Artificial Organs
34(2):E59–E64, Wiley Periodicals, Inc.
© 2010, Copyright the Authors
Journal compilation © 2010, International Center for Artificial Organs and Transplantation and Wiley Periodicals, Inc.
Prognostic Factors for Adult Patients Receiving
Extracorporeal Membrane Oxygenation as Mechanical
Circulatory Support—A 14-Year Experience at a
Medical Center
*Chinchun Lan, †Pi-Ru Tsai, †Yih-Sharng Chen, and †Wen-Je Ko
*Division of Critical Care, Changhua Christian Hospital; and †Department of Surgery, National Taiwan University
Hospital, Taipei, Taiwan
Abstract: Extracorporeal membrane oxygenation
(ECMO) is a resource-consuming and highly invasive
treatment. There were 1100 ECMO cases at the National
Taiwan University Hospital between August 1994 and
November 2008. Of these, 607 were adults (>18 years old)
who received ECMO as mechanical circulatory support. In
this study, patient characteristics and complications during
the ECMO course were evaluated for their correlation with
prognosis. The following factors were significantly different
between survivors and nonsurvivors: age, coronary artery
disease, diabetes mellitus, brain death, stroke during
ECMO, the need for dialysis during ECMO, pre-ECMO
Extracorporeal membrane oxygenation (ECMO)
is used for the management of life-threatening
cardiac or pulmonary failure (or both). Most com-
monly, it is implemented in an emergency or urgent
situation after failure of other treatment modalities.
The technique has become more reliable with
improving equipment and increased experience, and
this is reflected in improving results (1).
When the cardiac output of a patient is not suf-
ficient to meet the oxygen demand of the tissue,
inotropic agents are usually given as the first-line
treatment. If a failing heart does not respond well to
medical treatment, mechanical circulatory support
(MCS) may be required. Some patients require only
doi:10.1111/j.1525-1594.2009.00909.x
Received February 2009; revised June 2009.
Address correspondence and reprint requests to Dr. Wen-Je Ko,
Department of Surgery, National Taiwan University Hospital, 7,
Chung-Shan S. Road, Taipei 100, Taiwan. E-mail: kowj@ntu.edu.tw
E59
aor_909 59..64
infection, hypoglycemia, acidosis, alkalosis, the need for a
distal perfusion catheter, and the amount of red blood cells
transfused. Six independent predictors of mortality were
identified: age, stroke, the need for dialysis during ECMO,
pre-ECMO infection, hypoglycemia, and alkalosis. Our
institution has comparatively extensive experience with
adult patients, which may be quite different from other
medical centers with respect to distribution of patient age.
The findings should lead to better utilization of ECMO
for adult patients in the future. Key Words: Extracor-
poreal membrane oxygenation—Mechanical circulatory
support—Prognostic factor.
temporary support that can be discontinued when
there is recovery from the acute insult, whereas
others require prolonged support until more defini-
tive therapy, such as transplantation, is performed.
When MCS is indicated, an intra-aortic balloon
pump (IABP) is the first choice because of its relative
noninvasiveness compared with ECMO or ventricu-
lar assist device (VAD). IABP works by two major
mechanisms. First, aortic volume and afterload are
reduced during systole through a vacuum effect
created by rapid balloon deflation. Second, blood is
displaced to the proximal aorta by inflation during
diastole. The results are increased cardiac output
and decreased myocardial oxygen consumption.
However, the increase in cardiac output is limited in
patients with profound heart failure, tachyarrhyth-
mia, small body weight, or right heart failure. In
comparison, ECMO is a better MCS choice for
cardiogenic shock not amenable to IABP support
alone. ECMO can support both the right and left
sides of the heart, and provide greater cardiac output
E60 C. LAN ET AL.
than IABP. ECMO is not limited by body size and is
not affected by arrhythmia.
VAD is indicated for two categories of patients.
The first group is patients requiring support for
refractory heart failure as a “bridge to trans-
plantation.” These patients are usually on a heart
transplant list but deteriorate before a donor heart
becomes available. The second group is patients with
chronic heart failure who have a very poor long-term
prognosis and are not transplant candidates because
of advanced age, end-stage renal disease, or chronic
obstructive pulmonary disease. Such patients will
probably die from cardiac disease unless mechanical
circulatory assist systems become available for per-
manent use. Due to technical complexity, VAD is not
suitable for critical patients in emergencies. In such
cases, ECMO can be implemented because of its rela-
tive simplicity and lower cost (2), allowing time to
make a decision about the next therapeutic step and,
if necessary, bridging to heart transplantation or per-
manent VAD (3).
Various complications may occur during implanta-
tion and the course of ECMO. Longer duration of
ECMO support is associated with a higher probabil-
ity of complications. Prolonged ECMO is compli-
cated by systemic inflammatory state and may lead
to multiple organ failure, resistant vasoplegia, co-
agulopathy, and stroke. For these reasons, ECMO
support in adult patients is limited to short periods
until either myocardial or pulmonary recovery has
occurred or more definitive circulatory devices (for
example, VAD) can be placed (4).
As ECMO is a resource-consuming and highly
invasive treatment, the outcome and associated com-
plications deserve extensive research (2,5–9). Prog-
nostic factors, including age, gender, pre-ECMO
comorbidity, and ECMO-related complications, were
investigated in this study based on the 14 years of
experience at our institution. The following factors
were significantly different between survivors and
nonsurvivors: age, coronary artery disease, diabetes
mellitus, brain death, stroke during ECMO, the need
for dialysis during ECMO, pre-ECMO infection,
hypoglycemia, acidosis, alkalosis, the need for a distal
perfusion catheter, and the amount of red blood cells
(RBCs) transfused. Six independent predictors of
mortality were identified: age, stroke, requirement
of dialysis during ECMO, pre-ECMO infection,
hypoglycemia, and alkalosis.
PATIENTS AND METHODS
Between August 1994 and November 2008, there
were 1100 ECMO cases at the National Taiwan Uni-
Artif Organs, Vol. 34, No. 2, 2010
versity Hospital (NTUH). The ECMO team was
directly supervised by the director of the surgical
intensive care unit (ICU). They did not participate in
the cardiopulmonary bypass in the operating rooms.
They were full-time ECMO specialists, instead of tra-
ditional perfusionists who help to set up the ECMO
circuit only. The ECMO technicians were responsible
for ECMO care also.They worked in shifts so that the
service was always available. Their dedication led to
considerable success in extracorporeal cardiopulmo-
nary resuscitation (ECPR) in the hospital (10,11).
Interhospital ECMO transportation was also pro-
vided by the ECMO team (12).
In NTUH, indications for ECMO during the study
period included: MCS, acute respiratory distress
syndrome, replacement of cardiopulmonary bypass
needed in operations, organ preservation for
non-heart-beating donation, and others. This study
focused on the MCS category. Subgroups in MCS
included: postcardiotomy cardiogenic shock, acute
myocarditis, cardiomyopathy, acute myocardial
infarction, and acute rejection after heart
transplantation.
For quality assurance, data from each ECMO
patient were recorded in an ECMO database. Our
study was done by retrospective analysis of the pro-
spectively collected data. All adult patients (older
than 18 years old) who received ECMO for MCS
were included. Of the 1100 ECMO cases, 607 met the
inclusion criteria. During the same period, there were
284 ECMO patients younger than 18 years old. Our
institution has comparatively extensive experience
with adult patients, which may be quite different from
other medical centers with respect to distribution of
patient age.
There are two types of ECMO—venoarterial (VA)
and venovenous (VV). Both provide respiratory
support, but only VA ECMO provides hemodynamic
support. The femoral route was preferred to the open
sternotomy route for VA ECMO support because the
presence of an open sternotomy wound increases the
risk of bleeding and infection, and makes nursing
care more difficult. In the NTUH, we cannulated the
patients with a modified open Seldinger method (13).
The femoral vessels were dissected out and the can-
nulas were inserted with a guide wire under direct
vision. This method was particularly useful during
cardiopulmonary resuscitation with an impalpable
femoral pulse. Purse-string sutures were placed
around the cannula if possible to prevent bleeding.
With all the surgical instruments placed on a wheeled
ECMO cart, this technique was our standard
procedure. With this technique, femoral ECMO
support could be started securely within 15 min. The
 PROGNOSTIC FACTORS FOR ADULT PATIENTS RECEIVING ECMO AS MCS
main drawback of femoral access was ischemia to the
ipsilateral lower extremity. The likelihood of this
complication was decreased by inserting an addi-
tional arterial cannula distal to the femoral artery
cannula. A portion of the infused blood was redi-
rected into the additional cannula for perfusion of
the distal extremity.
In our analysis, two groups of possible prognostic
factors were taken into consideration. The first group
consisted of patient characteristics: age, gender,
and comorbidity. Comorbidity was divided into
cardiac comorbidity and noncardiac comorbidity.
Cardiac comorbidity included hypertension, coronary
artery disease, and history of myocardial infarction.
Noncardiac comorbidity included diabetes mellitus,
history of cerebral vascular accident, chronic obstruc-
tive pulmonary disease, end-stage renal disease under
maintenance hemodialysis, and liver cirrhosis. In total,
10 patient characteristics were taken into account.
The second group of possible prognostic factors
consisted of complications during the ECMO course.
Six categories of complications were considered in
this survey: neurologic, renal, infectious, metabolic,
vascular, and hemorrhagic. Cardiopulmonary compli-
cation was excluded because it would be difficult to
differentiate the effect of this complication from the
preexisting cardiopulmonary condition that led to
the installation of ECMO. Mechanical complication
was also excluded. The machine was robust and
equipped with a back-up battery. Pump breakdown
did not occur. The only mechanical complication was
blood leak from the oxygenator. We simply changed
the whole ECMO circuit immediately. No patient
died of this complication in this series.
The manifestations of neurological complications
included brain death, stroke, and seizure. Renal
complication was defined as the need for dialysis
during ECMO. Metabolic complications included
two major disorders: poor glucose control (blood
glucose < 40 mg/dL or blood glucose > 240 mg/dL)
and acid-base disturbance (pH < 7.2 or pH > 7.6).The
criterion for hyperglycemia was set at a relatively
high level because stress-induced hyperglycemia was
frequently observed in critically ill patients. Vascular
complications included digital gangrene, the neces-
sity for a distal perfusion catheter, or when fas-
ciotomy was required to relieve acute compartment
syndrome. A distal perfusion catheter was inserted
when mean arterial pressure distal to the femoral
cannulation site was less than 50 mm Hg (13).
The hemorrhagic complication was defined as the
transfusion of RBCs (including whole blood and
packed RBCs) on the first day after the ECMO setup.
ECMO duration varied greatly due to the difference
E61
in disease severity and patient characteristics. The
amount of blood loss is affected by many factors that
may change with time, such as coagulopathy. We
focused on the blood loss directly related to the
installation of ECMO. Because of heparin-bound
bioactive surface, systemic heparinization was not
needed on the first day of ECMO support. Therefore,
our definition of hemorrhagic complication can be
viewed as one associated with the implantation
procedure.
Chi-square test for categorical variables and t-test
for continuous variables were used to test the signifi-
cance of difference in prognostic factors between the
survivors and nonsurvivors. A P value less than 0.05
was considered statistically significant. Multivariate
logistic regression analysis was performed to identify
independent predictors of mortality.
RESULTS
In the 607 patients enrolled, 183 patients survived
to discharge. The survival rate was 30.1%, which was
comparable with the outcomes from the previous
reports of ECMO for cardiac failure (1,7). The
average age of our patients was 53.8 ⫾ 15.5 years old.
The youngest was 18 years old, as dictated by our
inclusion criteria. The oldest was 87 years old. Male
predominance was found in our study population,
making up more than 70% of the patients (Table 1).
Diabetes mellitus was the most prevalent comor-
bidity, accounting for almost one fourth of the
population. As to cardiac comorbidity, hypertension
was the most common one, followed by coronary
artery disease and history of myocardial infarction.
Age, coronary artery disease, and diabetes mellitus
were significantly different between survivors and
nonsurvivors (Table 1).
The incidences of the complications are shown
in Table 2. Hyperglycemia was the most common
complication. More than one-half of all patients had
hyperglycemia, even though we increased the thresh-
old of hyperglycemia to 240 mg/dL. The second was
the need for dialysis during ECMO, which occurred
in nearly half of the patients.
More than one-third of our patients needed a distal
perfusion catheter to prevent lower limb ischemia.
The beneficial effect of the distal perfusion catheter is
demonstrated by our results. Although 38% of the
patients had signs of ipsilateral limb ischemia and
required insertion of a distal perfusion catheter, only
5.4% developed digital gangrene (Table 2).
The following patient characteristics were found
significantly different between survivors and nonsur-
vivors by univariate analysis: age, coronary artery
Artif Organs, Vol. 34, No. 2, 2010
E62 C. LAN ET AL.

TABLE 1. Background data

Patient characteristic Total (n = 607) Survivors (n = 183) Nonsurvivors (n = 424) P value

Demography
Age (mean ⫾ SD) 53.8 ⫾ 15.5 48.8 ⫾ 15.3 56.0 ⫾ 15.1 <0.001
Men 434 (71.5%) 130 (71.0%) 304 (71.7%) 0.845
Cardiac comorbidity
Hypertension 104 (17.1%) 23 (12.6%) 81 (19.1%) 0.060
CAD 69 (11.4%) 12 (6.6%) 57 (13.4%) 0.017
Old MI 19 (3.1%) 2 (1.1%) 17 (4.0%) 0.074
Noncardiac comorbidity
DM 135 (22.2%) 30 (16.4%) 105 (24.8%) 0.025
Prior CVA 33 (5.4%) 6 (3.3%) 27 (6.4%) 0.171
COPD 8 (1.3%) 0 (0%) 8 (1.9%) 0.113
ESRD 30 (4.9%) 7 (3.8%) 23 (5.4%) 0.541
Liver cirrhosis 2 (0.3%) 1 (0.5%) 1 (0.2%) 0.512

P value was calculated by chi-square test or t-test, as appropriate.

CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; CVA, cerebral vascular accident; DM, diabetes mellitus;
ESRD, end-stage renal disease; MI, myocardial infarction.

disease, diabetes mellitus (Table 1). Among the com-
plications, the following were significant factors con-
tributing to difference in mortality rate: stroke, the
need for dialysis during ECMO, pre-ECMO infec-
tion, hypoglycemia, acidosis, alkalosis, necessity of
distal perfusion catheter, and the amount of RBCs
transfused (Table 2). Brain death always led to
mortality.
Multivariate logistic regression analysis with a
stepwise forward method was performed. The follow-
ing factors were identified as independent predictors
of mortality: age, stroke, the need for dialysis during
ECMO, pre-ECMO infection, hypoglycemia, and
alkalosis (Table 3). Alkalosis influenced the mortality
in a way opposite to the other variables. All other
variables increased the likelihood of mortality,
whereas alkalosis decreased the likelihood of
mortality.
DISCUSSION
The aim of this study was to determine the prog-
nostic factors that influence the outcome of patients
receiving ECMO as MCS. Based on the data

TABLE 2. Incidence of ECMO complications

Complication Total (n = 607) Survivors (n = 183) Nonsurvivors (n = 424) P value

Neurologic
Brain death 101 (16.6%) 0 (0%) 101 (23.8%) <0.001
Stroke 40 (6.6%) 5 (2.7%) 35 (8.3%) 0.012
Seizure 10 (0.2%) 1 (0.5%) 9 (2.1%) 0.296
Renal
Dialysis 301 (49.6%) 42 (23.0%) 259 (61.1%) <0.001
Infection
Pre-ECMO 79 (13.0%) 14 (7.7%) 65 (15.3%) 0.010
On-ECMO 132 (21.7%) 37 (20.2%) 95 (22.4%) 0.549
Metabolic
Hypoglycemia* 29 (4.8%) 1 (0.5%) 28 (6.6%) 0.001
Hyperglycemia† 304 (50.1%) 93 (50.8%) 211 (49.8%) 0.811
Acidosis (pH<7.2) 141 (23.2%) 24 (13.1%) 117 (27.6%) <0.001
Alkalosis (pH>7.6) 154 (25.3%) 62 (33.9%) 92 (21.7%) 0.002
Vascular
Digital gangrene 33 (5.4%) 7 (3.8%) 26 (6.1%) 0.250
Distal perfusion catheter 231 (38.0%) 55 (30.1%) 176 (41.5%) 0.008
Fasciotomy 19 (3.1%) 8 (4.4%) 11 (2.6%) 0.249
Hemorrhagic‡ (mean ⫾ SD) 5.35 ⫾ 7.05 4.26 ⫾ 6.42 5.82 ⫾ 7.27 0.009

*Blood glucose < 40 mg/dL.

†
Blood glucose > 240 mg/dL.
‡
Defined by the amount (units) of RBC transfusion (including whole blood and packed RBCs) required on the first day after ECMO setup.

Artif Organs, Vol. 34, No. 2, 2010

 PROGNOSTIC FACTORS FOR ADULT PATIENTS RECEIVING ECMO AS MCS
TABLE 3. Multivariate logistic-regression analysis: independent predictors of
mortality
Factor Adjusted odds ratio (95% confidence interval)
Age
Stroke
Dialysis
Pre-ECMO Infection
Hypoglycemia
Alkalosis
accumulated in the past 14 years, 24 factors were
considered.
Age is an important factor that affects the outcome
in almost all diseases. It is unlikely that ECMO is an
exception as ECMO implementation imposes great
physiologic stress on the patients. However, our
analysis yielded the odds ratio of age as only 1.04. A
study in Japan of 91 patients concluded that elderly
patients did as well as younger counterparts (9). A
plausible explanation is that selection bias is always a
factor when dealing with old patients.A senior citizen
will be recommended to undergo an invasive proce-
dure only if he or she has relatively less underlying
disease.
With regard to neurologic complications, our
analysis identified stroke, including ischemic and
hemorrhagic, as an independent predictor of
mortality. Seizure, a relatively minor form of neuro-
logic insult, was not a significant factor. Although
neurologic complications occurred less frequently
than other types of complications, they were most
devastating. Profound and irreversible neurologic
deficits were often the major reason for withdrawal
of ECMO support.
Although the need for dialysis was a strong predic-
tor of mortality, the role of renal function in the final
outcome was not clarified. Dialysis reflects extreme
impairment of renal function. The need for dialysis
implied either that the total cardiac output was still
not sufficient to maintain adequate renal perfusion
after implementation of ECMO (pre-renal failure),
or that the patient had sustained severe damage
of the kidney (intrinsic renal failure). Both could
contribute to a dismal outcome. It is interesting
that ESRD with maintenance hemodialysis before
ECMO institution was not a predictor of mortality.
In the modern ICU, infection control is a major
challenge for intensivists. This is no exception for an
ECMO team; 13% of our ECMO patients had infec-
tions before the implementation of ECMO and
21.7% were infected after the implantation of
ECMO (Table 2). In early experience, sepsis was con-
sidered a contraindication for the use of ECMO
E63
P value
1.04 (1.02–1.05) <0.001
4.94 (1.65–14.80) 0.004
6.49 (4.12–10.23) <0.001
2.10 (1.02–4.31) 0.044
17.31 (2.09–143.43) 0.008
0.39 (0.24–0.64) <0.001
support. Potential concerns were that cannula place-
ment might perpetuate bacteremia in the presence of
foreign bodies introduced into the vasculature, that
the inflammatory response might be aggravated by
foreign membranes, or that septic patients might
have a preexisting coagulopathy that would further
predispose them to ECMO bleeding complications
(14). However, several studies in the mid-1990s
refuted this hypothesis (15–17). Nonetheless, sepsis
with multiple organ failure has been associated with
poor outcomes in some reports for children (18–20).
Our data on adult patients indicated that the impact
of infection on mortality depended on the timing of
the onset of infection. It was interesting that the
infection that occurred after ECMO implementation
did not exert high impact on mortality. The difference
between survivors and nonsurvivors did not reach
statistical significance even in a cohort as large as
ours (Table 2). However, infection that already
existed before ECMO institution was not only signifi-
cantly different between survivors and nonsurvivors
(Table 2), but also an independent predictor of mor-
tality (Table 3).
Hypoglycemia turned out to have the highest odds
ratio among the independent predictors of mortality
(Table 3). The relationship between hypoglycemia
and ECMO-associated mortality warrants more
deliberation because patients with hypoglycemia
have higher ICU mortality than those who do not
(21). Therefore, although strict glucose control for
critical patients is recommended by the recent
literature, we should be cautious and avoid
over-treatment.
Alkalosis was found to be a reverse predictor of
ECMO-associated mortality. Patients for whom
ECMO is required are usually critically ill. Mild
elevation of the pH value due to respiratory alkalosis
may not be a bad sign for these patients because
respiratory alkalosis is a physiological response to
stress. However, the cut-off point of pH value for
alkalosis in our study was already heightened to 7.6.
Usually, prognosis is compromised in patients with
severe alkalosis. Therefore, more investigation is
Artif Organs, Vol. 34, No. 2, 2010
E64 C. LAN ET AL.
needed to determine the mechanism of alkalosis in
ECMO patients.
There are several limitations in this study. The
MCS group was a heterogenous population with a
great diversity of etiologies. The predicting power
of each of the pre-ECMO comorbidity or ECMO-
associated complications may be different among the
subgroups. Besides, mortality could occur at different
periods of the hospital course: died on ECMO;
weaned but died in ICU; or transferred to general
ward but died before discharge. The mechanisms
leading to mortality in each period may be different.
Therefore, more detailed study on the predictors of
mortality in each period may give more information
about how to reduce mortality by preventing fatal
complications.
CONCLUSION
ECMO is a very resource-consuming treatment. A
proper decision-making process is required to justify
its initiation. After ECMO institution, every possible
measure should be undertaken to prevent complica-
tions that may lead to failure of the treatment. As a
randomized study is difficult for an urgent therapy
such as ECMO, a large series as ours provides very
valuable information about this treatment modality.
Based on our prospectively collected and retrospec-
tively analyzed data, 12 factors were found signifi-
cantly different between survivors and nonsurvivors.
Six independent predictors of mortality were iden-
tified. Compared with other medical centers, our
institution has relatively extensive experience in
adult patients. The findings derived from this study
should help to utilize ECMO more effectively for
adult patients in the future.
Acknowledgments: Dr. Chui-Yu Chiu, Depart-
ment of Industrial Engineering and Management,
National Taipei University of Technology, was con-
sulted for the statistical method. The final version of
the manuscript was reviewed for English correction
by Dr. Alan R. Talbot, director of Department of
Medical Quality Assurance, Changhua Christian
Hospital. Dr. Talbot is the founder of the ICU of the
Changhua Christian Hospital and one of the pioneers
of critical care in Taiwan.
REFERENCES
1. Marasco SF, Lukas G, McDonald M, McMillan J, Ihle B.
Review of ECMO (Extra Corporeal Membrane Oxygenation)
Artif Organs, Vol. 34, No. 2, 2010
support in critically ill adult patients. Heart Lung Circ 2008;17
(Suppl 4):S41–7.
2. Ko WJ, Lin CY, Chen RJ, Wang SS, Lin FY, Chen YS. Extra-
corporeal membranous oxygenation support for adult postcar-
diotomy cardiogenic shock. Ann Thorac Surg 2002;73:538–45.
3. Chen YS, Ko WJ, Yu HY, et al. Analysis of the outcome for
patients experiencing myocardial infarction and cardiopulmo-
nary resuscitation refractory to conventional therapies neces-
sitating extracorporeal life support rescue. Crit Care Med
2006;34:950–7.
4. Kale P, Fang JC. Devices in acute heart failure. Crit Care Med
2008;36 (Suppl):S121–S128.
5. Zimpfer D, Heinisch B, Czerny M, et al. Late vascular compli-
cations after extracorporeal membrane oxygenation support.
Ann Thorac Surg 2006;81:892–5.
6. Bakhtiary F, Keller H, Dogan S, et al. Venoarterial extracor-
poreal membrane oxygenation for treatment of cardiogenic
shock: clinical experiences in 45 adult patients. J Thorac Car-
diovasc Surg 2008;135:382–8.
7. Smedira NG, Moazami N, Golding CM, et al. Clinical experi-
ence with 202 adults receiving extracorporeal membrane oxy-
genation for cardiac failure: survival at five years. J Thorac
Cardiovasc Surg 2001;122:92–102.
8. Combes A, Leprince P, Luyt CE, et al. Outcomes and long-
term quality-of-life of patients supported by extracorporeal
membrane oxygenation for refractory cardiogenic shock. Crit
Care Med 2008;36:1404–11.
9. Saito S, Nakatani T, Kobayashi J, et al. Is extracorporeal life
support contraindicated in elderly patients? Ann Thorac Surg
2007;83:140–5.
10. Chen YS, Lin JW, Yu HY, et al. Cardiopulmonary resuscita-
tion with assisted extracorporeal life-support versus conven-
tional cardiopulmonary resuscitation in adults with in-hospital
cardiac arrest: an observational study and propensity analysis.
Lancet 2008;372:554–61.
11. Chen YS, Yu HY, Huang SC, et al. Extracorporeal membrane
oxygenation support can extend the duration of cardiopulmo-
nary resuscitation. Crit Care Med 2008;36:2529–35.
12. Huang SC, Chen YS, Chi NH, et al. Out-of-center extracorpo-
real membrane oxygenation for adult cardiogenic shock
patients. Artif Organs 2006;30:24–8.
13. Huang SC, Yu HY, Ko WJ, Chen YS. Pressure criterion
for placement of distal perfusion catheter to prevent limb
ischemia during adult extracorporeal life support. J Thorac
Cardiovasc Surg 2004;128:776–7.
14. Fortenberry JD, Paden ML. Extracorporeal therapies in the
treatment of sepsis: experience and promise. Semin Pediatr
Infect Dis 2006;17:72–9.
15. Stewart DL, Dela Cruz TV, Ziagler C, Goldsmith LJ. The use
of extracorporeal membrane oxygenation in patients with
Gram-negative or viral sepsis. Perfusion 1997;12:3–8.
16. Goldman AP, Kerr SJ, Butt W, et al. Extracorporeal support
for intractable cardiorespiratory failure due to meningococcal
disease. Lancet 1997;349:466–9.
17. Rich PB, Younger JG, Soldes OS, Awad SS, Bartlett RH. Use
of extracorporeal life support for adult patients with respira-
tory failure and sepsis. ASAIO J 1998;44:263–6.
18. MacLaren G, Butt W, Best D, Donath S, Taylor A. Extra-
corporeal membrane oxygenation for refractory septic shock
in children: one institution’s experience. Pediatr Crit Care Med
2007;8:447–51.
19. Weber TR, Kountzman B. Extracorporeal membrane oxygen-
ation for nonneonatal pulmonary and multiple-organ failure.
J Pediatr Surg 1998;33:1605–9.
20. Luyt DK, Pridgeon J, Brown J, Peek G, Firmin R, Pandya HC.
Extracorporeal life support for children with meningococcal
septicaemia. Acta Paediatr 2004;93:1608–11.
21. Arabi YM, Dabbagh OC, Tamin HM, et al. Intensive versus
conventional insulin therapy: a randomized controlled trial
in medical and surgical critically ill patients. Crit Care Med
2008;36:3190–7.