INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Obliterative bronchiolitis (OB), presumably a form of chronic allograft rejection, remains the complication most frequently impairing long-term prognosis in lung and heart-lung transplantation (1, 2). Histologically, OB is a fibroproliferative process resulting in a progressive distortion and obstruction of the terminal and respiratory bronchioles, leading to allograft failure (3). Although OB was first described more than 10 yr ago, and despite its high prevalence, its pathogenesis is not clearly defined. Early episodes of acute rejection and a positive cytomegalovirus (CMV) serologic status have been shown to be risk factors for bronchiolitis obliterans syndrome (BOS) onset in multivariate modeling (6). Furthermore, human leukocyte antigen mismatch (6), lung infection, episodes of CMV infection (9), organizing pneumonia (8, 10), lymphocytic bronchitis, and decreased immunosuppression (cyclosporine A levels) (11) are thought to contribute to the development of BOS.

OB leads to an irreversible obstructive ventilatory defect. Histologic diagnosis of OB can be made by transbronchial lung biopsy (TBB). However, owing to the invasiveness of this procedure, the patchy distribution of lesions of OB within the lung, and the difficulties in obtaining representative specimens containing bronchioles (due to the samples' small size), significant deterioration of lung function may occur before pathologic evidence of the disease is obtained (3). Accordingly, the International Society for Heart and Lung Transplantation (ISHLT) proposed that a persistent decrease in FEV₁ to less than 20% of its baseline (B) value be considered a diagnostic criterion for BOS (4). The ISHLT grading system for assessing the extent of the disease, depending on the decline in FEV₁, is now widely accepted (Table 1). However, this system slightly overestimates the time of BOS onset.

Detecting the presence of OB at an earlier stage, before FEV₁ has dropped by 20%, could help in better understanding the origins of the disease, and may have therapeutic consequences. Changes in a patient's immunosuppressive regimen might be more efficient when airway damage is still minimal (9).

Because OB primarily affects the distal airways, we investigated the value, for diagnosing OB, of two tests that reflect small-airways dysfunction better than does FEV₁: maximal midexpiratory flow (FEF_25-75) and the single-breath nitrogen washout test (nitrogen slope, N₂), which reflects changes in ventilation distribution (12, 13). Furthermore, studies of bronchoalveolar lavage fluid (BALF) cells in patients with lung transplantation showed a strong association between increased neutrophil counts and OB (14, 15). We therefore also investigated whether BALF neutrophilia can assist the detection of BOS after lung transplantation. The main objective of the present, prospective cohort study was to compare functional markers of distal airways involvement and BALF neutrophilia in patients with and those without BOS at the end of follow-up, with intent to: (1) validate the respective agreement of these markers for the diagnosis of OB; and (2) to identify their potential usefulness for the detection of OB before the occurrence of a decline in FEV₁.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Transplant Population

From May 1989 to September 1996, we included in the study 45 consecutive patients, who received lung or heart-lung allografts at our institution and who survived for at least 1 yr after the transplant procedure and at least 1 yr after the diagnosis of OB when the latter was present. The patients' underlying diseases were cystic fibrosis (n = 34), bronchectasis (n = 1), emphysema (n = 3), primary pulmonary hypertension (n = 5), Eisenmenger's syndrome (n = 1), and idiopathic pulmonary fibrosis (n = 1). Patients received a standard triple-drug immunosuppressive regimen that consisted of cyclosporine, adjusted to maintain whole-blood trough levels between 200 and 400 ng/ ml; azathioprine (0.5 to 1 mg/kg/d), adjusted to white blood cell count (reducing the dose if white blood cell count fell below 5,000); and steroids, tapered to a dosage of 0.25 mg/kg/d over the first 3 mo and stopped at the sixth postoperative month. Prophylactic intravenous rabbit antithymocyte globulins (Pasteur Merieux, Lyon, France) were given for the first seven postoperative days for patients who underwent transplantation before January 1993 (n = 6), and for the first three postoperative days for patients who had transplantations after this. Episodes of acute cellular allograft rejection were treated with intravenous methylprednisolone (10 mg/kg/d for three consecutive days, and then rapidly reduced). No patient received inhaled steroids during the study.

Patients made regular visits to the hospital for clinical and radiologic evaluation. Fiberoptic bronchoscopic follow-up, with concurrent bronchoalveolar lavage (BAL) and transbronchial biopsy (TBB), was performed monthly for the first three postoperative months, and then every 3 mo for the first year and as needed in patients with clinical symptoms during the remainder of the follow-up period. The severity of acute and chronic organ rejection was graded according to ISHLT staging criteria (4, 5). Repeated BAL procedures in the same patient were considered as separate events, and the patients included in the study had at least six representative BALF specimens taken.

Pulmonary Function Tests

Pulmonary function tests (PFTs) were routinely conducted at our center on a monthly basis for the first 12 postoperative months, and at 3-mo intervals thereafter. In addition, PFT were conducted when patients had clinical symptoms or a decline in home spirometry values of at least 10%. Spirometry was done with an 830-L whole-body plethysmograph (Masterlab Jaeger, Wurzburg, Germany), and included measurement of FVC, FEV₁, RV, and TLC. The forced expiratory flow rate at 25% to 75% of FVC (FEF_25-75) was obtained from the best flow- curve volume. A single-breath N₂ washout test (SBN₂) was performed as previously described by Buist and Ross (16), and measurements of N₂ were made at mouth opening with a rapid N₂ analyzer (Model 4730; Hewlett-Packard, Waltham, MA). Expired volume was obtained with a bag-in-box system connected to a bell spirometer fitted with a potentiometer. Both N₂ concentration and expired volume were plotted on an X-Y recorder during a slow exhalation to RV, which followed a slow inhalation of pure oxygen to TLC. Flow was not recorded, but subjects were trained to keep constant flows (by observing the slope of the volume signal) below 0.5 L/s. The slope of the alveolar phase (Phase III) of the single-breath N₂ test (% N₂/L) was chosen as the most relevant parameter of the nitrogen washout test. Subjects produced at least three satisfactory tracings. Each trial was preceded by a maximal inspiration and followed by a 3-min pause.

Spirometric results were expressed as percentages of predicted values (P) (17, 18) and of baseline (B) values for an individual patient. BOS was defined according to the ISHLT staging criteria (4), using decline in FEV₁ as a criterion. Following the ISHLT proposals for the determination of FEV₁ B values, we determined B values for FEF_25-75, to which all subsequent measurements were referred, by averaging the two previous highest consecutive values before a sustained decline in function, making these measurements 4 to 8 wk apart. With reference to criteria used in a previous study (19), an FEF_25-75 < 70% of both P and B values, respectively, was considered abnormal, and values at or below this threshold value were chosen for investigation. The slope of Phase III of the N₂ washout curve was considered abnormal when it was more than 2 SD above the predicted values for nonsmoking men and women (13, 16). The threshold of > 3% for abnormality of the N₂ slope was extrapolated from Buist and Ross's analysis (16), and was chosen for assessment. FEF_25-75 and N₂ slope measurements were obtained at the same time as FEV₁ values.

Bronchoalveolar Cellularity

Results of 467 bronchoscopies with representative BAL and TBB performed during the same procedure were analyzed. Samples were taken from patients after sedation, through a flexible fiberoptic bronchoscope wedged into a subsegmental bronchus of either the lingula or the right middle lobe. BAL was done by instilling a total volume of 250 ml of isotonic saline solution, warmed to room temperature, in sequential 50-ml aliquots. The fluid fractions were aspirated with gentle low-pressure suction into a siliconized glass container. The retrieved fluid was pooled and mixed, and the total recovered volume was measured. Samples of fluid were processed immediately for cytologic and microbiologic analysis.

Both the total white cell count in BALF and a differential count were performed routinely. Red blood cells and nucleated cells from fresh, uncentrifuged, neat BALF were counted on a Kova Slide (Boehringer Mannheim, Mannheim, Germany). One-hundred-microliter aliquots of BALF were cytocentrifuged (700 rpm for 2 min) and spread on five silane-coated slides (Sigma Chemical Co.) with a Cytospin II (Shandon Southern Instruments, Sewickley, PA) to identify nucleated cells. Two of the cytospin preparations were stained with May-Grunwald-Giemsa, two others with Papanicolaou stain, and the last slide with Perls' method to detect ferric pigments. Results were expressed as percentages of total BALF cell counts. We defined as abnormal a level of relative neutrophilia 10-fold higher (> 20% of total BALF cell counts) than the expected values for nonsmoking healthy volunteers (20). All of the criteria for abnormality of FEF_25-75, N₂, and alveolar neutrophilia were a priori criteria rather than data-driven criteria.

Data Analysis

The threshold values for the three markers used in the study were considered as having been reached when they were obtained at least twice consecutively in an individual patient, and at least 4 to 8 wk apart, with the aim of excluding transient changes caused by an acute complication. This definition permitted us to determine whether each parameter was present or not in an individual patient during the study follow-up. The extensive longitudinal database permitted us to include only patients with fixed abnormalities in the study.

The outcome measure used in analysis was the time to initial manifestation of BOS (i.e., a  20% decrease in FEV₁) on the time scale starting from transplantation. The time point at which the studied parameters reached the chosen threshold values were recorded and associated with a delay (number of days) from surgery for each subject.

To assess the accuracy of these markers for the diagnosis of OB (i.e., to validate that the tests did in fact reflect the dimensions they were purported to reflect), we measured the overall degree of observed agreement of the results of each of the studied tests with the observed FEV₁ abnormalities. To incorporate chance expected agreement into the assessment of intertest reliability, we calculated the coefficient of agreement, kappa, which appears to reflect the measure of agreement with desirable properties (21). We evaluated sensitivity, specificity, and positive and negative predictive values of the three markers for the diagnosis of OB.

In the OB group, we compared the mean delay (number of days, expressed as median and range) from transplantation to the time point at which the threshold was reached for each parameter, and the mean delay from transplantation to the diagnosis of BOS. Analysis of differences in these two mean delays was done with the single-sample paired t test. A p value of p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During the follow-up period, we studied 765 PFT results and 467 representative BALF samples obtained from 45 patients who underwent 47 lung transplant procedures (double lung, n = 38; heart-lung transplantation, n = 9, including two retransplantations). Using the staging criteria defined by the ISHLT (4), we found that at the end of follow-up, 22 of the transplantation procedures were associated with evidence of BOS (three with Grade 1; two with Grade 2; and 17 with Grade 3 disease; of these, 17 cases overall were biopsy-proven) and 25 were not. The mean postoperative delay from the transplantation date to the initial manifestation of BOS (± confidence interval [CI]) was 578 ± 265 d (range: 122 to 2,619 d).

At the selected threshold values of the markers studied, the agreement on the diagnosis of BOS (established by the decline in FEV₁) was equally good with all three markers (FEF_25-75, the slope of Phase III of the N₂ washout curve, and the percentage of BALF neutrophils (kappa > 0.65 with p < 10⁵, respectively) (Table 2). The performances of these criteria for the diagnosis of BOS are summarized in Table 3. Sensitivity and negative predicted value reached 100% for FEF_25-75 < 70% B, N₂ > 3%, and BALF PMN > 20%. The diagnostic sensitivity of FEF_25-75 < 70% P was 86%. The specificities and positive predicted values of the different tests varied between 76% and 100%. These values were related to patient status at the end of follow-up. The effect of bacteriologic findings on BALF neutrophil counts was assessed in parallel. Among 22 patients affected with OB and who presented a sustained alveolar neutrophilia (PMN > 20%), 19 had a history of positive sputum cultures and three had no evidence of bacterial growth. Fourteen of these latter 19 patients had Pseudomonas aeruginosa cultured regularly, three of the 19 had Haemophilus influenzae cultured, and two had other gram-negative bacteria cultured. Only two of 25 patients considered as free of BOS at the end of the study had cultures that subsequently grew bacterial species.

For the 22 BOS patients, the mean numbers of days (± CI) from transplantation to the date when: (1) the FEF_25-75 fell below 70% B and P values, respectively; (2) N2 > 3%; and (3) alveolar neutrophilia > 20% were 405 (± 131) d and 467 (± 267) d, 427 (± 257) d, and 446 (± 184) d, respectively. These criteria for OB were present with mean delays of 173 d (p = 0.03), and 111 d (p = 0.09), 151 d (p = 0.003), and 132 d (p = 0.1), respectively, before there was a decline of 20% or more in FEV₁ (Table 4). Figure 1 shows representative patterns of change over time for the studied parameters in a lung allograft recipient who developed BOS.

View larger version (25K): [in this window] [in a new window]   View larger version (22K): [in this window] [in a new window]   Figure 1.   Representative patterns of change over time following transplantation for slope of N2 washout (A), FEF25-75 versus B (B), FEF25-75 versus P (C ) and relative BALF neutrophilia (D) as compared respectively with FEV1 in a lung allograft recipient who developed BOS.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical manifestations associated with OB may or may not parallel the pathologic findings with either TBB or open biopsy (3). The proposal made by the Lung Rejection Study Group for the grading of OB was that FEV₁ was the most reliable and consistent indicator of graft function, which permitted comparison of results from different centers (4).

One strength of the present study was the wealth of follow-up data for each patient. In accordance with the literature, we validated the usefulness of functional tests reflecting smaller airway dysfunction in detecting BOS. FEF_25-75, which is easy to measure, is a well-recognized indicator of peripheral bronchiolar involvement. It has been proposed as a more sensitive parameter than FEV₁ for determining the onset of OB (19, 22, 23). The reference work of Patterson and colleagues was focused on a group of patients fulfilling BOS criteria. In our cohort study, longitudinal follow-up of FEF_25-75 in the groups of patients with and without OB gave still more value to this marker.

We routinely made sequential measurements of indices obtained from the single-breath N₂ test to assess regional inhomogeneity and changes in ventilation distribution in the periphery of the lung, at the level of respiratory bronchioles (24). This test has been shown to be a highly sensitive test for small-airways disease preceding spirometric deterioration in smokers (12, 25). The slope of Phase III of the N₂ test is significantly correlated with the inflammation score for small airways in smokers (26). Although the changes in pulmonary functional variables associated with posttransplant OB have been well documented, there has been a lack of literature about the abnormalities in ventilation distribution in this setting. The single-breath N₂ test has been used to detect occurrence of transient airway obstruction as a result of an acute episode of lung infection or rejection (27, 28). We showed a steady and progressive increase in the slope of the N₂ washout curve in patients who gradually developed irreversible airflow obstruction. A previous study (27) supports our findings (29) on the effect of chronic allograft dysfunction on the distribution of ventilation as measured with the N₂ washout test. The present results emphasize the usefulness of this test in the clinical follow-up of BOS.

Our study demonstrated that increased BALF neutrophil counts showed good agreement with the diagnosis of OB and with bronchial infections. The sputum bacteriology results in the study were similar to other published data recorded for bronchectasis patients both having and not having lung transplantation (30, 31), in whom P. aeruginosa and H. influenzae were the major pathogens. In our infected allograft recipients, intermittent administration of systemic and/or nebulized antibiotic therapy decreased the density of bacteria in sputum, but had no significant effect on the underlying chronic bronchiolar inflammation, and therefore on BALF neutrophilia, when underlying OB was present. A strong association has previously been identified between increased percentages of neutrophils in BALF and OB (14, 15, 32, 33). Currently, it is difficult to define the exact contribution of neutrophil activation in BALF to the pathogenesis of OB, and to say which comes first, the infection, or inflammation with a sustained neutrophilia. Some investigators recently attempted to demonstrate the possible involvement of neutrophils in the development and progression of the disease process of OB. They reported that BALF levels of interleukin (IL)-8 (14), which is biologically chemoattractive for neutrophils, as well as of eosinophil cationic protein and myeloperoxidase (MPO) (15), correlated directly with the presence of PMN, and were significantly increased in BOS patients. Among these patients, both immunolocalization of IL-8 in peribronchial areas and neutrophil chemotaxis experiments confirmed that IL-8 was bioactive and accounted for the neutrophil presence in the lung (14). Similarly, increased numbers of activated neutrophils stained for MPO have been found in the bronchial wall of TBB specimens from patients with BOS (15). Consequently both IL-8 and neutrophils in the lung could produce a parallel release of reactive oxygen species, thereby playing a role in the genesis and perpetuation of chronic inflammation and in the development of pulmonary fibrosis in BOS.

At the selected threshold values, the change in FEF_25-75 from B, the slope of Phase III of the N₂ washout curve, and alveolar neutrophilia had a sensitivity and negative predictive value similar to those of FEV₁ for the diagnosis of BOS. The specificity of these parameters in detecting OB was slightly lower (range: 76% to 80%). However, one can begin to think that the actual rate of chronic rejection was perhaps somewhat underestimated with the use of FEV₁ alone. Thus, we can hypothesize that the specificity of the tests examined in our study could increase with time, and that patients remaining free of BOS at the end of follow-up, but fulfilling the three criteria examined in our study, might be moved into the BOS group in the future.

The chosen markers were found to be pertinent indicators of BO, and preceded (range: 173 to 110 d) a substantial decline in FEV₁. Few studies have succeeded in finding markers that can predate the physiologic changes of OB. Patterson and colleagues reported a 112-d average interval between a decrease in FEF_25-75 to 70% P and a 20% reduction in FEV₁ from B (19). In agreement with our results, a previous study showed that relative and absolute BALF neutrophilia were predictive of OB, and were found, on average, 11.4 mo before the diagnosis of OB was made (14). In the present study, despite being examined in the smallest number of BALF samples as compared with the PFT data base, alveolar neutrophilia was informative. Prior literature and our results emphasize the association of high PMN counts in BALF with chronic rejection, not only during the late phase but also during the earlier phase of bronchiolar involvement. Moreover, surveillance BAL for neutrophilia could be particularly useful in detecting OB after single-lung transplantation. In this situation, which applies to the majority of patients who undergo transplants around the world, clinical follow-up is difficult largely because PFT data cannot be extrapolated to the single transplanted organ when the native lung has an impairment.

In conclusion, we showed that irreversible changes in functional tests reflecting small-airways dysfunction, such as changes in FEF_25-75 from B values, changes in indices of ventilation distribution (rise in the slope of the N₂ washout curve), and increased percentages of neutrophils in BALF were more relevant than FEV₁ for the early detection of BOS. This suggests that in addition to monitoring FEV₁, sequential measurements of these parameters might be useful for clinical management of lung transplant recipients. Such markers could be taken into account as inclusion criteria for patients considered to have an early stage of OB in multicenter prospective studies, which are needed to know whether earlier enhancement or changes in immunosuppression prevent the development and/or progression of BOS.