| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;80:1872-1880
© 2005 The Society of Thoracic Surgeons
a Department of Surgery, Duke University Medical Center, Durham, North Carolina
b Department of Medicine, Duke University Medical Center, Durham, North Carolina
Accepted for publication April 26, 2005.
* Address correspondence to Dr Davis, Department of Surgery, Duke University Medical Center Box 2605, Durham, NC27710 (Email: davis053{at}mc.duke.edu).
Presented at the Forty-first Annual Meeting of The Society of Thoracic Surgeons, Tampa, FL, Jan 24–26, 2005.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
METHODS: We conducted an analysis of 522 patients who underwent lung transplantation from April 1992 to July 2004. Twenty-three (4.5%) patients required membrane oxygenation secondary to primary graft failure unresponsive to conventional treatment. Of these recipients, 15 (65%) were treated with venoarterial, while 8 (35%) underwent venovenous membrane oxygenation.
RESULTS: Median days to initiation and duration of membrane oxygenation did not differ between groups. Eight of 15 patients (53%) from the venoarterial group were successfully weaned from life support, with one surviving greater than 45 days. This lone long-term survivor required retransplantation 4 days after initial transplant. In contrast, all venovenous patients were weaned from support, with 7 of 8 surviving greater than 30 days. The 30-day survival for venovenous recipients (88%) approximates that of all lung recipients at our center (94%, p = 0.42). Noted complications for ECMO patients included renal failure (n = 16), neurologic catastrophes (n = 8), sepsis (n = 5), and hemorrhage (n = 10). The venoarterial recipients suffered 30 of 39 total complications. Most of the complications for venovenous recipients involved renal failure, but by hospital discharge these patients demonstrated a mean creatinine of 0.9 mg/dL.
CONCLUSIONS: For lung recipients with primary graft failure, venovenous membrane oxygenation provides better outcomes, with fewer complications, than venoarterial membrane oxygenation.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Patients suffering from IRI demonstrate worsening pulmonary compliance and hypoxemia. Radiographically, diffuse pulmonary infiltrates are uniformly seen on chest radiographs by postoperative day 3. Bronchoscopy may reveal copious amounts of airway secretions. The management of IRI consists primarily of preventive and supportive measures. Supportive actions include prolonged mechanical ventilation, often requiring sedation and paralysis, as well as aggressive diuresis. The utility of extracorporeal membrane oxygenation (ECMO) has also been described in this setting. Most commonly, venoarterial (VA) ECMO is utilized.
The early experience in our center with VA ECMO proved disappointing; therefore, we began preferentially utilizing venovenous (VV) ECMO for primary graft failure (PGF) after lung transplantation. This report details our experience with the use of ECMO posttransplant, comparing complications and survival rates between recipients undergoing VA and VV ECMO. Where applicable, medium and long-term outcomes including acute rejection, bronchiolitis obliterans syndrome (BOS) development, and mortality are presented.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Extracorporeal Membrane Oxygenation
Venoarterial cannulation proceeded by an extrathoracic or intrathoracic approach, depending on the specific situation. Bio-Medicus venous (23 French) and arterial (19 French) catheters were used for cannulation (Bio-Medicus, Eden Prairie, MN). For femoral vein and artery cannulation, distal catheter position involved the inferior vena cava-right atrial border for the venous cannula and the common femoral or external iliac artery for arterial return. In general, flow rates were maintained at 2.5–3.5 liters/minute, although they were titrated according to the level of support required. Circuit hardware consisted of the Medtronic Affinity oxygenator (Medtronic, Minneapolis, MN). Weaning from VA ECMO involved incrementally reducing flows by 0.5 L/minute until flow less than 1 L/minute was achieved while patient stability is monitored. Routinely, when flows were below 2 L/minute, heparin infusion was increased to maintain activated clotting time (ACT) greater than 200.
Venovenous cannulation typically occurred through the right femoral vein with a venous catheter (Bio-Medicus) and the left internal jugular vein with a pediatric arterial cannula (Medtronic). Cannulas were placed percutaneously using a modified Seldinger technique over a guide-wire after serial dilatations. Our current circuit consists of a hyaluron-based, heparin-coated 3/8 inch tubing (GISH Biomedical, Inc, Rancho Santa Margarita, CA) with a Jostra Quadrox hollow fiber membrane oxygenator and Jostra Rotaflow pump (MAQUET Cardiopulmonary, AG, Germany). The optimal placement of circuit in-flow and out-flow ports are determined by the level of recirculation noted in the system. Weaning from VV ECMO involved discontinuing membrane gas flow and increasing ventilatory parameters as needed. No increase in anticoagulation is required for VV ECMO weaning.
Our preservation and reperfusion techniques have evolved over time, so that we currently use an extracellular preservation solution (ie, Celsior [Imtix-SangStat, Lyon, France] or Perfadex [Medisan, Uppsala, Sweden]) with antegrade and retrograde flushing during procurement. We initially began using Celsior in the fall of 2000 and by the end of 2001 all donor lungs were procured using an extracellular preservation solution. Prior to 2000, University of Wisconsin solution was used for allograft flushing. Additionally, we now attempt to control reperfusion pressures by incrementally increasing flow over the first 10 to 15 minutes to the newly implanted pulmonary allograft. These maneuvers have not altered the incidence of ECMO use at our institution over time, which has remained constant at about 4.6 per 100 transplants, but they may partially abrogate IRI and allow for improved outcomes.
Severe reperfusion injury manifest by radiographic opacification and copious pulmonary secretions, coupled with worsening oxygenation and/or ventilation, provided the impetus for instituting extracorporeal support. All patients demonstrating severe reperfusion injury during this time period received inhaled nitric oxide in an attempt to ameliorate lung function. Inhaled prostacyclin was not utilized by our center for the prevention or treatment of reperfusion injury. Initiating ECMO support is considered when supporting ventilatory requirements reach peak inspiratory pressures (PIP) of 35 cm H2O and inspired oxygen content (FiO2) surpasses 0.60. However, no specific blood gas or ventilatory settings serve as criteria for the initiation of ECMO at our institution. In all instances, the suitability to initiate ECMO was 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 H20, positive end-expiratory pressure less than or equal to 10 cm H20, and FiO 2 less than or equal to 0.30. An ACT goal of 180 to 200 seconds was predetermined, but was adjusted as dictated by patient scenarios.
Statistical Analyses
Descriptive statistics were used for patient demographic information. Values are expressed as mean ± standard deviation for normally distributed data or median with interquartile range for results not normally distributed. Comparisons between groups were made using two sample t tests (parametric) or Wilcoxon rank-sum (nonparametric) for continuous data and 
2 or Fisher's exact test for categorical variables. Body surface area (BSA) calculations for the donor and recipient were made using the Mosteller BSA formula ([height (cm)·weight (kg)/3,600]1/2). A donor BSA-to-recipient BSA ratio was then determined for each transplant procedure. 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 software version 8.0 (SAS, Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
|
|
|
|
Survivors of ECMO were also analyzed for pulmonary function, severity of acute rejection, and BOS. The median discharge PaO 2/FiO 2 ratio for ECMO recipients surviving to hospital discharge was 338.1 (319.1 to 450.0). Median peak forced expiratory volume in one second (FEV1) value for patients surviving at least 3 months after ECMO was 1.71 liters (1.28 to 2.39). This is less than the median peak FEV1 for bilateral transplants from the NO ECMO group of 2.66 liters (2.24 to 3.24, p = 0.0061). A 6-month acute rejection score, the cumulative summation of each acute rejection episode [6], was computed for recipients surviving greater than 6 months and did not differ between NO ECMO and ECMO groups (Fig 2). Kaplan-Meier curves illustrating freedom from BOS for lung transplant recipients surviving 6 months are noted in Figure 3. Again, there appears to be no demonstrable difference between ECMO and NO ECMO groups.
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
The results of this current study support the practice of using ECMO in lung transplant recipients suffering from PGF. The prevalence of ECMO use at 4.6% is not dissimilar to other studies of lung transplant recipients that report prevalence between 2.7% and 3.6% [9, 10]. At our center, the two groups of recipients most likely to receive ECMO are retransplants and patients with PPH. Data from the Washington University experience [9] similarly suggests a preponderance of recipients with PPH requiring ECMO. However, we did not see the similar disproportionate number of women that was noted in their series. Differences between our ECMO and NO ECMO groups were also noted in ethnicity and type of transplant. Disparities of ethnic composition between groups may be secondary to a higher likelihood of Non-Caucasians to have interstitial pulmonary fibrosis and "other" diagnoses, or to have undergone retransplantation, which account for 58.3% of our ECMO recipients. Single lung transplant recipients are less likely to require ECMO posttransplant, but this is likely due to the association between single lung transplants and chronic obstructive pulmonary disorder as the underlying diagnosis.
Interestingly, there was a slightly increased risk for bilateral sequential living-related lobar recipients (LRLT) to require ECMO. All of our LRLT recipients were CF patients, which are not at increased risk of needing ECMO. In fact, only 2 CF patients out of 119 required ECMO, the 2 who underwent LRLT. Total ischemic times, thought to play a role in IRI, are minimal in LRLT. One theory is that lobar grafts, when confronted with the total cardiac output, may be subject to more severe IRI secondary to a relative size mismatch. Although BSA does not accurately represent size discrepancies between donors and recipients in LRLT, it may approximate size mismatching using cadaveric donors. Our results indicated that the mean donor-to-recipient BSA ratio for the ECMO group was 0.95, demonstrating that donors for these patients were, on average, of smaller size than the recipients. This is in contrast to the ratio for the NO ECMO group of 1.04. Although we do not know the clinical implications at this time, additional research in this area may be warranted if IRI could be abrogated by better size matching of graft to donor.
Besides being considered technically more difficult overall, secondary to adhesions and other postoperative changes, retransplant candidates also are more likely to be sensitized to common antigens. Therefore, we examined immunologic factors that may play a role in primary graft failure, especially the presence of pretransplant PRA and HLA mismatching. The number of major histocompatibility complex class I and II mismatches were nearly identical between groups. Although the PRA comparison did not reach statistical significance, only 36 patients overall had a detectable pretransplant PRA. The low prevalence makes this evaluation difficult. Considering there was a relative absolute difference in the prevalence of PRA between the ECMO and NO ECMO groups, further work delineating a role for PRA in IRI may prove beneficial.
Our lung transplant program preferentially used VA ECMO until 2001. It was originally thought that diverting cardiac output from a severely edematous and damaged lung would help alleviate already elevated pulmonary artery (PA) pressures and reduce the extent of IRI. Unfortunately, our initial experience with VA ECMO was quite disappointing. Therefore, we explored other methods of sustaining the small number of lung transplant recipients who would inevitably suffer from PGF and likely have irreversible injury on conventional therapy. As pretransplant evaluation excludes patients with significant cardiac dysfunction, most lung transplant recipients suffering severe IRI have isolated pulmonary failure. Therefore, we considered VV ECMO as a safer alternative to VA ECMO that may provide us with better results in this population.
There are a number of putative advantages of VV ECMO. Catheters for VV ECMO can be inserted using a simple, percutaneous method similar to other large bore central venous access. In our experience, catheter insertion is rapid, taking approximately 5 to 10 minutes, and can be performed easily at the bedside in the intensive care unit. The University of Michigan reviewed its experience with VV ECMO through percutaneous cannulation. In their series, 4 of 94 patients (4.3%) sustained vascular injuries after cannulation attempts. There were 2 arterial and 2 venous injuries. Three of these injuries could be repaired with a simple cutdown and vascular repair. One patient suffered a superior vena cava injury and subsequently died secondary to this complication. Although our series is smaller, we are fortunate to report no complications secondary to cannulation insertion in our VV ECMO recipients.
We originally utilized VA ECMO in an effort to minimize the capillary leak associated with severe IRI by unloading the pulmonary vasculature. When we transitioned to VV ECMO, we were concerned that continued high PA flows would worsen the capillary leak and contribute to graft dysfunction. However, this experience demonstrates that VV ECMO attenuated pulmonary hypertension, reduced the amount of pulmonary edema, and was associated with more rapid resolution of the capillary leak phenomenon. We hypothesize that VV ECMO offers the advantage of supplying appropriately oxygenated blood to the pulmonary circulation, which resuscitates the injured lung parenchyma and reduces hypoxic pulmonary vasoconstriction by altering PA hypoxemia and hypercarbia. Use of VV ECMO may be even more critical in the setting of lung transplantation, as pulmonary grafts lack bronchial artery circulation and are dependent upon the PA for circulatory support. In contradistinction, VA ECMO may lead to continued ischemic injury that threatens to cause anastomotic breakdown and further epithelial injury, as described by Glassman and colleagues [11] in the University of Pittsburg series of lung transplant recipients requiring ECMO.
Sepsis and central nervous system injuries were the most devastating complications in our ECMO group and the vast majority of them occurred in those receiving VA ECMO. Five of our ECMO recipients died secondary to overwhelming sepsis, all of which were in the VA group. Surprisingly, 4 of the 5 lethal infections were Candida species. Newer antifungal agents and more aggressive antimicrobial prophylaxes may assist in preventing these complications. Voriconazole is an approved fluconazole derivative with increased activity against resistant Candida species, such as C. glabrata and C. krusei [12]. Caspofungin (Cancidas, Merck and Co) is a member of a new class of antifungal agents, the echinocandins, which has activity against all species of Candida [12]. Because of the high rate of fungal infections in the early VA ECMO group, a more aggressive antifungal prophylaxis has been used in the VV ECMO recipients.
It has been demonstrated that the leading cause of death in some ECMO series is not cardiopulmonary failure, but rather irreversible cerebral injury from hemorrhage and/or infarction [13]. Similarly, our incidence of significant CNS injuries was high overall, as 8 of 23 (34.8%) ECMO patients sustained devastating neurologic complications. The actual incidence may be even higher, as 7 of the VA ECMO recipients were not completely assessed neurologically prior to their demise. However, it remains difficult to ascertain whether the neurologic insults occurred pre-ECMO or during ECMO. These patients are often unstable hemodynamically, severely hypoxic, and may have been fully anticoagulated for cardiopulmonary bypass or suffered an air embolus from the transplant procedure itself. At the same time, there are data from animals that indicate VA ECMO may decrease cerebral blood flow by 25% and cerebral oxygenation by 30%, while VV ECMO has no effect upon these parameters [14]. The one VV ECMO patient with a CNS injury had findings on CT scan consistent with a diffuse cerebral anoxic injury that likely occurred prior to ECMO initiation.
In conclusion, ECMO is an acceptable method of sustaining lung transplant recipients suffering from PGF. Although very effective at improving cardiopulmonary function, significant complications may arise during ECMO use. At our institution, VV ECMO is associated with improved outcomes and fewer complications than VA ECMO. Differences in survival are primarily related to neurologic catastrophes and severe sepsis. Most importantly, timely initiation of VV ECMO in a select population has resulted in survival that approximates that of lung recipients not requiring ECMO, with very few serious adverse events. We recommend VV ECMO support for all lung transplant recipients with severe, life-threatening IRI unless severe cardiac dysfunction refractory to VV ECMO is concomitantly present.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
My question to you is, was there a selection bias by the surgeon to use VA ECMO for only the most unstable patients? Many of us reserve VA ECMO for those patients who require immediate resuscitation. If the patients were so unstable are these results really so poor, given the fact that if you choose VA for the more unstable patients, perhaps this represents a much sicker group of patients.
DR HARTWIG: Initially, our institutional protocol involved VA ECMO to support all recipients. Similarly, if one reviews some of the early studies of ECMO use for lung transplant recipients, the vast majority described utilization of VA ECMO, regardless of patient status. The Washington University experience that they reported was exclusively VA ECMO, and Pittsburgh has also reported primarily using VA ECMO. We had such disappointing outcomes with our VA group that we wanted to try an alternative method of support. We thought that VV ECMO would be a much better way to support these patients because, as you mentioned, it provides oxygenated blood to a pulmonary allograft lacking bronchial artery circulation. Also, because VV ECMO can be performed quickly and safely at the bedside with minimal complications, we were able to minimize the delay in onset of extracorporeal life support. Therefore, our protocol was changed to initially use VV ECMO for all recipients, with VA ECMO being reserved for those recipients demonstrating overt cardiac failure and remain hemodynamically compromised despite VV ECMO. We have yet to encounter such a circumstance since our protocol change and we believe that VV ECMO effectively stabilizes the vast majority of recipients with severe acute graft dysfunction and that VA ECMO is rarely required.
DR ZWISCHENBERGER: So now that you feel that VV is superior, would you choose VV for a patient who appears unstable, since many of us have seen that once you stabilize the gas exchange with VV, the hemodynamics likewise stabilize.
DR HARTWIG: Correct. As you mentioned, and we demonstrated in this study, one does experience improved hemodynamics by regulating the gas exchange.
DR ZWISCHENBERGER: Excellent experience. Thank you.
DR P. MICHAEL MCFADDEN (New Orleans, LA): I was looking at the complications that you posted and I did not see any leg problems with the venoarterial perfusion listed. We have a fairly large pediatric cystic population, about 35%, which have small vessels, and we have been using a separate perfusion cannula into the distal extremity. Have you had any experience with this, and do you have any suggestions about the management of leg ischemia in small patients?
DR HARTWIG: It can indeed be a problem, particularly with the VA ECMO. Our experience with pediatric patients is a small component of our overall volume and we have been fortunate not to suffer significant peripheral vascular complications in our ECMO patients. As you mention, with smaller vessels, using a separate perfusion device is likely the best way to avoid lower extremity ischemia. Your question does highlight another significant advantage of VV ECMO, in that maldistribution of perfusion to the lower extremities does not occur.
DR THOMAS C. WOZNIAK (Indianapolis, IN): I have a technical question specifically pertaining to the use of ECMO in unilateral transplant. How do you manage the contralateral lung? Do you remain intubated with a single-lumen tube or do you go back to a double-lumen tube and use selective ventilation? When you are using ECMO for a single-lung transplant, how do you manage the other lung? Do you go back to selective ventilation or do you continue to ventilate the lung that's failing?
DR HARTWIG: Typically we maintain intubation with a single-lumen tube. Physiologically, we would prefer both lungs to be ventilated at "protective" levels in which PEEP (positive end-expiratory pressure), peak inspiratory pressures, and inspired oxygen are maintained at noninjurious levels, but are present to prevent atelectatic collapse.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Wang, S. D. Lick, X. Zhou, X. Liu, R. J. Benkowski, and J. B. Zwischenberger Ambulatory Oxygenator Right Ventricular Assist Device for Total Right Heart and Respiratory Support Ann. Thorac. Surg., November 1, 2007; 84(5): 1699 - 1703. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Walker, J. Primmer, B. E. Searles, and E. M. Darling The potential of accurate SVO2 monitoring during venovenous extracorporeal membrane oxygenation: an in vitro model using ultrasound dilution Perfusion, July 1, 2007; 22(4): 239 - 244. [Abstract] [PDF]
|
|
|
|
 |
|
 C. Aigner, W. Wisser, S. Taghavi, G. Lang, P. Jaksch, D. Czyzewski, and W. Klepetko Institutional experience with extracorporeal membrane oxygenation in lung transplantation Eur. J. Cardiothorac. Surg., March 1, 2007; 31(3): 468 - 474. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. P. Mason, D. J. Boffa, S. C. Murthy, T. R. Gildea, M. M. Budev, A. C. Mehta, A. M. McNeill, N. G. Smedira, J. Feng, T. W. Rice, et al. Extended use of extracorporeal membrane oxygenation after lung transplantation J. Thorac. Cardiovasc. Surg., October 1, 2006; 132(4): 954 - 960. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
