Impact of graft ischemic time on outcomes after bilateral sequential single-lung transplantation

Impact of Graft Ischemic Time on Outcomes After Bilateral Sequential Single-Lung Transplantation Tetsuya Ueno, MD, Gregory I. Snell, FRACP, Trevor J. Williams, FRACP, Thomas C. Kotsimbos, FRACP, Julian A. Smith, FRACS, Marc Rabinov, FRACS, and Donald S. Esmore, FRACS Heart and Lung Transplant Service, Alfred Hospital, Victoria, Australia

Background. Graft ischemic time (GIT) is a potential limiting factor in lung transplantation. Methods. Seventy-four patients who underwent bilateral sequential single-lung transplantation were divided into three groups: group I, GIT less than 5 hours (n 5 20); group II, GIT between 5 and 8 hours (n 5 39); and group III, GIT more than 8 hours (n 5 15). We compared early allograft function (ratio of arterial oxygen tension to inspired oxygen fraction and alveolar–arterial oxygen gradient), blood loss, the need for tracheostomy, the duration of ventilation, intensive care unit stay, and hospital stay. We also compared prevalences of acute and chronic rejection, airway complications, lung function test, and 2-year survival.

Results. Early allograft function in group III was significantly worse than those in groups I and II. However, there was no significant difference in any other variables of early and medium-term outcomes among the three groups. No significant correlation was detected between GIT and duration of intensive care unit stay or hospital stay. Conclusions. The limitation of acceptable GIT could be extended from the traditionally approved 4 to 5 hours, to 5 to 8 hours or even longer.

I

and compared pulmonary allograft function and early and medium-term outcomes among three groups.

n the last decade, lung transplantation has become a well-established therapeutic procedure for selected patients with end-stage pulmonary disease. However, a long-standing shortage of suitable donor organs and an undefined margin of acceptable graft ischemic time (GIT) have been major limitations for its more widespread application. Contrary to encouraging animal studies that demonstrate safe GITs of more than 24 hours [1], the acceptable upper limitation of GIT in clinical situations has been believed to be 4 to 5 hours [2]. There are a limited number of clinical reports in which the authors compared pulmonary graft function, the prevalence of posttransplant complications, and survival between groups with shorter and longer GITs, finding no significant difference between two groups [3–5]. However, they considered the results after single-lung transplantation (SLT), double-lung transplantation, and combined heartlung transplantation (HLT) all together. In our present study, we selected only the patients after bilateral sequential single-lung transplantation (BSSLT) for various indications except pulmonary hypertension and Eisenmenger’s syndrome. We divided them into three groups, based on duration of GIT, of less than 5 hours, between 5 and 8 hours, and more than 8 hours,

Accepted for publication Jan 15, 1999. Address reprint requests to Dr Esmore, Heart and Lung Transplant Service, Alfred Hospital, Commercial Rd, Prahran, Victoria 3181, Australia.

© 1999 by The Society of Thoracic Surgeons Published by Elsevier Science Inc

(Ann Thorac Surg 1999;67:1577– 82) © 1999 by The Society of Thoracic Surgeons

Patients and Methods Between August 1990 and December 1997, 97 BSSLTs were performed at the Alfred Hospital. Of these patients, 74 patients were included in this study. Thirteen patients with diagnosis of pulmonary hypertension or Eisenmenger’s syndrome and 10 patients with incomplete data were excluded. In BSSLT for the patients with pulmonary hypertension or Eisenmenger’s syndrome, the use of cardiopulmonary bypass (CPB) from the beginning of the operation is mandatory and consequently the adverse effect of long-term cardiopulmonary bypass support (mean period, 261 minutes, unpublished data) on posttransplant allograft function cannot be neglected. In addition, these patients had taken warfarin until the day of transplantation, which could cause postoperative coagulopathy problems, requiring an increased amount of blood products. The major reason to exclude another 10 patients was that blood sampling for blood gas analysis was not performed at one or two scheduled data points.

Donor Operation Donor organ procurement and preservation were standardized. For donor lungs, prostacyclin (Flolan, Wellcome, Sydney, Australia) was infused at 40 to 80 ng z kg21 z min21 for approximately 10 minutes intravenously and 4 to 6 L of cold modified Euro-Collins solution was 0003-4975/99/$20.00 PII S0003-4975(99)00309-4

1578

UENO ET AL EXTENDED LIMITATION OF GIT IN BSSLT

Ann Thorac Surg 1999;67:1577– 82

Table 1. Recipients’ Demographicsa Characteristics Age (y) Sex (M/F) Ischemic time (min) Anesthetic time (min) Previous thoracotomy CPB support Underlying disease Cystic fibrosis Bronchiectasis Obstructive disease ILD Others

Group I (n 5 20)

Group II (n 5 39)

Group III (n 5 15)

34 6 12 8/12 244 6 29 410 6 79 3 1

35 6 12 27/12 393 6 51b 454 6 79 4 4

37 6 13 9/6 527 6 40c 498 6 103d 4 1

9 5 4 0 2

24 7 6 1 1

8 4 2 1 0

a b Values are presented as the mean 6 standard deviation. p , c p , 0.0001, group II versus group 0.0001, group I versus group II; d p 5 0.0139, group I versus group III. II;

CPB 5 cardiopulmonary bypass; interstitial lung disease.

GIT 5 graft ischemic time;

ILD 5

administered at a pressure of 40 cm H2O through a cannula in the main pulmonary artery. The donor trachea was stapled after the lungs were inflated with 100% oxygen at a pressure of 5 cm H2O. Donor lungs were immersed into cold modified Euro-Collins solution for transport.

Recipient Demographics and Operation The recipients’ demographics are presented in Table 1. The GIT was defined as the average value of the ischemic time for the first and second transplanted lungs in our study. The ischemic time for each lung was that time between aortic cross-clamping of the donor heart and reperfusion of the transplanted lung. The technical details of our method of BSSLT have been described elsewhere [6]. Cardiopulmonary bypass was used for 6 patients when the first transplanted lung could not support the patient during the implantation of the second lung. Cardiopulmonary bypass was instituted with aortic and right atrial cannulation.

Postoperative Management The midthoracic epidural catheter was inserted before induction of anesthesia, and a mixed solution of local anesthetic and narcotic was continuously infused through this catheter for up to 5 days. The modes of ventilator setting were uniformly standardized in the intensive care unit (ICU). Percutaneous tracheostomy was performed in those patients who could not wean from the ventilator within 7 days of transplantation.

Immunosuppression and Prophylaxis of Infection Patients undergoing transplantation between August 1990 and October 1992 received cytolytic therapy with antithymocyte globulin for 7 to 10 days from the time of transplantation. Cyclosporine and azathioprine were commenced before transplantation and prednisolone added from day 7. Patients undergoing transplantation

after October 1992 were not given cytolytic induction, and all commenced triple therapy (cyclosporine, azathioprine, and prednisolone) immediately after operation. Maintenance therapy included cyclosporine (2 to 20 mg/kg per day), azathioprine (1 to 2 mg/kg per day), and prednisolone (0.1 to 0.2 mg/kg per day). Bronchoscopies with biopsy were performed to evaluate the airway on a regular basis at 2, 4, 8, 12, 26, 39, and 52 weeks or when there was a clinical suspicion of airway complications or allograft rejection. Antibiotics active against known or suspected organisms were given intravenously for a short period after transplant. The long-term prophylaxis for Pneumocystis carinii was achieved with lowdose oral trimethaprim-sulfamethoxazole. Intravenous ganciclovir was used for prophylaxis of Cytomegalovirus infection in serologically positive donors or recipients.

Statistical Analysis A computer program package, Statview 4.0 (Abacus Concepts, Berkeley, CA) for Macintosh, was used for statistical analysis. Data are presented as mean 6 standard deviation. Repeated two-way analysis of variance was used for comparisons of ratio of arterial oxygen tension to inspired oxygen fraction (Po2/Fio2) and alveolar–arterial oxygen gradient (A-aDo2), and Scheffe’s method was used as a post hoc test. This method was also used for comparisons of percent predicted forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV1). Comparisons of age, GIT, anesthetic time, blood loss, durations of ventilation, ICU stay, hospital stay, and follow-up were made by Kruskal-Wallis test, followed by Mann-Whitney U test. Correlation between GIT and duration of ICU stay or hospital stay was analyzed by Spearman’s rank test. Sex ratio, incidences of previous thoracotomy, cardiopulmonary bypass support, tracheostomy, hospital death, total death, sepsis and multiple organ failure, prevalences of rejection within 90 days after transplantation, airway complication, and bronchiolitis obliterans syndrome (stage III) were compared by x2 analysis. Two-year survival rate was estimated by the Kaplan-Meier method. A probability value of less than 0.05 was considered statistically significant.

Results No significant difference was found in patients’ age ( p 5 0.763) and sex ratio ( p 5 0.096) among the three groups. The mean GITs of the first and second transplanted lungs were, respectively, 184 and 304 minutes (average, 244 minutes) in group I, 327 and 459 minutes (average, 393 minutes) in group II, and 465 and 589 minutes (average, 527 minutes) in group III. The anesthetic time in group III was significantly longer than that in group I ( p 5 0.0139). There was no significant difference in prevalences of previous thoracotomy ( p 5 0.316) and cardiopulmonary bypass support during operation ( p 5 0.762) among the three groups. We have routinely accepted marginal donor lungs with Pao 2 approximately 300 mm Hg at Fio2 of 1.0. The Pao2 of the donors in groups I, II, and III were 474 6 61, 449 6 96, and 473 6

Ann Thorac Surg 1999;67:1577– 82

UENO ET AL EXTENDED LIMITATION OF GIT IN BSSLT

1579

Table 2. Early Outcomes After Transplantationa Variables Blood loss (mL/24 h) Ventilation (h) ICU stay (h) Hospital stay (days) Tracheostomy Hospital death a

Group I (n 5 20)

Group II (n 5 39)

Group III (n 5 15)

1302 6 1029

1341 6 923

1562 6 947

113 6 143 149 6 343 211 6 587 (median, 38) (median, 27) (median, 25) 189 6 256 186 6 339 281 6 579 (median, 76) (median, 72) (median, 69) 29 6 17 37 6 28 34 6 23 (median, 22) (median, 26) (median, 28) 6 6 3 0 3 1

Values are presented as the mean 6 standard deviation.

There were no significant differences in any variables among the three groups.

Fig 1. Changes in the ratio of arterial oxygen tension to inspired oxygen fraction (Po2/Fio2) after transplantation. Each point represents the mean 6 standard deviation. The Po2/Fio2 in group III was significantly lower than those in groups I (p 5 0.0126) and II (p 5 0.0121). (ANOVA 5 analysis of variance.)

73 mm Hg at Fio2 of 1.0, respectively, which was not significantly different ( p 5 0.474).

Allograft Function Allograft function evaluated as Po2/Fio2 and A-aDo2 within 24 hours after ICU admission is shown in Figures 1 and 2. Repeated measures two-way analysis of variance

showed significant differences in Po2/Fio2 and A-aDo2 among the three groups ( p 5 0.023 for Po2/Fio2 and 0.009 for A-aDo2). The Po2/Fio2 in group III was significantly lower than those of groups I ( p 5 0.013) and II ( p 5 0.012). The A-aDo2 in group III was significantly higher than those of groups I ( p 5 0.005) and II ( p 5 0.006). No significant difference either in Po2/Fio2 ( p 5 0.763) or A-ao2 ( p 5 0.650) was found between groups I and II.

Early Outcomes The early outcomes after transplantation are presented in Table 2. There was no significant difference in blood loss within 24 hours after ICU admission among the three groups ( p 5 0.486). The need for tracheostomy did not differ among the three groups ( p 5 0.417). No statistically significant difference was found in durations of ventilation ( p 5 0.558), ICU stay ( p 5 0.728), and hospital stay ( p 5 0.575) among the three groups. There was no significant difference in incidence of hospital death among the three groups ( p 5 0.452). There were four hospital deaths; three in group II and one in group III. The causes of death were sepsis related to airway complication in 2 patients and accidental air embolism through central venous line in one in group II, and sepsis followed by multiple organ failure in 1 patient in group III.

Correlation Between Graft Ischemic Time and Intensive Care Unit Stay or Hospital Stay The correlations between GIT and duration of ICU stay or hospital stay in all patients included in three groups are presented in Figures 3 and 4. There was no statistically significant correlation between the length of GIT and duration of ICU stay ( p 5 0.2338) or hospital stay ( p 5 0.2725). Fig 2. Changes in alveolar–arterial oxygen gradient (A-aDo2) after transplantation. Each point represents the mean standard 6 deviation. The A-aDo2 in group III was significantly higher than those in groups I (p 5 0.0046) and II (p 5 0.0057). (ANOVA 5 analysis of variance.)

Pulmonary Function Test The percentages of predicted FVC and FEV1 at 1, 3, and 6 months after transplantation in three groups are presented in Figure 5. Both FVC and FEV1 in each group

1580

UENO ET AL EXTENDED LIMITATION OF GIT IN BSSLT

Ann Thorac Surg 1999;67:1577– 82

Fig 3. Correlation between graft ischemic time (GIT) and duration of intensive care unit (ICU) stay. There was no significant correlation between GIT and duration of ICU stay (p 5 0.2338).

improved with time after transplantation. However, there was no significant difference in FVC ( p 5 0.5557) and FEV1 ( p 5 0.1774) among the three groups.

Medium-Term Outcomes The medium-term outcomes after transplantation are presented in Table 3. There was no significant difference in follow-up time among the three groups ( p 5 0.725). There was no significant difference in prevalences of rejection within 90 days ( p 5 0.265), airway complication such as stricture more than 50% or extensive local ischemia at the airway anastomosis ( p 5 0.081), and stage III bronchiolitis obliterans syndrome ( p 5 0.246) among the three groups. There was no significant difference in incidences of total death ( p 5 0.074), sepsis ( p 5 0.399), and multiple organ failure ( p 5 0.483) among the three groups. In addition to the patients who died in the

Fig 5. (A) Percent of predicted forced vital capacity (FVC) after transplantation. There was no significant difference in predicted FVC (p 5 0.5557) among the three groups. (B) Percent of predicted forced expiratory volume in 1 second (FEV1) after transplantation. There was no significant difference in predicted FEV1 (p 5 0.1774) among the three groups. (ANOVA 5 analysis of variance.)

Fig 4. Correlation between graft ischemic time (GIT) and duration of hospital stay. There was no significant correlation between GIT and duration of hospital stay (p 5 0.2725).

hospital, there were one medium-term death in group I, nine in group II, and two in group III. The patient in group I died of nonspecific graft failure. Five patients in group II and 2 in group III died of sepsis. Another 4 patients in group II died of multiple organ failure, rejection, subarachnoid hemorrhage, and accidental hemorrhage from granulation tissue at airway anastomosis. Two-year survival (Fig 6) showed no significant difference among the three groups by log-rank test (x2 5 3.833, two degrees of freedom, p 5 0.147).

Ann Thorac Surg 1999;67:1577– 82

UENO ET AL EXTENDED LIMITATION OF GIT IN BSSLT

Table 3. Medium-Term Outcomes After Transplantationa Variables Follow-up period (mo) Rejection Airway complicationb BOS (grade III) Total death Death from Sepsis MOF

Group I (n 5 20)

Group II (n 5 39)

Group III (n 5 15)

18 6 15 8 8 0 1

22 6 18 23 18 3 12

17 6 16 6 2 0 3

1 0

5 1

3 1

Values for follow-up period are presented as the mean 6 standard deviation. There were no significant differences in any variables b Airway complication (stricture . 50% or among the three groups. extensive local ischemia at airway anastomosis). a

BOS 5 bronchiolitis obliterans syndrome (grade III); organ failure.

MOF 5 multiple

Comment Despite numerous experimental investigations that have been performed to extend preservation time of pulmonary allograft, acceptable GIT is still suggested to be limited to 4 to 5 hours in clinical practice [2]. However, to the best of our knowledge, only a few clinical studies show a direct relationship between GIT and graft function and early and late outcomes after transplantation. Winton and colleagues [3] found no significant difference in allograft function, airway healing, rejection rate, and survival after 51 SLT and 52 double-lung transplantation between two groups: one with GIT less than 5 hours and the other with GIT in excess of 5 hours. Wahlers and colleagues [4] showed that GIT had no significant effect on postoperative oxygenation at 6 hours and also showed an excellent lung function early postoperatively with GIT up to 6.5 hours. Kshettry and coworkers [5] demonstrated that prolonged GIT more than 6 hours had no negative impact on ICU and hospital stay and actuarial survival. However, they considered the

Fig 6. Two-year survival rates in the three groups. There was no significant difference in 2-year survival rates among the three groups (p 5 0.1471).

1581

combined results of SLT, double-lung transplantation, and HLT all together in their studies. Because only one pathologic lung is replaced and the other native lung remains in situ after SLT, pulmonary gas exchange capacity and actual survival may be affected to various degrees by the condition of the remaining native lung, possibly making the direct impact of GIT of the transplanted lung somewhat inaccurate. In heartlung transplantation, impaired function of the transplanted heart may elevate pulmonary venous pressure and affect pulmonary allograft function. For these reasons, it is possible to evaluate the actual and direct impact of GIT on immediate allograft function only in the settings of double-lung transplantation because both pathologic lungs are replaced by donor lungs. In our present study, we defined the early allograft function as gas exchange capacity within 24 hours after admission to the ICU because other factors such as acute rejection and infection may affect the allograft function to a greater extent at a time later than the first 24 hours. The average GITs in groups I and II were 241 6 28 and 396 6 51 minutes in our study. These were similar to those in the two groups with GIT more or less than 5 hours in the study by Winton and colleagues [3]. In comparing groups I and II, we reached a conclusion, similar to those reported by these authors, that allograft function and early posttransplant outcome were not significantly affected by GIT in excess of 5 hours [3]. It may imply that our present method of lung preservation, infusion of prostacyclin and flush by modified EuroCollins solution, could extend the period of acceptable GIT from the traditionally approved 4 to 5 hours to 5 to 8 hours in clinical lung transplantation. Although pulmonary gas exchange capacity within 24 hours after ICU admission was significantly affected by longer GIT in group III, such significantly impaired allograft function did not translate into a negative impact on any variables of early outcome. This was also supported by our further statistical analyses, which showed no significant correlation between GIT and duration of ICU or hospital stay. The absence of statistical differences in the percentage of predicted FVC and FEV1 among the three groups also suggested that longer GIT did not have any deleterious influence on allograft function after discharge from the hospital. There are several possible explanations for these results. First, the pulmonary allograft may have an unexpected quick recovery in its function from ischemia-reperfusion injury even after longer GIT. Several clinical and experimental studies may support our hypothesis. Pasque and colleagues [7] reported that immediate postoperative perfusion scan demonstrated 80% perfusion to the first transplanted lung (GIT, 6 hours) and 20% to the second lung (GIT, 9 hours 15 minutes), but perfusion was equal to each lung by the following day. Ultrastructural studies have shown capillary edema, as well as detachment of endothelial cells and type I pneumocytes from the basement membrane demonstrated immediately after reperfusion, started improving by 4 hours after reperfusion [8]. Second, we believe that GIT is one of the most impor-

1582

UENO ET AL EXTENDED LIMITATION OF GIT IN BSSLT

tant contributors to allograft function in solid organ preservation including heart, liver, and kidney, but may not have such a close time-related impact on the transplanted lung. Other factors influencing allograft function including pretransplant condition, reperfusion edema, acute rejection, infection, and lymphostasis may have greater influences on pulmonary allograft function after transplantation than have been previously thought, and consequently also have a greater influence on early postoperative outcomes. Further analyses of these contributors are necessary to support this hypothesis. Third, as we have described before, we selected only patients having BSSLT and excluded patients having SLT and HLT to avoid any deleterious effects of the remaining native lung in SLT and simultaneously transplanted heart in HLT on the transplanted lung. Therefore, once bilateral allografts have recovered from ischemic injury and from other possibly deleterious effects, such as reperfusion edema, acute rejection, and infection, which frequently occur in any sorts of lung transplantation in early stage, the postoperative course of almost all patients without major complications after BSSLT is expected to be relatively uniform under totally standardized respiratory care in ICU and general wards as well as in outpatient clinic. Considering that GIT was presented as the average ischemic time of the first and second transplanted lung in our study, which was 526 6 37 minutes in group III, it is surprising that the second transplanted lung with GIT almost up to 10 hours or even longer resumed their function to an acceptable degree without occurrence of primary graft dysfunction and functional deterioration. These results will encourage members of transplantation centers throughout the world to use donor’s lungs with GIT around 8 hours or even longer, which have been abandoned for use for fear of primary graft dysfunction, and contribute to increase the numbers of patients who can benefit from BSSLT. However, we stress that these results of varying GIT in BSSLT cannot directly be applied in a similar fashion to SLT and HLT because of the reasons stated before. As far as medium-term results are concerned, we found no significant differences in any of the variables among the three groups. Some investigators showed that acute [9] or chronic [10] allograft rejection was related to GIT, which was not detected in our study. Despite the lack of statistically significant difference in 2-year survival among the three groups, it cannot be instantly ignored that group I with GIT less than 5 hours showed a relatively better 2-year survival. These results are similar to those reported previously by our group that showed that GIT greater than 5 hours was associated with significant reduction in survival at 2 years in a study of patients after SLT, BSSLT, and HLT [11]. Although the increased anesthetic time as a marker of surgical difficulity might partially contribute to the prolongation of GIT,

Ann Thorac Surg 1999;67:1577– 82

this does not appear to affect survival. Because the number of patients in each group was small and the follow-up period was not long enough in our present study, further analyses about long-term outcomes among the three groups will be necessary to make a conclusive comment on clinical availability and safety of pulmonary allograft with GIT around 8 hours or even longer. In conclusion, our current lung preservation method (prostacyclin infusion combined with a modified EuroCollins solution flush) could extend the limitation of acceptable GIT, which had been recognized to be 4 to 5 hours, to 5 to 8 hours or even longer in the setting of BSSLT. Although allograft function was significantly impaired immediately (within 24 hours) after transplant in the longer GIT group (GIT more than 8 hours), allograft may recover more quickly than has been believed, providing no significantly negative impact on the early and medium-term outcomes after BSSLT. Further investigations with more accumulated cases and longer follow-up period are still required to conclude definitively about limitations of clinically acceptable GIT in lung transplantation.

References 1. Date H, Matsumura A, Manchester JK, et al. Evaluation of lung metabolism during successful twenty-four-hour canine lung preservation. J Thorac Cardiovasc Surg 1993;105:480–91. 2. Kirk AJ, Colquhoun IW, Dark JH. Lung preservation: a review of current practice and future directions. Ann Thorac Surg 1993;56:990 –1000. 3. Winton TL, Miller JD, deHoyos A, Snell G, Maurer J. Graft function, airway healing, rejection, and survival in pulmonary transplantation are not affected by graft ischemia in excess of 5 hours. Transplant Proc 1993;25:1649–50. 4. Wahlers T, Schafers HJ, Cremer J, et al. Organ preservation for heart-lung and lung transplantation. Thorac Cardiovasc Surg 1991;39:344– 8. 5. Kshettry VR, Kroshus TJ, Burdine J, Savik K, Bolman M III. Does donor organ ischemia over four hours affect long-term survival after lung transplantation? J Heart Lung Transplant 1996;15:169–74. 6. Esmore DS, Brown R, Buckland M, et al. Techniques and results in bilateral sequential single lung transplantation. J Cardiac Surg 1994;9:1–14. 7. Pasque MK, Cooper JD, Kaiser LR, et al. Improved technique for bilateral lung transplantation: rationale and initial clinical experience. Ann Thorac Surg 1990;49:785–91. 8. Mills AN, Hooper TL, Hall SM, McGregor CG, Hawthorn SG. Unilateral lung transplantation: ultrastructural studies of ischemia-reperfusion injury and repair in the canine pulmonary vasculature. J Heart Lung Transplant 1992;11: 58– 67. 9. Shennib H, Serrick C, Reis A, Giaid A. Prolonged ischemia is associated with more pronounced rejection in the lung allograft [Abstract]. J Heart Lung Transplant 1994;13:A1097. 10. Paradis IL, Yousem SA, Griffith BP. Airway obstruction and bronchiolitis obliterans after lung transplantation. Clin Chest Med 1993;14:751– 63. 11. Snell GI, Rabinov M, Griffiths A, et al. Pulmonary allograft ischemia time: an important predictor of survival after lung transplantation. J Heart Lung Transplant 1996;15:160– 8.