Extracorporeal Membrane Oxygenation for Acute Respiratory Distress Syndrome After Pneumonectomy

Extracorporeal Membrane Oxygenation for Acute Respiratory Distress Syndrome After Pneumonectomy

Extracorporeal Membrane Oxygenation for Acute Respiratory Distress Syndrome After Pneumonectomy Jeremie Reeb, MD, MS, Anne Olland, MD, PhD, Julien Pot...

891KB Sizes 0 Downloads 29 Views

Extracorporeal Membrane Oxygenation for Acute Respiratory Distress Syndrome After Pneumonectomy Jeremie Reeb, MD, MS, Anne Olland, MD, PhD, Julien Pottecher, MD, PhD, Xavier Delabranche, MD, PhD, Mickael Schaeffer, MS, Stephane Renaud, MD, PhD, Nicola Santelmo, MD, Romain Kessler, MD, PhD, Gilbert Massard, MD, PhD, and Pierre-Emmanuel Falcoz, MD, PhD Division of Thoracic Surgery and Strasbourg Lung Transplant Program, Surgical Intensive Care Unit, Medical Intensive Care Unit, Department of Biostatistics, and Division of Respirology and Strasbourg Lung Transplant Program, Hopitaux Universitaires de Strasbourg, Strasbourg, France ˇ

Background. Postpneumonectomy acute respiratory distress syndrome (ppARDS) is a life-threatening condition with a disastrous prognosis. This study assessed the efficacy of venovenous extracorporeal membrane oxygenation (VV-ECMO) in adult patients with unresponsive severe ppARDS. Methods. We retrospectively reviewed data of all patients treated with VV-ECMO for ppARDS from January 2009 to December 2015. We calculated the Sequential Organ Failure Assessment score before ECMO insertion and monitored the subsequent mechanical ventilation settings. The primary end point was hospital survival. The secondary end point was the ability to achieve a protective ventilatory strategy allowing lung recovery on ECMO. Results. VV-ECMO was indicated in 8 ppARDS patients for refractory hypoxemia (median partial pressure of arterial oxygen/fraction of inspired oxygen: 68 [range,

60 to 75] mm Hg). Median Sequential Organ Failure Assessment before ECMO was 15 (range, 12 to 17), predicting a mortality rate greater than 80%. Median duration of ECMO was 9.5 (range, 5 to 16) days. Tidal volumes and plateau pressures both decreased on ECMO (preECMO tidal volume: 412 [range, 250 to 450 mL] vs ECMO tidal volume: 277 [range, 105 to 367 mL], p [ 0.0156; preECMO plateau pressure: 34 [range, 32 to 40] cm H2O vs ECMO plateau pressure: 24.5 [range, 23.3 to 27.3] cm H2O, p [ 0.0195). ECMO could be weaned in 7 patients (87.5%). Hospital survival was 50%. Conclusions. Hospital survival was better than predicted before ECMO insertion. In severe and refractory ppARDS, VV-ECMO allows lung recovery and therefore increased survival.


Hg) ARDS. These stages of mild, moderate, and severe ARDS are correlated with increasing hospital mortality of 27%, 32%, and 45%, respectively [3]. Prognosis of ARDS after lung resection, and particularly after postpneumonectomy ARDS (ppARDS), seems even more pessimistic. The hospital mortality rates of ppARDS ranged from 33% to 88% in small cohort studies [4–10]. Venovenous extracorporeal membrane oxygenation (VV-ECMO) is an extracorporeal life support technique that allows for the substitution the failing respiratory function. VV-ECMO is frequently associated with a protective ventilation strategy in severe ARDS with the dual rationale (1) to ascertain gas exchanges and (2) to promote lung recovery by resting the lungs [11, 12]. Some case reports describe the successful use of VV-ECMO in ppARDS associated with fistula [13–15]. To the best of our knowledge, no cohort study has been specifically devoted to VV-ECMO in ppARDS. The aim of this study was to assess the efficacy of VV-ECMO in adult patients with severe ARDS after pneumonectomy.

cute respiratory distress syndrome (ARDS) is a rare but a life-threatening complication after pneumonectomy. ARDS was first described in 1967 and consensually defined 27 years later by the AmericanEuropean Consensus Conference definition as an acute onset of hypoxemia (partial pressure of arterial oxygen–to–fraction of inspired oxygen ratio [PaO2/FIO2] 200 mm Hg) with bilateral infiltrates on frontal chest roentgenogram without evidence of associated heart failure [1, 2]. In 2012 the definition of ARDS was revised by a task force in Berlin, which proposed three categories according to the degree of hypoxemia: mild (Pao2/FIO2 <200 to 300 mm Hg), moderate (PaO2/FIO2 <100 to 200 mm Hg), and severe (PaO2/FIO2 100 mm

Accepted for publication Nov 8, 2016. Address correspondence to Dr Falcoz, Division of Thoracic Surgery and Strasbourg Lung Transplant Program, Hopitaux Universitaires de Strasbourg, 1 place de l’Hopital, 67091 Strasbourg Cedex, France; email: [email protected] ˇ


Ó 2017 by The Society of Thoracic Surgeons Published by Elsevier

(Ann Thorac Surg 2017;-:-–-) Ó 2017 by The Society of Thoracic Surgeons

0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2016.11.038



Abbreviations and Acronyms ADC ANOVA ARDS ASA

= = = =


= =








= = = = = = = = =


= = = =

TV = VO2max = VV =

adenocarcinoma analysis of variance acute respiratory distress syndrome American Society of Anesthesiologists body mass index chronic obstructive pulmonary disease diffusing capacity of the lung for carbon monoxide extracorporeal membrane oxygenation forced expiratory volume in 1 second femoral-femoral fraction of inspired oxygen femoral-jugular intensive care unit malignant mesothelioma multiorgan failure not applicable non-small cell lung cancer partial pressure of arterial carbon dioxide partial pressure of arterial oxygen positive end-expiratory pressure postpneumonectomy acute respiratory distress syndrome red blood cells right internal jugular vein squamous cell carcinoma Sequential Organ Failure Assessment tidal volume maximal oxygen consumption venovenous

Patients and Methods The French Society of Thoracic and Cardio-Vascular Surgery Ethical Committee approved this study and waived the need for informed consent.

Study Design We retrospectively reviewed all patients treated with VVECMO for severe and refractory ARDS after pneumonectomy or completion pneumonectomy between January 2009 and December 2015.

Pre-ECMO Management All pneumonectomies or completion pneumonectomies were performed as planned curative operations for nonsmall cell lung cancer or pleural malignancy. All patients were preoperatively assessed by the British Thoracic Society’s guidelines for the selection of patients with lung cancer for resection [16]. The anesthetic management protocol of the pneumonectomy was based on sufentanil, propofol or etomidate, and curare. Thoracic epidural analgesia was performed for every patient of the

Ann Thorac Surg 2017;-:-–-

cohort. The ventilation settings before pneumonectomy were (1) positive end-expiratory pressure (PEEP) 8 cm H2O, (2) tidal volume (TV) 5 to 6 mL/kg, (3) respiratory ratio depending on the end tidal CO2 (target ¼ 30 to 55 mm Hg), and (4) FIO2 enough to ensure an oxygen saturation of 90% or higher. All patients with ARDS had been receiving optimal conventional management at the time when the decision for ECMO was made in a multidisciplinary discussion.

ECMO Initiation Patients were eligible for ECMO in case of severe ARDS according the Berlin Definition of ARDS [3]. Patients were considered for VV-ECMO if they were unresponsive to optimal medical treatment and hemodynamically stable (mean arterial pressure >60 mm Hg; normal left- and right-sided heart functions and sizes) with minimal vasopressive support. Optimal medical therapy was defined by a combination of (1) protective mechanical ventilation (TV: 4 mL/kg ideal body weight, plateau pressure: 30 cm H2O, PEEP titrated according to pressure-volume loops or recruited lung volume), (2) a reduction of instrumental dead space by avoiding heat-moisture exchangers, (3) empiric broad-spectrum antibiotics in case of clinical suspicion of ventilator-associated pneumonia or positive Gram stain on bronchoalveolar lavage, (4) prone positioning, (5) repeated bronchoscopic sanitation, followed by recruitment maneuvers in case of clinical or radiologic suspicion of atelectasis, (6) neuromuscular blocking agent (curare) for at least 6 hours, and (7) inhaled nitric oxide (20 ppm) more or less combined with almitrine therapy (Fig 1) [17]. VV-ECMO was used as a first-line technique because of its higher efficiency in lungs, heart, and brain oxygenation compared with other extracorporeal life support techniques [11, 12]. The overall status and predicted death of the patients before the ECMO insertion was estimated using the Sequential Organ Failure Assessment (SOFA) score, which quantifies the severity of illness with the degree of hypoxemia, platelet level, liver function (bilirubin), level of the cardiovascular support (mean arterial pressure and vasopressor requirement), kidney function (creatinine level or urine output), and Glasgow Coma Score [15]. SOFA scores of 15 and above are correlated with mortality rates exceeding 80% and 90%, respectively [18].

Insertion of VV-ECMO VV-ECMO was implanted in the intensive care unit (ICU) using the Seldinger technique. Cannula size depended on the site of cannulation. For double-site cannulation, the outflow cannula was sized 21F to 25F, and the inflow cannula was 18F to 24F. In this setting, the drainage sites were the right or the left femoral veins, and the injection site was preferentially the right internal jugular vein [19]. Accurate positioning of the 2 cannulas was evaluated clinically and by chest roentgenogram. Single-site cannulation was achieved with a 27F Avalon Elite double-lumen catheter (Maquet Cardiopulmonary GmbH, Rastatt, Germany) implanted into the right

Ann Thorac Surg 2017;-:-–-



Fig 1. Treatment protocol for postpneumonectomy acute respiratory distress syndrome (ppARDS). (PEEP ¼ positive endexpiratory pressure; VV ECMO ¼ venovenous extracorporeal membrane oxygenation.)

internal jugular vein. Single-site cannulation was performed with transthoracic or transesophageal ultrasound and Doppler guidance to ascertain accurate positioning of the distal tip in the inferior vena cava and inflow directed through the tricuspid valve. We administered a heparin bolus of 5,000 IU before cannulation, targeting an activated partial thromboplastin time of 2.5-times to 3-times normal. The centrifugal pumps used were the Revolution pump (Sorin Group, Milan, Italy) or the BioMedicus pump (Medtronic Inc, Minneapolis, MN). We exclusively used the Eurosets oxygenators (Eurosets S.r.1., Medolla, Italy).

Mechanical Ventilation Settings During ECMO Once ECMO was started, we moved to protective ventilation. The initial ventilator settings were (1) pressure control mode, (2) maximum TV, 3 to 4 mL/kg; (3) PEEP, 5 to 10 cm H2O; and (4) plateau pressure of less than 30 cm H2O. Patients were sedated and paralyzed initially. We stopped curare and decreased stepwise sedation levels when patients were stabilized on VV-ECMO and when the systemic inflammatory response came under control. We switched to pressure support ventilation once the lung compliances increased with PEEP of 5 cm H2O or less and monitored plateau pressure of less than 30 cm H2O. Tracheostomy was performed after removal of ECMO when first-line weaning of mechanical ventilation failed.

Overall Management Patients During ECMO ECMO flows were adjusted to radial PaO2 values. The delivered fraction of oxygen ranged from 50% to 100%. The target for radial PaO2 was 60 mm Hg or higher, with a decreasing or normalized level of lactate or a mixed venous oxygen saturation exceeding 70%, or both. The values of sweep gas flow were titrated to the radial partial pressure of arterial carbon dioxide (PaCO2). The target for radial PaCO2 ranged from 35 to 60 mm Hg with a normal pH. Hypercapnia was decreased stepwise; permissive hypercapnia was tolerated in case of previous chronic hypercapnia. The continuous monitoring of the patients and the ECMO was performed by intensivists, thoracic surgeons, perfusionists, dedicated nurses, and physiotherapists. Particularly, daily multidisciplinary rounds were scheduled to assess the circuits, define care, and improve the clinical status of the patients. All patients on ECMO were heparinized with an activated partial thromboplastin time target of 2.0-times to 2.5-times normal. In case of bleeding, the activated partial thromboplastin time target was decreased to 1.5-times to 2.0-times normal. Intravenous heparin infusions were stopped in cases of severe bleeding for no longer than 48 hours. The policy for blood product administration was to transfuse (1) packed red blood cells if hemoglobin levels were less than 8 to less than 10 g/dL in patients with cardiovascular comorbidity, (2) platelets if platelets



counts were less than 50  109/L, and (3) fresh frozen plasma if the international normalization ratio exceeded 1.5 or in case of coagulopathy, or both. Mean arterial pressures were continuously monitored with a target of 65 mm Hg or higher. Vasopressor requirements were assessed with norepinephrine flow before and 6 hours after starting on ECMO. Renal replacement therapy was performed in case of renal failure or fluid overload with oliguria, or both. Patients supported with ECMO were considered for weaning when their pulmonary compliance increased and their clinical status improved (mechanical ventilation with FiO2 <50%, PEEP <5 cm H2O, plateau pressure <30 cm H2O, with TV >4 mL/kg). Sweep gas flow was progressively decreased and switched off in the presence of constant ECMO blood flow. If arterial blood gases remained satisfactory for 4 hours, ECMO was discontinued. Chest roentgenogram was performed on demand, based on clinical or biological arguments.

Study End Points The primary end point was in-hospital survival. Secondary end points were the ability to safely perform protective ventilation on ECMO and the weaning of ECMO and mechanical ventilation, proving the lung recovery.

Data Collection Collected data were demographic data (age, sex, body mass index), characteristics before VV-ECMO (peripneumonectomy characteristics, interval from pneumonectomy, nature of the ARDS, SOFA score, and respiratory and hemodynamic support), VV-ECMO characteristics (ECMO configuration, ECMO settings, associated therapies, and particularly mechanical ventilation variables), and outcomes (hospital survival, ECMO weaning, mechanical ventilation weaning, ICU survival).

Statistical Analysis Quantitative data are expressed as median and range and qualitative data are represented with absolute values and percentages. Continuous variables were compared using the nonparametric Mann-Whitney test. Comparisons between groups were made with one-way analysis of variance. Statistical significance was set at p values of less than 0.05 (with 95% interval confidence). We used GraphPad Prism 6 software (GraphPad Software Inc, La Jolla, CA).

Results Patients The patient characteristics are included in Tables 1 and 2. We performed 232 pneumonectomies for malignancy from January 2009 to December 2015, and 8 patients required VV-ECMO for refractory severe ARDS. No surgical complications or anesthetic issues occurred during the operations of these 8 patients. ARDS

Ann Thorac Surg 2017;-:-–-

complicated a documented pneumonia in the first 7 patients. ARDS developed in the eighth patient after aspiration. Patients were hemodynamically stable before ECMO. We never changed the circuit or switched the configuration of the ECMO. Transfusion rates of blood products were 0.97 (range, 0 to 4.29), 0.52 (range, 0 to 2.86), and 0 (range, 0 to 0.43) U/d for red blood cells, fresh frozen plasmas, and apheresis platelet concentrates, respectively. No adverse events related to the cannulation procedures or the ECMO circuits were observed.

Clinical Outcomes of the Patients Patient outcomes are summarized in Tables 1 and 2. ICU outcomes are illustrated in Figure 2. Because of the severe systemic inflammatory response syndrome in ARDS, none of the patients were weaned from mechanical ventilation before removal of ECMO. One patient supported with ECMO died of multiorgan failure. Two patients (25%) died on mechanical ventilation while they were weaned from ECMO: causes of death were arrhythmia in the first patient and multiorgan failure in the second. We successfully weaned 5 patients (62.5%) from mechanical ventilation. The median duration of mechanical ventilation was 17 (range, 10 to 19) days. One of the patients we successfully weaned from the mechanical ventilation died in the ICU of septic shock 22 days after weaning. Four patients (50%) were discharged alive from the ICU after a median stay of 27 (range, 20 to 38) days. All patients discharged alive from the ICU were discharged alive from the hospital with a median hospital stay of 46 (range, 25 to 121) days.

Lung Recovery On admission to the ICU, the median PaO2/FIO2 and PaCO2 were 94 (range, 66 to 63) and 46.1 (range, 38.1 to 63) mm Hg, respectively. Before insertion of ECMO, all patients were on 100% FIO2. Median TV was 412 (range, 250 to 450) mL, median PEEP was 5.5 (range, 4 to 8) cm H2O, and median plateau pressure was 34 (range, 32 to 40) cm H2O. VV-ECMO was indicated by respiratory acidosis (median pH: 7.32 [range, 7.18 to 7.38]; median PaCO2: 52.5 [range, 46 to 66] mm Hg) and severe and refractory hypoxemia (median PaO2/FIO2: 68 [range, 60 to 75] mm Hg). During ECMO support, median ECMO flow and sweep gas flow were 2.7 (range, 2.43 to 4.4) L/min and 4.6 (range, 2.8 to 8.7) L/min, respectively. Hypoxemia and hypercapnia were both corrected (median PaO2 on ECMO: 92.7 [range, 72 to 136] mm Hg vs pre-ECMO PaO2: 67.5 [range, 60 to 75] mm Hg, p ¼ 0.0038; median PaCO2 on ECMO: 44.0 [range, 33.0 to 48.6] mm Hg vs preECMO PaCO2: 52.5 [range, 46 to 66] mm Hg, p ¼ 0.0032). At the time of ECMO removal, native lung function had recovered: median PaO2/FIO2 was 310 (range, 64 to 514) mm Hg, significantly better than before ECMO support (p ¼ 0.0065), and median PaCO2 was 46.5 (range, 36.2 to 56.9) mm Hg (Fig 3). FIO2 on ECMO was 50% while patients were ventilated on pressure control mode with a median PEEP of 4.7 (range, 2.8 to 9.3) cm H2O. ECMO associated with the

Ann Thorac Surg 2017;-:-–-



Table 1. Patient Characteristics Patient Number Variable Age, y Sex BMI, kg/m2 ASA class Malignancy Postoperative TNM stages Preoperative chemotherapy Preoperative radiotherapy Pneumonectomy Completion pneumonectomy Lung perfusion, % Right Left Predictive postoperative FEV1 Liters Percentage DLCO, % VO2 max, mL/kg/min Lymph node dissection Estimated blood loss, mL Operating time, min Delay operation—ECMO, d ECMO mode SOFA Renal replacement therapy ECMO duration, d ECMO weaning Hospital survival Cause of death









59 Male 40 III SCC IIIA Yes No Left No

53 Male 22 III ADC IIIA No No Left Yes

82 Male 24 III SCC IIA No No Left No

61 Male 31 III SCC IIIA Yes No Left No

67 Male 19 II SCC IIB No No Right No

66 Male 23 II MM II No No Left No

57 Male 22 IV ADC IIIA No No Right No

70 Male 22 III SCC IIIA No No Right Yes

90 10


66 34


36 64


36 64

40 60

1.7 59 85 16 Radical 200 150 2 FF 17 No 7 Yes Yes NA

1.9 50 NA NA Radical 250 135 2 FJ 15 No 10 Yes Yes NA

1.2 42 NA 19 Radical 100 90 2 FJ 17 Yes 14 Yes No Arrhythmia

NA NA NA NA Radical 300 175 3 FJ 16 Yes 7 Yes Yes NA

0.9 40 30 16 Radical 100 90 411 RIJ 13 No 9 Yes No Septic shock

1.1 47 42 NA Radical 300 160 2 RIJ 12 No 5 Yes No MOF

0.8 37 NA NA Radical 200 60 0 RIJ 15 Yes 16 No No MOF

1.1 35 NA NA Radical 100 200 1,560 FJ 14 No 10 Yes Yes NA

a The pneumonectomy of patient 4 was performed in a district hospital. He was postoperatively transferred to our referred academic hospital. Preoperative data were not available.

ADC ¼ adenocarcinoma; ASA ¼ American Society of Anesthesiologists; BMI ¼ body mass index; DLCO ¼ diffusing capacity of the lung for FF ¼ femoral-femoral; carbon monoxide; ECMO ¼ extracorporeal membrane oxygenation; FEV1 ¼ forced expiratory volume in 1 second; FJ ¼ femoral-jugular; MM ¼ malignant mesothelioma; MOF ¼ multiorgan failure; NA ¼ not applicable; RIJ ¼ right internal jugular vein; SCC ¼ squamous cell carcinoma; SOFA ¼ Sequential Organ Failure Assessment; TNM ¼ tumor nodes metastasis; VO2max ¼ maximal oxygen consumption.

pressure control mode allowed a significant decrease in the patients TV (pre-ECMO TV: 412 [range, 250 to 450] mL vs ECMO TV: 277 [range, 105 to 367] mL, p ¼ 0.0156) as well as the plateau pressure levels (pre-ECMO: 34.0 [range, 32.0 to 40.0] cm H2O vs ECMO 24.5 [range, 23.3 to 27.3] cm H2O, p ¼ 0.0195; Fig 4).

Comment We demonstrated in this study that VV-ECMO applied to ppADRS decreased the expected mortality rate from 80% to 50%. To the best of our knowledge, no other cohort study has focused specifically on ECMO for parts, and no large multicenter study has evaluated the mortality rate of ppARDS.

Comparing the results of ARDS studies is hampered by the heterogeneity of the published cohorts. Before the Berlin criteria, the setoff defining ARDS was a PaO2/FIO2 of less than 200 mm Hg, which encompasses a large panel with variable prognosis [2]. The Berlin task force recognized that the adverse prognosis of ARDS is closely related to the severity of hypoxemia. It defined a high-risk group as having a PaO2/FIO2 of less than 100 mm Hg (identified as severe ARDS), for whom the mortality rate averages 45% [3]. These are actually the candidates for ECMO. This mortality rate appears to be even more elevated after pneumonectomy. Most available studies on ARDS after pulmonary resection are biased by diagnoses of ARDS with criteria before the Berlin consensus, and by confusing 30-day



Ann Thorac Surg 2017;-:-–-

Table 2. Continued

Table 2. Cohort Characteristics Median (range) or No. (%) Variables Prevenovenous ECMO characteristics Age, y Male sex Body mass index, kg/m2 Postpneumonectomy day Postpneumonectomy period <5 days Infectious ARDS Mechanical ventilation before ECMO, d SOFA score Mean arterial pressure, mm Hg Norepinephrine, mg/kg/min Prepneumonectomy and pneumonectomy characteristics NSCLC Adenocarcinoma Squamous cell carcinoma NSCLC TNM stages IIA IIB IIIA Pleural mesothelioma Induction therapy Chemotherapy Radiotherapy Pneumonectomy Right Left Completion pneumonectomy Right Left Prepneumonectomy COPD diagnosis FEV1 Liters Percentage Smoking history, pack-years Smoking quit <6 months DLCO, % VO2max, mL/kg/min ASA Cardiovascular disease history Lung transplantation history Venovenous ECMO characteristics Cannulation site Femoral-femoral Femoral-RIJ Mono-RIJa ECMO flow, L/min Sweep gas flow L/min Activated partial thromboplastin time, s Mean arterial pressure, mm Hg Norepinephrine 6 hours, mg/kg/min

(N ¼ 8) 63 8 22.6 2 6 7 0.5 15 80 0.24

(53–82) (100) (19.2–40.0) (0–1560) (75) (87.5) (0–6) (12–17) (44–94) (0.10–0.50)

7 (87.5) 2 5 1 1 5 1 (12.5) 2 0 3 5 2 (25) 1 1 4 (50) 1.78 63 30 2 83 16 3 2 1

(1.18–3.00) (57–82) (0–80) (25) (47–94) (16–19) (2–4) (25) (12.5)

2.73 4.6 53 78 0.27

1 4 3 (2.43–4.40) (2.8–8.7) (38–67) (68–89) (0.00–0.50) (Continued)

Median (range) or No. (%) Variables RBC, units transfused per day of ECMO Renal replacement therapy ECMO duration, d ECMO weaning a

(N ¼ 8) 0.97 3 9.5 7

(0.00–4.29) (37.5) (5.0–16.0) (87.5)

A single dual-lumen bicaval cannula was inserted.

ARDS ¼ acute respiratory distress syndrome; ASA ¼ American Society of Anesthesiologists; COPD ¼ chronic obstructive pulmonary disease; DLCO ¼ diffusing capacity of the lung for carbon monoxide; ECMO ¼ extracorporeal membrane oxygenation; FEV1 ¼ forced expiratory volume in 1 second; NSCLC ¼ non-small cell lung cancer; RBC ¼ red blood cells; RIJ ¼ right internal jugular vein; SOFA ¼ Sequential Organ Failure Assessment; TNM ¼ tumor nodes metastasis; VO2max ¼ maximal oxygen consumption.

mortality with in-hospital mortality. One of the first reports addressing mortality of acute lung injury (PaO2/FIO2 <300 mm Hg) and ARDS (PaO2/FIO2 <200 mm Hg) reviewed a 7-year experience at the Royal Brompton Hospital. ARDS occurred in 36 patients (3.1%), and their mortality rate was 72.2%. More specifically, the mortality rate for ppARDS was 80% (8 of 10 patients). The treatment strategy for this cohort did not include ECMO [4]. In an updated study including patients from 2000 to 2005 and which maintains the definition of ARDS as PaO2/FIO2 of less than 200 mm Hg, the authors identified 22 patients with a mortality rate of 45%. Three patients were managed without intubation, demonstrating that the PaO2/FIO2 setoff at 200 is too permissive. Three patients were managed with venoarterial ECMO, with 1 death. Death occurred in 5 of the 10 patients (50%) with ppARDS [7]. Another study pooling together acute lung injury and ARDS described a 40% mortality rate of ARDS after lung resection in general, increasing to 50% after pneumonectomy [5]. A study that pooled postoperative pneumonia, acute lung injury, and ARDS, without documenting use of ECMO, reported a mortality rate of 25% overall and 43% after pneumonectomy [6]. The hospital mortality rate observed 17 ppARDS patients (PaO2/FIO2 <200 mm Hg) was 88% [9]. Three other studies, published from 2009 to 2013, more liberally defined ARDS with a PaO2/FIO2 setoff of less than 300 mm Hg. Those three studies included 18, 6, and 9 patients with ppARDS, and mortality rates were 67%, 33%, and 55%, respectively. Use of ECMO was not described [8, 10, 11]. Iglesias and colleagues [20] evaluated the efficacy of the pumpless interventional lung-assist membrane oxygenator (iLA; Novalung, Hechingen, Germany) in 9 patients (5 pneumonectomies) with ARDS after pulmonary resection. Mean PaO2/FIO2 on admission at the ICU was 286 mm Hg. Respiratory failures were due to hypercapnia, with a mean PaCO2 at the ECMO insertion greater than 350 mm Hg. Only 1 patient (11%) died [20]. Another variable responsible for confusion in interpretation of data is introduced by mode of management.

Ann Thorac Surg 2017;-:-–-



Fig 2. Intensive care unit outcomes. (ECMO ¼ extracorporeal membrane oxygenation.)

ECMO is an interesting and potentially life-saving approach. The SOFA score we used in our cohort is a recognized marker of organ dysfunction and a validated indicator of predictive death. As such, it allows

comparison of observed outcomes and the estimated prognosis and a standardized interpretation of results [15, 18, 21]. A SOFA score exceeding 11 predicted a mortality rate of 80% [21]. Another study prospectively Fig 3. Dynamics of gas exchange: (A) ratio of the partial pressure of oxygen in the radial arterial blood to the fraction of inspired oxygen (PaO2/FIO2) and (B) partial pressure of carbon dioxide in the radial arterial blood (PaCO2). Data are illustrated as median and maximal value. (ANOVA ¼ analysis of variance; ECMO ¼ extracorporeal membrane oxygenation; ICU ¼ intensive care unit.)



Ann Thorac Surg 2017;-:-–-

Fig 4. Protective ventilation of postpneumonectomy acute respiratory distress syndrome patients supported with venovenous extracorporeal membrane oxygenation (ECMO). Comparison between pre-ECMO and ECMO (A) fraction of inspired oxygen (FIO2), (B) tidal volume (TV), (C) positive end-expiratory pressure (PEEP), and (D) measured plateau pressures. Data are illustrated as median and maximal value. #No statistical analysis was performed to compare pre-ECMO FIO2 vs ECMO FIO2. Indeed, pre-ECMO FIO2 had no variance: pre-ECMO FIO2 was 100% for every patient of the cohort.

monitored 85 patients with ARDS supported by ECMO. The median pre-ECMO SOFA score of the 48 patients (56%) who died in the hospital was 10 (interquartile range, 7 to 12). All patients with a SOFA score exceeding 12 died in the hospital [22]. In our cohort, the median SOFA score calculated immediately before ECMO was 15 (range, 12 to 17), and the SOFA score of our 4 survivors ranged from 14 to 17. The available evidence shows the SOFA score of our observed 50% hospital survival rate was far better than expected, which encourages us to continue our practice. We hypothesized that outcome of ppARDS might be favorably modulated by associating VV-ECMO and protective ventilation [23]. VV-ECMO is thought to put the damaged lung to rest and to switch to a pressure control mode of ventilation, decreasing both TV and plateau pressure. On the pressure control mode, while hematosis was supported by the VV-ECMO, lungs were recovering. Our study is limited by a small sample size of 8 patients, with ARDS developing in 2 patients more than 1 year after pneumonectomy, and might be continued by a multicenter study. Moreover, identifying pre-ECMO or intra-ECMO support prognosis factors would constitute a major tool for the clinicians taking care of those critically ill patients. Use of the SOFA score could help stratify patients by spontaneous prognosis and further evaluate the benefit of ECMO and any other innovative approach. In conclusion, the present study reports encouraging results of VV-ECMO in the management of severe and

refractory ppARDS. Despite the small number of patients included, our data suggest that VV-ECMO safely guarantees adequate gas exchange, puts the damaged lung to rest, allows recovery of the lung, and ultimately, decreases related death. The authors wish to thank Ilana Adleson for the expert editorial review of the manuscript.

References 1. Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute respiratory distress in adults. Lancet 1967;2:319–23. 2. Bernard GR, Artigas A, Brigham KL, et al. The AmericanEuropean Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818–24. 3. ARDS definition task force, Ranieri VM, Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin definition. JAMA 2012;307:2526–33. 4. Kutlu CA, Williams EA, Evans TW, Pastorino U, Goldstraw P. Acute lung injury and acute respiratory distress syndrome after pulmonary resection. Ann Thorac Surg 2000;69:376–80. 5. Dulu A, Pastores SM, Park B, Riedel E, Rusch V, Halpern NA. Prevalence and mortality of acute lung injury and ARDS after lung resection. Chest 2006;130:73–8. 6. Alam N, Park BJ, Wilton A, et al. Incidence and risk factors for lung injury after lung cancer resection. Ann Thorac Surg 2007;84:1085–91. 7. Tang SS, Redmond K, Griffiths M, Ladas G, Goldstraw P, Dusmet M. The mortality from acute respiratory distress syndrome after pulmonary resection is reducing: a 10-year

Ann Thorac Surg 2017;-:-–-




11. 12. 13.



single institutional experience. Eur J Cardiothorac Surg 2008;34:898–902. Jeon K, Yoon JW, Suh GY, et al. Risk factors for postpneumonectomy acute lung injury/acute respiratory distress syndrome in primary lung cancer patients. Anesth Intensive Care 2009;37:14–9. Kim JB, Lee SW, Park SI, Kim YH, Kim DK. Risk factor analysis for postoperative acute respiratory distress syndrome and early mortality after pneumonectomy: the predictive value of preoperative lung perfusion distribution. J Thorac Cardiovasc Surg 2010;140:26–31. Sen S, Sen S, Sent€ urk E, Kuman NK. Postresectional lung injury in thoracic surgery pre and intraoperative risk factors: a retrospective clinical study of a hundred forty-three cases. J Cardiothorac Surg 2010;5:62. Brodie D, Bacchetta M. Extracorporeal membrane oxygenation for ARDS in adults. N Engl J Med 2011;365: 1905–14. Del Sorbo L, Cypel M, Fan E. Extracorporeal life support for adults with severe acute respiratory failure. Lancet Respir Med 2014;2:154–64. D€ unser M, Hasibeder W, Rieger M, Mayr AJ. Successful therapy of severe pneumonia-associated ARDS after pneumonectomy with ECMO and steroids. Ann Thorac Surg 2004;78:335–7. Fica M, Suarez F, Aparicio R, Suarez C. Single site venovenous extracorporeal membrane oxygenation as an alternative to invasive ventilation in post-pneumonectomy fistula with acute respiratory failure. Eur J Cardiothorac Surg 2012;41: 950–2. Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsisrelated Organ Failure Assessment) score to describe





19. 20.

21. 22.



organ dysfunction/failure. Intensive Care Med 1996;22: 707–10. British Thoracic Society; Society of Cardiothoracic Surgeons of Great Britain and Ireland Working Party. BTS guidelines on the selection of patients with lung cancer for surgery. Thorax 2001;56:89–108. Roch A, Papazian L, Bregeon F, et al. High or low doses of almitrine bismesylate in ARDS patients responding to inhaled NO and receiving norepinephrine. Intensive Care Med 2001;27:1737–43. Vincent JL, de Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/ failure in intensive care units: results of a multicenter, prospective study. Crit Care Med 1998;26:1793–800. Reeb J, Olland A, Renaud S, et al. Vascular access for extracorporeal life support: tips and tricks. J Thorac Dis 2016;8:S353–63. Iglesias M, Martinez E, Badia JR, Macchiarini P. Extrapulmonary ventilation for unresponsive severe acute respiratory distress syndrome after pulmonary resection. Ann Thorac Surg 2008;85:237–44. Ferreira FL, Bota DP, Bross A, Melot C, Vincent JL. Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA 2001;286:1754–8. Roch A, Hraiech S, Masson E, et al. Outcome of acute respiratory distress syndrome patients treated with extracorporeal membrane oxygenation and brought to a referral center. Intensive Care Med 2014;40:74–83. The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301–8.