INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As lung transplantation has become an accepted mode of therapy for patients with end-stage parenchymal or vascular disease, it has become clear that bronchiolitis obliterans syndrome (BOS) (2) is a serious complication in about 40% of patients by 2 yr post-transplant (3). Multiple factors have been implicated in the cause of BOS (7), including number and severity of acute rejection episodes, cytomegalovirus (CMV) infection, and HLA mismatching. Prior studies have considered the frequency of episodes of symptomatic rejection or a single episode of moderate-grade rejection as risk factors for the development of BOS (5, 12). Our purpose in this study was to compare the relative importance of acute perivascular rejection and lymphocytic bronchitis/bronchiolitis (LBB) on the development of BOS. Issues examined include timing of acute rejection episodes and adequacy of immunosuppression. In the past there has been no standardized method for independently assessing and quantitating perivascular rejection and airway inflammation. In 1995 the lung transplant study group revised the grading system and provided a method to evaluate them separately. We regraded all our prior biopsies in which infection was excluded. Length of follow-up and number of biopsies varied from patient to patient. Hence, we devised a quantitative method for analyzing biopsies in terms of the severity of histopathologic changes and the number of episodes, thus taking into account differences in the length of follow-up and number of biopsies.

In this study we have shown that acute perivascular rejection, LBB, late acute rejection, and decreased immunosuppression correlate strongly with the development of BOS. Controversy exists about whether LBB represents acute airway rejection (15, 16). The current analysis strongly supports the contention that LBB is, in fact, a manifestation of acute rejection and carries with it the same implications as conventional acute perivascular rejection.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All patients who received lung allografts at Loyola University Medical Center between April 1990 and November 1995, and survived at least 90 d, form the basis for this report. Retransplant surgery was considered as a separate event from the first procedure for the purpose of this study. Immunosuppression with cyclosporine A, azathioprine, and prednisone was delivered by a previously described protocol (17) and was similar in all patients. Transbronchial biopsies were performed once prior to postoperative Day 30 in all patients and then as needed during the remainder of the follow-up period. Indications for biopsy included the clinical suspicion of presence of rejection or infection based on chest radiographs with new infiltrates, drop in FEV₁ on spirometry, decreased exercise tolerance, the presence of fever or cough or for the follow-up of prior rejection or infection. Biopsies were fixed and processed according to our previously reported protocol (18).

Patients were divided into those with BOS and those without BOS (NBOS) by presence or absence of the clinical syndrome, with or without histologic confirmation. BOS is the term used to designate clinical allograft functional deterioration after transplantation, which may be accompanied by a histologic diagnosis of obliterative bronchiolitis. After a post-transplant maximum FEV₁ has been established, BOS is graded progressively from 0 to 3, depending on decline in FEV₁ (2). The following data were tabulated for each patient: primary indication for transplant, age at transplantation, sex, length of follow-up, presence or absence of CMV pneumonitis on biopsy, multiple episodes of CMV pneumonitis, airway stenosis, prolonged (i.e., > 3 d) post-transplant ventilation, whole-blood cyclosporine A level < 200 ng/ml (monoclonal whole-blood assay; Abbot Laboratories, Abbot Park, IL) on two or more occasions, cause of death, and survival. Airway stenosis was defined by the need for bronchial dilatation or stent placement. All lung biopsies in which infection was excluded were regraded according to the 1996 formulation (1). Quantitative scores (Grade 0 to 4) for both acute perivascular rejection and LBB were totaled at 90, 180, and 365 d and at the end of the study period to give cumulative scores for each type of damage, respectively, e.g., a patient with A2 at 14 d, A3 at 97 d, and A2 at 180 d post-transplant would have cumulative score of A7 at 180 d. Mean cumulative scores were then tabulated for BOS and NBOS groups. Patients who died between 90 and 365 d post-transplant were analyzed in a manner similar to survivors and were not excluded from the study. In addition, we examined whether late acute rejection or late LBB might influence BOS by tabulating scores for patients with rejection and LBB episodes occurring more than 180 d after transplant in those who had little or no rejection or LBB during the first 180 d. HLA matching was analyzed when available, and reported by the number of matches at two A, B, and DR loci, respectively, with a maximum of six matches possible in any recipient. Absence, or a blank, at any locus was considered to be a mismatch.

Statistical analyses were performed using Student's unpaired t test, the Mann-Whitney U test, Chi-square with Yates correction, Fisher's exact probability test, and Spearman's rank correlation coefficients. Probability values of p  0.05 were considered to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Between April 1990 and November 1995, 130 patients underwent 134 lung transplantation procedures (including four retransplantations) and survived at least 90 d. Of these transplantation procedures, 66 developed BOS (36 biopsy-proven) and 68 did not. Patient summary data for the BOS and NBOS groups are shown in Table 1. Primary diagnosis, age, sex, type of transplant, and CMV pneumonitis were similarly distributed between the two groups, but follow-up time was significantly longer in the BOS group (p << 0.001, Mann-Whitney U test). There was no significant difference in CMV pneumonitis that was seen histologically in 28 of 66 (42%) BOS patients and 21 of 68 (31%) NBOS patients (² = 1.46, p = 0.23). There were two or more episodes of CMV disease in 12 of 66 (23%) BOS and seven of 68 (10%) NBOS patients (² = 1.13, p = 0.29). Three of the BOS and 10 of the NBOS group had significant bronchial stenosis that needed dilatation and/ or stent placement (² = 2.87, p = 0.09). Mechanical ventilation for longer than 3 d was needed in 13 BOS and in 15 NBOS patients (² = 0.015, p = 0.9). Cyclosporine A levels were less than 200 ng/ml on two or more occasions in 21 of 66 BOS patients and in three of 68 NBOS patients (² = 15.3, p << 0.001). In the above 21 BOS patients immunosuppression was decreased because of post-transplant lymphoproliferative disorder (PTLD) (one), bronchioloalveolar carcinoma (one), and infection (three). The remaining 16 were due to presumptive noncompliance. Twenty-five of 66 (38%) BOS patients have died (seven in < 365 d), and 14 of 68 (21%) NBOS patients have died (10 in < 365 d) (² = 4.05, p = 0.044). Ten of the former died with infection in addition to BOS. Of the 14 NBOS deaths, eight were due to infection, three to PTLDs, and three unrelated to the transplant.

Mean cumulative acute rejection and LBB scores are presented in Table 2. Spearman's rank correlation showed no statistically significant correlation between the cumulative acute perivascular rejection (A) and LBB (B) scores and the length of follow-up in the two study groups: (BOS-A: r = 0.06, BOS-B: r = 0.21; NBOS-A: r = 0.20, NBOS-B: r = 0.26). There were statistically significant differences between the BOS and NBOS groups in the severity of perivascular rejection and LBB at 180 and 365 d and at the end of the study period, with BOS patients having higher mean cumulative scores (see Table 2). The total scores at 90 d approached a significant difference (A: p = 0.22, B: p = 0.12), and at 180 d there was a clear separation of the two groups (A: p = 0.04, B: p = 0.01).

In order to evaluate the impact of late acute rejection or LBB, we analyzed all patients whose perivascular and airway scores were  2 at 180 d. There were 70 patients whose A score was  2 (irrespective of their B score), of whom 22 patients demonstrated an increase in A score  2 by the end of the study period. Of these 22 patients, 15 subsequently developed BOS, but the other seven did not. Of the remaining 48 patients who did not show a similar increase in late acute perivascular rejection score, 15 developed BOS and 33 did not ² = 6.96, p = 0.008). Similarly, there were 104 patients whose B scores were  2 (irrespective of their A score). Of these, 37 had an increase in their B score  2 by the end of the study period. Of these 37 patients, 28 developed BOS and nine did not. Of the remaining 67 patients who did not show a similar increase in B scores, 19 developed BOS and 48 did not (² = 19.67, p = 0.000009). Among the 11 BOS patients who were completely free of rejection and LBB at 180 d (A = 0, B = 0), 10 developed late acute rejection or airway scores of 2 or more within the first post-transplant year. None of the nine NBOS patients who were rejection-free and LBB-free at 180 d developed any acute rejection of LBB over the same interval (p < 0.001, Fisher's exact test).

HLA typing (A, B, and DR loci) was available on 58 of 66 BOS, and 54 of 68 NBOS patients. HLA analysis revealed no significant difference in the incidence of matching in BOS versus NBOS groups. There were 0 to 1 matches in 49 BOS (84%) and in 43 NBOS (80%) patients. In addition, there were two or more matches in nine BOS and 11 NBOS, with one patient from each group having four HLA matches. No patient had five or six matches at the HLA loci. The number of HLA-A, HLA-B, and HLA-DR mismatches were compared in 21 patients with no acute rejection or LBB (Group 1, M:F = 13:7; mean age, 50 yr; 12 emphysema, one cystic fibrosis [CF], and eight others) and 34 patients with three or more episodes of acute rejection or LBB (Group 2, M:F = 13:21; mean age, 41 yr; 15 emphysema, 13 CF [38%, p = 0.009], and six others) in the first 180 post-transplant days. There was no difference in the number of HLA-A and HLA-B mismatches, whereas HLA-DR showed 30 mismatches out of a possible 42 (71%) in Group 1 and 64 out of a possible 68 (94%) in Group 2 (p = 0.003). BOS occurred in nine (43%) of Group 1 and 21 (62%) of Group 2 patients.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Acute rejection has been shown to be a significant risk factor for the development of BOS in several prior reports (7). However, these took into account only the number and severity of acute perivascular rejection episodes. Yousem and colleagues (15, 16) suggest that LBB is a result of acute rejection and responds to augmented immunosuppression. Use of the new working formulation allows analysis of LBB, absent infection, independent of blood vessels in alveolated lung (1). The current study was undertaken to ascertain whether LBB, independent of perivascular rejection, marked patients for development of BOS. This study showed that at 180 and 365 d, LBB scores in BOS patients were more than twice that in NBOS patients (p << 0.001), whereas perivascular scores were 1.7 times greater in BOS than NBOS patients (p = 0.003). Thus, both LBB and acute perivascular rejection correlate with higher risk of developing BOS.

This study also addresses the problem of whether adequate tissue has been obtained for histologic analysis. Our protocol for rejection surveillance calls for a minimum of five to six pieces of alveolated lung tissue, but in some cases, adequate number of lung tissue fragments are not submitted and significant pathologic changes are noted only in airway mucosa (14). Possible causes for the inability to obtain appropriate samples include presence of bleeding, hypoxemia, or respiratory distress. Because both perivascular rejection and LBB provide comparable information, in our opinion, acute rejection can be treated based on what would otherwise be considered an inadequate specimen.

We also examined the significance of late acute rejection and late LBB on the development of BOS. In patients with little or no early rejection or LBB, late rejection episodes and LBB were more common in BOS than in NBOS patients. In fact, in contrast to the suggestion of Baz and colleagues (19) that patients who are rejection-free in the first few post-transplant months remain so, we have found that a significant fraction of patients who were rejection-free and LBB-free at 180 d and developed late acute rejection or late LBB subsequently developed BOS. This supports the claim by Girgis and colleagues (12) that late rejection is a major risk factor for development of BOS. This is similar to literature in renal transplantation that has also concluded that late acute rejection increases disproportionately the risk for chronic rejection (20).

We also show that decreased immunosuppression as measured by whole-blood cyclosporine A levels < 200 ng/ml was substantially more common among patients who subsequently developed BOS than among those who did not (² = 15.3, p = 0.0001). This finding, whether due to noncompliance or treatment of malignancy or infection, may, in part, explain the greater frequency of acute rejection and LBB in the BOS group. It is not known if decreased cyclosporine A levels predispose to development of BOS independent of acute rejection or LBB.

Because of the concern over the length of follow-up as an independent variable, we evaluated total rejection and LBB scores for BOS and NBOS groups at fixed intervals. We showed that the mean cumulative rejection and LBB scores at 180 and 365 d could be used to predict development of BOS since a statistically significant difference in rejection and LBB scores can be demonstrated between BOS and NBOS patients within 6 mo after transplantation. In addition, Spearman's rank correlation test demonstrated no correlation between length of follow-up and total acute rejection and LBB scores. In other words, the mean cumulative acute rejection and LBB scores are not directly related to the length of follow-up. However, clinical experience demonstrates that the incidence of BOS continues to increase with increasing post-transplant survival time (12, 14). This means that as follow-up of our patients proceeds, more patients will move from the NBOS to the BOS group, dependent on their early experience with rejection or LBB. Recent patients are reasonably included because BOS has been noted very shortly post-transplant, and also because their inclusion is a conservative way to manage the data. The early rejecters still free of BOS may well move to the BOS group in the future.

Limitation in the present study design was the fact that, after the first 30 d, transbronchial biopsies (TBB) were performed for clinical indication rather than surveillance, resulting in unequal number of biopsies per patient. However, using the cumulative method of comparing patients takes into account the total symptomatic rejection experienced by that patient. In this method, negative biopsies do not add to the cumulative score. In other words, the patients with higher number of biopsies had more symptomatic episodes and thus were at higher risk of developing BOS. In addition the total rejection and LBB scores at 365 d for each patient would vary from patient to patient depending on whether the patient was alive or dead prior to reaching 365 d. The mean values thus err on the side of being too conservative in the prediction of development of BOS.

Post-transplantation complications such as prolonged mechanical ventilation, ischemia-reperfusion injuries, bronchial stenosis, and recurrent/prolonged CMV infections were not statistically different between the BOS and NBOS groups, and therefore did not appear to play an important role in this study, in contrast to some previous investigations. (7)

The HLA analysis showed a very low incidence of matching donor-to-recipient in our population as a whole. This is consistent with random donor-recipient matching, as is expected. Although mismatching at the HLA-DR locus appears to be associated with a higher incidence of acute rejection and BOS, difference in degree of mismatching alone does not fully account for the incidence of rejection or LBB in lung transplant recipients.

In our center, patients who are thought to be at high risk for developing BOS are treated very aggressively. For every episode of acute perivascular rejection and/or LBB, including the first one, especially with clinical signs out of proportion to histologic grade, we use not only enhanced immunosuppression but may also alter maintenance therapy by use of tacrolimus or methotrexate and with the possible addition of other therapies, including antithymocyte globulin, total lymphoid irradiation, plasmapheresis with high dose immunoglobulin replacement, or photopheresis. The aim now, more than ever, is to prevent the development of BOS in long-term survivors of lung transplantation. However, we do not know that therapy for acute rejection or LBB prevents BOS, and a multicenter prospective study is needed to answer such a question.

In conclusion, we show that both perivascular rejection and LBB are associated with a higher risk of development of BOS. In our opinion, LBB absent infection, is indeed a manifestation of acute rejection, and may be an even more significant factor in the development of BOS than is acute perivascular rejection. In our study, decreased immunosuppression, whether caused by physician instruction or patient noncompliance, was a major risk for BOS. In addition, late acute rejection and LBB, even if mild, appeared to disproportionately increase the risk of BOS. Of particular interest, morbid events, including CMV infection, prolonged mechanical ventilation, and bronchial stenosis did not correlate with the development of BOS. HLA analysis showed essentially random matching and mismatching, as expected. Finally, we demonstrate that a quantitative method of scoring TBB may be used to compare acute rejection experience between groups of patients.