Advertisement

Clinical characterisation of pneumonia caused by atypical pathogens combining classic and novel predictors

      Abstract

      The aim of this study was to characterise community-acquired pneumonia (CAP) caused by atypical pathogens by combining distinctive clinical and epidemiological features and novel biological markers. A population-based prospective study of consecutive patients with CAP included investigation of biomarkers of bacterial infection, e.g., procalcitonin, C-reactive protein and lipopolysaccharide-binding protein (LBP) levels. Clinical, radiological and laboratory data for patients with CAP caused by atypical pathogens were compared by univariate and multivariate analysis with data for patients with typical pathogens and patients from whom no organisms were identified. Two predictive scoring models were developed with the most discriminatory variables from multivariate analysis. Of 493 patients, 94 had CAP caused by atypical pathogens. According to multivariate analysis, patients with atypical pneumonia were more likely to have normal white blood cell counts, have repetitive air-conditioning exposure, be aged <65 years, have elevated aspartate aminotransferase levels, have been exposed to birds, and have lower serum levels of LBP. Two different scoring systems were developed that predicted atypical pathogens with sensitivities of 35.2% and 48.8%, and specificities of 93% and 91%, respectively. The combination of selected patient characteristics and laboratory data identified up to half of the cases of atypical pneumonia with high specificity, which should help clinicians to optimise initial empirical therapy for CAP.

      Keywords

      INTRODUCTION

      Selection of empirical therapy for community-acquired pneumonia (CAP) has become complicated by the marked increase in β-lactam and macrolide resistance among strains of Streptococcus pneumoniae, and by the concerns about atypical pathogens (e.g., Mycoplasma pneumoniae, Chlamydophila spp. and Legionella spp.) [
      • Marston BJ
      • Plouffe JF
      • File TM
      • et al.
      Incidence of community‐acquired pneumonia requiring hospitalization. Results of a population‐based active surveillance study in Ohio. The Community‐Based Pneumonia Incidence Study Group.
      ,
      • Lieberman D
      • Schlaeffer F
      • Boldur I
      • et al.
      Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
      ,
      • Bochud PY
      • Moser F
      • Erard P
      • et al.
      Community‐acquired pneumonia. A prospective outpatient study.
      ,
      • Marrie TJ
      • Peeling RW
      • Fine MJ
      • et al.
      Ambulatory patients with community‐acquired pneumonia: the frequency of atypical agents and clinical course.
      ,
      • Mandell LA
      • Bartlett JG
      • Dowell SF
      • et al.
      Update of practice guidelines for the management of community‐acquired pneumonia in immunocompetent adults.
      ,
      • Lonks JR
      • Garau J
      • Medeiros AA
      Implications of antimicrobial resistance in the empirical treatment of community‐acquired respiratory tract infections: the case of macrolides.
      ]. No general agreement currently exists concerning the selection of the antimicrobial regimen for all patient groups. UK guidelines for outpatients with non-severe pneumonia advocate initial therapy with amoxycillin (http://www.brit-thoracic.org/guidelines), despite the high frequency of atypical organisms [
      • Marston BJ
      • Plouffe JF
      • File TM
      • et al.
      Incidence of community‐acquired pneumonia requiring hospitalization. Results of a population‐based active surveillance study in Ohio. The Community‐Based Pneumonia Incidence Study Group.
      ,
      • Lieberman D
      • Schlaeffer F
      • Boldur I
      • et al.
      Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
      ,
      • Bochud PY
      • Moser F
      • Erard P
      • et al.
      Community‐acquired pneumonia. A prospective outpatient study.
      ,
      • Marrie TJ
      • Peeling RW
      • Fine MJ
      • et al.
      Ambulatory patients with community‐acquired pneumonia: the frequency of atypical agents and clinical course.
      ]. In contrast, North American guidelines still recommend monotherapy with macrolides for many outpatients [
      • Mandell LA
      • Bartlett JG
      • Dowell SF
      • et al.
      Update of practice guidelines for the management of community‐acquired pneumonia in immunocompetent adults.
      ], which is an approach that can be questioned in view of the increasing rates of resistance among pneumococci [
      • Lonks JR
      • Garau J
      • Medeiros AA
      Implications of antimicrobial resistance in the empirical treatment of community‐acquired respiratory tract infections: the case of macrolides.
      ]. In view of these uncertainties, clinicians may decide to provide empirical therapy targeted against both standard pathogens and atypical organisms for patients with CAP by prescribing combined therapy with a β-lactam and a macrolide, or monotherapy with a respiratory fluoroquinolone. Although initial empirical therapy for patients who require admission to hospital may require broad-spectrum coverage, overuse of antibiotics for all patients with CAP might lead to increasing drug resistance during the next few years [
      • Lonks JR
      • Garau J
      • Medeiros AA
      Implications of antimicrobial resistance in the empirical treatment of community‐acquired respiratory tract infections: the case of macrolides.
      ,
      • Ho PL
      • Tse WS
      • Tsang KW
      • et al.
      Risk factors for acquisition of levofloxacin‐resistant Streptococcus pneumoniae: a case‐control study.
      ].
      Atypical organisms are now considered to be an important cause of CAP, being implicated in 20–40% of CAP cases [
      • Marston BJ
      • Plouffe JF
      • File TM
      • et al.
      Incidence of community‐acquired pneumonia requiring hospitalization. Results of a population‐based active surveillance study in Ohio. The Community‐Based Pneumonia Incidence Study Group.
      ,
      • Lieberman D
      • Schlaeffer F
      • Boldur I
      • et al.
      Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
      ,
      • Bochud PY
      • Moser F
      • Erard P
      • et al.
      Community‐acquired pneumonia. A prospective outpatient study.
      ,
      • Marrie TJ
      • Peeling RW
      • Fine MJ
      • et al.
      Ambulatory patients with community‐acquired pneumonia: the frequency of atypical agents and clinical course.
      ]. Unfortunately, coverage of atypical pathogens remains empirical in most cases because of an absence of rapid, standardised diagnostic tests. Although molecular techniques, e.g., PCR with respiratory secretions, are promising [
      • Mandell LA
      • Bartlett JG
      • Dowell SF
      • et al.
      Update of practice guidelines for the management of community‐acquired pneumonia in immunocompetent adults.
      ], their accuracy and reproducibility have yet to be established, and no commercial assays are currently available for use by clinical microbiology laboratories. In this scenario, the availability of predictors of infection by atypical organisms would be of interest, since they could help determine the initial therapy for cases of CAP. However, epidemiological studies have shown that single clinical, radiological or laboratory parameters have limited value in predicting the microbial aetiology of CAP and in characterising atypical pneumonia [
      • Marrie TJ
      • Peeling RW
      • Fine MJ
      • et al.
      Ambulatory patients with community‐acquired pneumonia: the frequency of atypical agents and clinical course.
      ,
      • Farr BM
      • Kaiser DL
      • Harrison BD
      • Connolly CK
      Prediction of microbial aetiology at admission to hospital for pneumonia from the presenting clinical features. British Thoracic Society Pneumonia Research Subcommittee.
      ,
      • Schneeberger PM
      • Dorigo‐Zetsma JW
      • Van Der Zee A
      • Van Bon M
      • Van Opstal JL
      Diagnosis of atypical pathogens in patients hospitalized with community‐acquired respiratory infection.
      ,
      • Sopena N
      • Pedro‐Botet ML
      • Sabria M
      • et al.
      Comparative study of community‐acquired pneumonia caused by Streptococcus pneumoniae, Legionella pneumophila or Chlamydia pneumoniae.
      ].
      Rapid diagnostic tests for bacterial pneumonia, e.g., the pneumococcal urinary antigen test [
      • Murdoch DR
      • Laing RT
      • Mills GD
      • et al.
      Evaluation of a rapid immunochromatographic test for detection of Streptococcus pneumoniae antigen in urine samples from adults with community‐acquired pneumonia.
      ], have become available recently, and novel serum biomarkers of bacterial infection, e.g., procalcitonin (PCT), have been described [
      • Hedlund J
      • Hansson LO
      Procalcitonin and C‐reactive protein levels in community‐acquired pneumonia: correlation with etiology and prognosis.
      ,
      • 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.
      ]. The information provided by these tests could be useful as additional criteria for differentiating between atypical and classical bacterial aetiology in CAP. The present study describes a large prospective investigation of CAP in which the clinical features of the patients were recorded, and extensive laboratory investigations, including determinations of pneumococcal urinary antigen and serum biomarkers of bacterial infection, were performed [
      • Gutierrez F
      • Masia M
      • Rodriguez JC
      • et al.
      Epidemiology of community‐acquired pneumonia in adult patients at the dawn of the 21st century: a prospective study on the Mediterranean coast of Spain.
      ]. Three previous reports have evaluated the Binax immunochromatographic assay for detection of S. pneumoniae urinary antigen in the same patient cohort [
      • Gutiérrez F
      • Masiá M
      • Rodríguez JC
      • et al.
      Evaluation of the immunochromatographic Binax NOW assay for detection of Streptococcus pneumoniae urinary antigen in a prospective study of community‐acquired pneumonia in Spain.
      ], together with the usefulness of lipopolysaccharide-binding protein (LBP) [
      • Masiá M
      • Gutiérrez F
      • Llorca B
      • et al.
      Serum concentrations of lipopolysaccharide‐binding protein as a biochemical marker to differentiate microbial etiology in patients with community‐acquired pneumonia.
      ] and PCT [
      • Masiá M
      • Gutiérrez F
      • Shum C
      • et al.
      Usefulness of procalcitonin levels in community‐acquired pneumonia according to Patients Outcome Research Team Pneumonia Severity Index.
      ] as predictors of aetiology and prognosis. The objective of the present study was to characterise atypical pneumonia by combining laboratory data with the epidemiological and clinical features of the patients. In addition, a scoring system was devised to compare CAP caused by atypical pathogens with other causes of CAP in order to determine the variables that were most effective in discriminating atypical pathogens from other organisms.

      PATIENTS AND METHODS

       Setting and population studied

      This prospective study was conducted at Hospital General Universitario de Elche, a 430-bed university-affiliated teaching hospital serving a population of 250 000 in Alicante, a province on the Mediterranean coast of Spain. All adult patients (aged ≥15 years) with signs and symptoms compatible with pneumonia during two consecutive periods of 12 months (from 15 October 1999 to 14 October 2000, and from 15 October 2000 to 14 October 2001) were eligible for inclusion in the study. The study was approved by the local ethical committee, and informed consent was obtained from all the patients. CAP was defined as an acute illness associated with at least one of the following signs or symptoms: fever (measured by axillary temperature, which is common clinical practice in our centre); new cough, with or without sputum production; pleuritic chest pain; dyspnoea; and altered breath sound on auscultation, plus a chest radiograph showing an opacity compatible with the presence of acute pneumonia. Patients with a provisional diagnosis of CAP were seen within 48 h by a study investigator to confirm the diagnosis. Patients with previous hospitalisation within 2 weeks of the current diagnosis of pneumonia were excluded.
      Demographical and clinical data were collected using a written standardised questionnaire. Among the clinical data, air-conditioning exposure was defined as repetitive and prolonged (several hours a day) exposure at home or at work, and exposure to birds was defined as having birds at home or at work, or frequent contact with birds as a hobby. The severity of pneumonia was calculated using the Pneumonia Patient Outcome Research Team (PORT) severity index (PSI) [
      • Fine MJ
      • Auble TE
      • Yealy DM
      • et al.
      A prediction rule to identify low‐risk patients with community‐acquired pneumonia.
      ], which classifies patients, according to outcome, in five risk classes (class I includes patients with the most favourable prognosis, and class V those with the poorest prognosis). All patients were followed for at least 4 weeks or until death. A repeat chest radiograph and a blood sample were obtained 2–4 weeks after the initial diagnosis of CAP.

       Microbiological investigations

      The laboratory investigation for a patient with CAP has been described previously [
      • Gutierrez F
      • Masia M
      • Rodriguez JC
      • et al.
      Epidemiology of community‐acquired pneumonia in adult patients at the dawn of the 21st century: a prospective study on the Mediterranean coast of Spain.
      ]. In brief, it included obtaining sputum samples for Gram's stain and culture, two blood samples for culture, a urine sample for detection of Legionella pneumophila and S. pneumoniae antigens, and serum samples for detection of antibodies against atypical pathogens and viruses (taken during the acute stage of illness and at least 2 weeks later).
      A complement fixation test was performed to detect antibodies against M. pneumoniae, Chlamydophila spp., Coxiella burnetii, influenza viruses A and B, respiratory syncytial virus and adenovirus. An indirect immunofluorescence test was used to detect antibodies to L. pneumophila, and a microimmunofluorescence test was used to detect antibodies against Chlamydophila pneumoniae, Chlamydophila psittaci and Chlamydophila trachomatis.

       Criteria for aetiological diagnosis

      The following criteria were used to classify a pneumonia as being of known aetiology: (i) a four-fold or greater antibody rise for M. pneumoniae, Chl. psittaci, Cox. burnetti, influenza viruses A and B, parainfluenza virus, respiratory syncytial virus or adenovirus; (ii) a four-fold rise in microimmunofluorescence antibody titre to ≥1:128, or the presence of IgM antibodies (>1:20) for Chl. pneumoniae; (iii) isolation of the organism from respiratory samples, or Legionella antigen detected in urine, or a four-fold or greater rise in immunofluorescence antibody titre for L. pneumophila; (iv) isolation of the organism from blood or pleural fluid, or isolation as the predominant organism from a qualified sputum sample, or antigen detected in urine for S. pneumoniae; and (v) isolation of the organism from blood or pleural fluid, or isolation as the predominant organism from a qualified sputum sample for Haemophilus influenzae, Moraxella catarrhalis, Staphylococcus aureus and other bacteria, including Gram-negative enterobacteria. Cases that fulfilled the aetiological diagnostic criteria described above for more than one pathogen were considered to be mixed pneumonia, and included the following combinations of two or more pathogens: standard bacterial pathogens plus atypical organisms; atypical organisms plus viruses; two or more standard bacterial pathogens; two or more atypical organisms; and two or more viruses. Cases that did not fulfil the aetiological diagnostic criteria described above were considered to be pneumonia of unknown aetiology.

       Detection of LBP, C-reactive protein (CRP) and PCT

      During the first 12-month period (15 October 1999 to 14 October 2000), patients included in the study had a blood sample collected within 24 h of fulfilling the diagnostic criteria for pneumonia in order to measure LBP, CRP and PCT levels. Serum samples were collected and stored at -80°C until analysed in May 2003. LBP levels were measured using a commercially available assay (Immulite LBP; DPC, Los Angeles, CA, USA) according to the manufacturer's guidelines, using an Immulite immunoanalyser (DPC) with a limit of detection of 0.2 mg/L [
      • Babson AL
      • Olson DR
      • Palmieri T
      • et al.
      The IMMULITE assay tube: a new approach to heterogeneous ligand assay.
      ]. CRP was measured using a Behring nephelometer II (Behring Co., Marburg, Germany), with a limit of detection of 3.67 mg/L [
      • Ledue TB
      • Weiner DL
      • Sipe JD
      • et al.
      Analytical evaluation of particle‐enhanced immunonephelometric assays for C‐reactive protein, serum amyloid A and mannose‐binding protein in human serum.
      ]. PCT levels were measured with a monoclonal immunoluminometric assay (Liaison Brahms PCT; Brahms Diagnostica, Berlin, Germany), with a limit of detection of 0.1 μg/L [
      • Meisner M
      • Tschaikowsky K
      • Schnabel S
      • et al.
      Procalcitonin‐influence of temperature, storage, anticoagulation and arterial or venous asservation of blood samples on procalcitonin concentrations.
      ]. All biochemical tests were performed in a blinded fashion, without knowing the results of other microbiological investigations.

       Data analysis

      Statistical analysis was performed using SPSS v. 11.5 (SPSS Inc., Chicago, IL, USA). For the purposes of the calculations, patients were classified into three groups: (i) patients with pneumonia caused by atypical pathogens, including three patients with mixed pneumonia caused by two atypical organisms; (ii) patients with pneumonia caused by standard pathogens; and (iii) patients with pneumonia of unknown aetiology. Clinical and laboratory data were compared as dichotomous variables using the chi-square test, with p <0.05 considered as significant. Variables found to be statistically significant in the univariate analysis were included in a multivariate stepwise logistic regression model. Two logistic regression models were developed, the first comparing atypical with other microbial causes of pneumonia, and the second comparing atypical pneumonia with pneumonia of unknown aetiology. A scoring system was then constructed that included four significant variables from the multivariate analyses, which were allocated one point each. Different combinations were assessed, until the best sensitivity and specificity for the scoring system were obtained. Two separate models were assessed, with the first comprising all 493 patients enrolled during the two study periods, and the second including only the 240 patients enrolled during the first 12-month study period, for whom serum biomarkers of bacterial infection were available. Receiver operating characteristic curves were used to describe the sensitivity and specificity of the classification of the aetiology of pneumonia for each scoring system.

      RESULTS

      Of 516 patients with signs and symptoms compatible with pneumonia, 23 were found subsequently not to have CAP (eight had lower respiratory tract infection without pneumonia, seven had heart failure, two had pulmonary embolism, two had lung cancer, one had bronchiectasis, one had atelectasis, one had pulmonary fibrosis, and one had pulmonary haemorrhage), leaving 493 patients in the study cohort. The first 12-month period (15 October 1999 to 14 October 2000) included 240 patients, and the second 12-month period (15 October 2000 to 14 October 2001) included 253 patients. Baseline characteristics of the patients, listed according to aetiological group, are summarised in Table 1.
      TABLE 1Demographical and clinical characteristics of patients with community-acquired pneumonia (n = 493), listed according to aetiological group
      CharacteristicAtypical (n = 94
      Three patients with mixed pneumonia caused by two atypical organisms were also included.
      ) pathogens
      Standard bacterial pathogens (n = 111)Other pathogens
      Mixed pneumonia (n = 25) + viral pneumonia (n = 20).
      (n = 45)
      Unknown aetiology (n = 243)
      Male, no. patients (%)63 (67)70 (63.1)26 (57.8)149 (61.3)
      Age (years), mean (range)46 (15–86)59.1 (15–90)58.8 (15–93)59.1 (15–94)
      PSI score, median (IQR)48 (23.5–72.3)79 (44–102.3)78 (42–101)69 (42–90)
       Risk classes I–II70 (74.5)50 (45)20 (44.4)128 (52.6)
       Risk class III11 (11.7)24 (21.6)9 (20)59 (24.3)
       Risk classes IV–V13 (13.8)36 (32.4)16 (35.5)56 (23)
      Co-morbidity, no. patients (%)25 (26.6)65 (58.6)28 (62.2)109 (44.9)
       Diabetes mellitus17 (18.1)26 (23.4)7 (15.6)48 (19.8)
       COPD10 (10.4)29 (26.1)10 (22.2)50 (20.6)
       Other
      Congestive heart failure, chronic liver disease, neoplasia, immunosuppression, altered mental status.
      8 (8.5)50 (45)16 (35.6)48 (19.8)
      Risk-factors, no. patients (%)
       Smoking23 (24.5)24 (21.6)17 (37.8)47 (19.3)
       Alcohol10 (10.6)17 (15.3)7 (15.6)29 (11.9)
       Exposure to birds35 (36.5)20 (18)7 (15.6)61 (25.1)
       Air-conditioning exposure
      Repetitive exposure at work or at home.
      14 (14.6)5 (4.5)018 (7.4)
      Hospital admission, no. patients (%)56 (59.6)87 (78.4)36 (80)182 (74.9)
      Previous antibiotic treatment, no. patients (%)29 (30.9)16 (14.4)8 (17.8)61 (25.1)
      Clinical presentation and laboratory data, no. patients (%)
       Dyspnoea, moderate or severe23 (24.5)47 (42.3)22 (48.9)95 (39.1)
       Pleuritic chest pain, moderate or severe21 (22.4)51 (45.9)21 (46.6)99 (40.7)
       Headache, moderate or severe19 (20.2)12 (10.8)6 (13.3)24 (9.9)
       Diarrhoea12 (12.8)4 (3.6)3 (6.6)16 (6.6)
       Confusion5 (5.3)17 (15.3)12 (26.7)24 (9.9)
       Purulent sputum production35 (37.2)70 (63)34 (75.5)121 (49.8)
       WBC/mm38200 (6575–12 100)13 550 (8963–17 930)12 900 (9785–16 500)11 600 (8208–16 000)
       Neutrophil count/mm36610 (4700–10 150)11 645 (7575–15 673)11 200 (7835–14 550)9270 (6000–13 455)
       LBP
      Data were available for 196 patients.
      serum levels (mg/L)
      9.2 (6.9–13.6)20.2 (10.7–32.7)19.8 (11.9–36.8)14.7 (9.6–31)
       CRP
      Data were available for 196 patients.
      serum levels (mg/L)
      3.67 (3.67–12.1)13.4 (4.4–115.5)40 (9.8–122)12 (3.7–80.7)
       Procalcitonin
      Data were available for 185 patients.
      serum levels (μg/L)
      0.10 (0.10–0.11)0.10 (0.10–0.31)0.10 (0.10–0.16)0.10 (0.10–0.15)
      Antibiotic treatment prescribed, no. patients (%)
       β-Lactam alone15 (16)27 (24.3)17 (37.8)49 (20.2)
       Macrolide alone12 (12.8)8 (7.2)026 (10.7)
       Fluoroquinolone
      Levofloxacin or moxifloxacin.
      22 (23.4)18 (16.2)4 (8.9)41 (16.9)
       Combined therapy
      β-Lactam plus a macrolide.
      40 (42.6)48 (43.2)17 (37.8)115 (47.3)
       Other/unknown4 (4.3)8 (7.2)7 (15.6)12 (4.9)
      Outcome, no. patients (%)
       Pleural effusion5 (5.3)19 (17.1)8 (17.8)27 (11.1)
       Atelectasis5 (5.3)9 (8.1)1 (2.2)14 (5.8)
       Mechanical ventilation1 (1.1)2 (1.8)3 (6.7)3 (1.2)
       Death08 (7.2)3 (6.7)13 (5.3)
      Data for continuous variables are median values (inter-quartile range), except for age.
      PSI, pneumonia PORT (patient outcome research team) severity index; IQR, inter-quartile range; COPD, chronic obstructive pulmonary disease; WBC, white blood cell count; LBP, lipopolysaccharide-binding protein; CRP, C-reactive protein.
      a Three patients with mixed pneumonia caused by two atypical organisms were also included.
      b Mixed pneumonia (n = 25) + viral pneumonia (n = 20).
      c Congestive heart failure, chronic liver disease, neoplasia, immunosuppression, altered mental status.
      d Repetitive exposure at work or at home.
      e Data were available for 196 patients.
      f Data were available for 185 patients.
      g Levofloxacin or moxifloxacin.
      h β-Lactam plus a macrolide.

       Aetiological agents identified

      Aetiological agents were found for 250 (50.7%) patients, including 140 bacterial, 110 atypical and 30 viral pathogens. In 243 (49.3%) cases, the aetiology remained unknown after microbiological investigation; in these patients, the clinical diagnosis of pneumonia was confirmed by subsequent follow-up evaluations. Table 2 shows the aetiological organisms identified.
      TABLE 2Microorganisms identified in patients (n = 493) with community-acquired pneumonia
      Patient groupNo. of patients (%)
      Bacterial pneumonia111 (22.5)
      Streptococcus pneumoniae83 (16.8)
      Pseudomonas spp.11 (2.2)
      Haemophilus spp.9 (1.8)
       Gram-negative bacilli other than Pseudomonas spp.
      Klebsiella spp. (two cases), Escherichia coli (one case), Citrobacter spp. (one case), Stenotrophomonas spp. (one case).
      5 (1.0)
      Staphylococcus aureus2 (0.4)
      Moraxella catarrhalis1 (0.2)
      Pneumonia caused by atypical pathogens91 (18.5)
      Mycoplasma pneumoniae38 (7.7)
      Legionella pneumophila21 (4.3)
      Chlamydophila spp.
      Chlamydophila pneumoniae (15 cases), Chlamydophila psittaci (nine cases). In six cases, there was a four-fold rise in antibodies against both Chl. psittaci (complement fixation test) and Chl. pneumoniae (microimmunofluorescence test).
      30 (6.1)
      Coxiella burnetii2 (0.4)
      Viral pneumonia20 (4.1)
       Influenza virus14 (2.8)
       Viruses other than influenza6 (1.2)
      Mixed pneumonia
      Streptococcus pneumoniae and Legionella pneumophila (three cases), Strep. pneumoniae and Pseudomonas spp. (three cases), Strep. pneumoniae and Mycoplasma pneumoniae (two cases), Strep. pneumoniae and influenza virus (two cases), M. pneumoniae and influenza virus (two cases), L. pneumophila and Chl. pneumoniae (two cases), Strep. pneumoniae and Haemophilus spp. (one case), Strep. pneumoniae and Staphyloccus aureus (one case), Strep. pneumoniae and Klebsiella spp. (one case), Strep. pneumoniae and Enterobacter spp. (one case), Strep. pneumoniae and Moraxella catarrhalis (one case), Strep. pneumoniae and Coxiella burnetii (one case), M. pneumoniae and Haemophilus influenzae (one case), Chl. psittaci and Listeria monocytogenes (one case), Chl. pneumoniae and Cox. burnetti (one case), L. pneumophila and influenza virus (one case), influenza virus and varicella-zoster virus (one case), influenza virus and respiratory syncytial virus (one case), Strep. pneumoniae, M. pneumoniae and influenza virus (one case), M. pneumoniae, Strep. pneumoniae and Staph. aureus (one case).
      28 (5.7)
      Pneumonia, with no pathogen identified243 (49.3)
      Total493
      a Klebsiella spp. (two cases), Escherichia coli (one case), Citrobacter spp. (one case), Stenotrophomonas spp. (one case).
      b Chlamydophila pneumoniae (15 cases), Chlamydophila psittaci (nine cases). In six cases, there was a four-fold rise in antibodies against both Chl. psittaci (complement fixation test) and Chl. pneumoniae (microimmunofluorescence test).
      c Streptococcus pneumoniae and Legionella pneumophila (three cases), Strep. pneumoniae and Pseudomonas spp. (three cases), Strep. pneumoniae and Mycoplasma pneumoniae (two cases), Strep. pneumoniae and influenza virus (two cases), M. pneumoniae and influenza virus (two cases), L. pneumophila and Chl. pneumoniae (two cases), Strep. pneumoniae and Haemophilus spp. (one case), Strep. pneumoniae and Staphyloccus aureus (one case), Strep. pneumoniae and Klebsiella spp. (one case), Strep. pneumoniae and Enterobacter spp. (one case), Strep. pneumoniae and Moraxella catarrhalis (one case), Strep. pneumoniae and Coxiella burnetii (one case), M. pneumoniae and Haemophilus influenzae (one case), Chl. psittaci and Listeria monocytogenes (one case), Chl. pneumoniae and Cox. burnetti (one case), L. pneumophila and influenza virus (one case), influenza virus and varicella-zoster virus (one case), influenza virus and respiratory syncytial virus (one case), Strep. pneumoniae, M. pneumoniae and influenza virus (one case), M. pneumoniae, Strep. pneumoniae and Staph. aureus (one case).

       Analysis group

      Of 95 patients with pneumonia caused by atypical organisms, one was excluded because of an indeterminate pneumococcal urinary antigen, leaving 94 patients in the analysis group, of whom three had mixed infections caused by two atypical organisms. Forty-seven of these patients were enrolled during the first period of the study. The atypical pathogens identified are listed in Table 2, and the demographical and clinical characteristics of the patients are summarised in Table 1. No patient with atypical pneumonia died within the 4-week follow-up period.
      Since pneumonia caused by Legionella spp. might have a distinctive clinical presentation, clinical and laboratory variables of patients with CAP caused by Legionella spp. were compared with those of patients with other causes of atypical CAP. Patients with Legionella pneumonia were more likely to have chronic obstructive pulmonary disease (p 0.01), exposure to air-conditioning (p <0.001) and elevated aspartate aminotransferase (AST) levels (p 0.03), and were less likely to belong to low PSI risk classes (I–II) (p 0.05). One case of Legionella pneumonia was a severe illness that required mechanical ventilation, but had a favourable outcome.
      Significant variables associated with CAP caused by atypical pathogens, compared with variables associated with other microbial causes of pneumonia and pneumonia of unknown aetiology, by univariate analysis are shown in Table 3. PCT serum levels were determined for 185 patients. In patients with atypical CAP, mean (SD) and median values were 0.30 μg/L (0.88) and 0.1 μg/L, respectively, whereas mean and median values for patients with CAP caused by other pathogens were 0.55 μg/L (1.50) and 0.1 μg/L, respectively (p 0.3).
      TABLE 3Univariate analysis of variables associated with community-acquired pneumonia (CAP) caused by atypical vs. standard pathogens (n = 250), and atypical pneumonia vs. pneumonia of unknown aetiology (n = 337)
      Clinical featureAtypical pathogens (n = 94)
      Three patients with mixed pneumonia caused by two or more atypical organisms were also included.
      Standard pathogens (n = 156)OR (95% CI)
      Comparison between patients with CAP caused by atypical vs. standard pathogens.
      p
      Comparison between patients with CAP caused by atypical vs. standard pathogens.
      Unknown aetiology (n = 243)OR (95% CI)
      Comparison between patients with CAP caused by atypical pathogens vs. CAP of unknown aetiology.
      p
      Comparison between patients with CAP caused by atypical pathogens vs. CAP of unknown aetiology.
      Risk-factor
       Age <65 years72/94 (76.6)75/156 (48.1)3.54 (2.0–6.26)<0.001126/243 (51.9)3.21 (1.85–5.52)<0.001
       COPD10/94 (10.6)39/155 (25.2)0.35 (0.17–0.75)0.00550/243 (20.6)0.46 (0.22–0.95)0.03
       Aspiration0/92 (0)10/152 (6.6)0.79 (0.76–0.83)0.00416/235 (6.8)0.15 (0.02–1.15)0.04
       Immunosuppression1/94 (1.1)17/155 (11)0.09 (0.01–0.67)0.0028/243 (3.3)0.32 (0.04–2.56)0.4
       No co-morbidity70/94 (74.5)75/155 (48.1)3.11 (1.78–5.45)<0.001145/243 (59.7)1.97 (1.16–3.35)0.01
       Exposure to air-conditioning
      Repetitive exposure at work or at home.
      14/92 (15.2)5/152 (3.3)5.26 (1.83–15.2)0.00118/235 (7.7)2.16 (1.03–4.57)0.06
       Exposure to birds34/92 (37)27/152 (17.8)2.72 (1.50–4.90)0.00161/235 (26)1.67 (1–2.79)0.05
      Signs/symptoms
       Fever
      Axillary body temperature ≥ 38°C.
      58/94 (61.7)69/155 (44.5)2.01 (1.19–3.39)0.009112/243 (46.1)1.88 (1.16–3.07)0.01
       Tachypnoea
      Respiratory frequency >18 respirations/min.
      21/94 (22.3)66/155 (42.6)0.41 (0.23–0.73)0.00399/243 (40.7)0.44 (0.26–0.76)0.003
       Headache39/94 (41.5)32/156 (20.5)2.75 (1.56–4.84)<0.00156/243 (23)2.37 (1.43–3.93)0.001
       Absence of purulent sputum59/94 (62.8)51/155 (32.9)3.4 (2.01–5.87)< 0.001121/242 (50)1.69 (1.04–2.75)0.03
      Laboratory findings
       Leukocytosis
      White blood cell count >10 × 109 cells/L.
      34/94 (36.2)115/156 (73.7)0.20 (0.12–0.35)<0.001153/243 (63)0.33 (0.20–0.55)<0.001
       Neutrophilia
      Neutrophil count >5 × 109 cells/L.
      64/93 (68.8)137/155 (88.4)0.29 (0.15–0.56)<0.001197/242 (81.4)0.5 (0.29–0.87)0.01
       AST >35 U/L27/93 (29.0)26/156 (16.7)2.05 (1.11–3.78)0.0234/243 (14)2.51 (1.41–4.74)0.002
       LBP <14 mg/L31/41 (75.6)24/70 (34.3)5.94 (2.40–14.14)<0.00139/85 (45.9)3.66 (1.59–8.39)0.002
       CRP <50 mg/L35/41 (85.4)40/68 (58.8)4.08 (1.51–11.01)0.00558/84 (69)2.62 (0.98–6.98)0.05
      PSI score <70 (I/II)69/94 (73.4)67/155 (43.2)3.63 (2.10–6.33)<0.001124/243 (51)2.65 (1.57–4.47)<0.001
      Data are number (%) of patients. All variables are dichotomous, and their presence is compared with their absence. An OR >1 indicates an association with CAP caused by atypical pathogens, and an OR <1 indicates an association with other causes of CAP.
      COPD, chronic obstructive pulmonary disease; AST, aspartate aminotransferase; LBP, lipopolysaccharide-binding protein; CRP, C-reactive protein; PSI, pneumonia PORT (patient outcome research team) severity index.
      a Three patients with mixed pneumonia caused by two or more atypical organisms were also included.
      b Comparison between patients with CAP caused by atypical vs. standard pathogens.
      c Comparison between patients with CAP caused by atypical pathogens vs. CAP of unknown aetiology.
      d Repetitive exposure at work or at home.
      e Axillary body temperature ≥ 38°C.
      f Respiratory frequency >18 respirations/min.
      g White blood cell count >10 × 109 cells/L.
      h Neutrophil count >5 × 109 cells/L.
      Multivariate analysis in which the 94 patients with CAP caused by atypical pathogens were compared with the 156 patients with other microbial causes of CAP revealed that patients with pneumonia caused by atypical pathogens were more likely to have repeated exposure to air-conditioning (OR 9.09), normal white blood cell (WBC) counts (OR 7.57), exposure to birds (OR 3.73), an absence of purulent sputum (OR 3.47), and an absence of co-morbidity (OR 3.10) (Table 4). Multivariate analysis in which patients with atypical CAP were compared with patients with CAP of unknown aetiology revealed that patients with atypical CAP were more likely to have elevated AST levels (OR 2.65), to be aged <65 years (OR 2.52) and to have normal WBC counts (OR 2.36), but were less likely to have tachypnoea (OR 0.52).
      TABLE 4Multivariate analysis comparing clinical and laboratory features of patients with community-acquired pneumonia caused by atypical
      Three patients with mixed pneumonia caused by two atypical organisms were also included.
      vs. standard pathogens
      Comparison between patients with pneumonia caused by atypical pathogens vs. other microbial causes of pneumonia.
      , and atypical pneumonia vs. pneumonia of unknown aetiology
      VariableOR
      Comparison between patients with pneumonia caused by atypical pathogens vs. other microbial causes of pneumonia.
      95% CI
      Comparison between patients with pneumonia caused by atypical pathogens vs. other microbial causes of pneumonia.
      p
      Comparison between patients with pneumonia caused by atypical pathogens vs. other microbial causes of pneumonia.
      OR
      Comparison between patients with pneumonia caused by atypical pathogens vs. pneumonia of unknown aetiology.
      95% CI
      Comparison between patients with pneumonia caused by atypical pathogens vs. pneumonia of unknown aetiology.
      p
      Comparison between patients with pneumonia caused by atypical pathogens vs. pneumonia of unknown aetiology.
      Exposure to air-conditioning
      Repetitive exposure at work or at home.
      9.092.56–32.250.001
      WBC count <10 000/mm37.573.81–15.01<0.001
      Exposure to birds3.731.75–7.950.001
      Absence of purulent sputum3.471.79–6.76<0.001
      Absence of co-morbidity3.101.57–6.100.001
      AST >35 U/L2.651.40–5.00.003
      Age <65 years2.521.37–4.640.003
      Tachypnoea
      Respiratory frequency >18 respirations/min.
      0.520.28–0.960.03
      WBC, white blood cell count; AST, aspartate aminotransferase.
      a Three patients with mixed pneumonia caused by two atypical organisms were also included.
      b Comparison between patients with pneumonia caused by atypical pathogens vs. other microbial causes of pneumonia.
      c Comparison between patients with pneumonia caused by atypical pathogens vs. pneumonia of unknown aetiology.
      d Repetitive exposure at work or at home.
      e Respiratory frequency >18 respirations/min.
      When only the 240 patients from the first period of the study were included in the analysis, and LBP and CRP serum levels were added to the multivariable logistic regression model, multivariate analysis revealed that patients with pneumonia caused by atypical pathogens were more likely than patients with standard pathogens to have repeated exposure to air-conditioning (OR 15.32), exposure to birds (OR 8.31), a PSI score <70 (OR 8.33), an LBP value <14 mg/L (OR 7.47), an absence of purulent sputum (OR 6.17) and elevated AST levels (OR 4.24) (Table 5). When patients with pneumonia caused by atypical pathogens were compared with patients with pneumonia of unknown aetiology, multivariate analysis showed that patients with atypical pneumonia were more likely to be aged <65 years (OR 6.59), to have an LBP value <14 mg/L (OR 3.94) and elevated AST levels (OR 3.32).
      TABLE 5Multivariate analysis of the clinical and laboratory features of 240 patients with community-acquired pneumonia caused by atypical
      Three patients with mixed pneumonia caused by two atypical organisms were also included.
      pathogens, compared with patients with pneumonia caused by standard pathogens
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with other microbial causes of pneumonia.
      , and with patients with pneumonia of unknown aetiology
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with pneumonia of unknown aetiology.
      VariableOR
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with other microbial causes of pneumonia.
      95% CI
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with other microbial causes of pneumonia.
      p
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with other microbial causes of pneumonia.
      OR
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with pneumonia of unknown aetiology.
      95% CI
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with pneumonia of unknown aetiology.
      p
      Comparison between patients with pneumonia caused by atypical pathogens vs. patients with pneumonia of unknown aetiology.
      Exposure to air- conditioning
      Repeated exposure at work or at home.
      15.322.02–116.030.008
      PSI score <70 (risk classes I–II)8.332.43–28.570.001
      Exposure to birds8.312.14–32.370.002
      LBP <14 mg/L7.472.32–24.090.0013.941.55–10.020.004
      Absence of purulent sputum6.171.95–19.610.002
      AST >35 U/L4.241.08–16.670.033.321.13–9.800.02
      Age <65 years6.592.52–17.25<0.001
      PSI, pneumonia PORT (patient outcome research team) severity index; LBP, lipopolysaccharide-binding protein; AST, aspartate aminotransferase.
      a Three patients with mixed pneumonia caused by two atypical organisms were also included.
      b Comparison between patients with pneumonia caused by atypical pathogens vs. patients with other microbial causes of pneumonia.
      c Comparison between patients with pneumonia caused by atypical pathogens vs. patients with pneumonia of unknown aetiology.
      d Repeated exposure at work or at home.

       Predictive modelling

      A scoring system was devised that included four of the significant variables revealed by the two multivariate analyses. When age <65 years, a normal WBC count, exposure to birds, and an elevated AST level were included, and the entire population of 493 patients was analysed, a score of ≥3 captured 32 of 91 patients with pneumonia caused by atypical organisms (35.2% sensitivity), and only 27 of 386 patients with other causes of pneumonia (93% specificity). The area under the curve was 0.73 (95% CI 0.67–0.79), and 477 (96.7%) patients were classified.
      A second scoring system was used to analyse only the 240 patients from the first period of the study. An additional variable from the univariate analysis was incorporated, as it improved the sensitivity and specificity of the model. When age <65 years, LBP levels <14 mg/L, exposure to birds and a normal WBC count were included, a score of ≥3 captured 20 of 41 patients with pneumonia caused by atypical pathogens (48.8% sensitivity), and only 14 of 155 patients with other causes of pneumonia (91% specificity). The area under the curve was 0.78 (95% CI 0.70–0.86), and 196 (81.7%) of the patients were classified.
      According to the PSI, the median punctuation of patients with a score suggestive of atypical aetiology (≥3) was 46. In patients with a score <3, suggestive of standard aetiology, the median PSI punctuation was 72 (p <0.001). Likewise, 83.1% of patients with a score suggestive of atypical aetiology had a PSI of <70 (risk classes I–II), compared with only 48.1% of patients with a score suggestive of non-standard aetiology (p <0.001).

      DISCUSSION

      This study identified several clinical and biological predictors of atypical pneumonia. Although atypical pneumonia was difficult to differentiate from other causes of pneumonia on clinical grounds, a combination of epidemiological, clinical and laboratory data greatly improved the ability to predict CAP caused by atypical pathogens. An age <65 years, exposure to birds, a normal WBC count, an elevated AST level and an LBP serum level of <14 mg/L were the best variables to include in a predictive index for atypical pneumonia. In addition, as all the patients classified as having atypical CAP were negative with the pneumococcal urinary antigen assay, this test was also incorporated into the predictive models. Using these variables, two highly specific clinical indices for predicting atypical aetiology in patients with CAP were developed. Although the sensitivity was low, up to half of the candidates for receiving targeted therapy with macrolides could be identified, thereby avoiding unnecessary use of β-lactams or fluoroquinolones and decreasing the risks of antibiotic resistance and side-effects, as well as healthcare-related costs. Since the incidence of CAP in developed countries is 11–12 cases/1000 individuals/year [
      • Marrie TJ
      Community‐acquired pneumonia.
      ], and atypical organisms may account for up to 50% of all cases of CAP [
      • Lieberman D
      • Schlaeffer F
      • Boldur I
      • et al.
      Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
      ], this proportion represents a significant number of patients.
      Interestingly, repeated exposure to air-conditioning was one of the independent variables associated with atypical aetiology. This association has not been reported previously. When the clinical characteristics of the patients with pneumonia caused by Legionella spp. were analysed individually, exposure to air-conditioning was associated strongly with infection caused by this microorganism (42.9% of exposed patients had Legionella pneumonia vs. 7% of exposed patients with pneumonia caused by other atypical pathogens; data not shown). Although it is well-known that cooling towers constitute a source of Legionnaires’ disease in the community, domestic air-conditioning exposure has not been identified previously as a risk-factor. In southern European countries, exposure to air-conditioning is common at work or at home, especially during the summer months, and most of the patients exposed had acquired pneumonia between May and October (data not shown). The higher frequency of exposure to air-conditioning among patients with atypical CAP might therefore reflect the influence of the group of patients with Legionella pneumonia. In addition, exposure to birds was one of the risk-factors associated with atypical pneumonia. As expected, patients with pneumonia caused by Chl. psittaci constituted the majority of patients exposed to birds among the sample of patients with atypical CAP (69.2% vs. 31.6%, respectively, p 0.01; data not shown).
      The results of the present study revealed that 36% of all diagnosed cases of CAP were caused by atypical pathogens, in line with the figures reported by other investigators [
      • Marston BJ
      • Plouffe JF
      • File TM
      • et al.
      Incidence of community‐acquired pneumonia requiring hospitalization. Results of a population‐based active surveillance study in Ohio. The Community‐Based Pneumonia Incidence Study Group.
      ,
      • Lieberman D
      • Schlaeffer F
      • Boldur I
      • et al.
      Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
      ,
      • Bochud PY
      • Moser F
      • Erard P
      • et al.
      Community‐acquired pneumonia. A prospective outpatient study.
      ,
      • Marrie TJ
      • Peeling RW
      • Fine MJ
      • et al.
      Ambulatory patients with community‐acquired pneumonia: the frequency of atypical agents and clinical course.
      ]. It is of note that 23% of patients with atypical pneumonia were aged >65 years, and 25% had an associated co-morbid condition. Atypical pneumonia is usually considered to be more frequent among younger patients without co-morbidity, but significant incidence rates have also been described in the elderly [
      • Riquelme R
      • Torres A
      • El‐Ebiary M
      • et al.
      Community‐acquired pneumonia in the elderly: a multivariate analysis of risk and prognostic factors.
      ,
      • Alvarez Gutierrez FJ
      • Garcia Fernandez A
      • Elias Hernandez T
      • et al.
      Community acquired pneumonia in patients older than 60 years. Incidence of atypical agents and clinical–radiological progression.
      ] and in patients with comorbidity [
      • Falguera M
      • Sacristan O
      • Nogues A
      • et al.
      Non‐severe community‐acquired pneumonia: correlation between cause and severity or comorbidity.
      ]. Likewise, although atypical pathogens, other than Legionella spp., have usually been associated with mild-to-moderate illness, severe cases of CAP caused by atypical organisms, other than Legionella spp., that require admission to an intensive care unit have been reported [
      • Ruiz M
      • Ewig S
      • Torres A
      • et al.
      Severe community‐acquired pneumonia. Risk factors and follow‐up epidemiology.
      ]. In the present study, patients with atypical pneumonia had a disease of low severity, as assessed by the PSI, and all had a favourable outcome.
      Among the atypical pathogens, M. pneumoniae, Chl. pneumoniae and L. pneumophila caused the majority of cases. Although M. pneumoniae has long been considered to be the main causative atypical organism [
      • Lieberman D
      • Schlaeffer F
      • Boldur I
      • et al.
      Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
      ,
      • Bochud PY
      • Moser F
      • Erard P
      • et al.
      Community‐acquired pneumonia. A prospective outpatient study.
      ], Chl. pneumoniae has reached a proportion comparable to, or higher, than that of Mycoplasma in some studies [
      • Almirall J
      • Bolibar I
      • Vidal J
      • et al.
      Epidemiology of community‐acquired pneumonia in adults: a population‐based study.
      ], whereas in other cohorts it has been observed much more rarely [
      • Bochud PY
      • Moser F
      • Erard P
      • et al.
      Community‐acquired pneumonia. A prospective outpatient study.
      ]. The exact significance of Chl. pneumoniae remains to be elucidated, although the lack of a reference standard for the diagnosis may account, in part, for the wide variation in the reported incidence rates. In the present cohort, using the currently recommended criteria to diagnose Chl. pneumoniae infection [
      • Mandell LA
      • Bartlett JG
      • Dowell SF
      • et al.
      Update of practice guidelines for the management of community‐acquired pneumonia in immunocompetent adults.
      ], Chl. pneumoniae was found to be a significant pathogen in CAP, but its frequency was lower than that of M. pneumoniae, and much lower than that reported in other studies of pneumonia [
      • Marrie TJ
      • Peeling RW
      • Fine MJ
      • et al.
      Ambulatory patients with community‐acquired pneumonia: the frequency of atypical agents and clinical course.
      ,
      • Almirall J
      • Bolibar I
      • Vidal J
      • et al.
      Epidemiology of community‐acquired pneumonia in adults: a population‐based study.
      ].
      Increasing reports of mixed organisms causing pneumonia [
      • Lieberman D
      • Schlaeffer F
      • Boldur I
      • et al.
      Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
      ,
      • Gleason PP
      The emerging role of atypical pathogens in community‐acquired pneumonia.
      ,
      • Jokinen C
      • Heiskanen L
      • Juvonen H
      • et al.
      Microbial etiology of community‐acquired pneumonia in the adult population of 4 municipalities in eastern Finland.
      ], together with recent retrospective studies suggesting that dual antibiotic therapy may reduce mortality in patients with CAP admitted to hospital [
      • Brown RB
      • Iannini P
      • Gross P
      • Kunkel M
      Impact of initial antibiotic choice on clinical outcomes in community‐acquired pneumonia: analysis of a hospital claims‐made database.
      ,
      • Gleason PP
      • Meehan TP
      • Fine JM
      • Galusha DH
      • Fine MJ
      Associations between initial antimicrobial therapy and medical outcomes for hospitalized elderly patients with pneumonia.
      ], could cast doubt on the wisdom of monotherapy with macrolides. The present analysis only investigated patients with monomicrobial pneumonia caused by atypical pathogens, and excluded cases of mixed infection. Nevertheless, despite the extensive microbiological investigations performed in the study, an associated bacterial infection could have been missed in some patients classified as having pneumonia caused by atypical organisms alone. Additionally, the use of alternative diagnostic techniques, e.g., PCR or viral cultures from respiratory secretions, or serological tests other than complement fixation, might have contributed to a decrease in the high proportion of patients without a final microbial diagnosis. This may have led to greater accuracy, as patients with pneumonia of unknown aetiology represent a miscellaneous aetiological group for which interpretation of data is uncertain.
      The study also had other limitations. Although the predictors were generated from a heterogeneous study population with a broad clinical and microbial spectrum of CAP, some subgroups may not have been represented sufficiently; thus, the predictors may not be applicable to all patient populations. Some of the biomarkers, e.g., LBP, currently have limited practical usefulness; LBP determination is a time-consuming test, and experience in its interpretation in the field of pneumonia is very limited. Finally, the clinical and biological predictors of atypical pneumonia, and the estimation of the sensitivity and specificity of their combination, were determined using the same sample of patients. Therefore, the current predictive models should be treated as developmental rather than definitive.
      In summary, several clinical and epidemiological features appear to characterise infections caused by atypical pathogens, but very few are useful for predicting atypical CAP. Indeed, some of these features are linked more frequently with particular atypical microorganisms. However, the combination of selected patient characteristics with additional diagnostic tests could help clinicians to discriminate CAP caused by atypical pathogens with a high specificity. The study confirmed that microbial diagnosis of CAP remains problematic. New microbiological techniques are needed urgently to improve the detection of microorganisms causing CAP. Meanwhile, the presumptive identification of patients with pneumonia caused by atypical pathogens might help to optimise initial therapy for CAP by allowing tailored antimicrobial coverage with macrolides.

      References

        • Marston BJ
        • Plouffe JF
        • File TM
        • et al.
        Incidence of community‐acquired pneumonia requiring hospitalization. Results of a population‐based active surveillance study in Ohio. The Community‐Based Pneumonia Incidence Study Group.
        Arch Intern Med. 1997; 157: 1709-1718
        • Lieberman D
        • Schlaeffer F
        • Boldur I
        • et al.
        Multiple pathogens in adult patients admitted with community‐acquired pneumonia: a one year prospective study of 346 consecutive patients.
        Thorax. 1996; 51: 179-184
        • Bochud PY
        • Moser F
        • Erard P
        • et al.
        Community‐acquired pneumonia. A prospective outpatient study.
        Medicine (Baltimore). 2001; 80: 75-87
        • Marrie TJ
        • Peeling RW
        • Fine MJ
        • et al.
        Ambulatory patients with community‐acquired pneumonia: the frequency of atypical agents and clinical course.
        Am J Med. 1996; 101: 508-515
        • Mandell LA
        • Bartlett JG
        • Dowell SF
        • et al.
        Update of practice guidelines for the management of community‐acquired pneumonia in immunocompetent adults.
        Clin Infect Dis. 2003; 37: 1405-1433
        • Lonks JR
        • Garau J
        • Medeiros AA
        Implications of antimicrobial resistance in the empirical treatment of community‐acquired respiratory tract infections: the case of macrolides.
        J Antimicrob Chemother. 2002; 2: 87-92
        • Ho PL
        • Tse WS
        • Tsang KW
        • et al.
        Risk factors for acquisition of levofloxacin‐resistant Streptococcus pneumoniae: a case‐control study.
        Clin Infect Dis. 2001; 32: 701-707
        • Farr BM
        • Kaiser DL
        • Harrison BD
        • Connolly CK
        Prediction of microbial aetiology at admission to hospital for pneumonia from the presenting clinical features. British Thoracic Society Pneumonia Research Subcommittee.
        Thorax. 1989; 44: 1031-1035
        • Schneeberger PM
        • Dorigo‐Zetsma JW
        • Van Der Zee A
        • Van Bon M
        • Van Opstal JL
        Diagnosis of atypical pathogens in patients hospitalized with community‐acquired respiratory infection.
        Scand J Infect Dis. 2004; 36: 269-273
        • Sopena N
        • Pedro‐Botet ML
        • Sabria M
        • et al.
        Comparative study of community‐acquired pneumonia caused by Streptococcus pneumoniae, Legionella pneumophila or Chlamydia pneumoniae.
        Scand J Infect Dis. 2004; 36: 330-334
        • Murdoch DR
        • Laing RT
        • Mills GD
        • et al.
        Evaluation of a rapid immunochromatographic test for detection of Streptococcus pneumoniae antigen in urine samples from adults with community‐acquired pneumonia.
        J Clin Microbiol. 2001; 39: 3495-3498
        • Hedlund J
        • Hansson LO
        Procalcitonin and C‐reactive protein levels in community‐acquired pneumonia: correlation with etiology and prognosis.
        Infection. 2000; 28: 68-73
        • 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
        • Gutierrez F
        • Masia M
        • Rodriguez JC
        • et al.
        Epidemiology of community‐acquired pneumonia in adult patients at the dawn of the 21st century: a prospective study on the Mediterranean coast of Spain.
        Clin Microbiol Infect. 2005; 11: 788-800
        • Gutiérrez F
        • Masiá M
        • Rodríguez JC
        • et al.
        Evaluation of the immunochromatographic Binax NOW assay for detection of Streptococcus pneumoniae urinary antigen in a prospective study of community‐acquired pneumonia in Spain.
        Clin Infect Dis. 2003; 36: 286-292
        • Masiá M
        • Gutiérrez F
        • Llorca B
        • et al.
        Serum concentrations of lipopolysaccharide‐binding protein as a biochemical marker to differentiate microbial etiology in patients with community‐acquired pneumonia.
        Clin Chem. 2004; 50: 1661-1664
        • Masiá M
        • Gutiérrez F
        • Shum C
        • et al.
        Usefulness of procalcitonin levels in community‐acquired pneumonia according to Patients Outcome Research Team Pneumonia Severity Index.
        Chest. 2005; 128: 2223-2229
        • Fine MJ
        • Auble TE
        • Yealy DM
        • et al.
        A prediction rule to identify low‐risk patients with community‐acquired pneumonia.
        N Engl J Med. 1997; 336: 243-250
        • Babson AL
        • Olson DR
        • Palmieri T
        • et al.
        The IMMULITE assay tube: a new approach to heterogeneous ligand assay.
        Clin Chem. 1991; 37: 1521-1522
        • Ledue TB
        • Weiner DL
        • Sipe JD
        • et al.
        Analytical evaluation of particle‐enhanced immunonephelometric assays for C‐reactive protein, serum amyloid A and mannose‐binding protein in human serum.
        Ann Clin Biochem. 1998; 35: 745-753
        • Meisner M
        • Tschaikowsky K
        • Schnabel S
        • et al.
        Procalcitonin‐influence of temperature, storage, anticoagulation and arterial or venous asservation of blood samples on procalcitonin concentrations.
        Eur J Clin Chem Clin Biochem. 1997; 35: 597-601
        • Marrie TJ
        Community‐acquired pneumonia.
        Clin Infect Dis. 1994; 18: 501-513
        • Riquelme R
        • Torres A
        • El‐Ebiary M
        • et al.
        Community‐acquired pneumonia in the elderly: a multivariate analysis of risk and prognostic factors.
        Am J Respir Crit Care Med. 1996; 154: 1450-1455
        • Alvarez Gutierrez FJ
        • Garcia Fernandez A
        • Elias Hernandez T
        • et al.
        Community acquired pneumonia in patients older than 60 years. Incidence of atypical agents and clinical–radiological progression.
        Med Clin (Barc). 2001; 117: 441-445
        • Falguera M
        • Sacristan O
        • Nogues A
        • et al.
        Non‐severe community‐acquired pneumonia: correlation between cause and severity or comorbidity.
        Arch Intern Med. 2001; 161: 1866-1872
        • Ruiz M
        • Ewig S
        • Torres A
        • et al.
        Severe community‐acquired pneumonia. Risk factors and follow‐up epidemiology.
        Am J Respir Crit Care Med. 1999; 160: 923-929
        • Almirall J
        • Bolibar I
        • Vidal J
        • et al.
        Epidemiology of community‐acquired pneumonia in adults: a population‐based study.
        Eur Respir J. 2000; 15: 757-763
        • Gleason PP
        The emerging role of atypical pathogens in community‐acquired pneumonia.
        Pharmacotherapy. 2002; 22: 2S-11S
        • Jokinen C
        • Heiskanen L
        • Juvonen H
        • et al.
        Microbial etiology of community‐acquired pneumonia in the adult population of 4 municipalities in eastern Finland.
        Clin Infect Dis. 2001; 32: 1141-1154
        • Brown RB
        • Iannini P
        • Gross P
        • Kunkel M
        Impact of initial antibiotic choice on clinical outcomes in community‐acquired pneumonia: analysis of a hospital claims‐made database.
        Chest. 2003; 123: 1503-1511
        • Gleason PP
        • Meehan TP
        • Fine JM
        • Galusha DH
        • Fine MJ
        Associations between initial antimicrobial therapy and medical outcomes for hospitalized elderly patients with pneumonia.
        Arch Intern Med. 1999; 159: 2562-2572