Improved Survival but Marginal Allograft Function in Patients Treated With Extracorporeal Membrane Oxygenation After Lung Transplantation

Transcription

1 GENERAL THORACIC ORIGINAL ARTICLES: GENERAL THORACIC CARDIOTHORACIC ANESTHESIOLOGY: The Annals of Thoracic Surgery CME Program is located online at To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal. Improved Survival but Marginal Allograft Function in Patients Treated With Extracorporeal Membrane Oxygenation After Lung Transplantation Matthew G. Hartwig, MD, Richard Walczak, BS, Shu S. Lin, MD, PhD, and R. Duane Davis, MD Division of Thoracic Surgery, Department of Surgery, Duke University Medical Center, Durham, North Carolina Background. Previous reports demonstrate that 1-year survival is severely compromised in patients with severe primary graft dysfunction (PGD) after lung transplantation. We examined if advances in extracorporeal membrane oxygenation (ECMO) support, including polymethylpentene oxygenators and reliance on venovenous (VV) ECMO have improved outcomes in patients with severe PGD after lung transplantation. Methods. The analysis included data prospectively collected on all single-lung or double-lung transplants between November 2001 and December Heart-lung transplants were excluded. Comparisons were made between recipients who did and did not require ECMO for PGD after transplant. Results. Since November 2001, when VV ECMO became the routine treatment for severe PGD after transplant at our center, 28 of 498 patients (6%) have required VV ECMO support. Successful weaning occurred in 27 of 28 (96%). Support was withdrawn for 1 patient with irreversible neurologic injury. Survival was substantially better than in previous reports: 30 days, 82%; 1 year, 64%; and 5 years, 49%. Freedom from bronchiolitis obliterans syndrome was 88% in the ECMO survivors at 3 years, but maximum allograft function was considerably worse than in transplant recipients not requiring ECMO (peak forced expiratory volume in 1 second: 58% in ECMO vs 83% in non-ecmo, p 0.001). Conclusions. Advances in ECMO technology, particularly VV ECMO, have greatly improved the ability to support patients with severe PGD after lung transplantation. VV ECMO is an important tool in the armamentarium of any lung transplant program to optimize patient outcomes; however, strategies to improve lung allograft function in patients experiencing severe PGD are still needed. (Ann Thorac Surg 2012;93:366 71) 2012 by The Society of Thoracic Surgeons Primary graft dysfunction (PGD) remains a common clinical problem after lung transplantation. Depending on how it is defined, PGD is thought to occur between 15% and 60% of the time [1, 2]. Despite being the subject of many published, peer-reviewed research efforts that have received considerable research monies, PGD persists as the leading cause of early death after lung transplantation. According to the International Society for Heart and Lung Transplantation (ISHLT) registry data, PGD accounts for nearly one-third of all deaths in the first 90 days after transplant [3]. Studies looking at potential therapies for PGD have yet to discern effective modalities for the prevention or treatment, but an ISHLT working group published a grading scale for determining Accepted for publication May 3, Presented at the Poster Session of the Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31 Feb 2, Address correspondence to Dr Davis, Department of Surgery, Duke University Medical Center, Box 3864, Durham, NC 27710; davis053@ mc.duke.edu. the severity of PGD to assist in clinical management, as well as the research efforts, of this vexing problem [4 6]. Extracorporeal membrane oxygenation (ECMO) is one method of supporting patients with PGD after lung transplantation. Although often assumed to be a treatment option of last resort, several published reports describe the experience with ECMO for PGD and the early posttransplant outcomes. These studies indicated acceptable perioperative outcomes when ECMO was instituted in a timely fashion [7, 8]. Our own center s experience suggests an advantage to using venovenous (VV) ECMO in patients with PGD compared with a venoarterial cannulation strategy. [9] In that study, our early experience indicated a 30-day patient survival of 88% for those receiving VV- ECMO, comparing favorably to the 94% for lung recipients who did not require ECMO for PGD. Although the efficacy of ECMO in extending early survival for recipients with PGD is documented, very little has been determined about the long-term consequences of surviving the most severe cases of PGD. We hypothesized that with the utilization of modern VV by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg HARTWIG ET AL 2012;93: IMPROVED SURVIVAL AFTER ECMO USE ECMO technology, recipients with the most severe forms of PGD would continue to demonstrate remarkable early recoverability of graft function, but that later outcomes, such as 5-year survival, the development of bronchiolitis obliterans syndrome (BOS), and overall graft function, would not be the same as for recipients who never sustained such severe PGD. Material and Methods Approval for the study was obtained from Duke University Hospital (DUH) Institutional Review Board for Human Studies. The requirement for informed consent was waived. Patient Population and Transplantation We retrospectively evaluated prospectively collected data on all patients who underwent single-lung or bilateral-lung transplantation at DUH between November 2001 and December The starting date was selected secondary to our center s routine use of VV-ECMO support of severe PGD being initiated at that time. Heart-lung transplants were excluded from the analysis. Redo transplants were included in this analysis and were considered a separate transplant procedure; that is, pulmonary allograft outcomes were analyzed separately for each patient. The transplant procedure was performed in a standard fashion, as previously described, with an institutional bias toward bilateral sequential lung transplantation through a bilateral transverse sternal thoracotomy or clamshell incision [9]. We use an extracellular preservation solution, such as Celsior (Genzyme Corp Cambridge, MA) or Perfadex (Medisan, Uppsala, Sweden), with antegrade and retrograde flushing during procurement, and all donor lungs were procured using the same techniques. We also attempt to control reperfusion pressures by incrementally increasing flow to the newly implanted pulmonary allograft over the first 10 to 15 minutes. 367 ECMO Protocol Venovenous cannulation typically occurred through the right femoral vein with a venous catheter and the left internal jugular vein with a pediatric arterial cannula. Cannulas were placed percutaneously using a modified Seldinger technique over a guidewire after serial dilatations. Placement is confirmed with echocardiography or radiographic imaging, or both. In more recent cases, the Avalon Elite Dual Lumen Catheter (Avalon Laboratories, LLC, Rancho Dominguez, CA) has been used. Our original circuit consisted of hyaluron-based, heparin-coated inch tubing, with a Jostra Quadrox hollow-fiber membrane oxygenator and a Jostra Rotaflow pump (MAQUET Cardiopulmonary, AG, Rastatt, Germany). Since 2005, we have been using the Quadrox D polymethylpentene oxygenator. The optimal placement of circuit inflow and outflow ports is determined by the level of recirculation noted in the system. Weaning from VV ECMO involved discontinuing membrane gas flow and increasing ventilator indicators as needed. No increase in anticoagulation is required for VV ECMO weaning. Severe PGD manifesting as radiographic opacification and copious pulmonary secretions, coupled with worsening oxygenation or ventilation, or both, was the impetus for instituting extracorporeal support. All patients demonstrating PGD during this period received inhaled nitric oxide in an attempt to ameliorate lung injury. Inhaled prostacyclin was not used by our center for the prevention or treatment of reperfusion injury. Initiating ECMO support is considered when supporting ventilator requirements reach peak inspiratory pressures (PIP) of 35 cm H 2 O and fraction of inspired oxygen (Fio 2 ) content surpasses 0.60, or whenever copious pulmonary edema develops that does not clear easily with bronchoscopic suctioning. However, no specific blood gas or ventilator settings serve as criteria for the initiation of ECMO at our institution. In all instances, the suitability to initiate ECMO is determined by the attending transplant surgeon. After commencement of ECMO, ventilator tidal volumes and rates are minimized to rest levels, which typically include PIP of 20 to 25 cm H 2 0, positive endexpiratory pressure of 10 cm H 2 O or less, and a Fio 2 of 0.30 or less. An activated clotting time goal of 180 to 200 seconds was predetermined but was adjusted as dictated by patient scenarios. Immunosuppression and Antimicrobial Agents Immunosuppression regimens and microbial prophylaxis were determined by institutional policies. All patients received standard triple-drug therapy postoperatively that included corticosteroids, tacrolimus, and azathioprine or mycophenolate mofetil. Mycophenolate mofetil served as the antimetabolite for a very few patients as part of a prior randomized clinical trial. Otherwise, all patients received azathioprine. Most patients received the monoclonal interleukin-2 receptor immunoglobulin, daclizumab, as part of the induction treatment protocol. Anti-CD52 monoclonal antibody (Campath, Genzyme) was used in some patients with baseline renal insufficiency or retransplants. Pulmonary Function Testing, Transbronchial Biopsies, and Acute Rejections All spirometry data were collected at DUH. We used the peak predicted forced expiratory volume in 1 second (FEV 1 ) as a representation of maximal function attained by each allograft. Standard ISHLT guidelines for determining BOS after transplant were used for each recipient based on the two highest absolute peak FEV 1 measurements. Surveillance transbronchial biopsies (TBBx) were routinely performed on all patients at 1, 3, 6, and 12 months after transplantation. Recipients routinely underwent annual surveillance biopsies after the first 12 months. Patients also underwent TBBx if clinically warranted by evidence of allograft dysfunction. Recipients with episodes of rejection grade III or higher received a follow-up TBBx 6 weeks after the initial diagnosis. Min- GENERAL THORACIC

3 GENERAL THORACIC 368 HARTWIG ET AL Ann Thorac Surg IMPROVED SURVIVAL AFTER ECMO USE 2012;93: Table 1. Demographics for Patients With and Without Extracorporeal Membrane Oxygenation (ECMO) Characteristics a No-ECMO ECMO p Value Transplants 470 (94) 28 (6) Follow-up, days 1, Gender 0.43 Female 276 (59) 14 (50) Male 194 (41) 14 (50) Ethnicity 0.2 White 423 (90) 24 (86) Nonwhite 47 (10) 4 (14) Bilateral transplant 448 (95) 25 (89) Age, median Underlying disease 0.15 COPD/A1A deficiency 154 (33) 5 (18) CF/bronchiectasis 93 (20) 5 (18) IPF 165 (35) 11 (39) Other 58 (12) 7 (25) a Categoric data are presented as number (%), continuous data are presented as the mean standard deviation or median and interquartile range. The Standard Deviation for mean days is 846 for the No-ECMO and 799 for ECMO. The Interquartile Range for age for No-ECMO is years, and for ECMO is years. A1A -1 antitrypsin deficiency; CF cystic fibrosis; COPD chronic obstructive pulmonary disorder; IPF interstitial pulmonary fibrosis. imal acute rejection (grade A1) was not treated unless symptomatic. Patients with grade II or greater acute rejection episodes, as documented by the biopsy specimen, were treated with methylprednisolone (500 mg/day for 3 days), followed by an oral prednisone taper over the next 2 to 3 weeks. Statistical Analyses Standard descriptive statistics were used for patient demographic information. Values were calculated as mean standard deviation for normally distributed data or median with interquartile range (IQR) for nonnormally distributed results. Comparisons between groups were made using two-sample t tests (parametric) or Wilcoxon rank sum (nonparametric) for continuous data and the 2 or Fisher exact test for categoric variables. Kaplan-Meier analysis was used to determine freedom from BOS and survival. Comparisons between groups were made using the log-rank statistic. All analyses were performed using SAS 9.1 software (SAS Institute, Cary, NC). transplants. Demographics for each group are summarized in Table 1. The mean duration of ECMO support was 3.6 days (IQR, 0.2 to 7.6 days). Cardiopulmonary bypass (CPB) was required during 13 of the 28 transplant procedures (46.4%) that resulted in ECMO use and in 81 of 409 transplant operations (19.8%) not requiring ECMO. CPB details for 61 patients, none of whom required ECMO, were lacking in the database. Survival For the purpose of this study, we evaluated allograft survival and patient survival. We did this solely to demonstrate that improved patient survival did not depend on retransplanting affected organs. One ECMO patient required retransplant for PGD and 1 patient was unable to be weaned successfully from ECMO. Therefore, overall patient survival was not too dissimilar from graft survival. The 30-day graft survival was 82% in the ECMO group compared with 97% in the no-ecmo group. Graft survival was 49% for both 3 years and 5 years in the ECMO group, but was 74% and 61% for the no-ecmo recipients, respectively. Figure 1 depicts Kaplan-Meier survival proportion results. The no-ecmo survival curve appears nearly linear in fashion over the time of this study. However, the ECMO group demonstrates a significant early decrement in graft survival that levels off and parallels that of the no-ecmo group after approximately 90 to 120 days. Anecdotally in our experience, blood stream infections appeared to be the significant source of morbidity and death in patients requiring extracorporeal support, and therefore, we evaluated the development of blood stream infections between the ECMO and no-ecmo groups to 90 days. These data are reported in Table 2. There were 15 isolates cultured from blood in 10 patients (35.7%) of the ECMO group, and 79 isolates were identified in 35 no-ecmo patients (12.1%; odds ratio, 6.5; p ). Fungal blood stream infections were present in 2 of 15 Results Patients From November 2001 to December 2009, 498 lung transplants were performed at DUH and included for analysis in this study. Of those, 28 patients required VV ECMO treatment for severe PGD (ECMO group), whereas 470 did not require ECMO (no-ecmo group). During this period, the incidence of ECMO for PGD was 5.6/100 Fig 1. Kaplan-Meier curves depict 5-year survival proportions for patients with extracorporeal membrane oxygenation (ECMO, dashed line) and without ECMO (No ECMO, solid line). After 120 days, the curves appear very similar in the rate of pulmonary allograft attrition.

4 Ann Thorac Surg HARTWIG ET AL 2012;93: IMPROVED SURVIVAL AFTER ECMO USE Table 2. Details of Blood Stream Infections for Groups With and Without Extracorporeal Membrane Oxygenation (ECMO) 90-Day Blood Stream Infections No-ECMO (n 470) ECMO (n 28) p Value Patients 37 (12) 10 (36) Isolates Bacterial Fungal 27 2 Mycobacterial 1 0 Total GENERAL THORACIC isolates (13.3%) in the ECMO group vs in 27 of 79 isolates (34%) in the no-ecmo recipients. Mycobacterial blood stream infections occurred rarely, developing in only 1 recipient in either group. Pulmonary Function Functional recovery after severe PGD was evaluated using the absolute maximum and peak percent-predicted FEV 1 attained by each allograft. Only those recipients with at least 6 months of survival or follow-up were studied for peak FEV 1 comparisons (Fig 2). The mean time to the peak FEV 1 was 410 days in the ECMO group compared with 558 days in the no-ecmo group. Mean peak FEV 1 was 2.0 liters in the ECMO group, for a mean peak percent-predicted FEV 1 of 58%, and was 2.7 liters in the no-ecmo group, for a mean peak percent-predicted FEV 1 of 83% (p 0.006). The development of chronic allograft dysfunction, or BOS, was evaluated, and Kaplan-Meier curves depicting the freedom from BOS are noted in Figure 3. The rate at which BOS develops appears similar between the two groups, with 3-year and 5-year freedom from BOS of 88% in the ECMO group, and 82% and 62%, respectively, in the no-ecmo group (p 0.43). The median time to the development of any grade of BOS was 3,030 days in the ECMO group and 2,462 days for no-ecmo recipients. Acute Rejection An acute rejection sum was determined for each patient that included a summation of each TBBx specimen demonstrating rejection, graded according to the severity of the rejection. The acute rejection sum was divided by the Fig 3. Kaplan-Meier curves illustrate freedom from bronchiolitis obliterans syndrome (BOS) at 5 years for patients with extracorporeal membrane oxygenation (ECMO) and without ECMO (No ECMO). No appreciable difference in the development of any grade of BOS could be detected. number of biopsies performed to normalize for the frequency of biopsies, and this determined the acute rejection ratio. The mean acute rejection sum per month of follow-up was in the ECMO group and in the no-ecmo patients. The acute rejection ratio was 0.33 in the ECMO group compared with 0.39 for the no- ECMO group (p 0.51; Fig 4). When the analysis was limited to the 1-year rejection data in only those recipients surviving to 1 year, the mean acute rejection sum per recipient was 1.9 for the ECMO patients and 2.2 for the no-ecmo group (p 0.62), and the acute rejection ratio was 0.32 and 0.39, respectively. Comment The peritransplant utility of ECMO in supporting lung transplant recipients with severe PGD has been demonstrated previously. Although our early outcomes, including a 30-day graft survival of more than 80%, are promising, the eventual ability of the transplanted lung to recover and the long-term durability of these grafts have not been thoroughly studied. The purposes of this article include describing expected perioperative outcomes Fig 2. Comparing the (A) peak forced expiratory volume in 1 second (FEV 1 ) and (B) peak percent-predicted FEV 1 values for patients with extracorporeal membrane oxygenation (ECMO) and without ECMO (No ECMO) demonstrates a significant decrement in allograft spirometric values. The error bars show the standard error of the mean.

5 GENERAL THORACIC 370 HARTWIG ET AL Ann Thorac Surg IMPROVED SURVIVAL AFTER ECMO USE 2012;93: Fig 4. Analysis of acute rejection data included calculating an acute rejection ratio (ARR). The ARR did not differ between patients with extracorporeal membrane oxygenation (ECMO) and without ECMO (No ECMO). The error bars show the standard error of the mean. when using modern, but widely available, extracorporeal support technologies. In addition, we wanted to better understand the natural history of survival and function in pulmonary allografts that suffer from the most severe forms of PGD. Unfortunately, our series is limited by its retrospective, single-center nature and small sample size. Also, the factors involved in the development of severe PGD are not explored here because they have been previously described, including the interplay between cardiopulmonary bypass and ECMO. However, our data do indicate that regardless of the etiology, even the most affected patients suffering from PGD can typically be supported safely with VV ECMO until the transplanted lungs have time to recover. Not unexpectedly, the overall allograft survival for recipients requiring ECMO support does not equal the survival of allografts in recipients who never required ECMO after transplant. However, with the advent of polymethylpentene oxygenators, biocoated circuitry, and aggressive antimicrobial prophylaxis, mortality can be reduced and one should expect improved survival when compared with the previously reported 30-day mortality rate of nearly 50% [10]. In our early ECMO experience, most deaths followed significant neurologic insults or overwhelming sepsis, particularly from fungemia. The use of VV ECMO compared with venoarterial ECMO appears to decrease the incidence and the severity of neurologic injury. [9] Although we continue to see a significantly higher incidence of blood stream infections in the ECMO group, fewer of those have been fungal isolates. The decrease in fungal isolates in our recent ECMO experience, especially compared with the relatively high incidence of fungemia in recipients not requiring ECMO, coincides with the use of aggressive antifungal prophylaxis when ECMO is initiated. Our current regimen includes voriconazole or caspofungin, or both, with particularly good activity against Candida species. Continuing to develop ways to prevent blood stream infections and neurologic injury while patients require ECMO remains a critical component of improving outcomes for these patients. Consistent with other published series, it appears that once the early hazard period resolves, allograft survival should be approximately equal [10]. In our series, this transition occurs 90 to 120 days after transplant, after which no difference in the Kaplan-Meier survival estimates between the groups can be discerned. Because the development of BOS remains the best predictor of late recipient death, it should not be surprising that the development of BOS over 5 years did not differ between the ECMO and no-ecmo groups. However, this is contrary to previous studies that identified ischemiareperfusion injury as a risk factor for developing BOS [2]. Additional studies using the more recently developed PGD grading scale may be able to help discern the role of various levels of PGD on BOS development. Another possibility may be that the rate of allograft decline (ie, BOS development) is not affected by PGD, but rather only the starting point (ie, peak FEV 1 ) is lowered. We could not detect a noticeable increase in BOS development for patients requiring ECMO support; however, there does appear to be permanent damage to the allograft that attenuates maximum graft function. Initially, we believed 1-year spirometry might be an objective measurement of graft function, as other groups have reported. However, upon analysis of our series, it became clear that recipients, regardless of PGD, were not reaching maximum allograft function until well after 1-year after transplant. Therefore, we chose the peak predicted FEV 1 as the most objective measurement of maximum allograft function attained. There was an absolute decrease of 26% in peak FEV 1 and 30% in the peak percent-predicted FEV 1 in recipients requiring ECMO support. This is corroborated by other studies that have demonstrated a negative relationship between the severity of PGD and the overall graft function attained after transplant [11]. Analogous to the survival difference between singlelung and double-lung transplants, it may be that with enough patients monitored over a long enough interval that this decrement in peak allograft function may lead to a detectable graft survival difference. Unfortunately, the effect on survival of this difference in pulmonary function remains to be determined, but alternative methods of evaluation, such as quality of life indicators, may better clarify differences between the groups. It does not appear that the decrease in lung allograft function after ECMO support can be attributed to an increase in acute rejections. The evaluation of acute rejections after lung transplantation remains a difficult task. The number of biopsy specimens and time from transplant are two important factors in determining the severity and frequency of acute rejections after lung transplantation. To account for these biases, we calculated an acute rejection sum that was used to determine the acute rejection ratio by dividing by the number of biopsies. Likewise, by including only those patients who survived 12 months with their respective 12-month rejection data, we controlled for the effect of time from

6 Ann Thorac Surg HARTWIG ET AL 2012;93: IMPROVED SURVIVAL AFTER ECMO USE transplant on acute rejection. The use of these methods revealed no difference in biopsy specimen proven acute rejection secondary to severe PGD. However, this aspect of the analysis may be misleading. Biopsy samples taken from lungs with severe PGD notoriously have considerable artifact, and less severe episodes of acute rejection may be missed in this setting. A systemic immune assay would likely provide better insight into this question. In conclusion, ECMO provides a reasonable support option for lung allografts suffering from severe PGD. Modern ECMO technologies and programs should provide considerably better survival than those documented in previously published series, and it should no longer be viewed as a salvage intervention. Our data also demonstrate that most affected allografts will recover from the initial insult and that retransplant should only be necessary in rare circumstances for PGD. The use of VV ECMO and aggressive antimicrobial prophylaxis may minimize the two most common causes of early death in these recipients. Despite improved perioperative survival, pulmonary allografts have attenuated function, as reflected in diminished spirometry values. This might, over time, lead to a detectable difference in survival or BOS development, but more likely will be reflected in measurements of quality of life or functional assessments. In the future, more detailed studies should evaluate the quality of life and cost-effectiveness of this time-intensive and expensive therapy. References 1. Thabut G, Vinatier I, Stern JB, et al. Primary graft failure following lung transplantation: predictive factors of mortality. Chest 2002;121: Christie JD, Kotloff RM, Pochettino A, et al. Clinical risk factors for primary graft failure following lung transplantation. Chest 2003;124: Christie JD, Edwards LB, Kucheryavaya AY, et al. The Registry of the International Society for Heart and Lung Transplantation: twenty-seventh official lung and heart-lung transplant report J Heart Lung Transplant 2010; 29: Meade MO, Granton JT, Matte-Martyn A, et al. A randomized trial of inhaled nitric oxide to prevent ischemiareperfusion injury after lung transplantation. Am J Respir Crit Care Med 2003;167: Keshavjee S, Davis RD, Zamora MR, et al. A randomized, placebo-controlled trial of complement inhibition in ischemia-reperfusion injury after lung transplantation in human beings. J Thorac Cardiovasc Surg 2005;129: Christie JD, Carby M, Bag R, et al. Report of the ISHLT Working Group on Primary Lung Graft Dysfunction Part II: definition. A consensus statement of the International Society for Heart and Lung Transplantation. J Heart Lung Transplant 2005;24: Meyers BF, Sundt TM, Henry S, et al. Selective use of extracorporeal membrane oxygenation is warranted after lung transplantation. J Thorac Cardiovasc Surg 2000;120: Glassman LR, Keenan RJ, Fabrizio MC, et al. Extracorporeal membrane oxygenation as an adjunct treatment for primary graft failure in adult lung transplant recipients. J Thorac Cardiovasc Surg 1995;110: Hartwig MG, Appel JZ, Cantu E, et al. Improved results treating lung allograft failure with venovenous extracorporeal membrane oxygenation. Ann Thorac Surg 2005;80: Bermudez CA, Adusumilli PS, McCurry KR, et al. Extracorporeal membrame oxygenation for primary graft dysfunction after lung transplantation: long-term survival. Ann Thorac Surg 2009;87: King RC, Binns OA, Rodriguez F, et al. Reperfusion injury significantly impacts clinical outcome after pulmonary transplantation. Ann Thorac Surg 2000;69: GENERAL THORACIC INVITED COMMENTARY The authors [1] report a retrospective single-center analysis about their extracorporeal membrane oxygenation (ECMO) experience to support patients with severe primary graft dysfunction. Twenty-eight of 498 patients (6%) required ECMO support. Congratulations to the authors for their outcomes that were superior to other published series. This study addresses an important problem for the transplant team: what to do when the graft fails? What is the best time to start ECMO? Is venovenous ECMO better than venoarterial ECMO? The question about long-term consequences after severe primary graft dysfunction and ECMO support is answered in the present work and provides some important advice for practice. Is it possibly better to establish ECMO when inspired oxygen content (Fio 2 ) surpasses 0.60? Is this a significant limit? In my experience physicians in most centers are more patient and start ECMO later (Fio ). The low threshold used by the Duke University group can be 1 reason for the good results. However other centers perhaps wait too long to initiate ECMO. The best time to begin ECMO support is not exactly defined. Nevertheless the indication for ECMO support must be decided individually in every patient. I, as well as the authors, seek more studies regarding evaluation of quality of life and not only functional assessments. Analysis of the cost-effectiveness is necessary, as mentioned by the authors, is a matter of taste. Torsten Bossert, MD Department of Cardiac and Vascular Surgery Heartcenter Coswig Lerchenfeld 1 Coswig D-06869, Germany tbossert11@aol.com Reference 1. Hartwig MG, Walczak R, Lin SS, Davis RD. Improved survival but marginal allograft function in patients treated with extracorporeal membrane oxygenation after lung transplantation. Ann Thorac Surg 2012;93: by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur