Role of Bronchoalveolar Lavage in Immunocompromised Patients With

Thorax 2001;56:115–120 115

Role of bronchoalveolar lavage in Thorax: first published as 10.1136/thorax.56.2.115 on 1 February 2001. Downloaded from immunocompromised patients with pneumonia treated with a broad spectrum antibiotic and antifungal regimen

I A Hohenadel, M Kiworr, R Genitsariotis, D Zeidler, J Lorenz

Abstract Keywords: bronchoalveolar lavage; pneumonia; Background—In a retrospective study the immunosuppression value of bronchoalveolar lavage (BAL) in the diagnosis of pneumonia was investi- Pneumoniaisamajorcauseofdeathinimmuno- gated in 95 immunocompromised patients compromised patients with haematological suVering from haematological disorders disorders; the mortality associated with un- and receiving a regimen of broad spectrum treated infection in antibiotics and antifungal agents (BSAR). Legionella pneumophila immunocompromised patients has been re- Methods—With the exception of four ported to be 80%.1 Identification of the causa- afebrile patients, all had fever, raised C tive pathogen and initiation of eVective treat- reactive protein (CRP) levels, and new ment is therefore essential. We have infiltrates visible on chest radiography. All investigated the diagnostic value of broncho- patients underwent BAL to identify the alveolar lavage (BAL) in 95 immunocompro- organism causing the pneumonia and sur- mised patients with haematological disorders veillance cultures were performed regu- and clinical signs of acute pneumonia, compar- larly for pathogens at diVerent sites. ing BAL with non-invasive procedures such as Following classification of the isolates, surveillance cultures at diVerent sites. We patients with positive cultures were sub- determined whether the BAL findings are divided into two groups, pathogenic or associated with the therapeutic outcome for contaminated. We investigated whether immunocompromised patients routinely re- relevant pathogens were cultured only ceiving a broad spectrum antibiotic and from the BAL fluid and whether they were antifungal regimen (BSAR). All BAL related susceptible to BSAR. complications were documented, including Results—Although 77 of the 95 patients Department of those observed for patients with thrombocyto- http://thorax.bmj.com/ were thrombocytopenic, bleeding during Pulmonary Medicine, penia. This study assesses the eVect of the BAL occurred in only 15% of all patients. Academic Teaching severity of immunodeficiency on a spectrum of Ten days after the procedure the fever Hospital pathogens as well as on the clinical course of Köln-Merheim, improved in 88% of patients, radiographic pneumonia. D-51109 Köln, findings improved in 71%, and CRP levels Germany improved in 75% of patients; 22% of I A Hohenadel D Zeidler patients died within 28 days. Pathologi- Methods cally relevant isolates were found in 65% of The retrospective study took place in the Department of Pulmonary Medicine in the

Department of all patients. Respiratory pathogens were on September 26, 2021 by guest. Protected copyright. Pulmonary Medicine, detected only in the BAL fluid of 29 of the University Hospital in Mainz, Germany over a University Hospital, 95 patients (35% Gram positive species, period of 4 years and 2 months. All patients Mainz, Germany 40% Gram negative species, 11% Myco- with haematological disorders and neutropenia M Kiworr bacterium, 11% fungi, and 3% cytomega- were enrolled in the study if they presented Department of lovirus). In 16 of these 29 patients (55%) with signs of acute pneumonia. The median Microbiology, the pathogens cultured only from the BAL (SD) age of the 95 immunocompromised University Hospital, fluid were resistant to treatment. Patho- patients (68 men, 27 women) was 49.4 (14.5) Mainz, Germany gens detected only in the BAL fluid were years. Underlying diseases included acute R Genitsariotis not susceptible to a standard broad spec- myeloid leukaemia (45), acute lymphatic leukaemia (10), chronic myeloid leukaemia (10), Department of trum antibiotic and antifungal regimen Pulmonary Medicine, including teicoplanin, ceftriaxon, tobra- non-Hodgkin’s lymphoma (10), myelodysplas- Academic Teaching mycin, and amphotericin B in 12 of the 29 tic syndrome (7), Hodgkin’s disease (5), Hospital, Lüdenscheid, patients (41%). chronic lymphatic leukaemia (3), aplastic Germany Conclusions—Our data suggest that 12 anaemia or agranulocytosis (5). Patients were J Lorenz patients were treated with broad spectrum divided into two groups according to the neutrophil granulocyte count: group A included 71 Correspondence to: antimicrobial agents which were not di- Dr I A Hohenadel rected at the appropriate organism on in patients with a neutrophil granulocyte count of [email protected] vitro sensitivity tests without BAL. BAL is less than 1000/µl and group B included 24 patients with a neutrophil granulocyte count of Received 24 February 2000 a relatively safe procedure in the diagnosis Returned to authors of pneumonia, supplying important infor- 1000–3500/µl. With the exception of four 17 April 2000 mation in immunocompromised patients afebrile patients, all patients presented with Revised version received respiratory symptoms (such as dyspnoea and 6 October 2000 as well as in immunocompromised pa- Accepted for publication tients receiving BSAR. cough), fever, raised C reactive protein (CRP) 27 October 2000 (Thorax 2001;56:115–120) levels, and new infiltrates as shown by chest

www.thoraxjnl.com 116 Hohenadel, Kiworr, Genitsariotis, et al

radiography. BSAR was administered to all Results Thorax: first published as 10.1136/thorax.56.2.115 on 1 February 2001. Downloaded from those except the patients with renal dysfunc- PATIENT CHARACTERISTICS tion for whom the antibiotic and antifungal Seventy one immunocompromised patients prescription was changed. were allocated to group A (neutrophil granulo- All patients underwent BAL in our depart- cyte count less than 1000/µl) and 24 to group B ment. After inspection of the transbronchial (neutrophil granulocyte count 1000–3500/µl) tree to the subsegmental level, the broncho- according to the severity of the underlying scope was wedged into a subsegment bronchus condition. The mean value of the lowest in the area with the most marked radiological granulocyte count was 590/µl (range 0–3400/ abnormality or, in the presence of diVuse infil- µl). Seventy six patients had undergone chemo- tration, into the lingula or middle lobe.2 BAL therapy during the six months preceding BAL was performed according to a standard tech- (mean 23 days). Nine patients had received a nique3 using 5–7 20 ml aliquots of normal bone marrow transplant. Admission to hospital saline. Suctioning after each bolus yielded a was on average 18.7 days before BAL (range total of 40–100 ml of fluid. Aliquots of lavage 0–70 days); 71 patients were admitted for a fluid were submitted for microscopic examina- minimum of five days before the procedure. A tion (Gram, acridin and acid-fast stains) and total of 71 patients were therefore at risk of cultured for aerobic and anaerobic bacteria, “hospital acquired pneumonia” with more fungi, mycobacteria, and viruses quantitatively. virulent pathogens. Onset of fever was noted on Standard culture methods were used. Direct average 9.4 days and pulmonary infiltrates immunofluorescence was employed for the were detected 7.7 days preceding BAL. There detection of Legionella pneumophila as well as were no significant diVerences between groups cultures and polymerase chain reaction in the A and B with respect to age, sex, mean duration last year of the study. For mycobacteria Ziehl of BAL, preceding chemotherapy, or the ratio Nielsen acid-fast stains and cultures were per- of patients with chemotherapy and clinical formed. Detection of cytomegalovirus was per- parameters. formed by immunofluorescence techniques Intravenous antibiotics and/or antifungal and human cell cultures. For all pathogenic agents were administered to 82 of the 95 stains isolated antibiograms were primarily patients (86%) 9.6 days before BAL. In 80 of determined. Transbronchial biopsy specimens the 95 patients (84%) the antibiotic and/or could only be taken from 11 patients; all the antifungal regimen was changed during the 10 remaining patients had thrombocytopenia days after the procedure as a result of BAL and/or coagulopathy. The biopsy specimens findings and/or clinical observations. In 73 of were taken from the area of interest or from the the 95 patients antifungal agents were given lower lobes in the presence of a diVuse infiltra- intravenously. Most patients received the de- tion. Only two patients underwent open lung scribed standard BSAR regimen. There was no biopsies using an invasive procedure. Radio- significant diVerence in the percentage of graphs were routinely taken during the 24 patients treated with BSAR and the duration of http://thorax.bmj.com/ hours after the transbronchial biopsy speci- antibiotic therapy between groups A and B. mens to rule out pneumothorax. Fever, labora- Group A included more patients with underly- tory parameters, radiographs, medication in- ing acute myeloid leukaemia (p=0.004), while cluding preceding chemotherapy, and duration the number of patients with non-Hodgkin’s of neutropenia were recorded from admission lymphoma (p=0.015) and Hodgkin’s disease to hospital to discharge. Cultures from BAL (p=0.014) was greater in group B. Surprisingly, fluid, sputum, blood, urine, pharyngeal smears, there was no diVerence in the time from the and smears from other sites were performed onset of neutropenia to BAL between the two regularly for all patients. Patients with positive groups. on September 26, 2021 by guest. Protected copyright. cultures were subdivided following classification of the isolates into pathogenic or COMPLICATIONS OF BAL contaminated groups during the 14 days before No episodes of collapse, allergic reaction to and after BAL. While colonies of Legionella anaesthesia, or pneumothorax occurred during were considered to be pathogenic, coagulase or after bronchoscopy. Seventy seven of the 95 negative Staphylococcus or Candida were con- patients (81%) were thrombocytopenic (me- sidered to be pathogenic only when cultured dian platelet count 38 000/µl). The ratio of several times from diVerent sources in the patients with thrombocytopenia was signifi- presence of a consistent clinical course. cantly higher in group A (p=0.002). The plate- To determine the diagnostic value of BAL we let count was below 50 000/µl in 58 patients investigated whether the respiratory pathogens and below 20 000/µl in 16. No patient had were sensitive to the treatment as well as to a major bleeding complications. Minor limited BSAR consisting of teicoplanin, ceftriaxon, bleeding was observed in 15% of the patients. tobramycin, and amphotericin B. This stand- Twenty ml norepinephrine (noradrenaline) ard BSAR regimen was selected on the basis of were administered locally to only one patient our clinical experience and according to with a normal platelet count to stop non- published recommendations.4 haemodynamic bleeding after transbronchial biopsies. One of the 11 patients who under- STATISTICAL ANALYSIS went a transbronchial biopsy developed mini- The Wilcoxon test for two samples, the ÷2 test, mal bleeding after the procedure. Additional and Fisher’s exact test (two sided) were used. anaesthetics or sedatives were administered to Values are expressed as mean (SD). A value of seven patients with severe coughing. Broncho- p<0.05 was considered significant.5 scopic examination could not be carried out in

www.thoraxjnl.com Role of BAL in immunocompromised patients with pneumonia 117

Table 1 Characteristics of patients who died during the 28 days following bronchoalveolar lavage (BAL) Thorax: first published as 10.1136/thorax.56.2.115 on 1 February 2001. Downloaded from

Death (days Patient no Group Isolates from BAL ﬂuid Isolates from other sources after BAL) Additional comments

1BEnterococcus faecalis Enterococcus (ph) 4 5 A Coagulase negative Coagulase negative Staphylococcus 15 Staphylococcus (ph, b) 10 A Multiresistant Staphylococcus; Multiresistant Staphylococcus (b); 5 Enterococcus faecalis Staphylococcus aureus (b); Aspergillus (sp, u); Candida (sp, rare) 12 A Legionella pneumophila Negative 16 14 B Acinetobacter; E hermani Candida (sp, ph) 18 24 B Candida Candida (ph, sp) 21 Patient refused antifugal therapy after 3 days of treatment 27 A Negative Negative 6 28 A Legionella pneumophila Acinetobacter (sp); Staphylococcus 24 aureus (sp, ph) 33 B M tuberculosis Negative 3 34 A Negative Aspergillus (sp); Candida (sp, ph, f) 3 39 A Stenotrophomonas maltophilia; Staphylococcus (ph) 5 Capnocytophaga; Staphylococcus 44 B Aspergillus fumigatus; Candida Aspergillus fumigatus (sp); Candida 20 (sp) 47 A Aspergillus fumigatus; Candida; Candida (sp, f) 15 Necropsy: fungal Streptococcus; Lactobacillus pneumonia 48 A Multiresistant Staphylococcus Negative 26 60 B Negative Negative 2 64 A Enterococcus faecalis Negative 12 70 B Candida Candida (ph); Klebsiella 1 pneumoniae (ph) 71 A Negative Negative 20 ARDS, pneumothorax related to mechanical ventilation 78 B Candida Candida (ph, f) Staph aureus (ph) 19 90 B Cytomegalovirus; Candida; Candida (u,f,sp,ph); 7 Necropsy: Staphylococcus; Lactobacillus Staphylococcus (sp, ph) cytomegalovirus pneumonia

ph = pharyngeal smears; sp = sputum; b = blood; f = faeces; u = urine.

one patient with persistent coughing. There patients and multiple granulomas with Langer- were no diVerences in coughing or bleeding hans’ cells suspected of Mycobacterium tubercu- episodes between the two groups and no corre- losis infection were seen in one. In 11 of the 95 lation was established between the platelet patients transbronchial biopsy specimens were

count and bleeding. taken and malignant lymphoma cells were http://thorax.bmj.com/ found in three patients (underlying disease CLINICAL COURSE non-Hodgkin’s lymphoma in all three), peri- Ten days after BAL 88% of the patients were bronchial fibrosis was noted in one patient, and afebrile and the radiographic findings and CRP signs of a graft-versus-host reaction were seen levels were improved or unremarkable in 71% in another patient. One transbronchial biopsy and 75%, respectively. A positive outcome specimen confirmed the BAL finding of Myco- (decline of fever together with improved radio- bacterium tuberculosis. In three patients in whom graphy) occurred in 55% of patients, a negative no pathogens were found in the BAL fluid, outcome (either no decline or increase in tem- blood, sputum, or other sites, signs of malig- on September 26, 2021 by guest. Protected copyright. perature or unchanged radiography) occurred nancy were observed in the transbronchial in 23% of patients, and 22% of patients died biopsy specimens of two and on cytological during the 28 days following BAL (table 1). examination of the BAL fluid in one; malig- There were no significant diVerences in these nancy was therefore the most likely cause of the parameters between groups A and B, although pulmonary infiltrates and fever in these three the number of patients who died during the 28 patients. Bronchoscopy therefore provided days after BAL was significantly higher in important information in the diagnosis of pul- group B (p=0.02). No significant correlation monary infiltrates in 65 of the 95 patients was found between the outcome and either the (68.4%). type of pathogen or the age of the patient. Two of the three patients in whom Legionella COMPARISON OF BAL FINDINGS WITH NECROPSIES pneumophila was cultured only in the BAL fluid AND OPEN LUNG BIOPSY SPECIMENS died within 28 days of the procedure. Necroscopic examination confirmed the BAL findings in two patients with a diagnosis of CYTOLOGICAL EXAMINATION OF BAL FLUID AND Candida pneumonia, in one patient with TRANSBRONCHIAL BIOPSY SPECIMENS squamous cell carcinoma and concomitant An additional cytological examination of the bronchopneumonia, and in one patient with BAL fluid was performed at the local depart- severe interstitial cytomegalovirus pneumonia. ment of pathology for 66 of the 95 patients. Open lung biopsy demonstrated aspergillosis The results revealed signs of inflammation in with a cavitation filled with hyphae of Aspergil- 30 of the 66 patients and malignancy was seen lus fumigatus (aspergilloma) and accompanying in three; hyphae consistent with necroscopic bronchopneumonia in two patients in whom findings confirmed Candida pneumonia in two BAL had detected no evidence of Aspergillus.In

www.thoraxjnl.com 118 Hohenadel, Kiworr, Genitsariotis, et al

view of the fact that open lung biopsy is an biological findings showed that pathogens cul- Thorax: first published as 10.1136/thorax.56.2.115 on 1 February 2001. Downloaded from invasive procedure, it is rarely performed in tured only in the BAL fluid were resistant to patients with leukopenia because of concomi- the treatment given in 16 of the 29 patients tant thrombocytopenia and the risk of bleed- (55%). Thus, without BAL, 44% of the ing. A statistical analysis could therefore not be patients in group A and all of the patients in performed. A strong correlation was estab- group B would have been treated with antimi- lished between the BAL fluid and necroscopic crobial agents which were not directed at the findings, although pathogens such as Aspergil- appropriate organism (p=0.02). If all 29 lus are not detected by BAL if they are localised patients had received a BSAR consisting of in cavitations. teicoplanin, ceftriaxon, tobramycin, and amphotericin B, the regimen would have been MICROBIOLOGICAL FINDINGS ineVective in 12 of the 29 patients (41%). With Surveillance cultures at diVerent sites were BSAR, 13% of all patients would have received regularly performed for 95 immunocompro- insuYcient treatment. mised patients with signs of acute pneumonia. There were no diVerences between the two Positive cultures were found in BAL fluid groups with regard to positive cultures or (91%), blood (13%), sputum (93%), urine pathogens with the exception of Candida (12%), pharyngeal smears (88%), and other detected in BAL fluid. The significantly lower sources including smears obtained from central incidence of Candida in group A than in group vein catheters, endotracheal tubes, and pleura B (p=0.005) may be because antifungal fluid (47%). Pathogenically relevant isolates therapy was administered to more patients in from the respective sources were detected in this group. Antifungal agents were given to BAL fluid (65%), blood (11%), sputum 61% of patients in group A and to only 25% of (70%), urine (6%), pharyngeal smears (59%), patients in group B. No significant correlation and smears from other sources (30%). The was observed between the individual pathogen results of cultured pathogenic colonies from and the outcome or the severity of immuno- BAL fluid are presented in table 2. suppression or duration of neutropenia with In 62 of the 95 patients (65%) a pathogen the exception of mycobacteria which were likely to be responsible for pneumonia was cul- associated with a longer duration of neutro- tured from the BAL fluid. In 36 of these 62 penia (p=0.03). We could not establish a rela- patients (58%) microbiological findings tionship between the location of pathogens and showed the pathogens from the BAL fluid to be the susceptibility to antibiotics or antifungal resistant to the treatment given. The ratio of agents and the outcome of treatment. patients receiving inadequate treatment was significantly higher (p=0.006) in group B than Discussion in group A. If all 62 patients had undergone We have compared the diagnostic value of BAL treatment with a BSAR consisting of teico- as an invasive procedure in 95 immunocom- planin, ceftriaxon, tobramycin, and amphoter- promised patients with haematological disor- http://thorax.bmj.com/ icin B, the treatment would have been ineVec- ders and signs of pneumonia with non-invasive tive in 20 of the 62 patients (32%) in whom procedures such as surveillance cultures at dif- pathogens were detected in the BAL fluid. The ferent sites. BAL and sputum had the highest following pathogens were resistant to BSAR: diagnostic yields. However, in assessing this Mycobacterium tuberculosis or M xenopii, result consideration has to be given to the fact Capnocytophaga, multiresistant Staphylococcus, that, in our study, only one culture was taken Staphylococcus aureus, Stenotrophomonas, from the BAL fluid compared with multiple , , , and sputum samples (up to 12) and samples from

Acinetobacter Legionella Chlamydia on September 26, 2021 by guest. Protected copyright. cytomegalovirus. diVerent sources. It is also of interest to deter- In 29 of the 95 patients (31%) the organism mine whether pathogens observed in BAL fluid responsible for pneumonia was detected only were also present in sputum. Our findings show in the BAL fluid (35% Gram positive species, that organisms responsible for pneumonia were 40% Gram negative species, 11% Mycobacte- detected only in the BAL fluid in 29 of the 95 rium, 11% fungi, 3% cytomegalovirus). Micro- patients (31%) while all other samples, including sputum, had no diagnostic yield. The Table 2 Pathogens cultured from bronchoalveolar lavage causative organisms were therefore not de- (BAL) fluid in 95 immunocompromised patients with pneumonia tected by less invasive procedures in 29 patients. Sputum was found to be useful in Frequency Classification Frequency diagnosis in only 14% of the immunocompro- Pathogens (%) of pathogens (%) mised patients with pneumonia compared with Staphylococcus 23 (25.3%) Gram positive 40.7% 30% bronchial washings, 38% brushings, and Enterococcus 9 (9.9%) 73% transbronchial biopsy specimens reported Streptococcus 3 (3.3%) 6 S aureus 2 (2.2%) by Chopra. Transbronchial biopsy specimens Acinetobacter 1 (1.1%) Gram negative 16.5% were of diagnostic value in seven of 11 patients Capnocytophaga 4 (4.4%) (64%) and relevant pathogens were cultured Chlamydia 2 (2.2%) Enterobacteriacea 2 (2.2%) from BAL fluid in 65% of all patients in our Haemophilus 2 (2.2%) study. Legionella 3 (3.3%) The value of BAL in the diagnosis of Pseudomonas 1 (1.1%) Mycobacterium 6 (6.6%) Mycobacteria 6.6% pneumonia in immunocompromised patients Candida 25 (27.5%) Fungi 35.2% with diVerent underlying diseases cited in the Aspergillus 7 (7.7%) literature ranges from 33%7 to 83%.8 A Cytomegalovirus 1 (1.1%) Viruses 1.1% diagnostic yield of 81%9 has been reported in

www.thoraxjnl.com Role of BAL in immunocompromised patients with pneumonia 119

patients with and without immunosuppression, Aspergillus pneumonia. In an investigation Thorax: first published as 10.1136/thorax.56.2.115 on 1 February 2001. Downloaded from with a yield ranging from 60%10 to 66%11 in performed by Stover et al8 BAL had a diagnos- patients with haematological disorders exclu- tic yield in 83% of all patients with invasive sively and of only 50%12 in patients with a bone pulmonary aspergillosis confirmed by open marrow transplant. Young reported a positive lung biopsy, transbronchial biopsies, and result for BAL in the diagnosis of pneumonia necropsy. Von EiV et al25 reported a sensitivity or malignancy in 93% of all patients.13 of 87% and a specificity of 100% for BAL. A Most of the studies on pneumonia in immu- single positive sputum culture for Aspergillus nocompromised patients do not diVerentiate has to be viewed critically, while a positive BAL between the diVerent underlying diseases. culture in immunocompromised patients While Streptococcus pneumoniae is the most should not be underestimated as contamina- common pathogen responsible for pneumonia tion but should, in view of a mortality rate of in immunocompetent patients14 and Pneumo- approximately 14.8% reported for invasive cystis carinii is one of the main causes of pneu- pulmonary aspergillosis,26 serve as an indica- monia in HIV infected patients,15 16 von EiV11 tion for the early onset of treatment with reported Gram negative bacteria (in 38 of 90 amphotericin B.26 Aspergillus was cultured from patients) and fungi (in 34 of 90 patients) to be BAL fluid in five of seven patients in our study. the most common pathogens in patients with The two patients in whom BAL was not diag- haematological malignancies and pneumonia. nostic were subsequently shown to have an We found both Gram positive (35%) and Gram aspergilloma which cannot be detected by negative species (40%) as well as fungi and BAL. One patient with a positive sputum mycobacteria (11% each) in the BAL fluid. culture for Aspergillus and a negative BAL find- The most diYcult problem encountered in ing underwent open lung biopsy with no this study was deciding whether a positive cul- evidence of invasive pulmonary aspergillosis, ture was consistent with a pathogen responsible confirming the accuracy of the BAL findings for pneumonia or whether it was caused by obtained by our investigation. Marra et al10 contamination of the nasopharynx. In particu- found that Aspergillus was responsible for lar, for Candida and coagulase negative Staphy- pneumonia in 20% of all patients with lococcus there are no firm recommendations as lymphoma or leukaemia. Analysis of the 62 to whether a positive culture in the BAL fluid patients in our study in whom a pathogen was of immunocompromised patients with signs of cultured from the BAL fluid showed that pneumonia necessitates treatment. A distinct Aspergillus was responsible for pneumonia in increase in infections in immunocompromised 11% of patients. patients caused by multiresistant or methicillin Mycobacteria other than tuberculosis resistant coagulase negative Staphylococcus has (MOTT)—for example, M xenopii—are con- been described in the literature.17 Immuno- sidered to be pathogenic in immunocompro- compromised patients are known to develop mised patients when cultured in a sterile severe pneumonia.18 The diagnosis of coagu- medium such as BAL fluid or found in a biopsy http://thorax.bmj.com/ lase negative Staphylococcus infection (not con- specimen compatible with the clinical course. tamination) is suggested when the organism is MOTT detected in a “non-sterile” medium cultured from normally “sterile” materials such such as sputum were considered to be as BAL fluid or blood,12 when it is cultured pathogenic when more than 100 colonies could several times in diVerent blood samples,19 or be seen in each of four diVerent samples.27 when it is resistant to five or more diVerent Although M xenopii may occur in immuno- antibiotics.20 In this study we accepted coagu- competent subjects as a contaminant, this is lase negative Staphylococcus as a pathogen only not the case in immunocompromised pa- when it was detected in at least two diVerent tients.27 on September 26, 2021 by guest. Protected copyright. sources or twice in diVerent samples taken on An important question remains if BAL find- diVerent days and when it was resistant to at ings are to be of therapeutic use in patients with least five antibiotics. BSAR, most of whom routinely receive this Candida is a common organism of the treatment. We detected pathogens responsible mucous membrane,1 but damage to the for pneumonia in the BAL fluid of 62 of the 95 mucous membrane resulting from immuno- patients (65%). Microbiological findings suppression leads to disseminated severe infec- showed that the pathogens cultured from the tion.21 We therefore accepted Candida as a BAL fluid were resistant to treatment in 58% of pathogen in immunocompromised patients patients and to standard BSAR consisting of with pneumonia and a concordant clinical teicoplanin, ceftriaxon, tobramycin, and am- course only when it was detected in diVerent photericin B in 32% of all patients. There are sources, or twice in diVerent samples taken on few data in the literature to indicate whether diVerent days, or in combination with a positive BSAR consisting of teicoplanin, ceftriaxon, cytological result.22 tobramycin, and amphotericin B is eVective in There are no definite recommendations for the treatment of pathogens, or if the treatment the diagnosis of an Aspergillus infection and the regimen is changed after BAL. Marra et al10 diagnosis of invasive pulmonary aspergillosis reported that the standard BSAR was changed can only be established by biopsy.23 However, in 18% of patients as a result of BAL findings. this cannot be performed in most immuno- Pathogens from the BAL fluid of patients with compromised patients because of coexisting hospital acquired pneumonia were resistant to thrombocytopenia. Although contamination of BSAR in 67% of cases10 and pathogens BAL fluid cannot be excluded, it is cultured from the BAL fluid of patients with nevertheless a valuable diagnostic tool for immunosuppression were resistant in 43% of

www.thoraxjnl.com 120 Hohenadel, Kiworr, Genitsariotis, et al

9 4 Maschmeyer G, Link H, Hiddemann W, . Empirische cases. Empirical antimicrobial treatment was et al Thorax: first published as 10.1136/thorax.56.2.115 on 1 February 2001. Downloaded from antimikrobielle Therapie bei neutropenischen Patienten. verified by bronchoscopy in 25 of 90 patients Med Klin 1994;89:114–23. and was successfully changed in 34 patients.11 5 The GLM procedure. SAS user’s guide. Cary, NC: SAS Insti- tute, 1985. Our findings showed that causative pathogens 6 Chopra SK, Mohsenifar Z. Fiberoptic bronchoscopy in detected exclusively in the BAL fluid in 29 of diagnosis of opportunistic lung infections. West J Med our 95 patients were resistant to the treatment 1979;131:4–8. 7 Martin WJ, Smith TF, Brutinel WM, et al. Role of broncho- given in 16 patients who would therefore have alveolar lavage in the assessment of opportunistic pulmo- received inadequate treatment without BAL. nary infections: utility and complications. Mayo Clin Proc 1987;62:549–57. Treatment with standard BSAR including 8 Stover DE, Muhammad BZ, Steven IH, et al. Broncho- teicoplanin, ceftriaxon, tobramycin, and am- alveolar lavage in diagnosis of diVuse pulmonary infiltrates in the immunosuppressed host. Ann Intern Med 1984;101: photericin B would have been ineVective in 12 1–7. of 29 patients (41%) with pathogens detected 9 DalhoV K, Braun J, Hollandt H, et al. Diagnostic value of bronchoalveolar lavage in patients with opportunistic and in BAL fluid only. nonopprtunistic bacterial pneumonia. Infection 1993;21: Complications related to BAL consisted of 291/17–296/22. 10 Marra R, Pagano L, Pagliari G, et al. The yield of broncho- minor self-limiting bleeding in 14.7% of alveolar lavage in the etiologically diagnosis of pneumonia patients; only one patient had more severe, in leukemia and lymphoma patients. Eur J Haematol 1993; 51:256–8. non-haemodynamically relevant bleeding, al- 11 Von EiV M, Zuhlsdorf M, Roos N, et al. Pulmonary though 77 of 95 patients were thrombocyto- infiltrates in patients with hematologic malignancies: clinical usefulness of non-invasive bronchoscopic procedures. penic. No pneumothorax occurred. Studies on Eur J Haematol 1995;54:157–62. BAL related complications with a particular 12 Cordonnier C, Bernaudin JF, Fleury J, et al. Diagnostic yield of bronchoalveolar lavage in pneumonitis occurring after focus on immunocompromised patients have allogenic bone marrow transplantation. Am Rev Respir Dis reported self-limiting bleeding, bronchial 1985;132:1118–23. 28 13 Young JA, Hopkin JM, Cuthbertson WP. Pulmonary spasm, and coughing in 33% of patients. Von infiltrates in immunocompromised patients. Diagnosis by 25 EiV et al noted that complications such as cytologic examination of bronchoalveolar lavage fluid. J Clin Pathol 1984;37:390–7. pneumothorax, respiratory symptoms requir- 14 Østergaard L, Andersen PL. Etiology of community ing treatment in the intensive care unit, and acquired pneumonia. Chest 1993;104:1400–7. self-limiting bleeding occurred in 5% of 15 Lerner CW, Tapper ML. Opportunistic infection complicat- ing acquired immune deficiency syndrome. Medicine 1984; patients with haematological disorders, trans- 63:155–64. plant recipients, or those with AIDS. Zavala29 16 Müller IA, Gloger J, Genitsariotis R, et al. The role of bronchoalveolar lavage in HIV-infected patients with pneu- reported bleeding following bronchoscopy in monia in comparison to non-invasive procedures. Am J 26% of immunocompromised patients. Respir Crit Care Med 1997;155:A-472 17 Vincent JL, Bihari DJ, Suter PM, et al. The prevalence of Thus, BAL may be considered to be a nosocomial infections in intensive care units in Europe. relatively safe procedure even for patients with JAMA 1995;274:639–44. 18 Jones RN. Impact of changing pathogens and antimicrobial a low platelet count when compared with the susceptibility patterns in the treatment of serious infections risk of untreated or unsuccessfully treated in hospitalized patients. Am J Med 1996;100:3–12S 19 Rubio M, Palau L, Vivas JR, et al. Predominance of pneumonia in patients with neutropenia. In gram-positive microorganisms as a cause of septicemia in view of the cost of bronchoscopy (about 360 patients with hematological malignancies. Infect Control

30 Hosp Epidemiol 1994;15:101–4. http://thorax.bmj.com/ German marks or 180 Euro), this is a very 20 Christensen GD, Bisno AL, Parisi JT, et al. Nosocomial sep- small sum of money compared with the costs of ticemia due to multiply antibiotic-resistant Staphylococcus epidermidis. Ann Intern Med 1982;96:1–10. a longer stay in hospital for severely ill patients. 21 Rosenow EC, Wilson WR, Cockerill FR. Pulmonary disease However, the question of cost eVectiveness in the immunocompromised host. Mayo Clin Proc 1985;60: 473–87. goes beyond the scope of this study. Our results 22 Saito H, Anaissie EJ, Morice RC, et al. Bronchoalveolar lav- suggest that BAL may be essential in the diag- age in the diagnosis of pulmonary inﬁltrates in patients with acute leukemia. Chest 1988;94:745–9. nosis of pneumonia in immunocompromised 23 Maschmeyer G, Link H, Hiddemann W, et al. Pulmonary patients. inﬁltrations in febrile patients with neutropenia. Cancer In conclusion, we have shown that the 1994;73:2296–304.

24 Kahn FW, Jones JM, England DM. The role of broncho- on September 26, 2021 by guest. Protected copyright. causative organisms were not detected by less alveolar lavage in the diagnosis of invasive pulmonary aspergillosis. Am J Clin Pathol 1986;86:518–23. invasive tools in 29 of 95 patients (31%) and 25 von EiV M, Steinemann R, Roos N, et al. Pneumonien bei that the pathogens found were resistant to abwehrgeschwächten Patienten: Stellenwert nicht biopt- ischer bronchoskopischer Untersuchungsverfahren in der BSAR in 12 of the 29 patients (41%). Even Erregerdiagnostik. Klin Wochenschr 1990;68:372–9. though 77 of 95 patients were thrombocyto- 26 Yu VL, Muder RR, Poorsatter A. Significance of isolation of Aspergillus from the respiratory tract in diagnosis of invasive penic, no major bleeding occurred. BAL is pulmonary aspergillosis. Results from a three-year prospec- therefore a relatively safe procedure in the tive study. Am J Med 1986;81:249–54. 27 Tenholder MF, Moser RJ, Tellis CJ. Mycobacteria other diagnosis of pneumonia in immunocompro- than tuberculosis. Pulmonary involvement in patients with mised patients and in those treated with BSAR. acquired immunodeficiency syndrome. Arch Intern Med 1988;148:953–6. 28 Kahn WF, Jones JM. Analysis of bronchoalveolar lavage 1 Edelstein H. Antimicrobial chemotherapy for Legionnaire’s specimens from immunocompromised patients with a pro- disease: a review. Clin Infect Dis 1995;2–3:265–76. tocol applicable in the microbiology laboratory. J Clin 2 Reynolds HY. State of art: bronchoalveolar lavage. Am Rev Microbiol 1988;26:1150–5. Respir Dis 1987;135:250–63. 29 Zavala DC. Pulmonary hemorrhage in fiberoptic transbron- 3 Hunninghake GW, Gadek JE, Kawanami O, et al. Inflamma- chial biopsy. Chest 1976;70:584–8. tory and immune processes in the human lung in health 30 Broglie MG, Pranschke-Schade S, Schade H-J. In: Gebühr- and disease: evaluation by bronchoalveolar lavage. Am J enhandbuch. Kommentar für Ärzte. Wiesbaden: Medical Clin Pathol 1979;97:149–206. Tribune Verlagsgesellschaft mbH, 1996.