Long Survival Following Lung Transplantation: What Matters?

: A retrospective review of the UNOS/OPTN Database was performed from 1 October 1987–31 December 2019. Recipients were classiﬁed as LSu (15+ years survival without GF/ReTx), normal survival (3–15 years) and short survival (<3 years). In total, 22,646 patients were identiﬁed. Groups were assessed with comparative statistics in addition to a multivariate analysis which included recipient, donor, transplant characteristics and select post-transplant complications. LSu recipients were younger, more commonly female, healthier and more commonly had cystic ﬁbrosis, pulmonary vascular disease or bilateral lung transplantation. LSu donors were younger, healthier and lacked clinical infection. Recipients with restrictive lung disease, single lung transplant and dialysis postoperatively were less likely to be LSu. Several recipient, donor and transplant characteristics are associated with long lung transplantation survival. While some factors cannot be altered, others related to donor selection and posttransplant management can potentially be inﬂuenced. Understanding these characteristics and employing discretion in donor selection, in appropriate recipients, may optimize the longevity of transplanted lungs.


Introduction
Lung transplantation is a viable treatment for patients with end-stage lung disease.Advances in donor-recipient matching, peri-operative management and immunosuppression have improved post-transplant survival [1,2].Despite favorable survival trends recently, longevity following lung transplantation compared to other solid organ transplantation remains significantly shorter.Current median survival following lung transplant approaching 6.5 years [3], compared to 12.8 years for heart transplant and 11.1 and 12.4 years following liver and kidney transplant, respectively [4,5].This difference is multifactorial, and frequently recipients suffer from primary graft dysfunction, which has deleterious impacts on short and long-term survival [6,7].An ideal donor-to-recipient allograft match can maximize organ utilization and project recipient survival.However, geography, logistical concerns and variations in risk acceptability limit the ability to perform such matching [8].While ideal donor criteria have been defined, these are often too strict as extended criteria donors can expand the donor pool and promote organ stewardship [9].
As such, our primary objective was to identify characteristics associated with prolonged lung transplant survival without graft failure/re-transplantation to allow providers to better identify donor organs for recipients and provide prognostic information for transplant candidates.We hypothesized that a combination of recipient, donor and transplant characteristics would be independently associated with long-term survival and that understanding these factors will better inform providers when counseling recipients and selecting suitable donor allografts.

Materials and Methods
A retrospective review of the Organ Procurement and Transplantation Network (OPTN) database administered by the United Network for Organ Sharing (UNOS) was performed from 1 October 1987 to 31 December 2021.The UNOS/OPTN database provides recipient, donor and transplant information as well as long-term survival data.The study was exempt from institutional review board approval (IRB: #2018H0079).
Recipients were categorized into three groups: long, normal and short survivors.Those with 15+ years survival (>90th percentile for survival) without graft failure or retransplantation were considered long survivors (LSu).Recipients with survival ranging from 3 to 15 years were classified as normal survivors (NSu), while those with survival <3 years were short survivors (SSu).Inclusion criteria included all lung transplant recipients ≥ 18 years old with known survival data.Multi-organ transplants (liver, kidney, heart) or redo-lung transplantation were excluded.Additionally, patients alive without graft failure or re-transplantation but with less than 15 years of survival were excluded from analysis as their grouping would not be clear and would bias the results (e.g., a recipient alive without retransplant with 10 years of follow-up would be excluded; Figure 1).

Results
After querying the UNOS database, 22,646 lung transplants were identified from 1 October 1987 to 31 December 2021.There were 1163 (5.13%) in the LSu group, 10,164 (44.88%)NSu and 11,319 (49.98%) in the SSu group.LSu recipients were more often younger (47 years, Inter-Quartile Range [IQR]: 36-54; p < 0.01) (Figure 2A), less often male (47.89%, p < 0.01) and had significantly lower BMI (22.36 kg/m 2 , IQR: 19.47-25.83kg/m 2 ).LSu recipients were less often smokers (59.26%, p < 0.01) or diabetics (7.96%, p < 0.01), and had the highest glomerular filtration rate (GFR) (99.07 mL/min/1.73m 2 , IQR: 82.26, 111.28 mL/min/1.73m 2 ).LSu recipients had the longest days on the waitlist, often required the least amount of pre-operative ventilator or ECMO use and were most likely to not be hospitalized (p < 0.01).Cystic fibrosis (CF)/pulmonary vascular disease (PVD) recipients were more commonly LSu, relative to NSu and SSu (p < 0.01).Additional recipient characteristics can be seen in Table 1.Continuous data was assessed for normality and parametric data, and are presented as mean ± standard deviation, while non-parametric data are presented as median (interquartile range).Missingness was calculated and missing data was assumed to be at random.Data was compared across groups for univariate analysis using an analysis of variance (parametric) or a Kruskal-Wallis (non-parametric) test for continuous variables and Chi-square/Fisher's test, as appropriate, for categorical variables.To identify variables associated with LSu a hierarchical, multivariable logistic regression model was created and recipients were classified as long survivors and non-long survivors (NSu and SSu).Recipient variables in the model were: age (modeled categorically due to nonlinearity), gender, race, body mass index (BMI), diabetes, glomerular filtration rate (GFR), diagnosis, blood type, preoperative ventilator/extracorporeal membrane oxygenation (ECMO) use, days on the waitlist and hospitalization status.Donor variables included: age, gender, race, BMI, classification of increased risk by the Centers for Disease Control (CDC), coronary artery disease (CAD), criteria cigarette history (>20 pack years [PY]), diabetes, hypertension, cause of death, presence of clinical infection, chest X-ray (CXR) results (normal vs. abnormal vs. none) and bronchoscopy results (normal vs. abnormal vs. none).Transplant characteristics were: gender mismatch, lung transplant type (single vs. bilateral), the difference in recipient-to-donor age, whether the donor was younger than the recipient, distance traveled in nautical miles (NM), ischemic time (hours), length of stay (LOS) (days), post-operative dialysis, post-operative stroke, post-operative airway dehiscence, presence of acute rejection before discharge, treated rejection in the first year, transplant era and cause of death.Variables were chosen a priori based on perceived clinical significance.Additionally, since many donor and transplant variables were only collected beginning in April 1994, only recipients of the initial cohort between 1 April 1994 and 31 December 2021 were included.SAS v. 9.4 (Cary, NC, USA) was used for all analyses and a p-value < 0.05 was considered significant.

Discussion
While multiple studies have focused on factors associated with survival, few have specifically examined factors which are associated with long survival in excess of 15 years [3,10,11].With the median survival of lung transplant recipients approaching seven years, recipients are living longer than ever before [8].However, particularly for young patients or those without additional significant comorbidities, this length of time can still be seen as discouraging.In order to better serve our patients, as well as promote recipient longevity, our study identified several factors associated with long-term survival.
Donor selection is crucial to setting up recipients for success.Providers are often quite selective with donors for recipients with expected long survival, to ideally provide them with a better long-term result (i.e., younger donor, high PF, no history of smoking, normal bronchoscopy).However, as more extended criteria donors are being used to increase the donor pool, prototypical donors can be hard to come by.Selection can ultimately be difficult and relies largely on the surgeon's experience or macroscopic appearance in the operating room [12,13].In order to offer some objectivity to this process, our unadjusted analysis revealed that donors who were younger, had fewer medical co-morbidities, more often died from neurologic causes and had normal CXR attributed to long-term surviving recipients.Following adjustment, donor age, diabetes, BMI and presence of a clinical infection were associated with decreased long-term survival.Some of these risk factors for improved longevity have been described before [14,15], and our study adds to this body of evidence in a selection of donors.Radiologic and bronchoscopy evidence of infection are associated with increased mortality following lung transplantation [16].Interestingly our study showed that a complete lack of bronchoscopy was significantly associated with long-term survival.Although it is difficult to discern the exact reasons, one may be that

Discussion
While multiple studies have focused on factors associated with survival, few have specifically examined factors which are associated with long survival in excess of 15 years [3,10,11].With the median survival of lung transplant recipients approaching seven years, recipients are living longer than ever before [8].However, particularly for young patients or those without additional significant comorbidities, this length of time can still be seen as discouraging.In order to better serve our patients, as well as promote recipient longevity, our study identified several factors associated with long-term survival.
Donor selection is crucial to setting up recipients for success.Providers are often quite selective with donors for recipients with expected long survival, to ideally provide them with a better long-term result (i.e., younger donor, high PF, no history of smoking, normal bronchoscopy).However, as more extended criteria donors are being used to increase the donor pool, prototypical donors can be hard to come by.Selection can ultimately be difficult and relies largely on the surgeon's experience or macroscopic appearance in the operating room [12,13].In order to offer some objectivity to this process, our unadjusted analysis revealed that donors who were younger, had fewer medical co-morbidities, more often died from neurologic causes and had normal CXR attributed to long-term surviving recipients.Following adjustment, donor age, diabetes, BMI and presence of a clinical infection were associated with decreased long-term survival.Some of these risk factors for improved longevity have been described before [14,15], and our study adds to this body of evidence in a selection of donors.Radiologic and bronchoscopy evidence of infection are associated with increased mortality following lung transplantation [16].Interestingly our study showed that a complete lack of bronchoscopy was significantly associated with long-term survival.Although it is difficult to discern the exact reasons, one may be that these donors are otherwise considered excellent (e.g., young, excellent PF ratio, etc.) and therefore felt to not need a bronchoscopy.
Our analysis also demonstrates that traditional factors in donor selection may be less important when considering potential donor allografts, such as smoking history or donor death from trauma, especially as donor cigarette use was in fact associated with increased long-term survival in our adjusted model.Data regarding donor smoking has been conflicted with some studies stating that significant smoking history is not associated with long-term survival [17][18][19], while other studies have suggested a negative impact on survival and increased risk of chronic lung allograft dysfunction (CLAD) [14].A recent meta-analysis however concluded that donor smoking, either ever or with >20 PY history, was not associated with greater 1-year mortality [20].It is important to consider that the smoking stratification in this analysis is based on >20 PY, whereas some previous studies have classified smoking history by pack years or any smoking history.One possible explanation is that donors with significant smoking history may have been excellent donors in all other respects, which ultimately led to their use (i.e., high PF ratio, no imaging or bronchoscopy abnormalities, etc.).Thus, smoking history alone should not be a major impediment to acceptance of lungs, if the lungs otherwise appear excellent.
In the management of recipients, some factors that are linked to decreased long-term survival might be in fact modifiable, such as diabetes.To a degree underlying kidney disease is also modifiable, and attention should be placed on preserving renal function.Associated risk factors with worsening renal function, such as hypertension or diabetes, can also be accounted for and treated as well.Importantly, it should be acknowledged that most of the factors identified that were associated with decreased long-term survival are non-modifiable (recipient age, diagnosis, hospitalization status or blood type).In these situations, these findings serve as a prognostic tool for providers to advise recipients about their potential for long survival.In the selection of operation, bilateral lung transplantation was associated with long-term survival in both unadjusted and adjusted analyses.Although the decision between single versus bilateral lung transplant is a complex topic with many patient-specific nuances [21], our study does confirm recent studies from the International Society for Heart and Lung Transplantation [8].
Post-operatively, a strong predictor of long-term survival was the lack of post-transplant dialysis, which was significant in both unadjusted and adjusted analyses.While it is unknown if dialysis was temporary or permanent, it is clear that dialysis post-transplant significantly limits the opportunity to achieve long-term survival.A recent pooled analysis revealed that approximately 50% of patients suffer from acute kidney injury (AKI) after lung transplantation, and approximately 10% of these patients develop an AKI that requires dialysis [22].Several risk factors for the development of AKI are bilateral transplantation, pulmonary vascular disease, mechanical ventilation, mechanical-circulatory support, and hemodynamic instability [23].AKI requiring dialysis also results in significant short-and long-term mortality [23,24].Intraoperative adjustments to help preserve renal function including minimization of periods of hypotension and avoiding cardiopulmonary bypass (or minimizing its time of use [25]) can be useful to mitigate renal failure.Additionally, post-operatively, providers can focus on weaning off vasoactive medications [26] and accepting potentially prolonged ventilation by promoting adequate volume resuscitation with judicious fluid administration [27] to avoid renal failure.Notably, although postoperative stroke and airway dehiscence were associated with decreased survival in our unadjusted model, this did not hold true for the adjusted model.
While this study was illuminating in several aspects, questions remain unanswered.Our data lacks information regarding glycemic control, thus indicating a potential avenue for further research as improved glycemic control may improve outcomes.Additionally, the literature suggests that connections may exist between frailty and the major complications of lung transplantation, which can impact survival [28].Due to data limitations within the UNOS/OPTN database, we were unable to examine the impact of frailty on long-term survival.Finally, in our unadjusted analysis, certain blood groups were associated with long-term survival.It is unclear why a particular blood group would impact the long-term, however, given the complex nature of transplantation surgery, there might be a complex interplay between donor and recipient immune systems and complexes.Thus, further investigation into potential immunologic factors driving this finding is warranted.
While we highlight several key concepts to incorporate into lung transplantation management, it is important to recognize our analysis is not meant to discourage the selection of recipients or donors at risk for only short-term success.Due to the comparatively short longevity of lung allografts, numerous recipients require re-listing for transplantation sooner than expected or do not survive due to graft failure.Moreover, not every recipient needs a 15+ year survival, particularly as older individuals (i.e., age > 70 years), are being transplanted with greater frequency and success [29].Additionally, limitations exist within the lung allocation score (LAS), which was derived from a formula designed to optimize allocation while weighing the risk of waitlist mortality and the potential for one-year posttransplant survival [3,30].For example, a 65-year-old recipient with pulmonary fibrosis with diabetes and moderate CAD will likely have a higher LAS than a 35-year-old recipient with cystic fibrosis and no other significant comorbidities.The older recipient would be more likely to receive offers from younger or "more ideal" donors.However, those "more ideal" donors may be better suited for the younger recipient.Meanwhile, donors that are considered extended criteria may be rejected for younger recipients and ultimately not used, though these may be appropriate for older recipients.Recent data demonstrate acceptable outcomes with such donors [31].With a better understanding of which donors and recipients are associated with longer survival, providers will have greater objectivity to aid in donor and recipient matching which may help increase donor utilization and improve recipient outcomes.Providing organ procurement organizations with more latitude to work with accepting centers may help to increase utilization and matching.In order to improve allograft placement, the upcoming change in allocation will use a composite allocation score (CAS) for continuous lung distribution, incorporate expected five-year survival and focus on allocation based on biological profile and placement efficiency [32].These changes are an important step to help with overall lung utilization; their impact on donor-recipient matching remains to be seen.
This study has inherent limitations that affect any large database, including lack of granular data, its retrospective nature and is subject to information and selection bias.Due to our definition of long survival, recipients are limited to those transplanted before 2007 and ultimately do not account for recent management and surgical advancements.Furthermore, a recipient doing well at 10 years without graft failure/death is undifferentiated with respect to the defined cohorts and was excluded-which may introduce further selection bias into the results presented.Additionally, primary graft dysfunction is a known contributor to short-and long-term morbidity and mortality [6,7], and was unaccounted for in this analysis.Finally, limitations of data limited our ability to comment on post-transplant events such as rejection and hospitalization, limiting inferences about how these events impact the potential for long survival as well as impact on quality of life.

Conclusions
Several characteristics correlate with long-term lung transplantation survival.While some of these factors cannot be altered, such as recipient age and diagnosis, others related to donor selection and transplant procedure can potentially be influenced.Factors such as age, donor clinical infection and post-transplant dialysis are negatively associated with long-term survival and should be considered if long survival is expected (i.e., young and otherwise 'healthy' recipients).Understanding these characteristics and employing discretion in donor selection and post-transplant management may optimize graft longevity, particularly in younger, clinically stable patients.Our hope is providers will be able to use these results to guide difficult decisions, to not only help transplant candidates but also to increase organ utilization.

Institutional Review Board Statement:
The study was exempt from institutional review board approval (IRB: #2018H0079).
Informed Consent Statement: Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Transplantology 2023, 4 ,Figure 1 .
Figure 1.Consort diagram of patients included for analysis based on inclusion and exclusion criteria.

Figure 1 .
Figure 1.Consort diagram of patients included for analysis based on inclusion and exclusion criteria.

Figure 2 .
Figure 2. (A) Percentage of each survivor group stratified by recipient age.(B) Percentage of each survivor group stratified by donor age.(C) Percentage of each survivor group stratified by transplant type.(D) Percentage of each survivor group stratified by ischemic time.

Figure 2 .
Figure 2. (A) Percentage of each survivor group stratified by recipient age.(B) Percentage of each survivor group stratified by donor age.(C) Percentage of each survivor group stratified by transplant type.(D) Percentage of each survivor group stratified by ischemic time.

Figure 3 .
Figure 3. Forest plot demonstrating odds ratio and 95% confidence interval of variables and their association with long survival.

Figure 3 .
Figure 3. Forest plot demonstrating odds ratio and 95% confidence interval of variables and their association with long survival.

Funding:
This work was supported in part by the Health Resources and Services Administration contract HHSH250-2019-00001C.
Data presented as mean +/− standard deviation or median (interquartile range) for continuous variables and number (%) for categorical variables.BMI, Body Mass Index; ECMO, Extracorporeal Membrane Oxygenation; GFR, Glomerular Filtration Rate; ICU, Intensive Care Unit; PRA, Panel Reactive Antigen.
Data presented as mean +/− standard deviation or median (interquartile range) for continuous variables and number (%) for categorical variables.BMI, Body Mass Index; CDC, Centers for Disease Control; CVA, Cardiovascular Accident; CXR, Chest X-ray; GSW, Gunshot Wound.
Data presented as mean +/− standard deviation or median (interquartile range) for continuous variables and number (%) for categorical variables.