Dangerous liaisons: the peri-transplant respiratory microbiota and their impact on lung transplantation outcomes Authors: Maximilian Vorstandlechner, Julia Walter, Mara A Bauer, David Steybe, Jan M Fertmann, Christian P Schneider, Sebastian Michel, Jürgen Barton, Ines Schroeder, Gökce Yavuz, Olaf M Glueck, Alexandra Weber, Wenko Smolka, Teresa Kauke, Katharina Theresa Obermeier Published: 2026 ABSTRACT AbstractBackgroundDespite advances in immunosuppression and perioperative care, lung transplantation (LuTX) outcome continues to be limited by allograft rejection and infectious complications. Given the allograft’s direct exposure to the environment, peri-transplant microbial presence and dynamics may critically shape outcomes, yet its impact remains incompletely understood. This study aimed to evaluate peri-transplant airway microbiological findings in lung transplant recipients and determine their associations with acute cellular rejection (ACR), chronic lung allograft dysfunction (CLAD), and survival.MethodsWe retrospectively analyzed 345 LuTX recipients treated between 2014 and 2019 with follow-up through 2024. Standardized microbiological sampling was performed pre-, intra- and postoperatively. Associations with ACR, CLAD and survival were evaluated.ResultsCandida species (spp.) (22.6% and 18.3%, respectively) and Pseudomonas aeruginosa (10.7% and 7.2%) were most frequently detected in pre- and postoperative sputum samples. Donor lungs showed positive microbial samples in 62.3% of cases, predominantly with Candida spp. (31.3%) and Staphylococcus spp. (26.1%). Multivariate analysis identified donor-derived Escherichia coli as an independent predictor of ACR (P=0.01; odds ratio 3.2), CLAD [P=0.03; hazard ratio (HR) 2.6] and impaired survival (P<0.001; HR 3.2). Furthermore, donor-derived Achromobacter spp. (P=0.01; HR 13.6) and postoperative deep airway Candida detection (P=0.045; HR 1.9) predicted mortality. While Klebsiella was associated with adverse outcomes in univariate analysis, this did not remain significant after adjustment.ConclusionsPeri-transplant airway microbiota are significantly associated with graft rejection and survival after LuTX. Donor-derived E. coli and Achromobacter spp., as well as postoperative Candida detection, emerged as robust predictors of poor outcome. Systematic microbiological surveillance and tailored prophylaxis are essential to mitigate microbial-driven allograft dysfunction and improve long-term survival.Keywords: Lung transplantation (LuTX), acute cellular rejection (ACR), chronic lung allograft dysfunction (CLAD), microbiota Highlight box Key findings • In a cohort of 345 lung transplant recipients, donor-derived Escherichia coli (E. coli) was the most consistent independent predictor of acute cellular rejection, chronic lung allograft dysfunction (CLAD) and reduced long-term survival. Donor-derived Achromobacter species (spp.) and postoperative Candida spp., particularly de novo and persistent detection, were independently associated with increased mortality. What is known and what is new? • Airway microbiota influence rejection, graft dysfunction, and survival after lung transplantation but the clinical relevance of donor-derived microorganisms remains unclear. • This study identifies donor lung microbiota, especially E. coli, as key determinants of rejection, CLAD and survival, exceeding the prognostic impact of recipient colonization alone. What is the implication, and what should we change now? • Routine donor lung microbiological screening and early postoperative airway surveillance should be integrated into risk stratification. High-risk microbial patterns may justify intensified monitoring and individualized antimicrobial and immunosuppressive management to improve long-term outcomes. Introduction Lung transplantation (LuTX) serves as a critical therapeutic option for patients with chronic progressive pulmonary diseases, including interstitial lung disease (ILD), chronic obstructive pulmonary disease (COPD) and cystic fibrosis (CF) (1,2). For these patients LuTX represents a pivotal turning point, offering significant potential improvements in both survival expectancy and quality of life (2). Despite advancements in the diagnosis and management of graft failure (GF), acute cellular rejection (ACR) and primary graft dysfunction (PGD) remain prevalent complications during the first weeks following LuTX (3). For patients who survive this critical early post-transplant period, chronic lung allograft dysfunction (CLAD) emerges as one of the leading causes of mortality (4). Another significant risk factor is the transplant recipient’s heightened susceptibility to infectious complications. This vulnerability is exacerbated by the necessary and often stringent adherence to immunosuppressive medication regimens (5,6). Due to its constant exposure to the external environment, the respiratory tract is one of the most common sites of infection. Factors such as the compromised immune system, atelectasis during the transplantation procedure and the onset of muscle atrophy due to intensive care unit-acquired weakness collectively contribute to an increased risk of post-transplant pneumonia, a complication often triggered by pathogenic organisms from the tracheobronchial tree (7,8). It is a widely accepted understanding that the microbiological composition to which an individual is exposed can play a significant role in the onset and progression of various diseases. To address this potential risk and respond appropriately within this particularly vulnerable patient population, we have developed a center-specific protocol that includes comprehensive screening and testing for pathogenic microorganisms at multiple stages of the transplantation process. Screening occurs at key points, including pre-, intra- and postoperative phases. This approach allows us to identify potentially harmful pathogens that may fall outside the typical microbiological spectrum. Objective The objective of the underlying retrospective cohort study was to collect microbiological samples from both the native and donor lungs at various stages throughout the perioperative course of LuTX and to investigate the extent to which specific, potentially pathogenic microbiota may influence postoperative survival and the development of allograft rejection. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0260/rc). Methods The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional review board of LMU University Hospital, LMU Munich, Munich, Germany (project No. 22-0466). Owing to the retrospective design of the study and the absence of any additional interventions or diagnostic procedures, the requirement for patients’ informed consent was waived. The underlying retrospective cohort study includes patients who underwent LuTX at LMU University Hospital between 2014 and 2019. Follow-up data were collected until the conclusion of the study in 2024. Throughout the entire study period, perioperative management remained consistent and all patients initially received postoperative standard triple immunosuppressive therapy, consisting of tacrolimus, mycophenolate mofetil (MMF) and prednisolone. Standard antimicrobial prophylaxis included meropenem for one week post-transplantation. Induction therapy was not routinely administered at our institution during the study period. Cotrimoxazole was administered as Pneumocystis jirovecii pneumonia prophylaxis. Antifungal prophylaxis included inhaled amphotericin B and oral itraconazole. If microbial organisms were identified that were not covered by standard prophylaxis and additionally considered clinically relevant, antimicrobial treatment was promptly adjusted based on resistance profiles provided in antibiograms, in close collaboration with the institutional Antibiotic Stewardship team. Strict inclusion criteria were applied, which mandated the collection of intraoperative bronchial swabs from both the recipient’s native lung and the donor lung, as well as pre- and postoperative deep airway secretion samples (sputum). Preoperative sputum samples were routinely collected during lung allocation score (LAS) update visits, which are routinely performed at 3-month intervals for patients on the waiting list; accordingly, the timing of the last available sample prior to transplantation reflects this standardized schedule. Postoperative sputum samples were routinely obtained within the early post-transplant period, typically following transfer from the intensive care unit to the general ward. In addition, bronchoalveolar lavage (BAL) and endotracheal aspiration (ENTA) samples were collected during postoperative bronchoscopy as part of routine transbronchial biopsy (TBB) protocols. This approach enabled correlation of microbiological findings from BAL/ENTA samples with concurrent ACR assessment based on TBB. TBB was generally performed once the maintenance dose of prednisolone (15 milligrams/day) was reached. Although not obtained on the same day, these sputum and BAL/ENTA samples were generally collected within a close temporal window of a few days to weeks, reflecting routine clinical practice. Microbiological analyses included routine bacterial and fungal cultures obtained from respiratory specimens; acid-fast bacilli (AFB) testing was not routinely performed and was therefore not included in this analysis. Samples were considered positive when cultivable bacteria or fungi were detected in respiratory specimens, as reported by the Department of Microbiology. To assess perioperative microbiota dynamics, patients were classified into five groups based on the presence (+) or absence (−) of microorganisms before (pre-transplant sputum and native bronchial swab) and after transplantation (postoperative sputum and BAL/ENTA), as well as the contribution of donor-derived microorganisms: ❖ −/−: no detectable microorganisms before and after transplantation; ❖ −/−: no detectable microorganisms before and after transplantation; ❖ −/+: new postoperative acquisition of microorganisms not present pre-transplant (donor-independent); ❖ −/+: new postoperative acquisition of microorganisms not present pre-transplant (donor-independent); ❖ −D+: new postoperative detection attributable to donor-derived microorganisms (donor-dependent); ❖ −D+: new postoperative detection attributable to donor-derived microorganisms (donor-dependent); ❖ +/−: microorganisms detected pre-transplant but no longer detected post-transplant (clearance); ❖ +/−: microorganisms detected pre-transplant but no longer detected post-transplant (clearance); ❖ +/+: persistence of microorganisms detected both before and after transplantation. ❖ +/+: persistence of microorganisms detected both before and after transplantation. All patient records were retrospectively reviewed. Data collected included the diagnosis, underlying disease leading to LuTX listing, surgery date, transplantation type, mechanical ventilation, graft function, cause of death (COD) and survival. CLAD was defined as a substantial and persistent decline (≥20%) in forced expiratory volume in one second (FEV1) when compared with the post-transplant baseline more than 3 months after transplantation (9). The occurrence of pyloric stenosis after transplantation was assessed based on clinically indicated esophagogastroduodenoscopy (EGD) in patients presenting with symptoms suggestive of gastric outlet obstruction (e.g., postprandial fullness, nausea, vomiting, or radiographic evidence of a distended gastric bubble), to explore potential aspiration-related mechanisms. Statistical analysis Statistical analyses were conducted using R Studio version 1.3.1093 (Free Software Foundation, Boston, MA, USA). Descriptive statistics summarized demographic and clinical characteristics. Normally distributed data were reported as mean ± standard deviation (SD), non-normally distributed data as median and interquartile range (IQR). To compare distributions between two groups, Fisher’s exact test was used for categorical variables, with odds ratios (ORs) reported where appropriate. For continuous variables, Mann-Whitney-Wilcoxon test was used for two-group comparisons and Kruskal-Wallis for multi-group comparisons. The primary aim of the analysis was to explore associations between peri-transplant microbial patterns and clinical outcomes rather than to construct a fully adjusted predictive model. For binary outcomes such as ACR, univariate and multivariate logistic regression models were applied, reporting OR with corresponding P values. For time-to-event outcomes, including CLAD and overall survival, Kaplan-Meier estimates were used for survival analysis and differences between groups were assessed using Cox proportional hazards regression models. Both univariate and multivariate Cox regression analyses were performed, with hazard ratios (HRs) reported. Censoring was applied automatically in all time-to-event analyses. A P value of <0.05 was considered statistically significant. Results Study population and patient characteristics Between 2014 and 2019, 411 patients underwent LuTX at our institution. Sixty-six patients were excluded due to incomplete microbiological data, yielding a final cohort of 345 patients. The demographic characteristics (Table 1) indicate a balanced sex distribution [male/female: n=181/164 (52.5%/47.5%)]. The majority of patients (n=251, 72.2%) underwent bilateral LuTX. The leading LuTX-indication was ILD (n=156, 45.2%), followed by COPD (n=87, 25.2%) and CF (n=47, 13.6%). Fifteen patients (4.3%) underwent retransplantation due to bronchiolitis obliterans syndrome (BOS). Of the 345 patients, 11 died within the first 30 days (3.2%), 47 within the first year (13.6%) and overall 144 during follow-up (41.7%). The most common COD was multi-organ failure (MOF) and sepsis (n=50; 34.7%) or GF (n=26; 18.1%). Sepsis and MOF were analyzed as a combined category due to substantial clinical overlap and limitations in reliably distinguishing these entities in the retrospective dataset, as sepsis frequently represents a precipitating factor for MOF. Table 1Study population and patient characteristicsCharacteristicsValuePatients345 (100.0)Age (years)55 [46–61]Sex Male181 (52.5) Female164 (47.5)Diagnosis CF47 (13.6) COPD87 (25.2) ILD156 (45.2) BOS15 (4.3) PH17 (4.9) Sarcoidosis9 (2.6) Other14 (4.1)Type of LuTX Single94 (27.2) Double251 (72.8)Mechanical ventilation (hours)129 [47–763]ACR A0272 (78.8) A126 (7.5) A23 (0.9) A31 (0.3) A40 (0.0)CLAD76 (22.0)Time to CLAD (years)2.7±2.0Overall survival201 (58.3) ≥30 days334 (96.8) ≥1 year298 (86.4) ≥3 years253 (73.3) ≥5 years219 (63.5)Cause of death GF26 (7.5) Malignancy14 (4.1) MOF/sepsis50 (14.5) Hemorrhage/circulatory11 (3.2) Other32 (9.3) COVID11 (3.2)Data are presented as n (%), median [interquartile range], or mean ± standard deviation. ACR, acute cellular rejection; BOS, bronchiolitis obliterans syndrome; CF, cystic fibrosis; CLAD, chronic lung allograft dysfunction; COPD, chronic obstructive pulmonary disease; COVID, coronavirus disease; GF, graft failure; ILD, interstitial lung disease; LuTX, lung transplantation; MOF, multi-organ failure; PH, pulmonary hypertension. Study population and patient characteristics Data are presented as n (%), median [interquartile range], or mean ± standard deviation. ACR, acute cellular rejection; BOS, bronchiolitis obliterans syndrome; CF, cystic fibrosis; CLAD, chronic lung allograft dysfunction; COPD, chronic obstructive pulmonary disease; COVID, coronavirus disease; GF, graft failure; ILD, interstitial lung disease; LuTX, lung transplantation; MOF, multi-organ failure; PH, pulmonary hypertension. Microbiological findings Pre-transplant The most common findings in recipient’s pre-LuTX sputum samples were Candida species (spp.) (n=78; 22.6%), followed by Pseudomonas aeruginosa (P. aeruginosa) (n=37; 10.7%) and Staphylococcus spp. (n=23; 6.7%) (Figure 1, Table 2). In the bronchial swab of the native lung Gram-negative bacteria (GNB) were the most commonly identified organisms (n=74, 21.4%), with P. aeruginosa (n=34, 9.9%) and Achromobacter spp. (n=14, 4.1%) being the most frequently detected. While the prevalence of P. aeruginosa and Staphylococcus spp. in intraoperative samples was roughly consistent with their occurrence in preoperative sputum cultures, we observed a substantial discrepancy in the detection of Candida spp., with a markedly higher prevalence in preoperative sputum samples (n=78; 22.6%) compared to intraoperative bronchial swabs (n=23; 6.7%). This notable difference in Candida spp. detection suggests potential variation in microbial load or sampling methodology between the upper and lower airways Microbiological findings. BAL, bronchoalveolar lavage; ENTA, endotracheal aspiration; spp., species. Table 2Microbiological findingsSamples (345 patients overall)Preoperative sputumIntraoperative bronchus swabPostoperative sputumBAL/ENTARecipientRecipientDonorRecipientRecipientn%n%n%n%n%Patients with positive findings13739.712435.921562.310329.910831.3Gram-positive298.45014.510229.6113.2185.2 Actinomyces odontolyticus 00.041.200.000.000.0 Enterococcus faecalis 00.020.620.600.010.3Staphylococcus spp.236.7298.49026.1102.9164.6Streptococcus spp.30.9133.882.310.310.3Gram-negative8324.17421.410129.35215.16117.7Achromobacter spp.92.6144.120.641.2123.5 Acinetobacter baumannii 30.930.961.720.610.3Burkholderia spp.20.610.310.330.930.9Citrobacter spp.30.900.092.600.010.3 Enterobacter aerogenes 41.210.341.200.000.0 Enterobacter cloacae 20.620.6133.820.630.9 Escherichia coli 82.320.6174.961.761.7 Haemophilus influenzae 10.320.6164.600.010.3Klebsiella spp.72.010.3123.551.420.6 Moraxella catarrhalis 41.210.310.300.000.0Prevotella spp.00.010.330.900.000.0 Proteus mirabilis 30.930.941.200.010.3 Pseudomonas aeruginosa 3710.7349.9144.1257.2257.2 Serratia marcescens 92.641.292.661.772.0 Stenotrophomonas maltophilia 113.272.061.730.9113.2Fungi8424.33710.711733.96819.75315.4Aspergillus spp.123.582.3113.251.4123.5Candida spp.7822.6236.710831.36318.34312.5BAL, bronchoalveolar lavage; ENTA, endotracheal aspiration; spp., species. Microbiological findings BAL, bronchoalveolar lavage; ENTA, endotracheal aspiration; spp., species. As some patients received immunosuppressive medication for underlying chronic lung diseases prior to LuTX, we investigated its effect on the microbial spectrum. Data were available for 282 patients (81.7%), of whom 190 were actively receiving immunosuppressive therapy preoperatively. Agents used were prednisolone (n=184), azathioprine (n=26), MMF (n=25), tacrolimus (n=14), cyclophosphamide (n=7) and rituximab (n=3). Patients receiving azathioprine were significantly more likely to have positive bronchial swabs for Candida spp. (P=0.003; OR=3.8), while those on cyclophosphamide had a higher incidence of P. aeruginosa in preoperative sputum samples (P=0.03; OR=6.8). No significant differences were found for other aforementioned immunosuppressive agents. Bronchus swab of the donor lung A total of 215 donor lungs (62.3%) exhibited detection bacterial and/or fungal pathogens (Figure 1, Table 2). Candida spp. (n=108, 31.3%) and Staphylococcus spp. (n=90, 26.1%) were the most frequently detected organisms, followed by Escherichia coli (E. coli) (n=17, 4.9%), Haemophilus influenzae (n=16, 4.6%) and P. aeruginosa (n=14, 4.1%). While data on antibiotic resistance were limited, susceptibility information for Staphylococcus aureus was available in a subset of donor bronchial swabs. Among 64 of 266 isolates analyzed, 3 cases (4.7%) were identified as methicillin-resistant Staphylococcus aureus. Post-transplant Consistent with preoperative findings, Candida spp. (n=63, 18.3%), P. aeruginosa (n=25, 7.2%), and Staphylococcus spp. (n=10, 2.9%) represented the majority of detected microorganisms. However, in comparison to other sample types analyzed in this study, postoperative sputum samples exhibited a relatively lower proportion of microbiota detection rates, resulting in a reduced overall microbial burden (n=103; 29.9%). BAL and ENTA samples were analyzed to assess endobronchial microbiota in lung transplant recipients. Among bacterial species, GNB were predominant (n=61; 17.7%) [P. aeruginosa (n=25, 7.2%), Achromobacter spp. (n=12, 3.5%), Stenotrophomonas maltophilia (n=11, 3.2%)]. Gram-positive bacteria (GPB) were detected in 5.2% of cases (n=18), predominantly Staphylococcus spp. (n=16, 4.6%). Candida spp. (n=43, 12.5%) and Aspergillus spp. (n=12, 3.5%) accounted for the majority of fungal isolates. Additional analyses comparing unilateral and bilateral LuTX did not demonstrate a significant difference in postoperative detection rates of the investigated microorganisms. Perioperative microbial dynamics in LuTX-recipients Ninety-eight patients (28.4%) acquired new, donor-independent, microbiota postoperatively (−/+). Donor-derived microbial transmission was identified in 35 patients (−D+) (10.1%), while 87 patients (+/−) (25.2%) demonstrated clearance of preoperatively detected microorganisms. Persistent microbial detection was observed in 37 patients (+/+) (10.7%). As summarized in Figure 2, GPB were newly acquired postoperatively in 15 patients (4.3%), independent of the donor, while seven patients (2.0%) acquired GPB from the donor lung. Clearance of GPB was observed in 66 patients (19.1%), while five patients (1.4%) exhibited persistent detection. Staphylococcus spp. persisted in only 1.4% of cases and were newly acquired postoperatively in 3.5% of patients. Perioperative microbial dynamics. Legend (pre- and post-transplant status regarding detection of microbiota): –/–: no detectable microorganisms before and after transplantation; –/+ (donor-independent): new postoperative acquisition not present pre-transplant; –D+ (donor-derived microbiota): microorganisms not detected in recipient samples prior to transplantation, identified in donor bronchial swabs, and subsequently detected postoperatively in recipient samples, consistent with potential donor-to-recipient transmission; +/–: clearance of pre-transplant detection; +/+: persistence of preoperatively detected microbiota. spp., species. Regarding GNB, P. aeruginosa new-acquisition was observed in 23 patients (6.7%), while only two patients (0.6%) showed donor-dependent occurrence. Clearance of P. aeruginosa occurred in 33 patients (9.6%), whereas 19 patients (5.5%) exhibited persistence postoperatively. Other clinically relevant GNB included Achromobacter spp., which persisted in 2.9% of cases and S. maltophilia, which showed similar rates of new, donor-independent detection and post-LuTX clearance. Interestingly, no donor-derived cases were observed for Achromobacter spp., E. coli, or Klebsiella pneumoniae (0/2, 0/17 and 0/12 donor-positive cases, respectively), indicating that recipients of donor lungs positive for these organisms did not demonstrate corresponding postoperative detection, whereas donor-associated detection was observed in 20 of 108 Candida-positive donor cases (18.5%). Clinical outcome ACR The median time to TBB after LuTX was 23 days (IQR, 15–37 days). Only TBBs conducted during the immediate postoperative hospital stay were included in the analysis. Retrieved biopsy samples were then analyzed according to the International Society of Heart and Lung Transplantation (ISHLT) grading scheme for the grading of pulmonary allograft rejection (10). ACR was diagnosed in 41 patients (11.9%), while the majority of recipients showed no evidence of rejection (A0: n=272; 78.8%). Mild rejection (A1) was observed in 26 patients (7.5%), whereas moderate (A2) and severe rejection (A3) were infrequent, occurring in 3 (0.9%) and 1 (0.3%) patients, respectively. Peri-transplant detection of Achromobacter spp. showed a signal for increased risk of ACR. While preoperative sputum isolates of Achromobacter spp. showed a non-significant trend toward ACR (P=0.07; OR 4.8), intraoperative isolates from native bronchial swabs were significantly associated with ACR in univariate logistic regression (P=0.03; OR 4.9) as depicted in Table 3. Similarly, donor-derived Klebsiella spp. conferred a markedly higher risk of ACR (P=0.03; OR 4.9). Donor-derived E. coli also emerged as a potential risk factor for ACR (Fisher’s exact test P=0.02), although the univariate regression model did not confirm significance (P=0.72; OR 1.3). However, in the multivariate model adjusted for baseline covariates, donor E. coli was identified as an independent predictor of ACR (P=0.01; OR 3.2), while donor-derived Klebsiella only showed a statistical trend towards significance (P=0.059; OR 4.9) and Achromobacter lost significance. Table 3Univariate analysis of clinical outcome by microbiological findingsSampling timepointMicroorganismACRCLADSurvivalPORPHRPHRIntraoperative bronchial swab Recipient Achromobacter spp. 0.034.9n.sn.sn.sn.s. Donor Achromobacter spp. n.sn.sn.sn.s0.00110.3 Escherichia coli 0.032.7<0.0013.6 Klebsiella spp. 0.034.9n.s.n.s.n.s.n.s.Postoperative sputum Klebsiella spp. n.s.n.s0.032.90.032.9BAL/ENTACandida spp.n.sn.sn.s.n.s.<0.0012.3ACR, acute cellular rejection; BAL, bronchoalveolar lavage; CLAD, chronic lung allograft dysfunction; ENTA, endotracheal aspiration; HR, hazard ratio; n.s., not significant; OR, odds ratio; spp., species. Univariate analysis of clinical outcome by microbiological findings ACR, acute cellular rejection; BAL, bronchoalveolar lavage; CLAD, chronic lung allograft dysfunction; ENTA, endotracheal aspiration; HR, hazard ratio; n.s., not significant; OR, odds ratio; spp., species. CLAD Overall, 76 patients (22.0%) developed CLAD during the follow-up period. Mean time to onset of CLAD was 2.7±2.0 years. In univariate analysis, postoperative detection with Klebsiella spp. was associated with an increased risk of CLAD (P=0.03; HR 2.9; Table 3) but this effect did not persist after adjustment for clinical covariates. In contrast, donor-derived E. coli was consistently predictive of CLAD, with univariate analysis showing a 2.7-fold increased hazard (P=0.03; HR 2.7). The association remained significant in the multivariate model (P=0.03; HR 2.6). Survival Impaired survival was observed in association with donor-derived GNB [Achromobacter spp. (P=0.001; HR 10.3), E. coli (P<0.001; HR 3.6)]. The same applied to Klebsiella spp.-positive post-transplant sputum samples (P=0.03; HR 2.9) and postoperative Candida spp. detected in BAL/ENTA samples (P<0.001; HR 2.3) (Table 3). Both de-novo (−/+) and persistent (+/+) Candida occurrence patterns were particularly detrimental (P=0.003 and 0.002; HR 2.2 and 2.1, respectively). Multivariate analysis confirmed most of these organisms as independent predictors of mortality [(−D+) Achromobacter spp. (P=0.01; HR 13.6), (−D+) E. coli (P<0.001; HR 3.2), Candida in BAL/ENTA samples (P=0.045; HR 1.9) and (−/+) Candida (P=0.045; HR 1.9)]. Associations with Klebsiella spp. were borderline significant (P=0.053; HR 2.2) after adjustment. To explore a potential contribution of microaspiration-related mechanisms, the occurrence of pyloric stenosis after transplantation was assessed. Pyloric stenosis was identified in 33 patients based on clinically indicated EGD. No significant association was observed between pyloric stenosis and postoperative microbiological findings; however, a non-significant trend was noted for E. coli detection in postoperative sputum samples (P=0.08; OR: 2.1). Discussion Over the past three decades, LuTX has been established as a life-saving therapy for patients with end-stage lung disease, significantly improving quality of life after years of chronic illness and long-term oxygen therapy (8). However, the success of LuTX is influenced by numerous factors, including donor organ availability, immune-mediated rejection, and infectious complications. Despite advances in immunosuppressive therapy and antimicrobial prophylaxis, infections remain a major threat to post-transplant survival. Unlike other solid organ transplants, the lung allograft is continuously exposed to environmental microorganisms, making bacterial, fungal, and viral infections a frequent occurrence (11-13). Several factors contribute to this vulnerability, including the disruption of small bronchial vessels, denervation, and impaired mucociliary clearance following transplantation (14,15). These physiological changes facilitate microbial colonization and increase the risk of infections, particularly in the early postoperative period. In former days, early infection after transplantation was found to be associated with an elevated risk for the occurrence of pneumonia, posing a substantial threat to postoperative survival (16). Studies performed on canine lung recipients suggested that clinically silent contamination of donor lungs progresses into pneumonia and it was deduced that prophylactic antibiotic treatment of human donors may decrease the risk of early bacterial pneumonia (17). Nowadays, anti-infective prophylactic regimens have been implemented. However, the composition and administration vary considerably depending on the transplant center, with each center pursuing its own strategy. Our study, which included a large cohort of 345 LuTX-recipients, provided a comprehensive analysis of microbial detection patterns and dynamics before, during, and after LuTX. To our knowledge, few studies have analyzed longitudinal microbiological data at multiple perioperative time points in a similarly sized patient population. The dominant microbial species identified in preoperative sputum samples—Candida spp., P. aeruginosa, and Staphylococcus spp.—are consistent with findings from previous studies (18). As some patients received pre-transplant immunosuppressive therapy, we investigated whether this may contribute to shaping the airway microbiological profile. In our cohort, only selected agents were associated with specific microbial patterns, with azathioprine linked to Candida spp. detection and cyclophosphamide to P. aeruginosa. This suggests that immunosuppression does not uniformly increase microbial burden but may selectively influence microbial composition depending on the agent and underlying disease. Importantly, these findings relate to pre-transplant immunosuppressive therapy and are therefore not influenced by post-transplant immunosuppressive regimens such as triple therapy. The observed associations likely reflect agent-specific effects and/or underlying disease characteristics rather than cumulative immunosuppressive burden. Notably, CF patients exhibited a distinctly different microbial profile compared to other transplant indications, characterized by a markedly higher prevalence of gram-negative bacteria, particularly P. aeruginosa and Achromobacter spp. This pattern was consistently observed across sampling modalities, including both preoperative sputum and intraoperative bronchial swabs, where CF patients showed a substantially higher burden of gram-negative organisms (e.g., P. aeruginosa 28.8% in sputum and 56.9% in bronchial swabs vs. 4–7% in COPD and ILD patients; Achromobacter spp. 11.5% vs. 0–1%). These findings are consistent with the known microbiological landscape of CF, shaped by chronic airway colonization and repeated antimicrobial exposure. Interestingly, despite the immunosuppressive state and impaired mucociliary clearance in transplant recipients, the relative proportion of positive postoperative sputum cultures was lower than in preoperative samples. This finding may be attributable to the routine administration of broad-spectrum antimicrobial prophylaxis, which likely suppresses bacterial and fungal growth in the immediate post-transplant period. Previous research has highlighted the importance of airway microbiota in influencing allograft outcomes, demonstrating that donor and recipient microbial profiles can impact post-transplant infections and allograft function (19-23). Our findings confirm the negative influence of several known facultative bacterial pathogens while also identifying fungal species as significant contributors to adverse post-transplant outcomes. Among GNB, Achromobacter spp. in intraoperative samples showed signals for an association with acute rejection in univariate analyses. These associations, however, did not persist after multivariate adjustment. By contrast, donor-derived Achromobacter was one of the strongest independent predictors of impaired long-term survival, underscoring its clinical relevance for mortality rather than acute rejection, contrary to what had previously been suggested by Crone et al. (24). Donor-derived E. coli emerged as the most consistent and independent microbial predictor across outcomes, being significantly associated with ACR, CLAD and long-term survival in multivariate analyses. Affected recipients exhibited higher mortality rates, primarily due to MOF and sepsis. These results are consistent with previous studies indicating that Enterobacterales contribute to post-transplant pneumonia and hospital acquired pneumonia (25,26). These robust associations suggest that E. coli occurrence may not only serve as a marker of donor lung injury but also contribute actively to immunological and infectious pathways driving allograft dysfunction. While the exact mechanisms remain unclear, previous work supports the immunomodulatory potential of E. coli, as suggested by Kim et al., who reported E. coli-induced upregulation of proinflammatory cytokines in an in vitro model (27). Notably, we detected no postoperative donor-derived E. coli, suggesting that standard antibiotic prophylaxis may effectively prevent direct bacterial transmission. This interpretation is further supported by our findings that the identified E. coli strains were susceptible to meropenem and that therapeutic levels of the antibiotic were confirmed in blood tests during the prophylactic period, suggesting that poor outcomes were not due to inadequate antimicrobial coverage. These results imply that E. coli detection may reflect donor-related lung injury or exert indirect immunomodulatory effects, rather than directly causing active infection. Alternatively, the presence of E. coli in donor bronchial cultures may serve as a marker of preexisting lung injury rather than an active infection. Factors such as trauma, aspiration, infection or prolonged mechanical ventilation prior to organ retrieval likely contribute to microbial colonization. Furthermore, brain death itself has been shown to induce the upregulation of proinflammatory cytokines in donor lungs, potentially exacerbating the inflammatory response (8,28,29). Given these considerations, the administration of systemic and inhaled antibiotics at the time of brain death diagnosis may improve donor lung quality and ultimately enhance transplant outcomes. Our findings provide a new perspective by emphasizing the microbial role of the allograft itself, demonstrating that E. coli-positive donor lungs are associated with increased dysfunction and poorer long-term survival. Similar to findings from Abdulqawi et al. [2024], who reported increased CLAD and early mortality with multidrug-resistant Klebsiella pneumoniae, we observed that postoperative Klebsiella detection was associated with higher risks of CLAD and reduced survival in univariate analyses (30). However, these associations did not remain significant in the multivariate model. Donor-derived Klebsiella only showed a statistical trend towards ACR, suggesting that it may act as a cofactor in allograft dysfunction and mortality, but not as an independent driver. Although P. aeruginosa is a well-known pathogen in immunocompromised patients, our study did not identify a direct association between P. aeruginosa and CLAD, contradicting earlier findings that linked P. aeruginosa to an increased risk of BOS and allograft dysfunction (31,32). However, our data did reveal that patients with a high LAS were more likely to be P. aeruginosa-positive samples in both the upper and lower respiratory tract preoperatively. This suggests that critically ill patients awaiting transplantation are at higher risk of P. aeruginosa detection, though its direct role in post-transplant complications remains unclear. Fungal infections are a well-documented complication in lung transplant recipients, particularly affecting the bronchial anastomosis due to poor vascularization in this region (33,34). We demonstrated that especially de novo detection of Candida spp. after LuTX was a major risk factor for reduced survival. Additionally, Candida-positive BAL/ENTA-samples were associated with higher MOF/sepsis-related mortality (P=0.004; OR: 3.1). The impact of fungal colonization after LuTX on postoperative outcomes remains controversial, with conflicting evidence in the literature, particularly regarding Candida spp. (35,36). In our cohort, the observed association with impaired survival was limited to lower airway samples (BAL/ENTA), whereas sputum detection was not associated with adverse outcomes (37). This discrepancy is likely attributable to sampling characteristics, as sputum may not reliably reflect the microbiological composition of the lower respiratory tract and may therefore underestimate clinically relevant fungal detection. Furthermore, these findings raise questions regarding the adequacy of current antifungal prophylaxis, as existing strategies may not fully prevent fungal detection in the lower airways and potentially fail to capture patients at increased microbial risk. However, as Dhanani et al. have previously pointed out, fungal colonization may also represent a marker of disrupted airway microbiology or increased clinical vulnerability rather than a direct causal factor in adverse outcomes (38). A detailed species-level analysis of Candida was not consistently feasible, as microbiological reports did not uniformly distinguish between Candida subspecies and were in part reported as Candida spp. without further specification, limiting interpretation of species-specific differences in pathogenicity and antifungal susceptibility. Furthermore, the lack of systematic antimicrobial resistance analysis may have influenced the observed associations, as appropriate treatment of susceptible strains may have attenuated the observed associations, thereby potentially leading to an underestimation of the true effect size. To account for potential microaspiration-related influences, we evaluated pyloric stenosis as a surrogate marker. Although no significant association with postoperative microbiological findings was observed, a trend toward increased E. coli detection suggests that aspiration-related mechanisms may contribute to microbial patterns in selected cases. However, this observation should be interpreted with caution, as the analysis was exploratory and does not allow conclusions regarding causality. In addition, as pyloric stenosis was assessed based on clinically indicated investigations, the findings may preferentially reflect a symptomatic subgroup and may not be fully generalizable to the broader transplant population. It is therefore more likely that pyloric dysfunction represents one of several contributing factors within a multifactorial process influencing post-transplant microbial dynamics. The same applies to potential influence of uni- or bilateral LuTX. The retained native lung frequently harbors chronic infections or a dysbiotic microbiota. Functioning as a persistent bacterial reservoir, the native lung may facilitate the translocation of opportunistic pathogens into the newly transplanted lung. Structural abnormalities, including bronchiectatic changes, mucus plugging and impaired mucociliary clearance, further predispose the native lung to act as a source of aspiration, thereby introducing colonized bacteria into the allograft. Existing literature highlights that pathogens such as P. aeruginosa, Achromobacter spp., S. maltophilia, and Burkholderia spp. are frequently isolated in patients with chronic lung diseases and have been implicated in allograft colonization, potentially exacerbating post-transplant outcome (39-41). However, our findings do not support an increased detection rate of these microorganisms following unilateral LuTX. This study was designed to provide a detailed overview of the potentially harmful microbiological spectrum affecting outcomes and long-term survival after LuTX. A key strength lies in the comprehensive screening approach, capturing microbial dynamics at multiple perioperative stages, pre-, intra- and postoperatively, which to our knowledge has not been systematically reported in a similarly sized cohort. This multi-timepoint design allows for a nuanced analysis of microbial dynamics across the transplant trajectory. However, certain limitations should be considered. As a retrospective, single-center study, the findings may reflect center-specific practices and may not be universally generalizable. While our primary focus was on bacterial and fungal pathogens, co-infections were not systematically analyzed and viral pathogens such as cytomegalovirus and herpes simplex virus were not included. The timing of preoperative sampling varied slightly between individuals, which may have influenced microbial detection at the time of transplantation. In addition, data on concomitant chronic sinusitis, which may represent an additional reservoir for microbial persistence and recolonization, particularly in patients with CF, were not systematically available, and no sinonasal sampling was performed, representing a limitation of the present study. Furthermore, molecular strain typing and resistance profiling were beyond the scope of this analysis, limiting detailed characterization of pathogen persistence. Although microbiological profiles differed between sampling modalities, analyses were intentionally based on a consolidated approach to capture peri-transplant microbial dynamics across compartments and time points, prioritizing clinically relevant microbial trajectories over compartment-specific differences. Importantly, given the retrospective design of this study, microbiological findings were evaluated post hoc and did not systematically influence immunosuppressive management. During the study period, no standardized adjustments of immunosuppressive therapy based on microbiological detection were applied, as treatment decisions were primarily guided by clinical parameters such as rejection status. This reduces the likelihood that microbiology-driven changes in immunosuppression confounded the observed associations. One aspect that needs to be addressed is the study’s inability to clearly differentiate between detection of microbiota and active infection. Due to the retrospective design, microbiological findings cannot be definitively classified as true pathogens, colonizing organisms or antibiotic-selected artifacts and should therefore be interpreted as indicators of microbial presence rather than confirmed infection. Accordingly, these findings are associative rather than causal and may reflect underlying donor lung injury or host susceptibility rather than direct pathogenic effects. Given the multifactorial nature of post-transplant inflammation, we did not incorporate systemic infection markers such as C-reactive protein, leukocyte counts or interleukin-6, which can be elevated in the context of rejection, surgical trauma or infection. We chose to focus on microbiological findings rather than clinical infection diagnoses to ensure consistent and objective assessment across all patients. However, we acknowledge that this approach may overestimate the pathogenic role of certain organisms in the absence of corroborating inflammatory markers or clinical signs of infection. In addition, no systematic analysis of multidrug-resistant gram-negative organisms was performed. For instance, Stenotrophomonas maltophilia, a bacterium with high intrinsic resistance, including to carbapenems such as meropenem, may be selectively detected under antibiotic pressure, yet it often does not cause clinically relevant infection. Furthermore, Aspergillus was only assessed by direct microbiological detection and no additional fungal biomarkers such as galactomannan were included. Therefore, the true incidence of Aspergillus occurrence or infection may have been underestimated. Conclusions In conclusion, systematic microbiological surveillance across all peri-transplant stages should be considered an integral part of clinical management in LuTX. Our findings highlight donor-derived E. coli and Achromobacter as well as postoperative Candida in BAL/ENTA samples detection as robust predictors of adverse outcomes. These results emphasize the need for targeted screening of donor lungs and early postoperative airway specimens to promptly identify high-risk patterns. While standard prophylaxis already lowers microbial detection rates, further optimization is warranted. Clinically, this may include preoperative decolonization protocols, a shift towards more individualized antimicrobial prophylaxis and moving away from uniform treatment durations, while increasingly incorporating intraoperative microbiological findings to guide duration and spectrum of therapy, as well as allowing adjustment to center-specific resistance profiles and tailoring of immunosuppressive regimens in patients with high microbial risk. However, such strategies must be balanced against potential risks, including increased adverse drug events, higher healthcare costs, and the promotion of multidrug-resistant organisms, underscoring the need for careful, individualized implementation. Close microbiological monitoring can thereby support early therapeutic interventions, reduce the risk of allograft dysfunction and ultimately improve long-term survival. Supplementary The article’s supplementary files as Acknowledgments A preliminary version of this study was included in the doctoral thesis of the first author (Maximilian Vorstandlechner; MD), submitted to LMU Munich in 2025. The current manuscript presents substantially expanded analyses, including detailed microbiological profiling and clinical outcome associations, which have not been previously published or peer-reviewed. During the preparation of this manuscript, the authors used ChatGPT (OpenAI, San Francisco, CA, USA) exclusively for language editing and refinement. The tool was not used for data analysis, statistical evaluation or interpretation of results. The authors reviewed and edited the manuscript in full and accept complete responsibility for its scientific content. All use adhered to applicable ethical and institutional guidelines. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional review board of LMU University Hospital, LMU Munich, Munich, Germany (project No. 22-0466). Owing to the retrospective design of the study and the absence of any additional interventions or diagnostic procedures, the requirement for patients’ informed consent was waived. Footnotes Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0260/rc Funding: None. Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0260/coif). S.M. reports that he received consulting fees from Berlin Heart GmbH and serves in a leadership or fiduciary role on the Standards and Guidelines Committee of the ISHLT. The other authors have no conflicts of interest to declare. Data Sharing Statement Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0260/dss References Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0260/prf Supplementary Materials The article’s supplementary files as Data Availability Statement Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0260/dss