Performance of bilateral bronchoalveolar lavage during routine transplant surveillance bronchoscopy changes management

Abstract

Introduction

Surveillance bronchoscopy for lung transplant monitoring is a standard of care at most high-volume transplant centers, though methods vary based on institutional preferences. At our medical center, we perform bilateral bronchoalveolar lavage (BAL) and unilateral transbronchial biopsy. This study aims to determine how often bilateral BALs lead to a change in medical management.

Methods

This is a single-center, retrospective cohort study enrolling 237 patients who underwent either bilateral or single lung transplant and who underwent 759 surveillance bronchoscopies with bilateral BALs from 2020 to 2025. Data collected included demographic data, cultured pathogens bilaterally, and pathology from unilateral biopsies. Different pathogens (DP) were defined by having two positive contralateral results with differing infectious agents, or a positive result in one lung with a negative result in the contralateral lung. Change in management was defined as either alteration of antibiotic coverage or immunosuppressive drugs.

Results

Eighty-three bilateral BALs (10.9%) identified DP. Of these, 43 discordant findings (5.7% of all procedures) resulted in a change in management. All 43 of these changes involved initiation or adjustment of antimicrobial therapy, and 3 (7.0%) involved an adjustment in immunosuppression. Common pathogens that prompted management changes were Pseudomonas aeruginosa (14/43 32.6%) and Aspergillus species (12/43 27.9%).

Conclusions

Bilateral BALs identifies clinically significant DPs in over 1 in 10 surveillance procedures, with more than half of these findings directly altering patient care. Because the most common DP identified—P aeruginosa and Aspergillus species—are strongly associated with chronic lung allograft dysfunction, the results support the practice of performance of bilateral BAL for post-lung transplant surveillance.

Background

Lung transplantation represents the final therapeutic option for end-stage lung disease, with over 2,800 performed annually in the throughout the world.¹ Despite significant advances in surgical techniques and immunosuppressive therapy, long-term outcomes remain limited by complications including acute cellular rejection, antibody-mediated rejection, and chronic lung allograft dysfunction (CLAD).² Infectious complications represent a major source of morbidity and mortality in lung transplant recipients, with pulmonary infections occurring in 35%-66% of patients and contributing up to 30% of deaths in the first year post-transplant.3, 4

Surveillance bronchoscopy with transbronchial biopsy (TBBX) and bronchoalveolar lavage (BAL) allows for early detection of rejection, infection, and other pulmonary complications before the development of clinical symptoms.5, 6 Current practice varies widely among transplant centers, with protocols ranging from symptom-driven bronchoscopy to regularly scheduled surveillance procedures at predetermined intervals, leaving considerable heterogeneity in clinical practice.7, 8

A critical yet insufficiently studied question in surveillance bronchoscopy is whether bilateral BAL sampling is required or if unilateral sampling yields comparable diagnostic accuracy. In bilateral lung transplants (BLT) recipients, the assumption of uniform pathology across both lungs may be flawed. Anatomic differences, variable allograft perfusion, aspiration risk, and heterogeneous immune responses can result in discordant infectious or inflammatory processes. In single-lung transplant recipients, studies show that allografts and native lungs differ in infection susceptibility, with fatal infections demonstrating greater histopathologic involvement of the allograft.⁹ This inter-lung variability provides a biological rationale for bilateral BAL sampling to ensure comprehensive pathogen detection.

BAL is an inherently safe procedure, even in patients within the first-year post-lung transplantation. The practice of performance of bilateral BAL minimally increases sedation time and fluid instillation volumes but may compound the risk of hypoxemia, bleeding, or bronchospasm.⁵ However, if bilateral sampling identifies clinically significant pathogens that would be missed by unilateral sampling, the benefits could substantially outweigh these risks. Studies in mechanically ventilated patients have shown that bilateral BAL detects discordant pathogens in 25%-37% of cases, but the clinical significance of bilateral sampling in lung transplant surveillance—specifically whether discordant findings lead to changes in patient management—remains poorly characterized.10, 11, 12

This study aims to characterize the clinical utility of bilateral BAL during surveillance bronchoscopy in lung transplant recipients by determining the frequency with which bilateral sampling identifies discordant pathogens between lungs and, critically, how often these findings lead to actionable changes in patient management. By analyzing a large cohort of patients undergoing protocol surveillance bronchoscopy at a high-volume transplant center, we sought to provide evidence-based guidance on whether bilateral BAL should be considered standard practice in the post-transplant surveillance protocol.

Methods

This is a single-center, retrospective cohort study performed at a tertiary referral center that performs >100 lung transplantations annually.

Immunosuppression and anti-infective protocol

Lung transplant recipients at our center receive induction immunosuppression with methylprednisolone and basiliximab followed by a 3-drug regimen including tacrolimus, mycophenolate mofetil, and steroid taper. Initial anti-infective protocol includes trimethoprim-sulfamethoxazole for PJP prophylaxis, posaconazole for fungal prophylaxis, and valganciclovir for CMV prophylaxis, if warranted.

Surveillance bronchoscopy protocol

Surveillance bronchoscopy at our institution is performed according to a scheduled protocol regardless of clinical symptoms or imaging findings. Chest imaging prior to scheduled surveillance bronchoscopy is not protocolized, and bilateral BAL is performed irrespective of the presence or absence of radiographic abnormalities. All surveillance bronchoscopies were performed by experienced interventional pulmonologists or lung transplant surgeons following standardized techniques to minimize contamination, including avoiding suctioning during scope advancement and thoroughly irrigating the working channel with sterile saline between sequential BAL samples from each lung. Institutional protocol mandates performance of bilateral BAL (with culture data, PCR, cell count/differential, cytology, fungal serologies, and other as needed tests sent). BAL was performed using standardized technique with the bronchoscope wedged in the target subsegmental bronchus, usually the right lower lobe and left lower lobe. Sterile saline was instilled beginning with an initial 100 mL divided into a 60 mL aliquot followed by 30 mL aliquot. Further 30 mL aliquots were instilled as needed, with a maximum total instilled volume of 300 mL per sampling site. The target minimum return volume was at least 30 mL per site, consistent with ATS guidelines recommending >30% return of instilled volume for optimal diagnostic yield. While our institutional protocol uses different aliquot sizing than the ISHLT-recommended 2×50 mL approach, it achieves comparable total instilled volumes (90-300 mL) and consistently meets guidelines for adequate return volume (>30 mL), thereby ensuring adequate sampling of both airway and alveolar compartments for microbiologic analysis. Samples were sent for bacterial culture, fungal culture, viral PCR panel, and galactomannan testing as clinically indicated. Unilateral TBBX was performed with forceps to obtain >7 adequate tissue samples as determined by our pathologists.

Patient population

Patient data was obtained retrospectively from our institutional Lung Transplant Biorepository database, a prospective real-time enrollment of subjects who are undergoing or have undergone lung transplantation, for the collection of blood and biopsy samples. The biorepository is institutional review board approved (IRB #00212120), and patients provide informed consent to the collection of biospecimen and clinical data. Each patient undergoing lung transplantation at our institution is approached to enroll and provide informed consent for this prospective registry. Data was collected from eligible patients who underwent surveillance bronchoscopy from November 2020 through January 2025. The data was collected in a combination of already extracted data into a secure REDCap database,¹³ as well as manual review/chart extraction.

This dataset was refined to the following exclusion criteria:

• 1)Removal of patients enrolled but who did not receive a transplant,
• 2)Those without a surveillance bronchoscopy on record,
• 3)Those who only underwent a single lung BAL.

This resulted in the enrollment of 237 patients with a total of 759 bilateral BALs performed (full demographics in Table 1).

Table 1.

Proportion of Bilateral Bronchoalveolar Lavages That Isolated Different Pathogens With Proportion of Discrepant Results That Lead to Change in Management

	Total	% of Total
Total Bilateral BALs (BiBAL)	759
BiBAL, Different Pathogen	83	10.9%
BiBAL, Different Pathogen, Change in Management (CIM)	43	5.7%

Data collection and definitions

Each record was examined individually and cataloged for microbiologic data from cultures and galactomannan. Lavage samples were marked as positive for the resulting pathogen if culture, multiplex PCR assay, or galactomannan testing resulted positive. Organisms were considered contaminants and excluded from analysis if they were recovered at concentrations below 10⁴ CFU/mL or represented typical oropharyngeal flora. Common oral flora and known contaminants such as Strep viridans and Candida species were not included as positive results (full list of excluded pathogens included in Supplement A). Each individual bronchoscopy was flagged as different pathogen (DP) if there were 2 positive results with differing pathogens or a positive result in one lung with a negative result in the contralateral lung.

Each instance of DP was flagged as “Change in Management” if the discovery of the pathogen led to a change in drug therapy as clinically adjudicated by 2 transplant pulmonologists. Changes in drug therapy were identified as one of the following: change or addition of antimicrobial drug or reduction in immunosuppression. Each DP instance was then reviewed by a second observer to ensure agreement.

Statistical analysis

Data were analyzed using the most recent versions of Stata (version 19, StataCorp, College Station, TX) and Microsoft Excel (Microsoft, Redmond, WA). Descriptive statistics were calculated for all variables, including means and standard deviations for continuous data and frequencies with percentages for categorical data. Group comparisons were performed using 2-tailed independent t-tests for continuous variables. A significance level of p < 0.05 was considered statistically significant. Excel was used for initial data cleaning, organization, and visualization, while Stata was employed for all inferential analyses to ensure accuracy and reproducibility.

Results

Two hundred thirty-seven patients underwent a total of 759 bilateral BALs. Patients enrolled in the database had an average of 3.2 bronchoscopies performed and cataloged.

Of the 759 bilateral BALs, 83 (10.9%) were flagged as DP. Of these, 43 (5.7%) led to a change in antimicrobial therapy, with 3 (7.0%) of those patients also having a change in immunosuppression (Table 1). These management changes occurred across 34 individual patients (Figure 1). This represents a number needed to harm of 18 bronchoscopies (1/0.057).

Figure 1.

Flow chart illustrating the workflow of chart abstraction and analysis of data for this retrospective study.

Of the 83 different patients with “DP” identified, the most common organisms were: Pseudomonas aeruginosa (18, 21.7%), followed by Aspergillus (15, 18.1%) and Rhinovirus (12, 14.5%) (Chart 1). Of the 43 patients with organisms that led to a change in management, again P aeruginosa (14, 32.6%) was the most common, followed by Aspergillus (12, 27.9%) and “Other” (9, 20.9%) (Figure 2; full list of organisms contained in “Other” included in Supplement A).

Figure 2.

Commonly isolated pathogens from bilateral bronchoalveolar lavage.

Of the 43 pathogens that led to change in management, 17 (39.5%) were in single lung transplants (SLT) and 26 (60.5%) were in BLT. Of the 17 SLTs, the pathogen that led to a change in management was found in the transplanted lung in 11 (64.7%) cases. The proportion of different pathogen patients did not differ significantly across SLT (17/75, 22.7%) or BLT (26/162, 16.0%).

Discussion

This study represents the largest single-center analysis to date examining the clinical utility of bilateral BAL during surveillance bronchoscopy in lung transplant recipients. Our findings demonstrate that bilateral BALs identifies discordant pathogens between lungs in 10.9% of procedures, with 5.7% leading to changes in clinical management. This is a number needed to change the management of 1 in 18 bronchoscopies. These results support bilateral sampling as standard practice in post-transplant surveillance protocols.

The 10.9% rate of discordant pathogen detection in our cohort is notably lower than that described in ventilator-associated pneumonia, where discordance rates of 25%-37% have been reported.10, 11 This difference is not surprising given the distinct clinical contexts involved; ventilator-associated pneumonia involves acute, fulminant processes, whereas our surveillance cohort comprised mostly asymptomatic recipients with lower, more heterogeneous pathogen burdens.

The clinical significance of our findings is underscored by the fact that 5.7% of all bilateral BALs led to management changes after bronchoscopy results based on discordant organisms. This represents a substantial proportion of surveillance procedures where bilateral sampling directly impacted patient care through initiation of antimicrobial therapy, adjustment of existing regimens, or modification of immunosuppression. Given that lung transplant recipients at our institution typically undergo multiple surveillance bronchoscopies in the first post-transplant year and beyond, the cumulative impact of bilateral sampling on individual patient outcomes may be considerable.

Pseudomonas aeruginosa emerged as the most frequently identified discordant pathogen in our study. This finding has clinical relevance given the well-established association between P aeruginosa colonization and the development of bronchiolitis obliterans syndrome (BOS). Multiple studies have demonstrated that allograft colonization with Pseudomonas is strongly associated with subsequent BOS development, with colonization typically preceding BOS diagnosis by a median of 204 days.14, 15 If BAL sampling was solely unilateral, multiple important pathogens may be missed as we strive to prevent BOS.

Aspergillus species represented the second most common discordant pathogen in our cohort and frequently prompted management changes. The relationship between Aspergillus colonization and CLAD remains somewhat controversial in the literature, with conflicting data regarding its impact on BOS development. Despite this, most transplant centers adopt a strategy of surveillance for Aspergillus with consideration of antifungal prophylaxis or treatment in colonized patients, particularly given its association with invasive aspergillosis in immunosuppressed recipients.¹⁶ Our finding that Aspergillus was frequently detected in only one lung suggests that unilateral sampling could miss clinically significant colonization.¹⁶

Respiratory viruses, particularly rhinovirus, comprised the third major category of discordant pathogens in our study. Community-acquired respiratory viral infections have been increasingly recognized as important contributors to both acute rejection and chronic allograft dysfunction in lung transplant recipients, with 16% of patients developing acute rejection within 3 months of infection and 18% developing an FEV1 drop consistent with BOS.17, 18 Importantly, both symptomatic and asymptomatic viral infections can trigger acute rejection and BOS, emphasizing the importance of comprehensive surveillance sampling.¹⁹Interestingly, this finding in our study goes in contrary to recently published literature on the detection of viral infection on bilateral BAL specimens, where discrepant results were found in only 3.6% of the cases.²⁰

The biological rationale for discordant pathogen distribution between lungs is supported by recent evidence demonstrating differential susceptibilities to infection in allograft versus native lungs. One detailed histopathologic analysis of fatal infections in single-lung transplant recipients and found that allografts consistently showed more extensive infection than native lungs.⁹ This differential susceptibility likely reflects multiple factors, including denervation-related impaired mucociliary clearance, altered local immune responses, anastomotic complications, and ischemia-reperfusion injury affecting the allograft.3, 21 Even in BLTs, heterogeneous distribution of these risk factors between the two grafts could theoretically lead to asymmetric pathogen colonization.

Our exclusion of common oral flora and known contaminants such as S viridans and Candida species from the definition of "positive results" was intentional and clinically appropriate. These organisms frequently represent oropharyngeal contamination rather than true lower respiratory tract infection or colonization. However, this conservative approach may have led to underestimation of the true discordance rate, as some excluded organisms could represent genuine differences in airway colonization patterns between lungs. Future studies employing quantitative culture thresholds or molecular techniques to distinguish colonization from contamination may provide additional insights.

Limitations of this study include its retrospective, single-center design, which restricts generalizability to centers with different surveillance protocols or patient populations. Additionally, we used a non-exhaustive cohort that included only patients that provided informed consent for data collection and excluded a significant proportion of our lung transplantation cohort. Further, we did not systematically collect data on the specific location (right vs. left lung, upper vs. lower lobe) of BAL sampling, which could have provided insights into anatomic patterns of pathogen distribution – though our established protocols utilize the lingula and right middle lobe as lavage sampling sites most of the time. Additionally, our definition of "change in management" was based on documentation in the medical record and clinical judgment, introducing potential subjectivity that we attempted to mitigate by having two providers confirm the change. We did not analyze long-term outcomes such as BOS development or survival based on whether discordant organisms were detected and treated, which represents an important area for future investigation. Finally, we did not perform quantitative cultures or use standardized definitions for colonization versus infection, relying instead on clinical interpretation and provider decision-making regarding treatment to reflect real clinical practice patterns. We also did not collect safety data of the performance of a second BAL, though the overall complication rate from a single BAL is so low, this was not felt to be warranted and would be confounded by the concomitant TBBX.

The absence of long-term outcome data represents an important limitation of our study. Future studies should correlate bilateral BAL findings with these outcomes through prospective, multicenter registries with standardized surveillance protocols and extended follow-up. Cost-effectiveness analyses incorporating both short-term management changes and long-term outcomes would provide valuable data to inform practice guidelines.

Despite these limitations, our study provides compelling evidence that bilateral BAL during surveillance bronchoscopy identifies clinically significant pathogens that would be missed by unilateral sampling in approximately 11% of cases, with more than half of these findings prompting changes in patient management. The organisms most frequently identified—P aeruginosa, Aspergillus species, and respiratory viruses—are precisely those with established associations with CLAD and poor long-term outcomes. While bilateral BAL slightly increases procedural time, laboratory cost and effort, and perhaps risk, these considerations must be weighed against the potential benefit of early pathogen detection and intervention to preserve allograft function.

Disclosure statement

The authors do not have relevant disclosures.

Supplement A

Excluded Organisms

Strep viridans

Strep intermedius

Kluyvera

Candida species

Stomatococcus

Coagulase Negative Staph

Alternaria

Geotrichum

Saccharomyces cerevisiae

Granulicatella

Sophingobacterium spiritivorum

Raoultella planticola

Pichia kudriavzevii

Pathogens Included in "Other"

Burkholderia

Stenotrophomonas

Fusarium species

Acinetobacter

Paecilomyces

Actinomyces

Verticillium

Rhizopus

Citrobacter

Financial support

The authors declare that this research received no external funding. No financial support was provided by any organization, institution, or sponsor for the design, execution, analysis, or publication of this study.

Conflicts of Interest statement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.