Early posttransplant spirometry is associated with the development of bronchiolitis obliterans syndrome after allogeneic hematopoietic cell transplant

Abstract

Bronchiolitis obliterans syndrome (BOS) after allogeneic hematopoietic cell transplantation (allo-HCT) is often diagnosed at a late stage when lung dysfunction is severe and irreversible. Identifying patients early after transplant may offer improved strategies for early detection that could avert the morbidity and mortality of BOS. This study aimed to determine if decline in lung function pre and early post-transplant (days +80–100) are associated with risk of BOS at least 6 months after transplant. In a single center cohort of 2941 allo-HCT recipients, 186 (6%) met NIH criteria for BOS. Pre-transplant and day +80 spirometric parameters were analyzed as continuous variables and included in a multivariable model with other factors such as donor source, graft source, conditioning regimen, use of total body irradiation and immunoglobulin levels. Pre-transplant FEF_25–75 (forced expiratory flow between 25–75% maximum), day +80 FEV₁ (forced expiratory volume in 1 sec) and day +80 FEF_25–75 had the strongest association with increased risk of BOS. Assessment of the multivariable model showed that day +80 FEF_25–75 decline added additional risk to the day +80 FEV₁ model (p=0.03), while FEV₁ decline at day +80 added no additional risk to the day +80 FEF_25–75 model (p=0.645). Moreover, day +80 FEF_25–75 conferred additional risk when considered with pre-transplant FEF_25–75. These results suggest that day +80 FEF_25–75 may be more important than FEV₁ in predicting the development of BOS. This study highlights the importance of obtaining early posttransplant pulmonary function tests for the potential risk stratification of patients at risk for BOS.

Introduction

Bronchiolitis obliterans syndrome (BOS) is a devastating complication of allogeneic hematopoietic cell transplant (allo-HCT) and is associated with significant morbidity and mortality. BOS is usually not diagnosed until patients present with respiratory symptoms, at which point airflow obstruction is already advanced.^1,2 Detecting an early decline in lung function is challenging as it is often asymptomatic.^{3, 4} Once established, patients can experience progressive and irreversible respiratory impairment, repeated lung infections, and increased risk of non-relapse mortality.^5–7 Identifying patients at risk could aid in targeting patients for early detection or prophylactic strategies.

Approximately 5% of allo-HCT recipients will develop BOS.^{5, 8} While allo-HCT-related risk factors including extra-pulmonary chronic graft-versus-host disease (cGVHD), peripheral blood stem cell source, female donor, busulfan-based conditioning and post-HCT pulmonary infections, for BOS have been well described,^{5, 6, 8, 9} there has been comparatively little study of pre- and early post-HCT lung function parameters as risk factors.

It is well recognized that pre-existing lung dysfunction increases the risk for poor outcomes after allo-HCT. Pre-transplant lung function has been associated with acute respiratory failure and increased non-relapse mortality.^10–12 Lung dysfunction after the early post-transplant period has also been associated with increased late mortality.^13–15 Here we sought to determine if pre-transplant and early post-transplant lung dysfunction as measured by routine pulmonary function tests (PFT) are risk factors for the development of BOS at least six months after transplant. In addition to forced expiratory volume in one second (FEV₁), we examined spirometric parameters that have not been associated with the formal definition of BOS, including the forced expiratory flow between 25–75% maximum (FEF_25–75). The FEF_25–75 is considered a marker of small airways disease in obstructive lung diseases and appears to be an early marker of BOS after lung transplantation.^{16, 17}

Methods

Study Cohort

Patients at least 18 years of age receiving a first allogeneic HCT at Fred Hutchinson Cancer Research Center (FHCRC) between January 1, 2000-June 30, 2016 and who survived at least 6 months post-HCT were selected for chart review. PFTs through December 2016, demographic data, and clinical data were retrieved from institutional databases and clinical records. Standard of care practice at our institution for allogeneic HCT recipients includes obtaining PFT as part of the pre-transplant evaluation within a month of transplant as well as a post-transplant study around day +80–100 (hereafter known as day +80). Subjects with a missing pre-HCT or day +80 PFT, with no PFT beyond day +80 and who already met diagnostic criteria for BOS by day +80 were excluded from the analysis. All subjects had signed informed consent for the use of clinical data for research purposes. The protocol was approved by the FHCRC Institutional Review Board.

Definitions

BOS was defined as per modified 2014 National Institutes of Health (NIH) consensus spirometric diagnostic criteria, as follows: 1) FEV₁ <75% predicted with ≥10% decline in absolute FEV₁ from pre-HCT PFT, and 2) FEV₁/VC <0.7 (FVC was used if VC was not available).¹⁸ Bronchodilator response was not included as this parameter was not uniformly available. Chart review was undertaken for all subjects who met the spirometric criteria to ascertain whether BOS or an alternate etiology for obstructive lung disease was present. High resolution chest tomography, as suggested by the NIH criteria, was not required for the determination of a BOS diagnosis as this also was not uniformly available. For patients with new PFT abnormalities, our typical institutional practice includes an infectious workup, imaging and bronchoscopy as clinically appropriate.

Statistical Analysis

For descriptive data, chi-square, Fisher’s exact test and the Mann-Whitney U-test were used as appropriate to assess first-pass associations with BOS. We then fit univariate logistic regression models followed by multivariable logistic regression models among non-PFT variables relating to transplant variables (referred to as the non-PFT model) to model their association with the development of BOS. Candidate variables for the non-PFT model included those that have been previously shown to be associated with BOS or cGVHD, including recipient age and sex, recipient day +80 IgG level, conditioning regimen (myeloablative +/− total body irradiation, nonmyeloablative), stem cell source, donor type and donor sex.^{5, 6, 8} PFT parameters at baseline (pre-HCT) and at day +80 post-HCT were then examined in univariate models as potential risk factors for BOS (referred to as PFT models). PFT parameters were modeled as continuous linear variables, so that the log-odds of BOS was assumed to be a linear function of PFT. This assumption means that the relative change in the log-odds between patients with a percent predicted PFT of x-c% compared to patients with a percent predicted PFT of x% is the same for all values of x and a given value of c. For presentation purposes, odds ratios (OR) were presented for a difference of c=20% decrease in percent predicted PFT parameter, e.g., the odds ratio for patients whose PFT is 80% of predicted vs. patients whose PFT is 100% of predicted, and this odds ratio is assumed to be the same as that for patients whose PFT is 65% compared to patients whose PFT is 85%. Predicted values were calculated from National Health and Nutrition Examination Survey III reference equations.¹⁹ The PFT parameters examined included FEV₁, forced vital capacity (FVC), FEF_25–75, total lung capacity (TLC) and diffusing capacity for carbon monoxide (DLCO). PFT parameters with p < 0.1 on univariate analysis (baseline and day 80 FEF_25–75% and FEV₁) were then added to the multivariable non-PFT model (referred to as comprehensive models). These comprehensive models were then compared to determine the relative value of each PFT parameter with respect to their association with the development of BOS. Nested models were compared using the likelihood ratio test.

Results

Cohort

Between January 1, 2000-June 30, 2016, 2941 adult patients underwent first allogeneic HCT at FHCRC (Figure 1). Of these, 186 (6.3%) developed BOS. Five of these patients were excluded from analysis because they met spirometric criteria for BOS at day 80 post-HCT, resulting in a cohort of 181 patients who developed BOS at a median of 17 months (interquartile range 9–28 months). Of the remaining 2755 patients who did not develop BOS, 178 were excluded due to missing pre-HCT or day-80 post-HCT PFTs and 155 were excluded due to lack of PFTs beyond day-80 post-HCT, resulting in a control cohort of 2422 patients. Median follow-up of survivors as of last follow-up on November 21, 2017, was 67 months, interquartile range 35–114 months). Table 1 describes the characteristics of each cohort.

Figure 1.

Cohort Development

Table 1.

Characteristics of patients who developed bronchiolitis obliterans syndrome (BOS) and controls.

	No. (%) of Subjects
	BOS cohort (n=181)	Control (n=2422)
Median age at HCT (range)	51 (18–74)	50 (18–80)
Sex, Female (%)	78 (43.1%)	1033(42.7%)
Transplant Indication
AML	66(36.5%)	798(32.9%)
ALL	12(6.7%)	266(11.0%)
CML	20(11.0%)	186(7.7%)
CLL	8(4.4%)	102(4.2%)
MDS	37(20.4%)	513(21.1%)
NHL	17(9.4%)	233(9.6%)
Other	21(11.6%)	324(13.4%)
Donor Type*
Matched Related	69(38.1%)	965(39.8%)
Matched Unrelated	72(39.8%)	991(40.9%)
Mismatched Unrelated	34(18.8%)	248(10.2%)
Haploidentical	3(1.7%)	85(3.5%)
Graft Source
Peripheral Blood	162(89.5%)	1965(81.1%)
Bone Marrow	16(8.8%)	324(13.4%)
Cord	3(1.7%)	133(5.5%)
Recipient Sex/Donor Sex**
Female/Female	42(23.5%)	526(22.3%)
Female/Male	35(19.5%)	479(20.3%)
Male/Female	58(32.4%)	563(23.9%)
Male/Male	44(24.6%)	791(33.5%)
Conditioning Regimen
Myeloablative no TBI	72(39.8%)	832(34.4%)
Myeloablative + TBI	38(21.0%)	690(28.5%)
Nonmyeloablative***	71(39.2%)	900(37.1%)
Acute GVHD
Grade II-IV	130(71.8%)	1710(70.6%)

^*This omits donor type of cord blood, which is represented in Graft Source.

^**There were 2 BOS and 63 control patients who received cord blood transplants in which the sex of the donor was unknown.

^***This includes regimens that utilized total body irradiation not exceeding 300 cGray.

Abbreviations. HCT: hematopoietic cell transplant; AML: acute myelogeneous leukemia; ALL: acute lymphoblastic leukemia; CML: chronic myelogeneous leukemia; CLL: chronic lymphoblastic leukemia; MDS: myelodysplastic syndrome; NHL: non-Hodgkin lymphoma; TBI: total body irradiation; GVHD: graft-versus-host disease

Non-PFT Model

We first assessed factors related to baseline characteristics of the transplant and their association with BOS (Table 2). On univariate analysis, lower risk of BOS was associated with myeloablative conditioning without total body irradiation (TBI) (relative to myeloablative with TBI, OR 0.64, p=0.04), use of bone marrow or umbilical cord graft (relative to peripheral blood stem cells, OR 0.51, p=0.009) and donor type (relative to mismatched unrelated: matched unrelated (OR 0.55, p=0.007), matched related (OR 0.53, p=0.004), umbilical cord (OR 0.1, p=0.03) and haploidentical (OR 0.26, p=0.03)). Lower day +80 IgG was associated with a higher risk of BOS (OR 1.01, p=0.03).

Table 2.

Univariate and multivariable analyses of non-pulmonary function test variables as risk factors for development of bronchiolitis obliterans syndrome.

	Univariate Analysis			Multivariable Analysis*
	Odds Ratio	95% CI	P-value	Odds Ratio	95% CI	P-value
Donor Type (vs. mismatched unrelated)
Umbilical cord	0.10	0.01–0.75	0.03
Haploidentical	0.26	0.08–0.87	0.03	0.35	0.1–1.21	0.1
Matched related	0.53	0.34–0.82	0.004	0.55	0.35–0.86	0.009
Matched unrelated	0.55	0.36–0.85	0.007	0.56	0.36–0.87	0.01
Mismatched related	0.26	0.03–1.97	0.19
Conditioning (vs. myeloablative + TBI)
Myeloablative, no TBI	0.64	0.42–0.97	0.04	0.69	0.45–1.10	0.09
Non myeloablative**	0.92	0.65–1.29	0.62
CMV serostatus match (vs. recipient negative/donor negative)
Negative/positive	0.90	0.54–1.50	0.69
Positive/negative	0.82	0.55–1.24	0.35
Positive/positive	1.04	0.71–1.53	0.84
UC or BM graft (vs. PBSC graft)	0.51	0.30–0.85	0.009	0.58	0.33–1.03	0.06
Male donor (vs. female donor)	0.66	0.49–0.90	0.008
Recipient age	1.01	1.00–1.02	0.14
Number of respiratory viral infections	0.94	0.76–1.15	0.52
through day 80
Day +80–100 IgG level	1.01	1.0–1.01	0.03	1.01	1.0–1.01	0.03
Acute GVHD grade 2–4 (vs. grade 0–1)	1.06	0.76–1.49	0.72

^*Only variables with p ≤0.1 are shown.

^**This includes regimens utilizing total body irradiation not exceeding 300 cGray.

Abbreviations. CI: confidence interval; TBI: total body irradiation; CMV: cytomegalovirus; UC: umbilical cord; BM: bone marrow; PBSC: peripheral blood stem cell; GVHD: graft-versus-host disease

On multivariable analysis, lower risk of BOS was associated with matched related and matched unrelated donors (versus mismatched unrelated, OR 0.53, p=0.009 and OR 0.55, p=0.01, respectively). Higher risk of BOS was associated with lower day-80 IgG level (1.01, p=0.03). Borderline associations with lower risk of BOS included umbilical cord or bone marrow grafts (versus peripheral blood stem cells, OR 0.58, p=0.06), haploidentical donors (versus mismatched unrelated, OR 0.35, p=0.1) and myeloablative conditioning without total body irradiation (versus myeloablative with total body irradiation, OR 0.69, p=0.09).

PFT Model

Next, we assessed the impact of PFT parameters, modeled as continuous linear variables, on the risk of BOS after inclusion of the non-PFT factors summarized in the previous section. On univariate analysis, a higher risk of BOS was associated with pre-transplant reductions in FEF_25–75 (OR 1.13, p=0.01 for each 20% difference in FEF_25–75, e.g., 71% vs. 91%, 63% vs. 83%, etc.) and lower FEV₁ (OR 1.14, p=0.09 for each 20% difference in FEV₁) (Table 3). With respect to day +80 PFT parameters, a higher risk of BOS was associated with reduced FEF_25–75 (OR 1.20, p <0.001 for each 20% difference in FEF_25–75) and reduced FEV₁ (OR 1.27, p=0.008 for each 20% difference in FEV₁). Pre-HCT and day +80 FVC, TLC and DLCO were not statistically significantly associated with BOS.

Table 3.

Univariate analysis of pulmonary function test parameters as risk factors for development of bronchiolitis obliterans syndrome (Odds ratio for each 20% decrease relative to 100% predicted.)

Pre-HCT Parameter	Odds Ratio	95% CI	P-value
FEV1	1.14	0.98–1.31	0.09
FEF25–75	1.13	1.03–1.24	0.01
FVC	1.05	0.91–1.22	0.51
TLC	0.92	0.78–1.10	0.36
DLCO	0.94	0.80–1.10	0.43
Day 80 Parameter
FEV1	1.27	1.06–1.52	0.008
FEF25–75	1.20	1.08–1.34	<0.001
FVC	1.14	0.95–1.36	0.17
TLC	0.98	0.82–1.16	0.78
DLCO	0.94	0.79–1.12	0.47

Abbreviations. FEV₁: forced expiratory volume in 1 second; FEF_25–75: forced expiratory flow between 25–75% maximum; FVC: forced vital capacity; TLC: total lung capacity; DLCO: diffusing capacity of the lungs for carbon monoxide.

While the above considers PFT levels as continuous, we also performed a receiver-operating-curve (ROC) analysis on day +80 FEF_25–75 and FEV₁, where sensitivity and specificity were estimated as thresholds of each PFT varied. The area-under-the-curve (AUC) for FEF_25–75 was 0.591 and for FEV₁ the AUC was 0.582. Neither of these values is particularly large, but this perhaps reflects the observation above that the risk of BOS increases as the PFT declines farther from normative values.

Comprehensive Models

PFT parameters that were statistically significant for association with BOS on univariate analysis (pre-HCT and day +80 FEV₁ and FEF_25–75) were then included into the multivariable non-PFT model. These four PFT values (baseline and day +80 for each of the two parameters) are expected to have a fairly high degree of correlation, therefore each parameter was added individually to the non-PFT model, resulting in four distinct multivariable models. The PFT associations seen after adjustment for the non-PFT factors were similar to the PFT associations without adjustment (Table 4). Figure 2 depicts the relationship between odds of BOS development and day +80 FEF_25–75 modeled as a continuous linear variable (linear for the log-odds of BOS). Given the association of day +80 FEF_25–75 with respect to BOS risk, we examined the relative value of this parameter by assessing the change in model fit when adding day+80 FEF_25–75 to the multivariable models containing day +80 FEV₁ and pre-HCT FEF_25–75 and vice versa. When day +80 FEF_25–75 was added to the day +80 FEV₁ model, there was a significant improvement in model fit (p=0.03). However, when day +80 FEV₁ was added to the day +80 FEF_25–75 model, there was very little improvement in model fit (p=0.645). When pre-HCT FEF_25–75 was added to the model containing day +80 FEF_25–75, the model was scarcely improved (p=0.615). On the other hand, adding day +80 FEF_25–75 to the model containing pre-HCT FEF_25–75 improved the model (p=0.06). Taken together, these models suggest that day +80 FEF_25–75 adds significant value to knowledge of both day +80 FEV₁ and pre-HCT FEF_25–75 with respect to assessing the risk of developing BOS. Furthermore, once one already has knowledge of day +80 FEF_25–75, little additional information is gained by knowing day +80 FEV₁ or pre-HCT levels for FEF_25–75 or FEV₁.

Table 4.

Multivariable analysis of pulmonary function test parameters as risk factors for development of bronchiolitis obliterans syndrome, after adjustment of factors summarized in Table 2. (Odds ratio for each 20% decrease relative to 100% predicted.)

	Odds Ratio	95% CI	P-value
Pre-HCT FEF25–75%	1.15	1.04–1.27	0.007
Pre-HCT FEV1	1.17	1.01–1.36	0.04
Day +80 FEF25–75%	1.21	1.08–1.35	<0.001
Day +80 FEV1	1.28	1.07–1.54	0.007

Abbreviations. HCT: hematopoietic cell transplant; FEF_25–75: forced expiratory flow between 25–75% maximum; FVC: forced vital capacity; FEV₁: forced expiratory volume in 1 second.

Figure 2.

Odds ratio for development of bronchiolitis obliterans syndrome based on day +80 percent predicted FEF_25–75. For example, the OR to develop BOS for a person with 60% of predicted FEF_25–75 versus a person with 100% predicted FEF_25–75 on day +80 is 1.5. The predicted value of FEF_25–75 was calculated based on NHANESIII reference equations (Ref.19)

Discussion

Here we demonstrate the association between pre- and early post-allo-HCT lung function parameters and subsequent development of BOS in a large single institution cohort. Not only are reduced baseline and day +80 FEV₁ and FEF_25–75 significantly associated with BOS, there is a progressive increase in odds of BOS with progressive decline in the parameter below normative reference values. Our results also suggest that once one has knowledge of day +80 FEF_25–75, none of the other 3 parameters provided significant additional information in describing risk of BOS. On the other hand, knowledge of day +80 FEF25–75 provided additional information above and beyond what each of the other parameters provided in terms of describing the risk of BOS.

FEV₁ is the most reliable and validated parameter for detecting airflow decline and is central to the diagnosis of BOS.¹⁸ Decline in FEV₁ post-HCT has been associated with subsequent development of BOS and other non-infectious lung complications.²⁰ While the FEV₁ measures early forced expiratory flow volume, the FEF_25–75 is a measure of mid-expiratory flow volume and may be more representative of small airway function than FEV₁.²¹ Histopathologically, the major site of damage in BOS is the terminal bronchioles.^{3, 22} Pathologic changes in the small airways, however, can occur in the absence of FEV1 impairment, especially in early disease, as has been noted in patients with COPD in which loss of small airways precedes pathologic evidence of emphysema.^{23, 24} Thus it is biologically plausible that the FEF_25–75 may provide valuable diagnostic information beyond that provided by the FEV₁. Indeed, after lung transplantation, FEF_25–75 has been found to be a sensitive marker for subsequent development of BOS, declining significantly earlier than FEV₁,^{16, 17} although these findings have not been consistently replicated.^{25, 26}

In the allo-HCT setting, FEF_25–75 is emerging as a potentially useful spirometric biomarker for later lung disease. In a longitudinal cohort of prospectively followed allogeneic HCT recipients, Bergeron et al. found that a reduced FEF_25–75 at 100 days post-HCT, in addition to reduced FEV₁, was associated with any late-onset non-infectious pulmonary complication, almost half of which were BOS.²⁷ In applying criteria similar to a “pre-BOS” (BOS-0p) stage used in lung transplant,²⁸ Abedin et al. found in a retrospective cohort of allogeneic HCT patients that the development of BOS-0p at any point post-HCT (defined by a 10–19% decline in FEV₁ or a >25% decline in FEF_25–75 on consecutive PFTs) was associated with a significantly increased risk of BOS.²⁹ The FEF_25–75 criterion demonstrated somewhat better positive predictive value than the FEV₁ criterion. A recent report of 445 allo-HCT recipients who received reduced intensity conditioning revealed that lower pre-HCT mid-expiratory flow rates (such as FEF_25–75) were associated with an increased risk of BOS post-transplant.³⁰

Our report replicates and extends the above findings in a large and heterogeneous cohort using the 2014 NIH consensus diagnostic criteria for BOS. We confirm that low values of both FEV₁ and FEF_25–75 early post-HCT and pre-HCT are associated with an increased risk of subsequent development of BOS. Additionally, we demonstrate that: 1) a drop in FEF_25–75 is significantly associated and additive to FEV₁ with respect to subsequent risk of BOS, and 2) the greater the abnormality in the separate parameters of pre-HCT and day 80 FEV₁ and FEF_25–75 decline, the greater the odds of BOS. Taken together, these data reiterate the value of obtaining PFTs in the early post-transplant period. Modeling the FEV₁ and FEF_25–75 as continuous variables allowed us to demonstrate that the degree of decline at day +80 is highly informative and suggests that events in the early post-HCT period are likely to play a role in the pathogenesis of BOS. While patients who experience acute severe lung injury early post-transplant, such as idiopathic pneumonia syndrome, may not survive to develop BOS, it is possible that patients who develop BOS have sustained subclinical lung injury from other processes including respiratory viral infections, bacterial pneumonias, and aspiration events.

Early detection of those at high risk for development of BOS could allow for an enhanced monitoring strategy and pre-emptive or prompt therapy prior to significant loss of lung function. For example, portable handheld spirometry offers a novel approach to lung function monitoring that may allow for frequent, convenient and accurate lung function measurement that the patient may perform on their own.^{31, 32} While there is a paucity of novel therapies for BOS, current approaches have been associated with stabilization of lung function, which allows for the possibility that treatments applied in early-stage disease may halt lung function decline.^{1, 33, 34} Therefore the identification of individuals who would benefit from more intensive monitoring approaches, such as frequent handheld spirometry, is essential.

The large sample size of our cohort allowed for a comprehensive multivariate analysis that included both PFT and HCT variables. The apparent lower risk of BOS in recipients of haploidentical donor allo-HCT as compared to mismatched unrelated donors is a relatively novel finding, although this observed association warrants confirmation in future studies. This finding has been recently reported in a more contemporary subset of the present cohort,³⁵ and is consistent with the favourable cGVHD outcomes observed after haploidentical allo-HCT with post-transplant cyclophosphamide.^36–38

Our study has limitations. Given the retrospective nature of the cohort, and the fact that many patients leave our center for their care after the day +80–100 posttransplant period, we could not ensure that all patients had PFTs done at routine intervals during posttransplant follow up and therefore we cannot rule out missed cases of BOS. Our institutional guidelines do specify the optimal frequency of PFT after allo-HCT which may be variably followed by treating physicians once they are discharged after the early posttransplant period. We excluded a relatively small number of patients without a pre-HCT or post day + 80PFT. Reassuringly, we report an incidence of BOS that is entirely consistent with other large cohorts.^{5, 8} An important limitation of our study is that our model is inadequate for point of service evaluation of risk, as the relative odds of developing BOS based on the PFT parameters is relatively small based on the increments of PFT decline. FEF_25–75 is also a nonspecific marker of lung injury, and the relative risk may be diluted by other non-BOS conditions such as infectious pneumonia in the control group. Further work, including validation with other cohorts and prospective studies in which longitudinal lung function is collected at regular and frequent intervals after transplant, need to be done to assess the predictive value of early PFT parameters for BOS and other non-infectious lung complications.

In summary, examination of both the FEV₁ and FEF_25–75 early post allo-HCT may identify those at risk of subsequently developing BOS. These findings highlight the importance of obtaining routine PFTs on patients post-transplant. Those at high risk for BOS may benefit from enhanced screening programs and/or pre-emptive or prompt therapy of BOS prior to significant loss of lung function.

Highlights.

• Identifying patients at risk for bronchiolitis obliterans syndrome after allogeneic hematopoietic cell transplantation may offer improved strategies for early detection.

• Declines in pretransplant and early posttranpslant spirometric parameters are associated with risk of developing BOS.

• Reduction in FEF_25–75 at day +80 may be more important than FEV₁ in predicting BOS development.