- Email: [email protected]
Outcomes and Long-term Survival After Pulmonary Retransplantation: A Single-Center Experience
Andreas Wallinder, MD, PhD, Christian Danielsson, MD, Jesper Magnusson, MD, PhD, Gerdt C. Riise, MD, PhD, and G€ oran Dellgren, MD, PhD Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, Gothenburg; Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg; and Transplant Institute, Sahlgrenska University Hospital, Gothenburg, Sweden
Background. The median survival after lung retransplantation (ReLTx) reported to the International Society of Heart and Lung Transplantation is restricted to 2.5 years. We report the results after ReLTx from our center. Methods. A retrospective data collection was performed for the 635 patients who underwent lung transplantation between 1991 and 2017 at our center. Recipient variables were compared between patients undergoing only primary lung transplantation (PLTx) and those undergoing PLTx and later ReLTx. Time to death was compared using the Kaplan-Meier method. The risk of ReLTx was analyzed in Cox regression models. Any interaction between type of transplantation, single/double, and PLTx/ReLTx was investigated. Results. ReLTx was performed in 49 patients. Survival after ReLTx at 30 days and 1, 2, and 5 years was 90%, 76%, 71%, and 55%, respectively, and the corresponding survival after PLTx was 94%, 82%, 76%, and 61%,
respectively. A hazard ratio of 1.73 for ReLTx was shown (95% confidence interval [CI], 1.14 to 2.63; P [ .011). After adjustments for sex, age, diabetes, renal function, preoperative ventilator, and extracorporeal membrane oxygenation, the hazard ratio was 1.43 (95% CI, 0.90 to 2.26; P [ .13). ReLTx was performed in 8 patients (16%) within the first year after PLTx. The 1-year survival for this group was 50% compared with 81% (P [ .18) for patients who underwent ReLTx later than 1 year after the PLTx. Oneyear survival after double ReLTx was 60% (95% CI, 25% to 83%) compared with 79% (95% CI, 63% to 89%) for single ReLTx. Conclusions. ReLTx is a reasonable option for a selected group of patients. Ideally, a number of wellestablished risk factors are avoided and the ReLTx is performed more than 1 year after the PLTx.
L
annual number and proportion of lung retransplantation (ReLTx) increased during the latter part of this era, but the use of a second LTx was significantly less frequent than retransplantations for kidneys, livers, or hearts.3 The median survival after retransplantation was only 2.5 years during this time period. Retransplantation recipients were significantly younger than primary LTx (PLTx) recipients (46 vs 55 years) and more often female (46.6% vs 43.7%).2 Risk factors for poor outcome after ReLTx are the same as after PLTx, including recipient age, preoperative mechanical ventilation, and impaired renal function.4 In addition, early retransplantation and an indication for retransplantation other than bronchiolitis obliterans syndrome has been associated with poor outcome.4-7 We performed a retrospective review of our LTx program regarding the short- and long-term outcome after
ung transplantation (LTx) remains a treatment reserved for a small number of highly selected patients with end-stage lung disease. Long-term survival for the transplant population is limited to a median of 6 years, and is better for recipients of double lungs and for transplantations done after 1998.1 Many recipients are young when the transplant is performed, and hence, the question of a retransplantation may be raised if the transplanted lungs deteriorate and the patient is without major comorbidity. During the era January 1995 to June 2013, 5.1% (799 of 15,631) of single LTxs and 3.4% (925 of 27,213) of bilateral LTxs reported to the International Society of Heart and Lung Transplantation were retransplantations.2 The
(Ann Thorac Surg 2019;108:1037-44) Ó 2019 by The Society of Thoracic Surgeons
Accepted for publication Apr 8, 2019. Presented at the Thirty-ninth Annual Meeting and Scientific Sessions of The International Society for Heart and Lung Transplantation, Orlando, FL, April 3-6, 2019. Address correspondence to Dr Wallinder, Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, Blå Stråket 5, 413 45 Gothenburg, Sweden; email: [email protected].
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
The Supplemental Tables and Supplemental Figure can be viewed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2019.04.028] on http://www.annalsthoracicsurgery.org.
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2019.04.028
GENERAL THORACIC
Outcomes and Long-term Survival After Pulmonary Retransplantation: A SingleCenter Experience
GENERAL THORACIC
1038
WALLINDER ET AL PULMONARY RETRANSPLANTATION
ReLTx with the hypothesis that the results after ReLTx would be comparable to those after PLTx.
Patients and Method The University of Gothenburg Ethics Committee approved this study and waived the need for individual patient consent because only retrospective data were used in the analysis.
Patients LTx was performed in 673 patients who underwent 728 procedures at the Sahlgrenska University Hospital between 1990 and December 31, 2017. In all, 55 ReLTxs were performed. Heart-lung transplantations (n ¼ 37) were excluded from this analysis as were ReLTx procedures in the heart-lung transplantation cohort (n ¼ 4). One patient who had 2 ReLTxs was also excluded (n ¼ 3). The remaining LTx procedures (n ¼ 684) were included in the analysis (Supplemental Figure 1). The PLTx (n ¼ 635) and ReLTx (n ¼ 49) patients were monitored until April 26, 2018. Survival after PLTx and ReLTx were the primary outcomes of the analysis. Recipient variables from PLTx and ReLTx were collected from the medical records. Preoperative kidney function was categorized by estimated glomerular filtration rate (eGFR) of less than 30, 30 to 60, and more than 60 mL/min/1.73 m2. Primary graft dysfunction was categorized as 0, 1, and 2þ3 because the more severe grades 2 and 3 often vary with ventilator settings and both indicate severe graft dysfunction. No patient was lost during follow-up.
Ann Thorac Surg 2019;108:1037-44
After PLTx, each patient was monitored until death or was censored at ReLTx or at April 26, 2018; likewise, during ReLTx, the patients started being at risk at the time of the second surgery (ie, applying left censoring approach,) and were monitored until death or were censored at the end of follow-up. Stepwise adjustment was as follows: model 1 was unadjusted, model 2 included sex and time-updated age, model 3 additionally included time-updated diabetes, model 4 additionally included time-updated eGFR categories, and model 5 additionally included time-updated ventilator preoperatively and extracorporeal membrane oxygenation preoperatively. Missing values were handled as their own categories for each variable respectively in the adjustments and are described in Supplemental Table 1. The overall effects were described by hazard ratios (HRs) with 95% CIs and associated P values. The impact of the time from PLTx on the effect of ReLTx on death was investigated by including an interaction term between the 2 variables. The analysis of time-updated variable of ReLTx was graphically described using the same methodology as was used for Kaplan-Meier estimates but incorporating left-censoring of the starting point of the time period from ReLTx as a technique for performing time-updated analysis, and the 95% CIs using Greenwood’s formula for the SE calculation.8 An interaction between type of transplantation, single/ double, and PLTx/ReLTx was investigated. Missing values were not imputed. All analyses were performed using SAS 9.4 software, (SAS Institute Inc, Cary, NC). All tests were 2-tailed and conducted at the 0.05 significance level.
Statistical Analysis Recipient preoperative, intraoperative, and postoperative variables at the time of the PLTx were compared between patients undergoing only PLTx and those undergoing both PLTx and later ReLTx by using the Fisher exact test for dichotomous variables, Mantel-Haenszel c2 test for ordered categorical variables, c2 test for nonordered categorical variables, and Mann-Whitney U test for continuous variables. The same methods were also used for tests between data collected at the time of PLTx vs the time of ReLTx. The patients were independent within the 2 groups, resulting in conservative tests. Time to death from the surgery date of PLTx and ReLTx, respectively, were described in graphs by using Kaplan-Meier technique; survival probability with 95% confidence intervals (CI) is shown. The risk of ReLTx on time to death as an outcome variable was analyzed in unadjusted and adjusted Cox regression models by using the time-updated variable of ReLTx. Time-updated analysis was chosen to statistically compare the survival after a PLTx with the survival after a ReLTx. Handling the ReLTx group on a patient level including also the PLTx period in the ReLTx group would introduce a statistical bias called immortal time bias, because the future ReLTx is not known for a patient until that time point is considered in the analysis.
Results Baseline Characteristics The PLTx cohort consisted of 243 (38.3%) single LTxs with a retransplantation rate of 11% and 392 (61.7%) double LTxs with a retransplantation rate of 5.6%. Single LTx was the preference in the earlier years of the experience: 89 (65.4%) in 1991 to 2000, 126 (53.6%) in 2001 to 2010, and 28 (10.6%) in 2011 to 2017 (P < .001), which may skew data because chronic lung allograft dysfunction had time to develop in more patients in this cohort. The number of annual PLTxs and ReLTxs are presented in Figure 1.
PLTx Data The recipients who later underwent ReLTx were younger at the time of their initial transplantation than the group that did not (mean 47 vs 52 years, P < .001), had a higher eGFR (mean 100 vs 91 mL/min/1.73 m2, P ¼ .010), and more often had a single LTx (55.1% vs 36.9%, P ¼ .019). The mean donor age for the group that later had a second transplantation was lower (41 vs 45 years, P ¼ .047). The cold ischemic time for the second lung, in the event of double LTx, was significantly longer for the group that later had a ReLTx (mean 534 vs 396 minutes, P ¼ .005). Cold ischemic time data were only available from 15 of 22
1039
WALLINDER ET AL PULMONARY RETRANSPLANTATION
50
45
40
35
30
25
42
38 30
20
36
36
41
35
39
34
31 29 15
22 23
23
19
10
16 15 12
5
14
17 13
15
17 12
12
11
1 6 6 1 1 5 1 1 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 4
0 PTx
4
12
14
15
Re-DLTx Re-SLTx
13
17
11
23
15
12
12
23
17
16
1 1
19
29
1 1
1
2
1
2
22
31
1
1
2
2
30
36
35
38
41
42
1 6
2
3
6
5
34
36
39
1
1
3
2
3
Figure 1. Annual number of primary lung transplantations (PTx) and retransplantations over 27 years at a single center. (Re-DLTx, double-lung retransplantation; Re-SLTx, single-lung retransplantation.)
patients in this group. The number of treated episodes of rejection, verified by biopsy specimen or not, was not higher for PLTx patients who later had a second LTx. Recipient preoperative, intraoperative, and postoperative characteristics are presented in detail in Supplemental Table 1 and grouped for patients who later did or did not have a second LTx.
ReLTx Data Recipient characteristics are presented in Table 1. The median time from PLTx to ReLTx was 3.7 years (range, 0.06-18 years), and the median follow-up was 2.9 years (range, 0-20 years) in the ReLTx cohort vs 4.0 years (range, 0-27 years) in the PLTx cohort. ReLTx recipients had similar age as the PLTx cohort (52 vs 51 years, P ¼ .77). The ReLTx group had a lower eGFR (66 vs 91 mL/min/ 1.73 m2, P < .001) and a higher prevalence of diabetes mellitus (22% vs 9%, P ¼ .006), comorbidities that in many cases are unwanted adverse effects of the immunosuppressive treatment after PLTx. The mean donor age for the ReLTx group was older (53 vs 45 years, P ¼ .002). Primary graft dysfunction grade 2 or 3 was diagnosed in 19% after ReLTx and in 23% after PLTx (P ¼ .13). Renal replacement therapy was used in 20% in the ReLTx group and in 13% in the PLTx group (P ¼ .29). The median total ventilator time, including reintubations, was 5 hours in the ReLTx group and 6 hours in the PLTx group (P ¼ .50). The number of treated episodes of rejections, both clinical and verified by biopsy specimen, was similar between groups. Detailed data are presented in Supplemental Table 2.
Indication for the PLTx in Patients Who Later Underwent ReLTx In the group of patients who later underwent ReLTx, 11 different diagnoses were registered as the indication for the PLTx. Patients with pulmonary artery hypertension, a-1-antitrypsin deficiency, lymphangioleiomyomatosis, and cystic fibrosis were most likely to have a ReLTx, probably explained by the younger age at which patients with these diagnoses undergo PLTx. Detailed data are presented in Table 2.
Timing of ReLTx Restrictive use of ReLTx for primary graft dysfunction was implemented throughout the duration of the program. ReLTx was performed in 8 of the 49 patients (16%) within the first year after the PLTx, with a 1-year survival of 50% (4 of 8) compared with 81% (33 of 41, P ¼ .18), for patients who underwent ReLTx later than 1 year after the PLTx.
Survival Survival after ReLTx at 30 days and 1, 2, and 5 years was 90%, 76%, 71%, and 55%, respectively, and the corresponding survival after PLTx was 94%, 82%, 76%, and 61%, respectively (Figure 2). Unadjusted Cox regression analyses showed a significantly higher mortality risk of ReLTx compared with PLTx (HR, 1.73; 95% CI, 1.14 to 2.63; P ¼ .011). After adjustments for sex and timeupdated age, time-updated diabetes, and time-updated eGFR, the HR was reduced to 1.47 (95% CI, 0.93 to 2.32; P ¼ .10). After additional adjustments for preoperative need of ventilator or extracorporeal membrane
GENERAL THORACIC
Ann Thorac Surg 2019;108:1037-44
GENERAL THORACIC
1040
WALLINDER ET AL PULMONARY RETRANSPLANTATION
Table 1.
Ann Thorac Surg 2019;108:1037-44
Patient Characteristics at the Time of PLTx Vs at the Time of ReLTx
Variables Preoperative characteristics Years between first and second LTx Age at LTx, y
Male Diabetes Ventilator pre-op ECMO pre-op eGFR, mL/min/1.73 m2
eGFR categories <30 mL/min/1.73 m2 30-60 mL/min/1.73 m2 >60 mL/min/1.73 m2 FEV1%
6-min walk test, m
Intraoperative characteristics Mechanical circulatory support No Extracorporeal circulation Venovenous ECMO Venoarterial ECMO Lung 1 ischemia, min
Lung 2 ischemia, min
Donor age, y
Postoperative characteristics Time on ventilator, h
Dialysis Primary graft dysfunction Grade 0 Grade 1 Grade 2 and 3 Number of biopsy-verified rejections 0 1 2 3 4
Total (N ¼ 684)
PLTx (n ¼ 635)
ReLTx (n ¼ 49)
4.65 (3.89) 3.68 (0.06-18.04) n ¼ 49 51.3 (12.8) 54.5 (8.7-74.1) n ¼ 684 310 (45.3) 66 (9.9) 30 (4.5) 24 (3.6) 89.8 (22.3) 89.0 (12.0-180.0) n ¼ 641
51.3 (12.9) 54.6 (8.7-74.1) n ¼ 635 283 (44.6) 56 (9.1) 26 (4.2) 23 (3.7) 91.4 (21.7) 90.0 (12.0-180.0) n ¼ 601
4.65 (3.89) 3.68 (0.06-18.04) n ¼ 49 51.6 (11.1) 53.7 (19.8-66.2) n ¼ 49 27 (55.1) 10 (21.7) 4 (9.1) 1 (2.2) 65.9 (16.9) 65.0 (36.0-111.0) n ¼ 40
4 (0.6) 42 (6.6) 595 (92.8) 50.5 (24.7) 41.0 (2.0-100.0) n ¼ 625 285.3 (144.4) 270.0 (3.0-870.0) n ¼ 604
4 (0.7) 25 (4.2) 572 (95.2) 50.4 (25.2) 40.0 (2.0-100.0) n ¼ 585 286.3 (145.8) 270.0 (3.0-870.0) n ¼ 571
0 (0.0) 17 (42.5) 23 (57.5) 52.1 (16.1) 49.0 (26.0-80.0) n ¼ 40 267.8 (116.4) 246.0 (96.0-522.0) n ¼ 33
467 (68.6) 186 (27.3) 13 (1.9) 15 (2.2) 255.3 (100.0) 240.0 (83.0-869.0) n ¼ 550 403.6 (134.5) 385.0 (122.0-978.0) n ¼ 307 45.6 (16.5) 48.0 (5.0-76.0) n ¼ 684
429 (67.9) 178 (28.2) 11 (1.7) 14 (2.2) 254.0 (99.4) 240.0 (83.0-869.0) n ¼ 516 402.9 (134.9) 385.0 (122.0-978.0) n ¼ 301 45.1 (16.6) 48.0 (5.0-76.0) n ¼ 635
38 (77.6) 8 (16.3) 2 (4.1) 1 (2.0) 274.5 (108.8) 261.5 (92.0-672.0) n ¼ 34 439.3 (117.6) 418.0 (315.0-614.0) n¼6 52.5 (14.4) 56.0 (15.0-72.0) n ¼ 49
130.4 (325.8) 6.0 (1.0-2510.0) n ¼ 660 89 (13.7)
129.5 (327.2) 6.0 (1.0-2510.0) n ¼ 613 80 (13.2)
142.3 (310.1) 5.0 (1.0-1513.0) n ¼ 47 9 (20.0)
P Value
.77
.20 .006 .26 1.00 <.001
<.001 .15
.58
.23 .074
.39
.002
.50
.29
212 (33.9) 272 (43.5) 141 (22.6)
192 (33.0) 257 (44.2) 133 (22.9)
20 (46.5) 15 (34.9) 8 (18.6)
.13
533 (83.5) 80 (12.5) 21 (3.3) 3 (0.5) 1 (0.2)
496 (83.4) 76 (12.8) 19 (3.2) 3 (0.5) 1 (0.2)
37 4 2 0 0
.74
(86.0) (9.3) (4.7) (0.0) (0.0)
(Continued)
1041
WALLINDER ET AL PULMONARY RETRANSPLANTATION
Table 1. Continued Variables Number of treated unverified rejections 0 1 2 3 LTx procedure Unilateral Bilateral Follow-up, y
Total (N ¼ 684)
PLTx (n ¼ 635)
461 (80.3) 107 (18.6) 5 (0.9) 1 (0.2)
427 (79.5) 104 (19.4) 5 (0.9) 1 (0.2)
282 (41.2) 402 (58.8) 5.45 (5.23) 3.94 (0.00-27.23) n ¼ 684
243 (38.3) 392 (61.7) 5.56 (5.29) 4.02 (0.00-27.23) n ¼ 635
ReLTx (n ¼ 49) 34 3 0 0
P Value
(91.9) (8.1) (0.0) (0.0)
39 (79.6) 10 (20.4) 4.09 (4.16) 2.89 (0.00-19.79) n ¼ 49
.067
<.001 .066
Categorical variables are presented as n (%). Continuous variables are presented as mean (SD) and median (range) for the number of patients. ECMO, extracorporeal membrane oxygenation; eGFR, estimated glomerular filtration rage; FEV1, forced expiratory volume in 1 second; LTx, lung transplant; PLTx, primary lung transplantation; ReLTx, repeat lung transplantation.
oxygenation, the HR was further reduced to 1.43 (95% CI, 0.90 to 2.26; P ¼ .13, Table 3). A low P value with a negative coefficient for the interaction with time from surgery was found. This indicates that the increased risk of death after retransplantation was high early after surgery and significantly reduced over the studied period. This phenomenon is graphically illustrated in Figure 3, where Kaplan-Meier estimates for death from the unadjusted time-updated model with left-censored time points for ReLTx are presented. The starting point (day 1) of the ReLTx period for patients is not counted from the ReLTx itself, as is done in Figure 2, but from PLTx, to correspond to the methodology that is used in the time-updated analysis. The figure visualizes the increased risk of early retransplantation and explains the negative interaction with time after surgery.
Table 2. Percentage of All Transplant Recipients With a Specific Diagnosis Who Later Underwent Retransplantation
Diagnosisa COPD A1AT PAH CF IPF Sarcoidosis LAM GVHD OB Bronchiectasis
ReLTx (n ¼ 49)
PLTx (n ¼ 635)
ReLTx in the Diagnosis Group
No.
No.
%
12 13 2 5 10 1 3 1 1 1
171 110 43 53 153 14 8 8 7 8
7.0 11.8 4.6 9.4 6.5 7.1 37.5 12.5 14 12.5
a
Patients with an uncertain primary diagnosis are excluded.
A1AT, a-1-antitrypsin deficiency; CF, cystic fibrosis; COPD, chronic obstructive pulmonary disease; GVHD, graft vs host disease; IPF, idiopathic pulmonary fibrosis; LAM, lymphangioleiomyomatosis; No., number; OB, obliterative bronchiolitis; PAH, pulmonary artery hypertension; PLTx, primary lung transplantation; ReLTx, repeat lung transplantation.
The 4 patients with a ReLTx after a primary heart and lung transplantation were excluded from the principal analysis. One of these patients had a double ReLTx and died 57 days postoperatively. The remaining 3 patients had a single ReLTx. One patient died after 5 years, and 2 are alive after 10 and 17 years, respectively.
Type of Retransplantation The retransplantation for 39 of the patients (80%) was done as a single LTx, including 21 procedures on the contralateral side after an initial single LTx. Double PLTx was more frequently performed in the recent era, as was double ReLTx. The adjusted HR for a single ReLTx was 1.23 (95% CI, 0.74 to 2.05) vs the PLTx, and for a double lung was 3.50 (95% CI, 1.17 to 10.50) vs the PLTx. The 1-year survival after double ReLTx was 60% (95% CI, 25% to 83%) compared with 79% (95% CI, 63% to 89%) for single ReLTx. The types of PLTx and ReLTx are found in Figure 4.
Comment In this single-center retrospective review, ReLTx was undertaken in 49 patients, 7.7% of our total experience with LTx. Survival was 76% at 1 year and 55% at 5 years. In an unadjusted Cox regression analysis, a ReLTx imposed a higher risk for mortality compared with a PLTx (HR, 1.73; P ¼ .011). After adjustments for known transplantation risk factors the risk for retransplantation was reduced (HR, 1.43; P ¼ .13). ReLTx is the only option to achieve long-term survival for patients with established dysfunction of the allograft, but survival after a ReLTx is limited even for well-selected patients in experienced centers. One-year survival ranges from 47% to 71%,4,5,9,10 and 5-year survival from 21% to 49%.5,10-12 A potential problem in studies of outcomes after ReLTx compared with PLTx is the lack of adjusted comparisons and time-updated data. Kaplan-Meier curves provide a graphical illustration of survival, but the log-rank test associated with such presentation of data is not appropriate and could cause an immortal time bias issue. The patients should not be handled as 2
GENERAL THORACIC
Ann Thorac Surg 2019;108:1037-44
GENERAL THORACIC
1042
WALLINDER ET AL PULMONARY RETRANSPLANTATION
Ann Thorac Surg 2019;108:1037-44
Figure 2. Kaplan-Meier estimates for death by first/second lung transplantation (LTx). Time 0 is defined as the primary LTx (PLTx) in the PLTx cohort and the day of the retransplantation (ReLTx) in the ReLTx group. Comparison of the groups is presented in the Cox regression model in Table 3. The shaded areas indicate the 95% confidence intervals (CIs).
independent groups, those with only PLTx vs those with PLTxþReLTx, because the PLTxþReLTx cohort is a selection of individuals who survived the PLTx. Retransplantation imposes an increased risk compared with a primary transplantation, in this study reflected by an unadjusted HR of 1.73, but survival is encouraging compared with previous reports. The explanation could be patient selection for retransplantation. Our ReLTx patients were the same age as the PLTx group but had a lower eGFR as well as a higher prevalence of diabetes than the PLTx cohort. Diabetes and impaired GFR are both common complications from long-standing immunosuppressive treatment. Finally, when adjusted for age,
Table 3. Adjusted Cox Regression Analyses for Time-Updated Risk of Repeat Lung Transplantation on Mortality
Modela
HR (95% CI) Time-Updated ReLTx Vs PLTx
Model Model Model Model Model
1.73 1.59 1.54 1.47 1.43
1 2 3 4 5
(1.14-2.63) (1.04-2.43) (1.00-2.38) (0.93-2.32) (0.90-2.26)
P Value .011 .031 .048 .10 .13
P Value for Interaction With Time .041 .005 .007 .019 .063
Neg Neg Neg Neg Neg
a Model 1: unadjusted; model 2: adjusted for sex and time-updated age; model 3: additionally adjusted for time-updated diabetes; model 4: additionally adjusted for time-updated eGFR categories (<30, 30-60, >60 mL/ min/1.73 m2 or missing); model 5: additionally adjusted for time-updated ventilator and time-updated extracorporeal membrane oxygenation. Data are adjusted in multiple steps. The risk of ReLTx is increased through the model but does not meet statistical significance in model 4 and 5. In all models a significant interaction with time is present. This is further illustrated in Figure 3, with an increased risk if ReLTx is done early after PLTx.
CI, confidence interval; HR, hazard ratio; PLTx, primary lung transplantation; ReLTx, repeat lung transplantation.
Figure 3. Survival estimates for death from unadjusted timeupdated model with left-censored time-points for repeat lung transplantation (ReLTx). At-risk and event data were obtained from the time-updated Cox proportional hazards model stratified on primary lung transplant (PLTx) or ReLTx. The starting point for PLTx in the analysis is time of surgery and the end point is death, ReLTx, or the end of follow-up. The starting point for ReLTx is left-censored at the time of the ReLTx relative to the PLTx, and the end point is death or the end of follow-up. The significant effect with time in the study is demonstrated by the larger reduction in the survival curve when ReLTx is done in the beginning of the study period compared with later than 1 year after PLTx. The shaded areas show the 95% confidence intervals. (LTx, lung transplantation.)
diabetes, and eGFR in the Cox regression analysis, the retransplantation risk was elevated but not significantly different from that of a primary transplantation. In clinical practice, the patient with diabetes or impaired kidney function could still very well be a candidate for a second transplantation but with an awareness of the increased risks. In concurrence, propensity score-matched comparisons between ReLTx and PLTx have been attempted and could not demonstrate inferior survival after ReLTx.11 The donor age was significantly older for the ReLTx group. This could reflect the thought that lungs from ideal donors should be reserved for patients with a longer life expectancy (ie, PLTxs). The largest published material on ReLTx to date from Thomas and coworkers10 reported a 1- and 5-year survival of 71.1% and 34.5%. Compared with our study, recipient age, sex, and body mass index were similar. The proportion of patients with a preoperative ventilator requirement was higher in the study by Thomas and colleagues10 (21.1% vs 7.3%), as was the frequency of ReLTx within 1 year after the PLTx (24% vs 15%). Both factors are considered to have a negative effect on survival. The use of bilateral ReLTx was 54% in their study, whereas 80% of the patients in our analysis had a single ReLTx. For most patients in our cohort, a primary single LTx was followed by a retransplantation of the contralateral lung (Figure 4). An analysis based on the United Network of Organ Sharing registry in 2014 found a weak advantage in favor
WALLINDER ET AL PULMONARY RETRANSPLANTATION
1043
numbers in the early ReLTx group hampers a robust statistical analysis, but like previous studies, suggest a restrictive use of early ReLTx. Struber and colleagues9 reported similar findings with a 1-year survival of 50% after ReLTx for acute graft failure, airway complications, and for patients with preoperative ventilatory support. Novic and colleagues4 showed a 1-year survival of 40% for ReLTx within 2 years after the PLTx. Osho and colleagues11 demonstrated approximately 40% 1-year survival for ReLTx within 90 days and Aigner and colleagues7 a 1-year survival of 35% for patients with primary graft dysfunction. Taken together, these findings support very restrictive use of ReLTx within the first year after PLTx. Figure 4. Type of primary transplantation and retransplantation. The first letter indicates the type of primary lung transplantation and the second the type of retransplantation. (D, double; L, left; R, right.)
of double-lung ReLTx over single-lung ReLTx.13 A more recently published study, based on the same registry, showed no significant difference in graft survival after retransplantation with single or double lungs when stratified by previous transplant type.14 The authors of the latter study suggest that single ReLTx should be considered, regardless of the previous transplant type, in an effort to optimize organ resources. In the present study, the adjusted HR for a single ReLTx was 1.23 (95% CI, 0.74 to 2.05) vs the PLTx. The HR for a double ReLTx, however, was 3.50 (95% CI, 1.17 to 10.50). The interaction between the 2 groups did not reach significance (P ¼ .088) but is still an interesting finding, implying that the risk of a retransplantation could be lower for single LTx. In accordance with our study, Hall and colleagues5 demonstrated a very high early mortality (55%) during the first 6 months after double-lung ReLTx and recommended single LTx over double LTx in retransplantation. Double ReLTx is a more challenging technical procedure where pleural adhesions prolong the operative time and increase the risk of major bleeding as well as technical complications. This also implies that the cold ischemic time increases for the second lung in cases of double LTx. We found a nonsignificant trend for longer cold ischemic times in the ReLTx group. In the early years of our retransplantation experience, most patients evaluated had a single LTx as their primary transplantation. A retransplantation of the contralateral lung was then preferred. In recent years, most patients have had a double LTx as the primary procedure. Based on the result presented in this study, we intend to primarily use a single LTx also for this patient group. In the present study, patients who underwent ReLTx within 1 year after the PLTx had a 1-year survival limited to 50% compared with 81% in patients who underwent ReLTx later than 1 year after the PLTx. This finding is best illustrated in Figure 3, with an unadjusted time-updated model with left-censored time point for ReLTx. The survival curves diverge when the retransplantation is done early and then follow a more parallel course. The small
Limitation Relative to previous publications on ReLTx, this study provides additional insights. However, the small number of patients limits the power and possibilities of the statistical analysis. The median follow-up time of 2.9 years (range, 0-20 years) may seem short and is restricted as a result of the high incidence of early mortality in this cohort of patients. Very limited donor data have been presented in this study due to a high proportion of missing variables from the earlier years of the transplantation program. We are convinced that recipient-related factors influence the results far more than donor-related factors and that the impact of missing donor variables on the results is limited.
Conclusion ReLTx is a reasonable option for a selected group of patients. Ideally, a number of well-established risk factors are avoided and the ReLTx is performed more than 1 year after PLTx. If so, results may even be comparable to those of PLTx. The study was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG-771911 to AW).
References 1. Chambers DC, Yusen RD, Cherikh WS, et al. The Registry of the International Society for Heart and Lung Transplantation: thirty-fourth adult lung and heart-lung transplantation report-2017; focus theme: allograft ischemic time. J Heart Lung Transplant. 2017;36:1047-1059. 2. Yusen RD, Edwards LB, Kucheryavaya AY, et al. The registry of the International Society for Heart and Lung Transplantation: thirty-first adult lung and heart-lung transplant report-2014; focus theme: retransplantation. J Heart Lung Transplant. 2014;33:1009-1024. 3. Magee JC, Barr ML, Basadonna GP, et al. Repeat organ transplantation in the United States, 1996-2005. Am J Transplant. 2007;7:1424-1433. 4. Novick RJ, Stitt LW, Al-Kattan K, et al. Pulmonary retransplantation: predictors of graft function and survival in 230 patients. Pulmonary Retransplant Registry. Ann Thorac Surg. 1998;65:227-234.
GENERAL THORACIC
Ann Thorac Surg 2019;108:1037-44
GENERAL THORACIC
1044
WALLINDER ET AL PULMONARY RETRANSPLANTATION
5. Hall DJ, Belli EV, Gregg JA, et al. Two decades of lung retransplantation: a single-center experience. Ann Thorac Surg. 2017;103:1076-1083. 6. Verleden SE, Todd JL, Sato M, et al. Impact of CLAD phenotype on survival after lung retransplantation: a multicenter study. Am J Transplant. 2015;15:2223-2230. 7. Aigner C, Jaksch P, Taghavi S, et al. Pulmonary retransplantation: is it worth the effort? A long-term analysis of 46 cases. J Heart Lung Transplant. 2008;27:60-65. 8. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley & Sons; 1980:22-36. 9. Strueber M, Fischer S, Gottlieb J, et al. Long-term outcome after pulmonary retransplantation. J Thorac Cardiovasc Surg. 2006;132:407-412. 10. Thomas M, Belli EV, Rawal B, Agnew RC, Landolfo KP. Survival after lung retransplantation in the United States in
Ann Thorac Surg 2019;108:1037-44
11.
12. 13.
14.
the current era (2004 to 2013): better or worse? Ann Thorac Surg. 2015;100:452-457. Osho AA, Castleberry AW, Snyder LD, et al. Differential outcomes with early and late repeat transplantation in the era of the lung allocation score. Ann Thorac Surg. 2014;98: 1914-1920 [discussion: 1920-1921]. Lindstedt S, Dellgren G, Iversen M, et al. Pulmonary retransplantation in the Nordic countries. Ann Thorac Surg. 2015;99:1781-1787. Hayanga JA, Yang J, Aboagye J, et al. Risk factors associated with lung retransplantation: evaluation of a nationwide registry over a quarter century. Ann Thorac Surg. 2014;98:17421746 [discussion: 1746-1747]. Schumer EM, Rice JD, Kistler AM, et al. Single versus double lung retransplantation does not affect survival based on previous transplant type. Ann Thorac Surg. 2017;103:236-240.
ABTS Announcement for Maintenance of Certification The American Board of Thoracic Surgery’s Maintenance of Certification (MOC) program was adopted 11 years ago. Since that time, the Board has continuously evaluated the overall process, based upon internal discussions and input from our Diplomates. The input resulted in our decision to migrate from a purely knowledge-based, multiple-choice exam using a Pearson Testing Center to a mastery learning process using a SESATS format. Diplomates enrolled in the 10-year MOC process will now fulfill their Part III requirement by following the instructions on the ABTS website and conveniently completing the exam at their home or office. The MOC exam is composed of 100 questions that are based on SESATS. Diplomates choose their exam module (Adult Cardiac, General Thoracic, Cardiothoracic, or Congenital) by indicating their preference within the 10-year application. The exam is tailored to one’s practice— for example, if your practice is 100% adult cardiac, you may choose the Adult Cardiac exam, which will only have adult cardiac and some critical care questions on your exam.
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
Diplomates with approved applications will be able to take the 100-question MOC exam anytime during the months of September and October 2019. For those Diplomates who have used SESATS in the past, the process of working through the questions is the same. For those who are not familiar with SESATS, it may be beneficial to purchase and download SESATS and work through the specialty-specific module. This preparation will enable you to become familiar with the process. While SESATS is a helpful resource, it is not required. The Board and MOC Committee believe that reading the critique provided after each question is key to the learning process. The goal of this exam is to provide a learning opportunity using judgment, knowledge, and decision-making skills. The Board sincerely hopes that this new MOC exam format is viewed favorably by our Diplomates. The ABTS staff thank you for embracing the primary principle of MOC—life-long learning, which is consistent with our obligation to the public trust.
Ann Thorac Surg 2019;108:1044
0003-4975/$36.00
Background. The median survival after lung retransplantation (ReLTx) reported to the International Society of Heart and Lung Transplantation is restricted to 2.5 years. We report the results after ReLTx from our center. Methods. A retrospective data collection was performed for the 635 patients who underwent lung transplantation between 1991 and 2017 at our center. Recipient variables were compared between patients undergoing only primary lung transplantation (PLTx) and those undergoing PLTx and later ReLTx. Time to death was compared using the Kaplan-Meier method. The risk of ReLTx was analyzed in Cox regression models. Any interaction between type of transplantation, single/double, and PLTx/ReLTx was investigated. Results. ReLTx was performed in 49 patients. Survival after ReLTx at 30 days and 1, 2, and 5 years was 90%, 76%, 71%, and 55%, respectively, and the corresponding survival after PLTx was 94%, 82%, 76%, and 61%,
respectively. A hazard ratio of 1.73 for ReLTx was shown (95% confidence interval [CI], 1.14 to 2.63; P [ .011). After adjustments for sex, age, diabetes, renal function, preoperative ventilator, and extracorporeal membrane oxygenation, the hazard ratio was 1.43 (95% CI, 0.90 to 2.26; P [ .13). ReLTx was performed in 8 patients (16%) within the first year after PLTx. The 1-year survival for this group was 50% compared with 81% (P [ .18) for patients who underwent ReLTx later than 1 year after the PLTx. Oneyear survival after double ReLTx was 60% (95% CI, 25% to 83%) compared with 79% (95% CI, 63% to 89%) for single ReLTx. Conclusions. ReLTx is a reasonable option for a selected group of patients. Ideally, a number of wellestablished risk factors are avoided and the ReLTx is performed more than 1 year after the PLTx.
L
annual number and proportion of lung retransplantation (ReLTx) increased during the latter part of this era, but the use of a second LTx was significantly less frequent than retransplantations for kidneys, livers, or hearts.3 The median survival after retransplantation was only 2.5 years during this time period. Retransplantation recipients were significantly younger than primary LTx (PLTx) recipients (46 vs 55 years) and more often female (46.6% vs 43.7%).2 Risk factors for poor outcome after ReLTx are the same as after PLTx, including recipient age, preoperative mechanical ventilation, and impaired renal function.4 In addition, early retransplantation and an indication for retransplantation other than bronchiolitis obliterans syndrome has been associated with poor outcome.4-7 We performed a retrospective review of our LTx program regarding the short- and long-term outcome after
ung transplantation (LTx) remains a treatment reserved for a small number of highly selected patients with end-stage lung disease. Long-term survival for the transplant population is limited to a median of 6 years, and is better for recipients of double lungs and for transplantations done after 1998.1 Many recipients are young when the transplant is performed, and hence, the question of a retransplantation may be raised if the transplanted lungs deteriorate and the patient is without major comorbidity. During the era January 1995 to June 2013, 5.1% (799 of 15,631) of single LTxs and 3.4% (925 of 27,213) of bilateral LTxs reported to the International Society of Heart and Lung Transplantation were retransplantations.2 The
(Ann Thorac Surg 2019;108:1037-44) Ó 2019 by The Society of Thoracic Surgeons
Accepted for publication Apr 8, 2019. Presented at the Thirty-ninth Annual Meeting and Scientific Sessions of The International Society for Heart and Lung Transplantation, Orlando, FL, April 3-6, 2019. Address correspondence to Dr Wallinder, Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, Blå Stråket 5, 413 45 Gothenburg, Sweden; email: [email protected].
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
The Supplemental Tables and Supplemental Figure can be viewed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2019.04.028] on http://www.annalsthoracicsurgery.org.
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2019.04.028
GENERAL THORACIC
Outcomes and Long-term Survival After Pulmonary Retransplantation: A SingleCenter Experience
GENERAL THORACIC
1038
WALLINDER ET AL PULMONARY RETRANSPLANTATION
ReLTx with the hypothesis that the results after ReLTx would be comparable to those after PLTx.
Patients and Method The University of Gothenburg Ethics Committee approved this study and waived the need for individual patient consent because only retrospective data were used in the analysis.
Patients LTx was performed in 673 patients who underwent 728 procedures at the Sahlgrenska University Hospital between 1990 and December 31, 2017. In all, 55 ReLTxs were performed. Heart-lung transplantations (n ¼ 37) were excluded from this analysis as were ReLTx procedures in the heart-lung transplantation cohort (n ¼ 4). One patient who had 2 ReLTxs was also excluded (n ¼ 3). The remaining LTx procedures (n ¼ 684) were included in the analysis (Supplemental Figure 1). The PLTx (n ¼ 635) and ReLTx (n ¼ 49) patients were monitored until April 26, 2018. Survival after PLTx and ReLTx were the primary outcomes of the analysis. Recipient variables from PLTx and ReLTx were collected from the medical records. Preoperative kidney function was categorized by estimated glomerular filtration rate (eGFR) of less than 30, 30 to 60, and more than 60 mL/min/1.73 m2. Primary graft dysfunction was categorized as 0, 1, and 2þ3 because the more severe grades 2 and 3 often vary with ventilator settings and both indicate severe graft dysfunction. No patient was lost during follow-up.
Ann Thorac Surg 2019;108:1037-44
After PLTx, each patient was monitored until death or was censored at ReLTx or at April 26, 2018; likewise, during ReLTx, the patients started being at risk at the time of the second surgery (ie, applying left censoring approach,) and were monitored until death or were censored at the end of follow-up. Stepwise adjustment was as follows: model 1 was unadjusted, model 2 included sex and time-updated age, model 3 additionally included time-updated diabetes, model 4 additionally included time-updated eGFR categories, and model 5 additionally included time-updated ventilator preoperatively and extracorporeal membrane oxygenation preoperatively. Missing values were handled as their own categories for each variable respectively in the adjustments and are described in Supplemental Table 1. The overall effects were described by hazard ratios (HRs) with 95% CIs and associated P values. The impact of the time from PLTx on the effect of ReLTx on death was investigated by including an interaction term between the 2 variables. The analysis of time-updated variable of ReLTx was graphically described using the same methodology as was used for Kaplan-Meier estimates but incorporating left-censoring of the starting point of the time period from ReLTx as a technique for performing time-updated analysis, and the 95% CIs using Greenwood’s formula for the SE calculation.8 An interaction between type of transplantation, single/ double, and PLTx/ReLTx was investigated. Missing values were not imputed. All analyses were performed using SAS 9.4 software, (SAS Institute Inc, Cary, NC). All tests were 2-tailed and conducted at the 0.05 significance level.
Statistical Analysis Recipient preoperative, intraoperative, and postoperative variables at the time of the PLTx were compared between patients undergoing only PLTx and those undergoing both PLTx and later ReLTx by using the Fisher exact test for dichotomous variables, Mantel-Haenszel c2 test for ordered categorical variables, c2 test for nonordered categorical variables, and Mann-Whitney U test for continuous variables. The same methods were also used for tests between data collected at the time of PLTx vs the time of ReLTx. The patients were independent within the 2 groups, resulting in conservative tests. Time to death from the surgery date of PLTx and ReLTx, respectively, were described in graphs by using Kaplan-Meier technique; survival probability with 95% confidence intervals (CI) is shown. The risk of ReLTx on time to death as an outcome variable was analyzed in unadjusted and adjusted Cox regression models by using the time-updated variable of ReLTx. Time-updated analysis was chosen to statistically compare the survival after a PLTx with the survival after a ReLTx. Handling the ReLTx group on a patient level including also the PLTx period in the ReLTx group would introduce a statistical bias called immortal time bias, because the future ReLTx is not known for a patient until that time point is considered in the analysis.
Results Baseline Characteristics The PLTx cohort consisted of 243 (38.3%) single LTxs with a retransplantation rate of 11% and 392 (61.7%) double LTxs with a retransplantation rate of 5.6%. Single LTx was the preference in the earlier years of the experience: 89 (65.4%) in 1991 to 2000, 126 (53.6%) in 2001 to 2010, and 28 (10.6%) in 2011 to 2017 (P < .001), which may skew data because chronic lung allograft dysfunction had time to develop in more patients in this cohort. The number of annual PLTxs and ReLTxs are presented in Figure 1.
PLTx Data The recipients who later underwent ReLTx were younger at the time of their initial transplantation than the group that did not (mean 47 vs 52 years, P < .001), had a higher eGFR (mean 100 vs 91 mL/min/1.73 m2, P ¼ .010), and more often had a single LTx (55.1% vs 36.9%, P ¼ .019). The mean donor age for the group that later had a second transplantation was lower (41 vs 45 years, P ¼ .047). The cold ischemic time for the second lung, in the event of double LTx, was significantly longer for the group that later had a ReLTx (mean 534 vs 396 minutes, P ¼ .005). Cold ischemic time data were only available from 15 of 22
1039
WALLINDER ET AL PULMONARY RETRANSPLANTATION
50
45
40
35
30
25
42
38 30
20
36
36
41
35
39
34
31 29 15
22 23
23
19
10
16 15 12
5
14
17 13
15
17 12
12
11
1 6 6 1 1 5 1 1 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 4
0 PTx
4
12
14
15
Re-DLTx Re-SLTx
13
17
11
23
15
12
12
23
17
16
1 1
19
29
1 1
1
2
1
2
22
31
1
1
2
2
30
36
35
38
41
42
1 6
2
3
6
5
34
36
39
1
1
3
2
3
Figure 1. Annual number of primary lung transplantations (PTx) and retransplantations over 27 years at a single center. (Re-DLTx, double-lung retransplantation; Re-SLTx, single-lung retransplantation.)
patients in this group. The number of treated episodes of rejection, verified by biopsy specimen or not, was not higher for PLTx patients who later had a second LTx. Recipient preoperative, intraoperative, and postoperative characteristics are presented in detail in Supplemental Table 1 and grouped for patients who later did or did not have a second LTx.
ReLTx Data Recipient characteristics are presented in Table 1. The median time from PLTx to ReLTx was 3.7 years (range, 0.06-18 years), and the median follow-up was 2.9 years (range, 0-20 years) in the ReLTx cohort vs 4.0 years (range, 0-27 years) in the PLTx cohort. ReLTx recipients had similar age as the PLTx cohort (52 vs 51 years, P ¼ .77). The ReLTx group had a lower eGFR (66 vs 91 mL/min/ 1.73 m2, P < .001) and a higher prevalence of diabetes mellitus (22% vs 9%, P ¼ .006), comorbidities that in many cases are unwanted adverse effects of the immunosuppressive treatment after PLTx. The mean donor age for the ReLTx group was older (53 vs 45 years, P ¼ .002). Primary graft dysfunction grade 2 or 3 was diagnosed in 19% after ReLTx and in 23% after PLTx (P ¼ .13). Renal replacement therapy was used in 20% in the ReLTx group and in 13% in the PLTx group (P ¼ .29). The median total ventilator time, including reintubations, was 5 hours in the ReLTx group and 6 hours in the PLTx group (P ¼ .50). The number of treated episodes of rejections, both clinical and verified by biopsy specimen, was similar between groups. Detailed data are presented in Supplemental Table 2.
Indication for the PLTx in Patients Who Later Underwent ReLTx In the group of patients who later underwent ReLTx, 11 different diagnoses were registered as the indication for the PLTx. Patients with pulmonary artery hypertension, a-1-antitrypsin deficiency, lymphangioleiomyomatosis, and cystic fibrosis were most likely to have a ReLTx, probably explained by the younger age at which patients with these diagnoses undergo PLTx. Detailed data are presented in Table 2.
Timing of ReLTx Restrictive use of ReLTx for primary graft dysfunction was implemented throughout the duration of the program. ReLTx was performed in 8 of the 49 patients (16%) within the first year after the PLTx, with a 1-year survival of 50% (4 of 8) compared with 81% (33 of 41, P ¼ .18), for patients who underwent ReLTx later than 1 year after the PLTx.
Survival Survival after ReLTx at 30 days and 1, 2, and 5 years was 90%, 76%, 71%, and 55%, respectively, and the corresponding survival after PLTx was 94%, 82%, 76%, and 61%, respectively (Figure 2). Unadjusted Cox regression analyses showed a significantly higher mortality risk of ReLTx compared with PLTx (HR, 1.73; 95% CI, 1.14 to 2.63; P ¼ .011). After adjustments for sex and timeupdated age, time-updated diabetes, and time-updated eGFR, the HR was reduced to 1.47 (95% CI, 0.93 to 2.32; P ¼ .10). After additional adjustments for preoperative need of ventilator or extracorporeal membrane
GENERAL THORACIC
Ann Thorac Surg 2019;108:1037-44
GENERAL THORACIC
1040
WALLINDER ET AL PULMONARY RETRANSPLANTATION
Table 1.
Ann Thorac Surg 2019;108:1037-44
Patient Characteristics at the Time of PLTx Vs at the Time of ReLTx
Variables Preoperative characteristics Years between first and second LTx Age at LTx, y
Male Diabetes Ventilator pre-op ECMO pre-op eGFR, mL/min/1.73 m2
eGFR categories <30 mL/min/1.73 m2 30-60 mL/min/1.73 m2 >60 mL/min/1.73 m2 FEV1%
6-min walk test, m
Intraoperative characteristics Mechanical circulatory support No Extracorporeal circulation Venovenous ECMO Venoarterial ECMO Lung 1 ischemia, min
Lung 2 ischemia, min
Donor age, y
Postoperative characteristics Time on ventilator, h
Dialysis Primary graft dysfunction Grade 0 Grade 1 Grade 2 and 3 Number of biopsy-verified rejections 0 1 2 3 4
Total (N ¼ 684)
PLTx (n ¼ 635)
ReLTx (n ¼ 49)
4.65 (3.89) 3.68 (0.06-18.04) n ¼ 49 51.3 (12.8) 54.5 (8.7-74.1) n ¼ 684 310 (45.3) 66 (9.9) 30 (4.5) 24 (3.6) 89.8 (22.3) 89.0 (12.0-180.0) n ¼ 641
51.3 (12.9) 54.6 (8.7-74.1) n ¼ 635 283 (44.6) 56 (9.1) 26 (4.2) 23 (3.7) 91.4 (21.7) 90.0 (12.0-180.0) n ¼ 601
4.65 (3.89) 3.68 (0.06-18.04) n ¼ 49 51.6 (11.1) 53.7 (19.8-66.2) n ¼ 49 27 (55.1) 10 (21.7) 4 (9.1) 1 (2.2) 65.9 (16.9) 65.0 (36.0-111.0) n ¼ 40
4 (0.6) 42 (6.6) 595 (92.8) 50.5 (24.7) 41.0 (2.0-100.0) n ¼ 625 285.3 (144.4) 270.0 (3.0-870.0) n ¼ 604
4 (0.7) 25 (4.2) 572 (95.2) 50.4 (25.2) 40.0 (2.0-100.0) n ¼ 585 286.3 (145.8) 270.0 (3.0-870.0) n ¼ 571
0 (0.0) 17 (42.5) 23 (57.5) 52.1 (16.1) 49.0 (26.0-80.0) n ¼ 40 267.8 (116.4) 246.0 (96.0-522.0) n ¼ 33
467 (68.6) 186 (27.3) 13 (1.9) 15 (2.2) 255.3 (100.0) 240.0 (83.0-869.0) n ¼ 550 403.6 (134.5) 385.0 (122.0-978.0) n ¼ 307 45.6 (16.5) 48.0 (5.0-76.0) n ¼ 684
429 (67.9) 178 (28.2) 11 (1.7) 14 (2.2) 254.0 (99.4) 240.0 (83.0-869.0) n ¼ 516 402.9 (134.9) 385.0 (122.0-978.0) n ¼ 301 45.1 (16.6) 48.0 (5.0-76.0) n ¼ 635
38 (77.6) 8 (16.3) 2 (4.1) 1 (2.0) 274.5 (108.8) 261.5 (92.0-672.0) n ¼ 34 439.3 (117.6) 418.0 (315.0-614.0) n¼6 52.5 (14.4) 56.0 (15.0-72.0) n ¼ 49
130.4 (325.8) 6.0 (1.0-2510.0) n ¼ 660 89 (13.7)
129.5 (327.2) 6.0 (1.0-2510.0) n ¼ 613 80 (13.2)
142.3 (310.1) 5.0 (1.0-1513.0) n ¼ 47 9 (20.0)
P Value
.77
.20 .006 .26 1.00 <.001
<.001 .15
.58
.23 .074
.39
.002
.50
.29
212 (33.9) 272 (43.5) 141 (22.6)
192 (33.0) 257 (44.2) 133 (22.9)
20 (46.5) 15 (34.9) 8 (18.6)
.13
533 (83.5) 80 (12.5) 21 (3.3) 3 (0.5) 1 (0.2)
496 (83.4) 76 (12.8) 19 (3.2) 3 (0.5) 1 (0.2)
37 4 2 0 0
.74
(86.0) (9.3) (4.7) (0.0) (0.0)
(Continued)
1041
WALLINDER ET AL PULMONARY RETRANSPLANTATION
Table 1. Continued Variables Number of treated unverified rejections 0 1 2 3 LTx procedure Unilateral Bilateral Follow-up, y
Total (N ¼ 684)
PLTx (n ¼ 635)
461 (80.3) 107 (18.6) 5 (0.9) 1 (0.2)
427 (79.5) 104 (19.4) 5 (0.9) 1 (0.2)
282 (41.2) 402 (58.8) 5.45 (5.23) 3.94 (0.00-27.23) n ¼ 684
243 (38.3) 392 (61.7) 5.56 (5.29) 4.02 (0.00-27.23) n ¼ 635
ReLTx (n ¼ 49) 34 3 0 0
P Value
(91.9) (8.1) (0.0) (0.0)
39 (79.6) 10 (20.4) 4.09 (4.16) 2.89 (0.00-19.79) n ¼ 49
.067
<.001 .066
Categorical variables are presented as n (%). Continuous variables are presented as mean (SD) and median (range) for the number of patients. ECMO, extracorporeal membrane oxygenation; eGFR, estimated glomerular filtration rage; FEV1, forced expiratory volume in 1 second; LTx, lung transplant; PLTx, primary lung transplantation; ReLTx, repeat lung transplantation.
oxygenation, the HR was further reduced to 1.43 (95% CI, 0.90 to 2.26; P ¼ .13, Table 3). A low P value with a negative coefficient for the interaction with time from surgery was found. This indicates that the increased risk of death after retransplantation was high early after surgery and significantly reduced over the studied period. This phenomenon is graphically illustrated in Figure 3, where Kaplan-Meier estimates for death from the unadjusted time-updated model with left-censored time points for ReLTx are presented. The starting point (day 1) of the ReLTx period for patients is not counted from the ReLTx itself, as is done in Figure 2, but from PLTx, to correspond to the methodology that is used in the time-updated analysis. The figure visualizes the increased risk of early retransplantation and explains the negative interaction with time after surgery.
Table 2. Percentage of All Transplant Recipients With a Specific Diagnosis Who Later Underwent Retransplantation
Diagnosisa COPD A1AT PAH CF IPF Sarcoidosis LAM GVHD OB Bronchiectasis
ReLTx (n ¼ 49)
PLTx (n ¼ 635)
ReLTx in the Diagnosis Group
No.
No.
%
12 13 2 5 10 1 3 1 1 1
171 110 43 53 153 14 8 8 7 8
7.0 11.8 4.6 9.4 6.5 7.1 37.5 12.5 14 12.5
a
Patients with an uncertain primary diagnosis are excluded.
A1AT, a-1-antitrypsin deficiency; CF, cystic fibrosis; COPD, chronic obstructive pulmonary disease; GVHD, graft vs host disease; IPF, idiopathic pulmonary fibrosis; LAM, lymphangioleiomyomatosis; No., number; OB, obliterative bronchiolitis; PAH, pulmonary artery hypertension; PLTx, primary lung transplantation; ReLTx, repeat lung transplantation.
The 4 patients with a ReLTx after a primary heart and lung transplantation were excluded from the principal analysis. One of these patients had a double ReLTx and died 57 days postoperatively. The remaining 3 patients had a single ReLTx. One patient died after 5 years, and 2 are alive after 10 and 17 years, respectively.
Type of Retransplantation The retransplantation for 39 of the patients (80%) was done as a single LTx, including 21 procedures on the contralateral side after an initial single LTx. Double PLTx was more frequently performed in the recent era, as was double ReLTx. The adjusted HR for a single ReLTx was 1.23 (95% CI, 0.74 to 2.05) vs the PLTx, and for a double lung was 3.50 (95% CI, 1.17 to 10.50) vs the PLTx. The 1-year survival after double ReLTx was 60% (95% CI, 25% to 83%) compared with 79% (95% CI, 63% to 89%) for single ReLTx. The types of PLTx and ReLTx are found in Figure 4.
Comment In this single-center retrospective review, ReLTx was undertaken in 49 patients, 7.7% of our total experience with LTx. Survival was 76% at 1 year and 55% at 5 years. In an unadjusted Cox regression analysis, a ReLTx imposed a higher risk for mortality compared with a PLTx (HR, 1.73; P ¼ .011). After adjustments for known transplantation risk factors the risk for retransplantation was reduced (HR, 1.43; P ¼ .13). ReLTx is the only option to achieve long-term survival for patients with established dysfunction of the allograft, but survival after a ReLTx is limited even for well-selected patients in experienced centers. One-year survival ranges from 47% to 71%,4,5,9,10 and 5-year survival from 21% to 49%.5,10-12 A potential problem in studies of outcomes after ReLTx compared with PLTx is the lack of adjusted comparisons and time-updated data. Kaplan-Meier curves provide a graphical illustration of survival, but the log-rank test associated with such presentation of data is not appropriate and could cause an immortal time bias issue. The patients should not be handled as 2
GENERAL THORACIC
Ann Thorac Surg 2019;108:1037-44
GENERAL THORACIC
1042
WALLINDER ET AL PULMONARY RETRANSPLANTATION
Ann Thorac Surg 2019;108:1037-44
Figure 2. Kaplan-Meier estimates for death by first/second lung transplantation (LTx). Time 0 is defined as the primary LTx (PLTx) in the PLTx cohort and the day of the retransplantation (ReLTx) in the ReLTx group. Comparison of the groups is presented in the Cox regression model in Table 3. The shaded areas indicate the 95% confidence intervals (CIs).
independent groups, those with only PLTx vs those with PLTxþReLTx, because the PLTxþReLTx cohort is a selection of individuals who survived the PLTx. Retransplantation imposes an increased risk compared with a primary transplantation, in this study reflected by an unadjusted HR of 1.73, but survival is encouraging compared with previous reports. The explanation could be patient selection for retransplantation. Our ReLTx patients were the same age as the PLTx group but had a lower eGFR as well as a higher prevalence of diabetes than the PLTx cohort. Diabetes and impaired GFR are both common complications from long-standing immunosuppressive treatment. Finally, when adjusted for age,
Table 3. Adjusted Cox Regression Analyses for Time-Updated Risk of Repeat Lung Transplantation on Mortality
Modela
HR (95% CI) Time-Updated ReLTx Vs PLTx
Model Model Model Model Model
1.73 1.59 1.54 1.47 1.43
1 2 3 4 5
(1.14-2.63) (1.04-2.43) (1.00-2.38) (0.93-2.32) (0.90-2.26)
P Value .011 .031 .048 .10 .13
P Value for Interaction With Time .041 .005 .007 .019 .063
Neg Neg Neg Neg Neg
a Model 1: unadjusted; model 2: adjusted for sex and time-updated age; model 3: additionally adjusted for time-updated diabetes; model 4: additionally adjusted for time-updated eGFR categories (<30, 30-60, >60 mL/ min/1.73 m2 or missing); model 5: additionally adjusted for time-updated ventilator and time-updated extracorporeal membrane oxygenation. Data are adjusted in multiple steps. The risk of ReLTx is increased through the model but does not meet statistical significance in model 4 and 5. In all models a significant interaction with time is present. This is further illustrated in Figure 3, with an increased risk if ReLTx is done early after PLTx.
CI, confidence interval; HR, hazard ratio; PLTx, primary lung transplantation; ReLTx, repeat lung transplantation.
Figure 3. Survival estimates for death from unadjusted timeupdated model with left-censored time-points for repeat lung transplantation (ReLTx). At-risk and event data were obtained from the time-updated Cox proportional hazards model stratified on primary lung transplant (PLTx) or ReLTx. The starting point for PLTx in the analysis is time of surgery and the end point is death, ReLTx, or the end of follow-up. The starting point for ReLTx is left-censored at the time of the ReLTx relative to the PLTx, and the end point is death or the end of follow-up. The significant effect with time in the study is demonstrated by the larger reduction in the survival curve when ReLTx is done in the beginning of the study period compared with later than 1 year after PLTx. The shaded areas show the 95% confidence intervals. (LTx, lung transplantation.)
diabetes, and eGFR in the Cox regression analysis, the retransplantation risk was elevated but not significantly different from that of a primary transplantation. In clinical practice, the patient with diabetes or impaired kidney function could still very well be a candidate for a second transplantation but with an awareness of the increased risks. In concurrence, propensity score-matched comparisons between ReLTx and PLTx have been attempted and could not demonstrate inferior survival after ReLTx.11 The donor age was significantly older for the ReLTx group. This could reflect the thought that lungs from ideal donors should be reserved for patients with a longer life expectancy (ie, PLTxs). The largest published material on ReLTx to date from Thomas and coworkers10 reported a 1- and 5-year survival of 71.1% and 34.5%. Compared with our study, recipient age, sex, and body mass index were similar. The proportion of patients with a preoperative ventilator requirement was higher in the study by Thomas and colleagues10 (21.1% vs 7.3%), as was the frequency of ReLTx within 1 year after the PLTx (24% vs 15%). Both factors are considered to have a negative effect on survival. The use of bilateral ReLTx was 54% in their study, whereas 80% of the patients in our analysis had a single ReLTx. For most patients in our cohort, a primary single LTx was followed by a retransplantation of the contralateral lung (Figure 4). An analysis based on the United Network of Organ Sharing registry in 2014 found a weak advantage in favor
WALLINDER ET AL PULMONARY RETRANSPLANTATION
1043
numbers in the early ReLTx group hampers a robust statistical analysis, but like previous studies, suggest a restrictive use of early ReLTx. Struber and colleagues9 reported similar findings with a 1-year survival of 50% after ReLTx for acute graft failure, airway complications, and for patients with preoperative ventilatory support. Novic and colleagues4 showed a 1-year survival of 40% for ReLTx within 2 years after the PLTx. Osho and colleagues11 demonstrated approximately 40% 1-year survival for ReLTx within 90 days and Aigner and colleagues7 a 1-year survival of 35% for patients with primary graft dysfunction. Taken together, these findings support very restrictive use of ReLTx within the first year after PLTx. Figure 4. Type of primary transplantation and retransplantation. The first letter indicates the type of primary lung transplantation and the second the type of retransplantation. (D, double; L, left; R, right.)
of double-lung ReLTx over single-lung ReLTx.13 A more recently published study, based on the same registry, showed no significant difference in graft survival after retransplantation with single or double lungs when stratified by previous transplant type.14 The authors of the latter study suggest that single ReLTx should be considered, regardless of the previous transplant type, in an effort to optimize organ resources. In the present study, the adjusted HR for a single ReLTx was 1.23 (95% CI, 0.74 to 2.05) vs the PLTx. The HR for a double ReLTx, however, was 3.50 (95% CI, 1.17 to 10.50). The interaction between the 2 groups did not reach significance (P ¼ .088) but is still an interesting finding, implying that the risk of a retransplantation could be lower for single LTx. In accordance with our study, Hall and colleagues5 demonstrated a very high early mortality (55%) during the first 6 months after double-lung ReLTx and recommended single LTx over double LTx in retransplantation. Double ReLTx is a more challenging technical procedure where pleural adhesions prolong the operative time and increase the risk of major bleeding as well as technical complications. This also implies that the cold ischemic time increases for the second lung in cases of double LTx. We found a nonsignificant trend for longer cold ischemic times in the ReLTx group. In the early years of our retransplantation experience, most patients evaluated had a single LTx as their primary transplantation. A retransplantation of the contralateral lung was then preferred. In recent years, most patients have had a double LTx as the primary procedure. Based on the result presented in this study, we intend to primarily use a single LTx also for this patient group. In the present study, patients who underwent ReLTx within 1 year after the PLTx had a 1-year survival limited to 50% compared with 81% in patients who underwent ReLTx later than 1 year after the PLTx. This finding is best illustrated in Figure 3, with an unadjusted time-updated model with left-censored time point for ReLTx. The survival curves diverge when the retransplantation is done early and then follow a more parallel course. The small
Limitation Relative to previous publications on ReLTx, this study provides additional insights. However, the small number of patients limits the power and possibilities of the statistical analysis. The median follow-up time of 2.9 years (range, 0-20 years) may seem short and is restricted as a result of the high incidence of early mortality in this cohort of patients. Very limited donor data have been presented in this study due to a high proportion of missing variables from the earlier years of the transplantation program. We are convinced that recipient-related factors influence the results far more than donor-related factors and that the impact of missing donor variables on the results is limited.
Conclusion ReLTx is a reasonable option for a selected group of patients. Ideally, a number of well-established risk factors are avoided and the ReLTx is performed more than 1 year after PLTx. If so, results may even be comparable to those of PLTx. The study was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG-771911 to AW).
References 1. Chambers DC, Yusen RD, Cherikh WS, et al. The Registry of the International Society for Heart and Lung Transplantation: thirty-fourth adult lung and heart-lung transplantation report-2017; focus theme: allograft ischemic time. J Heart Lung Transplant. 2017;36:1047-1059. 2. Yusen RD, Edwards LB, Kucheryavaya AY, et al. The registry of the International Society for Heart and Lung Transplantation: thirty-first adult lung and heart-lung transplant report-2014; focus theme: retransplantation. J Heart Lung Transplant. 2014;33:1009-1024. 3. Magee JC, Barr ML, Basadonna GP, et al. Repeat organ transplantation in the United States, 1996-2005. Am J Transplant. 2007;7:1424-1433. 4. Novick RJ, Stitt LW, Al-Kattan K, et al. Pulmonary retransplantation: predictors of graft function and survival in 230 patients. Pulmonary Retransplant Registry. Ann Thorac Surg. 1998;65:227-234.
GENERAL THORACIC
Ann Thorac Surg 2019;108:1037-44
GENERAL THORACIC
1044
WALLINDER ET AL PULMONARY RETRANSPLANTATION
5. Hall DJ, Belli EV, Gregg JA, et al. Two decades of lung retransplantation: a single-center experience. Ann Thorac Surg. 2017;103:1076-1083. 6. Verleden SE, Todd JL, Sato M, et al. Impact of CLAD phenotype on survival after lung retransplantation: a multicenter study. Am J Transplant. 2015;15:2223-2230. 7. Aigner C, Jaksch P, Taghavi S, et al. Pulmonary retransplantation: is it worth the effort? A long-term analysis of 46 cases. J Heart Lung Transplant. 2008;27:60-65. 8. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley & Sons; 1980:22-36. 9. Strueber M, Fischer S, Gottlieb J, et al. Long-term outcome after pulmonary retransplantation. J Thorac Cardiovasc Surg. 2006;132:407-412. 10. Thomas M, Belli EV, Rawal B, Agnew RC, Landolfo KP. Survival after lung retransplantation in the United States in
Ann Thorac Surg 2019;108:1037-44
11.
12. 13.
14.
the current era (2004 to 2013): better or worse? Ann Thorac Surg. 2015;100:452-457. Osho AA, Castleberry AW, Snyder LD, et al. Differential outcomes with early and late repeat transplantation in the era of the lung allocation score. Ann Thorac Surg. 2014;98: 1914-1920 [discussion: 1920-1921]. Lindstedt S, Dellgren G, Iversen M, et al. Pulmonary retransplantation in the Nordic countries. Ann Thorac Surg. 2015;99:1781-1787. Hayanga JA, Yang J, Aboagye J, et al. Risk factors associated with lung retransplantation: evaluation of a nationwide registry over a quarter century. Ann Thorac Surg. 2014;98:17421746 [discussion: 1746-1747]. Schumer EM, Rice JD, Kistler AM, et al. Single versus double lung retransplantation does not affect survival based on previous transplant type. Ann Thorac Surg. 2017;103:236-240.
ABTS Announcement for Maintenance of Certification The American Board of Thoracic Surgery’s Maintenance of Certification (MOC) program was adopted 11 years ago. Since that time, the Board has continuously evaluated the overall process, based upon internal discussions and input from our Diplomates. The input resulted in our decision to migrate from a purely knowledge-based, multiple-choice exam using a Pearson Testing Center to a mastery learning process using a SESATS format. Diplomates enrolled in the 10-year MOC process will now fulfill their Part III requirement by following the instructions on the ABTS website and conveniently completing the exam at their home or office. The MOC exam is composed of 100 questions that are based on SESATS. Diplomates choose their exam module (Adult Cardiac, General Thoracic, Cardiothoracic, or Congenital) by indicating their preference within the 10-year application. The exam is tailored to one’s practice— for example, if your practice is 100% adult cardiac, you may choose the Adult Cardiac exam, which will only have adult cardiac and some critical care questions on your exam.
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
Diplomates with approved applications will be able to take the 100-question MOC exam anytime during the months of September and October 2019. For those Diplomates who have used SESATS in the past, the process of working through the questions is the same. For those who are not familiar with SESATS, it may be beneficial to purchase and download SESATS and work through the specialty-specific module. This preparation will enable you to become familiar with the process. While SESATS is a helpful resource, it is not required. The Board and MOC Committee believe that reading the critique provided after each question is key to the learning process. The goal of this exam is to provide a learning opportunity using judgment, knowledge, and decision-making skills. The Board sincerely hopes that this new MOC exam format is viewed favorably by our Diplomates. The ABTS staff thank you for embracing the primary principle of MOC—life-long learning, which is consistent with our obligation to the public trust.
Ann Thorac Surg 2019;108:1044
0003-4975/$36.00