| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Medicine (Mr. Henke and Drs. Golden, Yelin, and Blanc), Cardiovascular Research Institute (Drs. Golden and Blanc), and Department of Surgery (Dr. Keith), University of California San Francisco, San Francisco, CA.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Design: Retrospective review and analysis of 498 bronchoscopies with transbronchial biopsy (TBB) and BAL performed in 34 patients after lung transplantation.
Setting: University-based, tertiary referral medical center.
Patients: Thirty-four patients after lung
transplantation. The mean age at transplantation was 49 ± 9 years;
20 (59%) were female. Twenty-four (71%) underwent single and 10
(29%) underwent bilateral lung transplantation. The most common
pretransplantation diagnostic groups were emphysema/COPD without
concomitant 
1-antiprotease deficiency (n = 13) and
other obstructive disease processes (n = 10).
Interventions: Over follow-up, subjects underwent multiple bronchoscopies with TBB and BAL. The median number per subject was 15 (25 to 75% range 13 to 17).
Measurements and results: We calculated the overall median BAL WBCs and median percent neutrophils (polymorphonuclear leukocytes [PMNs]) among all of the BALs performed for each subject. We then calculated the mean ± SD of those median values. We used Cox proportionate hazards to assess mortality risk. The median overall follow-up observation period for the cohort was 560 days. There were 11 deaths during this period. Twenty-four subjects (71%) had acute rejection (AR) grades 2 to 4 (mild to severe), and nine (27%) had obliterative bronchiolitis (OB) diagnosed by TBB at any point. The mean value for BAL WBCs was 366 ± 145 x 103 per milliliter; for percentage PMNs, the mean was 7 ± 10%. Adjusting for age, gender, single vs bilateral lung transplantation, pretransplantation diagnostic group, presence of AR, presence of OB, BAL WBC concentration, and lymphocyte CD4/CD8 ratio, PMN percent was a significant predictor of mortality (p = 0.02).
Conclusions: Ongoing inflammation manifested by an increased percentage PMNs over repeated bronchoscopies predicts mortality following lung transplantation. Biopsy data alone may be insufficient to identify posttransplantation patients at risk of poor outcome.
Key Words: bronchoalveolar lavage • bronchoscopy • lung transplant • neutrophils • transbronchial biopsy
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The greatest threats to patient survival posttransplantation are infection, acute rejection (AR), and obliterative bronchiolitis (OB).2 ,3 ,4 ,5 ,6 The distinction between infection and AR can be difficult based on clinical symptoms alone, often requiring confirmation through transbronchial biopsy (TBB) specimens and BAL.6 ,7 ,8 ,9 In contrast to AR and infection, OB is often diagnosed on the basis of lung function, although TBB specimens can be confirmatory.2 ,10 ,11 The role of BAL in evaluating OB remains to be established. Nevertheless, a recent review noted that although research into markers of OB in BAL remains rudimentary, it is an area of considerable promise.2
We studied 34 patients after lung transplantation to evaluate whether findings from surveillance bronchoscopies, including both TBB and BAL, were predictive of patient survival.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Postoperative Bronchoscopic Evaluations
All lung transplantation patients, as part of a routine clinical
surveillance protocol, underwent regular, periodic surveillance
bronchoscopies with BAL and transbronchial lung biopsies. These
evaluations were carried out on a preset schedule of every 2 weeks for
the first 2 months after transplantation, followed by every month for
the next 3 months, then every 3 months for 6 months and, finally, at
6-month intervals. Eighteen-month follow-up, therefore, would typically
include 10 surveillance bronchoscopies. Additional bronchoscopies were
also carried out as indicated on clinical grounds.
Bronchoalveolar Lavage
The distal tip of the fiberoptic bronchoscope (Pentax FB-19 TX;
Pentax Precision Instrument Corp; Orangeburg, NY) was wedged into the
allograft in either the right middle lobe or left lingular bronchus.
Sterile normal saline solution was instilled in 20-mL boluses and
aspirated serially with collection in sequential suction traps
(Sherwood Medical; St. Louis, MO) for a total lavage instillation of
approximately 150 mL. Each of three sequential traps was filled with 20
to 30 mL of return (approximately 50% of instilled saline solution).
The first trap collected was used for microbiological cultures,
including cytomegalovirus (CMV) culture. Total cell count and
differential count were assessed from the second trap and were
performed by the clinical laboratories of UCSF using standard
hemocytometer and staining techniques, as previously
published.12
The third trap was used for lymphocyte
phenotyping, performed by the UCSF Clinical Immunology Laboratory using
automated flow cytometry with fluorescent-labeled monoclonal antibodies
directed against standard cell surface antibodies.
Examining microbiological culture results, we identified all subjects in whom BAL cultures tested positive for CMV at any time posttransplantation and those who were CMV positive within the first 3 months posttransplantation. We defined these latter subjects as early CMV positive for the purposes of this analysis. Because CMV diagnosis early posttransplantation has been found to be a risk factor for poorer outcomes13 and because survival until later culture positivity would, of necessity, be linked to overall survival, we examined only early BAL culture positivity as a predictor in our model.
Transbronchial Lung Biopsies
Eight to 10 transbronchial lung biopsy specimens were routinely
obtained immediately following BAL. Using biopsy forceps (Pentax
Crocodile, Large Cup Biopsy Forceps KA-2411s; Orangeburg, NY),
specimens were obtained from two different subsegments of the lower
lobe. AR was classified histologically as minimal to severe (grades 1
through 4) according to standard criteria of the "Working Formulation
of the Lung Study Group."14
For our data analysis, we
considered AR to have occurred if present histologically and graded
mild, moderate, or severe (grades 2 to 4) in any of the serial biopsy
specimens. OB was graded histologically based on the degree of scarring
and fibrosis of small peripheral airways and designated as total or
subtotal, active or inactive. We considered OB to have occurred if
present histologically in any designation in any of the serial biopsy
specimens.
Pulmonary Function Testing
All subjects underwent serial pulmonary function testing (PFT)
according to standard American Thoracic Society criteria15
with at least as much frequency as surveillance bronchoscopy. This
allowed us to develop an alternative definition of bronchiolitis
obliterans defined by decline in FEV1 rather than biopsy.
We employed the standard approach to the designation of physiologic
bronchiolitis obliterans syndrome (BOS) defining BOS dichotomously when
present as stage 2 or 3 (
65% of the baseline posttransplantation
FEV1 value).16
Data Collection
Demographic data and the clinicopathologic diagnosis leading to
lung transplantation were extracted from a clinical database maintained
on all transplantation patients. Based on these data, we collapsed the
patients into one of three pretransplantation diagnostic categories:
chronic obstructive lung disease associated with cigarette smoking and
without 1-antiprotease deficiency; other obstructive
lung diseases, including cystic fibrosis, bronchiectasis, and
1-antiprotease deficiency (with or without concomitant
cigarette smoking); and interstitial lung disease including idiopathic
pulmonary fibrosis. We also extracted from the database the results of
TBB and BAL studies, as well as date of transplantation surgery and the
date of death, if death occurred. PFT results were extracted from the
laboratory's computerized record system. We confirmed BAL and biopsy
data with a second, hospital-wide computerized laboratory database.
Data Analysis
We analyzed the data using a standard statistical package (SAS;
Cary, NC). We tested differences between survivors and nonsurvivors for
BAL polymorphonuclear leukocyte (PMN) percentage using the Wilcoxon
rank sum test. We used Fisher's Exact Test for the association between
BOS and OB. Using Cox proportionate hazard modeling, we studied the
relationship between the predictors of study interest and time elapsed
until death. This technique best utilizes data from subjects followed
up for varying lengths of time. Observations were censored if patients
were alive as of September 30, 1996. The predictive variables we
studied fell into four groups: core demographic and clinical data at
the time of transplantation; BAL cell and culture data; transbronchial
biopsy data; and pulmonary function results.
The following were core demographic and clinical predictors we
initially included in the Cox model: age at transplantation, gender,
single vs bilateral lung transplantation, and diagnostic group. The
diagnostic groups of smoking-related COPD exclusive of
1-antiprotease deficiency and emphysema as one category
and obstructive diseases including cystic fibrosis and
1-antiprotease deficiency as a second category were
treated as two separate indicator variables. In this way, fibrotic or
vascular lung disease served as the remaining, referent diagnostic
category. We chose these potential predictors on an a priori
basis.
The BAL cell variables analyzed as predictors of mortality were the absolute concentration of leukocytes and the percentage of PMNs relative to total leukocytes. We used the median rather than mean value of observed BAL leukocytes and PMN percentage over all of the BAL studies performed during the observation period for each patient. We did not include in the same model the percent macrophages because this was strongly inversely correlated with the percentage of PMNs. We also tested as additional predictors the median ratio of CD4 to CD8 BAL lymphocytes over the sequential studies and CMV-culture positivity within 3 months posttransplantation. As detailed previously, the TBB findings were defined dichotomously as two variables: rejection (present or absent) or OB (present or absent). The PFT-derived variable we tested was that of BOS defined by FEV1 decline.
We tested Cox proportionate hazard models containing the core
demographic clinical predictors with BAL leukocytes and percentage of
PMNs. We reestimated the model adding the biopsy variables of OB and AR
together with the CD4/CD8 ratio. We also tested the additional
explanatory power of CMV-culture positivity. We compared the predictive
power of models by testing the difference in model 
2
with and without the added variables in question. We also retested the
model substituting PFT-defined BOS for OB defined histologically.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antiprotease deficiency; other
obstructive lung diseases, including 1-antiprotease
deficiency (n = 6), bronchiectasis (n = 3), and cystic fibrosis
(n = 1); and restrictive or vascular disease, including idiopathic
pulmonary fibrosis (n = 5), primary pulmonary hypertension (n = 3),
and other conditions, including secondary pulmonary hypertension
(n = 3).
|
|
Table 2 also includes the results of surveillance TBBs. Although 79% of the subjects (n = 27) had evidence of either OB or AR at some point, only 6 subjects had both findings at any point in time. Twelve subjects (35%) had BOS (grade 2 or 3) by PFT at some point posttransplantation, including seven (21%) who also had OB by TBB specimen. BOS and OB were statistically associated (Fisher's Exact Test, p = 0.004).
A comparison of BAL data for median percentage of PMNs among those alive at follow-up as compared with nonsurvivors is shown in Figure 1 . The median value for the survivors was 2%, while for nonsurvivors it was 7%. This difference was statistically significant (Wilcoxon rank sum test, p = 0.0003).
|
The initial 2 of the Cox proportionate hazards model was
17.7 (p = 0.014) including only the seven core variables of median
BAL; ie, total WBC and PMN percentage and five key clinical
and demographic variables at baseline: age, gender, single vs bilateral
lung transplantation, and two of the three pretransplantation
diagnostic groups (with the third as the default referent). We then
built upon the model by testing the impact of five additional
variables: TBB AR, TBB OB, BAL lymphocyte CD4/CD8 ratio, early
CMV-culture positivity, and PFT BOS.
Adding the two biopsy variables of OB and AR to the model increased its
predictive power marginally (model 2 difference = 5.5,
2 df; p = 0.06). However, when the three additional
variables of CD4/CD8 ratio, OB, and AR were together added to the same
seven-variable core model, its overall explanatory power increased
significantly (model 2 difference = 9.3, 3
df; p = 0.03). Addition of early CMV BAL culture
positivity did not significantly increase the model's explanatory
power (model 2 difference = 1.3, 1 df;
p = 0.26). CMV positivity was not statistically predictive of
mortality (Wald 2 = 1.2; p = 0.26) and did not
account for the association between percent of PMNs and mortality in
the same model (p = 0.019 for PMN percentage after inclusion of CMV
in model).
Substitution of BOS defined by PFT for biopsy specimen-defined OB
provided less explanatory power, both for the risk factor itself
(direction of risk changed from positive to negative; Wald
2 = 0.2; p = 0.68) and for the model overall (model
2 21.3 vs 27.0, 10 df for each model).
Substitution of BOS did not substantively change the association
between PMN percentage and mortality.
The final model is presented in Table 3 . As shown, the concentration of BAL PMNs as a median proportion of leukocytes over all sequential studies over follow-up was a powerful predictor of poor survival. This takes into account all of the other variables in the model, many of which were also significant predictors of mortality.
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A great deal of research has focused on the diagnosis of discrete episodes of infection or AR among lung transplantation recipients.3 ,5 ,6 ,8 ,9 ,19 ,20 ,21 ,22 ,23 There has been little study, however, of the impact of persistent lung inflammation following transplantation, even during periods of relative health. Rather, studies of BAL findings have generally focused on isolated BAL results at a specific point in time.7 ,8 ,17 ,19 ,21 ,22 ,23
In this study, we analyzed the results of 498 bronchoscopies, including TBB and BAL, performed in 34 patients after lung transplantation. One strength of this study is the wealth of follow-up data on each patient. Our protocol specified nine surveillance bronchoscopies in the first year posttransplantation, with additional diagnostic bronchoscopies in response to clinical indications. We had available for analysis a median of 15 bronchoscopies per patient during the study period. Previously published research has been based on a much smaller selection of biopsy or BAL data per patient, ranging from 1.8 to 7.3 bronchoscopies per patient.3 ,6 ,7 ,8 ,9 ,10 ,19 ,20 ,21 ,22 ,23 ,24 The large number of observations for each patient allows us to better examine the overall pattern of bronchoscopic data, especially during periods without clinical complications, that might prompt diagnostic bronchoscopy.
For each patient, we analyzed the median value of the percentage of PMNs relative to all leukocytes, integrated over all of their follow-up bronchoscopies. A high PMN percentage observed in a single lavage could indeed reflect an episode of infection or of AR.7 Nevertheless, analysis of the entire posttransplantation follow-up period and the use of the median instead of the mean minimizes the influence of isolated extreme measurements due to discrete clinical events. It also dampens the impact of experimental or nonnormally distributed random variation.
Our analysis demonstrates that a persistently elevated neutrophil percentage is a statistically significant predictor of mortality following lung transplantation. In fact, the long-term pattern of BAL cell counts was at least as powerful a predictor of mortality risk as were biopsy diagnoses of AR or OB, complications generally considered the most potent threats to lung transplantation recipients. Moreover, the neutrophil-associated risk did not appear to be affected by the other potential confounding factors we included in our model, which accounted for key clinical and demographic covariates.
There are several potential explanations for the apparent risk imputed by chronic BAL inflammation independent of biopsy specimen-identified AR or OB.
Since PMN percentage increases substantially in the presence of infection, the observed elevated PMN percentage could reflect multiple episodes of acute infection.7 ,23 ,24 However, the use of the median PMN percentage value instead of the mean, derived from the large number of bronchoscopies being studied, makes this unlikely.
Previous research has indicated that early CMV infection following lung transplantation is a risk factor for worse outcome (BOS).13 Nevertheless, in our model, the presence of CMV in BAL culture early in the posttransplantation course failed to account for the predictive power of PMN percentage and was not itself a significant predictor of mortality, indicating that the elevated PMN percentage measured was not simply a marker of CMV infection.
TBB is recognized as the procedure of choice for the diagnosis of AR.3 ,9 Our model took biopsy specimen-detected AR into account, but this failed to explain the predictive power of BAL PMN percentage in the multivariate analysis. Moreover, as with infection, the impact of isolated episodes of AR would be unlikely to have a substantial effect of the median value of PMN percentage. Given our analytic strategy, this would not explain the observed relative weakness of rejection diagnosis in predicting mortality. Similarly, if PMN percentage were solely a marker of insufficient pharmacologic immunosuppression, inclusion of the lymphocyte phenotype data should have captured this effect.7 ,25
Previous research has indicated that the sensitivity of TBB for OB may be poor18 ,10 (although this observation has not been universal11 ). The limited power of an OB diagnosis as a predictor of mortality in our study may reflect the limitations of TBB as a diagnostic tool, to the extent that occurrences of OB among the study patients went undiagnosed prior to death. Our study showed a cumulative incidence of OB diagnosed by TBB specimens of 26% at any point during the study period, on the low end of the commonly reported range of 25 to 40%.26 ,27 Substituting BOS, defined as a decline in PFT FEV1, for OB provided even less explanatory power and still failed to account for the impact of the PMN percentage. Since elevated neutrophil percentages in BAL have been associated with the presence of infection, OB, or AR,7 ,23 ,24 the elevated PMN percentage that we observed may actually be a marker for episodes of OB undetected by simultaneous TBB. This suggests that surveillance BAL may indeed be a particularly valuable tool for early detection of OB, providing greater sensitivity than TBB or PFT alone.
DiGiovine and coinvestigators24 have made the intriguing observation that an elevated level of neutrophils both predicts the onset of OB and predates diagnosis by nearly a year. They also found that neutrophilia persisted even when patients were treated for OB. This relationship would be consistent with our findings.
The possibility remains that the elevated PMN percentage in fact represents a risk factor independent of acute infection, AR, or OB. For example, stimulation of proinflammatory cytokines through neural or other pathways could lead to an elevated PMN percentage. If so, new anti-inflammatory drugs that interfere with such cytokines may be employed in the future in situations with bronchoscopic evidence of elevated PMNs. Further research is needed to clarify the potential etiology and impact of ongoing, low-grade inflammatory cellular response following lung transplantation.
Our findings reinforce the potential importance of surveillance bronchoscopy with BAL following lung transplantation. Based on these data, persistent neutrophilia should be considered a significant potential threat to posttransplantation patient survival.
|
Footnotes |
|---|
Abbreviations: AR = acute rejection; BOS = bronchiolitis obliterans syndrome; CMV = cytomegalovirus; OB = obliterative bronchiolitis; PFT = pulmonary function test; PMNs = neutrophils (polymorphonuclear leukocytes); TBB = transbronchial biopsy; UCSF = University of California, San Francisco
Received for publication April 28, 1998. Accepted for publication September 19, 1998.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Liu, P. Drew, Y. Cheng, and G. A. Visner Pirfenidone inhibits inflammatory responses and ameliorates allograft injury in a rat lung transplant model J. Thorac. Cardiovasc. Surg., September 1, 2005; 130(3): 852 - 858. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. G. Gerhardt, J. F. McDyer, R. E. Girgis, J. V. Conte, S. C. Yang, and J. B. Orens Maintenance Azithromycin Therapy for Bronchiolitis Obliterans Syndrome: Results of a Pilot Study Am. J. Respir. Crit. Care Med., July 1, 2003; 168(1): 121 - 125. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Helmi, R. B. Love, D. Welter, R. D. Cornwell, and K. C. Meyer Aspergillus Infection in Lung Transplant Recipients With Cystic Fibrosis: Risk Factors and Outcomes Comparison to Other Types of Transplant Recipients Chest, March 1, 2003; 123(3): 800 - 808. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.-J. SLEBOS, J. SCHOLMA, H. MARIKE BOEZEN, G. H. KOETER, W. VAN DER BIJ, D. S. POSTMA, and H. F. KAUFFMAN Longitudinal Profile of Bronchoalveolar Lavage Cell Characteristics in Patients with a Good Outcome after Lung Transplantation Am. J. Respir. Crit. Care Med., February 15, 2002; 165(4): 501 - 507. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Reynaud-Gaubert, P. Thomas, R. Gregoire, M. Badier, P. Cau, J. Sampol, R. Giudicelli, and P. Fuentes Clinical utility of bronchoalveolar lavage cell phenotype analyses in the postoperative monitoring of lung transplant recipients Eur. J. Cardiothorac. Surg., January 1, 2002; 21(1): 60 - 66. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. C. MEYER, D. R. NUNLEY, J. H. DAUBER, A. T. IACONO, R. J. KEENAN, R. D. CORNWELL, and R. B. LOVE Neutrophils, Unopposed Neutrophil Elastase, and Alpha1-Antiprotease Defenses Following Human Lung Transplantation Am. J. Respir. Crit. Care Med., July 1, 2001; 164(1): 97 - 102. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K M Beeh, O Kornmann, J Lill, and R Buhl Induced sputum cell profiles in lung transplant recipients with or without chronic rejection: correlation with lung function Thorax, July 1, 2001; 56(7): 557 - 560. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Agostini, F. Calabrese, F. Rea, M. Facco, A. Tosoni, M. Loy, G. Binotto, M. Valente, L. Trentin, and G. Semenzato CXCR3 and Its Ligand CXCL10 Are Expressed by Inflammatory Cells Infiltrating Lung Allografts and Mediate Chemotaxis of T Cells at Sites of Rejection Am. J. Pathol., May 1, 2001; 158(5): 1703 - 1711. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Zheng, E H. Walters, C. Ward, N. Wang, B. Orsida, H. Whitford, T. J Williams, T. Kotsimbos, and G. I Snell Airway neutrophilia in stable and bronchiolitis obliterans syndrome patients following lung transplantation Thorax, January 1, 2000; 55(1): 53 - 59. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
