The role of neutropenia on outcomes of cancer patients with community-acquired pneumonia

Transcription

1 Er Respir J 2009; 33: DOI: / CopyrightßERS Jornals Ltd 2009 The role of netropenia on otcomes of cancer patients with commnity-acqired pnemonia S. Aliberti*,#, J.A. Myers ", P. Peyrani #, F. Blasi*, R. Menendez +, P. Rossi 1, R. Cosentini e, G. Lopardo**, L. de Vedia ## and J.A. Ramirez # ABSTRACT: Althogh the presence of netropenia may predispose cancer patients to develop commnity-acqired pnemonia, the role of netropenia on their otcomes has not been investigated. The prpose of the present stdy was to compare clinical otcomes of cancer commnity-acqired pnemonia patients with and withot netropenia. Patients with cancer, identified in the Commnity-Acqired Pnemonia Organization database, were divided into two grops according to the type of cancer and the presence of netropenia: patients with solid cancer withot netropenia verss those with fnctional or absolte netropenia. Among the 3,106 commnity-acqired pnemonia patients enrolled, 135 had cancer withot netropenia and 75 had cancer with netropenia. No significant difference was fond between patients with and withot netropenia regarding mean time to clinical stability ( verss days, respectively), mean length of hospital stay ( verss days) and in-hospital mortality (18 verss 15%, respectively). Using a mltiple logistic regression model, netropenia was not associated with mortality in cancer patients when adjsting for significant covariates (odds ratio 1.30). Lack of netropenia, dring the initial evalation of a cancer commnity-acqired pnemonia patient, shold not be considered an indicator of better clinical otcome. KEYWORDS: Cancer, commnity-acqired pnemonia, netropenia, otcomes Infectios diseases accont for high morbidity and mortality in patients with cancer and, among them, commnity-acqired pnemonia (CAP) represents the most common and lifethreatening disease [1 3]. The development of CAP in cancer patients can be de to an alteration of the immne-defence mechanisms reslting from either the malignancy or its treatment. The risk of infection, de to the natre of cancer, can be associated with hmoral immne deficit, celllar immne deficit or netrophil disorders. Haematological malignancies can predispose patients to infections becase of the replacement of the marrow with malignant cells. Conseqently, these patients have fnctional netropenia even thogh they may have normal or even increased nmbers of netrophils. Moreover, cancer patients may experience netropenia as a side effect of chemotherapy or radiotherapy (absolte netropenia). The degree of netropenia has been considered the single most important factor contribting to infection in cancer patients, particlarly when the netrophil cont drops below 500 cells?mm -3 [4, 5]. The overall mortality recorded in febrile netropenic patients with cancer is 30 50% [6]. Dring the past two decades, treatment of infections in the cancer poplation has been primarily focsed on the management of fever and netropenia, de to the fact that the site of infection cannot be determined in 50 80% of cancer patients [7, 8]. The American Thoracic Society gidelines for the management of CAP, pblished in 2001, sed netropenia in order to identify more severe cancer patients with CAP [9]. Patients with haematological malignancy, experiencing fnctional netropenia, or patients affected by any type of cancer and having absolte netropenia were exclded from the gidelines. The decision to inclde patients with solid cancer withot netropenia in the gidelines was based only on expert opinions. De to this, physicians may feel AFFILIATIONS *Institte of Respiratory Disease, University of Milan, Ospedale Maggiore Fondazione IRCCS Policlinico, Mangiagalli e Regina Elena, and e Emergency Medicine Dept, Ospedale Maggiore Fondazione IRCCS Policlinico, Mangiagalli e Regina Elena, Milan, and 1 Division of Internal Medicine, Dept of Medicine, Azienda Ospedaliera S. Maria della Misericordia, Udine, Italy. # Division of Infectios Diseases, Dept ofmedicine,universityofloisville,and " Dept of Bioinformatics and Biostatistics, School of Pblic Health and Information Sciences, University of Loisville, Loisville, Kentcky, USA. + Pnemology Service, La Fe University Hospital, Valencia, Spain. **Division of InfectiosDiseases, Dept of Medicine, Hospital Bernardo Hossay, and ## Hospital Francisco J. Mñiz, Benos Aires, Argentina. CORRESPONDENCE J.A. Ramirez, Division of Infectios Diseases, Dept of Medicine, University of Loisville, Loisville, KY, USA. Fax: j.ramirez@loisville.ed Received: December Accepted after revision: September This work won the Yong Researcher Award at the 2007 ERS Annal Congress. This poster was accepted and presented as a poster at the 2007 ERS Annal Congress. This manscript has spplementary data accessible from For editorial comments see page 6. STATEMENT OF INTEREST: A statement of interest for F. Blasi can be fond at Eropean Respiratory Jornal Print ISSN Online ISSN VOLUME 33 NUMBER 1 EUROPEAN RESPIRATORY JOURNAL

2 S. ALIBERTI ET AL. NEUTROPENIA IN CANCER AND CAP confident in treating cancer patients withot netropenia as reglar patients with CAP. However, the role that netropenia may have in clinical otcomes of cancer patients hospitalized with CAP has not yet been investigated. In order to better define clinical otcomes of cancer patients hospitalised for an episode of CAP, the present stdy had the following objectives: 1) to compare clinical otcomes in cancer patients verss immnocompetent patients; 2) to compare clinical otcomes in cancer patients with absolte or fnctional netropenia verss those withot netropenia. MATERIALS AND METHODS Stdy design and sbjects A secondary analysis of the Commnity-Acqired Pnemonia Organization (CAPO) database was performed. The database contains data retrospectively collected from 43 hospitals in 12 contries between Jne 2001 and Janary In each participating centre, primary investigators selected nonconsective adlt hospitalised patients diagnosed with CAP. All data were collected on a case report form and transferred electronically to the CAPO co-ordinating centre at the University of Loisville (Loisville, KY, USA). A sample of the stdy protocol and the data collection forms are available at the stdy website [10]. Discrepancies and inconsistencies in the data were determined at the co-ordinating centre. After all qeries were clarified, cases were entered into the database. Local instittional review board approval was obtained for each stdy site. Inclsion and exclsion criteria Patients aged o18 yrs and satisfying the criteria for CAP were inclded in the stdy. In order to investigate primarily immnosppression de to cancer, patients with a diagnosis of HIV infection were exclded. Stdy definition CAP was defined as the presence of a new plmonary infiltrate on chest radiograph at the time of hospitalisation associated with at least one of the following: 1) new or increased cogh; 2) an abnormal temperatre (,35.6 or.37.8c); 3) an abnormal serm lekocyte cont (lekocytosis, left shift or lekopenia defined by local laboratory vales). Cancer was defined as any type of malignancy that was diagnosed in the previos 12 months or as active cancer. Active cancer was considered to be present in patients who had received chemotherapy and/or radiotherapy in the previos 12 months or in patients who exhibited signs or symptoms of cancer dring the past year. Patients with basal or sqamos cell cancer of the skin were considered immnocompetent. Netropenia was defined as fnctional or absolte netropenia. Fnctional netropenia was considered present in all patients affected by haematological malignancy. Absolte netropenia was defined as an absolte netrophils cont (ANC),500 cells?mm -3 on admission. Netropenia was not considered in patients with a septic state. Stdy grops Patients with CAP were divided into two grops according to the presence of cancer: sbjects withot cancer, Grop 1, and sbjects with cancer, Grop 2. Patients in Grop 2 were divided into two sbgrops according to the type of cancer and the presence of netropenia: patients with solid cancer withot netropenia, Grop 2a, and patients with netropenia, Grop 2b. Otcomes Mortality, length of stay and time to clinical stability (TCS) were considered as otcomes and their definitions are presented in the online data spplement. Statistical analysis Descriptive statistics were reported at baseline with continos data expressed as a mean SD and categorical data expressed as conts. Patient characteristics were compared between Grop 1 and Grop 2 and between Grop 2a and Grop 2b. Smmary statistics for all continos explanatory variables are presented as means with differences between the two grops compared by means of independent npaired t- tests. Categorical explanatory variables are smmarized as freqencies and percentages with differences between the two grops analysed sing the chi-sqared test. A p-vale f0.05 was considered statistically significant. The one-way ANOVA techniqe followed by the Dnnett s post hoc comparison techniqe was initially sed to compare mortality rates between Grop 1, Grop 2a and Grop 2b. The associations between mortality and cancer among the stdy poplation and between mortality and netropenia among Grop 2 were also stdied sing mltiple logistic regression models. All the significant explanatory variables and those that were considered of clinical relevance were incorporated into the model by tilising mltiple logistic regression techniqes. Differences between Grop 1 and Grop 2 and between Grop 2a and Grop 2b were stdied regarding time to clinical stability (TCS) and length of stay (LOS). A nivariate analysis of comparing mean LOS and TCS was performed. Differences were frther assessed with the Kaplan Meier method, the logrank test and Cox regression analysis patients Grop patients 2960 patients 135 patients Grop 2a 146 patients exclded 310 patients Grop patients 73 patients Grop 2b 102 missing clinical or laboratory information FIGURE 1. The stdy flow chart. Grop 1: commnity-acqired pnemonia (CAP) patients withot cancer. Grop 2: CAP patients with cancer. Grop 2a: CAP patients with solid cancer withot netropenia. Grop 2b: CAP patients with netropenia. c EUROPEAN RESPIRATORY JOURNAL VOLUME 33 NUMBER 1 143

3 NEUTROPENIA IN CANCER AND CAP S. ALIBERTI ET AL. RESULTS A total of 3,106 patients with CAP were enrolled dring the stdy period. The stdy flow chart is shown in figre 1. Characteristics of patients with and withot cancer Demographics, comorbidities, severity of the disease, clinical, laboratory and radiological findings of the stdy poplation on admission according to the presence of cancer are smmarised in table 1. TABLE 1 Demographics, comorbidities, severity of the disease, clinical, laboratory and radiological findings of the stdy poplation on admission, according to the presence of cancer Characteristic Grop 1 Grop 2 p-vale Demographics Sbjects n Male 1604 (61) 214 (69) Age mean SD yrs Age.65 yrs 1535 (58) 215 (69) Age.75 yrs 1014 (38) 126 (41) Comorbidities Congestive heart failre 577 (22) 55 (18) COPD 795 (30) 90 (29) Diabetes mellits 529 (20) 55 (18) Cerebrovasclar accident 436 (17) 41 (13) Renal disease 301 (11) 45 (15) Liver disease 95 (3.6) 15 (4.8) Severity on admission Altered mental stats 349 (13) 40 (13) Admission to ICU 322 (12) 38 (12) MV on admission 83 (3.1) 15 (5.0) Physical findings Temperatre,35 or o40c 159 (6.0) 11 (3.6) Respiratory rate o30 breaths?min (22) 59 (19) Alteration of gas exchange # 1,011(38) 125 (41) Hypotension " 148 (5.6) 23 (7.5) Heart rate o125 beats?min (12) 33 (11) Laboratory vales Arterial ph, (6.3) 11 (3.7) Sodim,130 mmol?l (5.9) 27 (8.7) Haematocrit,30% 131 (4.9) 65 (21) BUN.30 mg?dl (20) 74 (24) Radiology findings on CXR Mltilobar involvement 714 (27) 86 (28) Pleral effsion 479 (18) 68 (22) Cavitation 28 (1.1) 4 (1.3) Data are presented as n (%), nless otherwise stated. Grop 1: commnityacqired pnemonia (CAP) patients withot cancer; Grop 2: CAP patients with cancer; COPD: chronic obstrctive plmonary disease; ICU: intensive care nit; MV: mechanical ventilation; BUN: blood rea nitrogen; CXR: chest radiograph. # : arterial oxygen tension,60, Pa,O 2/inspiratory oxygen fraction,300 or oxygen satration,90%; " : systolic blood pressre,90 mmhg or diastolic blood pressre,60 mmhg. Distribtion of the type of cancer among patients in Grop 2 is shown in table 2. A co-existence of malignancy affecting two different organs was fond in three patients. Otcomes of patients with and withot cancer The mortality rate was significantly higher in patients with cancer when compared with patients withot cancer (43 ot of 310 patients, 14% verss 213 ot of 2,650, 8.0%, respectively; p50.001). In the 13 (30%) ot of 43 patients among those with cancer who died, mortality was considered to be related to CAP. Using mltiple logistic regression methods, this reslt was consistent: the presence of cancer on admission among the stdy poplation was fond to significantly impact mortality when adjsting for significant covariates (odds ratio (OR) 1.63, 95% confidence interval (CI) ; p50.020). In comparison to patients withot cancer, those with cancer had a significantly longer TCS ( days verss days, respectively; p50.001) and LOS ( days verss days, respectively; p50.001). Using the Kaplan Meier and log-rank methods, cancer on admission was fond to significantly impact both TCS (Log rank Chisqared 12.45; p,0.0001) and LOS (Log-rank Chi-sqared 10.87; p50.001). Characteristics of cancer patients with and withot netropenia Baseline demographics, comorbidities, severity of the disease, clinical, laboratory and radiological findings among cancer patients stratified by netropenia are smmarised in table 3. Otcomes of cancer patients with and withot netropenia The mean TCS was similar between cancer patients with and withot netropenia ( verss days, respectively; p50.264). Using the Kaplan Meier and log-rank methods to evalate TCS, no statistical difference was fond between the two grops (Log rank Chi-sqared 0.859; p50.354; fig. 2). This reslt held consistent after adjstment for significant explanatory variables inclded in the logistic model (hazard ratio (HR) 0.911, 95% CI ; p50.566). TABLE 2 Solid tmor Type of malignancy in grop of patients with cancer (Grop 2) Patients n Haematological malignancy Patients n Lng cancer 60 Lekaemia 22 Prostate cancer 18 Non-Hodgkin s lymphoma 15 Breast cancer 14 Mltiple myeloma 12 Head and neck cancer 11 Myelodisplasia 10 Bladder cancer 8 Hodgkin s lymphoma 3 Colon cancer 8 Aplastic anaemia 2 Kidney cancer 4 Policytemia vera 2 Gastric cancer 3 Other 5 Cervical cancer 2 Pancreatic cancer 2 Other VOLUME 33 NUMBER 1 EUROPEAN RESPIRATORY JOURNAL

4 S. ALIBERTI ET AL. NEUTROPENIA IN CANCER AND CAP TABLE 3 Demographics, comorbidities, severity of the disease, clinical, laboratory and radiological findings on admission of the cancer poplation, according to the presence of netropenia Characteristic Grop 2a Grop 2b p-vale Demographics Sbjects n Male 105 (78) 51 (70) Age mean SD yrs Age.65 yrs 94 (70) 44 (60) Age.75 yrs 53 (39) 26 (36) Comorbidities Congestive heart failre 30 (22) 18 (25) COPD 47 (35) 14 (19) Diabetes mellits 26 (20) 12 (16) Cerebrovasclar accident 23 (17) 4 (6) Renal disease 26 (19) 11 (15) Liver disease 9 (6.7) 1 (1.4) Severity on admission Altered mental stats 21 (16) 5 (7) Admission to ICU 21 (16) 9 (12) MV on admission 7 (5.2) 4 (6.2) Physical findings Temperatre,35 or o40c 6 (4.5) 3 (4.2) Respiratory rate o30 breaths?min (20) 13 (19) Alteration of gas exchange # 56 (42) 29 (41) Hypotension " 11 (8.3) 6 (8.2) Heart rate o125 beats?min (15) 6 (11) Laboratory vales Arterial ph, (3.7) 1 (1.6) Sodim,130 mmol?l -1 9 (6.7) 11 (15.3) Haematocrit,30% 15 (11) 29 (41) BUN.30 mg?dl (20) 29 (40) Radiology findings on CXR Mltilobar involvement 36 (27) 26 (36) Pleral effsion 32 (24) 17 (23) Cavitation 3 (2.2) 0 (0) N/A Medical treatment ABT Compliant to local gidelines 116 (88) 51 (85) Floroqinolone 46 (34) 34 (47) b-lactam 108 (80) 49 (67) Macrolide 58 (43) 29 (40) Data are presented as n (%), nless otherwise stated. Grop 2a: commnityacqired pnemonia (CAP) patients with solid cancer withot netropenia; Grop 2b: CAP patients with fnctional or absolte netropenia; COPD: chronic obstrctive pimonary disease; ICU: intensive care nit; MV: mechanical ventilation; BUN: blood rea nitrogen; CXR: chest radiograph. # : arterial oxygen tension (Pa,O2),60, Pa,O2/inspiratory oxygen fraction,300 or oxygen satration,90%; " : systolic blood pressre,90 mmhg or diastolic blood pressre,60 mmhg. The mean LOS was similar between cancer patients with and withot netropenia ( verss days, respectively; p50.638). Using the Kaplan Meier and log rank methods to evalate LOS, no statistical difference was fond between the two grops (Log rank Chi-sqared 0.043; p50.837; fig. 3). This reslt held consistent after adjstment for Proportion of patients who did not reach TCS TCS days FIGURE 2. Kaplan Meier srvival crves for time to clinical stability (TCS) in commnity-acqired pnemonia patients with ( ) and withot (------) netropenia. Proportion of patients who were not discharged LOS days FIGURE 3. Kaplan Meier srvival crves for length of stay (LOS) in commnity-acqired pnemonia patients with ( ) and withot (------) netropenia. significant explanatory variables inclded in the logistic model (HR 1.055, 95% CI ; p50.728). The mortality of the stdy poplation according to the presence of netropenia is shown in figre 4. No significant difference in mortality rate was fond in cancer patients with and withot netropenia (18 verss 15%, respectively; p50.573). A significant difference in mortality was fond between patients withot cancer and both cancer patients withot netropenia (8 verss 15%, respectively; p50.006) and with netropenia (8 verss 18%, respectively; p50.003). Using a mltiple logistic regression model, netropenia was fond to not be significantly associated to mortality in cancer patients (OR 1.56, 95% CI ; p50.474), when adjsting for significant covariates. When the stepwise procedre was applied to the fll model, netropenia did not srvive the stepwise procedre, while CRB-65 (confsion, respiratory rate o30 breaths?min -1, low blood pressre (systolic vale,90 mmhg or diastolic vale f60 mmhg) and age o65 yrs) c EUROPEAN RESPIRATORY JOURNAL VOLUME 33 NUMBER 1 145

5 NEUTROPENIA IN CANCER AND CAP S. ALIBERTI ET AL. Mortality rate % Grop 1 # Grop 2a Grop 2b TABLE 4 Mltivariable analysis of mortality in patients with commnity-acqired pnemonia and cancer Characteristic OR (95% CI) p-vale Netropenia ( ) COPD ( ) Cerebrovasclar accident ( ) PSI Risk Class IV V ( ) CRB-65 Score ( ) BUN.30 mg?ml ( ) Haematocrit,30% ( ) Sodim,130 mmol?l ( ) Compliant ABT ( ) FIGURE 4. Mortality rate in the stdy poplation, presented in the cancer grop according to the netrophil stats. Grop 1: commnity-acqired pnemonia (CAP) patients withot cancer; Grop 2a: CAP patients with solid cancer withot netropenia; Grop 2b: CAP patients with fnctional or absolte netropenia. # :p50.006; " :p The nmber of deaths in Grop 1, 2a and 2b were 213, 20 and 13, respectively. The nmber of patients in Grop 1, 2a and 2b were 2,650, 135 and 73, respectively. did srvive as independently associated with mortality (OR 4.59, 95% CI ; p50.034; fig. 5 and table 4). DISCUSSION This stdy indicates that, in patients with CAP, the presence of malignancy significantly increases mortality and worsens clinical otcomes. Among CAP patients with cancer, those withot netropenia have similar otcomes compared to those with netropenia. The presence of netropenia on admission was not an independent risk factor for mortality in patients with cancer ndergoing an episode of CAP. The present findings sggest that physicians shold aggressively manage cancer patients with CAP, regardless of the type of cancer or the netrophil cont on admission. The lack of netropenia dring the initial evalation of a cancer patient with CAP shold not be considered an indicator of better clinical otcomes. Althogh netropenia is a key risk factor for infections, other factors seem to pose cancer patients with CAP at eqal risk for poor otcomes. At least three considerations can explain the findings of the present stdy. First, even thogh patients with solid tmors may have a normal netrophil cont, cytotoxic drgs sed for the treatment of cancer can affect chemotactic and phagocytic fnction, withot reflecting on the total cell cont. Secondly, the fnctional capabilities of phagocytes sch as netrophils, eosinophils and mononclear cells may be intrinsically defective even before the initiation of chemotherapy. A stdy performed by HÜBEL et al. [11] showed cancer patients having preactivated phagocytes with sppressed fnction prior to initiation of treatment. Thirdly, hmoral and celllar immne mechanisms mediated by B and T lymphocytes, respectively, can be profondly altered in patients with solid cancer. Mltiple aspects of the host defence system can be simltaneosly impaired in cancer patients hospitalised with CAP, ths inflencing their otcomes. This mixed pattern of immnodeficiency may explain the lack of improved otcomes in the present patients withot netropenia. OR: odds ratio; CI: confidence interval; COPD: chronic obstrctive plmonary disease; PSI: pnemonia severity index; CRB-65: confsion, respiratory rate o30 breaths?min -1, low blood pressre (systolic vale,90 mmhg or diastolic vale f60 mmhg) and age o65 yrs; BUN: blood rea nitrogen; ABT: antibiotic therapy. The reslts of the present stdy are consistent with recent literatre which has qestioned the role of netropenia in otcomes of both haematological and solid tmors patients ndergoing an infection. Netropenia was not fond to affect mortality in cancer patients with Streptococcs pnemoniae bacteraemia [12], as well as bacteraemia cased by other micro-organisms [13]. In line with these data, five other stdies evalating patients hospitalised in ICU showed netropenia to not be independently associated with mortality in cancer patients with different sites of infection [14 18]. Dring the past decades, the attention in management of infections in cancer patients has focsed on fever and netropenia [19]. ToHohe latest gidelines pblished by the National Comprehensive Cancer Network, however, recommend that immnocompromised non-netropenic cancer patients receive eqal attention as those with netropenia [20]. Netropenia COPD Cerebrovasclar accident PSI Risk Class IV-V CRB-65 Score 3 4 BUN >30 mg dl -1 Haematocrit <30 % Sodim <130 mmol L -1 Compliant ABT OR (95% CI) FIGURE 5. Mltivariable analysis of mortality in patients with commnityacqired pnemonia and cancer. All variables inclded in the model were dichotomised: yes verss no. OR: odds ratio; Cl: confidence interval; COPD: chronic obstrctive plmonary disease; PSI: pnemonia severity index; CRB-65: confsion, respiratory rate o30 breaths?min -1, low blood pressre (systolic vale,90 mmhg or diastolic vale f60 mmhg) and age o65 yrs; BUN: blood rea nitrogen; ABT: antibiotic therapy VOLUME 33 NUMBER 1 EUROPEAN RESPIRATORY JOURNAL

6 S. ALIBERTI ET AL. NEUTROPENIA IN CANCER AND CAP One important limitation of the present stdy, in evalating the effect of netropenia on clinical otcomes, is the lack of information regarding the dration of netropenia before admission to the hospital, the development of netropenia or the netropenia recovery dring hospitalisation. Moreover, in light of the fact that the present was a retrospective stdy, the athors were not able to collect more data on malignancies, neither on chemotherapies nor on bone marrow transplant. This stdy is strengthened by the large patient cohort involving 43 instittions in 12 different contries and by its se of an nselected poplation. The sample was composed of sbjects admitted to general hospitals and not by those referring to specialised cancer facilities only. In conclsion, the present data indicate that the lack of netropenia in cancer patients hospitalised for CAP does not prevent poor clinical otcomes; therefore, physicians cannot be reassred of a better otcome when netropenia is not present at the initial evalation of a cancer patient with CAP. Ths, physicians shold aggressively manage all cancer patients with CAP, regardless of the type of cancer or the netrophil cont. ACKNOWLEDGEMENTS TheCAPOinvestigatorsareasfollows.R.Nakamats(Veterans Affairs Medical Center); F. W. Arnold and M. Allen (University Hospital); G. Broch (Dept of Bioinformatics and Biostatistics, University of Loisville, all Loisville, KY, USA); J. Bordon (Providence Hospital, Washington, DC, USA); P. Gross (Hackensack University Medical Center, Hackensack, NJ, USA); K. Weiss (Maisonneve-Rosemont Hospital, University of Montreal, Montreal, Canada); D. Legnani (Ospedale L. Sacco, Milan, Italy); M. Bodi (Hospital Universitario Joan XXIII); J. Porras (Hospital Sant Pa i Santa Tecla, both Tarragona, Spain); A. Torres (Institto de Nemonologia y Cirgia Toracica, Barcelona, Spain); H. Lode (City Hosp. E.v.Behring/Lngenklinik Heckeshorn, Berlin, Germany); J. Roig (Hospital Nostra Senyora de Meritxell, Escaldes, Andorra); G. Benchetrit (IDIM A. Lanari, Benos Aires (BA)); J. Gonzalez (Hospital Enriqe Torn, BA); A. Videla (Hospital Universitario Astral, BA); J. Corral (Hospital Dr Oscar Alende, Mar del Plata); J. Martinez (Institto Medico Platense, La Plata); E. Rodrigez (Hospital Espanol de La Plata, La Plata); M. Rodrigez (Hospital Rodolfo Rossi, La Plata); C. Victorio (Clinica Urgay, Entre Rios, all Argentina); G. Levy (Hospital Universitario de Caracas, Caracas); F. Arteta (Hospital Lis Gomez Lopez-Ascardio, Barqisimeto, both Venezela); A. Diaz Fenzalida (Pontifica Universidad de Chile); M. Parada (Clinica las Condes, both Santiago, Chile); J. Lna (Hospital Nacional Roosevelt, Gatemala). The athors acknowledge the assistance of M. E. Allen (Division of Infectios Diseases) and E. Smigielski (Kornhaser Health Sciences Library; both University of Loisville, Loisville, KY, USA). REFERENCES 1 Annal report on trends in the health of Americans. Health, United States. National Center for Health Statistics, USA, Collin BA, Ramphal R. Pnemonia in the compromised host inclding cancer patients and transplant patients. Infect Dis Clin North Am 1998; 12: Rañó A, Agstí C, Benito N, et al. Prognostic factors of non- HIV immnocompromised patients with plmonary infiltrates. Chest 2002; 122: Pizzo PA. Fever in immnocompromised patients. N Engl J Med 1999; 341: Jones RN. Contemporary antimicrobial ssceptibility patterns of bacterial pathogens commonly associated with febrile patients with netropenia. Clin Infect Dis 1999; 29: Mandell LA, Wnderink RG, Anzeto A, et al. Consenss gidelines on the management of commnity-acqired pnemonia in adlts. Clin Infect Dis 2007; 44: Sppl. 2, S27 S72. 7 Rosenow EC III. Diffse plmonary infiltrates in the immnocompromised host. Clin Chest Med 1990; 11: Williams DM, Krick JA, Remington JS. Plmonary infection in the compromised host: part I. Am Rev Respir Dis 1976; 114: Niederman MS, Mandell LA, Anzeto A, et al. Gidelines for the management of adlts with commnity-acqired pnemonia. Diagnosis, assessment of severity, antimicrobial therapy, and prevention. Am J Respir Crit Care Med 2001; 163: Commnity-Acqired Pnemonia Organization (CAPO). Accessed and pdated daily. 11 Hübel K, Hegener K, Schnell R, et al. Sppressed netrophil fnction as a risk factor for severe infection after cytotoxic chemotherapy in patients with acte nonlymphocytic lekemia. Ann Hematol 1999; 78: Kmashi P, Girgawy E, Tarrand JJ, Rolston KV, Raad II, Safdar A. Streptococcs pnemoniae bacteremia in patients with cancer: disease characteristics and otcomes in the era of escalating drg resistance ( ). Medicine (Baltimore) 2005; 84: Velasco E, Byington R, Martins CA, Schirmer M, Dias LM, Goncalves VM. Comparative stdy of clinical characteristics of netropenic and non-netropenic adlt cancer patients with bloodstream infections. Er J Clin Microbiol Infect Dis 2006; 25: Kress JP, Christenson J, Pohlman AS, Linkin DR, Hall JB. Otcomes of critically ill cancer patients in a niversity hospital setting. Am J Respir Crit Care Med 1999; 160: Azolay E, Morea D, Alberti C, et al. Predictors of shortterm mortality in critically ill patients with solid malignancies. Intensive Care Med 2000; 26: Stadinger T, Stoiser B, Müllner M, et al. Otcome and prognostic factors in critically ill cancer patients admitted to the intensive care nit. Crit Care Med 2000; 28: Regazzoni CJ, Irrazabal C, Lna CM, Poderoso JJ. Cancer patients with septic shock: mortality predictors and netropenia. Spport Care Cancer 2004; 12: Brnet F, Lanore JJ, Dhainat JF, et al. Is intensive care jstified for patients with haematological malignancies? Intensive Care Med 1990; 16: Hghes WT, Armstrong D, Bodey GP, et al gidelines for the se of antimicrobial agents in netropenic patients with nexplained fever. Infectios Diseases Society of America. Clin Infect Dis 1997; 25: National Comprehensive Cancer Network. Prevention and treatment of cancer-related infections abot/news/newsinfo.asp?newsid5107 EUROPEAN RESPIRATORY JOURNAL VOLUME 33 NUMBER 1 147