Abstract
Extracorporeal membrane oxygenation (ECMO) as a bridge to lung transplant (BTT) has been used for critically ill candidates with excellent outcomes, but data on this strategy in older recipients remain limited. We compared outcomes of no BTT, mechanical ventilation (MV)-only BTT, and ECMO BTT in recipients ≥65 years. Lung-only recipients ≥65 years in the United Network for Organ Sharing database between 2008–2022 were included and stratified by bridging strategy. Of the 9,936 transplants included, 226 (2.3%) were MV-only BTT and 159 (1.6%) were ECMO BTT. ECMO BTT recipients were more likely to have restrictive disease pathology, had higher median lung allocation score, and spent fewer days on the waitlist (all p<0.001). Compared to no-BTT recipients, ECMO BTT recipients were more likely to be intubated or on ECMO at 72 hours post-transplant and had longer hospital lengths of stay (all p<0.001). ECMO BTT recipients had increased risk of 3-year mortality compared to both no-BTT (aHR 1.48 [95% CI: 1.14–1.91], p=0.003) and MV-only recipients (aHR 1.50 [95% CI: 1.08–2.07], p=0.02). Overall, we found that ECMO BTT in older recipients is associated with inferior post-transplant outcomes compared to MV-only or no BTT, but over half of recipients remained alive at 3 years post-transplant.
Keywords: lung transplant, extracorporeal membrane oxygenation, bridging strategy, outcomes
INTRODUCTION
Extracorporeal membrane oxygenation (ECMO) has increasingly been used as a bridge to lung transplantation (BTT) in recent years.1–3 This shift in clinical practice has been facilitated by significant improvements in technology as well as improved understanding of optimal management for patients on ECMO.1,4–8 While bridging with ECMO has been associated in some series with complications including longer hospital stay9,10 and increased usage of blood products,10 long-term survival has been comparable to recipients without pre-transplant support in several studies.1 The outcomes of ECMO BTT have also improved significantly over time, with outcomes of ECMO BTT shown to be equivalent to those of mechanical ventilation bridge to lung transplantation (MV BTT).2,3,11 However, while ECMO as BTT is becoming widely accepted, there remains significant variation between centers on criteria for patient selection, including cutoffs for advanced age.1
In a multi-center survey conducted by Tsiouris and colleagues in 2014, 45% (15/33) of responding institutions reported age >65 as an absolute contraindication for ECMO BTT.12 However, 36% (12/33) of centers, including all six high-volume respondents, had no maximum age restriction. The prevalence of age restrictions and selection bias towards younger patients are likely due to concerns that older recipients might have inferior outcomes after BTT, but there is a paucity of data to support such restrictions.13 Only two studies to our knowledge examined this relationship. Hayanga et al. found significantly increased mortality in recipients older than 35,3 while George et al. found that older patients showed decreased 1-year survival when stratified by age, determining that the optimal cutoff for 1-year survival was 62 years.14 However, both studies only included lung transplants prior to 2011, whereas the shift towards increasing utilization of ECMO BTT did not occur until around 2014.11
Given the aging lung transplant population,15–18 we sought to further explore the association of recipient age to ECMO BTT in this study by assessing post-transplant outcomes in patients ≥65 years who received no BTT, MV-only BTT, and ECMO BTT.
MATERIALS AND METHODS
Study Population
Lung transplant recipients ≥65 years old who were transplanted between 2008 and 2022 in the United Network for Organ Sharing (UNOS) Organ Procurement and Transplantation Network (OPTN) database were included in the study. Multi-organ transplants and re-transplants were excluded. Recipients were stratified into three groups based on BTT strategy: no BTT, MV-only BTT, and ECMO BTT. This study was approved by the Johns Hopkins Institutional Review Board (IRB00352819).
Center analysis
Characteristics of centers that performed ECMO BTT transplants were compared by whether they had performed at least one ECMO BTT transplant in a recipient ≥65 years using Wilcoxon’s rank sum tests. The waitlist size of each transplant center was defined as the median of the numbers of candidates on the waitlist on January 1st of each year during the study period (2008 to 2022).
Waitlist outcomes
For the waitlist analysis, candidates ≥65 years of age at listing who were bridged with ECMO between 2008 and 2022 with information available on ECMO cannulation date (80%) were included. The percent of candidates who were transplanted or experienced death/deterioration within 1 year of ECMO cannulation were determined. The number of waitlist days spent on ECMO support was compared between those who were transplanted and those who died/deteriorated using Wilcoxon rank-sum tests. Cause of death on the waitlist was categorized as infection, cardiovascular, pulmonary, cerebrovascular, hemorrhage, or other.
Baseline recipient, donor, and transplant characteristics
Baseline recipient, donor, and transplant characteristics of the unmatched cohorts were assessed. Normality of all variables in this study were evaluated using Shapiro-Wilk testing and histogram visualization. Parametric continuous variables were compared using one-way ANOVA and reported as mean (standard deviation). Nonparametric continuous variables were compared using Kruskal Wallis testing and reported as median (interquartile range). Categorical variables were assessed using Chi-squared testing and reported as number (percentage). Trends were assessed using Cuzick tests.
Perioperative outcomes
Predischarge acute rejection, intubation at 72 hours, ECMO support at 72 hours, post-operative airway dehiscence, post-operative dialysis, and post-operative stroke were compared between the unmatched cohorts using Chi-squared testing and multivariable logistic regression. Hospital length of stay was compared using Kruskal Wallis tests. All multivariable models adjusted for center volume, transplant era (2008–2012, 2013–2017, 2018–2022), and baseline recipient, donor, and transplant characteristics with p<0.2 on univariate analysis.
Post-transplant survival
Time-to-event analysis was performed on unmatched cohorts and outcomes were visualized using Kaplan-Meier curves. Univariate and multivariable Cox regression on unmatched cohorts were used to compare post-transplant survival at 1 year and 3 years post-transplant by bridging strategy. All multivariable models were adjusted for center volume, transplant era (2008–2012, 2013–2017, 2018–2022), and baseline recipient, donor, and transplant characteristics with p<0.2 on univariate analysis. Of the baseline characteristics potentially included in the multivariable models, recipient diagnosis, recipient diabetes, recipient prior cardiac surgery, donor cause of death, donor >20 pack-year smoking history, and ischemic time had low level of missingness (<2.5%); recipients missing these data were excluded from our models. The remainder of the variables had 0% missingness. Goodness-of-fit was evaluated using Gronnesby and Borgan tests, with models meeting a threshold p-value of >0.05. Recipients were followed to outcome of interest or administrative censorship on December 31, 2022.
Conditional survival
Time-to-event analysis was performed on unmatched cohorts to evaluate survival at 3 years post-transplant conditional on survival at 1 year post-transplant. For this analysis, the risk set population was restricted only to recipients who survived the first year following transplant, as previously described.19 Univariate and multivariable Cox regressions were used to compare survival by bridging strategy as described above.
Subgroup analysis of ECMO BTT recipients
A subgroup analysis of ventilated vs. non-ventilated ECMO BTT recipients was performed. Normality of all variables in this study were evaluated using Shapiro-Wilk testing and histogram visualization. Parametric continuous variables were compared using student’s t-tests and reported as mean (standard deviation). Nonparametric continuous variables were compared using Wilcoxon’s rank sum tests and reported as median (interquartile range). Categorical variables were assessed using Chi-squared testing (or Fisher’s exact testing for n<5) and reported as number (percentage). Predischarge acute rejection, intubation at 72 hours, ECMO support at 72 hours, post-operative airway dehiscence, post-operative dialysis, and post-operative stroke were compared using Chi-squared testing. Length of stay was assessed using Wilcoxon’s rank sum tests. Post-transplant survival at 1 and 3 years was analyzed using univariate and multivariable Cox regressions as described above.
Propensity-matched analysis
To further control for differences in baseline characteristics, ECMO BTT transplants were propensity-matched 1:1 to MV-only BTT transplants using nearest neighbor with no replacement and a caliper of 0.2 of the standard deviation of the logit of the propensity score.20 Transplants were matched on recipient age, ethnicity, diagnosis, and lung allocation score; donor age, sex, cause of death, >20 pack-year smoking history, and donation after circulatory death; ischemic time, and procedure type (bilateral vs. single). Balance of covariates after propensity matching was assessed and defined as a standardized mean difference of less than 0.1 between ECMO BTT transplants and MV-only BTT transplants.
Standardized mean difference of baseline characteristics were assessed. Intubation at 72 hours, ECMO at 72 hours, pre-discharge acute rejection, post-operative airway dehiscence, post-operative dialysis, and post-operative strokewere compared using McNemar tests. Hospital lengths of stay were compared using paired t-tests. Post-transplant survival at 1 and 3 years were compared using univariate Cox regression. All statistics were performed using StataSE 18 (StataCorp, College Station, Texas).
RESULTS
Of the total 9,936 recipients included in our study, 9,551 (96.1%) were no BTT, 226 (2.3%) were MV-only BTT, and 159 (1.6%) were ECMO BTT. Prior to 2018, utilization of MV-only BTT was more common than ECMO BTT in recipients ≥65 years of age (Figure 1). From 2018–2022, ECMO BTT was more commonly utilized than MV-only BTT every year.
Figure 1:
(A) Bridging strategy utilized as a percentage of total lung transplants in recipients ≥65 years. (B) Number of centers performing transplants in recipients ≥65 years by bridging strategy.
Center analysis
Of the 80 centers that performed at least one lung transplant during the study period, 69 (86.3%) transplanted at least one ECMO BTT recipient of any age and 45 (56.3%) transplanted at least one ECMO BTT recipient ≥65 years of age. Centers that performed at least one ECMO BTT transplant in a recipient ≥65 years, compared those that performed ECMO but not in recipients ≥65 years, performed more ECMO BTT transplants overall (median 19 [IQR: 13–39] vs. 5.5 [1.5–12] transplants during study period, p<0.001; Table 1) and began performing ECMO BTT earlier (median year of first ECMO BTT transplant 2013 [2010–2016] vs. 2015 [2013–2019], p=0.002). Of centers that performed ECMO BTT transplants in recipients ≥65 years of age, the median time between the first ECMO BTT transplant and the first ECMO BTT transplant in a recipient ≥65 years of age was 2 years. Centers performing ECMO BTT in recipients ≥65 years also had greater median waitlist size (10 [6–20] vs. 4 [0–9.5] candidates, p=0.002) with similar time spent on the waitlist (60 [43–69] vs. 63 [43–97] days, p=0.22).
Table 1:
Characteristics of centers performing extracorporeal membrane oxygenation (ECMO) bridge-to-transplant (BTT) from 2008 to 2022.
| Variables, median (IQR) | Performed no ECMO BTT in recipients ≥65 years | Performed at least one ECMO BTT in recipients ≥65 years | p-value |
|---|---|---|---|
| N=24 | N=45 | ||
|
| |||
| Number of ECMO BTT transplants during study period | 5.5 (1.5-12) | 19 (13-39) | <0.001 |
| Year of first ECMO BTT transplant | 2015 (2013-2019) | 2013 (2010-2016) | 0.002 |
| Waitlist size | 4 (0-9.5) | 10 (6-20) | 0.002 |
| Recipient age | 60 (55-61) | 61 (60-62) | 0.004 |
| Waitlist days | 63 (43-97) | 60 (43-69) | 0.22 |
Abbreviations: IQR, interquartile range; ECMO, extracorporeal membrane oxygenation; BTT, bridge to transplant.
Waitlist outcomes
Of candidates ≥65 years with available ECMO cannulation dates (n=229, 79%), 122 (53.3%) were transplanted within 1 year of cannulation, 98 (42.8%) died/deteriorated within 1 year of cannulation, and 9 (3.9%) were removed from the waitlist for other reasons. No cannulated candidates were removed from the waitlist due to improvement in condition, and no cannulated candidate remained on the waitlist beyond a year. Those who deteriorated/died spent more waitlist days on ECMO support compared to candidates who were transplanted (11 [5–17] vs. 4 [2–12] days, p<0.001). For candidates with data on waitlist causes of death, the most common cause of death was pulmonary (n=30, 48%), followed by cardiovascular (n=10, 16%), other (n=9, 15%), infection (n=7, 11%), hemorrhage (n=5, 8%), and cerebrovascular (n=1, 2%).
Recipient characteristics
Compared to MV-only BTT and no BTT recipients, ECMO BTT recipients were less likely to be of white race (86.4% no BTT vs. 86.3% MV-only vs. 76.7% ECMO, p=0.02; Table 2), more likely to have restrictive disease pathology (68.2% no BTT vs. 69.9% MV-only vs. 93.1% ECMO, p<0.001), had higher median lung allocation score at transplant (39.8 [34.7–49.0] no BTT vs. 75.3 [41.4–88.1] MV-only vs. 86.2 [84.5–88.9] ECMO, p<0.001), and were more likely to have had prior cardiac surgery (2.1% no BTT vs. 3.5% MV-only vs. 7.5% ECMO, p<0.001). ECMO BTT recipients also spent fewer days on the waitlist (42 [13–130] no BTT vs. 18 [6–69] MV-only vs. 13 [5–39] ECMO, p<0.001).
Table 2:
Baseline recipient, donor, and transplant characteristics of recipients ≥65 years old who received a lung transplant from 2008 to 2022, by bridge-to-transplant (BTT) strategy.
| Variable, N (%) | No BTT | MV-only BTT | ECMO BTT | p-value |
|---|---|---|---|---|
| N=9,551 | N=226 | N=159 | ||
|
| ||||
| Recipient characteristics | ||||
| Age (years), median (IQR) | 68 (66-70) | 67 (66-70) | 67 (66-69) | <0.001 |
| Male sex | 6,504 (68.1%) | 159 (70.4%) | 116 (73.0%) | 0.34 |
| Race/ethnicity | 0.02 | |||
| White | 8,252 (86.4%) | 195 (86.3%) | 122 (76.7%) | |
| Black | 436 (4.6%) | 7 (3.1%) | 12 (7.5%) | |
| Hispanic | 576 (6.0%) | 16 (7.1%) | 19 (11.9%) | |
| Other | 287 (3.0%) | 8 (3.5%) | 6 (3.8%) | |
| Diagnosis | <0.001 | |||
| Obstructive disease | 2,610 (27.3%) | 59 (26.1%) | 6 (3.8%) | |
| Pulmonary vascular disease | 207 (2.2%) | 4 (1.8%) | 3 (1.9%) | |
| Cystic fibrosis | 15 (0.2%) | 0 (0.0%) | 0 (0.0%) | |
| Restrictive disease | 6,511 (68.2%) | 158 (69.9%) | 148 (93.1%) | |
| Other | 208 (2.2%) | 5 (2.2%) | 2 (1.3%) | |
| Disease severity | ||||
| Pre-transplant mechanical ventilation | 0 (0.0%) | 226 (100.0%) | 82 (51.6%) | <0.001 |
| LAS at transplant, median (IQR) | 39.8 (34.7-49.0) | 75.3 (41.4-88.1) | 86.2 (84.5-88.9) | <0.001 |
| Comorbidities | ||||
| BMI (kg/m2), median (IQR) | 26.3 (23.3-29.0) | 25.9 (23.1-29.0) | 27.0 (24.2-30.1) | 0.01 |
| Diabetes | 1,571 (16.4%) | 35 (15.5%) | 31 (19.5%) | 0.54 |
| Prior cardiac surgery | 199 (2.1%) | 8 (3.5%) | 12 (7.5%) | <0.001 |
| Total waitlist days, median (IQR) | 42 (13-130) | 18 (6-69) | 13 (5-39) | <0.001 |
| Follow-up time (years), median (IQR) | 2.4 (1.0-4.9) | 3.0 (1.1-5.9) | 1.5 (0.4-3.3) | <0.001 |
| Donor characteristics | ||||
| Age (years), median (IQR) | 34 (24-47) | 36 (26-48) | 38 (26-49) | 0.15 |
| Male sex | 5,997 (62.8%) | 130 (57.5%) | 91 (57.2%) | 0.10 |
| White race | 5,802 (60.7%) | 143 (63.3%) | 99 (62.3%) | 0.69 |
| Cause of death | 0.02 | |||
| Anoxia | 2,725 (28.5%) | 46 (20.4%) | 41 (25.8%) | |
| Cerebrovascular/stroke | 2,762 (28.9%) | 82 (36.3%) | 54 (34.0%) | |
| Head trauma | 3,818 (40.0%) | 90 (39.8%) | 58 (36.5%) | |
| CNS tumor | 45 (0.5%) | 2 (0.9%) | 3 (1.9%) | |
| Unknown | 201 (2.1%) | 6 (2.7%) | 3 (1.9%) | |
| DCD | 403 (4.2%) | 6 (2.7%) | 11 (6.9%) | 0.12 |
| >20 pack-year smoking history | 771 (8.1%) | 32 (14.2%) | 16 (10.1%) | 0.003 |
| Transplant characteristics | ||||
| Ischemic time (hours), median (IQR) | 4.9 (3.9-6.1) | 5.3 (4.3-6.2) | 5.4 (4.4-6.3) | <0.001 |
| Bilateral (vs single) | 5,006 (52.4%) | 165 (73.0%) | 126 (79.2%) | <0.001 |
Abbreviations: BTT, bridge-to-transplant; MV, mechanical ventilation; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; LAS, lung allocation score; BMI, body mass index; CNS, central nervous system; DCD, donation after circulatory death
Donor and transplant characteristics
Donors between the three groups were similar in terms of age (p=0.15), sex (p=0.10), and ethnicity (p=0.69; Table 2). ECMO BTT donors were more likely than no bridge donors and less likely than MV-only donors to have >20 pack-year smoking history (8.1% no BTT vs. 14.2% MV-only vs. 10.1% ECMO, p<0.001). ECMO BTT transplants had longer ischemic times (4.9 [3.9–6.1] hours for no BTT vs. 5.3 [4.3–6.2] MV-only vs. 5.4 [4.4–6.3] ECMO, p<0.001) and were more likely to be bilateral (52.4% no-BTT vs. 73.0% MV-only vs. 79.2% ECMO, p<0.001).
Perioperative outcomes
Compared to no-BTT recipients, MV-only BTT recipients were more likely to be intubated at 72 hours post-transplant (67.4% vs. 23.4%; aOR 4.60 [95% CI: 2.86–7.38], p<0.001; Table 3) and had longer median hospital lengths of stay (27 [18–24] vs. 16 [11–27] days; p<0.001), but had similar likelihood of ECMO at 72 hours post-transplant (8.7% vs. 4.9%; aOR 1.06 [95% CI: 0.49–2.30], p=0.88), pre-discharge acute rejection (10.2% vs. 6.4%; aOR 1.33 [0.84–2.11], p=0.22), post-operative airway dehiscence (2.7% vs. 1.4%; aOR 1.15 [0.45–2.96]; p=0.78), post-operative dialysis (11.5% vs. 5.7%; aOR 1.47 [0.94–2.32], p=0.09), and post-operative stroke (4.1% vs. 2.6%; aOR 1.26 [0.62–2.57]; p=0.52).
Table 3:
Adjusted post-transplant outcomes of transplants in recipients ≥65 years of age from 2008 to 2022, by bridging strategy.
| Variables, n (%) | Group | Number (%) | Adjusted OR (95% CI) | Adjusted p-value |
|---|---|---|---|---|
|
| ||||
| Intubated at 72 hours | No BTT | 1,551 (23.4%) | Ref | Ref |
| MV-only BTT | 62 (67.4%) | 4.60 (2.86-7.38) | <0.001 | |
| ECMO BTT | 87 (69.0%) | 3.49 (2.29-5.33) | <0.001 | |
|
| ||||
| On ECMO at 72 hours | No BTT | 324 (4.9%) | Ref | Ref |
| MV-only BTT | 8 (8.7%) | 1.06 (0.49-2.30) | 0.88 | |
| ECMO BTT | 41 (32.5%) | 3.79 (2.37-6.06) | <0.001 | |
|
| ||||
| Post-transplant length of stay (days), median (IQR)* | No BTT | 16 (11-27) | Ref | Ref |
| MV-only BTT | 27 (18-49) | n/a | <0.001 | |
| ECMO BTT | 31.5 (19-64) | n/a | <0.001 | |
|
| ||||
| Pre-discharge acute rejection | No BTT | 614 (6.4%) | Ref | Ref |
| MV-only BTT | 23 (10.2%) | 1.33 (0.84-2.11) | 0.22 | |
| ECMO BTT | 6 (3.8%) | 0.45 (0.19-1.04) | 0.06 | |
|
| ||||
| Post-operative airway dehiscence | No BTT | 135 (1.4%) | Ref | Ref |
| MV-only BTT | 6 (2.7%) | 1.15 (0.45-2.96) | 0.78 | |
| ECMO BTT | 7 (4.4%) | 1.95 (0.81-4.68) | 0.13 | |
|
| ||||
| Post-operative dialysis | No BTT | 536 (5.7%) | Ref | Ref |
| MV-only BTT | 26 (11.5%) | 1.47 (0.94-2.32) | 0.09 | |
| ECMO BTT | 37 (23.4%) | 2.41 (1.56-3.71) | <0.001 | |
|
| ||||
| Post-operative stroke | No BTT | 239 (2.6%) | Ref | Ref |
| MV-only BTT | 9 (4.1%) | 1.26 (0.62-2.57) | 0.52 | |
| ECMO BTT | 8 (5.1%) | 1.42 (0.65-3.12) | 0.38 | |
Unadjusted Kruskal Wallis with pairwise comparison performed for post-transplant length of stay.
Models adjusted for center volume, transplant era (2008-2012, 2013-2017, 2018-2022), and baseline characteristics with p<0.2 on univariate analysis.
Abbreviations: OR, odds ratio; CI, confidence interval; MV, mechanical ventilation; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range.
Compared to no-BTT recipients, ECMO BTT recipients were more likely to be intubated (69.0% vs. 23.4%; aOR 3.49 [95% CI: 2.29–5.33], p<0.001) or on ECMO (32.5% vs. 4.9%; aOR 3.79 [95% CI: 2.37–6.06], p<0.001) at 72 hours post-transplant and had longer median hospital lengths of stay (31.5 [19–64] vs. 16 [11–27] days; p<0.001). ECMO BTT recipients also had greater likelihood of requirement for post-operative dialysis compared to no-BTT recipients (23.4% vs. 5.7%; aOR 2.41 [95% CI: 1.56–3.71]; p<0.001). ECMO BTT and no-BTT recipients had similar likelihood of pre-discharge acute rejection (3.8% vs. 6.4%; aOR 0.45 [95% CI: 0.19–1.04], p=0.06), post-operative airway dehiscence (4.4% vs. 1.4%; aOR 1.95 [95% CI: 0.81–4.68]; p=0.13), and post-operative stroke (5.1% vs. 2.6%; aOR 1.42 [95% CI: 0.65–3.12]; p=0.38).
Post-transplant survival
Post-transplant survival at 1 year was 85.9% for no-BTT, 82.1% for MV-only BTT, and 68.4% for ECMO BTT recipients. On univariate analysis, compared to no-BTT recipients, MV-only BTT recipients had similar risk of mortality at 1 year (HR 1.30 [95% CI: 0.95–1.79], p=0.11), while ECMO BTT recipients had increased risk of mortality at 1 year (HR 2.62 [95% CI: 1.96–3.49], p<0.001). Compared to MV-only BTT recipients, ECMO BTT recipients had increased risk of mortality at 1 year (HR 2.01 [95% CI: 1.32–3.07], p=0.001). These inferences regarding 1-year mortality risk remained unchanged on adjusted analysis (MV-only BTT vs. no-BTT aHR 0.98 [95% CI: 0.70–1.37], p=0.91; ECMO BTT vs. no-BTT aHR 2.09 [1.52–2.88], p<0.001; ECMO BTT vs. MV-only BTT aHR 2.13 [95% CI: 1.38–3.30], p=0.001; Supplemental Table 1).
Post-transplant survival at 3 years was 66.2% for no BTT, 60.8% for MV-only BTT, and 50.9% for ECMO BTT. Compared to no BTT, MV-only BTT had similar risk (HR 1.21 [95% CI: 0.97–1.51], p=0.09) and ECMO BTT had increased risk (HR 1.91 [95% CI: 1.50–2.42], p<0.001) of mortality at 3 years on univariate analysis (Figure 2). Compared to MV-only BTT, ECMO BTT had increased risk of mortality at 3 years [HR 1.58 [95% CI: 1.15–2.17], p=0.005). These findings regarding 3-year mortality were unchanged on unadjusted analysis (MV-only BTT vs. no BTT aHR 0.99 [95% CI: 0.78–1.25], p=0.92; ECMO BTT vs. no BTT aHR 1.48 [95% CI: 1.14–1.91], p=0.003; ECMO BTT vs. MV-only BTT aHR 1.50 [95% CI: 1.08–2.07], p=0.02; Supplemental Table 2).
Figure 2:
Unadjusted 3-year survival of lung transplant recipients ≥65 years of age from 2008 to 2022 in the United Network for Organ Sharing registry, by bridging strategy. 95% confidence intervals (CI) shown. Hazard ratio is adjusted and represents extracorporeal membrane oxygenation (ECMO) bridge-to-transplant (BTT) versus mechanical ventilation (MV)-only BTT recipients.
Conditional survival
For recipients who survived the first year after transplant, conditional survival at 3 years post-transplant was 77.1% for no BTT, 74.0% for MV-only BTT, and 73.7% for ECMO BTT. Compared to no BTT, MV-only BTT (HR 1.13 [95% CI: 0.84–1.53], p=0.42) and ECMO BTT had similar risk (HR 1.25 [95% CI: 0.83–1.89], p=0.29) of mortality at 3 years on univariate analysis (Figure 3). ECMO BTT also had similar risk of mortality as MV-only BTT on univariate analysis (HR 1.11 [95% CI: 0.67–1.84], p=0.70). On multivariable analysis, these findings were unchanged (MV-only BTT vs. no BTT aHR 1.00 [95% CI: 0.73–1.38], p=0.99; ECMO BTT vs. no BTT aHR 0.96 [95% CI: 0.62–1.49], p=0.99; ECMO BTT vs. MV-only BTT aHR 0.96 [95% CI: 0.58–1.60], p=0.88).
Figure 3:
Survival at 3 years post-transplant conditional on survival to 1 year post-transplant in lung transplant recipients ≥65 years of age, by bridging strategy. 95% confidence intervals (CI) shown. Hazard ratio is adjusted and represents extracorporeal membrane oxygenation (ECMO) bridge-to-transplant (BTT) versus mechanical ventilation (MV)-only BTT recipients.
Subgroup analysis of ECMO BTT recipients
On subgroup analysis of ventilated and non-ventilated ECMO recipients, baseline donor, recipient, and transplant characteristics were similar (Supplemental Table 3). Non-ventilated ECMO recipients had decreased likelihood of being intubated at 72 hours (55% vs. 83%, p<0.001), but similar likelihood of being on ECMO at 72 hours (35% vs. 30%, p=0.49), pre-discharge acute rejection (4% vs. 4%, p>0.99), post-operative airway dehiscence (4% vs. 5%, p>0.99), post-operative dialysis (17% vs. 29%, p=0.07), and post-operative stroke (4% vs. 6%, p=0.72). Non-ventilated vs. ventilated ECMO recipients had shorter hospital lengths of stay (26 [17–58] vs. 38 [24–68] days, p=0.03). At 1 year post-transplant, non-ventilated vs. ventilated ECMO recipients had similar unadjusted (73.1% vs. 63.4%; HR 0.70 [95% CI: 0.39–1.24], p=0.22) and adjusted (aHR 0.79 [95% CI: 0.44–1.46], p=0.46) survival. Non-ventilated vs. ventilated ECMO recipients also had similar unadjusted (51.8% vs. 50.0%; HR 0.86 [95% CI: 0.54–1.38], p=0.54) and adjusted (aHR 0.90 [95% CI: 0.55–1.47], p=0.67) survival at 3 years post-transplant.
Propensity-matched analysis
In total, 121 ECMO BTT transplants were propensity-matched to 121 MV-only transplants. Baseline characteristics of the propensity-matched cohorts are shown in Table 4. ECMO BTT recipients had higher likelihood of ECMO at 72 hours post-transplant (33%% vs. 11%, p<0.001; Table 5), post-operative airway dehiscence (5.0% vs. 3.4%, p<0.001), and post-operative dialysis (24.2% vs. 14.9%, p<0.001, but lower likelihood of pre-discharge acute rejection (3.3% vs. 6.6%, p<0.001) and post-operative stroke (5.1% vs. 6.7%, p<0.001). ECMO BTT and MV-only BTT recipients had similar likelihood of intubation at 72 hours (71% vs. 72%, p=0.09). ECMO BTT compared to MV-only BTT recipients had longer hospital lengths of stays (32.5.5 [19–64] vs. 27 [18–50] days, p=0.04). In propensity-matched cohorts, ECMO BTT recipients had 121% higher risk of mortality by 1 year (HR 2.21 [95% CI: 1.27–3.84], p=0.005) and 78% higher risk of mortality by 3 years (HR 1.78 [95% CI: 1.18–2.68], p=0.006) post-transplant compared to MV-only recipients.
Table 4:
Baseline recipient, donor, and transplant characteristics of propensity-matched groups.
| Variable, n (%) | MV-only BTT | ECMO BTT | Standardized mean difference |
|---|---|---|---|
| N=121 | N=121 | ||
|
| |||
| Recipient characteristics | |||
| Age (years), median (SD) | 68 (3) | 68 (2) | 0.046 |
| Male sex | 88 (72.7%) | 87 (71.9%) | 0.018 |
| White race | 99 (81.8%) | 100 (82.6%) | 0.022 |
| Restrictive disease | 112 (92.6%) | 111 (91.7%) | 0.031 |
| LAS at transplant | 83.2 (12.9) | 82.0 (12.8) | 0.090 |
| Donor characteristics | |||
| Age (years) | 36 (14) | 37 (15) | 0.043 |
| Male sex | 65 (53.7%) | 62 (51.2%) | 0.050 |
| White race | 79 (65.3%) | 79 (65.3%) | <0.001 |
| Cause of death | |||
| Anoxia | 29 (24.0%) | 29 (24.0%) | <0.001 |
| Cerebrovascular/stroke | 39 (32.2%) | 43 (35.5%) | 0.070 |
| Head trauma | 48 (39.7%) | 45 (37.2%) | 0.051 |
| Other | 5 (4.1%) | 4 (3.3%) | <0.001 |
| DCD | 4 (3.3%) | 4 (3.3%) | <0.001 |
| >20 pack year history | 11 (9.1%) | 12 (9.9%) | 0.028 |
| Transplant characteristics | |||
| Ischemic time (hours) | 5.1 (1.5) | 5.2 (1.3) | 0.017 |
| Bilateral (vs single) | 91 (75.2%) | 92 (76.0%) | 0.019 |
Abbreviations: BTT, bridge-to-transplant; MV, mechanical ventilation; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range; LAS, lung allocation score; BMI, body mass index; CNS, central nervous system; DCD, donation after circulatory death.
Table 5:
Post-transplant outcomes of propensity-matched groups.
| Variables, n (%) | MV-only BTT | ECMO BTT | p-value | |
|---|---|---|---|---|
| N=121 | N=121 | |||
|
| ||||
| 1-year survival | 83.6% | 67.8% | 0.005 | |
| 3-year survival | 64.3% | 48.7% | 0.006 | |
| Intubated at 72 hours | 41 (72%) | 67 (71%) | 0.09 | |
| On ECMO at 72 hours | 6 (11%) | 31 (33%) | <0.001 | |
| Hospital length of stay (days), median (IQR) | 27 (18-50) | 32.5 (19-64) | 0.04 | |
| Pre-discharge acute rejection | 8 (6.6%) | 4 (3.3%) | <0.001 | |
| Post-operative airway dehiscence | 3 (3.4%) | 6 (5.0%) | <0.001 | |
| Post-operative dialysis | 18 (14.9%) | 29 (24.2%) | <0.001 | |
| Post-operative stroke | 8 (6.7%) | 6 (5.1%) | <0.001 | |
Abbreviations: MV, mechanical ventilation; BTT, bridge-to-transplant; ECMO, extracorporeal membrane oxygenation; IQR, interquartile range.
DISCUSSION
In this national retrospective cohort study, we analyzed transplant center trends, waitlist outcomes, and post-transplant survival in ECMO BTT recipients ≥65 years of age. The majority of centers performing ECMO BTT performed at least one ECMO BTT transplant in a recipient ≥65 years, and centers that performed ECMO BTT in recipients ≥65 years displayed significant center-level differences, including larger waitlists and a longer, more robust history of ECMO BTT compared to centers that did not utilize ECMO BTT in older recipients. Our waitlist analysis showed that more than half of candidates placed on ECMO received a transplant within one year of cannulation. Finally, post-transplant survival in ECMO BTT recipients ≥65 years remained inferior to MV-only recipients, with an increased risk of 1- and 3- year mortality. However, for recipients who survive the first year after transplant, ECMO BTT is not associated with additional 3 year mortality risk.
A survey conducted by Tsiouris et al. in 2014 showed that 15 of 33 (45%) transplant centers implemented a hard age cutoff of 65 years for ECMO BTT.12 In our study, we demonstrated that 45 of 69 (65.2%) centers in our study performed at least one ECMO BTT transplant in a recipient ≥65 years, suggesting that some centers might have expanded their eligibility critieria over the past decade. The Tsiouris et al. study was limited to centers that responded to the survey (33 out of 57); however, potential selection bias toward respondents from larger transplant centers with ECMO experience in their study likely overestimated the practice of ECMO BTT in recipients ≥65 years old nationwide.12 Our study demonstrates that since 2014, more centers have reconsidered age as a contraindication to ECMO BTT as experience with ECMO BTT has grown and outcomes have improved and been shown to be equivalent to MV-only BTT.2,3,11,12,21 However, there still remain significant center-level differences. Centers that performed ECMO BTT in recipients ≥65 years of age performed more ECMO BTT transplants overall, had larger median waitlist size, and adopted ECMO BTT earlier compared to those who performed ECMO BTT in only younger recipients. This corresponds with the findings of Tsiouris et al., who described that every high-volume center (>50 transplants per year) that responded to the survey performed ECMO BTT on recipients ≥65 years old. Although centers nationwide appear to be increasingly willing to perform ECMO BTT on patients ≥65 years old, this practice has been driven by larger transplant centers with greater ECMO experience.
We showed that 122 (53.7%) candidates ≥65 years started on ECMO received a transplant, while 98 (43.2%) died or deteriorated. This proportion of successful bridging to transplantation was lower than that reported recently by Deitz et al., who presented a cohort of 634 adult candidates supported on ECMO at the time of listing between 2016 and 2021. Of these candidates, 445 (70%) were successfully bridged to transplant.22 In Deitz and colleagues’ ECMO BTT cohort, candidates who died or deteriorated on the waitlist were significantly older (54 vs. 47 years, p<0.001) than those successfully transplanted, and advanced age was identified as an independent risk factor of increased waitlist mortality.22 The median age of candidates in our waitlist analysis was 67 years. However, in addition to candidate age, a number of other factors may also have contributed to the higher waitlist mortality observed in our cohort. First, while our study captured transplants from 2008 to 2022, Deitz et al. included a more recent study period (2016–2021), when experience and outcomes in ECMO BTT were more well-established.22 Additionally, Deitz et al. included candidates who were listed with ECMO, while our study additionally included candidates who were started on ECMO while already being listed. It is possible that candidates who were cannulated while already listed represent candidates who decompensated and were more critically ill. Similar to the results of Deitz et al., we demonstrated that candidates who died or deteriorated spent significantly more time on the waitlist supported by ECMO than those with successful ECMO BTT (11 vs. 4 days, p<0.001). This suggests that extended duration of ECMO on the waitlist may be associated with increased waitlist mortality.
We found that post-transplant outcomes in recipients ≥65 years for ECMO BTT were significantly worse than those with MV-only BTT or no BTT. ECMO BTT recipients had greater post-transplant mortality at 1 and 3 years compared to MV-only BTT recipients. These results held true even on our propensity-matched analysis. This contrasts with the findings of Hayanga et al., who recently demonstrated equivalent post-transplant outcomes ECMO BTT and MV-only BTT in adult lung transplant recipients overall.11 In the analysis by Hayanga et al., the median age at transplant for ECMO BTT recipients was 48.5 years, which was younger than both MV-only BTT recipients (median age 53 years) and no BTT (median age 59 years). The findings of Hayanga et al. support the use of ECMO BTT in the overall lung transplant candidate population. Our findings of inferior outcomes of ECMO BTT vs. MV-only BTT in older candidates, however, suggests that providers should carefully select and counsel suitable older candidates before employing this bridge to transplant technique.
Importantly, however, through our conditional survival analysis, we found that the increased risk of mortality among ECMO BTT recipients could largely be attributed to increased risk in the immediate post-transplant period. For recipients who survived the first year following transplant, ECMO BTT was not associated with increased risk for mortality at three years. This suggests that the inferior outcomes we observed for ECMO vs. MV-only BTT in older recipients could be due to increased rates of perioperative complications, as well as increased debilitation from the greater sedation and paralysis utilized for ECMO. Recipients who survive to one year have likely completed rehabilitation and reconditioning, and thus might no longer be at greater risk for mid-term post-transplant mortality. Additionally, though limited by size, our subgroup analysis of non-ventilated vs. ventilated ECMO candidates found improved perioperative outcomes, including hospital length of stay and prolonged intubation, as well as a trend towards improved 1-year survival in non-ventilated ECMO candidates (73.1% vs. 63.4%), suggesting that utilizing a non-ventilated ECMO strategy can help prevent deconditioning and muscle atrophy associated with prolonged mechanical ventilation.23 Further studies should be performed to determine whether utilizing ECMO without mechanical ventilation can improve short-term post-transplant outcomes in this population.
There are several limitations to this study, due in part to the inherent methodologic constraints of retrospective studies and the lack of granularity in the data available using national databases. First, the small sample size of ECMO BTT recipients ≥65 years old might have limited our ability to detect differences in outcomes by use of BTT strategy. Second, we were unable to assess certain waitlist complications associated with ECMO use (e.g. bleeding). Third, while we were able to capture ECMO cannulation dates for 80% of candidates with ECMO BTT, we did not have mechanical ventilation start dates for the majority of candidates. This limited our ability to compare waitlist outcomes by bridging strategy. Additionally, ECMO is a more recently adopted bridging strategy in recipients ≥65 years. Our study may therefore be subject to chronology bias. We have attempted to control for this by adjusting for transplant era in our analysis (2008–2012, 2013–2017, 2018–2022). Finally, since the majority of ECMO cannulations occurred at high-volume centers, these results might not be fully generalizable to other, smaller centers with limited ECMO experience.
Overall, we found that the majority of lung transplant centers in the US have performed ECMO BTT transplants in recipients ≥65 years of age, particularly at centers with larger waitlists and a longer history of institutional ECMO experience. More than half of listed candidates ≥65 years on ECMO were successfully bridged to transplant. Among ECMO BTT recipients ≥65 years of age, post-transplant survival was 68.4% at 1 year and 50.9% at 3 years. While ECMO BTT has inferior outcomes to other bridging strategies among older candidates, over two thirds of transplanted candidates survived to 1 year post-transplant, while all of those who remained on the waitlist died or deteriorated within 1 year of cannulation. In receipients ≥65 years, ECMO BTT recipients who survive to one year post-transplant have no excess three year mortality compared to those bridged with other strategies. Moving forward, future research should focus on candidate selection and perioperative management in order to improve immediate post-operative morbidity and mortality in this population.
Supplementary Material
Sources of funding:
F32AG067642 (PI: Ruck) from the National Institute on Aging
Footnotes
Conflicts of interest: None
DISCLOSURES
None.
IRB: This study has been approved by the Johns Hopkins Institutional Review Board (IRB00352819).
Meeting presentation: Accepted for oral presentation at ASAIO 2023 conference.
REFERENCES
- 1.Patterson CM, Shah A, Rabin J, et al. Extracorporeal life support as a bridge to lung transplantation: Where are we now? J Heart Lung Transplant. 2022;41(11):1547–1555. [DOI] [PubMed] [Google Scholar]
- 2.Benazzo A, Schwarz S, Frommlet F, et al. Twenty-year experience with extracorporeal life support as bridge to lung transplantation. J Thorac Cardiovasc Surg. 2019;157(6):2515–2525.e10. [DOI] [PubMed] [Google Scholar]
- 3.Hayanga AJ, Aboagye J, Esper S, et al. Extracorporeal membrane oxygenation as a bridge to lung transplantation in the United States: An evolving strategy in the management of rapidly advancing pulmonary disease. J Thorac Cardiovasc Surg. 2015;149(1):291–296. [DOI] [PubMed] [Google Scholar]
- 4.Diaz-Guzman E, Hoopes CW, Zwischenberger JB. The Evolution of Extracorporeal Life Support as a Bridge to Lung Transplantation. ASAIO J. 2013;59(1):3. [DOI] [PubMed] [Google Scholar]
- 5.Daneshmand MA, Milano CA. ECMO: 40 Years Later. ASAIO J. 2017;63(6):693. [DOI] [PubMed] [Google Scholar]
- 6.Stokes JW, Gannon WD, Bacchetta M. Extracorporeal Membrane Oxygenation as a Bridge to Lung Transplant. Semin Respir Crit Care Med. 2021;42(3):380–391. [DOI] [PubMed] [Google Scholar]
- 7.Jang EJ, Jung SY, Lee HJ, et al. Trends in Extracorporeal Membrane Oxygenation Application and Outcomes in Korea. ASAIO J Am Soc Artif Intern Organs 1992. 2021;67(2):177–184. [DOI] [PubMed] [Google Scholar]
- 8.Biscotti M, Sonett J, Bacchetta M. ECMO as bridge to lung transplant. Thorac Surg Clin. 2015;25(1):17–25. [DOI] [PubMed] [Google Scholar]
- 9.Toyoda Y, Bhama JK, Shigemura N, et al. Efficacy of extracorporeal membrane oxygenation as a bridge to lung transplantation. J Thorac Cardiovasc Surg. 2013;145(4):1065–1071. [DOI] [PubMed] [Google Scholar]
- 10.Ius F, Natanov R, Salman J, et al. Extracorporeal membrane oxygenation as a bridge to lung transplantation may not impact overall mortality risk after transplantation: results from a 7-year single-centre experience†. Eur J Cardiothorac Surg. 2018;54(2):334–340. [DOI] [PubMed] [Google Scholar]
- 11.Hayanga JWA, Hayanga HK, Holmes SD, et al. Mechanical ventilation and extracorporeal membrane oxygenation as a bridge to lung transplantation: Closing the gap. J Heart Lung Transplant. 2019;38(10):1104–1111. [DOI] [PubMed] [Google Scholar]
- 12.Tsiouris A, Budev MM, Yun JJ. Extracorporeal Membrane Oxygenation as a Bridge to Lung Transplantation in the United States: A Multicenter Survey. ASAIO J. 2018;64(5):689. [DOI] [PubMed] [Google Scholar]
- 13.Nasir BS, Klapper J, Hartwig M. Lung Transplant from ECMO: Current Results and Predictors of Post-transplant Mortality. Curr Transplant Rep. 2021;8(2):140–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.George TJ, Beaty CA, Kilic A, Shah PD, Merlo CA, Shah AS. Outcomes and temporal trends among high-risk patients after lung transplantation in the United States. J Heart Lung Transplant. 2012;31(11):1182–1191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Valapour M, Lehr CJ, Skeans MA, et al. OPTN/SRTR 2017 Annual Data Report: Lung. Am J Transplant. 2019;19(S2):404–484. [DOI] [PubMed] [Google Scholar]
- 16.Valapour M, Lehr CJ, Schladt DP, et al. OPTN/SRTR 2021 Annual Data Report: Lung. Am J Transplant. 2023;23(2, Supplement 1):S379–S442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Arjuna A, Olson MT, Walia R. Current trends in candidate selection, contraindications, and indications for lung transplantation. J Thorac Dis. 2021;13(11). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Zhou AL, Karius AK, Ruck JM, et al. Outcomes of Lung Transplant Candidates Aged ≥70 Years During the Lung Allocation Score Era. Ann Thorac Surg. Published online June 3, 2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Suarez-Pierre A, Lui C, Zhou X, et al. Conditional Survival in Heart Transplantation: An Organ Procurement and Transplantation Network Database Analysis. Ann Thorac Surg. 2020;110(4):1339–1347. [DOI] [PubMed] [Google Scholar]
- 20.Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10(2):150–161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Tipograf Y, Salna M, Minko E, et al. Outcomes of Extracorporeal Membrane Oxygenation as a Bridge to Lung Transplantation. Ann Thorac Surg. 2019;107(5):1456–1463. [DOI] [PubMed] [Google Scholar]
- 22.Deitz RL, Emerel L, Chan EG, et al. Waitlist Mortality and Extracorporeal Membrane Oxygenation Bridge to Lung Transplant. Ann Thorac Surg. Published online March 31, 2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Schechter MA, Ganapathi AM, Englum BR, et al. Spontaneously Breathing Extracorporeal Membrane Oxygenation Support Provides the Optimal Bridge to Lung Transplantation. Transplantation. 2016;100(12):2699–2704. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.