|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Division of Pulmonary and Critical Care Medicine (Drs. Santo Tomas and Lipchik), Department of Medicine; Division of Epidemiology (Dr. Layde), Health Policy Institute; and International Bone Marrow Transplant Registry (Drs. Loberiza, Klein, Rizzo, Bredeson, and Horowitz), Medical College of Wisconsin, Milwaukee, WI.
Correspondence to: Linus H. Santo Tomas, MD, MS, Division of Pulmonary and Critical Care Medicine, Medical College of Wisconsin, 9200 West Wisconsin Ave, Milwaukee, WI 53226; e-mail: santo{at}mcw.edu
| Abstract |
|---|
|
|
|---|
Methods: Registry data on 6,275 adult patients with leukemia who received human leukocyte antigen-identical sibling transplants from 1989 to 1997 and survived at least 100 days after transplantation were evaluated for the study. Risk factors for BO were analyzed using proportional hazards regression.
Results: Seventy-six patients were found to have BO, with an incidence rate of 1.7% at 2 years after transplantation. The Kaplan-Meier estimate of median time to onset of BO was 431 days. Histologic evaluation was performed in 36 patients (47%). In 28 patients (37%), diagnosis was based on pulmonary function tests, CT scans of the chest, or a combination of both. On multivariate analysis, the factors that were associated with an increased risk for BO included the following: peripheral blood-derived stem cell, a busulfan-based conditioning regimen, interval from diagnosis to transplant
14 months, female donor to male recipient sex match, prior interstitial pneumonitis, and an episode of moderate-to-severe acute graft-vs-host disease (GVHD).
Conclusion: In addition to corroborating previously reported risk factors, such as acute GVHD and a busulfan-based conditioning regimen, we found that peripheral blood stem-cell transplantation, long duration to transplant, female donor to male recipient, and a prior episode of interstitial pneumonitis are associated with an increased risk for BO.
Key Words: bone marrow transplantation chronic airflow obstruction obstructive lung disease peripheral blood stem cell transplantation
| Introduction |
|---|
|
|
|---|
After the first case of "necrotizing obliterative bronchiolitis" in an allogeneic bone marrow transplant recipient was reported by Roca et al10 in 1982, the syndrome of progressive obstructive lung disease occurring after HSCT was increasingly recognized, with an incidence ranging from 2 to 26%.341112 While histologic diagnosis is available for only a minority of patients with posttransplant obstructive lung disease, BO is identified as the most common pathologic finding in patients undergoing biopsy or at autopsy.1314 Despite diagnostic uncertainties, there is a growing tendency to attribute most cases of posttransplant chronic airflow obstruction to BO even in the absence of histologic confirmation.
Previous studies of posttransplant obstructive lung disease and BO are either case reports or case series. The larger retrospective cohort studies121516 include patients treated in single institutions. Moreover, most previous studies on risk factors for BO lack adequate consideration of multiple potential predictors and adjustment for confounding variables (Table 1 ). Evaluating the occurrence of BO in HSCT recipients using a large observational database, like that maintained by the IBMTR, provides an opportunity to validate findings from previous smaller studies and offers increased statistical power. The objective of this study was to identify patient-, disease-, and transplant-related risk factors for BO in persons receiving allogeneic HSCT for leukemia. Some of the results presented in this article have been previously reported in the form of an abstract.31
|
| Materials and Methods |
|---|
|
|
|---|
The IBMTR collects data at two levels: registration and research. Registration data include disease type, age, sex, pretransplant disease stage and chemotherapy responsiveness, date of diagnosis, graft type (bone marrow- and/or blood-derived stem cells), high-dose conditioning regimen, posttransplant disease progression and survival, development of a new malignancy, and cause of death. Requests for data on progression or death for registered patients are at 6-month intervals. All IBMTR teams contribute registration data. Research data are collected on subsets of registered patients and includes comprehensive pretransplant and posttransplant clinical information. Computerized checks for errors, physician reviews of submitted data, and on-site audits of participating centers ensure the quality of data.32
Patient Selection
The study population included all patients
18 years of age receiving human leukocyte antigen (HLA)-identical sibling HSCT for leukemia from January 1989 to December 1997, who were reported to the IBMTR and who survived at least 100 days after transplantation. Only patients receiving their first transplant were considered. Patients with pretransplant diagnoses of asthma, COPD, emphysema, chronic bronchitis, and bronchiectasis were excluded from the analysis.
Outcomes
The main outcome of interest is the reported development of BO. The diagnosis of BO was made using any of the following methods: transbronchial biopsy, open-lung biopsy, or autopsy. Pulmonary function studies, CT scan of the chest, or a combination of both were also considered valid diagnostic tests. The time interval from the date of transplantation to the date of diagnosis of BO was noted.
Predictor Variables
Variables evaluated for association with BO are summarized in Table 2
. Acute graft-vs-host disease (GVHD) was classified as none to mild (grades 0 to I) vs moderate to severe (grades II to IV). Data on the development of chronic GVHD were also retrieved. However, since the questionnaire used for collecting the data listed BO as a possible manifestation of chronic GVHD, an unbiased assessment of the association between these two variables could not be undertaken. This confounding by definition was not a problem with acute GVHD and allowed us to analyze the association of BO with acute GVHD. The third quartile was used as the cutoff point to dichotomize the time interval from diagnosis to transplant based on previous findings that a longer interval is a risk factor for poorer HSCT outcome.733
|
Since status may change on follow-up for any given individual, interstitial pneumonitis and acute GVHD were entered as time-dependent variables. Time to the first onset of these variables was measured from date of transplantation. Missing data were handled using the complete-subject analysis method.3536 This led to deletion of 371 subjects (1 with a diagnosis of BO) in the final model due to missing covariate values.
| Results |
|---|
|
|
|---|
90. However, other coexisting disease or organ impairment was noted in 23%. A comparable number of recipients (62%) and donors (58%) were found to be seropositive for cytomegalovirus (CMV) prior to transplantation. Thirty-five percent of recipients had a positive smoking history, with a median of 10 pack-years of smoking.
|
Diagnosis of BO
Of the 6,275 cohort members, BO was diagnosed in 76 patients, yielding a probability of 1.7% at 2 years. The Kaplan-Meier estimate of median time interval from transplant to the diagnosis of BO was 431days. All but one case were diagnosed by 1,600 days (52 months). The earliest case was diagnosed at 65 days and the latest at 2,444 days after transplantation (Fig 1
). Five of the patients with a BO diagnosis did not have manifestations of chronic GVHD in other organs. As shown in Table 4
, a diagnostic procedure for the evaluation of BO was reported in 64 of the 76 cases (84%). Only 36 patients underwent an invasive procedure to attempt histologic confirmation. The remainder underwent pulmonary function studies, CT scans of the chest, or a combination of both to support the diagnosis of BO. The original reports and results of these diagnostic tests were not available for review.
|
|
14 months (HR, 1.93; 95% CI, 1.22 to 3.07; p = 0.0053); and female donor to male recipient sex match (HR, 1.78; 95% CI, 1.12 to 2.85; p = 0.0152). Posttransplant factors associated with an increased risk for BO after multivariate analysis included: moderate-to-severe acute GVHD (HR, 2.12; 95% CI, 1.34 to 3.37; p = 0.0014), and a previous episode of interstitial pneumonitis (HR, 2.28; 95% CI, 1.33 to 3.92; p = 0.0029). The median interval from diagnosis of interstitial pneumonitis to diagnosis of BO was 114 days (interquartile, 38 to 320 days).
|
| Discussion |
|---|
|
|
|---|
To our knowledge, a higher risk for developing BO with PBSCT, as compared to BMT, has not been reported previously. Selection of stem-cell source may be influenced by patient-donor characteristics. We used multivariate analyses to adjust for the potential confounding effects of the variables listed in Table 2. Some studies37383940 have reported a higher incidence of chronic GVHD or more severe chronic GVHD with PBSCT compared to BMT, but the incidence of BO was not different. However, these studies were not designed nor powered to detect this difference.
In consonance with previous studies,101213141617192124284142 we found that GVHD is a risk factor for BO. Acute GVHD is described as a risk factor not only for BO but also for subsequent chronic GVHD. Though we are able to satisfactorily estimate the association of acute GVHD with BO, we are unable to obtain an unbiased estimate of the effect of chronic GVHD because the registry questionnaire incorporates BO as a manifestation of chronic GVHD. This introduces a differential misclassification leading to a greater estimate of association between chronic GVHD and BO than might be otherwise observed. Nevertheless, 5 of the 76 patients (7%) with BO did not have manifestations of chronic GVHD in other organs. A study by Dudek et al9 found a higher proportion of patients (13%) who acquired BO without chronic GVHD.
It is unclear how many patients were classified as having chronic GVHD based solely on the diagnosis of BO. It is noteworthy that six patients received a diagnosis of BO prior to a diagnosis of chronic GVHD. Although BO preceding development of chronic GVHD has been reported, most other authors have observed the contrary.1316234344 It is possible that there are more patients who have chronic GVHD who also have undiagnosed BO. This could happen with subclinical airflow obstruction or mild disease. A more standardized criteria and strategy for screening may identify these patients better in future studies. Both BO and chronic GVHD can present insidiously, and the fact that they ultimately occur within a specified period of time may be more relevant clinically than their exact temporal relationship. It remains to be determined whether BO in HSCT recipients with chronic GVHD results from autoimmune injury to the bronchial tree or other complications that accompany chronic GVHD.
Though both busulfan and irradiation have been described to cause pulmonary toxicity, we found that a busulfan-based conditioning regimen increased the risk for BO at least twofold when compared to a TBI-based regimen. The possible detrimental effect of a busulfan-based regimen on development of late pulmonary impairment following allogeneic HSCT was previously described by Beinert et al.6 CML patients were more likely to receive a busulfan-based regimen, which may explain previously reported higher risk for BO in this subgroup.22
The adverse influence of a long interval from diagnosis to transplantation on overall posttransplant survival, as well as on interstitial pneumonitis, has been described in other investigations.733 An association with BO has not been previously reported. Some authors suggest that this reflects the cumulative toxic sequelae of chemotherapy administered prior to transplantation.3341
Interstitial pneumonitis usually occurs much earlier than BO, with a median onset at 50 to 80 days.3345 Up to 40% of transplantation-related deaths in the early posttransplant period (< 100 days) are attributed to interstitial pneumonitis.2433 It is likely that those who survive an episode of interstitial pneumonitis have residual scarring of both the interstitium and peribronchiolar structures. In a study of six patients with biopsy-proven BO reported by Holland et al,23 three patients had concomitant histologic evidence for interstitial pneumonitis. It is also possible that survivors of interstitial pneumonitis may have decreased respiratory function reserve, which can lead to more respiratory complaints and subsequent surveillance bias.
The association between BO and donor-recipient sex match has been reported previously by Chien et al,12 although the association did not persist in multivariate analysis. Several authors have determined the importance of sex-match status in other transplant outcomes such as GVHD and survival.46474849 It has been suggested that alloreactivity to Y-chromosome antigens and other minor sex-related antigens may play a role in the pathogenesis GVHD. In addition, previously parous women may have experienced maternal alloimmunization to unshared fetal antigens.505152 Thus, it is possible that sex-match status may influence the development of BO, at least in part, via mechanisms related to GVHD. Identifying sex-match status as an independent risk factor for BO may be considered as supportive evidence that an immune-mediated injury to the bronchiolar structures is involved in the pathogenesis of BO.
We limited our investigation to adult allogeneic transplant recipients since previous studies155354 have suggested that this population may be at greater risk for BO. However, the estimated risk for BO in this study, 1.7% at 2 years, is lower than noted in most previous investigations. This discrepancy may be due to differing diagnostic criteria and definition of the population at risk.55 Variability in the diagnostic surveillance practices of the institutions contributing data to the IBMTR may also play a role in this discrepancy.
The Kaplan-Meier estimate of median interval from transplant to diagnosis of BO (431 days) falls within the range of previous observations, as reflected in Table 1. It should be noted that there was one patient in our study in whom BO was diagnosed after 52 months. Using the proportional hazards model, this outlier may potentially bias the estimated risk effect toward the direction of the subjects covariate values. However, for diseases with small or more typical risk rates, as in BO, this effect is negligible.
The advantage of a strategy aimed toward early detection and treatment of BO after HSCT remains controversial.13162842 In a study of 29 patients with chronic respiratory impairment after allogeneic transplantation, Schwarer et al56 noted partial or complete improvement in 83% after a course of immunosuppressive agents. However, only 62% of the patients in that series had a predominantly obstructive ventilatory impairment; which raises the question of heterogeneity of disease. In another study, Sanchez et al30 followed the course of 20 patients with obstructive airway disease who were treated with immunosuppression; 60% had complete or partial improvement in symptoms or pulmonary function tests. The experiences of other investigators have been less encouraging.15182023
Our findings support the argument that BO after HSCT is a multifactorial disease. A survey of the risk factors identified in this study suggests that a variety of mechanismsimmune-mediated, and treatment-related toxicityalone or in combination, may be responsible for the pathogenesis of BO. It is likely that the responsiveness to treatment of BO depends not only on the histologic stage (early lymphocytic bronchiolitis vs complete cicatricial obliteration of the bronchiolar lumen), but also on the pathophysiology involved. The findings in this study can be used to prospectively stratify patients into risk categories to facilitate investigating the merits of an aggressive diagnostic and treatment strategy in particular risk groups.
In accordance with previous descriptions of BO as a new-onset obstructive lung disease in HSCT recipients, patients with obstructive lung disease prior to transplant were excluded from the analysis. This was also done to avoid surveillance and misclassification bias. The latter is especially a concern since we relied on the reported diagnosis of BO, with the criteria defined by individual centers. In the absence of prespecified criteria to distinguish BO from progression or exacerbation of prior obstructive lung disease, it would not be possible to remove the apparent effects of such a bias.
The analysis of this registry data, while providing a large number of patients from different institutions, also presents limitations. One concern is the underestimation of the incidence of BO since the registry questionnaire was not specifically designed to address this. This study used the reported diagnosis of BO as the main outcome of interest. Fewer than half of the patients diagnosed with BO had a diagnostic procedure to obtain histologic confirmation. This is comparable with the experience of other investigators (Table 1). As illustrated by these previous investigations, histologic verification of diagnosis is not always practical. In part, the difficulty of histologic diagnosis relates to its invasiveness in the context of patients that may be severely ill. Moreover, even if histologic specimens are obtained, BO may not be found due to the patchy distribution of the disease and tissue sampling limitations.57 Thus, even in lung transplant recipients, in whom BO eventually develops in 50%, clinical and physiologic criteria are used in management and research.58596061
Unlike studies on BO in lung transplant recipients, there is less uniformity in clinical criteria used for the diagnosis of BO in HSCT at present. A liberal spirometric criteria identifying airflow obstruction as an annualized FEV1 decline of
5% and FEV1/FVC ratio < 0.8 as proposed by Chien et al12 may overestimate the incidence of patients who have histologic BO. However, using these spirometric criteria to define probable BO may have the benefit of identifying progressive disease at an earlier stage, and will allow conduct of prospective studies to determine if early intervention can limit deterioration of obstruction and alter other outcomes. Chien et al62 have recommended closer monitoring of spirometry during the first year based on their findings that suggest patients who have the highest rate of decline in lung function during the first year have the highest mortality. Considering the importance of BO in influencing the course of HSCT outcomes, a consensus on a diagnostic strategy, and clinical criteria that includes spirometric and radiologic findings needs to be established.
We conclude that several pretransplant and posttransplant factors were associated with an increased risk for BO. Our results corroborate previously reported risk factors, such as acute GVHD and a busulfan-based conditioning regimen. In addition, we found that PBSCT, long duration to transplant, female donor to male recipient, and a prior episode of interstitial pneumonitis are associated with an increased risk for BO.
| Footnotes |
|---|
Received for publication September 24, 2004. Accepted for publication December 14, 2004.
| References |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |