Chest Email Content Delivery
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
First published online on June 15, 2007
Chest, doi:10.1378/chest.07-0175
doi:10.1378/chest.07-0175
(Chest. 2007; 132:504-514)
© 2007 American College of Chest Physicians
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow All Versions of this Article:
chest.07-0175v1
132/2/504    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stolz, D.
Right arrow Articles by Tamm, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stolz, D.
Right arrow Articles by Tamm, M.

BAL Neutrophils, Serum Procalcitonin, and C-Reactive Protein To Predict Bacterial Infection in the Immunocompromised Host*

Daiana Stolz, MD; Andreas Stulz; Beat Müller, MD; Alois Gratwohl, MD and Michael Tamm, MD

* From the Clinics of Pneumology and Pulmonary Cell Research (Drs. Stolz, Stulz, and Tamm), Endocrinology, Diabetes, and Clinical Nutrition (Dr. Müller), and Hematology (Dr. Gratwohl), University Hospital Basel, Basel, Switzerland.

Correspondence to: Daiana Stolz, MD, Assistant Professor, University of Massachusetts Medical School, 55 Lake Ave North, Worcester, MA 01655; e-mail: dstolz{at}uhbs.ch

Abstract

Background: Bacterial pulmonary infection is a common life-threatening complication in immunocompromised patients. The results of BAL cultures are not immediately available, and their microbiological yield might be limited by empiric antibiotic prescriptions. We evaluated clinical signs and symptoms, leukocyte counts, C-reactive protein (CRP) levels, procalcitonin levels, and BAL fluid neutrophil percentages as potential markers for bacterial infection in a cohort of immunocompromised patients with pulmonary complications.

Methods: One hundred seven consecutive patients who had been referred for bronchoscopy due to suspected pulmonary infection were included in this study. Based on clinical, laboratory, radiologic, microbiological, and histologic results, patients were classified as having proven bacterial infection (n = 27), possible bacterial infection (n = 11), and no bacterial infection (n = 69).

Results: Most common underlying conditions were hematologic malignancy (n = 62) and solid organ transplantation (n = 20). Clinical parameters were similar in patients with and without bacterial infection (difference was not significant). The percentage of BAL fluid neutrophils had the highest area under the curve (0.818; 95% confidence interval [CI], 0.700 to 0.935; p < 0.001), followed by absolute neutrophil counts (0.797; 95% CI, 0.678 to 0.916; p < 0.001), procalcitonin level (0.746; 95% CI, 0.602 to 0.889; p = 0.001), and CRP level (0.688; 95% CI, 0.555 to 0.821; p = 0.015) to predict proven bacterial infection (in opposition to no or possible bacterial infection) in the receiver operating characteristic analysis. Conversely, neither infiltrates (p = 0.123) nor leukocyte counts (p = 0.429) were useful in diagnosing bacterial infection. The percentage of BAL fluid neutrophils and procalcitonin level were independent predictors of bacterial infection in the multivariate regression.

Conclusions: Neutrophil percentage in BAL fluid, procalcitonin level, and CRP level might be potentially useful to differentiate bacterial infection from nonbacterial conditions in immunocompromised hosts with pulmonary complications.

Key Words: biomarker • immunosuppression • stem cell transplantation

Pulmonary complications are a major cause of morbidity and mortality among immunocompromised patients.1234567 It has been suggested89 that clinical or radiologic pulmonary abnormalities warranting therapeutic consideration will develop in 30 to 60% of all immunocompromised patients. The accurate diagnosis of the underlying pulmonary disorder is pivotal as it influences medical therapy and outcome.5710 Although the lung is the organ that is most commonly affected by noninfectious abnormalities,11 the identification and severity assessment of bacterial infection is of particular interest, as it represents a major cause of mortality.1213

Diagnostic bronchoscopy is routinely performed in immunocompromised patients.571415 Because of the perceived severity of the baseline condition and the worsened prognosis associated with a delay in antibiotic therapy, direct respiratory tract sampling usually occurs after antibiotic therapy has been instituted.1617 The microbiological yield of BAL fluid cultures is therefore expected to be compromised, ranging widely from 15 to 93%.10141819 Differential cell counts of BAL fluid may serve as a complementary examination to microbiological studies. Herein, higher neutrophil counts have been reported in immunocompromised patients with bacterial pneumonia compared to those patients with no detectable pathogens.20

Circulating proinflammatory mediators such as C-reactive protein (CRP) and, more recently, procalcitonin, have been suggested to be predictive for invasive bacterial infection in the immunocompromised host.212223 Some studies,2425 but not all,26 investigating immunocompromised patients have shown that these patients are capable of producing high serum concentrations of both mediators during severe systemic bacterial infection. While there have been no interventional studies using CRP, procalcitonin was shown to safely support the decision to start antibiotic therapy in respiratory infections in nonimmunocompromised patients.272829 However, in the immunocompromised host, it is thought that there is no sufficient evidence to base a therapeutic decision on.30

We hypothesize that BAL fluid neutrophils and serum biomarkers (CRP and procalcitonin) might be more useful than routinely used signs and symptoms for the diagnosis of bacterial infection in the immunocompromised host. Therefore, we performed a prospective, explorative, observational study to evaluate the diagnostic accuracy of clinical signs and symptoms, leukocyte counts, CRP levels, procalcitonin levels, and neutrophil percentages in BAL fluid as potential markers for bacterial infection in a cohort of immunocompromised patients with pulmonary complications.

Materials and Methods

Patients
This study was approved by the institutional review board and performed at the University Hospital Basel, a 784-bed tertiary care hospital located in Basel, Switzerland. One hundred seven consecutive hospitalized immunocompromised patients who were referred to the pulmonary division for diagnostic bronchoscopy within a period of 8 months in 2005 were included in this prospective cohort study. Eligibility criteria included the following: (1) immunocompromised state; (2) age > 18 years; (3) a decision by the pulmonary consultant to perform bronchoscopy due to suspected pulmonary infection; and (4) informed consent by the patient to undergo flexible bronchoscopy and related data analysis. Suspected pulmonary infection was defined by the presence of respiratory symptoms (eg, cough, sputum, and dyspnea), fever, and/or new or progressive radiologic findings, as judged by the attending physician.

Immunocompromised state was defined by the following conditions: hematologic or other malignancy treated by high-dose chemotherapy and/or stem cell transplantation; solid organ transplantation; AIDS; long-term corticosteroid therapy (ie, ≥ 20 mg/d prednisone equivalent for ≥ 2 months); and current use of immunosuppressive or cytotoxic medication for indications other than organ or stem cell transplantation.

Data Collection
Data collection including demographics, underlying diseases, immunosuppressive drugs, antimicrobial treatment for suspected or documented infection, laboratory findings, and radiologic results were recorded at the time of bronchoscopy using a standardized study form. The collected data also included information related to bronchoscopy. Further data on final diagnosis, clinical outcome, and, if applicable, cause of death including autopsy results were collected after hospital discharge. A review of medical records was performed by two independent, board-certified pulmonary specialists. Discrepancies were settled by consensus.

Bronchoscopy
Flexible bronchoscopy was performed with the patient under conscious sedation using hydrocodone and midazolam according to the British Thoracic Society guidelines.313233 BAL was performed by the three installations of 50 mL each of a pyrogen-free, sterile, 0.9% NaCL solution over the working channel of the bronchoscope according to standard guidelines and as described earlier.143435 In patients with diffuse pulmonary infiltrates, BAL was performed either in the right middle lobe or the lingula. For patients with focal lung infiltrates, BAL was performed in the pulmonary segment corresponding to the radiologic abnormality. BAL fluid was recovered by suction. Besides BAL, the choice of further sampling techniques used during bronchoscopy was at the pulmonologist’s discretion. Major complications were defined as major bleeding, worsening hypoxemia requiring endotracheal intubation, any adverse event requiring early discontinuation of the procedure, and death. Major bleeding was defined as the need for additional interventions such as the placement of a temporary bronchus-blocker, the application of a fibrin sealant, admission to a critical care unit, or the need for treatment with blood products.36

The microbiological diagnostic workup included a search for Legionella pneumophila, Chlamydia pneumoniae, Mycoplasma pneumoniae, and respiratory viruses (ie, herpes simplex virus, cytomegalovirus, respiratory syncytial virus, and adenovirus) by polymerase chain reaction, culture, or immunofluorescence. Appropriate stains and cultures for bacteria, mycobacteria, fungi, and Pneumocystis jiroveci were performed. Positive bacterial cultures were counted as the number of colony-forming units per milliliter, and identification and susceptibility tests were performed according to standard methods. Aspergillus infection was defined by the cytologic or histopathologic demonstration of the presence of Aspergillus or the isolation of Aspergillus in a respiratory specimen in the presence of a compatible clinical and radiographic pattern. Histologic lung biopsy specimens, recovered by transbronchial biopsy, endobronchial biopsy, video-assisted thoracic surgery, and autopsy, were obtained according to the clinical setting. Cytologic analysis of BAL fluid was performed with hematoxylin-eosin staining. Cell differentiation in BAL fluid was reported as absolute numbers and as a percentage of the total cell count.

Diagnostic Criteria
Patients were examined, treated, and followed up according to the usual practice of the institution. Chest radiographs were performed in all cases, and CT scans were obtained as indicated by the treating physicians. Clinical diagnoses were determined from physician notes, hospital discharge summaries, laboratory studies, radiologic examinations, and pathologic reports.

Severe neutropenia was defined as an absolute neutrophil count of < 0.5 x 109 cells/L. For the clinical diagnosis of noninfectious pulmonary complications, we used the criteria outlined in the definitions given in this paragraph. Nonspecific interstitial lung disease was defined as a pattern of interstitial chronic inflammation with type II pneumocyte hyperplasia and a varying degree of fibrosis with no evidence of lung infection, and without alternative histologic diagnosis.37 Organizing pneumonia was diagnosed by the presence of patchy intraluminal organizing fibrosis in distal airspaces with mild interstitial chronic inflammation and the preservation of lung architecture.37 Bronchiolitis obliterans syndrome was defined as progressive airflow limitation due to small airway obstruction after allogeneic stem cell transplantation or lung transplantation.38 Histology was characterized by constrictive bronchiolitis obliterans.139 Respiratory bronchiolitis was defined as a patchy pattern of dusty brown macrophages clustering and a patchy submucosal and peribronchiolar infiltrate of lymphocytes and histiocytes.37 Diffuse alveolar hemorrhage was defined by the presence of a widespread alveolar injury seen on a thoracic CT scan (as evidenced by multilobar infiltrates) and by BAL fluid returns becoming progressively bloodier or showing ≥ 20% iron-laden macrophages in the absence of infection.40 Diffuse alveolar damage was defined as a pattern of acute lung injury, which was characterized by an exudative phase, with cell necrosis, edema, and hyaline membrane formation, and a proliferative phase with organizing interstitial fibrosis.3741 Invasive pulmonary aspergillosis was diagnosed in the context of compatible clinical and radiographic features according to a consensus statement.42 Pulmonary edema was defined by bibasilar rales, radiologic evidence of pulmonary edema, and the resolution of symptoms by therapy with diuretic agents. Acute bronchitis was defined by the self-limited presence of acute cough associated with bronchial system erythema in the absence of infiltrates and bacterial infection. Granuloma was defined by the typical appearance of small nodules (with or without calcification) as evidenced on a chest CT scan without change over time in the absence of bacterial or mycobacterial infection.

Diagnostic Groups
Proven bacterial infection was diagnosed in cases with positive bacteriology results in the BAL fluid or histologic specimens (ie, a culture yielding a single pathogenic bacterial microorganism at the minimum concentration of 103 cfu/mL or any microorganism excluding mouth flora above the minimum concentration of 104 cfu/mL5; or the identification of Legionella spp, C pneumoniae, or M pneumoniae regardless of colony counts). Possible bacterial infection was characterized as a clinically defined infection without a proven microbial pathogen (ie, new infiltrate seen on the chest radiograph accompanied by respiratory symptoms, abnormal breath sounds on auscultation, leukocytosis or leukopenia; and no identification of bacterial, viral, or fungal pathogens in the respiratory tract, cultures, specific stains, polymerase chain reaction, or immunofluorescence, as appropriate). No bacterial infection was considered to be present when an alternative cause for pulmonary symptoms and/or infiltrates was established without evidence of bacteria in microbiological studies (ie, nonspecific interstitial lung disease, organizing pneumonia, bronchiolitis obliterans, graft-vs-host disease, respiratory bronchiolitis, diffuse alveolar hemorrhage, diffuse alveolar damage, invasive pulmonary aspergillosis, pulmonary edema, acute bronchitis, granuloma, Wegener granulomatosis, malignant pulmonary involvement, or fungal or viral infections). Cases fulfilling the diagnostic criteria for proven bacterial infection were classified as such, irrespective of the presence of any alternative causes for the pulmonary symptoms.

Measurement of CRP and Procalcitonin
CRP levels, procalcitonin levels, and leukocyte counts were assessed on the day of bronchoscopy. Procalcitonin was measured using 100 µL of serum by an ultrasensitive chemiluminometric assay (PCT sensitive LIA; Brahms AG; Hennigsdorf, Germany).43 The assay has a lowest standard of 0.005 ng/L and a functional assay sensitivity of 0.02 µg/L, which is within the normal range of healthy persons.44 CRP was measured using an enzyme immunoassay having a detection limit of < 5 mg/dL in heparin plasma (Hitachi Instrument 917; Roche Diagnostics, Rotkreuz, Switzerland) using reagents (Wako Chemicals GmbH; Neuss, Germany) [detection limit, < 5 mg/dL].

Statistical Analysis
Discrete variables are expressed as the count (percentage), and continuous variables are expressed as the mean ± SD, unless otherwise stated. The frequency comparison was performed with the {chi}2 test. Two-group comparisons of normally distributed data were performed using the Student t test. For data not normally distributed, the Mann-Whitney U test was used if only two groups were compared, and the Kruskal-Wallis one-way analysis of variance was used if more than two groups were being compared. To evaluate the relationship among procalcitonin level, CRP level, and BAL neutrophil percentage and proven bacterial infection, a logistic multivariate regression model analysis was performed. We log-transformed skewed variables for the regression analysis. We constructed receiver operating characteristic (ROC) curves and determined the area under the curve (AUC) for the ROC. Significant predictors for bacterial infection (BAL neutrophil percentage log x procalcitonin log) were combined, and the predicted probability derived from logistic regression was used to construct a ROC curve. For these binary analyses, patients classified as having proven bacterial infection were plotted against the other diagnostic groups. Correlation analyses were performed by using Spearman rank correlation. Standard definitions of sensitivity, specificity, and likelihood ratio were used. The AUC was considered to be clinically useful if it was ≥ 0.8.45 All statistical tests were two-tailed; a p value of < .05 was considered to be significant. Data were analyzed using a statistical software package (SPSS, version 14 for Windows; SPSS Inc; Chicago, IL).

Results

Baseline Characteristics of Patients
Underlying conditions and baseline characteristics of the 107 patients are presented in Tables 1 and 2. Most common underlying conditions were hematologic malignancies that were treated by high dose chemotherapy alone (n = 29) or combined with allogeneic stem cell transplantation (n = 26) followed by solid organ transplantation (n = 20). Overall, 20 patients (18.7%) had neutropenia at the time of bronchoscopy.


View this table:
[in this window]
[in a new window]

  		Table 1.. Underlying Diagnosis of 107 Immunocompromised Patients

 

View this table:
[in this window]
[in a new window]

  		Table 2.. Characteristics of 107 Patients at the Time of Bronchoscopy*

 
The median length of hospital stay before bronchoscopy was 2 days (interquartile range [IQR], 1 to 12.5). Fever and cough were the most commonly observed symptoms. Infiltrates were documented in the majority of cases (71%). The most common immunosuppressive medications used were steroids (43.9%), cyclosporine (30.8%), and mycophenolate mofetil (28.0%).

A total of 78 patients (72.8%) received therapy with antibiotics for either prophylactic or therapeutic reasons before bronchoscopy. In patients receiving a therapeutic regimen, the median duration of antibiotic therapy at the time of bronchoscopy was 1 day (IQR, 1 to 4 days). Fifteen patients (19.2%) were receiving piperacillin/tazobactam, 7 patients (9.0%) cefepime, 4 patients (5.1%) meropenem, 3 patients (3.8%) ceftriaxone, 3 patients (3.8%) amoxicillin/clavulanic acid (1 of those patients was also receiving prophylactic trimethoprim/sulfamethoxazole), 3 patients (3.8%) amikacin and cefepime, 3 patients (3.8%) piperacillin/tazobactam and cefepime, 2 patients (2.6%) clarithromycin, 2 patients (2.6%) piperacillin/tazobactam and levofloxacin, 2 patients (2.6%) obracin and cefepime, 2 patients (2.6%) amoxicillin/clavulanic acid and clarithromycin (1 of those patients was also receiving prophylactic trimethoprim/sulfamethoxazole), 2 patients (2.6%) piperacillin/tazobactam, cefepime, and amikacin, 1 patient (1.3%) trimethoprim/sulfamethoxazole, 1 patient (1.3%) levofloxacin, 1 patient (1.3%) imipenem and cilastatin, 1 patient (1.3%) vancomycin and cefepime, 1 patient (1.3%) piperacillin/tazobactam and vancomycin, 1 patient (1.3%) piperacillin/tazobactam and meropenem, 1 patient (1.3%) ceftriaxone and clarithromycin, 1 patient (1.3%) cefepime and amikacin, and 1 patient (1.3%) rifampin and isoniazid and pyrazinamide. Furthermore, 20 patients (25.6%) were receiving trimethoprim/sulfamethoxazole prophylaxis, and 1 patient (1.3%) was receiving continuous therapy with clarithromycin as an antiinflammatory drug.

Table 3 presents median values for leukocyte counts, CRP level, procalcitonin level, BAL fluid neutrophil percentage, and absolute counts in the different immunocompromised patient subgroups. While the median CRP level, procalcitonin level, and BAL fluid neutrophil percentage did not significantly differ in the different subgroups of immunocompromised patients, leukocyte counts and BAL fluid neutrophil absolute counts were significantly affected by the underlying immunocompromised status. Patients who were admitted to the hospital because of respiratory symptoms and were undergoing bronchoscopy within 24 h presented with lower procalcitonin levels (0.08 ng/mL; IQR, 0.03 to 0.23 ng/mL) compared to those patients who had a longer duration of symptoms during the hospital stay preceding bronchoscopy (0.17 ng/mL; IQR, 0.06 to 0.61 ng/mL; p = 0.045). Microbiological data for BAL fluid are presented in Table 4 . Overall, 38 patients (35.5%) had a microbiological study result that was positive for bacteria, viruses, and/or fungi. The most common bacterial pathogens were Streptococcus pneumoniae, Pseudomonas aeruginosa, L pneumophila, and Enterococci.


View this table:
[in this window]
[in a new window]

  		Table 3.. Laboratory Parameters in the Different Immunocompromised Patient Subgroups*

 

View this table:
[in this window]
[in a new window]

  		Table 4.. Documentation of Bacteria, Viruses, and Fungi in the BAL Fluid of 107 Immunocompromised Patients

 
In addition to BAL fluid specimens, transbronchial biopsy specimens were obtained in 15 cases (14.0%) and endobronchial biopsy specimens in 3 cases (2.8%). There were no major complications during bronchoscopy. Histologic pulmonary specimens obtained by video-assisted thoracic surgery were available in seven cases (6.5%). The in-hospital mortality rate was 5.6% (six cases). Autopsy results were available in four of these six cases.

Distribution of Diagnostic Groups
Proven bacterial infection was diagnosed in 27 patients (25%). In four patients (15%), the diagnosis was based on a polymerase chain reaction finding that was positive for L pneumophila, and in three patients (11%) bacterial growth in the BAL fluid was > 103 cfu/mL but < 104 cfu/mL (P aeruginosa, S pneumoniae, and Enterocci spp in one patient each). In one patient (4%), a negative finding for Staphylococcus coagulase. was found both in a BAL fluid specimen (> 100,000 cfu/mL) and in blood culture. Gram stains on BAL fluid specimens failed to show evidence of bacteria in culture from 9 of 27 patients with proven bacterial infection.

Eleven patients (11%) were classified as having possible bacterial infection, and a nonbacterial condition was identified in 69 patients (65%). The distribution of cases classified as proven, possible, and no bacterial infection within the different subgroups of immunocompromised patients is shown in Table 5 . Among the 20 neutropenic patients, 4 (20%) had proven bacterial infection, 1 (5%) had possible bacterial infection, and 15 (75%) had no bacterial infection. Table 6 presents the final pulmonary diagnosis of the 69 patients considered to have a nonbacterial condition.


View this table:
[in this window]
[in a new window]

  		Table 5.. Distribution of Diagnostic Groups Among the Different Immunocompromised Patient Subgroups*

 

View this table:
[in this window]
[in a new window]

  		Table 6.. Final Pulmonary Diagnosis of the 69 Patients Considered To Have a Nonbacterial Condition Causing Pulmonary Signs and Symptoms*

 
The clinical parameters were similar in patients with and without bacterial infection (Table 7 ). Infiltrates were present significantly more often in patients with proven bacterial infection (89%) or possible bacterial infection (82%), compared to nonbacterial conditions (62%; p = 0.04).


View this table:
[in this window]
[in a new window]

  		Table 7.. Signs, Symptoms, and Radiologic Findings in 107 Immunocompromised Patients, According to Diagnostic Group Classification*

 
Table 8 presents leukocyte counts, CRP levels, procalcitonin levels, and BAL fluid neutrophil counts (absolute and percentage) in patients with proven bacterial infection, possible bacterial infection, and no bacterial infection. While there was a statistically significant difference among the three diagnostic groups, laboratory parameters and neutrophil counts in the BAL fluid were not significantly different in cases of proven and possible bacterial infection (difference not significant [determined by Mann-Whitney U test]).


View this table:
[in this window]
[in a new window]

  		Table 8.. Laboratory Parameters in the 107 Immunocompromised Patients, According to the Diagnostic Group Classification*

 
CRP and procalcitonin levels were similar in patients with proven or suspected (probable and possible) invasive pulmonary aspergillosis and in those patients in whom other nonbacterial conditions were identified (p = 0.221 and p = 0.162, respectively). There was no difference in CRP level, procalcitonin level, BAL fluid neutrophil percentage and absolute counts in patients with and without neutropenia and proven bacterial infection (difference not significant for all). Steroids and cyclosporin intake did not significantly affect BAL fluid neutrophil results. The median BAL fluid neutrophil count was 8% (IQR, 2 to 22.5%) in patients not receiving steroids compared to 10.5% (IQR, 4 to 27.5%) in patients receiving therapy with steroids (p = 0.414). Similarly, the median BAL neutrophil count was 10% (IQR, 3 to 37.5%) in patients not receiving therapy with cyclosporin compared to 8% (IQR, 2 to 22%) in patients receiving therapy with cyclosporin (p = 0.212). Similarly, median procalcitonin levels were not influenced by the following immunosuppressive drugs: steroids (n = 47; p = 0.673); cyclosporin (n = 33; p = 0.881); mycophenolate (n = 30 p = 0.603); azathioprine (n = 10; p = 0.996); sirolimus (n = 6; p = 0.096); and tacrolimus (n = 3.0; p = 0.166).

There was a moderate correlation between procalcitonin and CRP levels (Spearman rho = 0.595; p < 0.001) and a poor negative correlation between procalcitonin levels and leukocyte counts (Spearman rho = –0.213; p = 0.031). There was no correlation between procalcitonin levels and BAL neutrophil counts, both absolute and as a percentage (p = 0.204 and p = 0.171, respectively).

Diagnostic Accuracy of Clinical and Laboratory Parameters
To analyze the diagnostic accuracy of clinical and laboratory parameters to predict bacterial infection, a ROC analysis was performed (Fig 1 ). The percentage of neutrophils in BAL fluid had the highest AUC (0.818; 95% confidence interval [CI], 0.700 to 0.935; p < 0.001), followed by absolute neutrophil counts (AUC, 0.797; 95% CI, 0.678 to 0.916; p < 0.001), procalcitonin level (AUC, 0.746; 95% CI, 0.602 to 0.889; p = 0.001), and CRP level (AUC, 0.688; 95% CI, 0.555 to 0.821; p = 0.015). The AUC was not statistically significant for infiltrates (AUC, 0.619; 95% CI, 0.484 to 0.754; p = 0.123) and leukocyte counts (AUC, 0.561; 95% CI, 0.403 to 0.719; p = 0.429).


Figure 1
View larger version (21K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1.. ROC analysis for CRP level (AUC, 0.688; 95% CI, 0.555 to 0.821; p = 0.015), BAL fluid neutrophil counts by percentage (AUC, 0.818; 95% CI, 0.700 to 0.935; p < 0.001) and absolute numbers (AUC, 0.797; 95% CI, 0.678 to 0.916; p < 0.001), leukocyte count (AUC, 0.561, 95% CI, 0.403 to 0.719; p = 0.429), procalcitonin level (AUC, 0.746; 95% CI, 0.602 to 0.889; p = 0.001), and infiltrates (AUC, 0.619; 95% CI, 0.484 to 0.754; p = 0.123).

 
Compared to BAL fluid neutrophil percentage alone, the AUC for the combination of BAL fluid neutrophil percentage and either CRP level (AUC, 0.844; 95% CI, 0.737 to 0.951; p < 0.001) or procalcitonin level (AUC, 0.872; 95% CI, 0.778 to 0.965; p < 0.001) increased the AUC for the diagnosis of bacterial infection (Fig 2 ). However, this difference did not reach statistical significance for either situation (p = 0.078 and p = 0.118, respectively). The combination of BAL fluid neutrophil percentage with both CRP and procalcitonin levels did not further improve the AUC of the model.


Figure 2
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2.. ROC analysis for a model combining BAL fluid neutrophil counts in percentages combined with CRP level (AUC, 0.844; 95% CI, 0.737 to 0.9551; p < 0.001) or procalcitonin level (AUC, 0.872; 95% CI, 0.778 to 0.965; p < 0.001).

 
In a multivariate logistic regression model, the log-transformed percentage of neutrophils in BAL fluid (p = 0.001; odds ratio [OR], 2.508; 95% CI, 1.424 to 4.420) and the log-transformed procalcitonin level (p = 0.038; OR, 1.557; 95% CI, 1.025 to 2.364) proved to be independent factors predicting bacterial infection. We found no significance for log-transformed CRP level (p = 0.455; OR, 1.191; 95% CI, 0.752 to 1.887).

Sensitivity, specificity, and positive and negative likelihood ratio point estimates for CRP level, procalcitonin level, and BAL fluid neutrophil percentage are presented in Table 9 . At the threshold value of 20 mg/L, CRP level had a sensitivity of 84% and a specificity of 48%. Procalcitonin level had the highest specificity (84%) at the threshold value of 0.5 ng/mL. BAL fluid neutrophil percentage had the highest overall performance at the threshold of 15% (sensitivity, 84%; specificity, 77%).


View this table:
[in this window]
[in a new window]

  		Table 9.. Point Estimates for CRP, Procalcitonin, and BAL Fluid Neutrophil Counts To Diagnose Bacterial Infection*

 
Discussion

In this study, we found that clinical parameters were not useful for the differential diagnosis of pulmonary complications in immunocompromised patients. In contrast, neutrophilia in the BAL fluid as well as increased serum levels of CRP and procalcitonin were significantly associated with bacterial pulmonary infection. The percentage of neutrophils in BAL fluid showed the best diagnostic accuracy for predicting bacterial infection in the ROC curve analysis, followed by procalcitonin level and CRP level. While the combination of BAL fluid neutrophil percentage and the level of either biomarker increased the AUC of the model, this improvement failed to reach statistical significance when compared to the use of BAL fluid neutrophil percentage alone. This suggests that the evaluation of BAL fluid still represent the best diagnostic modality in immunocompromised patients with pulmonary complications. However, if there are contraindications to performing BAL or bronchoscopy, and its results are not available within a reasonable time frame, the determination of the CRP or procalcitonin level might represent a feasible approach for gathering evidence of bacterial infection. Furthermore, as procalcitonin level proved to be an independent predictor of bacterial infection in a multivariate regression analysis, the combination of BAL fluid neutrophil percentage and procalcitonin level, including a cost-benefit analysis, might warrant evaluation in future studies.

Pulmonary bacterial infections in immunocompromised patients are a challenge to physicians as both clinical signs and radiologic appearance are nonspecific.46 BAL has been established as a reliable technique for the diagnosis of pulmonary infection in the immunocompromised host, specifically for detecting opportunistic infectious such as P jiroveci and also bacteria.141547 Previous reports518484950 have suggested that BAL might be capable of identifying bacterial pathogens in immunocompromised patients despite antibiotic therapy due to the high bacterial load in lower respiratory tract secretions or to the higher incidence of infection with virulent organisms that respond slowly to antimicrobial therapy. However, empiric antimicrobial therapy has been shown to impair the diagnostic yield of bronchoscopy in HIV-infected patients and in those with nosocomial infections.5152 Alternatively, BAL fluid cytology has proved to be useful in the diagnosis of infectious pulmonary infiltrates in immunocompromised adults and children.2046 Thereby, significantly higher total neutrophil cell counts and percentages were found in the group of patients in whom bacterial organisms were detected in BAL fluid, compared with the group of patients in whom no bacterial agent was detected.46 In accordance with previous studies,535455 we found BAL fluid neutrophilia to be present in neutropenic patients, suggesting that even patients who are undergoing high-dose chemotherapy and stem cell transplantation with neutropenia can generate a local immune reaction in response to infectious agents. In the current study, neutrophilia in the BAL fluid at a threshold of 15% had a sensitivity of 84% and a specificity of 77% to predict bacterial infection. These data indicate that the differential cell count in BAL fluid may be of value in the differential diagnosis of pulmonary complications in immunocompromised hosts.

The use of inflammatory biomarkers for the early detection of bacterial infection might be a useful approach for distinguishing febrile episodes and for guiding the choice of specific antibiotic therapies, even before culture results are available.23 The present findings illustrate that CRP and procalcitonin serum levels are significantly higher in patients with proven and possible bacterial infection, compared to those patients with a nonbacterial condition. Thus, circulating biomarkers might give evidence of bacterial infection even in those cases in which microbiology results turned negative due to antimicrobial therapy. Interestingly, the combination of BAL fluid neutrophils with serum biomarkers tended to improve the diagnostic accuracy of the ROC model, indicating that both parameters might have additive value. This is particularly important in the setting of immunocompromised patients due to the high morbidity and mortality if the infection is treated inappropriately or recognized too late.

The fundamental goal of diagnostic bronchoscopy is to establish a specific diagnosis that will allow more targeted therapy and improve survival. While noninfectious pulmonary complications tend to be less common than infectious complications, fewer than half of potentially treatable, nonbacterial pulmonary infections have been reported to be diagnosed prior to death.5657 The poor outcome of pulmonary complications in immunocompromised patients is claimed to be associated with the delay in the administration of specific drugs in the evolution of the illness.5 Hence, it is tempting to speculate that a diagnostic approach that rules out bacterial infection (eg, the determination of serum biomarkers) could also lead to an earlier diagnostic workup targeting noninfectious etiologies and, thus, to a better outcome.

Several limitations of our study should also be noted. First, we have included a heterogeneous population of immunocompromised patients. While this diversity is routine for the pulmonary consultant in a tertiary care setting, we might have missed findings peculiar to a particular subpopulation of patients. Second, the majority of our study population was composed of patients with hematologic malignancies being treated with high-dose chemotherapy and/or allogeneic stem cell transplantation. This might explain the relatively high incidence of noninfectious pulmonary complications found in our cohort. It would be of additional clinical and scientific interest to evaluate the combination of BAL neutrophil counts with procalcitonin or CRP levels in larger subgroups of immunocompromised patients with a higher incidence of infectious complications, such as lung, kidney, heart, and liver transplant recipients. Third, we have included in the study only hospitalized patients who were treated in the medical ward. Because of the exclusion of critically ill patients requiring mechanical ventilation, mortality rates in our study were lower than those in previous reports. Fourth, we have evaluated clinical and laboratory parameters in regard to their diagnostic value on the day of bronchoscopy (not at the initiation of the symptoms), and the kinetics of both biomarkers have not been analyzed. Finally, we cannot exclude the possibility that previous antibiotic therapy might have decreased the microbiological yield of BAL fluid specimens. However, 96% of patients (26 of 27 patients) with proven bacterial infection were receiving therapy with antibiotics at the time of bronchoscopy, and most patients were started on therapy with antibiotics within 24 h of undergoing bronchoscopy. Thus, due to the high bacterial load and slower antibiotic response observed in immunocompromised patients,518484950 we believe that previous antibiotic therapy did not considerably influence the microbiological yield of BAL fluid specimens in our study. In summary, our findings indicate that BAL fluid neutrophil percentage alone or in combination with serum levels of procalcitonin or CRP at the time of bronchoscopy may potentially be used to guide diagnostic and therapeutic decisions in immunocompromised hosts with pulmonary complications.

Footnotes

Abbreviations: AUC = area under the curve; CI = confidence interval; CRP = C-reactive protein; IQR = interquartile range; OR = odds ratio; ROC = receiver operating characteristic

The study was presented in part at the 16th Annual Congress of European Respiratory Society, Munich, 2006 and was honored with the Research Excellence Award in Respiratory Infections.

Drs. Stolz and Müller have received payments from BRAHMS (the manufacturer of procalcitonin assays) to attend advisory board meetings and speaker engagements, or for research. Drs. Stulz, Gratwohl, and Tamm have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Received for publication January 19, 2007. Accepted for publication April 18, 2007.

References

  1. Kotloff, RM, Ahya, VN, Crawford, SW (2004) Pulmonary complications of solid organ and hematopoietic stem cell transplantation. Am J Respir Crit Care Med 170,22-48[Abstract/Free Full Text]
  2. Afessa, B, Peters, SG Major complications following hematopoietic stem cell transplantation. Semin Respir Crit Care Med 2006;27,297-309[CrossRef][ISI][Medline]
  3. Waite, S, Jeudy, J, White, CS Acute lung infections in normal and immunocompromised hosts. Radiol Clin North Am 2006;44,295-315ix[CrossRef][ISI][Medline]
  4. Fishman, JA, Rubin, RH Infection in organ-transplant recipients. N Engl J Med 1998;338,1741-1751[Free Full Text]
  5. Rano, A, Agusti, C, Jimenez, P, et al Pulmonary infiltrates in non-HIV immunocompromised patients: a diagnostic approach using non-invasive and bronchoscopic procedures. Thorax 2001;56,379-387[Abstract/Free Full Text]
  6. Shorr, AF, Susla, GM, O’Grady, NP Pulmonary infiltrates in the non-HIV-infected immunocompromised patient: etiologies, diagnostic strategies, and outcomes. Chest 2004;125,260-271[Abstract/Free Full Text]
  7. Jain, P, Sandur, S, Meli, Y, et al Role of flexible bronchoscopy in immunocompromised patients with lung infiltrates. Chest 2004;125,712-722[Abstract/Free Full Text]
  8. Cordonnier, C, Bernaudin, JF, Bierling, P, et al Pulmonary complications occurring after allogeneic bone marrow transplantation: a study of 130 consecutive transplanted patients. Cancer 1986;58,1047-1054[CrossRef][ISI][Medline]
  9. Jules-Elysee, K, Stover, DE, Yahalom, J, et al Pulmonary complications in lymphoma patients treated with high-dose therapy autologous bone marrow transplantation. Am Rev Respir Dis 1992;146,485-491[ISI][Medline]
  10. Patel, NR, Lee, PS, Kim, JH, et al The influence of diagnostic bronchoscopy on clinical outcomes comparing adult autologous and allogeneic bone marrow transplant patients. Chest 2005;127,1388-1396[Abstract/Free Full Text]
  11. Chandrasekar, PH, Weinmann, A, Shearer, C Autopsy-identified infections among bone marrow transplant recipients: a clinico-pathologic study of 56 patients: Bone Marrow Transplantation Team. Bone Marrow Transplant 1995;16,675-681[ISI][Medline]
  12. Baughman, RP The lung in the immunocompromised patient: infectious complications part 1. Respiration 1999;66,95-109[CrossRef][ISI][Medline]
  13. Tamm, M The lung in the immunocompromised patient: infectious complications part 2. Respiration 1999;66,199-207[CrossRef][ISI][Medline]
  14. Joos, L, Chhajed, PN, Wallner, J, et al Pulmonary infections diagnosed by BAL: a 12-year experience in 1066 immunocompromised patients. Respir Med 2006;
  15. Stover, DE, Zaman, MB, Hajdu, SI, et al Bronchoalveolar lavage in the diagnosis of diffuse pulmonary infiltrates in the immunosuppressed host. Ann Intern Med 1984;101,1-7[ISI][Medline]
  16. Bodey, GP, Rolston, KV Management of fever in neutropenic patients. J Infect Chemother 2001;7,1-9[CrossRef][Medline]
  17. Pizzo, PA Fever in immunocompromised patients. N Engl J Med 1999;341,893-900[Free Full Text]
  18. Dunagan, DP, Baker, AM, Hurd, DD, et al Bronchoscopic evaluation of pulmonary infiltrates following bone marrow transplantation. Chest 1997;111,135-141[Abstract/Free Full Text]
  19. White, P, Bonacum, JT, Miller, CB Utility of fiberoptic bronchoscopy in bone marrow transplant patients. Bone Marrow Transplant 1997;20,681-687[CrossRef][ISI][Medline]
  20. Sternberg, RI, Baughman, RP, Dohn, MN, et al Utility of bronchoalveolar lavage in assessing pneumonia in immunosuppressed renal transplant recipients. Am J Med 1993;95,358-364[CrossRef][ISI][Medline]
  21. Ortega, M, Rovira, M, Almela, M, et al Measurement of C-reactive protein in adults with febrile neutropenia after hematopoietic cell transplantation. Bone Marrow Transplant 2004;33,741-744[CrossRef][ISI][Medline]
  22. Ortega, M, Rovira, M, Filella, X, et al Prospective evaluation of procalcitonin in adults with febrile neutropenia after haematopoietic stem cell transplantation. Br J Haematol 2004;126,372-376[CrossRef][ISI][Medline]
  23. Persson, L, Engervall, P, Magnuson, A, et al Use of inflammatory markers for early detection of bacteraemia in patients with febrile neutropenia. Scand J Infect Dis 2004;36,365-371[CrossRef][ISI][Medline]
  24. Bernard, L, Ferriere, F, Casassus, P, et al Procalcitonin as an early marker of bacterial infection in severely neutropenic febrile adults. Clin Infect Dis 1998;27,914-915[ISI][Medline]
  25. Engel, A, Mack, E, Kern, P, et al An analysis of interleukin-8, interleukin-6 and C-reactive protein serum concentrations to predict fever, Gram-negative bacteremia and complicated infection in neutropenic cancer patients. Infection 1998;26,213-221[ISI][Medline]
  26. Svaldi, M, Hirber, J, Lanthaler, AI, et al Procalcitonin-reduced sensitivity and specificity in heavily leucopenic and immunosuppressed patients. Br J Haematol 2001;115,53-57[CrossRef][ISI][Medline]
  27. Stolz, D, Christ-Crain, M, Bingisser, R, et al Antibiotic treatment of exacerbations of COPD: a randomized, controlled trial comparing procalcitonin-guidance with standard therapy. Chest 2007;131,9-19[Abstract/Free Full Text]
  28. Christ-Crain, M, Stolz, D, Bingisser, R, et al Procalcitonin guidance of antibiotic therapy in community-acquired pneumonia: a randomized trial. Am J Respir Crit Care Med 2006;174,84-93[Abstract/Free Full Text]
  29. Christ-Crain, M, Jaccard-Stolz, D, Bingisser, R, et al Effect of procalcitonin-guided treatment on antibiotic use and outcome in lower respiratory tract infections: cluster-randomised, single-blinded intervention trial. Lancet 2004;363,600-607[CrossRef][ISI][Medline]
  30. von Lilienfeld-Toal, M, Schneider, A, Orlopp, K, et al Change of procalcitonin predicts clinical outcome of febrile episodes in patients with hematological malignancies. Support Care Cancer 2006;14,1241-1245[CrossRef][ISI][Medline]
  31. Stolz, D, Chhajed, PN, Leuppi, J, et al Nebulized lidocaine for flexible bronchoscopy: a randomized, double-blind, placebo-controlled trial. Chest 2005;128,1756-1760[Abstract/Free Full Text]
  32. Stolz, D, Chhajed, PN, Leuppi, JD, et al Cough suppression during flexible bronchoscopy using combined sedation with midazolam and hydrocodone: a randomised, double blind, placebo controlled trial. Thorax 2004;59,773-776[Abstract/Free Full Text]
  33. British Thoracic Society Guidelines Committee.. British Thoracic Society guidelines on diagnostic flexible bronchoscopy. Thorax 2001;56(suppl),i1-21[Free Full Text]
  34. European Society of Pneumology.. Clinical guidelines and indications for bronchoalveolar lavage (BAL): report of the European Society of Pneumology Task Group on BAL. Eur Respir J 1990;3,937-976[Medline]
  35. Tamm, M, Traenkle, P, Grilli, B, et al Pulmonary cytomegalovirus infection in immunocompromised patients. Chest 2001;119,838-843[Abstract/Free Full Text]
  36. Herth, FJ, Becker, HD, Ernst, A Aspirin does not increase bleeding complications after transbronchial biopsy. Chest 2002;122,1461-1464[Abstract/Free Full Text]
  37. American Thoracic Society, European Respiratory Society.. American Thoracic Society/European Respiratory Society International Multidisciplinary Consensus Classification of the Idiopathic Interstitial Pneumonias: this joint statement of the American Thoracic Society (ATS), and the European Respiratory Society (ERS) was adopted by the ATS board of directors, June 2001 and by the ERS Executive Committee, June 2001. Am J Respir Crit Care Med 2002;165,277-304[Free Full Text]
  38. Estenne, M, Maurer, JR, Boehler, A, et al Bronchiolitis obliterans syndrome 2001: an update of the diagnostic criteria. J Heart Lung Transplant 2002;21,297-310[CrossRef][ISI][Medline]
  39. Afessa, B, Litzow, MR, Tefferi, A Bronchiolitis obliterans and other late onset non-infectious pulmonary complications in hematopoietic stem cell transplantation. Bone Marrow Transplant 2001;28,425-434[CrossRef][ISI][Medline]
  40. Afessa, B, Tefferi, A, Litzow, MR, et al Diffuse alveolar hemorrhage in hematopoietic stem cell transplant recipients. Am J Respir Crit Care Med 2002;166,641-645[Free Full Text]
  41. Katzenstein, AL, Bloor, CM, Leibow, AA Diffuse alveolar damage: the role of oxygen, shock, and related factors; a review. Am J Pathol 1976;85,209-228[Medline]
  42. Ascioglu, S, Rex, JH, de Pauw, B, et al Defining opportunistic invasive fungal infections in immunocompromised patients with cancer and hematopoietic stem cell transplants: an international consensus. Clin Infect Dis 2002;34,7-14[CrossRef][ISI][Medline]
  43. Linscheid, P, Seboek, D, Schaer, DJ, et al Expression and secretion of procalcitonin and calcitonin gene-related peptide by adherent monocytes and by macrophage-activated adipocytes. Crit Care Med 2004;32,1715-1721[CrossRef][ISI][Medline]
  44. Nylen, ES, Muller, B, Becker, KL, et al The future diagnostic role of procalcitonin levels: the need for improved sensitivity. Clin Infect Dis 2003;36,823-824[CrossRef][ISI][Medline]
  45. Metz, CE Basic principles of ROC analysis. Semin Nucl Med 1978;8,283-298[ISI][Medline]
  46. Ratjen, F, Costabel, U, Havers, W Differential cytology of bronchoalveolar lavage fluid in immunosuppressed children with pulmonary infiltrates. Arch Dis Child 1996;74,507-511[Abstract]
  47. Xaubet, A, Torres, A, Marco, F, et al Pulmonary infiltrates in immunocompromised patients: diagnostic value of telescoping plugged catheter and bronchoalveolar lavage. Chest 1989;95,130-135[Abstract/Free Full Text]
  48. Kahn, FW, Jones, JM Diagnosing bacterial respiratory infection by bronchoalveolar lavage. J Infect Dis 1987;155,862-869[ISI][Medline]
  49. von Eiff, M, Zuhlsdorf, M, Roos, N, et al Pulmonary infiltrates in patients with haematologic malignancies: clinical usefulness of non-invasive bronchoscopic procedures. Eur J Haematol 1995;54,157-162[ISI][Medline]
  50. Eriksson, BM, Dahl, H, Wang, FZ, et al Diagnosis of pulmonary infections in immunocompromised patients by fiber-optic bronchoscopy with bronchoalveolar lavage and serology. Scand J Infect Dis 1996;28,479-485[ISI][Medline]
  51. Gracia, JD, Miravitlles, M, Mayordomo, C, et al Empiric treatments impair the diagnostic yield of BAL in HIV-positive patients. Chest 1997;111,1180-1186[Abstract/Free Full Text]
  52. Souweine, B, Veber, B, Bedos, JP, et al Diagnostic accuracy of protected specimen brush and bronchoalveolar lavage in nosocomial pneumonia: impact of previous antimicrobial treatments. Crit Care Med 1998;26,236-244[CrossRef][ISI][Medline]
  53. Cordonnier, C, Escudier, E, Verra, F, et al Bronchoalveolar lavage during neutropenic episodes: diagnostic yield and cellular pattern. Eur Respir J 1994;7,114-120[Abstract]
  54. Milburn, HJ, Du Bois, RM, Prentice, HG, et al Pneumonitis in bone marrow transplant recipients results from a local immune response. Clin Exp Immunol 1990;81,232-237[ISI][Medline]
  55. Leskinen, R, Taskinen, E, Volin, L, et al Use of bronchoalveolar lavage cytology and determination of protein contents in pulmonary complications of bone marrow transplant recipients. Bone Marrow Transplant 1990;5,241-245[ISI][Medline]
  56. Yokoi, T, Hirabayashi, N, Ito, M, et al Broncho-bronchiolitis obliterans as a complication of bone marrow transplantation: a clinicopathological study of eight autopsy cases: Nagoya BMT Group. Virchows Arch 1997;431,275-282[CrossRef][ISI][Medline]
  57. Sharma, S, Nadrous, HF, Peters, SG, et al Pulmonary complications in adult blood and marrow transplant recipients: autopsy findings. Chest 2005;128,1385-1392[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow All Versions of this Article:
chest.07-0175v1
132/2/504    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stolz, D.
Right arrow Articles by Tamm, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stolz, D.
Right arrow Articles by Tamm, M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS