Community-Acquired Pneumonia
A Prospective Outpatient Study
PIERRE-YVES BOCHUD, FRANÇOIS MOSER, PHILIPPE ERARD, FRANÇOIS VERDON, JEAN-PAUL STUDER, GILBERT VILLARD, ALAIN COSENDAI, MARTINE COTTING, FREDY HEIM, JACQUELINE TISSOT, YVES STRUB, MARC PAZELLER, LAYLEE SAGHAFI, ALINE WENGER, DANIEL GERMANN, LUCAS MATTER, JACQUES BILLE, LAURENT PFISTER, AND PATRICK FRANCIOLI

Introduction Pneumonia constitutes the sixth cause of death and the first cause of infectious deaths in the United States (5). The incidence of pneumonia in community studies ranges from 2.6 to 16.8 per 1,000 adults per year (1, 7, 24, 30, ). Most information and recommendations for clinical practice are based on studies of hospitalized patients, who represent only 5%-16% of pneumonia cases (24, 39). Only 11 clinical studies of community pneumonia have been performed in ambulatory practice (1, 9, 10, 16, 17, 23, 36, 38, 42, 51, ). Most of these studies have focused on etiology and provide little information about clinical outcome (9, 16, 42, ). In addition, interpretation of the results is hampered by the small number of patients (9, 16, 51), by incomplete microbiologic documentation (17, 23, 38, 42), and especially by the absence of clear radiographic criteria (16, 36, 42, 51, 55). Understanding the outcome of community pneumonia is important because it allows physiciansto evaluate the risk for potential complications and the natural history of symptom resolution. Few reports have assessed the prognosis of community-acquired pneumonia, especially in the ambulatory setting (13, 22, ). The purpose of this prospective study is to establish the etiology, clinical and radiographic characteristics, and prognosis of community pneumonia in a population of patients seeking care from practitioners on an ambulatory basis.

From Division of Hospital Preventive Medicine (PYB, LS, PF) and Division of Infectious Diseases (PYB, JB, PF), Centre Hospitalier Universitaire Vaudois, Lausanne; Institute of Microbiology (AW, JB) of Lausanne; Institute of Microbiology (DG, LM) of St. Gallen; Hospital of Cadolles (LP), Neuchatel; and private practitioner (FM, PE, FV, JPS, GV, AC, MC, FH, JT, YS, MP), Neuchatel, Switzerland. Part of this study was supported by an educational grant from Pfizer. Address reprint requests to: Prof. Patrick Francioli, Division autonome de médecine préventive hospitalière, CHUV, CH-1011 Lausanne, Switzerland. Fax: 41 21 314 02 62; e-mail: Patrick.Francioli@chuv.hospvd.ch.

Patients and Methods
Potential patients were recruited consecutively during ambulatory consultation by 11 practitioners. The inclusion criteria were as follows: 1) age over 15 years, 2) presence of recent symptoms of lower respiratory infection, 3) presence of pulmonary infiltrate on chest X-rays, and 4) patient’s consent. Patients coming from a nursing home or hospitalized duringthe month before the consultation were excluded from the study. The total number of patient visits during the period of the study was recorded by 5 practitioners.

Patients
Case history data were recorded in a standard questionnaire for each patient, comprising age, sex, occupation, comorbidities requiring regular medical follow-up and/or chronic medication (chronic obstructive pulmonary disease, heart failure, alcoholism, diabetes mellitus) and other relevant information. The following paraclinical exams were obtained: chest X-ray, differential blood count, virus cultures from a throat swab placed in a viral transport medium (0.2 M sucrose-phosphate containing 1% bovine serum albumin, 0.1 mg gentamicin per mL, and 2.5 mg amphotericin per mL), and serum samples at 0 and 4 weeks. Sputum sample for direct examination and culture and blood cultures were obtained before the initiation of antimicrobial therapy. The patients were treated with a macrolide antibiotic, josamycin with a loading dose of 1 g followed by 500 mg twice daily for 8 days or longer if infection with Legionella or Mycoplasma was diagnosed or suspected. The patients were followed up by the same practitioner, and seen at least once after 4 weeks. Symptoms and signs present during these consultations (feeling unwell, shivering, dyspnea, cough, expectoration, pleurodynia, headache, myalgia) were noted on a standard form and evaluated on a severity scale (1 mild, 2 moderate, 3 severe). A chest X-ray was obtained at the firstconsultation and at least once more during a later consultation.

Radiographic interpretation
X-rays obtained for each patient were evaluated by a radiologist according to predefined criteria including the localization of the infiltrate (6 predefined zones: upper, middle and lower parts on the left and right lungs), and the presence of adenopathy, pleural reaction, or abscess formation. The pattern of infiltrates was categorized as predominantly alveolar, predominantly interstitial, or mixed alveolar and interstitial (53). Any other particular finding was also recorded, such as suspicion of tumor. The radiologic outcome was classified into 4 categories: 1) complete cure, 2) presence of minimal residual infiltrate, 3) improvement but significant residual lesion, or 4) no change or deterioration.

MicrobiologyThe serum specimens collected at 0 and 4 weeks were tested for antibodies by the following methods: standard complement fixation tests were used for Mycoplasma pneumoniae, Chlamydia spp., Coxiella burnetii, influenza viruses A and B, parainfluenza viruses types 1–3, adenoviruses, and respiratory syncytial virus. Indirect immunofluorescent antibody tests were used for Coxiella burnetii and L. pneumophila of serogroups 1–10. Chlamydia pneumoniae antibodies were tested by microimmunofluorescence (32). ELISA test was used to detect antibodies against Streptococcus pneumoniae. Sputum specimens were smeared for Gram stain and streaked on blood agar, chocolate, and MacConkey plates by the practitioner and incubated immediately at 37 C. After a minimum of 12 hours, slides, agar plates, and the rest of the original specimen kept at 4 C were sent to the reference laboratory by express mail. Gram stain was assessed in 3 different conditions: slide prepared and read by the practitioner, slide prepared by the practitioner but read by the microbiologist, and slide prepared with the original specimen and read by the microbiologist. A sputum was considered to be of good quality if it contained less than 25 epithelial cells per high power field (54). Pneumococcal antigen was also tested in the sputum.

Results During the 4 years of the study, the 11 practitioners enrolled 184 patients, of whom 14 were excluded from analysis due to incomplete data or absence of pulmonary infiltrates on the chest X-ray uponreview by the radiologist. This represented approximately 0.6 case per 1,000 patient visits. Etiology One-hundred seven etiologic agents were identified in 92 patients (54.1%) (Table 1). Evidence of bacterial infection due to pneumococci (34 cases) or H. influenzae (3 cases) was observed in 37 cases, “atypical” bacterial infection in 37 cases, and viral infection in 18 cases. The diagnosis was considered as definite in 44 cases and as presumptive in 48 cases. The most common organism was S. pneumoniae, followed by M. pneumoniae, influenza A virus, and C. pneumoniae (see Table 1). Polymicrobial infection was observed in 15 cases (11%). Among the 34 S. pneumoniae infections, the diagnosis was based on only 1 diagnostic test in 25 cases and on at least 2 diagnostic tests in 9 cases (Figure 1). The diagnosis was considered definite in 6 cases with positive blood cultures, and presumptive in the others. Since serologic tests could be performed in only 108 patients because of lack of available serum, the number of pneumonia episodes due to this organism may have been underestimated. S. pneumoniae was combined with another pathogen in 10 cases (virus, 6 cases; atypical pathogen, 4 cases). No blood cultures were positive for an organism other than S. pneumoniae. All 3 cases of H. influenzae pneumonia were diagnosed by sputum culture and Gram strain. All “atypical” pneumonia episodes were diagnosed by serologic tests. Among 23 M. pneumoniae pneumonias, 21 were considered definite. M. pneumoniae was combinedwith another pathogen in 5 cases: S. pneumoniae (1 case), adenovirus (1 case), influenza B virus (1 case), parainfluenza virus (1 case), and betahemolytic Streptococcus (1 case). Pneumonia due to Chlamydia pneumoniae was diagnosed in 9 cases

Definition of the etiologic diagnosis
The etiologic diagnosis was considered as definite if there was a blood culture positive for a respiratory pathogen, a throat culture positive for a viral respiratory pathogen, or a fourfold or greater increase of serum antibody titers between the 2 samples obtained 4 weeks apart. The diagnosis was considered presumptive if there was a positive sputum culture for a respiratory pathogen or a high antibody titer on the initial specimen in a patient symptomatic for more than a week. For S. pneumoniae a positive pneumococcal antigen test in the sputum associated with a Gram stain showing Gram-positive diplococci was also considered as presumptive. When a patient met the criteria for a pneumonia due to 2 pathogens, the pathogen associated with a definite diagnosis was considered as dominant over the presumptive pathogen. For influenza A and B, high titers were accepted as evidence of recent infection only if the patient was not vaccinated and if the episode was compatible with the epidemiology. For most analyses, etiologic agents were grouped as follows: “pyogenic bacteria” (S. pneumoniae and Haemophilus influenzae), “atypical bacteria” (Mycoplasma pneumoniae, Chlamydia spp., Coxiella burnetii, and Legionella), “virus,” and“undetermined.”

Statistical analysis
Testing procedures for differences between specific groups included chi-square for categorical variables and the Student t- test or ANOVA for continuous variables. Associations were considered to be statistically significant if the p value was 0.05 using a 2tailed test. A Gram stain was considered as a true positive or negative when putting the patient in the correct etiologic category.



No organism identified — — — — 78 45.9 Total episodes with organisms* 44 25.9 48 28.2 92 54.1 “Pyogenic” bacteria 6 3.5 31 18.2 37 21.8 Streptococcus pneumoniae 6 3.5 28 16.5 34 20.0 Haemophilus influenzae — — 3 1.8 3 1.8 “Atypical” bacteria 28 16.5 9 5.3 37 21.8 Mycoplasma pneumoniae 21 12.4 2 1.2 23 13.5 Chlamydia spp. 5 2.9 4 2.4 9 5.3 Legionella spp. — — 1 0.6 1 0.6 Coxiella Burnetti 2 1.2 2 1.2 4 2.4 Viruses 8 4.7 10 5.9 18 10.6 Influenza A 5 2.9 7 4.1 12 7.1 Influenza B 2 1.2 2 1.2 4 2.4 Parainfluenza 1 0.6 — — 1 0.6 Adenovirus — — 1 0.6 1 0.6 *Includes 15 patients with mixed infections (16.3%): S. pneumoniae and influenza A virus (3), S. pneumoniae and influenza B virus (2), S. pneumoniae and Legionella spp. (2), S. pneumoniae and respiratory syncitial virus (1), S. pneumoniae and Mycoplasma pneumoniae (1), S. pneumoniae and Chlamydia spp. (1), Mycoplasma pneumoniae and adenovirus (1), Mycoplasma pneumoniae andinfluenza B virus (1), Mycoplasma pneumoniae and parainfluenza virus (1), Mycoplasma pneumoniae and beta-hemolytic Streptococcus (1), influenza B and influenza A virus (1).

of which 5 were considered definite and 4 presumptive. Pneumonia due to Coxiella burnetii was diagnosed in 4 cases (2 certain, 2 presumptive), and due to Legionella spp., in only 1 case. All viral cultures of throat swabs were negative in the first 100 patients, and the method was abandoned. The 18 cases of viral pneumonia were diagnosed by serologic tests. Among the 16 pneumonia cases due to influenza A or B virus, 7 diagnoses were considered definite and 9 presumptive. Among these 16 patients, evidence of superinfection by S. pneumoniae was observed in 5 cases. Pneumonia due to parainfluenza virus and adenovirus was diagnosed in 1 case each. Sex and age Among the 170 patients analyzed, 82 (48.2%) were men. Median age was 43.1 years (range, 15–96 yr) (Table 2). Seventy percent of the patients had no comorbidity. The most common underlying conditions observed in the remaining patients were cardiac disease (10.6%), chronic obstructive pulmonary disease (6.5%), and diabetes (2.9%); 35.3% of the patients were smokers and 12.9% were alcoholics. The median age of patients suffering from bacterial or viral pneumonia was significantly higher than that of patients presenting with atypical pneumonia (see Table 2). S. pneumoniae caused 28.2% of pneumonias in patients over 45 years, and 13.0% in younger patients (p 0.01). All 3 H.influenzae episodes were observed in patients over 45 years. M. pneumoniae was responsible for 22.8% of pneumonias in patients under 45 years and 2.6% in patients

over that age (p 0.001). Similarly, C. pneumoniae accounted for 8.7% of pneumonias in patients under 45 years, compared with only 1.3% in older patients (p 0.03) (Figure 2). Comorbidities Comorbidities were more frequent in patients with viral pneumonia and less frequent in patients suffering from atypical pneumonias (see Table 2). No patients with cardiac insufficiency presented with pneumonia due to M. pneumoniae, whereas this pathogen caused 24% of pneumonias in the other patients (p 0.03). Influenza A virus caused 25% of pneumonic episodes among patients with cardiac insufficiency and 5.2% of episodes in patients without heart disease (p 0.003). No significant difference was observed in the etiology of pneumonia in relation to the presence or absence of chronic obstructive pulmonary disease, diabetes, smoking, or alcoholism.

BOCHUD ET AL TABLE 2. Underlying conditions according to etiologic categories Total “Pyogenic” Bacteria “Atypical” Bacteria Viruses 18 (10.6%) 33.3 56.3 44.4 11.1 33.3 5.6 5.6 0.0 0.0 5.6 5.6 0.0 0.0 5.6 11.1 44.4 0.0 16.7 33.3 0.0 11.1 0.0 0.0 22.2 Undetermined p Value (chi-square or ANOVA) 78 (45.9%) 51.3 44.6 35.9 7.7 11.5 3.8 2.6 1.3 2.6 1.3 0.0 1.3 0.0 2.6 14.1 33.3 5.1 14.1 21.8 10.3 11.5 7.7 3.8 25.6NS 0.001 NS (0.07) NS 0.002 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

Number of patients (%) 170 Characteristic Male (%) 48.2 Median age (yr) 43.1 Comorbidities (%) 30.0 COPD (%) 6.5 Cardiovascular disease (%) 10.6 Diabetes mellitus (%) 2.9 Hypertension (%) 1.8 Splenectomy (%) 1.2 HIV infection (%) 1.2 Digestive disease (%) 1.8 Renal disease (%) 1.2 Cancer (%) 0.6 Neuropsychologic 1.2 disease (%) Dental pyorrhea (%) 2.4 Habits Alcoholism (%) 12.9 Tobacco use (%) 35.3 Other characteristics Pneumococcal vaccine (%) 4.1 Influenza vaccine (%)* 10.6 24.1 Contact with animals (%)† Positive travel history (%)* 8.8 Recent infection (%)† 10.6 Recent antibiotic treatment (%)† 4.7 Recent hospitalization (%) 2.9 Upper respiratory tract 24.7 infection in contacts (%)† Abbreviations: NS not significant; COPD *During the last 12 months. †During the month prior to diagnosis.

37 (21.8%) 45.9 57.2 24.3 2.7 8.1 2.7 0.0 0.0 0.0 2.7 2.7 0.0 2.7 2.7 16.2 43.2 5.4 8.1 18.9 8.1 13.5 2.7 2.7 27.0

37 (21.8%) 51.4 34.2 16.2 5.4 0.0 0.0 0.0 2.7 0.0 0.0 0.0 0.0 2.7 0.0 8.1 27.0 2.7 2.7 29.7 10.8 5.4 2.7 2.7 21.6

chronic obstructive pulmonary disease; HIV

human immunodeficiency virus.

Among 7 patients vaccinated against S. pneumoniae in the previous 5 years, 2 presented with pneumococcal pneumonia. Among 6 patients vaccinated against influenza, 2 developed pneumonia due to influenza A virus. Seasonality The seasonal variation is shown in Figure 3. Twenty-one percent of pneumonias occurredin spring, 15% in summer, 29% in autumn, % in winter (p 0.001). Among the S. pneumoniae pneumonias, 41% were diagnosed in spring (p 0.03). Seventythree percent of pneumonias of indeterminate origin occurred in autumn and winter (p 0.001). Ninetyone percent of pneumonias due to M. pneumoniae were diagnosed during the first 2 years of the study (p 0.001), and 83% of cases due to influenza A virus, during the fourth year (p 0.01). Although pneumonias of indeterminate origin occurred throughout the study, there was a concomitant rise in frequency during the epidemics of pneumonia due to mycoplasma and influenza A.

Gram stain Sputum was obtained from 105 patients (62%) (Table 3). The sensitivity and specificity of the sputum Gram stain to predict the final etiologic category (“pyogenic” [mostly pneumococci], “atypical,” “viral,” and “undetermined”) were evaluated for 3 different methquality, but the sensitivity and specificity were higher. When the slide was prepared and read by the microbiologist, 51 sputa were judged of sufficient quality, the sensitivity was low but the specificity was high. Clinical signs and symptoms The symptoms encountered most frequently were cough (96%), malaise (90%), fever (81%), chills (59%), headache (58%), myalgia (58%), dyspnea (46%), and pleurodynia (37%) (Table 4). The most common clinical observations were rales (69%), expectoration (52%), and dullpercussion sounds (19%) (Table 5). Bacterial pneumonias were characterized by pleural rubbing (never observed in the other etiologic categories) and more frequent pleuritic pain and sputum production (p 0.07), as well as higher leukocyte counts. The leukocyte count was significantly less elevated among patients with atypical pneumonia (p 0.001). Radiography Initial chest X-rays were available for interpretation by the radiologist in 159 cases, with follow-up Xrays for 136 cases (Table 6). The presence of an infiltrate was considered to be uncertain in 4% of cases, and to be absent in 1% of cases (these cases were excluded from analysis). The infiltrates were often located predominantly in the left lung (69%) (p 0.001), especially the middle (45%) and lower (49%) regions. Bilateral infiltrates were observed in 14% of cases. The infiltrate was alveolar in 52%, and interstitial in 25%. In 95% of the episodes the infiltrate did not follow a segmental or lobar distribution. Pleural effusion, adenopathy, or abscess formation were rare. Viral pneumonias were significantly associated with an interstitial infiltrate (56%), with bilateral involvement (35%), and with a right median region involvement (18%). The infiltrate was more

FIG. 3. Etiology in 170 outpatients with pneumonia over 4 years.

ods. When the slide prepared by the practitioner was read by him, 62 of the 105 sputa were judged of sufficient quality, but the sensitivity and specificity to predict the etiologic diagnosis were low. When the sameslide was interpreted by the reference laboratory microbiologist, only 37 sputa were judged of sufficient

TABLE 3. Sensitivity and specificity of the Gram stain according to 3 different procedures Gram Stain 1 Slide preparation Slide examination Number of patients Number (%) of sputum samples obtained Sputum judged of sufficient quality ( 25 epithelial cells/high power field) True positive (n) False positive (n) False negative (n) True negative (n) Sensitivity (%) Specificity (%) Positive predictive value (%) Negative predictive value (%) Physician Physician 170 105 (62%) 62 (36%) 12 16 7 27 63 63 43 79 Gram Stain 2 Physician Microbiologist 170 105 (62%) 37 (22%) 9 6 2 20 82 77 56 91 Gram Stain 3 Microbiologist Microbiologist 170 105 (62%) 51 (30%) 9 9 8 25 53 74 50 76

Number of patients Cough at diagnosis (%) Number of days before diagnosis (median) Number of days on treatment (median) Malaise at diagnosis (%) Number of days beforediagnosis (median) Number of days on treatment (median) Temperature 38 °C at diagnosis (%) Median temperature Number of days before diagnosis (median) Number of days on treatment (median) Chills at diagnosis (%) Number of days before diagnosis (median) Number of days on treatment (median) Headache at diagnosis (%) Number of days before diagnosis (median) Number of days on treatment (median) Myalgia at diagnosis (%) Number of days before diagnosis (median) Number of days on treatment (median) Dyspnea at diagnosis (%) Number of days before diagnosis (median) Number of days on treatment (median) Pleurodynia at diagnosis (%) Number of days before diagnosis (median) Number of days on treatment (median) Abbreviations: NS not significant.

170 96 4 10 90 4 6 81 39 3 2 59 3.5 1 58 3 3 58 3 3 46 3 5 37 2 5

frequently described as alveolar in cases of “atypical” (74%) or “pyogenic” bacterial pneumonia (56%). Treatment and outcome Treatment with josamycin was started in 158 of the 170 studied patients (93%). Another antibiotic treatment was administered at the start of the illness in 12 cases, including 9 patients for whom hospitalization was decided during the consultation and 3 patients for other reasons. In most patients the outcome was favorable after 4 weeks, with disappearance of fever (98%) and other signs (95%-99%), improvement of the chest X-ray (95%), and improvement of laboratory parameters (see Tables 5 and 6). Patients with viral

pneumonia did not improve as fast as those with otheretiologies after the initiation of antibiotic therapy: the median duration of malaise (13 days) and headache (8 days) on treatment was significantly longer, and a complete radiologic cure at 4 weeks was observed in only 15% of cases, a proportion that is significantly lower than in the other etiologic groups (p 0.004). Among the 158 patients treated with josamycin, the treatment was changed in 11 cases (6.9%) due to diarrhea (1 case) or no improvement of symptoms after a few days of treatment (10 cases). Cancer was diagnosed in 6 patients (small-cell carcinoma in 3, bronchial carcinoma in 2, and malignant non-Hodgkin lymphoma in 1). The diagnosis was


COMMUNITY-ACQUIRED PNEUMONIA TABLE 5. Physical signs and laboratory values according to etiologic categories Total First consultation Number of patients Temperature 38 C (%) Rales (%) Sputum production (%) Dullness on percussion (%) Bronchial breath sounds (%) Pleural rub (%) White blood cells/mm3 (median) Segmented neutrophil granulocytes/mm3 (median) Nonsegmented neutrophil granulocytes/mm3 (median) Sedimentation rate (median) “Pyogenic” Bacteria 37 74 78 70 17 6 11 13,000 5,464 1,865 38.5 34 5 3 0 0 0 3 5,600 3,093 271 7.0 “Atypical” Bacteria 37 68 57 44 19 6 0 6,700 2,810 902 34.5 37 0 6 6 3 0 0 6,000 2,484 162 5.5 Viruses Undetermined

Last consultation Number of patients 164 Temperature 38 C (%) 2 Rales (%) 4 Sputum production (%) 5 Dullness on percussion (%) 3 Bronchial breath sounds (%) 1 Pleural rub (%) 1 6,150 White blood cells/mm3 (median) Segmented neutrophil 3,072 granulocytes/mm3 (median) Nonsegmented neutrophil 166 granulocytes/mm3 (median) Sedimentation rate (median) 7.0 Abbreviations: NS not significant.

NS NS NS NS NS NS NS NS NS NS

TABLE 6. Radiologic findings according to etiologic categories Total Number of patients Number of X-rays evaluated Extent Only 1 lobe involved (%) More than 3 lobes involved (%) Bilateral involvement (%) Characteristics of infiltrate Alveolar (%) Mixed alveolar and interstitial (%) Interstitial (%) Segmental or lobar distribution (%) Other characteristics Pleural effusion (%) Hilar adenopathies (%) Abscess (%) Evolution at 4 weeks Cure (%) Minimal residual lesions (%) Improvement (%) No improvement or worsening (%) Neoplasia Suspected (no., [%]) Confirmed (no.) Abbreviations: NS not significant. 170 159 60 6 14 52 23 25 5 20 20 2 45 38 12 5 9 (7) 6 “Pyogenic” Bacteria 37 37 69 3 6 56 22 22 9 22 24 0 33 47 13 7 2 (6) 1 “Atypical” Bacteria 37 36 66 0 9 74 11 14 3 17 25 3 68 26 6 0 1 (3) 1 Viruses 18 17 41 18 35 44 0 56 0 18 29 6 15 62 8 15 0 (0) 0 Undetermined 78 69 57 7 16 39 36 25 7 22 13 1 45 35 15 5 6 (10) 4 NS NS 0.02 0.002 0.02 0.02 NS NS NS NS 0.004 NS NS NS NS NS p Value (chi-squareor ANOVA)

suspected at the time of the first consultation in 4 patients on the basis of the chest X-ray (3 cases) or manifestations suggestive of metastases (1 case). In the other 2 cases, the diagnosis was suspected on the basis of the control chest X-ray, associated with poor clinical response in 1 of them. In another 3 patients, a tumor was suspected on the chest X-ray but was not confirmed upon further investigation and follow-up. A total of 14 patients required hospitalization: in 9 of them, the hospitalization was decided during the initial visit and in 5, at a later stage due to poor response to antibiotic therapy. The median age of the hospitalized patients (67 yr) was higher than that of the other patients (41 yr) (p 0.005), and they presented more frequently with comorbidities (64% versus 27%, p 0.03), especially with chronic obstructive pulmonary disease (21% versus 5%, p 0.02) and alcoholism (36% versus 11%, p 0.01). Initial blood cultures were positive in 3 of the 14 hospitalized patients (21%), and in only 3 of the 156 non-hospitalized patients (2%) (p 0.001). The 2 groups did not differ significantly with regard to vaccination against S. pneumoniae or influenza, contact with animals, recent travel, prior infection, or infection in contacts. Two patients died (1.2%). Both had pneumonia of indeterminate origin. The first patient, aged 83 years, was hospitalized for bilateral pneumonia and died from liver failure due to hepatitis B-related cirrhosis. Thesecond patient, aged 86 years, died at home a few days after the diagnosis of severe pneumonia associated with end-stage heart failure. Discussion This study identified 170 patients with communityacquired pneumonia diagnosed by private practitioners at their offices. Except for the Pneumonia Patient Outcomes Research Team (PORT) cohort study (23), only a minority of patients with lower respiratory tract infections described in previous large outpatient reports actually had chest X-rays performed and/or presented with an infiltrate (36, 55). However, the PORT cohort study was not designed to determine the cause of pneumonia, and the proportion of patients with microbiologic documentation was low. In the present study, a large battery of diagnostic tests were used. Strict definitions were applied for the diagnosis which were classified as “definite” or “presumptive,” taking into account the specificity of the various tests. This should allow for easier comparisons with other studies (18). Finally, particular attention was paid to patient follow-up, which included a systematic visit at 4 weeks. The organism identified most often was S. pneumoniae. In studies performed in hospitalized patients, S. pneumoniae is invariably the most common etiologic agent, representing up to 75% of causes of pneumonia (5, 18). Although S. pneumoniae was also the organism most frequently isolated in many community studies (10, 16, 36, 51, ), some have found atypical organisms such as M. pneumoniae (9, 17, 38), C.pneumoniae (1), or viruses (42) to be more frequent. Various factors may explain these differences. In particular, tests to diagnose pneumococcal pneumonia were not applied equally extensively in all the studies (Table 7). In studies where S. pneumoniae was frequently identified, at least 3 diagnostic methods were used (sputum culture, blood culture, test for pneumococcal antigen at multiple sites) (10, 16, 36, 55). The low specificity of certain tests may also lead to overestimation of the number of pneumococcal infections (17, 38). The pneumococcal antigen test of the sputum could be performed in 22 patients diagnosed with pneumococcal pneumonia, and was positive in 10 (45%). A good correlation between the detection of pneumococcal antigen in the sputum by immunoelectrophoresis and the rise in antipneumococcal antibody titers in pneumonia patients suggests that the detection of this antigen is a predictor of infection rather than colonization (8). However, the sensitivity and specificity of this test have not been evaluated sufficiently (44). In our study only 3 of the 34 cases of pneumococcal pneumonia were diagnosed solely on the basis of this examination and the presence of diplococci on Gram stain, so that it could not have contributed greatly to an overestimation of the rate of pneumococcal pneumonia. Sixteen pneumococcal pneumonias were diagnosed solely on the basis of a fourfold rise in antibody titers, but the tests could be performed in only 108 of the 170 cases. Blood cultures werepositive for S. pneumoniae in 6 cases. Sputum cultures were positive in 11 cases, and 2 of the 34 pneumococcal pneumonias were diagnosed solely on the basis of this examination. Other authors have shown that only 44%-50% of patients with pneumococcal pneumonia with bacteremia have a positive sputum culture (2, 39). Several investigators have suggested that routine sputum cultures should not be performed in less seriously affected patients (33, 49). Despite their rather poor sensitivity, cultures allow examiners to perform antibiotic susceptibility testing, the result of which might be important in certain geographic areas (4). Indeed, the emergence of pneumococci with reduced sensitivity to several common antibiotics has been noted in several countries during the last decade (6, 29, ). M. pneumoniae is the second etiologic agent in terms of incidence and was responsible for almost 15% of the pneumonic episodes. This organism was the most common etiologic agent in 2 studies, where it represented 9%-37% of cases of pneumonia (9, 17, ). On the other hand, it was rarely found in other studies (1, 55). These differences may be due to dif-


TABLE 7. Review of studies of pneumonia in outpatients
Dulake et al 1982 (16) Plymouth, UK 106 (15/91) 54 (0/54) 236 (22/214) 510 (462/48) 117 (0/117) 315 (4/311) 105 (53/52) 149 (8/141) Göteborg, Sweden Nottingham, UK Valencia, Spain Tromsø, Norway Nottingham, UK Barcelona, Spain Halifax, Canada Everett et al 1983 (17) Berntsson et al 1986 (9) Woodheadet al 1987 (55) Blanquer et al 1991 (10) Melbye et al 1992 (42) Macfarlane et al 1993 (36) Almirall et al 1993 (1) Marrie et al 1996 (38) Fine et al 1999 (23) Present Report

ferences in diagnostic tests and lack of consensus on diagnostic criteria. This discrepancy may also be explained by seasonal variation and by epidemic peaks occurring every 3–7 years (14, 25, 34, ). The present study supports this hypothesis: almost all the M. pneumoniae infections occurred in the first 2 years of the study. In some studies, C. pneumoniae caused up to 15% of the pneumonic episodes (1, 38), whereas it was observed much more rarely in other studies (see Table 7) (10, 50). Serologic methods used may partly account for these differences (41). In our study, C. pneumoniae infections were relatively rare (4.1%). Seasonal variations in C. pneumoniae infections have been reported (27, 31, ). Viral pneumonia was found in 11% of patients, and almost all were attributed to influenza viruses A (8%) or B(2%). Bacterial superinfection in viral pneumonia is most commonly caused by S. pneumoniae and Staphylococcus aureus (40). Both S. pneumoniae and influenza A or B virus were found in 5 of our cases, but we detected no infections due to Staph. aureus. Viral pneumonia was seen most often in patients presenting with comorbidities, especially heart failure. Multiple pathogens were found in 10.6% of pneumonic episodes, and all types of combination were observed, as reported by others (1, 38). Despite multiple combinations of diagnostic methods, the proportion of pneumonia of indeterminate origin has remained high in most studies, ranging from 23% to 80%. In our study, these episodes accounted for 45.9% of the episodes and presented characteristics that suggested bacterial pneumonia in some cases (elevated nonsegmented neutrophil count, pleuritic pain) and atypical or viral pneumonia in other cases (bilateral infiltrates). The frequency of these episodes increased during peaks of pneumonia due to Mycoplasma pneumoniae and influenza A virus. This observation suggests that part of the pneumonias of undetermined origin are caused by the same etiologic agents, which may be underdiagnosed. This may be favored by the early administration of antibiotics (38), the absence of sputum available for cultures (17, 38, 42), or inadequate sensitivity of serologic tests (41). Alternatively, some pneumonias of undetermined origin may be caused by emerging pathogens. In this regard, anaerobic organisms are theprincipal pathogens involved in aspiration pneumonia, pulmonary abscess, and empyema, but their role in uncomplicated pneumonia has not been fully elucidated. Bartlett et al (3) suggested that a substantial proportion of communityacquired pneumonias of indeterminate origin might be due to anaerobes. In 2 studies utilizing tracheal aspirations, anaerobes were encountered in 33% and 22% of patients presenting with community-acquired pneumonia, respectively (45, 48).

56 Fine et al 1999 (23)

S. pneumoniae C. pneumoniae M. pneumoniae S. pneumoniae (30) (15) (26) (2) H. influenzae S. pneumoniae C. pneumoniae H. influenzae (8) (12) (15) (1) Influenza B (3) M. pneuInfluenza A (4) Atypical moniae (8) pathogens (1) Influenza A Parainfluenza Caxiella Enterobacter(2) virus (5) burnetti (3) iaceae (0.2) RSV (2) Adenovirus Adenovirus B. catharralis (4) (3) (0.2), M. tuberculosis (0.2) S. pneumoniae S. pneumoniae M. pneumoniae M. pneumoniae S. pneumoniae S. pneumoniae Influenza A/B (?) (32) (9) (37) (36) (12) (14) Chlamydia H. influenzae Influenza A (4) H. influenzae H. influenzae Legionella RSV (11) spp. (?) (23) (12) (10) spp. (12) H. influenzae K. pneuChlamydia spp. S. pneumoniae Influenza A (6) M. pneu S. pneu(?) moniae (4) (3) (9) moniae (12) moniae (10) Fourth Influenza A/B P. multocida Measles (1) Influenza A (6) Influenza A Influenza A M. pneupathogen (%) (?) (4) (2), RSV (2) (8) moniae (6) Fifth Staph. aureus M. pneuM. tuberculosis C. psittaci (4) Adenovirus Influenza B (6) Parainfluenza pathogen (%) (?) moniae (4)(1) (2) virus (4)

There were only minor differences in the clinical features of the different etiologic groups. Patients with S. pneumoniae pneumonia presented with pleural involvement and produced sputum significantly more often, and had higher leukocyte counts. Studies have suggested that the speed of onset of symptoms, sputum production, and pleural pain allow to distinguish between the etiologic categories (35, 49), and that leukocyte counts are higher in S. pneumoniae pneumonia (19). This has not been confirmed in all studies (18). Although some characteristics have been anecdotally associated with certain etiologic agents, a computer analysis based on age, duration of the disease, sputum characteristics, leukocyte count, and radiographic appearance applied to 441 patients grouped into 4etiologic groups (pneumococci, mycoplasmas, others, and undetermined) placed only 42% of cases in the correct category (19). The value of direct examination of Gram-stained sputum has been debated for many years. The IDSA guidelines still recommend to perform sputum Gram stain, at least for hospitalized patients (4), whereas the ATS guidelines do not recommend it at all (43). Several combinations of criteria, including the polymorphonuclear cell count, epithelial cell count and quantity of mucus, have been proposed for assessing the quality of the specimen, but none has been found to be superior to the others (54). Some authors have suggested that the presence of numerous polymorphonuclear leukocytes with large numbers of bacteria constitutes an adequate basis for undertaking empirical treatment (11, 12, ). In a study performed at the John Hopkins Hospital in Baltimore, Maryland (5), about half the sputum specimens showed a possible pulmonary pathogen by direct examination, and there was a 90% correlation between the organisms identified initially by Gram-stain observation and those subsequently identified by culture. However, other studies have questioned the usefulness of Gram stain (2, 33, ). Among 12 different studies on the utility of sputum Gram stain, the sensitivity ranged from 15% to 100% and the specificity from 11% to 100%, and results appeared partly related to the observer (46). Our findings support these observations, since the results were different in 3 different conditionsof observation, with different slide preparations and different observers. Moreover, sputum specimens could be obtained in only 62% of patients and were considered to be of good quality in only one-third of them. The presence of a radiologic infiltrate was confirmed by the radiologist in 95% of cases. In 2,287 patients participating in the PORT cohort study, an infiltrate was detected by at least 2 of 3 radiologists in 87% of cases (28). It was striking that the left lung was affected significantly more often than the right, a finding which to our knowledge has not been re-

ported. Viral pneumonias were significantly more often characterized by bilateral and interstitial infiltrates. The infiltrates were rarely confined to a definite segment or lobe. By considering the number of lobes involved (38) or the presence of hilar adenopathies (37), some studies have revealed associations between certain radiographic characteristics and a given etiologic agent. Despite these associations, there are no radiologic criteria reliable enough to suggest firmly an etiologic diagnosis (20). However, the chest X-ray may be of prognostic value: multivariate analysis of the clinical and radiographic characteristics of 1,906 patients has shown that the presence of bilateral pleural effusion is significantly associated with lethality (28). Like other studies performed in patients treated on an ambulatory basis, most patients were young but the distribution of the pathogens varied according to the age. S. pneumoniaepneumonia was significantly more common among elderly patients, as classically described (4), as opposed to M. pneumoniae and C. pneumoniae, confirming the findings of other studies (43). Hospitalized patients (8%) had a median age of 67 years as compared to 41 years for outpatients; in addition, the rate of comorbidities was much higher (64% versus 27%). Similar observations were reported in the PORT cohort study, in which heart and pulmonary diseases were 4 and 2 times more frequent among hospitalized patients, respectively (23). In another study of 359 patients hospitalized for community pneumonia in Pittsburgh, Pennsylvania, the incidence of solid tumors (22.6%) and hematologic conditions (5.8%) was high, whereas these conditions were very rarely encountered in outpatients (18). In our series, a neoplasia was found to be associated with the pneumonic episodes in 6 cases (3.4%). Most patients in this study had a very good prognosis. At the time of the last consultation, only 2% of patients were still febrile and less than 5% of them presented signs of pneumonia on physical examination. The clinical symptoms and radiologic infiltrates in patients with viral pneumonia lasted longer than in the other etiologic groups. At 1 month, the leukocyte count was normal in 91% of cases and the mean erythrocyte sedimentation rate had fallen from 37 to 7 mm/h. Radiographic examination showed improvement or cure in 95% of cases. These observations are similar to those of others. Among 105 patients withcommunity-acquired pneumonia, radiologic signs of improvement were observed in 82% of the cases after 5 days of treatment, and residual radiographic signs in only 3% after 4 weeks (1). In the PORT cohort study, nearly 90% of 944 outpatients with communityacquired pneumonia had returned to usual activities or work during a 30-day follow-up. At 1 month of follow-up, only 2 of 170 patients had died (1.2%), both over 80 years old and with severe concomitant diseases. This result is similar to thePORT cohort study, in which only 6 of 944 outpatients with community-acquired pneumonia died (0.6%). Five of them were designated as high-risk patients based on a validated prediction rule for prognosis in pneumonia (21). This is also consistent with recent studies in outpatients reporting a lethality of 0% and 3% (23), although the patient population might differ according to the setting of recruitment (see Table 7) (1, 10, 20, 36, 38). This contrasts with the lethality of 8% (26) and 14% (18) reported in 2 large studies of patients hospitalized for communityacquired pneumonia. The present study confirms that communityacquired pneumonia has a favorable outcome and can be managed successfully in an outpatient setting. Patients hospitalized for community-acquired pneumonia constitute only a small subset of patients with pneumonia, with very different characteristics, including higher age and more frequent and severe comorbid conditions. In a small proportion of patients (3.4% inthis series), the pneumonia episode is the first overt manifestation of a local neoplastic process. Summary We initiated a prospective study with a group of practitioners to assess the etiology, clinical presentation, and outcome of community-acquired pneumonia in patients diagnosed in the outpatient setting. All patients with signs and symptoms suggestive of pneumonia and an infiltrate on chest X-ray underwent an extensive standard workup and were followed over 4 weeks. Over a 4-year period, 184 patients were eligible, of whom 170 (age range, 15–96 yr; median, 43 yr) were included and analyzed. In 78 (46%), no etiologic agent could be demonstrated. In the remaining 92 patients, 107 etiologic agents were implicated: 43 were due to “pyogenic” bacteria (39 Streptococcus pneumoniae, 3 Haemophilus spp., 1 Streptococcus spp.), 39 were due to “atypical” bacteria (24 Mycoplasma pneumoniae, 9 Chlamydia pneumoniae, 4 Coxiella burnetii, 2 Legionella spp.), and 25 were due to viruses (20 influenza viruses and 5 other respiratory viruses). There were only a few statistically significant clinical differences between the different etiologic categories (higher age and comorbidities in viral or in episodes of undetermined etiology, higher neutrophil counts in “pyogenic” episodes, more frequent bilateral and interstitial infiltrates in viral episodes). There were 2 deaths, both in patients with advanced age (83 and 86 years old), and several comorbidities. Only 14 patients (8.2%) required hospitalization. In 6patients (3.4%), the pneumonia episode uncovered a local neoplasia. This study shows that most cases of community-acquired pneumonia have a

favorable outcome and can be successfully managed in an outpatient setting. Moreover, in the absence of rapid and reliable clinical or laboratory tests to establish a definite etiologic diagnosis at presentation, the spectrum of the etiologic agents suggest that initial antibiotic therapy should cover both S. pneumoniae and atypical bacteria, as well as possible influenza viruses during the epidemic season. References
1. Almirall J, Morato I, Riera F, Verdaguer A, Priu R, Coll P, Vidal J, Murgui L, Valls F, Catalan F, et al. Incidence of community-acquired pneumonia and Chlamydia pneumoniae infection: A prospective multicentre study. Eur Resp J 6: 14–8, 1993. 2. Barrett-Connor E. The nonvalue of sputum culture in the diagnosis of pneumococcal pneumonia. Am Rev Resp Dis 103: 845–8, 1971. 3. Bartlett JG. Anaerobic bacterial pneumonitis. Am Rev Resp Dis 119: 19–23, 1979. 4. Bartlett JG, Breiman RF, Mandell LA, File TM Jr. Community-acquired pneumonia in adults: Guidelines for management. The Infectious Diseases Society of America. Clin Infect Dis 26: 811–38, 1998. 5. Bartlett JG, Mundy LM. Community-acquired pneumonia. N Engl J Med 333: 1618–24, 1995. 6. Bedos JP, Chevret S, Chastang C, Geslin P, Regnier B. Epidemiological features of and risk factors for infection by Streptococcus pneumoniae strains with diminished susceptibility to penicillin: Findingsof a French survey. Clin Infect Dis 22: 63–72, 1996. 7. Benson V, Marano MA. Current estimates from the National Health Interview Survey, 1992. Vital Health Stat 10 Jan(189): 1–269, 1994. 8. Berntsson E, Blomberg J, Lagergard T, Trollfors B. Etiology of community-acquired pneumonia in patients requiring hospitalization. Eur J Clin Microbiol 4: 268–72, 1985. 9. Berntsson E, Lagergard T, Strannegard O, Trollfors B. Etiology of community-acquired pneumonia in out-patients. Eur J Clin Microbiol 5: 446–7, 1986. 10. Blanquer J, Blanquer R, Borras R, Nauffal D, Morales P, Menendez R, Subias I, Herrero L, Redon J, Pascual J. Aetiology of community acquired pneumonia in Valencia, Spain: A multicentre prospective study. Thorax 46: 508–11, 1991. 11. Boerner DF, Zwadyk P. The value of the sputum gram’s stain in community-acquired pneumonia. JAMA 247: 642–5, 1982. 12. Bohte R, Hermans J, van den Broek PJ. Early recognition of Streptococcus pneumoniae in patients with community-acquired pneumonia. Eur J Clin Microbiol Infect Dis 15: 201–5, 1996. 13. Carson CA, Fine MJ, Smith MA, Weissfeld LA, Huber JT, Kapoor WN. Quality of published reports of the prognosis of community-acquired pneumonia. J Gen Intern Med 9: 13–9, 1994. 14. Chanock RM. Mycoplasma infections of man. N Engl J Med 273: 1257–64 concl, 1965. 15. Dans PE, Charache P, Fahey M, Otter SE. Management of pneumonia in the prospective payment era. A need for more clinician and support service interaction. Arch Intern Med 144: 1392–7, 1984. 16.Dulake C, Selkon J. The incidence of pneumonia in the UK. Royal Society Medicine London Symposium Series: 87–94, 1982. 17. Everett MT. Major chest infection managed at home. Practitioner 227: 1743–54, 1983. 18. Fang GD, Fine M, Orloff J, Arisumi D, Yu VL, Kapoor W, Grayston JT, Wang SP, Kohler R, Muder RR, Yee YC, Rihs JD, Vickers RM. New and emerging etiologies for community-acquired pneumonia with implications for therapy. A prospective multicenter study of 359 cases. Medicine (Baltimore) 69: 307–16, 1990. 19. 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 44: 1031–5, 1989. 20. Finch RG, Woodhead MA. Practical considerations and guidelines for the management of community-acquired pneumonia. Drugs 55: 31–45, 1998. 21. Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, Coley CM, Marrie TJ, Kapoor WN. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 336: 243–50, 1997.


COMMUNITY-ACQUIRED PNEUMONIA
22. Fine MJ, Smith MA, Carson CA, Mutha SS, Sankey SS, Weissfeld LA, Kapoor WN. Prognosis and outcomes of patients with communityacquired pneumonia. A meta-analysis. JAMA 275: 134–41, 1996. 23. Fine MJ, Stone RA, Singer DE, Coley CM, Marrie TJ, Lave JR, Hough LJ, Obrosky DS, Schulz R, Ricci EM, Rogers JC, Kapoor WN. Processes and outcomes of care for patients withcommunity-acquired pneumonia: Results from the Pneumonia Patient Outcomes Research Team (PORT) cohort study. Arch Intern Med 159: 970–80, 1999. 24. Foy HM, Kenny GE, Cooney MK, Allan ID. Long-term epidemiology of infections with Mycoplasma pneumoniae. J Infect Dis 139: 681–7, 1979. 25. Foy HM, Kenny GE, McMahan R, Mansy AM, Grayston JT. Mycoplasma pneumoniae pneumonia in an urban area. Five years of surveillance. JAMA 214: 1666–72, 1970. 26. Gilbert K, Gleason PP, Singer DE, Marrie TJ, Coley CM, Obrosky DS, Lave JR, Kapoor WN, Fine MJ. Variations in antimicrobial use and cost in more than 2, 000 patients with community-acquired pneumonia. Am J Med 104: 17–27, 1998. 27. Grayston JT, Aldous MB, Easton A, Wang SP, Kuo CC, Campbell LA, Altman J. Evidence that Chlamydia pneumoniae causes pneumonia and bronchitis. J Infect Dis 168: 1231–5, 1993. 28. Hasley PB, Albaum MN, Li YH, Fuhrman CR, Britton CA, Marrie TJ, Singer, DE, Coley CM, Kapoor WN, Fine MJ. Do pulmonary radiographic findings at presentation predict mortality in patients with community-acquired pneumonia? Arch Intern Med 156: 2206–12, 1996. 29. Hofmann J, Cetron MS, Farley MM, Baughman WS, Facklam RR, Elliott JA, Deaver KA, Breiman RF. The prevalence of drug-resistant Streptococcus pneumoniae in Atlanta. N Engl J Med 333: 481–6, 1995. 30. Jokinen C, Heiskanen L, Juvonen H, Kallinen S, Karkola K, Korppi M, Kurki S, Ronnberg PR, Seppa A, Soimakallio S, et al. Incidence of community-acquired pneumonia in the population of fourmunicipalities in eastern Finland. Am J Epidemiol 137: 977–88, 31. Kleemola M, Saikku P, Visakorpi R, Wang SP, Grayston JT. Epidemics of pneumonia caused by TWAR, a new Chlamydia organism, in military trainees in Finland. J Infect Dis 157: 230–6, 1988. 32. Kuo CC, Jackson LA, Campbell LA, Grayston JT. Chlamydia pneumoniae (TWAR). Clin Microbiol Rev 8: 451–61, 1995. 33. Lentino JR, Lucks DA. Nonvalue of sputum culture in the management of lower respiratory tract infections. J Clin Microbiol 25: 758–62, 1987. 34. Lieberman D, Porath A. Seasonal variation in community-acquired pneumonia. Eur Resp J 9: 2630–4, 1996. 35. Lode H. Initial therapy in pneumonia. Clinical, radiographic, and laboratory data important for the choice. Am J Med 80: 70–4, 1986. 36. Macfarlane JT, Colville A, Guion A, Macfarlane RM, Rose DH. Prospective study of aetiology and outcome of adult lower- respiratory-tract infections in the community. Lancet 341: 511–4, 1993. 37. Macfarlane JT, Miller AC, Roderick Smith WH, Morris AH, Rose DH. Comparative radiographic features of community acquired Legionnaires’ disease, pneumococcal pneumonia, mycoplasma pneumonia, and psittacosis. Thorax 39: 28–33, 1984. 38. Marrie TJ, Peeling RW, Fine MJ, Singer DE, Coley CM, Kapoor WN. Ambulatory patients with community-acquired pneumonia: The frequency of atypical agents and clinical course. Am J Med 101: 508–15, 1996. 39. Marston BJ, Plouffe JF, File TM Jr, Hackman BA, Salstrom SJ, Lipman HB, Kolczak MS, Breiman RF. 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 157: 1709–18, 1997. Martin CM, Kunin CM, Gottlieb LS. Asian Influenza A in Boston 1957–58. Arch Intern Med 103: 515, 1999. Mayaud C. Asthma and Chlamydia pneumoniae. A future prospect for macrolides in general and roxithromycin in particular? Presse Med 26 Suppl 2: 27–9, 1997. Melbye H, Berdal BP, Straume B, Russell H, Vorland L, Thacker WL. Pneumonia—a clinical or radiographic diagnosis? Etiology and clinical features of lower respiratory tract infection in adults in general practice. Scand J Infect Dis 24: 647–55, 1992. Niederman MS, Bass JB Jr, Campbell GD, Fein AM, Grossman RF, Mandell LA, Marrie TJ, Sarosi GA, Torres A, Yu VL. Guidelines for the initial management of adults with community- acquired pneumonia: Diagnosis, assessment of severity, and initial antimicrobial therapy. American Thoracic Society. Medical Section of the American Lung Association. Am Rev Resp Dis 148: 1418–26, 1993. Palmer DL, Jones CC. Diagnosis of pneumococcal pneumonia. Semin Respir Infect 3: 131–9, 1988. Pollock HM, Hawkins EL, Bonner JR, Sparkman T, Bass JB Jr. Diagnosis of bacterial pulmonary infections with quantitative protected catheter cultures obtained during bronchoscopy. J Clin Microbiol 17: 255–9, 1983. Reed WW, Byrd GS, Gates RH Jr,Howard RS, Weaver MJ. Sputum gram’s stain in community-acquired pneumococcal pneumonia. A meta-analysis. West J Med 165: 197–204, 1996. Rein MF, Gwaltney JM Jr, O’Brien WM, Jennings RH, Mandell GL. Accuracy of Gram’s stain in identifying pneumococci in sputum. JAMA 239: 2671–3, 1978. Ries K, Levison ME, Kaye D. Transtracheal aspiration in pulmonary infection. Arch Intern Med 133: 453–8, 1974. Rodnick JE, Gude JK. Diagnosis and antibiotic treatment of community-acquired pneumonia. West J Med 154: 405–9, 1991. Saikku P. The epidemiology and significance of Chlamydia pneumoniae. J Infect 25(Suppl 1): 27–34, 1992. Shaw AB, Fry J. Acute infections of the chest in general practice. Br Med J 2: 1577–86, 1955. Taylor-Robinson D. Infections due to species of Mycoplasma and Ureaplasma: An update. Clin Infect Dis 23: 671–82; quiz 683–4, 1996. Tuddenham WJ. Glossary of terms for thoracic radiology: Recommendations of the Nomenclature Committee of the Fleischner Society. Am J Roentgenol 143: 509–17, Wong LK, Barry AL, Horgan SM. Comparison of six different criteria for judging the acceptability of sputum specimens. J Clin Microbiol 16: 627–31, 1982. Woodhead MA, Macfarlane JT, McCracken JS, Rose DH, Finch RG. Prospective study of the aetiology and outcome of pneumonia in the community. Lancet 1: 671–4, 1987. Wust J, Huf E, Kayser FH. Antimicrobial susceptibilities and serotypes of invasive Streptococcus pneumoniae strains in Switzerland. J Clin Microbiol 33: 3159–63, 1995.