Seven-day graft loss after lung transplantation occurred in 2.1% of adult recipients in a national registry analysis covering 2015–2024. ECMO support at 72 hours was the strongest adjusted association with both early graft loss and early death, emphasizing the clinical importance of unresolved postoperative instability. These findings describe associations and do not establish that ECMO causes either outcome.
Study Design and Outcome Definitions
Khan and colleagues analyzed Organ Procurement and Transplantation Network data for 25,593 adult deceased donor lung transplant recipients. The primary endpoint, seven-day graft loss, combined graft failure and recipient death within seven days, including death regardless of graft function. Seven-day posttransplant death was assessed separately.
The graft-loss analysis compared 526 recipients with a first-week event against 18,191 recipients with at least one year of graft survival. Recipients with graft loss on days 8–364 were excluded. The mortality analysis compared 216 first-week deaths with 18,283 recipients who survived at least one year, excluding deaths on days 8–364. These selected comparisons are important: the models do not estimate first-week risk against all other transplant recipients.
Seven-Day Graft Loss After Lung Transplantation: Key Findings
The authors reported 526 early graft-loss events among the 25,593 recipients. Registry-coded primary nonfunction accounted for 63.5% of these events, other or unknown etiologies for 35.4%, and acute rejection for 1.1%. Primary nonfunction coding should not be treated as interchangeable with a standardized, adjudicated diagnosis of primary graft dysfunction.
Cardiovascular failure was the leading reported cause of first-week death, accounting for 24.1%. Graft-failure coding and cause-of-death coding describe different aspects of an event and should not be combined into a single etiologic classification.
Adjusted Associations With Early Graft Loss
| Factor | Adjusted odds ratio | 95% confidence interval |
|---|---|---|
| ECMO use at 72 hours | 9.04 | 7.17–11.40 |
| Retransplantation | 1.80 | 1.15–2.83 |
| Higher serum total bilirubin at transplant | 1.16 | 1.09–1.23 |
| Higher physiologic severity marker | 1.02 | 1.02–1.03 |
| ECMO at listing | 0.61 | 0.39–0.97 |
In the separate mortality model, ECMO at 72 hours had an adjusted odds ratio of 21.97. Higher physiologic severity and bilateral transplantation were also associated with higher odds of first-week death; male sex and higher match-time lung allocation score were associated with lower odds. These findings require cautious interpretation because selection and unmeasured clinical differences may influence the estimates.
Interpreting the ECMO Signal
ECMO at listing and ECMO at 72 hours represent different clinical states. Pretransplant support may bridge a selected patient to transplantation, while continued postoperative support can accompany severe graft dysfunction or cardiopulmonary compromise. The inverse adjusted association for ECMO at listing does not establish a protective effect.
Likewise, the strong postoperative association does not justify withholding or discontinuing ECMO. Confounding by indication is central: patients requiring rescue support are often the sickest patients. Because ECMO was recorded at 72 hours while the outcome window began at transplantation, the exposure is also a postoperative marker within the outcome period, rather than a baseline predictor. The analysis cannot establish event-level temporal ordering for all cases or provide a validated preoperative risk calculator.
Relevance for Perfusion and Transplant Teams
The study supports attention to the first postoperative week as a distinct period for multidisciplinary review. For patients who remain dependent on extracorporeal support, the clinical context—including graft recovery, oxygenation, hemodynamic stability, and systemic organ dysfunction—matters more than treating ECMO status as an isolated prognostic label.
The registry did not capture the detailed perfusion strategies needed to compare circuit management, flow targets, cannulation approaches, or weaning protocols. It also did not specifically test intraoperative ECMO against cardiopulmonary bypass, or evaluate the effectiveness of a particular rescue strategy. These results therefore cannot select an optimal perfusion technique.
For quality improvement, the reporting gaps identify a practical opportunity: record the timing and indication for support, distinguish graft dysfunction from systemic causes of deterioration, and document the sequence of complications. These are implications for better data collection, rather than interventions shown by this study to improve survival.
Limitations and Interpretation
This retrospective registry study remains vulnerable to residual confounding, missing data, and variation in center reporting. The complete-case graft-loss model included 12,602 of the 18,717 recipients in that analytic comparison; the mortality model included 11,644 of 18,499. Missing covariate data may affect representativeness.
Donor age, donation after circulatory death, and ischemic time were examined only in unadjusted comparisons and were not included in the prespecified multivariable models. The study therefore does not establish their independent effects. Excluding intermediate outcomes also limits application of the reported odds ratios to the overall transplant population.
Frequent other or unknown coding limits mechanistic conclusions, particularly when early death occurs without documented graft failure. The data span changes in allocation practice and do not fully represent broader contemporary adoption of machine perfusion. The authors report no research funding or grants and no potential conflicts of interest.
Original Article
Khan S, Toro D, McAfee G, et al. Incidence, Etiologies, and Predictors of Seven-Day Graft Loss in Adult Lung Transplantation: A National Registry Analysis. Experimental and Clinical Transplantation. 2026;10:841–852. doi:10.6002/ect.2026.0167.
The source PDF is unchanged. The cover is an AI-generated editorial illustration.





