SaO2, ScvO2, and O2ER mea-
before weaning trial, 251 had been MV [12 ⫾ 3.2 kPa] vs. 103 ⫾ 35 torr [13.7 ⫾ surements between ES and EF from MV
dependent for ⬍48 hrs, 86 declined 4.7 kPa], p ⬍ .001; 95 ⫾ 3% vs. 97 ⫾ to 30th min of SBT are shown in Figure 2
consent, 98 patients had no central ve- 2%, p ⬍ .001; 66 ⫾ 9% vs. 69 ⫾ 7%, and Table 2.
nous catheter at the moment of wean- p ⬍ .001, respectively) (Table 2). Heart A ⬎4.5% reduction of ScvO2 was asso-
ing trial, and 54 were submitted to early rate increased significantly at the 30th ciated with greater risk for reintubation
tracheotomy. Seventy-three consecu- min of SBT (95 ⫾ 19 beats/min to 98 ⫾ (odds ratio ⫽ 49.5, 95% confidence inter-
tive MV patients were enrolled in the 19 beats/min, p ⫽ .009) and mean ar- val ⫽ 12.1–201.5, p ⬍ .0001). A receiver
study (Fig. 1). The mean age of these terial pressure remained stable (92 ⫾ operating characteristic curve was ob-
ptients was 57 ⫾ 19 yrs, mean APACHE 14 mm Hg to 95 ⫾ 13 mm Hg, p ⫽ .42) tained with this value and demonstrated
II score in first 24 hrs of ICU stay was when compared with MV measurement. 90% of sensitivity, 84% of specificity, pos-
17 ⫾ 6, and 53.4% were male. The most Ventilatory parameters as well as other itive predictive value of 0.86, and negative
frequent diagnosis at the time of ICU hemodynamic parameters and blood predictive value of 0.88 for EF (Fig. 3). In
admission was sepsis (49%) and ICU gas measurements were similar during our study, a reduction of ⬎17% in ScvO2
sample mortality rate was 18%. All pa- MV and at 30th min of SBT. predicted 100% of EF.
tients underwent a two-step weaning Mortality rate, ICU stay, and days on
protocol, but the intubation rate was Extubation Outcome MV were significantly higher in the EF
42.5%. The number of failed SBT before (Comparing Extubation Failure group than in the ES group (Table 4).
a successful SBT was equal in ES and and Extubation Success)
EF groups (1.2 ⫾ 0.5 vs. 1.3 ⫾ 0.5, p ⫽
.92). Patient characteristics, ventilatory Admission APACHE II, COPD diag- DISCUSSION
settings, ventilatory and hemodynamic nosis, SaO2, ScvO2, and O2ER at the 30th
parameters are shown in Table 1. min of SBT were associated with EF A greater risk of reintubation was
(Tables 1 and 2). However, multivariate associated with a ⬎4.5% reduction in
All Patients (Comparing MV and analysis demonstrated that only ScvO2 ScvO2 (88% sensitivity, 95% of specific-
(ES ⫽ 70 ⫾ 7 vs. EF ⫽ 60 ⫾ 8; p ⫽ .009) ity, positive predictive value of 0.93,
30th Minute of SBT)
and O2ER (ES ⫽ 27 ⫾ 7 vs. EF ⫽ 36 ⫾ 8, and negative predictive value of 0.90),
In the sample, PaO2, SaO2, and ScvO2 p ⫽ .003) at the 30th min of SBT could and with a higher calculated O2ER in
evaluated at 30th min of SBT were lower discriminate EF from ES (Table 3). difficult-to-wean patients.

Crit Care Med 2010 Vol. 38, No. 2 493

Table 2. Ventilatory and hemodynamic parameters, and arterial and central venous blood gases data Several investigators (1, 8, 18 –21) re-
during MV and at 30th min of SBT ported that formalizing weaning steps
into a protocol might improve the out-
All Patients Extubation Success Extubation Failure
come. However, approximately 25% of
Variable (n ⫽ 73) (n ⫽ 42) (n ⫽ 31) pa
patients have EF when followed during
During mechanical ventilation 48 hrs to 72 hrs (1). Our patients were
Blood gases extubated based on a rigid two-step wean-
pH 7.39 ⫾ 0.9 7.38 ⫾ 0.9 7.4 ⫾ 0.8 .42 ing institutional protocol, but the reintu-
PaCO2, torr 35 ⫾ 11 34 ⫾ 9 37 ⫾ 13 .22 bation rate was high (42.5%). Our pa-
PaO2, torr 103 ⫾ 35 103 ⫾ 32 103 ⫾ 39 .41 tients already had failed in previous SBT,
SaO2, % 97 ⫾ 2 97 ⫾ 2 96 ⫾ 3 .46 placing them in a subgroup of difficult-
ScvO2, % 69 ⫾ 7 69 ⫾ 7 68 ⫾ 8 .36
to-wean patients (1, 22–24).
O2ER, % 28.8 ⫾ 7 28.6 ⫾ 7 29.3 ⫾ 9 .1
Ventilatory parameters measured According to the Fick principle, ox-
at 1st min of SBT ygen uptake (V̇O2) depends on ḊO2 and
f/VT, breath/min/L 73 ⫾ 27 68 ⫾ 21 78 ⫾ 25 .45 O2ER (10, 25). Jubran et al (11) showed
RR, beat/min 28 ⫾ 6 29 ⫾ 5 26 ⫾ 7 .17 that successfully weaned patients had
MIP, cm H2O 41 ⫾ 13 42 ⫾ 12 41 ⫾ 15 .37 an increase in cardiac index between
MEP, cm H2O 30 ⫾ 11 32 ⫾ 10 27 ⫾ 10 .05 MV and end of the trial. This was not
CROP index 41 ⫾ 35 47 ⫾ 41 36 ⫾ 29 .09
found in the failure group. The combi-
Hemodynamic parameters
HR, beat/min 95 ⫾ 18 93 ⫾ 16 98 ⫾ 21 .21 nation of greater venous admixture and
SAP, mm Hg 132 ⫾ 21 131 ⫾ 19 134 ⫾ 23 .49 low SvO2 can lead to rapid arterial de-
DAP, mm Hg 72 ⫾ 14 72 ⫾ 13 71 ⫾ 16 .92 saturation and a relative decrease of
MAP, mm Hg 92 ⫾ 14 91 ⫾ 13 92 ⫾ 15 .78 oxygen supplied to the tissues (26). In
At 30th min SBT our study, ḊO2 was not measured, but
Blood gases similar results between groups (EF and
pH 7.38 ⫾ 0.6 7.38 ⫾ 0.5 7.37 ⫾ 0.7 .69
ES) were been estimated based on the
PaCO2, torr 35 ⫾ 11 34 ⫾ 9 39 ⫾ 13 .09
PaO2, torr 92 ⫾ 24b 94 ⫾ 24 88 ⫾ 24 .31 same hemoglobin level, hemodynamic
SaO2, % 95 ⫾ 3b 96 ⫾ 3 94 ⫾ 4 .02 parameters, and SaO2 and PaO2. There-
ScvO2, % 65 ⫾ 9b 70 ⫾ 7 60 ⫾ 8 .0001 fore, we believe that the drop in ScvO2
O2ER, % 31.2 ⫾ 8 27.1 ⫾ 7 37 ⫾ 8 .0001 could reflect the increase of respiratory
Ventilatory parameters muscles V̇O2 observed in EF patients
f/VT, breath/min/L 80 ⫾ 36 73 ⫾ 24 89 ⫾ 18 .52 during the SBT (11, 26).
RR, beat/min 28 ⫾ 7 27 ⫾ 6 29 ⫾ 8 .16
MIP, cm H2O 39 ⫾ 14 41 ⫾ 14 36.8 ⫾ 13 .16
Use of SvO2 during the weaning pe-
MEP, cm H2O 30 ⫾ 10 31 ⫾ 9 27 ⫾ 11 .15 riod has been previously studied (10,
Hemodynamic parameters 19 –23). Jubran et al (11) demonstrated
HR, beat/min 98 ⫾ 19c 95 ⫾ 14 102 ⫾ 23 .15 that SvO2 decreased in weaning failure
SAP, mm Hg 135 ⫾ 20 131 ⫾ 17 139 ⫾ 23 .12 patients, probably due to increased re-
DAP, mm Hg 75 ⫾ 13 75 ⫾ 12 75 ⫾ 14 .85 spiratory muscles O2ER. Noll and Byes
MAP, mm Hg 95 ⫾ 13 94 ⫾ 12 97 ⫾ 15 .36 (27) showed correlation of SvO2, vital
MV, mechanical ventilation; SBT, spontaneous breathing trial; PaCO2, carbon dioxide arterial
signs, and arterial blood gases in 30
pressure; HCO3 arterial bicarbonate; PaO2 oxygen arterial pressure; SaO2 arterial oxygen saturation; consecutive postoperative coronary ar-
ScvO2 central venous oxygen saturation; O2ER, oxygen extraction rate; f/VT, frequency to tidal volume tery bypass graft cases, but only SpO2
index; RR, respiratory rate; MIP, maximal inspiratory pressure; MEP, maximal expiratory pressure; HR, and respiratory rate correlated with
heart rate; SAP, systolic blood pressure; DAP, diastolic blood pressure; MAP, mean arterial blood weaning failure. Cason et al (28), in ten
pressure. postoperative coronary artery bypass
a
Comparing extubation failure and extubation success; bp ⬍ .01 compared with MV measurements; graft patients, evaluated SvO2 and SpO2
c
p ⬍ .05 compared with MV measurements. during SBT and showed that weaning
failure occurred when SvO2 was ⬍60%.
Armaganidis and Dhainaut (29) moni-
Table 3. Results of univariate and multivariate analyses for sample weaning predictors tored SvO2 in postoperative coronary ar-
tery bypass graft patients and demon-
Extubation Success Extubation Failure Univariate Multivariate
strated that SvO2 of ⬎60% was the best
Variable (n ⫽ 42) (n ⫽ 31) Analysis Analysis
weaning success predictor studied and
Admission APACHE II 16 ⫾ 5 20 ⫾ 7 ⬍.006 .08 depended on O2ER measurements. Dif-
COPD, % 3 (7) 9 (29) .02 .08 ferent from these previous studies, our
SaO2 at 30th min of SBT, % 96 ⫾ 3 94 ⫾ 4 .02 .34 study group was comprised predomi-
ScvO2 at 30th min of SBT, % 70 ⫾ 7 60 ⫾ 8 .0001 .009 nantly of critical ill nonsurgical MV-
O2ER at 30th min of SBT, % 27.4 ⫾ 7 36.1 ⫾ 8 .0001 .003
dependent patients, and our results
APACHE, Acute Physiology and Chronic Health Evaluation; COPD, chronic obstructive pulmonary demonstrated that ScvO2 was an effec-
disease; SaO2 on SBT, arterial oxygen saturation measured at 30th min of spontaneous breathing trial; tive EF predictor in difficult-to-wean
ScvO2, central venous oxygen saturation measured at 30th min of spontaneous breathing trial; O2ER, patients who underwent a successful T-
oxygen extraction rate measured at 30th min of spontaneous breathing trial. tube trial.

494 Crit Care Med 2010 Vol. 38, No. 2

 The choice of ScvO2 instead of SvO2
was due to limited use of a pulmonary
artery catheter during weaning period,
reflecting our everyday clinical practice.
Pulmonary artery catheterization is
costly and has inherent risks. In compar-
ison, ScvO2 is part of the standard care of
critically ill patients and is easier and
safer. Others investigators (13–15) had
demonstrated adequate correlation be-
tween ScvO2 and SvO2, except in critically
ill patients. Rivers et al (12) and Vallet et
al (15) previously showed that early goal-
directed therapy based on ScvO2 reduces
mortality in patients with severe sepsis
and septic shock. Measurement of ScvO2
is a potentially reliable and convenient
tool, which could rapidly warn about
acute change in the oxygen supply and
demand of critically ill patients. Our data
showed that, during MV (immediately be-
fore SBT), ScvO2 was not different be-
tween EF and ES patients, but that ScvO2
reduction during T-tube trial was able to
predict EF in 86% of the cases. ScvO2
remained unchanged in the ES group.
Overall mortality was 18%, with a
mean APACHE II score of 17— corrob-
orating the findings of Knaus et al (17).
It is noteworthy that the APACHE II
score was higher in the EF group (16 ⫾
5 vs. 20 ⫾ 7, p ⫽ .006) and could
contribute to failure and elevated mor-
tality in this group as well as presence
of COPD (7% vs. 29%, p ⫽ .02) and a
large number of septic patients at the
time of ICU admission (42%). Patients
with COPD present reduced pulmonary
function and develop dynamic hyperin-
flation during spontaneous ventilation,
which increases the intrinsic positive
end-expiratory pressure and makes
weaning from MV difficult (1). Further-
more, the large presence of COPD pa-
tients in the EF group could contribute
to prolong MV duration and ICU days in
this group because noninvasive ventila-
tion was only used after a successful
SBT and extubation in this subgroup of
patients. However, at multivariate analysis,

Table 4. Outcomes

Extubation Extubation
Success Failure
(n ⫽ 42) (n ⫽ 31) p

ICU days 12 ⫾ 10 24 ⫾ 14 ⬍.0001

MV days 7⫾7 9⫾5 .02
Figure 2. Changes in blood gases, from mechanical ventilation (MV) to 30th min of spontaneous Mortality in 1 (2) 12 (39) ⬍.0001
breathing trial (SBT), comparing extubation failure (dashed line) and extubation success (continuous ICU, %
line). *p ⬍ .001 compared with MV measurement; #p ⫽ .02 comparing extubation failure and
extubation success; §p ⬍ .0001 comparing extubation failure and extubation success. SaO2, arterial ICU, intensive care unit; MV, mechanical ven-
oxygen saturation; ScvO2, central venous oxygen saturation; O2ER, oxygen extraction rate. tilation.

Figure 3. Receiver operating characteristic curve for central venous oxygen saturation (A) and simplified oxygen extraction ratio (B) variations (30th min
of spontaneous breathing trial measured—mechanical ventilation measured). A ⬎ 4.5% reduction of central venous oxygen saturation was associated with
greater risk of reintubation (odds ratio 49.5, 95% confidence interval 12.1–201.5, p ⬍ .0001). Receiver operating characteristic curve demonstrated 88%
of sensitivity, 95% of specificity, positive predictive value of 0.93, and negative predictive value of 0.90 for extubation failure. In addition, a ⬎8% increase
of simplified oxygen extraction ratio demonstrated 95% of sensitivity, 63% of specificity, positive predictive value of 0.76 and negative predictive value of
0.91 to detect extubation failure (odds ratio 32, 95% confidence interval 7–163, p ⬍ .0002). AUC, area under the curve.

Crit Care Med 2010 Vol. 38, No. 2 495

the ScvO2 reduction during SBT was inde-
pendent from the other variables.
This study’s main limitation is the
nonrandom design. In difficult-to-wean
patients, the tolerance to SBT did not
guarantee successful extubation. In
these cases, reduction of ScvO2 was able
to predict extubation outcome. There-
fore, we believe that measurements of
ScvO2 during T-tube trial could be con-
sidered as a new parameter for difficult-
to-wean patients.
REFERENCES
1. MacIntyre NR, Cook DJ, Ely EW Jr, et al:
Evidence-based guidelines for weaning and
discontinuing ventilatory support: A collec-
tive task force facilitated by the American
College of Chest Physicians; the American
Association for Respiratory Care; and the
American College of Critical Care Medicine.
Chest 2001; 120(6 Suppl):375S–395S
2. Esteban A, Frutos F, Tobin MJ, et al: A com-
parison of four methods of weaning patients
from mechanical ventilation. N Eng J Med
1995; 332:345–350
3. Torres A, Serra-Batlles J, Ros E, et al: Pul-
monary aspiration of gastric contents in pa-
tients receiving mechanical ventilation: The
effect of body position. Ann Intern Med 1992;
116:540 –543
4. Epstein SK, Ciubotaru RL, Wong JB: Effect
of failed extubation on the outcome of me-
chanical ventilation. Chest 1997; 112:
186 –192
5. Esteban A, Alia I, Tobin MJ, et al: Effect of
spontaneous breathing trial duration on out-
come of attempts to discontinue mechanical
ventilation. Am J Respir Crit Care Med 1999;
159:512–518
6. Vallverdú I, Calaf N, Subirana M, et al: Clin-
ical characteristics, respiratory function pa-
rameters, and outcome of two-hour T-piece
trial in patients weaning from mechanical
ventilation. Am J Respir Crit Care Med 1998;
158:1855–1862
7. Lee KH, Hui KP, Chan TB, et al: Rapid shal-
low breathing (frequency-tidal volume ra-
496
tion) did not predict extubation outcome.
Chest 1994; 105:540 –543
8. Meade MO, Guyatt G, Cook D, et al: Predicting
success in weaning from mechanical ventila-
tion. Chest 2001; 120(6 Suppl):400S– 424S
9. Conti G, Montini L, Pennisi MA, et al: A
prospective, blinded evaluation of indexes
proposed to predict weaning from mechani-
cal ventilation. Intensive Care Med 2004; 30:
830 – 838
10. Reinhart K, Kuhn HJ, Hartog C, et al: Con-
tinuous central venous and pulmonary artery
oxygen saturation monitoring in the criti-
cally ill. Intensive Care Med 2004; 30:
1572–1578
11. Jubran A, Mathru M, Dries D, et al: Contin-
uous monitoring of mixed venous oxygen
saturation during weaning from mechanical
ventilation and the ramifications thereof.
Am J Respir Crit Care Med 1998; 158:
1763–1769
12. Rivers E, Nguyen B, Havstad S, et al: Early
goal-direct therapy in the treatment of severe
sepsis and septic shock. N Eng J Med 2001;
345:1365–1377
13. Sandham JD, Hull RD, Brant RF, et al: A
randomized controlled trial of the use of pul-
monary-artery catheters in high-risk surgical
patients. N Eng J Med 2003; 348:5–14
14. Chawla LS, Zia H, Gutierrez G, et al: Lack of
equivalence between central and mixed ve-
nous oxygen saturation. Chest 2004; 126:
1891–1896
15. Vallet B, Wiel E, Lebuffe G: Resuscitation
from circulatory shock. In: Textbook of Crit-
ical Care. Fifth Edition. Fink MP, Abraham E,
Vincent JL, Kochanek PM (Eds). Philadel-
phia, Elsevier Saunders, 2005, pp 905–910
16. American Thoracic Society/European Respi-
ratory Society: ATS/ERS Statement on respi-
ratory muscle testing. Am J Respir Crit Care
Med 2002; 166:518 – 624
17. Knaus WA, Draper EA, Wagner DP, et al:
APACHE II: A severity of disease classifica-
tion system. Crit Care Med 1985; 13:
818 – 829
18. Saura P, Blanch L, Mestre J, et al: Clinical
consequences of the implementation of a
weaning protocol. Intensive Care Med 1996;
22:1052–1056
19. Ely EW, Baker AM, Dunagan DP, et al: Effect
on the duration of mechanical ventilation of
identifying patients capable of breathing
spontaneously. N Engl J Med 1996; 154:
1647–1652
20. Kollef MH, Shapiro SD, Silver P, et al: A
randomized, controlled trial of protocol-
directed versus physician-directed weaning
from mechanical ventilation. Crit Care Med
1997; 25:567–574
21. Marelich GP, Murin S, Battistella F, et al:
Protocol weaning of mechanical ventilation
in medical and surgical patients by respira-
tory care practitioners and nurses: Effect on
weaning time and incidence of ventilator-
associated pneumonia. Chest 2000; 118:
459 – 467
22. Upadya A, Tilluckdharry L, Muralidharan V,
et al: Fluid balance and weaning outcomes.
Intensive Care Med 2005; 31:1643–1647
23. Mekontso-Dessap A, Prost N, Girou E, et al:
B-type natriuretic peptide and weaning from
mechanical ventilation. Intensive Care Med
2006; 32:1529 –1536
24. Wysocki M, Cracco C, Teixeira A, et al: Re-
duced breathing variability as a predictor of
unsuccessful patient separation from me-
chanical ventilation. Crit Care Med 2006; 34:
2076 –2083
25. Silance PG, Simon C, Vicent JL: The relation
between cardiac index and oxygen extraction
in acutely ill patients. Chest 1994; 105:
1190 –1197
26. Tobin MJ, Jubran A: Weaning from mechan-
ical ventilation. In: Principles and Practice of
Mechanical Ventilation. Second Edition. To-
bin MJ (Ed). New York, McGraw-Hill, 2006,
pp 1185–1220
27. Noll ML, Byers JF: Usefulness of measures of
Svo2, Spo2, vital signs, and derived dual
oximetry parameters as indicators of arterial
blood gas variables during weaning of cardiac
surgery patients from mechanical ventila-
tion. Heart Lung 1995; 24:220 –227
28. Cason CL, De Salvo SK, Ray WT: Changes in
oxygen saturation during weaning from
short-term ventilator support after coronary
bypass graft surgery. Heart Lung 1994; 23:
368 –375
29. Armaganidis A, Dhainaut JF: Weaning from
artificial respiration: Value of continuous
monitoring of mixed venous oxygen satura-
tion. [English Abstract]. Ann Fr Anesth Re-
anim 1989; 8:708 –715
Crit Care Med 2010 Vol. 38, No. 2