Impact of Mild-to-Moderate Donor Lung Infiltration on Post-lung Transplant Survival Outcomes and Risk Factors: A National Cohort Study

Abstract

Objectives

Amid donor shortages, chest X-ray (CXR) is widely used for lung donor screening due to its accessibility, low cost, and rapid imaging. We evaluated the impact of donor CXR infiltrates on post-transplant survival to inform donor lung utilization.

Methods

This retrospective cohort study included adult lung transplant recipients from the United Network for Organ Sharing (UNOS) registry from May 5, 2005 to December 31, 2022. Patients were grouped by the presence or absence of donor CXR infiltrates. Using propensity score matching (PSM) to control variable balance, Kaplan-Meier survival analysis estimated post-transplant survival (at 30 days, 90 days, 1 year, 5 years, and 10 years). Multivariable logistic regression models were applied to identify independent risk donor factors of CXR infiltration.

Results

Unadjusted analyses showed higher short-term survival (30-day/90-day/1-year) in the presence of chest X-ray infiltration group (PCIG) versus the absence of chest X-ray infiltration group (ACIG) (all P < .05), but no differences in 5-year/10-year survival (P > .05) or median survival (P = .567). After PSM, the 2 groups had similar recipient causes of death and rates of postoperative complications. Obesity, prior pulmonary infection, ratio of arterial oxygen partial pressure to fractional inspired oxygen (PaO₂/FiO₂)<300 mmHg, and positive bronchoscopy independently predicted donor infiltrates (all P < .001).

Conclusions

Mild-to-moderate donor lung CXR infiltrates do not independently impair long-term survival or increase major complications. These findings support rational utilization of such lungs within standardized multivariable assessment.

Lung transplantation (LTx) is crucial for end-stage lung disease patients, but donor organ shortages significantly limit its application.

Graphical abstract

INTRODUCTION

Lung transplantation (LTx) is crucial for end-stage lung disease patients, but donor organ shortages significantly limit its application. In 2022, the non-utilization rate of donor lungs in the United States was 9.0%, with a pretransplant mortality rate of 18.8 per 100 patient-years,¹ suggesting current donor evaluation system may overly reject marginal donors.

The diagnostic criteria for extended criteria donors (ECDs) include advanced donor age, decreased oxygenation index, radiographic abnormalities, prolonged ischaemic time, etc.² Bedside chest X-ray (CXR) is commonly used in clinical practice for the preliminary assessment of donor lungs, which is convenient, inexpensive, fast in imaging, and facilitates bedside re-examination. CT, despite its better predictive ability and sensitivity for transplant prognosis, is limited by time consumption and high cost.³ Abnormal donor chest films, especially pulmonary infiltrates, are common in clinical practice but lack unified and quantitative evaluation standards.⁴ Although some studies have indicated that ECDs and standard criteria donors (SCDs) show no significant differences in long-term survival and postoperative outcomes,^5–8 donors with severe aspiration or infectious radiographic evidence are usually rejected. About 12% of potential donors are rejected due to abnormal chest films, mainly pulmonary infiltrates/contusions.⁹

A few single-centre studies have raised similar doubts: Bozovic et al¹⁰ found no significant correlation between radiographic abnormalities and short- or long-term outcomes after lung transplantation, but the study sample included only 35 lung donors from a single institution. Tanaka et al¹¹ confirmed that focal consolidation in pneumonia donors also does not affect 5-year survival, although the definition of pneumonia donors may not align with the standards thoracic surgeons use to assess ECD suitability. However, CXR infiltrates are still empirically used as an evaluation criterion for ideal donor lungs, with low standardization in acceptance or rejection, which does not meet the current donor pool and lacks large cohort studies to assess the association between chest film infiltrate features and transplant prognosis.

The United Network for Organ Sharing (UNOS) database⁶ shows the proportion of donors with abnormal chest films increased from 27% to 70% from 2005 to 2018, reflecting a trend towards increased ECD use. Given the low donor lung utilization rate and the uncertainty of CXR infiltrates in clinical decision-making, this study is the first to systematically evaluate the impact and prognosis of donor lung CXR infiltrates on LTx recipients using the multicentre UNOS database (2005-2022).

METHODS

Study methodology and exclusion criteria

This retrospective cohort study included adult lung transplant recipients in the UNOS database from May 5, 2005 to December 31, 2022. It was exempt from institutional review board approval, followed the International Society for Heart and Lung Transplantation (ISHLT) ethical guidelines, and adhered to the STROBE statement (complete checklist in Table S3).

Records in the time window were screened, with exclusions for: (1) missing survival data, (2) paediatric (<18 years) transplants, (3) repeat lung transplants, (4) single-lung transplants, (5) multi-organ transplants, (6) recorded ex vivo lung perfusion (EVLP), and (7) missing donor CXR data.

Variables with a missing data rate <10% were addressed using multiple imputation by chained equations (MICE) in Stata; those with >30% missing data were removed. Recipient was classified into the absence of chest X-ray infiltration group (ACIG) or the presence of chest X-ray infiltration group (PCIG) based on donor lung CXR infiltration status.

Null hypothesis (H₀)

For adult bilateral lung transplant recipients, donor CXR infiltrates are not independently associated with 5-year post-transplant survival after adjusting for potential confounders.

Study outcomes

Primary outcome is cumulative mortality. Secondary outcomes include post-transplant death causes and postoperative complications. Data collection ended December 31, 2022, with patients categorized as post-transplant alive or deceased.

Statistical analysis

Donor and recipient baseline characteristics and survival data were reported. Categorical variables were presented as frequencies (%), continuous variables as medians (interquartile ranges). Chi-squared tests compared categorical groups with n > 5; Shapiro-Wilk tests assessed continuous variable normality. For non-normal continuous variables, bootstrap (BCa, 1000 replications) estimated between-group median differences and 95% CIs, with 2-sided permutation test-derived P-values. Confounding was controlled via Stata psmatch2-based propensity score matching (PSM): 1:1 nearest-neighbour matching without replacement (caliper 0.2), with pre-matching random seed-setting and sample shuffling for randomness/reproducibility. Covariate balance was confirmed by standardized mean difference (SMD) <10% (Table S2) and kernel density plots (Figure S1). Survival curves were generated using the Kaplan-Meier method. The log-rank test was applied to the unmatched cohort, while the stratified log-rank test was used for the matched cohort to account for the paired structure. Multivariable logistic regression identified donor factors associated with CXR infiltrates. All analyses used STATA/MP 18.0 (StataCorp, College Station, TX).

RESULTS

Baseline characteristics of study population

A total of 21 978 lung transplant recipients were categorized into the ACIG (n = 8911) and the PCIG (n = 13 067) (Figure 1). Before PSM, baseline characteristics differed (Table 1): PCIG recipients were older and had higher body mass index (BMI). PCIG accounting for an increasing proportion over time (13.5% in 2005-2010 to 53.4% in 2017-2022). PCIG recipients showed better pulmonary function (FEV₁, FVC, FEV₁/FVC). Idiopathic pulmonary fibrosis (IPF) was the most common diagnosis in both groups. There were no significant differences in female proportion, racial distribution, history of diabetes, pan-resistant bacterial pulmonary infection, smoking history, ventilator support, human leukocyte antigen (HLA) mismatch, lung allocation score at listing, or pretransplant inpatient status.

Figure 1.

Flowchart of Cohort Buildup

Table 1.

Recipient and Donor Characteristics

		Unmatched sample (N = 21 978)			Propensity 1:1 matching (N = 14 538)
Characteristics	Total (N = 21 978)	ACIG (N = 8911)	PCIG (N = 13 067)	Standardized mean difference	ACIG (N = 7269)	PCIG (N = 7269)	Standardized mean difference
Recipient variables
Age, year	58 [48; 64]	57 [47; 63]	59 [49; 64]	0.124	58 [48; 64]	57 [47; 63]	0.036
Body mass index, kg/m2	25.3 [21.6; 28.8]	25.2 [21.4; 28.7]	25.5 [21.8; 28.8]	0.056	25.3 [21.6; 28.8]	25.2 [21.4; 28.7]	0.028
Female sex	9078 (41.3%)	3574 (40.1%)	5504 (42.1%)	0.041	2946 (40.5%)	2855 (39.3%)	0.026
Race
White	17 255 (78.5%)	7100 (79.7%)	10 155 (77.7%)	0.061	5731 (78.8%)	5690 (78.3%)	0.029
Black	2220 (10.1%)	905 (10.2%)	1315 (10.1%)		754 (10.4%)	706 (9.7%)
Asian	475 (2.2%)	165 (1.9%)	310 (2.4%)		146 (2.0%)	171 (2.4%)
Other	2028 (9.2%)	741 (8.3%)	1287 (9.9%)		638 (8.8%)	702 (9.7%)
Transplant year
2005-2010	4220 (19.2%)	2458 (27.6%)	1762 (13.5%)	0.401	1404 (19.3%)	1514 (20.8%)	0.095
2011-2016	7463 (34.0%)	3142 (35.3%)	4321 (33.1%)		2809 (38.6%)	3102 (42.7%)
2017-2022	10 293 (46.8%)	3309 (37.1%)	6984 (53.4%)		3056 (42.0%)	2653 (36.5%)
Medical history
Diabetes	4309 (19.6%)	1780 (20.0%)	2529 (19.4%)	0.015	1461 (20.1%)	1452 (20.0%)	0.003
Pan-resistant bacterial lung	786 (3.6%)	356 (4.0%)	430 (3.3%)	0.018	265 (3.7%)	259 (3.6%)	0.016
History of cigarette use	12 176 (55.4%)	4910 (55.1%)	7266 (55.6%)	0.009	4061 (55.9%)	3989 (54.9%)	0.02
Respiratory function
FEV1	35 [22; 54]	34 [21; 53]	36 [22; 54]	0.069	35 [22; 53]	35 [21; 53]	0.007
FVC	46 [36; 59]	46 [35; 59]	47 [36; 60]	0.045	46 [36; 59]	46 [35; 58]	0.022
FEV1/FVC	0.92 [0.50; 1.11]	0.89 [0.50; 1.10]	0.94 [0.51; 1.11]	0.048	0.91 [0.50; 1.10]	0.92 [0.50; 1.10]	0.012
Life support ventilator	1628 (7.4%)	647 (7.3%)	981 (7.5%)	0.009	530 (7.3%)	514 (7.1%)	0.009
Lung allocation score diagnosis group
COPD/emphysema	4692 (21.4%)	1866 (20.9%)	2826 (21.6%)	0.029	1536 (21.1%)	1559 (21.5%)	0.021
Cystic fibrosis	2776 (12.6%)	1286 (14.4%)	1490 (11.4%)		941 (13.0%)	993 (13.7%)
IPF	6738 (30.7%)	2692 (30.2%)	4046 (31.0%)		2238 (30.8%)	2239 (30.8%)
Other	7772 (35.4%)	3067 (34.5%)	4705 (36.0%)		2554 (35.1%)	2478 (34.1%)
HLA mismatch ≥ 5	12 918 (58.8%)	5236 (58.8%)	7682 (58.9%)	<0.001	6185 (85.1%)	6142 (84.5%)	0.016
Lung allocation score at match	41.8 [35.5; 57.1]	41.8 [35.5; 56.7]	41.8 [35.5; 57.3]	0.01	42.0 [35.5; 57.7]	42.0 [35.6; 58.2]	0.007
Medical condition before transplant
Not hospitalized	16 494 (75.1%)	6741 (75.7%)	9753 (74.6%)	0.02	5456 (75.1%)	5388 (74.1%)	0.027
In ICU	3233 (14.7%)	1275 (14.3%)	1958 (15.0%)		1072 (14.8%)	1079 (14.8%)
Hospitalized not in ICU	2251 (10.2%)	895 (10.0%)	1356 (10.4%)		741 (10.2%)	802 (11.0%)
Donor variables
Age, years	33 [23; 46]	32 [22; 45]	34 [23; 47]	0.069	32 [23; 45]	33 [22; 47]	0.032
Body mass index, kg/m2	25.4 [22.5; 29.1]	24.7 [22.0; 28.0]	26.0 [22.8; 29.8]	0.247	25.0 [22.2; 28.4]	24.5 [21.9; 27.8]	0.121
Female sex	8916 (40.6%)	3606 (40.5%)	5310 (40.6%)	0.003	2933 (40.4%)	2866 (39.4%)	0.019
Medical history
Diabetes	1653 (7.5%)	605 (6.8%)	1048 (8.0%)	0.063	517 (7.1%)	506 (7.0%)	0.017
History of cigarette use	1888 (8.6%)	840 (9.4%)	1048 (8.0%)	0.003	656 (9.0%)	585 (8.1%)	0.011
Hypertension	5156 (23.5%)	1985 (22.3%)	3171 (24.3%)	0.063	1660 (23.0%)	1634 (22.5%)	0.011
Pulmonary infection	13 622 (62.0%)	4911 (55.1%)	8711 (66.7%)	0.238	4397 (60.5%)	4156 (57.2%)	0.067
Clinical infection	15 219 (69.3%)	5629 (63.2%)	9590 (73.4%)	0.204	4937 (67.9%)	4757 (65.4%)	0.055
Anti-CMV serology	13 783 (62.7%)	5626 (63.1%)	8157 (62.4%)	0.013	4591 (63.2%)	4603 (63.3%)	0.006
Ischaemic time (hours)	5.6 [4.6; 6.6]	5.5 [4.6; 6.6]	5.6 [4.7; 6.6]	0.051	5.5 [4.6; 6.5]	5.5 [4.6; 6.6]	0.025
PaO2/FiO2	443 [380; 500]	460 [400; 515]	431 [370; 488]	0.035	458 [399; 514]	440 [383; 494]	0.03
Positive bronchoscopy findings	3413 (16.2%)	1040 (12.2%)	2372 (18.8%)	0.183	965 (13.3%)	795 (10.9%)	0.072
Donation after cardiac death	772 (3.5%)	288 (3.2%)	484 (3.7%)	0.026	184 (2.5%)	134 (1.8%)	0.047
Cause of death
Anoxia	5591 (25.4%)	1877 (21.1%)	3714 (28.4%)	0.008	1686 (23.2%)	1501 (20.7%)	0.017
Cerebrovascular/stroke	6705 (30.5%)	2918 (32.8%)	3787 (29.0%)		2275 (31.3%)	2352 (32.4%)
Head trauma	9082 (41.3%)	3870 (43.4%)	5212 (39.9%)		3115 (42.9%)	3235 (44.5%)
Neurological tumour	128 (0.6%)	61 (0.7%)	67 (0.5%)		41 (0.6%)	46 (0.6%)
Other	472 (2.2%)	185 (2.0%)	287 (2.2%)		152 (2.1%)	135 (1.9%)

Values are n (%) or median [Q1, Q3].

Abbreviations: ACIG: absence of chest X-ray infiltration group; FEV₁: forced expiratory volume in 1 second; FVC: forced vital capacity; HLA: human leukocyte antigen; COPD: chronic obstructive pulmonary disease; IPF: idiopathic pulmonary fibrosis; ICU: Intensive Care Unit; PCIG: presence of chest X-ray infiltration group.

PCIG donors were older (median 34 [23-47] vs 32 [22-45]; median difference 2.0, 95% CI 1.2-2.8; P < .001) with higher BMI (median 26.0 [22.8-29.8] vs 24.7 [22.0-28.0]; median difference 1.2, 95% CI 1.1-1.4; P < .001), and higher rates of pulmonary infection (66.7% vs 55.1%; SMD = 0.238) and clinical infection (73.4% vs 63.2%; SMD = 0.204). They also had poorer oxygenation (ratio of arterial oxygen partial pressure to fractional inspired oxygen [PaO₂/FiO₂] ratio: median 431 [370-488] vs 460 [400-515]; median difference −29.0, 95% CI −31.9 to −26.1; P < .001), more positive bronchoscopy findings (18.8% vs 12.2%; SMD = 0.183), and higher hypoxia-related mortality. Head trauma was the leading cause of donor death in both groups. Donor sex, diabetes, hypertension, smoking history, Cytomegalovirus (CMV) seropositivity, and donation after circulatory death (DCD) proportion showed no differences.

After PSM, most baseline variables were balanced, except for higher donor BMI in ACIG (Figure S1).

Survival rates and causes of death

PCIG recipients had slightly higher short-term survival rates than ACIG (30-day: 97.5% vs 96.8%; 90-day: 95.2% vs 94.2%; 1-year: 88.6% vs 87.4%; all P < .05) (Table 2). Consistent with our null hypothesis, no differences were seen in 5-year survival (61.1% vs 60.7%, P = .318; Figure 2), 10-year survival (35.9% vs 35.8%, P = .394), or median survival time (7.14 vs 7.01 years, P = .567). Overall, 43.7% (9606) of recipients died by study end. After PSM, the 2 groups showed similar distributions of specific death causes (Table 3).

Table 2.

Survival Rates Before and After Matching

		Unmatched sample (N = 21 978)			Propensity 1:1 matching (N = 14 538)
Survival rate	Total (N = 21 978)	ACIG (N = 8911)	PCIG (N = 13 067)	P-value	ACIG (N = 7269)	PCIG (N = 7269)	P-value
30-day survival, % (95% CI)	97.2 (97.0; 97.4)	96.8 (96.4; 97.1)	97.5 (97.2; 97.7)	.003	97.0 (96.0; 97.5)	97.2 (96.5; 97.7)	.302
90-day survival, % (95% CI)	94.8 (94.5; 95.1)	94.2 (93.7; 94.7)	95.2 (94.8; 95.6)	<.001	94.6 (92.9; 94.8)	94.8 (93.8; 95.3)	.520
1-year survival, % (95% CI)	88.1 (87.7; 88.6)	87.4 (86.7; 88.1)	88.6 (88.0; 89.2)	.007	88.2 (85.4; 88.2)	88.1 (87.5; 89.7)	.879
5-year survival, % (95% CI)	61.0 (60.2; 61.7)	60.7 (59.6; 61.9)	61.1 (60.1; 62.1)	.318	61.4 (59.6; 64.0)	61.1 (58.4; 62.4)	.417
10-year survival, % (95% CI)	35.9 (34.9; 36.9)	35.8 (34.4; 37.2)	35.9 (34.5; 37.3)	.394	35.2 (34.4; 39.8)	36.4 (29.9; 35.6)	.878
Median survival, years (95% CI)	7.07 (6.90; 7.23)	7.01 (6.77; 7.23)	7.14 (6.89; 7.33)	.567	7.07 (6.79; 7.34)	7.15 (6.86; 7.38)	.877

Abbreviations: ACIG: absence of chest X-ray infiltration group; PCIG: presence of chest X-ray infiltration group.

Figure 2.

Kaplan-Meier Survival Estimates Stratified by Chest X-ray Infiltration Status. Kaplan-Meier analysis showed no statistical differences between the ACIG and PCIG in 5-year overall survival, both before (A) and after (B) propensity score matching. Comparisons by log-rank test. Statistical significance was set to P < .05. Abbreviations: ACIG: absence of chest X-ray infiltration group; PCIG: presence of chest X-ray infiltration group

Table 3.

Outcomes Before and After Matching

		Unmatched sample (N = 21 978)			Propensity 1:1 matching (N = 14 538)
Variables	Total (N = 21 978)	ACIG (N = 8911)	PCIG (N = 13 067)	P-value	ACIG (N = 7269)	PCIG (N = 7269)	P-value
Cause of death
Graft failure	1780 (18.5%)	782 (17.8%)	998 (19.1%)	.083	588 (17.7%)	653 (19.3%)	.145
Infection	1763 (18.4%)	794 (18.1%)	969 (18.6%)		582 (17.6%)	586 (17.3%)
Cardiovascular	961 (10.0%)	453 (10.3%)	508 (9.7%)		348 (10.5%)	310 (9.2%)
Lung diseases	1726 (18.0%)	769 (17.5%)	957 (18.3%)		590 (17.8%)	612 (18.1%)
Haemorrhage	152 (1.6%)	63 (1.4%)	89 (1.7%)		44 (1.3%)	57 (1.7%)
Post-transplant malignancy	859 (8.9%)	409 (9.3%)	450 (8.6%)		303 (9.1%)	327 (9.7%)
Multiple organ failure	471 (4.9%)	217 (5.0%)	254 (4.9%)		164 (5.0%)	160 (4.7%)
Renal failure	180 (1.9%)	77 (1.8%)	103 (2.0%)		56 (1.7%)	75 (2.2%)
Other	739 (7.7%)	331 (7.6%)	408 (7.8%)		260 (7.9%)	270 (8.0%)
Unknown\missing	975 (10.2%)	489 (11.2%)	486 (9.3%)		379 (11.4%)	336 (9.9%)
Outcomes from transplant to discharge
Length of stay Tx to discharge	15 [13; 31]	18 [13; 31]	19 [13; 31]	.005	18 [13; 30]	18 [13; 30]	1
Ventilator support
No ventilator support	571 (2.6%)	245 (2.8%)	326 (2.5%)	<.001	192 (2.6%)	222 (3.1%)	.098
<48 h	11 987 (54.5%)	4887 (54.9%)	7096 (54.3%)		4029 (55.5%)	3965 (54.6%)
48 h-5 days	3875 (17.6%)	1494 (16.8%)	2380 (18.2%)		1236 (17.0%)	1329 (18.3%)
≥5 days	5059 (23.0%)	2016 (22.7%)	3038 (23.3%)		1627 (22.4%)	1594 (21.9%)
Unknown/missing	486 (2.2%)	264 (3.0%)	222 (1.7%)		180 (2.5%)	158 (2.2%)
Complications morbidity
Acute rejection	1646 (7.5%)	709 (8.0%)	936 (7.2%)	.03	555 (7.6%)	522 (7.2%)	.292
Airway dehiscence	376 (1.7%)	155 (1.74%)	221 (1.69%)	.002	121 (1.7%)	129 (1.8%)	.715
Stroke	589 (2.7%)	230 (2.6%)	359 (2.8%)	.017	191 (2.6%)	194 (2.7%)	.356
Tracheostomy	995 (4.5%)	382 (4.3%)	613 (4.7%)	.03	320 (4.4%)	315 (4.3%)	.901
Dialysis	1781 (8.1%)	693 (7.8%)	1088 (8.3%)	<.001	567 (7.8%)	580 (8.0%)	.012
Reintubation	4304 (19.6%)	1754 (19.7%)	2550 (19.5%)	.01	1405 (19.3%)	1375 (18.9%)	.813
Treated for rejection within 1 year	4031 (18.3%)	1740 (19.5%)	2291 (17.5%)	.001	1399 (19.3%)	1353 (18.6%)	.167

Values are n (%) or median [Q1, Q3].

Abbreviations: ACIG: absence of chest X-ray infiltration group; PCIG: presence of chest X-ray infiltration group.

Postoperative outcomes and complications morbidity

PCIG recipients had a slightly longer post-transplant hospital stay (median 19 [13; 31] vs 18 [13; 31]; median difference 1, 95% CI 0.3 to 1.7; P = .005). The ACIG group showed higher rates of acute rejection and airway dehiscence, among others (Table 3). After PSM, the previously observed differences in length of stay and most complications were eliminated. Except for a higher rate of postoperative dialysis in the PCIG group (8.0% vs 7.8%, P = .012), differences in other postoperative complications between groups were not statistically significant.

Risk factors for donor lung CXR infiltration

Multivariable logistic regression identified several donor factors significantly associated with CXR infiltrates (Table 4). Higher BMI (especially obesity, BMI ≥30 kg/m²) demonstrated the strongest association, while underweight was protective. Other significant risk factors included pulmonary or clinical infection, impaired oxygenation (PaO₂/FiO₂ <300 mmHg), positive bronchoscopy findings, prolonged ischaemic time, and cause of death (anoxia or head trauma). In contrast, a history of cigarette use was associated with reduced risk. Donor sex, diabetes, hypertension, CMV serology, and DCD status showed no significant associations.

Table 4.

Multivariable Logistic Regression: Donor Factors Associated with CXR Infiltrates

Variables	OR	95% CI	P-value
Age ≥ 55 y	1.19	1.08-1.31	<.001
Body mass index, kg/m2
18.5-25	Ref	Ref	Ref
<18.5	0.82	0.71-0.96	.012
25-30	1.35	1.26-1.44	<.001
≥30	1.83	1.69-1.98	<.001
Female sex	1.01	0.95-1.07	.802
Diabetes	1.04	0.92-1.17	.527
History of cigarette use	0.83	0.75-0.92	<.001
Hypertension	1.06	0.98-1.15	.144
Pulmonary infection	1.40	1.25-1.56	<.001
Clinical infection	1.13	1.01-1.27	.039
Anti-CMV serology	1.00	0.94-1.06	.938
Ischaemic time (hours)	1.02	1.01-1.04	.008
PaO2/FiO2<300 mmHg	1.51	1.34-1.70	<.001
Positive bronchoscopy findings	1.59	1.46-1.72	<.001
Donation after cardiac death	1.01	0.85-1.21	.871
Cause of death
Cerebrovascular/stroke	Ref	Ref	Ref
Anoxia	1.51	1.39-1.64	<.001
Head trauma	1.14	1.06-1.23	.001
Neurological tumour	0.90	0.63-1.29	.578
Other	1.28	1.05-1.55	.016

DISCUSSION

The global shortage of donor lungs has made ECD a key solution, with around 80% of lung transplant recipients receiving lungs meeting at least 1 ECD criterion.^6,12 Bedside CXR is widely used for donor lung screening due to their convenience and low cost.³ CXR infiltrates may indicate contusions, infections, or other lesions, and could be linked to poor post-transplant outcomes.¹³ Ideally, donor lungs should have no radiographic infiltrates—especially for low-risk or selected recipients—but this standard relies heavily on clinical experience^2,3 and lacks large-scale evidence.

In our multicentre large-cohort assessment based on UNOS data, we found that although chest X-ray infiltrates were associated with several donor characteristics (such as high BMI, prior pulmonary infection, reduced PaO₂/FiO₂, and positive bronchoscopy), after adjusting for confounders, mild to moderate radiographic infiltrates were not an independent risk factor for 5-year overall survival. Propensity-matched analysis also showed no significant differences between the groups in overall survival, major causes of death, or postoperative complications, suggesting that with comprehensive, standardized evaluation and appropriate perioperative management, mild to moderate radiographic infiltrates do not necessarily worsen long-term transplant outcomes.¹⁴

Our study confirms that using ECD lungs does not affect transplant outcomes,^5–8 and CXR infiltrates were not significantly associated with long-term survival, postoperative causes of death, or adverse events like stroke. The matched cohort analysis further supports this: after accounting for donor-recipient confounders, ACIG and PCIG showed no differences in survival, death causes, or postoperative outcomes, consistent with single-centre studies by Bozovic et al¹⁰ and Tanaka et al.¹¹

The PCIG group’s short-term survival advantage may be related to better preoperative pulmonary function in its recipients, consistent with Fakhro et al,¹⁵ who found FEV₁ levels correlate positively with post-transplant survival. Notably, while there were differences in recipient age, BMI between groups, these did not affect clinical grading, suggesting that while preoperative pulmonary function impacts short-term survival, long-term survival may be influenced by a combination of multiple factors.¹⁶ The cohort’s high proportion of IPF patients may have limited the association between preoperative lung function and long-term survival. Additionally, more intensive monitoring of PCIG recipients might explain their short-term survival advantage, a hypothesis that needs prospective confirmation.

In unmatched donor baseline comparisons, ACIG and PCIG differed statistically in variables like age and BMI, but these differences were small in the overall magnitude and all met current donor lung assessment standards. PSM achieved balance on most key variables, though residual differences remained in BMI, pulmonary/clinical infection history, PaO₂/FiO₂, and bronchoscopy positivity. Combined with the multivariable logistic regression results, these factors are interrelated and jointly influence the occurrence of radiographic infiltrates. This phenomenon is consistent with the clinical characteristics of the PCIG population and validates the effectiveness of modern donor management systems: systematic imaging and microbiological screening, rapid diagnostic techniques, and targeted perioperative anti-infective strategies can convert controllable or focal infections into acceptable transplant resources,¹⁷ reducing postoperative complication risks and ensuring long-term outcomes.

Hypoxia-related death and lower PaO₂/FiO₂ were more common in PCIG donors, and PaO₂/FiO₂ < 300 mmHg independently predicted infiltrates. Hypoxia increases microvascular permeability, producing high-protein oedema visible on CXR,¹⁸ while obesity reduces functional residual capacity (FRC/ERV), inducing small airway closure and ventilation inhomogeneity, leading to imaging changes.¹⁹ Thus, radiographic shadows reflect combined effects of impaired oxygenation, infection, and lung function rather than a single pathology, supporting multiparameter donor assessment and discouraging isolated CXR exclusion.

Although not specifically documented in this study, EVLP can assess and recondition marginal donor lungs, improving function in cases of embolism, hypoxia, or infection to expand the donor pool.²⁰

Substantial inter-surgeon variability in CXR interpretation²¹ underscores the limitations of single-film decision-making. Aligning with a recent UNOS analysis,²² we advocate standardizing donor assessment through multifactorial, objective models (eg, LUNDON score) that integrate age, smoking history, infection markers, and PaO₂/FiO₂ to minimize subjectivity.

Over time, PCIG donor use increased significantly (2017-2022), reflecting greater acceptance of ECD. This likely stems from optimized donor selection, improved perioperative management and anti-infective strategies, advances in immunosuppression and organ preservation,²³ and techniques such as EVLP, improving organ utilization and recipient outcomes without compromising long-term survival.

Overall, our study shows that mild-to-moderate CXR infiltrates are a collective sign of various donor factors rather than an independent risk for long-term survival. Therefore, through strict multiparameter evaluation and targeted optimization, these lungs can be prudently utilized to expand the donor pool. Future studies should refine imaging grading and prospectively validate assessment protocols to guide clinical practice and avoid unnecessary organ rejection.

Limitations of this study include: (1) UNOS does not provide original chest radiographs or CT images nor inflammatory markers such as CRP/PCT, and variables like primary graft dysfunction (PGD) and EVLP usage lack standardized definitions, so the data may have missing entries or coding errors; (2) in clinical practice lungs with severe or diffuse consolidation are often excluded during selection, and the cohort in this study mainly includes mild-to-moderate, reversible infiltrates, so the conclusions cannot be extrapolated to severe imaging abnormalities; (3) the retrospective study design may introduce recall bias; chest radiograph and bronchoscopy findings are subject to interpretive variability, which may lead to information bias; (4) although more than 20 donor-recipient-operative variables were included and PSM was used, unmeasured or unrecorded confounders cannot be completely ruled out.

CONCLUSION

Donor lung CXR infiltrates, under multidimensional assessment and modern perioperative management, do not independently compromise survival or transplant outcomes. This supports abandoning their use as a sole exclusion criterion in favour of multifactorial, objective assessment to expand the donor pool. Future efforts should standardize evaluation protocols and leverage technologies like EVLP to optimize organ utilization.

AUTHOR CONTRIBUTIONS

J.H., C.L., and X.L. contributed equally to data collection and analysis. C.L., G.P., J.S., and X.X. conceptualized the study. X.L., C.L., and W.H. collected the data. X.L., C.L., Y.P., J.Z., Y.L., Z.W., C.Y., and J.S. did the analysis and the interpretation of data. All authors assisted in the writing of the article and approved the final article.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

This work was supported by Specific Project of Guangzhou National Laboratory, Grant GZNL2024A03006.

CONFLICTS OF INTEREST

None declared.

DATA AVAILABILITY

Data are available on reasonable request.

PROVENANCE AND PEER REVIEW

Not commissioned; externally peer reviewed.

ETHICS STATEMENTS

This is a publicly available database that does not contain information such as patient identity. The Institutional Review Board of the First Affiliated Hospital of Guangzhou Medical University determined that formal review was not required for this study.