Usefulness of routine laboratory parameters in the decision to treat refractory cardiac arrest with extracorporeal life support

Usefulness of routine laboratory parameters in the decision to treat refractory cardiac arrest with extracorporeal life support

Resuscitation 82 (2011) 1154–1161 Contents lists available at ScienceDirect Resuscitation journal homepage: www.elsevier.com/locate/resuscitation C...

444KB Sizes 8 Downloads 78 Views

Resuscitation 82 (2011) 1154–1161

Contents lists available at ScienceDirect

Resuscitation journal homepage: www.elsevier.com/locate/resuscitation

Clinical paper

Usefulness of routine laboratory parameters in the decision to treat refractory cardiac arrest with extracorporeal life support夽 Bruno Mégarbane a,∗ , Nicolas Deye a,b , Mounir Aout c , Isabelle Malissin a , Dabor Résière a , Hakim Haouache a , Pierre Brun a , William Haik a , Pascal Leprince d , Eric Vicaut c , Frédéric J. Baud a a

Assistance Publique – Hôpitaux de Paris, Lariboisière Hospital, Medical and Toxicological Critical Care Department, Paris-Diderot University, 75010 Paris, France INSERM U 689, Paris-Diderot University, 75010 Paris, France c Assistance Publique – Hôpitaux de Paris, Lariboisière Hospital, Biostatistics Department, Paris-Diderot University, 75010 Paris, France d Assistance Publique – Hôpitaux de Paris, Pitié Salpétrière Hospital, Department of Cardiothoracic Surgery, Paris-Pierre and Marie Curie University, 75013 Paris, France b

a r t i c l e

i n f o

Article history: Received 24 January 2011 Received in revised form 3 April 2011 Accepted 2 May 2011

Keywords: Cardiac arrest Cardiopulmonary resuscitation Extracorporeal life support Peripheral venous oxygen saturation Capillary leak syndrome Multiorgan failure

a b s t r a c t Aim: To evaluate the usefulness of routine laboratory parameters in the decision to treat refractory cardiac arrest patients with extracorporeal life support (ECLS). Methods: Sixty-six adults with witnessed cardiac arrest of cardiac origin unrelated to poisoning or hypothermia undergoing cardiopulmonary resuscitation without return of spontaneous circulation (duration: 155 min [120–180], median, [25–75%-percentiles]) were included in a prospective cohortstudy. ECLS was implemented under cardiac massage, using a centrifugal pump connected to a hollow-fiber membrane-oxygenator, aiming to maintain ECLS flow ≥2.5 l/min and mean arterial pressure ≥60 mm Hg. Results: Forty-seven of 66 patients died within 24 h from multiorgan failure and massive capillary leak. Of 19/66 patients who survived ≥24 h with stable circulatory conditions permitting neurological evaluation, four became conscious and were transferred for further cardiac assistance, while three became organ donors. Ultimately, one patient survived without neurologic sequelae after cardiac transplantation. Using multivariate analysis, only pre-cannulation peripheral venous oxygen saturation (SpvO2 , 28% [15–52]) independently predicted inability to maintain targeted ECLS conditions ≥24 h (odds ratio for each 10%-decrease [95%-confidence interval]: 1.65 [1.21; 2.25], p = 0.002). The area under the receiveroperating-characteristics curve was 0.78 [0.63; 0.93]. SpvO2 cut-off value of 33% was associated with a sensitivity of 0.68 [0.50; 0.83] and specificity of 0.81 [0.54; 0.96]. SpvO2 ≤8%, lactate concentration ≥21 mmol/l, fibrinogen ≤0.8 g/l, and prothrombin index ≤11% predicted premature ECLS discontinuation with a specificity of 1. Conclusion: SpvO2 is useful to predict the inability of maintaining refractory cardiac arrest victims on ECLS without detrimental capillary leak and multiorgan failure until neurological evaluation. © 2011 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Extracorporeal life support (ECLS) has been proposed as the ultimate heroic rescue measure in prolonged cardiac arrest unresponsive to conventional cardiopulmonary resuscitation (CPR).1–5 In selected patients, ECLS implantation has resulted in some unexpected survivals with excellent neurological outcome, mainly in severe accidental hypothermia6 or poisonings involving cardiotoxicants.2,7 Additionally, ECLS has been shown to be life-

夽 A Spanish translated version of the abstract of this article appears as Appendix in the final online version at doi:10.1016/j.resuscitation.2011.05.007. ∗ Corresponding author. Tel.: +33 149 956 491; fax: +33 149 956 578. E-mail addresses: [email protected], [email protected] (B. Mégarbane). 0300-9572/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.resuscitation.2011.05.007

saving in terminating ventricular tachycardia8 and reversing cardiac arrest complicating myocardial infarction in the catheterization laboratory.3 However, all these reports have been based on limited series. Recently, a larger observational study established the survival benefit of ECLS-based CPR over conventional CPR in in-hospital cardiac arrests of cardiac origin.1 ECLS effectiveness in out-of-hospital cardiac arrest remains debated: a recent meta-analysis suggested possible good neurological recovery using extracorporeal CPR;9 however, in persistent cardiac arrests, outcome was not clearly addressed. Another recent study showed that early attainment of a core temperature of 34 ◦ C had neurological benefits for patients with out-of-hospital cardiac arrest who underwent ECLS plus intra-aortic balloon pumping, with subsequent percutaneous coronary intervention if needed.10 In all reports, decision to discontinue CPR due to medical futility is based upon presumed prolonged anoxia, with existing

B. Mégarbane et al. / Resuscitation 82 (2011) 1154–1161

guidelines for termination.11–13 Early criteria predictive of survival or death with cardiopulmonary bypass following persistent cardiac arrest are unknown. French guidelines regarding ECLS indications in refractory cardiac arrest were proposed based on expert consensus.14 However, even when ECLS is implemented, failure to maintain stable hemodynamic conditions due to marked capillary leak frequently results in patient’s death. This study was designed to assess the usefulness of rapidly available routine laboratory parameters on cannulation for predicting the development of early multiple organ failure and thus ECLS premature discontinuation, if implemented in refractory cardiac arrest. Our aim was to provide quantitative parameters which may be helpful in deciding whether or not it would be worthwhile to perform ECLS, thus limiting the procedure to patients for whom it could be really beneficial.

2. Methods 2.1. Experience in ECLS Our medical intensive care unit (ICU) is located in a university hospital, with the largest emergency department in Paris area, but no cardiovascular surgery department. Annually, we admit approximately 1000 patients, including 100 cardiac arrests. Based on CEDIT’s (an international agency for new health technology) recommendations,15 we trained our team to perform emergent ECLS in our ICU in tight collaboration with the cardiothoracic surgery department of a neighboring hospital.2 2.2. Refractory cardiac arrest definition Patients presenting in cardiac arrest who had received continuous CPR for at least 30 min without return in spontaneous circulation (ROSC) were considered to be refractory and therefore candidates for ECLS support. In France, resuscitation of at least 30 min of continuous cardiac massage is mandatory before cardiac arrest is determined to be refractory, whether in- or out-ofhospital.14 Therefore, transport to hospital of cardiac arrest victims without ROSC cannot be considered until after 30 min of CPR, in contrast to other systems (“scoop and run” concept). 2.3. Patient selection All adults admitted consecutively over a four-year period for witnessed refractory in- or out-of-hospital cardiac arrest presumed or confirmed to be of cardiac etiology, except if related to poisoning (based on history) or hypothermia (body core temperature < 30 ◦ C on admission) were included if treated with ECLS. Victims of refractory cardiac arrest occurring in other hospitals and transferred to our institution for ECLS were considered as in-hospital. Poisoned and severely hypothermic patients were not included, as sufficient data exist to support ECLS use in these patients, primarily due to the potential reversibility of cardiac arrest once the poisoning resolves or hypothermia is corrected. Decision to perform ECLS was taken case-by-case by a senior physician in our ICU, based on patient’s history and presentation (presence of signs of life including movements, gasping, or constricted pupils) in accordance with recommendations.14 No laboratory parameter was used in the decision. This prospective study, approved by our institutional review board, was conducted according to Helsinki’s principles. Whenever possible, verbal consent was obtained from the next-of-kin for both patient’s management with ECLS and study enrollment, after information regarding the purposes and risks of ECLS procedure had been discussed.

1155

2.4. Cardiopulmonary resuscitation CPR was initiated by bystanders and continued by mobile medical team or firefighters at the scene and en-route to hospital. Advanced cardiac life-support (ACLS) was performed according to international guidelines. During transport, cardiac massage was performed either manually or using an automated loaddistributing chest compression device (AutoPulse® , Zoll, France). During ECLS implantation, cardiac massage was provided by the Lund University Cardiac Arrest System (LUCAS® , Jolife, Sweden), a gas-driven sternal compression device that incorporates a suction cup for active decompression. 2.5. ECLS management during ECLS ECLS was set up using a portable centrifugal pump (Rotaflow® , Jostra-Maquet, France) with hollow-fiber membrane oxygenator (Quadrox) and dedicated circuits (BEHQV 50600). Femorofemoral cannulation (15- to 17-Fr arterial and 23- to 29-Fr venous Jostra® cannulae) was performed using Seldinger technique. An additional 7-Fr catheter was inserted for distal limb perfusion. Venous cannula position was confirmed by echocardiography during the procedure while under CPR and confirmed radiologically when a stable ECLS flow was achieved. Extracorporeal blood flow was adjusted to maintain adequate systemic blood flow and oxygen supply as monitored by mean arterial pressure, urine output, and lactate concentrations. Arterial pressure tracing was strictly monitored for reappearing pulsatile systemic blood flow, indicating residual left ventricular myocardial contractility facilitating left ventricular drainage. Fluids and vasopressors were infused to maintain a mean blood pressure of ≥60 mm Hg and flow rate of 3.5–4 l/min to preserve organ function. Dobutamine (10 ␮g/kg per minute) was infused to facilitate left ventricular decompression, minimizing the risks of pulmonary and left ventricle blood stasis as well as intracardiac clotting. Heparin was infused to maintain activated clotting time 2.0–2.5 times higher than control. Patients were mechanically ventilated with 5–6 ml/kg tidal volume and 10-cmH2 O positive end-expiratory pressure. Mild hypothermia (33 ◦ C) was induced by 1 l ice-cold saline infused on ICU admission then maintained during 12–24 h using a heater–cooler unit. No sedation was continued until neurological evaluation. Echocardiography was performed twice daily to assess myocardial contractility. ECLS status was considered to be unstable if a mean blood pressure of ≥60 mm Hg and flow rate of ≥2.5 l/min were impossible to maintain despite fluids and vasopressors. Unstable ECLS resulted in multiorgan failure, alveolar hemorrhage, and capillary leak syndrome, defined as generalized edema with persistent hypovolemia despite vascular repletion. Under these conditions, ECLS was prematurely discontinued. Clinical evaluation of neurological status was assessed at 24 h if hemodynamic conditions were stable after patient rewarming. Serial electroencephalograms were obtained if the patient remained unresponsive. After 24 h, decision to discontinue ECLS support was based upon evidence of persistent multiorgan failure, overwhelming sepsis, or severe neurological injury. If the patient survived, prerequisites to wean from ECLS were echocardiography assessment of myocardial function (left ventricular ejection fraction > 50%) and radial PaO2 /FiO2 > 150 mm Hg. If evaluation excluded severe brain damage independently of ROSC achievement, the patient was transferred under ECLS to the cardio-surgical ward to be considered for ventricular-assist device or cardiac transplantation. If brain death was determined, organ donation was considered in agreement with the French law.

1156

B. Mégarbane et al. / Resuscitation 82 (2011) 1154–1161

2.6. Measurements Utstein template variables regarding resuscitation were collected. Venous and arterial blood samplings were performed at the jugular and radial sites, respectively, under cardiac massage before surgical cut-down and cannula implantation. Peripheral venous oxygen saturation (SpvO2 ), arterial blood gases, and lactate concentration were determined at the bedside (RapidSystemsTM , Siemens, UK) and collected. Serum potassium, creatinine, aminotransferases, platelet count, prothrombin index, activated clotting time ratio, and serum fibrinogen were also collected. Organ system dysfunction was assessed using standard definitions. Physiologic variables measured at admission were used to calculate the Simplified Acute Physiology Score (SAPS)II.16 For analysis, the patient population was split into two subgroups according to ECLS duration: <24 h (premature discontinuation) and >24 h (stable hemodynamic conditions allowing objective neurological evaluation). Outcomes evaluated included survival at 24 h, transfer to cardiothoracic department, and follow-up at one-year after discharge with Glasgow–Pittsburgh Cerebral Performance Category (CPC) determination.17 2.7. Statistical analysis Data are presented as medians [25–75% percentiles] or percentages. Comparisons between groups were performed using Student t-tests or, if not applicable, Mann–Whitney U-tests for continuous variables. Categorical variables were compared using Chi-square tests or, if not applicable, Fischer’s exact tests. Failure to reach stable ECLS ≥24 h to allow objective neurological evaluation was considered as resulting in rapid death and thus rendering ECLS “futile”. A logistic regression model was performed to evaluate the potential contribution of each parameter obtained before cannulation in predicting further premature discontinuation. Statistically significant variables at a 10%-threshold in the univariate analysis were introduced into a stepwise multivariate model to select independent predictive factors of this endpoint. The odds ratios were reported with their corresponding 95%-exact confidence intervals; for SpvO2 , we reported the odds ratios corresponding to the 10% of decrease. For each significant continuous variable, the areaunder-the receiver-operating-characteristics curve (AUC-ROC) was calculated. We calculated several cut-offs which maximize sensitivity, specificity, sum of sensitivity and specificity, and accuracy. For all binary variables, we reported their sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and their corresponding 95%-exact confidence intervals. We chose the cut-offs maximizing specificity as we aimed to be sure that the decision not to implant ECLS would not change the patient’s final outcome. We calculated several scores to save the same specificity level while trying to increase sensitivity. Analyses were two-sided and differences with p-values < 0.05 were considered statistically significant. Analyses were carried out using the SAS-9.2 software (SAS Inc., USA) 3. Results From 2005 to 2008, sixty-six consecutive refractory cardiac arrests (51M/15F, 46 years [39–55]) undergoing continuous cardiac massage were treated with ECLS (Fig. 1). Cardiac arrest occurred either out-of- (71%) or in-hospital (29%), including the emergency room (N = 3), medical wards (N = 9), ICU (N = 5), operating (N = 1), and labor room (N = 1). The presumed cardiac cause of arrest was confirmed based on patient’s history (cardiac disease: 23%) or autopsy, including acute myocardial infarction (39%), pulmonary embolism (11%), end-stage cardiomyopathy (8%),

cardiomyopathy-related or idiopathic ventricular dysrrhythmia (8%), and myocarditis (3%); while remained undetermined (32%). The no-flow period was estimated at 2 min [0–6]; bystanders initiated CPR immediately in 29 patients (43%). Initial rhythm was asystole (53%), ventricular fibrillation (45%), or electromechanical dissociation (2%). The total epinephrine bolus-dose was 13 mg [8–20]. Patients received amiodarone (32%) and/or lidocaine (2%). On cannulation, all patients presented in asystole or electromechanical dissociation with enlarged ventricular complexes. CPR duration before ECLS initiation was 155 min [120–180], with a 110 min [80–135]-delay for ICU transfer. Cannulation was successful in all but one patient, found on autopsy to have major atherosclerosis-related narrowing of the femoral artery diameter. Early complications included bleeding at the cannulation site requiring surgical revision (N = 1) and massive transfusions (4 units [2–6] of erythrocytes and 4 units [2–6] of fresh frozen plasma in the first 24 h). In one patient, fasciotomy was required to treat the ischemic cannulated lower limb. ECLS duration was 8 h [3–26]. Targeted hemodynamic conditions (ECLS-generated flow rate of ≥2.5 l/min with mean blood pressure of ≥60 mm Hg) were initially achieved in 57/66 patients (86%). However, these conditions were maintained beyond 24 h in only 19/66 (29%) patients. In all other patients, death resulted from multiorgan failure associated with major capillary leak and pulmonary hemorrhage. Among the 19 survivors at 24 h, brain death was declared in six patients (9%) permitting organ donation in three cases (Fig. 1). Four patients (6%) became progressively conscious (Glasgow coma score: 15) and were transported under ECLS to the collaborating cardiothoracic surgery department, for consideration of further ventricular assistance or cardiac transplantation, due to irreversible cardiac injury. Ultimately, one 65-year old female patient suffering from end-stage dilated cardiomyopathy (out-of-hospital arrest, CPR duration: 130 min, SpvO2 : 79%, lactate: 14.4 mmol/l, prothrombin index: 11%, activated clotting time: 2.84, and fibrinogen: 1.92 g/l) survived cardiac transplantation, with CPC 1 at one-year follow-up. The three other patients died from hospital-acquired pneumonia (N = 1), hemorrhage (N = 1), and myocardial infarction-related intraventricular communication (N = 1). Pre-cannulation laboratory parameters are presented in Table 1 with univariate comparisons according to patient’s outcome at 24 h. The corresponding odds ratios are given in Table 2. Using a multivariate analysis, SpvO2 was the only independent factor that could predict the development of early multiple organ failure (odds ratio for each 10%-decrease [95%-confidence interval]: 1.65 [1.21; 2.25], p = 0.002). We determined the ROC curves and their areas-under-the-curve for all five parameters that were significantly different in the univariate analysis and could be measured before cannulation (Fig. 2). We determined the ability of various threshold values of these parameters to predict further the development of early multiple organ failure (Table 3). Regarding SpvO2 , a threshold of 33% maximized both sensitivity (0.68) and specificity (0.81). A threshold of 80% was associated with a sensitivity of 1, while a threshold of 8% with a specificity of 1. Combining the two bedside parameters SpvO2 ≤8% and lactate ≥21 mmol/l allowed to reach a sensitivity of 0.27, while combining SpvO2 ≤8% and fibrinogen ≥0.8 g/l a sensitivity of 0.47.

4. Discussion Over a four-year period, sixty-six patients with refractory cardiac arrests of presumed cardiac origin and unrelated to poisoning or hypothermia were managed using ECLS in our ICU. Four conscious patients were discharged to the cardiosurgical ward for further assistance and ultimately, one patient

B. Mégarbane et al. / Resuscitation 82 (2011) 1154–1161

1157

66 patients with refractory cardiac arrest admitted in our medical ICU

1 patient with cannulation failure

65 patients with successful femoral cannulation

46 patients with unstable hemodynamic conditions leading to ECLS discontinuation within the first 24h

« Futile » ECLS

19 patients with stable hemodynamic conditions 24 h

Possible neurological evaluation

4 conscious patients transferred to a cardiothoracic surgery department (for bridge to ventricular assist device or transplantation)

6 brain deaths

47 deaths with multiorgan failure and capillary leak syndrome

12 deaths with severe anoxic brain damage

3 patients allowing organ donation

3 deaths with severe sepsis or complication of surgical procedures

1 survivor

Fig. 1. Outcome of sixty-six victims of refractory cardiac arrests treated with extracorporeal life support (ECLS) in a medical intensive care unit (ICU).

Table 1 Pre-cannulation parameters according to the patients’ outcome at 24 h: patients who survived >24 h versus patients who died within 24 h. Quantitative variables are expressed as median [25–75% percentiles].

Location of cardiac arrest (out-of versus in-hospital) Initial rhythm (ventricular fibrillation versus asystole + electromechanical dissociation) No flow period (min) Delay to ICU admission (min)a Duration of cardiac massage (min) SAPS II value on admissionb Arterial pH PaCO2 (mm Hg) PaO2 /FiO2 ratio (mm Hg) SpvO2 (%) Serum bicarbonate concentration (mmol/l) Serum potassium concentration (mmol/l) Plasma lactate concentration (mmol/l) Serum creatinine concentration (␮mol/l) Platelets (G/l) Prothrombin indexc (%) Activated clotting timed Fibrinogen (g/l) Aspartate aminotransferase (IU/l) Alanine aminotransferase (IU/l) Bold indicates p value < 0.05. a Intensive care unit. b Simplified Acute Physiology Score. c Expressed as percentage of normal value. d Expressed as ratio of the patient’s time to the normal time.

All patients (N = 66)

Survivors ≥24 h (N = 19)

Deaths within 24 h (N = 47)

p

47/19 30/36

13/6 11/8

34/13 19/28

0.8 0.3

2 [0–6] 110 [80–135] 155 [120–180] 90 [79–95] 6.93 [6.79–7.08] 40 [28–59] 114 [69–276] 28 [15–52] 8.3 [6.6–15.1] 5.0 [4.4–6.3] 15.4 [13.2–18.7] 138 [116–155] 82 [58–130] 31 [14–44] 4.2 [2.0–7.2] 1.3 [0.5–2.4] 353 [195–811] 261 [156–428]

1 [0–5] 102 [60–134] 138 [90–170] 82 [67–95] 6.93 [6.91–7.08] 37 [24–56] 197 [82–302] 53 [34–76] 8.2 [5.4–13.2] 5.1 [4.3–5.7] 13.8 [11.8–15.3] 141 [125–159] 103 [59–145] 41 [32–49] 2.1 [1.6–4.2] 1.9 [1.3–3.0] 373 [195–795] 264 [146–412]

3 [0–6] 114 [86–135] 165 [137–180] 92 [84–97] 6.95 [6.73–7.09] 46 [30–59] 89 [62–214] 20 [12–43] 9.1 [6.8–15.7] 5.0 [4.5–6.4] 16.8 [13.9–19.6] 135 [111–155] 74 [57–124] 23 [10–41] 5.1 [2.5–7.2] 0.9 [0.2–2.0] 347 [194–837] 257 [159–451]

0.6 0.3 0.05 0.04 0.2 0.1 0.2 0.0009 0.5 0.4 0.02 0.2 0.2 0.02 0.002 0.003 0.8 0.7

1158

B. Mégarbane et al. / Resuscitation 82 (2011) 1154–1161

Table 2 Univariate logistic regression. Parameters

Odds ratio [95%-confidence interval]

p

Location of cardiac arrest (out-of versus intra-hospital) Initial rhythm (asystole versus ventricular fibrillation) No flow period Delay between cardiac arrest and cannulation decision Duration of cardiac massage SAPS II value on admissiona Arterial pH PaCO2 PaO2 /FiO2 ratio SpvO2 (for each 10%-decrease) Serum bicarbonate concentration Serum potassium concentration Plasma lactate concentration Serum creatinine concentration Platelets Prothrombin index Activated clotting time Serum fibrinogen concentration Aspartate aminotransferase Alanine aminotransferase

0.83 [0.26; 2.64] 0.49 [0.17; 1.45] 0.95 [0.83; 1.08] 0.99 [0.98; 1.01] 0.99 [0.97; 1.00] 0.96 [0.92; 0.99] 5.03 [0.46; 55.43] 0.98 [0.96; 1.01] 1.00 [0.99; 1.00] 1.65 [1.21; 2.25] 0.96 [0.88; 1.04] 0.79 [0.54; 1.16] 0.84 [0.72; 0.99] 1.01 [0.99; 1.01] 1.00 [0.99; 1.01] 1.03 [1.00; 1.07] 0.71 [0.54; 0.92] 1.51 [0.96; 2.37] 1.00 [0.99; 1.00] 0.99 [0.99; 1.00]

0.7 0.3 0.4 0.3 0.05 0.03 0.2 0.2 0.4 0.002 0.3 0.2 0.03 0.2 0.4 0.04 0.009 0.08 0.4 0.4

Bold indicates p value < 0.05. a Simplified Acute Physiology Score.

survived without significant neurological sequelae. Our results clearly confirmed that emergent ECLS implementation using femoral cannulation is feasible in medical ICU with limited complications.2 However, despite optimization of ECLS management in our ICU, delays to ECLS implementation remained

significantly prolonged in relation to the required delays to refer patients with refractory cardiac arrest to our hospital, as medical teams respected the recommended 30 min of ACLS at the scene14 before transporting the patient under cardiac massage. Thus, concerns regarding ECLS futility in cardiac arrest victims with

Fig. 2. Receiver operating characteristics (ROC) curves showing the relationship between sensitivity (true positive) and 1-specificity (true negative) in determining the interest of peripheral venous oxygen saturation (SpvO2 ), plasma lactate concentration, prothrombin index, activated clotting time, and serum fibrinogen concentration obtained before cannulation to predict the development of early multiple organ failure and thus premature extracorporeal flow discontinuation within the first 24 h. The areas under the ROC curves are given at the upper part of each graph. All were significantly different (p < 0.001) from the no-discrimination curve (plain line).

B. Mégarbane et al. / Resuscitation 82 (2011) 1154–1161

1159

Table 3 Interest of various pre-cannulation cut-off levels of five routine parameters to predict the development of early multiple organ failure in cannulated victims of refractory cardiac arrest. Values are expressed with their 95%-confidence intervals. Parameters

Cut-offs

Sensitivity

Specificity

Positive predictive value

Negative predictive value

Accuracy

Peripheral venous oxygen saturation

33%a 80%b 8%c 55%d 0.8 g/la 3.2 g/lb 0.8 g/lc 3.2 g/ld 7.1a 1.2b 7.1c 2.1d 14.5 mmol/la 9.9 mmol/lb 21.0 mmol/lc 10.4 mmol/ld 32%a 62%b 11%c 62%d

0.68 [0.50; 0.83] 1.00 [0.89; 1.00] 0.18 [0.07; 0.35] 0.97 [0.85; 1.00] 0.47 [0.31; 0.62] 0.98 [0.88; 1.00] 0.47 (0.31; 0.62] 0.98 [0.88; 1.00] 0.48 [0.33; 0.63] 0.98 [0.89; 1.00] 0.48 [0.33; 0.63] 0.83 [0.69; 0.92] 0.72 [0.57; 0.84] 1.00 [0.93; 1.00] 0.17 [0.07; 0.31] 0.98 [0.89; 1.00] 0.64 [0.49; 0.78] 0.98 [0.88; 1.00] 0.31 [0.18; 0.47] 0.98 [0.88; 1.00]

0.81 [0.54; 0.96] 0.18 [0.04; 0.46] 1.00 [0.79; 1.00] 0.50 [0.25; 0.75] 1.00 [0.82; 1.00] 0.16 [0.03; 0.40] 1.00 [0.82; 1.00] 0.16 [0.34; 0.40] 0.95 [0.74; 1.00] 0.11 [0.01; 0.33] 0.95 [0.74; 1.00] 0.53 [0.29; 0.76] 0.68 [0.44; 0.87] 0.05 [0.00; 0.26] 1.00 [0.82; 1.00] 0.11 [0.01; 0.33] 0.79 [0.54; 0.94] 0.05 [0.00; 0.26] 1.00 [0.82; 1.00] 0.05 [0.00; 0.26]

0.89 [0.70; 0.98] 0.72 [0.57; 0.84] 1.00 [0.54; 1.00] 0.81 [0.65; 0.91] 1.00 [0.83; 1.00] 0.72 [0.59; 0.83] 1.00 [0.83; 1.00] 0.72 [0.59; 0.83] 0.96 [0.78; 1.00] 0.73 [0.60; 0.83] 0.96 [0.78; 1.00] 0.81 [0.67; 0.91] 0.85 [0.70; 0.94] 0.72 [0.60; 0.83] 1.00 [0.63; 1.00] 0.73 [0.60; 0.83] 0.88 [0.72; 0.97] 0.71 [0.58; 0.82] 1.00 [0.77; 1.00] 0.71 [0.58; 0.82]

0.54 [0.33; 0.75] 1.00 [0.29; 1.00] 0.36 [0.22; 0.52] 0.89 [0.52; 1.00] 0.45 [0.30; 0.61] 0.75 [0.19; 0.99] 0.45 [0.30; 0.61] 0.75 [0.19; 0.99] 0.43 [0.28; 0.59] 0.67 [0.09; 0.99] 0.43 [0.28; 0.59] 0.56 [0.31; 0.79] 0.50 [0.30; 0.70] 1.00 [0.03; 1.00] 0.33 [0.21; 0.46] 0.67 [0.09; 0.99] 0.48 [0.30; 0.67] 0.50 [0.0; 0.99] 0.38 [0.25; 0.53] 0.50 [0.0; 0.99]

0.72 [0.58; 0.84] 0.74 [0.60; 0.85] 0.44 [0.30; 0.59] 0.82 [0.69; 0.91] 0.63 [0.50; 0.75] 0.73 [0.60; 0.83] 0.63 [0.50; 0.75] 0.73 [0.60; 0.83] 0.62 [0.49; 0.73] 0.72 [0.60; 0.83] 0.62 [0.49; 0.73] 0.74 [0.62; 0.84] 0.71 [0.59; 0.82] 0.73 [0.60; 0.83] 0.41 [0.29; 0.54] 0.73 [0.60; 0.83] 0.69 [0.56; 0.80] 0.70 [0.58; 0.81] 0.52 [0.39; 0.64] 0.70 [0.58; 0.81]

Serum fibrinogen concentration

Activated clotting time

Plasma lactate concentration

Prothrombin index

a b c d

Cut-off to maximize sensitivity + specificity sum. Cut-off to maximize sensitivity. Cut-off to maximize specificity. Cut-off to maximize accuracy.

presumably irreversible cardiac dysfunction following prolonged CPR, along with pos-anoxic neurologic injury from prolonged anoxia, represent a serious limitation for application of this technique. Variable survival rates (0–64%) have been reported with cardiopulmonary bypass following cardiac arrest.1–5,9,10,18–25 Differences in outcome are related to the small cohorts, specific etiologies, incident location, variable time delays until initiation of circulatory support, and varying experiences with this technique. In patients experiencing myocardial infarction with cardiac arrest in the catheterization laboratory, ECLS resulted in one survivor out of eight patients.3 In victims of in-hospital cardiac arrest of cardiac origin undergoing CPR for ≥10 min, treatment with ECLS resulted in 15.3% of survivors at one-year follow-up with CPC 1 or 2.1 In selected out-of-hospital cardiac arrests, ECLS may be lifesaving, provided that the patient has not sustained hypoxic cerebral damage. Consistently, Chen et al.25 hesitated to recommend ECLS for out-of hospital CPR because of the uncertain duration of arrest. However, they recognized that ECLS may offer an acceptable survival rate in prolonged CPR up to 60 min with 30%-probability of survival. In our study, duration of cardiac massage until ECLS implementation (155 min [120–180]) was longer than in others: With a 105 ± 44 min duration, three patients survived in Massetti’s study despite irreversible cardiac dysfunction, while bridged to a ventricular-assist device (N = 2) and cardiac transplantation (N = 1).4 However, the majority of Massetti’s cardiac arrests occurred and were treated within the hospital in charge, with a shorter time to cannulation. Interestingly, our survival rate was in the probability range (<10%) reported in ECLS-treated intra-hospital cardiac arrests when CPR duration was >90 min.25 In patients who died before 24 h, capillary leak developed and multiorgan failure persisted. Massive edema which worsens vital organ dysfunction is first related to the prolonged cardiac arrest, as reported in our series. Additionally, ECLS-induced leak has been attributed to permeability and lymph edema, as fluid was found relatively protein-poor in experimental models.26 Other authors have suggested leukocyte activation and endothelial cell dysfunction in response to prolonged blood contact with synthetic ECLS surfaces.27,28 Here, both mechanisms were possible due to cardiac

arrest-related endothelial cell damage and vascular tone dysfunction. In cardiac arrests, decision to terminate CPR is based upon presumed prolonged anoxia, with existing advisory rules for termination.11,12 Probability of survival is low, but especially grim in out-of-hospital arrests.. The time delay from cardiac arrest to admission in our ICU and finally to ECLS implementation was very long. Many other centres would consider initiation of ECLS futile in these conditions. We are convinced that shortening time delay to ECLS remains the key-priority to improve outcome; however, this objective appeared as a serious limitation to the French concept that transport to hospital of cardiac arrest victims without ROSC cannot be considered until after 30 min of CPR.14 Nevertheless, as with our 65-year old patient, long-term survival is conceivable after ECLS in selected patients with underlying cardio-circulatory diseases amenable to corrective intervention (angioplasty, surgery, or transplantation).2,3,5 ECLS allows neurological evaluation before shifting to more sophisticated cost-generating cardiac supports. In order to improve decision-making for implementing ECLS and to avoid “futile” ECLS efforts, we aimed to identify readily laboratory parameters that could predict the occurrence of capillary leak and multiorgan failure, resulting in ECLS discontinuation. We found that SpvO2 is useful to help clarify ECLS indications in prolonged cardiac arrests. Using multivariate analysis, SpvO2 was the only independent factor to predict further stable hemodynamic conditions ≥24 h, mirroring prolonged or insufficient resuscitation. SpvO2 ≤8% predicted the development of early multiple organ failure with a specificity of 1. Similarly, lactate concentration ≥21 mmol/l, fibrinogen ≤0.8 g/l, and prothrombin index ≤11% suggested ECLS futility. Using these criteria, 26/66 (39%) patients should not have been cannulated. However, among these parameters, only SpvO2 and lactate are obtainable within reasonable amounts of time to be useful, especially if hand-held analyzers are available.29 To our knowledge, routine measurement of fibrinogen, prothrombin index, or activated clotting time requires at least 20–30 min, although accurate point-of-care testing is under development.30 We therefore advise that SpvO2 , completed when possible by lactate concentration should be obtained at the bedside prior to patient cannulation. Additionally, we suggest that a strategy based on fluid infusion and optimal cardiac massage deliv-

1160

B. Mégarbane et al. / Resuscitation 82 (2011) 1154–1161

ery during prolonged CPR aimed at increasing SpvO2 may result in better outcome. To date, no definitive criteria to identify optimal candidates exist, making the decision to undertake ECLS difficult and the line between life and death thin.31 As uncontrolled applications of ECLS for out-of-hospital cardiac arrests may ultimately lead to its abandonment because of poor results, guidelines to help physicians’ decisions are needed. Prerequisites for successful ECLS rescue of out-of hospital cardiac arrest should be listed: (1) CPR quality remains the critical step for neurological outcome32 even though long delays to ROSC have resulted in good cognitive outcomes, confirming that duration should not limit CPR efforts.33–36 (2) Although a weak association between cardiac arrest duration and lactate levels on ICU admission has been shown,37 functional neurological recovery was more likely to be unfavorable with increasing lactate concentrations. High arterial lactate levels (>16.3 mmol/l) are predictive of unfavorable neurological recovery, with 100%-specificity. In our series, lactate >21 mmol/l resulted in premature ECLS discontinuation. However, in poisoning-related refractory cardiac arrests, survivals with very elevated lactate concentrations have been reported.2 (3) An endtidal carbon dioxide (EtCO2 ) ≤10 mm Hg measured 20 min after ACLS initiation has been shown to accurately predict death in cardiac arrests with electrical activity but no pulse.38 Here, we did not include this parameter as not systematically obtained by the pre-hospital medical services. However, we believe that in patients arriving with EtCO2 ≤10 mm Hg, ECLS should reasonably not be considered. Limitations to this single-center study exist. Our study was underpowered to yield statistically significant findings, especially regarding the identification of prognosticators of the absence of severe brain damage like in our four patients who were discharged from our ICU to the cardiosurgical ward. Inclusion of various cardiac arrest etiologies with various degrees of irreversible heart dysfunction may also be a limitation. Decisions to stop CPR and not to transport patients under massage to our ICU were independent of our team, making any estimation of the proportion of refractory cardiac arrests included in this study difficult. Due to the prolonged delay in ECLS implementation in our series which resulted in a small number of survivors (1 out of 66), our results preclude overarching recommendations regarding the use of SpvO2 in the setting of early ECLS implementation. Moreover, time to ROSC under ECLS was not evaluated. Another limitation is the inability to evaluate the potential value of CPR assist devices, although they appeared to be helpful in facilitating patient transport and ECLS management. Finally, we are aware of the need for evidence of cost-effectiveness.39,40

5. Conclusion Based on our experience, ECLS remains an investigational and compassionate treatment in patients with refractory cardiac arrest from a non-toxic and non-hypothermic origin, following prolonged CPR. SpvO2 is helpful in identifying patients who will likely die within the first 24 h due to capillary leak and multiorgan failure. When SpvO2 is ≤8%, plasma lactate concentration ≥21 mmol/l, fibrinogen ≤0.8 g/l, or prothrombin index ≤11% after prolonged CPR, we do not recommend implementing ECLS which could be considered futile under these conditions.

Conflict of interest statement None declared.

Acknowledgement The authors would like to acknowledge Jenny Lu, MD, from the Dept of Emergency Medicine and Division of Toxicology, Cook County Hospital, Chicago, USA for her helpful review of this manuscript.

References 1. Chen YS, Lin JW, Yu HY, et al. Cardiopulmonary resuscitation with assisted extracorporeal life-support versus conventional cardiopulmonary resuscitation in adults with in-hospital cardiac arrest: an observational study and propensity analysis. Lancet 2008;372:554–61. 2. Mégarbane B, Leprince P, Deye N, et al. Emergency feasibility in medical intensive care unit of extracorporeal life support for refractory cardiac arrest. Intensive Care Med 2007;33:758–64. 3. Chen JS, Ko WJ, Yu HY, et al. Analysis of the outcome for patients experiencing myocardial infarction and cardiopulmonary resuscitation refractory to conventional therapies necessitating extracorporeal life support rescue. Crit Care Med 2006;34:950–7. 4. Massetti M, Tasle M, Le Page O, et al. Back from irreversibility: extracorporeal life support for prolonged cardiac arrest. Ann Thorac Surg 2005;79:178–83. 5. Schwarz B, Mair P, Margreiter J, et al. Experience with percutaneous venoarterial cardiopulmonary bypass for emergency circulatory support. Crit Care Med 2003;31:758–64. 6. Gilbert M, Busund R, Skagseth A, Nilsen PA, Solbø JP. Resuscitation from accidental hypothermia of 13.7 ◦ C with circulatory arrest. Lancet 2000;355:375–6. 7. Daubin C, Lehoux P, Ivascau C, et al. Extracorporeal life support in severe drug intoxication: a retrospective cohort study of seventeen cases. Crit Care 2009;13:R138. 8. Tsai FC, Wang YC, Huang YK, et al. Extracorporeal life support to terminate refractory ventricular tachycardia. Crit Care Med 2007;35:1673–6. 9. Morimura N, Sakamoto T, Nagao K, et al. Extracorporeal cardiopulmonary resuscitation for out-of-hospital cardiac arrest: a review of the Japanese literature. Resuscitation 2011;82:10–4. 10. Nagao K, Kikushima K, Watanabe K, et al. Early induction of hypothermia during cardiac arrest improves neurological outcomes in patients with out-of-hospital cardiac arrest who undergo emergency cardiopulmonary bypass and percutaneous coronary intervention. Circ J 2010;74:77–85. 11. Horsted TI, Rasmussen LS, Lippert FK, Nielsen SL. Outcome of out-of-hospital cardiac arrest—why do physicians withhold resuscitation attempts? Resuscitation 2004;63:287–93. 12. Morrison LJ, Visentin LM, Kiss A, et al. Validation of a rule for termination of resuscitation in out-of-hospital cardiac arrest. N Engl J Med 2006;355:478–87. 13. Hill JG, Bruhn PS, Cohen SE, et al. Emergent applications of cardiopulmonary support: a multiinstitutional experience. Ann Thorac Surg 1992;54:699–704. 14. Conseil franc¸ais de réanimation cardiopulmonaire, Société franc¸aise d’anesthésie et de réanimation, Société franc¸aise de cardiologie, et al. Guidelines for indications for the use of extracorporeal life support in refractory cardiac arrest. Ann Fr Anesth Reanim 2009;28:182–90. 15. Comité d’Evaluation et de Diffusion des Innovations Technlogiques, CEDIT, http://cedit.aphp.fr/servlet/siteCeditGB?Destination=reco&numArticle=03.03/ Re1/04; 2004. 16. Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA 1993;270:2957–63. 17. Booth CM, Boone RH, Tomlinson G, Detsky AS. Is this patient dead, vegetative, or severely neurologically impaired? Assessing outcome for comatose survivors of cardiac arrest. JAMA 2004;291:870–9. 18. Nagao K, Hayashi N, Kanmatsuse K, et al. Cardiopulmonary cerebral resuscitation using emergency cardiopulmonary bypass, coronary reperfusion therapy and mild hypothermia in patients with cardiac arrest outside the hospital. J Am Coll Cardiol 2000;36:776–83. 19. Younger JG, Schreiner RJ, Swaniker F, Hirschl RB, Chapman RA, Bartlett RH. Extracorporeal resuscitation of cardiac arrest. Acad Emerg Med 1999;6:700–7. 20. Martin GB, Rivers EP, Paradis NA, Goetting MG, Morris DC, Nowak RM. Emergency department cardiopulmonary bypass in the treatment of human cardiac arrest. Chest 1998;113:743–51. 21. Mair P, Hoermann C, Moertl M, Bonatti J, Falbesoner C, Balogh D. Percutaneous venoarterial extracorporeal membrane oxygenation for emergency mechanical circulatory support. Resuscitation 1996;33:29–34. 22. Hartz R, LoCicero 3rd J, Sanders Jr JH, et al. Clinical experience with portable cardiopulmonary bypass in cardiac arrest patients. Ann Thorac Surg 1990;50:437–41. 23. Reichman RT, Joyo CI, Dembitsky WP, et al. Improved patient survival after cardiac arrest using a cardiopulmonary support system. Ann Thorac Surg 1990;49:101–4. 24. Mooney MR, Arom KV, Joyce LD, et al. Emergency cardiopulmonary bypass support in patients with cardiac arrest. J Thorac Cardiovasc Surg 1991;101:450–4. 25. Chen YS, Yu HY, Huang SC, et al. Extracorporeal membrane oxygenation support can extend the duration of cardiopulmonary resuscitation. Crit Care Med 2008;36:2529–35.

B. Mégarbane et al. / Resuscitation 82 (2011) 1154–1161 26. Heltne JK, Koller ME, Lund T, et al. Studies on fluid extravasation related to induced hypothermia during cardiopulmonary bypass in piglets. Acta Anaesthesiol Scand 2001;45:720–8. 27. Sonntag J, Dahnert I, Stiller B, Hetzer R, Lange PE. Complement and contact activation during cardiovascular operations in infants. Ann Thorac Surg 1998;65:525–31. 28. Graulich J, Walzog B, Marcinkowski M, et al. Leukocyte and endothelial activation in a laboratory model of extracorporeal membrane oxygenation (ECMO). Pediatr Res 2000;48:679–84. 29. Dempfle CE, Borggrefe M. Point of care coagulation tests in critically ill patients. Semin Thromb Hemost 2008;34:445–50. 30. Steinfelder-Visscher J, Weerwind PW, Teerenstra S, Brouwer MH. Reliability of point-of-care hematocrit, blood gas, electrolyte, lactate and glucose measurement during cardiopulmonary bypass. Perfusion 2006;21:33–7. 31. Bracco D, Noiseux N, Hemmerling TM. The thin line between life and death. Intensive Care Med 2007;33:751–4. 32. Wik L, Kramer-Johansen J, Myklebust H, et al. Quality of cardiopulmonary resuscitation during out-of-hospital cardiac arrest. JAMA 2005;293:299–304. 33. Cooper S, Macnaughton P. Prolonged resuscitation: a case report. Resuscitation 2001;50:349–51.

1161

34. Freysz M, Honnart D, Besancon A, Wilkening M. Cardiac arrest after beta-blocker poisoning. Crit Care Med 1986;14:837–8. 35. Vukmir RB. Survival from prehospital cardiac arrest is critically dependent upon response time. Resuscitation 2006;69:229–34. 36. Van Alem AP, de Vos R, Schmand B, Koster RW. Cognitive impairment in survivors of out-of-hospital cardiac arrest. Am Heart J 2004;148: 416–21. 37. Mullner M, Sterz F, Domanovits H, Behringer W, Binder M, Laggner AN. The association between blood lactate concentration on admission, duration of cardiac arrest, and functional neurological recovery in patients resuscitated from ventricular fibrillation. Intensive Care Med 1997;23: 1138–43. 38. Levine RL, Wayne MA, Miller CC. End-tidal carbon dioxide and outcome of outof-hospital cardiac arrest. N Engl J Med 1997;337:301–6. 39. Chen B, Chang YM. CPR with assisted extracorporeal life support. Lancet 2008;372:1879. 40. Mahle WT, Forbess JM, Kirshbom PM, Cuadrado AR, Simsic JM, Kanter KR. Costutility analysis of salvage cardiac extracorporeal membrane oxygenation in children. J Thorac Cardiovasc Surg 2005;129:1084–90.