Systematic fungal biomarker or polymerase chain reaction testing in lung transplant recipients: Retrospective analysis to optimize their use

Abstract

Background:

Among lung transplant recipients, serial bronchoscopies are performed frequently. Often, serial galactomannan (GM), 1,3-β-d-glucan (BDG), and Pneumocystis jirovecii (PJ) testing is performed with these broncho-alveolar lavages (BALs) as standard of care with limited data to support their routine use.

Methods:

After Institutional Review Board approval, we retrospectively collected all blood and BAL GM, BDG, and PJ test results from January 2015 to July 20, 2022. Primary data collection from the Northwestern Medicine EDW was supplemented by manual chart review.

Results:

During the study period, 236 lung transplant recipients were cared for by our center. Of these patients, 217 (91.9%) had 1418 GM tests performed; 61 (4.3%) were positive (index ≥1). Fungal cultures were requested for most BAL-GM (90.7%). Out of duplicates in same BAL, results discrepancy was minimal (3.4%). 172 (72.9%) had BDG tests were performed; 25.6% were positive. Thirteen patients had multiple BDG during one hospitalization (mean 2.3 tests); none of the negative test repeated became positive. Eleven negative BDG were seen in patients with invasive aspergillosis (IA). Note that, 577 PJ testing were performed (direct fluorescent antibody [n = 494] or polymerase chain reaction [PCR] [n = 80], or both [n = 3]) in 174 different patients. None were positive.

Conclusion:

Despite supplemental GM, BDG, and Pneumocystis jirovecii pneumonia PCR being performed routinely on lung transplant recipients undergoing BAL at our center, the data suggests a more tailored approach may be appropriate. There is no role for routine serial testing with these assays during a single hospitalization. BDG confers no added-value over GM with cultures for IA diagnosis.

1 |. BACKGROUND

Sequential broncho-alveolar lavages (BAL) and serum biomarkers are used as fungal microbiologic surveillance tools after lung transplantation. With the availability of commercial enzyme-linked immunosorbent assay tests and the development of specific in-house polymerase chain reaction (PCR), reflex collection of these supplemental tests has widened adding to the total cost of care and discrepant results. As such, reflex collection of galactomannan (GM) in any BAL is now frequent and seen as a preemptive tool for early diagnosis of aspergillosis pneumonia.¹ Serum 1,3-β-d-glucan (BDG) is notable for being resource-consuming and its lack of specificity questions the utility of a positive result especially in the asymptomatic recipient.² Pneumocystis jirovecii (PJ) qualitative or quantitative PCR is also being increasingly used and has now replaced direct fluorescent antibody (DFA) in many centers because of its high sensitivity despite variable specify for clinically significant disease in the absence of a predefined cutoff value for heterogeneous respiratory specimens.^3–5 The yield of these tests is known to be markedly reduced among patients who are receiving active preventative therapy.^2,6 Lastly, all diagnostic tests have reduced positive predictive value when the pre-test probability is low, as would be the case of asymptomatic patients with non-typical findings on imaging.⁷

At our center, serum and BAL GM, serum BDG, and BAL PJ PCR have become standard whenever a BAL is performed among our lung transplant recipients. This study was conducted to estimate the diagnostic yield of these supplemental tests. We hypothesized that systematic or repeat GM and PJ testing in surveillance BAL is rarely associated with diagnosis of clinically significant disease and that BDG testing does not add value in the diagnosis of invasive fungal infection (IFI) over fungal stains, culture, or GM.

2 |. METHODS

After Institutional Review Board approval, we retrospectively collected all GM and PJ DFA or PCR performed on BAL; and all GM and BDG performed on serum from adults lung transplant recipients at our large academic center from January 2015 – July 20, 2022. Primary data was abstracted from the Northwestern Medicine Enterprise Data Warehouse and then manual chart review was used to inform clinical likelihood of infection.⁸ The following data were included in extraction: demographics, date of transplantation, prophylactic antimicrobials used by the patient within 30 days of test, clinical laboratory data including date and result of respiratory cultures and biomarkers or PCR, and microbiological data including organism identification. Local practices at our institution include 6 months of universal mold-active azole prophylaxis (voriconazole until 2016 and posaconazole thereafter) and lifelong TMP-SMX (or atovaquone). As part of a surveillance protocol, our lung transplant recipients undergo bronchoscopy at 1, 3, 6, 9, and 12 months following transplantation; all of which include a GM (including those performed on posaconazole prophylaxis) and a PJ PCR. Reported IFI diagnoses were based on the EORTC/MSGERC revised consensus definitions.⁹

The primary objective was to evaluate if systematic GM, BDG, and PJ testing lead to the diagnosis of clinically significant fungal infections regardless of bronchoscopy indication. To do so, a correlation between conventional fungal cultures and clinical diagnoses was attempted. The secondary objective was to describe the context in which these tests are prescribed at our institution (timing post-transplantation and patient setting) and document repeat testing and their added value. Descriptive statistics were used to describe incidence rates. Cost estimates of the laboratory tests were based on local pricing of referral testing with Quest Diagnostics.

3 |. RESULTS

Between January 2015 and July 20, 2022, 236 adult lung transplant recipients were cared for by our program, for which we collected more than 800 different BAL specimens.

3.1 |. Galactomannan

A total of 1418 GM tests were performed in-house during our study period in 217 different lung transplant recipients; 1267 (89.4%) from BAL and 151 (10.6%) from serum. The 1267 tests performed on BAL were from 884 different bronchoscopy procedures; one GM from the right lung and one from the left lung were sent in almost half of the procedures (383/884; 43.3%). GM testing was performed at a mean of 8.5 months after transplantation; 60.6% (859/1418) were performed at the time of mold-active prophylaxis. Sixty-one (61/1418; 4.3%) GM were positive (index ≥1), all from BAL specimens except one, in 35 different recipients. 59.0% (36/61) of the positive results were seen in 23 recipients during ongoing antifungal prophylaxis; 20 of them had azoles dosing with only three being below the suggested level for prophylactic activity. The positive serum GM was seen in a neutropenic patient with concomitant positive BAL GM and fungal cultures growing Aspergillus spp. 5.8% (51/884) of the bronchoscopy procedures had at least one positive GM. Out of BAL with duplicates (e.g., right and left lungs) a discrepancy (positive and negative) was seen in 3.4% (13/378). Among all positive GM, the mean quantitative result was 2.6. The laboratory cost associated with this set of 1418 GM tests represents approximately USD 53,884 (unit pricing estimate $38).

Fungal smears and cultures were requested for most BAL for which a GM was ordered (802/884; 90.7%) and were positive for mold in 21 incidences including nine Aspergillus spp. strains. BAL GM was positive in seven of the nine BAL-growing Aspergillus spp; the two negative GMs were seen in patients without mold-active azole at the time of testing. Among recipients with positive BAL GM and Aspergillus spp. positive cultures; 57.1% (4/7) were on voriconazole at the time of testing. A significant number of BAL had GM testing without fungal culture (82/884; 9.3%).

Clinical diagnoses of invasive aspergillosis (IA) including possible and probable cases were made in 63 of our recipients during our study period (63/236; 26.7%). Of those, 35 had positive BAL GM (35/63; 55.6%) with the remaining probable diagnoses based on positive cultures. Possible cases included those with a GM index between 0.5 and 1 (not meeting the mycologic criteria) in combination with radiological features. Out of the IA diagnoses with positive GM; 62.9% (22/35) were receiving mold-active azole at the time of testing.

3.2 |. β-D-Glucan

We recovered a total of 172 serum BDG tests (Fungitell (1–3)-ß-D-Glucan Assay) among our cohort (Quest Diagnostics). These were recovered from 90 different patients, at a mean of 15.1 months after transplantation. One-quarter of all tests (44/172; 25.6%) were positive with a mean value of 320 pg/ml, with 27.3% (12/44) being above our upper limit of detection of 500 pg/ml. The positive tests were recovered from 24 different recipients. Three tests were invalid because of interfering substances and one was indeterminate. Thirteen patients had serial testing during the same hospitalization; with a mean of 2.3 tests. None of the negative BDG tests repeated during the same stay became positive. The total laboratory-associated cost for the 172 tests is approximately USD 12,900 (unit pricing estimate $75).

Most BDG testing was requested for hospitalized patients (127/172; 73.8%) with most being admitted to the ICU (88/127; 69.3%) and on mechanical ventilation (67/88; 76.1%). Among the testing of hospitalized patients, the mean duration of stay was 16.3 days at the time of BDG testing. Among outpatient BDG testing, 24.4 % (11/45) were positive, compared to 20.5% (8/39) for those performed in patients admitted to the floor and 28.4% (25/88) to the ICU.

Out of all the IA diagnoses, 19 had a BDG test within one month of the IA diagnosis with 5 of them being positive (5/19; 26.3%). Three cases of IA had a positive BDG while having a negative BAL GM but all 3 had positive Aspergillus spp. cultures. Eleven negative BDGs were seen in patients with IA. Out of those IA cases with negative BDG, four were receiving a mold-active azole at the time of BDG testing (4/11; 36.4%). In the remaining cases of positive BDG (39), three were in the context of fungemia, and 13 of yeast lung colonization.

3.3 |. PJ testing

We recovered 577 PJ assays during our study period (DFA (n = 494) or PCR (n = 80), or both (n = 3)). Of note, PJ qualitative PCR was implemented as a replacement for DFA at our institution starting in May 2022. One DFA was inconclusive because of an unsatisfactory specimen quality and one PCR was indeterminate. None were positive and no clinical diagnosis of PJ pneumonia (PJP) was made during the study period; with most of our recipients receiving lifelong PJP prophylaxis. The associated costs are approximately USD 4940 and USD 4240 for DFA and PCR, respectively (unit pricing estimates $10 and $53), with PCR being up to 5 times more expensive than DFA.

Most PJ testing on BAL was requested when recipients were admitted to the ICU (403/577; 69.8%) and on mechanical ventilation (313/403; 77.7%) while 57 (57/577; 9.9%) were requested as outpatient. When documented, steroids were usually active at the time of PJ testing (497/528; 94.1%). Among hospitalized patients, the mean duration of stay was 31.2 days at the time of bronchoscopy and many had repeated testing with a mean of 1.8 tests per stay. One patient had 14 different BALs tested for PJ during a 4-month stay.

4 |. DISCUSSION

In this single-center study, serum GM, serum BDG, and routine PJP testing did not lead to a new diagnosis of IFI in addition to routine testing. We did not have documentation of indication for bronchoscopy (routine testing in asymptomatic patients, symptomatic patients with a range of clinical presentations), but we did not see an impact on the diagnostic yield of the supplemental testing for patients in the outpatient or inpatient settings. BAL GM testing was not positive in half of patients diagnosed with IA and most of the positive BDG results were not associated with proven IFI. These data suggest that supplemental GM is useful in a targeted approach and should not be repeated in the absence of new clinical data over the course of a hospitalization. BDG is likely not needed in most patients and PJ testing likely can be limited to symptomatic patients. Laboratory stewardship has the potential to be cost-effective without impacting the diagnosis of IFI in lung transplant recipients.

Bronchoscopy procedures after lung transplantation are performed for the surveillance or diagnosis of allograft dysfunction or rejection and respiratory infections.^10,11 The ISHLT consensus statement on BAL suggests that surveillance BAL at 1, 3, 6, and 12 months post-transplant is an acceptable procedure to identify asymptomatic infection or colonization, however, the utility of these has not been shown to reduce infection or mortality.^11,12 No recommendation is made after the first year post-transplantation. Specifically, serial GM from BAL has been proposed as a preemptive tool for early diagnosis of pulmonary aspergillosis among new lung transplant recipients but is now part of most BAL bundle orders regardless of timing after transplantation.^1,13 GM should always be performed with fungal smear and cultures, especially in the context of azole active prophylaxis given the possibility of false positive results or breakthrough infection in this context. This was identified as a missed opportunity in almost a tenth of our cohort for which no microbiological data was available. Possible explanations include clinicians’ accidental omission or lost specimens or insufficient volume. A specimen from an area of suspected infection (e.g., left specimen if left cavity) should be prioritized, and/or a single GM testing on a composite specimen (i.e., the pool of left and right specimens) could be explored to limit cost. Our study does not deviate from previous studies that documented the limited sensitivity of serum GM in the non-neutropenic lung transplant recipient.¹⁴ Only one positive serum GM was documented in a neutropenic patient with positive BAL GM, which further argues against its added value when a BAL GM was performed and resulted in negative.

With the exception of excluding PJP, the utility of serum BDG testing has been shown to be limited among lung transplant recipients because of its very low specificity.² In our study, BDG testing showed a low predictive value for the diagnosis of IA and other IFIs. Serum BDG appears to confer no added-value over BAL GM combined with fungal cultures for the diagnosis of IA and likely can be deferred in most cases. There is especially very limited diagnostic value in repeating a negative BDG tests during a stay and see no role in the outpatient setting. Given the poor clinical utility, it is difficult to justify the cost and technical time allocated to prepare the shipment as a referred or send-out test. Educational strategies and/or restrictions on ordering could be explored to optimize its use.¹⁵ Further, with the enhanced sensitivity of PCR, the utility of supplemental BDG would be further lessened.

Overall, diagnoses of PJP among lung transplant recipients are rare compared to other infectious complications, mostly because of the wide use of long-term effective prophylaxis. Our analysis revealed no positive DFA or PJ PCR among both acutely sick and stable recipients as almost all our cohort is receiving lifelong prophylaxis. Therefore, there is limited value for systematic PJ testing on BAL for the recipients at our institution outside of those patients with strong clinical suspicion of PJP. There is also likely little added value in repeating the same PJ testing after a negative result for a hospitalized patient; except when suspicion for PJP is high in which case a PCR would be more sensitive. We advocate for a tailored selection of patients for PJP testing.

Our study results are limited by their retrospective and descriptive nature. The data are highly influenced by local protocols including diagnostic approaches, tests availability, and antifungal prophylaxis use. We were not able to recover the indication (surveillance vs. diagnostic) of the discussed tests which influenced their yield, as does pre-test probability. The large number of analyzed tests however allowed us to better understand the prescription habits of the transplant teams at our institution, which is the first step in implementing targeted diagnostic stewardship strategies. Cost analyses were estimated based on our center’s specific data and are limited by variations in patient billing and among different reference labs.

Given the widespread use of EMR, standard testing panels are increasingly common and likely to result in suboptimal testing strategies. Expensive, novel tests are often ordered while stains and cultures are often underestimated or forgotten. In the context of the diagnosis of fungal infections among lung transplant recipients, we emphasize the need to prioritize cultures when feasible and interpret biomarkers based on clinical presentation and risk factors. Additional testing should be personalized based on our understanding of the host’s net state of immunosuppression, prior antifungal exposure, and site(s) of infection. There is a clear role for laboratory stewardship to optimize diagnostic yield in lung transplant recipients.

List of Abbreviations:

BAL

broncho-alveolar lavage

BDG

1,3-β-d-glucan

DFA

direct fluorescent antibody

GM

galactomannan

IA

invasive aspergillosis

IFI

invasive fungal infection

PCR

polymerase chain reaction

PJ

Pneumocystis jirovecii

PJP

Pneumocystis jirovecii pneumonia

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.