Published ahead of print on September 27, 2007, doi:10.1164/rccm.200707-1132OC
© 2008 American Thoracic Society doi: 10.1164/rccm.200707-1132OC
Outcomes after Lung Retransplantation in the Modern Era1 Department of Medicine, College of Physicians and Surgeons, and 2 Department of Epidemiology, Joseph L. Mailman School of Public Health, Columbia University, New York, New York; 3 Division of Thoracic Surgery, University of Toronto, Toronto, Ontario, Canada; 4 Department of Surgery, College of Physicians and Surgeons, Columbia University, New York, New York; and 5 Department of Cardiothoracic Surgery, University of Southern California, Los Angeles, California Correspondence and requests for reprints should be addressed to Steven Kawut, M.D., M.S., Division of Pulmonary, Allergy, and Critical Care Medicine, Columbia University College of Physicians and Surgeons, PH 8E, Room 101, 622 West 168th Street, New York, NY 10032. E-mail: sk2097{at}columbia.edu
Rationale: Characteristics of and survival estimates for recipients of lung retransplantation in the modern era are unknown. Objectives: To compare lung retransplant patients in the modern era with historical retransplant patients, to compare retransplant patients with initial transplant patients in the modern era, and to determine the predictors of the risk of death after lung retransplantation. Methods: We performed a retrospective cohort study of patients who underwent lung retransplantation between January 2001 and May 2006 in the United States (modern retransplant cohort). The characteristics and survival of this cohort were compared with those of patients who underwent first lung retransplantation between January 1990 and December 2000 (historical retransplant cohort) and patients who underwent initial lung transplantation between January 2001 and May 2006 (modern initial transplant cohort). Measurements and Main Results: Modern retransplant recipients (n = 205) had a lower risk of death compared with that of the historical retransplant cohort (n = 184) (hazard ratio, 0.7; 95% confidence interval, 0.5–0.9; P = 0.006). However, modern retransplant recipients had a higher risk of death than that of patients who underwent initial lung transplantation (n = 5,657) (hazard ratio, 1.3; 95% confidence interval, 1.2–1.5; P = 0.001), which appeared to be explained by a higher prevalence of certain comorbidities. Retransplantation at less than 30 days after the initial transplant procedure was associated with worse survival. Conclusions: Outcomes after lung retransplantation have improved; however, retransplantation continues to pose an increased risk of death compared with the initial transplant procedure. Retransplantation early after the initial transplant poses a particularly high mortality risk.
Key Words: lung transplantation • cohort study • retransplantation • solid organ
Lung transplantation is a therapeutic option for patients with advanced lung disease. Despite surgical and medical innovations, the outcomes after lung transplantation lag behind those of other solid organ recipients. Primary graft dysfunction and technical problems may lead to death early after transplantation, whereas acute and chronic rejection (manifest by bronchiolitis obliterans syndrome [BOS]) may contribute to later demise. There are no established medical therapies for many of these complications. Lung retransplantation may be used to treat these conditions if refractory to other measures. The recipient of lung retransplantation often suffers from the systemic sequelae of short- or long-term intensive immunosuppression, infection, and technical issues attributable to the previous transplant procedure. Not surprisingly, outcomes after retransplantation have historically been worse than after the initial transplant procedure. The largest cohort study that focused on outcomes after lung retransplantation was published almost a decade ago (1). More recently, investigators have shown that survival of certain recipient groups after retransplantation may rival that after initial transplant (2, 3). However, the small sample size and single-center design of these studies make it difficult to generalize the findings to a broader population. Given the limited supply of donor lungs, the higher risk of retransplantation is complicated by the ethical issues of granting the opportunity for two (or more) lung transplant procedures to one recipient while potentially depriving another individual of even one pulmonary allograft. The recent changes in lung allocation in the United States to a priority-based system predicated on net survival benefit may also affect the number of, or waiting time for, lung retransplantation procedures. Therefore, we aimed to compare the recipient characteristics and outcomes for modern lung retransplantation in the United States with both lung retransplantation in the previous decade and primary transplantation in the modern era. We also aimed to determine predictors of the risk of death after modern lung retransplantation. Some of these results have been previously reported in the form of an abstract (4).
Study Design and Patient Population We performed a retrospective cohort study of consecutive patients who underwent lung transplantation in the United States. The "modern retransplant cohort" included patients who were registered with the United Network for Organ Sharing (UNOS) for a first lung retransplantation procedure between January 2001 and May 2006. Patients who underwent a first lung transplant procedure during this time period were included in the "modern initial transplant cohort." Patients who underwent lung retransplantation between January 1990 and December 2000 comprised the "historical retransplant cohort." We excluded records subsequent to the first retransplant for patients with multiple retransplant procedures. That is, if a patient underwent more than one retransplant procedure, only the first retransplant procedure was included, and subsequent procedures were considered endpoints. We also excluded patients who received transplantation from a living donor, who underwent heart–lung retransplantation, or who had missing data for survival time. The initial lung transplant records of modern retransplant recipients were excluded from the modern primary transplant cohort to satisfy the assumption of independence necessary to compare the cohorts. The study was approved by the Columbia University Institutional Review Board.
Data Sources and Study Variables The primary endpoint was time until death. A patient was censored as alive at the last medical contact with UNOS. We analyzed several characteristics, including demographic data from recipients and donors, recipient body mass index, the initial pulmonary diagnosis of the recipients, indication for retransplantation, the presence of diabetes, drug-treated hypertension, renal failure (defined by an estimated glomerular filtration rate < 60 ml/min per 1.73 m2 using the Simplified Modification of Diet in Renal Disease equation) (5), corticosteroid use, procedural data (including the type of original and retransplant surgery and ischemic time for the retransplantation procedure), and blood group matching.
Statistical Analysis Survival was assessed using the Kaplan-Meier estimator and model-based estimates. We used Cox proportional hazards models to compare the risks of death between the modern retransplant cohort and both the historical retransplant cohort and the modern initial transplant cohort in bivariate analyses, expressed by hazard ratios (HRs). To determine whether certain characteristics of the cohorts might explain the differences in survival, we included in the models factors believed to be potential confounders or mediators. Change in the coefficient of transplant cohort by more than 20% after inclusion of these factors in the multivariate models was considered to indicate confounding. We compared the risks of BOS in the modern retransplant and initial transplant cohorts while accounting for death as a competing risk. To determine the predictors of survival within the modern retransplant cohort, we constructed Cox proportional hazards models including each variable of interest. Multivariate models were constructed by purposeful selection of covariates, including variables with P values less than 0.20 in the bivariate analyses as the primary risk factors of interest. Bivariate analyses were performed with available data; multivariate analyses used multiple imputation for variables with missing data (6). Sensitivity analyses were performed after imputing a distinct term for missing data. The proportional hazards assumption and model fit were assessed for the cohorts of interest using log–log plots and Schoenfeld residuals, respectively. P values less than 0.05 were considered statistically significant. Stata/IC version 10.0 for Windows (StataCorp LP, College Station, TX) and R 2.5.1 (R Foundation for Statistical Computing, Vienna, Austria) were used for all analyses.
Two hundred and fourteen patients underwent a first lung retransplantation procedure between January 2001 and May 2006. Seven of these had missing data for survival (all from 2006), and two underwent heart–lung retransplantation, leaving 205 patients in the final modern retransplant cohort. The historical retransplant cohort included 184 patients, and the modern primary transplant cohort included 5,657 patients. The characteristics of the final cohorts are shown in Table 1. The modern retransplant cohort was significantly older than the historical retransplant cohort and younger than the modern initial transplant cohort. Approximately half of the modern retransplant cohort was female, and 89% were non-Hispanic whites. Modern retransplant patients had somewhat lower body mass index than initial transplant patients. Most of the patients in all cohorts had either chronic obstructive pulmonary disease or diffuse parenchymal lung disease as the original indication for lung transplantation. These diagnoses were less frequent in the modern retransplant patients than in the initial transplant patients, and pulmonary arterial hypertension and cystic fibrosis/bronchiectasis were more frequent. Diabetes mellitus was present in 36% of the modern retransplant cohort, which was significantly more common than in the historical retransplant cohort (17%) and the modern initial transplant cohort (11%). Hypertension, renal failure, and corticosteroid use were more common in the retransplant patients than in the initial transplant patients.
Significantly more modern and historical retransplant patients than initial transplant recipients were receiving mechanical ventilation at the time of their procedures (20 and 25% vs. 3%, respectively; P < 0.001). More than half of the modern and historical retransplant procedures were performed for BOS; retransplants were frequently performed for other indications or the indication was missing. Donor and procedural data are shown in Table 2. The modern retransplant cohort received lungs from donors with characteristics similar to those of the modern initial transplant cohort; however, there were some differences compared with the historical retransplant cohort in terms of age, race/ethnicity, and mechanism of death. Cytomegalovirus (CMV) mismatch and blood group matching did not differ between the groups (data not shown).
Half of the modern retransplant patients received an initial double lung transplant or heart–lung transplantation (n = 3); most of these patients went on to receive double lung retransplantation. Modern retransplant patients more often underwent single lung retransplantation after previous double lung transplant than did historical retransplant patients; modern retransplant patients less frequently underwent ipsilateral single lung retransplantation. Mean ischemic time for modern retransplantation was longer than that for initial lung transplantation. The number of retransplantations performed each year in the modern cohort gradually increased until 2005 (the year of the introduction of the lung allocation score [LAS] priority system), when the number of retransplant procedures performed more than doubled from the previous year (Figure 1). The percentage of lung transplants that were retransplants increased from 2.8% (95% CI, 2.3–3.3%) between 2001 and 2004, to 4.5% (95% CI, 3.7–5.3%) in 2005 and 2006 (P < 0.001) (based on OPTN data as of July 13, 2007). Data collected to date from 2007 follow the trend of lung allografts increasingly being used for retransplantation (6.3%; 95% CI, 4.3–8.9%). Not surprisingly, the median wait time on the list for retransplantation was much longer for patients listed before May 2005 (the start date of the LAS priority system) than after this date (180 d [IQR, 32–569 d] vs. 25 d [IQR, 3–66 d], respectively; P < 0.001).
The modern era retransplantation cohort had a 1-year survival of 62% (95% CI, 53–70%), 3-year survival of 49% (95% CI, 38–58%), and 5-year survival of 45% (95% CI, 34–56%) (Figure 2). There were 39 deaths within 180 days of retransplantation; the most common causes of early demise were infection (n = 13), respiratory failure (n = 9), and multiorgan system failure (n = 8). Patients undergoing modern lung retransplantation had a much lower risk of death after the procedure than did those in the historical retransplant cohort (Table 3) (HR, 0.7; 95% CI, 0.5–0.9; P = 0.006). Adjustment for recipient and donor variables, pulmonary diagnosis, and mechanical ventilation at the time of transplant did not affect these results. On the other hand, patients undergoing modern lung retransplantation had a 30% higher risk of death compared with the risk of patients undergoing initial transplantation during the same period (HR, 1.3; 95% CI, 1.2–1.5; P = 0.001). Adjustment for recipient and surgical factors attenuated the results somewhat, indicating that differences in these factors explained some (but not all) of the increased risk associated with retransplantation. Further adjustment for renal failure reduced the HR even further, showing that this factor accounted for much of the increased risk in the retransplantation recipients. Figure 3 shows the model-based adjusted survival estimates of patients undergoing modern retransplantation, historical retransplantation, or initial transplantation (P < 0.05 for all comparisons). Imputation of a variable for missing data did not change the results.
When we examined patients with data for BOS status (modern retransplant cohort, n = 91, and initial transplant cohort, n = 3,809), we found that retransplant patients had a higher risk of BOS than initial transplant patients (HR, 2.0; 95% CI, 1.4–3.0; P < 0.001). The cumulative incidences of BOS at 2 years after retransplant and initial transplant were 22 and 12%, respectively. The cumulative incidences at 4 years were 46 and 30%, respectively. We then assessed the predictors of the risk of death in the modern retransplant cohort (Table 4). Recipient demographics, body mass index, and pulmonary diagnosis were not associated with outcomes. Mechanical ventilation at the time of transplantation was associated with an increased risk of death (HR, 2.0; 95% CI, 1.1–3.4; P = 0.02). The presence of diabetes mellitus, hypertension, CMV mismatch, or matched blood group was not associated with survival (data not shown). Male donor sex was associated with an increased risk of death as was early retransplantation (<30 d from the initial transplant) (HR, 2.6; 95% CI, 1.4–4.9; P = 0.003). One-year survival for the 22 patients who underwent retransplantation within 30 days of the initial transplant was only 31% (95% CI, 11–55%). One- and five-year survival was 66% (95% CI, 57–74%) and 47% (95% CI, 34–59%), respectively, for those who underwent retransplantation beyond 30 days. Initial and retransplant procedure choice was not associated with outcomes. Multivariate analysis confirmed that early retransplantation and male donor sex were independently associated with an increased risk of death (Table 5). Further adjustment for donor–recipient size mismatch or insertion of a donor–recipient sex interaction term in the model did not affect the results (data not shown).
The proportional hazards assumption was met by the variables of interest in the final models. The fit of the final models was adequate, and there were no particularly influential subjects.
Lung retransplantation is a treatment option for acute or chronic failure of the pulmonary allograft. Modern lung retransplant recipients are younger than initial transplant recipients; however, they shoulder a higher burden of medical complications due to chronic immunosuppressive medications. Despite improvements in outcome after lung retransplantation since the previous decade, the survival after lung retransplantation is still not as good as that after the initial lung transplant procedure, mostly attributable to a higher prevalence of renal failure and other comorbidities in the retransplant recipients. In addition, modern retransplant recipients appear to have a risk of BOS higher than that of initial transplant recipients. We have shown that the proportion of lung transplant procedures that are retransplantations has increased by 60% or more since the initiation of the LAS system in the United States. Although lung retransplantation still constitutes only approximately 5% of lung transplant procedures performed, the new allocation system may facilitate retransplantation by significantly shortening waiting times for the procedure. The best-known study of lung retransplantation was a prospective multicenter registry that enrolled 230 patients from 47 centers between 1985 and 1996 (1, 7–11). Several publications resulted from these data, and the final outcome assessment showed a 1-year survival of only 47% and 3-year survival of 33% (1). This final analysis showed that (1) either nonambulatory status or ventilator dependence and (2) retransplantation performed before 1992 independently increased the risk of death. These authors concluded that patients who did not meet these criteria had a "low risk" profile with survival that could equal that of patients undergoing initial transplantation. Only two small cohort studies of lung retransplantation have been published since the multicenter registry. Brugiere and colleagues (3) reviewed their experience with 15 patients who received lung retransplantation for BOS between 1988 and 2002. All were clinically stable at the time of retransplantation, although 40% required tracheostomy and assisted ventilation and six were not ambulatory. All patients underwent single lung retransplantation. The 1-year survival was 60% and 5-year survival was 45%. Infection of the retained graft (in those undergoing single lung retransplantation contralateral to the initial transplant or after double lung transplantation) was a common cause of death. Investigators from another center recently published a cohort of 54 consecutive patients who underwent lung retransplantation before January 1, 2004 (2). Survival in those patients retransplanted for "chronic" graft failure mirrored that of patients undergoing initial transplant from this center. However, patients with acute graft failure or airway problems who underwent retransplantation had a much worse survival, leading to avoidance of retransplantation for acute indications in the latter portion of the study period and, potentially, to confounding by time. Also, patients who underwent retransplantation for chronic graft failure were more likely to have cystic fibrosis (and be younger) than those who were retransplanted for acute graft failure, possibly leading to confounding by age or underlying lung disease. Our results answer many of the questions raised by these single-center studies. Retransplantation within 30 days of the initial transplant procedure poses a high risk for a poor outcome, independent of many other recipient and procedure characteristics. Although use of mechanical ventilation was associated with an increased risk of death in our study, this factor related to timing of, and therefore indication for, retransplantation. For example, retransplantation within 30 days of initial transplant is often used for primary graft dysfunction, technical airway problems, or refractory acute rejection, which also usually require mechanical ventilation. After adjusting for timing of retransplantation, however, mechanical ventilation was no longer associated with the risk of death. This finding supports the conclusion of Brugiere and colleagues that, for patients who are farther out from transplant (e.g., with BOS), mechanical ventilation may not be associated with an increased risk of death after retransplantation. Choice of retransplantation procedure does not seem related to outcomes. Indeed, retransplant in the modern cohort involved "leaving an allograft behind" more often than in the historical cohort, but with better outcomes. Of course, the decision regarding which procedure to perform is based on recipient characteristics, such as presence of superinfection; therefore, these analyses may be confounded by indication. Although a randomized clinical trial of retransplantation procedures is likely not feasible, these data suggest that the choice of retransplant procedure is not associated with outcome. Donor male sex was associated with an increased risk of death after retransplantation, independent of size mismatch, mode of death, and other characteristics that may be sex specific. Because this factor has not been documented previously, this finding should be studied further in other modern cohorts. These data also elucidate why retransplantation continues to carry a worse prognosis than the initial transplant procedure. The increased risk with retransplantation was somewhat explained by factors such as a higher prevalence of mechanical ventilation in this group; however, retransplant recipients still had a higher risk of death even after adjusting for this variable. On the other hand, adjustment for the presence of renal failure reduced the HR of retransplantation versus initial transplant almost to unity, which was not statistically significant. The presence of renal failure in retransplant recipients is likely attributable to long-standing use of calcineurin inhibitors and has been associated with poor outcomes in nonrenal solid organ transplant recipients (12). Renal failure itself is associated with hypertension, anemia, malnutrition, neuropathy, and abnormal bone metabolism, which could lead to adverse events. Alternatively, the presence of renal failure may just serve as a surrogate marker for either the duration and intensity of immunosuppression or the degree of disease of the microcirculation in other organs (e.g., the heart), leading to higher risk after retransplantation. In either case, we have documented that this significant comorbidity may potentially explain why outcomes after lung retransplantation are not as good as those after the initial transplant. Although the risk of BOS was higher in retransplant recipients, providing another explanation for worse survival, the data regarding BOS status were not sufficiently complete to make a definitive statement. The important controversy not addressed by these findings is the ethical conundrum of retransplantation in an era of limited organs (13, 14). The opposing forces are the competing goals of utilitarianism (directing organs toward recipients who will reap the greatest benefit) and egalitarianism (all of those in need should have equal access to a critical resource). Unfortunately, many of the arguments regarding this issue are imperfect and imprecise. The argument that retransplantation recipients continue to have worse outcomes than initial lung transplants, justifying abandoning retransplantation in favor of utilitarianism, is faulty. Using this same line of reasoning, initial lung transplantation would never be performed for patients with higher postoperative risk and would only be performed for patients with low risk. On the other hand, it sounds unfair to provide retransplant patients the opportunity for more than one transplant while potentially depriving another patient of even a single transplant procedure. Of course, human disease is usually unfair, striking some individuals during childhood (such as with cystic fibrosis) and resulting in potentially many years of life lost, while affecting others later in life after a childhood and long adulthood of excellent health. Does fairness justify depriving a lung recipient with a failed allograft the opportunity for another in favor of performing initial lung transplantation on an individual 40 years senior, when retransplantation could potentially save more years of life otherwise lost? Clearly, the ethical stance of retransplantation lags behind our ability to perform the procedure successfully. Several factors have increased the urgency for patients and the lung transplant community to address these issues. First, the new LAS priority system in the United States has shortened waiting times for retransplantation and increased the numbers of procedures performed and the proportion of lung allografts being used for retransplantation. The LAS system ranks potential recipients based on the weighted probabilities of survival after transplant and survival without transplant; this system may need to incorporate the medical and ethical issues specific to retransplantation. Second, outcomes for all lung transplant recipients have improved. Although survival for retransplant patients is still worse than for those receiving initial transplantation, outcomes are not dissimilar from certain diagnoses that are traditional indications for transplantation, such as pulmonary arterial hypertension. In addition, selection of patients who are greater than 1 month out from their initial transplant may minimize the risk. Considering the increased numbers of retransplantations being performed and the better outcomes, it is imperative to promptly address the issues of fairness and allocation. Our study has some limitations. Lung retransplantation is an uncommon procedure, leading to relatively small sample size and few endpoints. There were some missing data in this retrospective study, and many of the variables were not measured in a standardized fashion across centers. Data regarding center and center volume were not available. Although a multicenter source, the UNOS registry only includes data from centers in the United States, so that these conclusions may not be generalizable to other countries. Of course, the only other multicenter cohort study of retransplantation included patients transplanted more than a decade ago, likely making these current data useful for clinical decision making in the modern era.
Conclusions
The authors appreciate the assistance of Leah Bennett Edwards, Ph.D., and the late Myron Kauffman, M.D., in completing this study.
Supported in part by Health Resources and Services Administration contract 234-2005-370011C. The content of this article is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. Originally Published in Press as DOI: 10.1164/rccm.200707-1132OC on September 27, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form July 30, 2007; accepted in final form September 27, 2007
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
