Pneumococcal capsular antigen detection and pneumococcal serology in patients with community acquired pneumonia

Evolving Understanding of the Causes of Pneumonia in Adults, With Special Attention to the Role of Pneumococcus

Before 1945, Streptococcus pneumoniae caused more than 90% of cases of pneumonia in adults. After 1950, the proportion of pneumonia caused by pneumococcus began to decline. Pneumococcus has continued to decline; at present, this organism is identified in fewer than fewer10%-15% of cases. This proportion is higher in Europe, a finding likely related to differences in vaccination practices and smoking. Gram-negative bacilli, Staphylococcus aureus, Chlamydia, Mycoplasma, and Legionella are each identified in 2%-5% of patients with pneumonia who require hospitalization. Viruses are found in 25% of patients, up to one-third of these have bacterial coinfection. Recent studies fail to identify a causative organism in more than 50% of cases, which remains the most important challenge to understanding lower respiratory infection. Our findings have important implications for antibiotic stewardship and should be considered as new policies for empiric pneumonia management are developed.

Atypical aetiology in patients hospitalised with community-acquired pneumonia is associated with age, gender and season; a data-analysis on four Dutch cohorts

BACKGROUND

Microorganisms causing community-acquired pneumonia (CAP) can be categorised into viral, typical and atypical (Legionella species, Coxiella burnetii, Mycoplasma pneumoniae, and Chlamydia species). Extensive microbiological testing to identify the causative microorganism is not standardly recommended, and empiric treatment does not always cover atypical pathogens. In order to optimize epidemiologic knowledge of CAP and to improve empiric antibiotic choice, we investigated whether atypical microorganisms are associated with a particular season or with the patient characteristics age, gender, or chronic obstructive pulmonary disease (COPD).

METHODS

A data-analysis was performed on databases from four prospective studies, which all included adult patients hospitalised with CAP in the Netherlands (N = 980). All studies performed extensive microbiological testing.

RESULTS

A main causative agent was identified in 565/980 (57.7 %) patients. Of these, 117 (20.7 %) were atypical microorganisms. This percentage was 40.4 % (57/141) during the non-respiratory season (week 20 to week 39, early May to early October), and 67.2 % (41/61) for patients under the age of 60 during this season. Factors that were associated with atypical causative agents were: CAP acquired in the non-respiratory season (odds ratio (OR) 4.3, 95 % CI 2.68-6.84), age <60 year (OR 2.9, 95 % CI 1.83-4.66), male gender (OR 1.7, 95 % CI 1.06-2.71) and absence of COPD (OR 0.2, 95 % CI 0.12-0.52).

CONCLUSIONS

Atypical causative agents in CAP are associated with respectively non-respiratory season, age <60 years, male gender and absence of COPD. Therefore, to maximise its yield, extensive microbiological testing should be considered in patients <60 years old who are admitted with CAP from early May to early October.

TRIAL REGISTRATION

NCT00471640 , NCT00170196 (numbers of original studies).

Results of a cohort model analysis of the cost-effectiveness of routine immunization with 13-valent pneumococcal conjugate vaccine of those aged > or =65 years in the Netherlands

BACKGROUND

Community-acquired pneumonia and invasive pneumococcal disease are common among older people (ie, those aged > or =65 years). A new 13-valent pneumococcal conjugate vaccine (PCV-13) is under study in the Netherlands.

OBJECTIVE

The aim of this work was to model the cost-effectiveness of PCV-13 vaccination among those aged > or =65 years in the Netherlands, both in the total population and in those at increased risk for pneumonia, for various levels of efficacy (30%-90%) assumed.

METHODS

Our previously published cost-effectiveness model was updated to include age-specific epidemiologic data and health-care utilization and costs for a hypothetical cohort of adults aged > or =65 years in the Netherlands. This cohort was followed twice-once as unvaccinated and once as vaccinated-over a time period of 5 years, with differences between both analyses reported. Outcome measures included costs, life-years gained (LYGs), quality-adjusted life-years, and incremental cost-effectiveness ratios (ICERs). All analyses were performed from a societal perspective.

RESULTS

In the model, the ICER for vaccination remained below euro80,000/LYG, except when the vaccine was assumed to protect only against bacteremic pneumonia, with a relatively low effectiveness (40%) in combination with a high vaccine price (euro65), and indirect effects of serotype replacement would largely offset the direct effect of vaccination. For various assumptions, introduction of widespread PCV-13 vaccination (assuming a 60% efficacy against invasive and noninvasive disease because of vaccine serotypes, and a cost of euro50 per vaccinated person) was associated with the ICERs varying from cost-saving to euro50,676/LYG.

CONCLUSIONS

In this model analysis of a hypothetical cohort in the Netherlands, vaccination with PCV-13 might be considered cost-effective, both for the total population and for the high-risk population aged > or =65 years, from a societal perspective, over a 5-year time horizon. The main limitation of this study was uncertainty regarding how great a proportion of pneumonia could be attributed to pneumococcal disease.

Evaluating the Use of a Streptococcus pneumoniae Urinary Antigen Detection Kit for the Management of Community-Acquired Pneumonia in Japan

BACKGROUND

The urinary antigen detection kit for Streptococcus pneumoniae was tested.

OBJECTIVES

It was our aim to evaluate the usefulness of the immunochromatographic membrane test by doing a large prospective study of community-acquired pneumonia (CAP) in Japan.

METHODS

We prospectively evaluated the use of the S. pneumoniae urinary antigen detection kit and analyzed the treatment and clinical effect seen in patients with positive test kit results. One hundred and fifty-six patients with CAP admitted to our hospital between October 2001 and September 2003 were evaluated.

RESULTS

In 49% of these CAP patients, the causative microorganisms were isolated. S. pneumoniae was suspected to be the causative microorganism in 15%, but positive results of the urinary antigen detection kit indicated S. pneumoniae to be a probable microorganism in 28%, even though antibiotics had previously been administered to half of the patients. The kit was particularly useful for diagnosing patients with poor quality sputum in whom antibiotics treatment nevertheless had to be selected. Antibiotics appropriate for S. pneumoniae (mainly penicillin) were given. The treatment was found to have excellent clinical results in 89% of the CAP patients.

CONCLUSIONS

The S. pneumoniae urinary antigen detection kit was considered to be useful in selecting treatment since there was a high level of clinical effectiveness when the most suitable antibiotics were immediately administered to positive patients. The use of the S. pneumoniae urinary antigen kit is rapid and simple compared with conventional microbiological procedures.

Diagnostic value of C reactive protein in infections of the lower respiratory tract: systematic review

OBJECTIVES

To evaluate the diagnostic accuracy of C reactive protein in detecting radiologically proved pneumonia and to evaluate how well it can discriminate between bacterial and viral infections of the lower respiratory tract.

DATA SOURCES

Medline and Embase (January 1966 to April 2004), with reference checking.

STUDY SELECTION

We included articles comparing C reactive protein with a chest radiograph or with microbiological work-up as a reference test. Two authors independently assessed methodological items.

RESULTS

None of the studies met all validity criteria. Six studies used an infiltrate on chest radiograph as reference test. Sensitivities ranged from 10% to 98%, specificities from 44% to 99%. For adults, the relation of C reactive protein with an infiltrate (in a subgroup analysis of five studies) showed an area under the curve of 0.80 (95% confidence interval 0.75 to 0.85). In 12 studies, the relation of C reactive protein with a bacterial aetiology of infection of the lower respiratory tract was studied. Sensitivities ranged from 8% to 99%, specificities from 27% to 95%. These data were epidemiologically and statistically heterogeneous, so overall outcomes could not be calculated.

CONCLUSION

Testing for C reactive protein is neither sufficiently sensitive to rule out nor sufficiently specific to rule in an infiltrate on chest radiograph and bacterial aetiology of lower respiratory tract infection. The methodological quality of the diagnostic studies is generally poor. The evidence not consistently and sufficiently supports a wide introduction of C reactive protein as a rapid test to guide antibiotics prescription.

Diagnosis of parapneumonic pleural effusion by polymerase chain reaction in children

PURPOSE

Most pleural effusions are associated with bacterial pneumonia, and the identification of the pathogen will assist the therapeutic decision. A specific method that is not affected by previous antibiotic therapy is sought to detect the main causative agents of pneumonia in infants and children (Streptococcus pneumoniae, Haemophilus influenzae, and Staphylococcus aureus). The aim of the present study was to compare the polymerase chain reaction (PCR) technique with standard culture methods in identifying bacterial infections in infants' and children's pleural effusion.

METHODS

Samples obtained from pediatric patients (n = 37) with a diagnosis of pneumonia associated to pleural effusion, submitted to thoracentesis, were analyzed by PCR with specific primers.

RESULTS

The PCR technique identified the presence of bacterial infection in a larger proportion (95.2%) than the standard culture method (33.3%) on complicated pleural effusion samples. The microorganism detection on uncomplicated pleural effusion samples was positive only by the PCR method (31.3%). The frequencies of microorganisms identified on complicated pleural effusion were 57.1% of all patients for methicillin-resistant Staphylococcus; 52.4%, S pneumoniae; 28.6%, S aureus; and 23.8%, H influenzae. The previous use of antibiotics interferes with standard culture method, but it did not interfere with the PCR results.

CONCLUSIONS

The molecular diagnosis by PCR method could improve the etiologic diagnosis and might help to guide the treatment of parapneumonic effusion in children.

Evaluation of a rapid immunochromatographic test for detection of Streptococcus pneumoniae antigen in urine samples from adults with community-acquired pneumonia

Streptococcus pneumoniae is the most common cause of community-acquired pneumonia but is undoubtedly underdiagnosed. Isolation of S. pneumoniae from blood is specific but lacks sensitivity, while isolation of S. pneumoniae from sputum may represent colonization. We evaluated a new immunochromatographic test (NOW S. pneumoniae urinary antigen test; Binax, Portland, Maine) that is simple to perform and that can detect S. pneumoniae antigen in urine within 15 min. Urine samples from 420 adults with community-acquired pneumonia and 169 control patients who did not have pneumonia were tested. Urine from 315 (75%) of the pneumonia patients and all controls was tested both before and after 25-fold concentration, while the remaining 105 samples were only tested without concentration. S. pneumoniae urinary antigen tests were positive for 120 (29%) patients with pneumonia and for none of the controls. Of the urine samples tested with and without concentration, 96 were positive, of which 6 were positive only after concentration. S. pneumoniae antigen was detected in the urine from 16 of the 20 (80%) patients with blood cultures positive for S. pneumoniae and from 28 of the 54 (52%) patients with sputum cultures positive for S. pneumoniae. The absence of S. pneumoniae antigen in the urine from controls suggests that the specificity is high. Concentration of urine prior to testing resulted in a small increase in yield. The NOW S. pneumoniae urinary antigen test should be a useful adjunct to culture for determining the etiology of community-acquired pneumonia in adults.

Diagnostic testing for community-acquired pneumonia

The microbial cause of community-acquired pneumonia can be identified by noninvasive means in the majority of cases, usually within a few days of presentation. The Gram stain and culture of a pretreatment sputum sample are the most useful tests, but have significant limitations. Methods for detecting pneumococcal antigen in respiratory secretions are particularly helpful in patients who have received antibiotics before evaluation. Testing for specific pathogens such as L. pneumophila, M. pneumoniae, or C. pneumoniae should be guided by clinical suspicion in individual circumstances. Invasive procedures are most helpful in patients suspected of having infection with opportunistic or resistant pathogens, and in those whose initial management has been unsuccessful.

Severe community-acquired pneumonia

Severe CAP is a life-threatening condition defined by the presence of respiratory failure or symptoms of severe sepsis or septic shock. It accounts for approximately 10% of hospitalized patients with CAP. The majority of patients with severe pneumonia have underlying comorbid illnesses, with COPD, alcoholism, chronic heart disease, and diabetes mellitus being the most frequent. S. pneumoniae, Legionella spp, GNEB (especially K. pneumoniae), H. influenzae, S. aureus/spp, Mycoplasma pneumoniae, respiratory viruses (especially influenza viruses), and P. aeruginosa represent the most important causative organisms of severe CAP. Rapid initiation of appropriate antimicrobial treatment is crucial for a favorable outcome. Initial antimicrobial treatment should be based on an epidemiological (empiric) approach. Microbial investigation may be helpful in the individual case but is probably more useful to define local antimicrobial policies based on local epidemiologic and susceptibility patterns. Mortality rates range from 21% to 54%. The most important prognostic factors include general health state of the patient, appropriateness of initial antimicrobial treatment, and the existence of bacteremia, as well as factors reflecting severe respiratory failure, severe sepsis, septic hypotension or shock, and the extent of infiltrates in chest radiograph. Initial antimicrobial treatment should consist of a second (or third) generation cephalosporin and erythromycin. Modifications of this basic regimen should be considered in the presence of distinct comorbid conditions and risk factors for distinct pathogens. Promising new approaches of nonantimicrobial treatment, including noninvasive ventilation, treatment of hypoxemia, and immunomodulation, are under investigation.

Value of the polymerase chain reaction assay in noninvasive respiratory samples for diagnosis of community-acquired pneumonia

We studied the causes of community-acquired pneumonia (CAP) in 184 patients. Microbiologic evaluation included sputum examination, blood culture, assessment of acute and convalescent antibody titers for Legionella pneumophila, Mycoplasma pneumoniae, Chlamydia pneumoniae, Coxiella psitacci, Coxiella burnetii and respiratory viruses, polymerase chain reaction (PCR) assays for M. pneumoniae and C. pneumoniae in throat swab, and PCR assay based on the amplification of pneumolysin gene fragment in sera. The causative pathogen was identified in 78 patients (Streptococcus pneumoniae, 44; M. pneumoniae, 26; C. pneumoniae, 1; others, 7). S. pneumoniae was detected in serum by the PCR assay in 41 patients, five of whom also had a positive blood culture. PCR assay was negative in two patients with positive blood culture for S. pneumoniae. C. pneumoniae was detected by PCR in nine patients, but only one showed seroconversion. M. pneumoniae was detected by PCR in only three patients (two without seroconversion). The diagnosis of pneumonia caused by S. pneumoniae was five times greater using PCR in serum than with blood culture. Detection of C. pneumoniae by PCR without fulfilling criteria for acute infection may be considered a prior infection. The PCR assay for the diagnosis of M. pneumoniae has a lower sensitivity than serologic methods.

Is Streptococcus pneumoniae the leading cause of pneumonia of unknown etiology? A microbiologic study of lung aspirates in consecutive patients with community-acquired pneumonia

PURPOSE

Although a wide variety of recognized pathogens can cause community-acquired pneumonia, in many patients the etiology remains unknown after routine diagnostic workup. The aim of this study was to identify the causal agent in these patients by obtaining lung aspirates with transthoracic needle aspiration.

SUBJECTS AND METHODS

During a 15-month period, all consecutive patients with community-acquired pneumonia who were eligible for transthoracic needle aspiration were enrolled in the study. In addition to conventional microbial methods (culture of blood and sputum, serologic studies), we performed cultures and genetic and antigen tests for common respiratory pathogens in lung aspirates.

RESULTS

The study group consisted of 109 patients. Conventional microbial studies identified an etiology in 54 patients (50%), including Mycoplasma pneumoniae in 19 patients, Chlamydia pneumoniae in 9 patients, and Streptococcus pneumoniae in 9 patients. Among the remaining 55 patients, study of the lung aspiration provided evidence of the causal agent in 36 (65%). In 4 additional patients with a single microbial diagnosis by conventional methods, the lung sample provided evidence of an additional microorganism. The new pathogens detected by lung aspiration were S. pneumoniae in 18 patients, Haemophilus influenzae in 6 patients, Pneumocystis carinii in 4 patients, and C. pneumoniae in 3 patients; other organisms were identified in 4 patients.

CONCLUSIONS

In our study, S. pneumoniae was the leading cause of community-acquired pneumonia, accounting for 25% of all cases, including about one-third of the cases the cause of which could not be ascertained with routine diagnostic methods.

Pneumococcal community-acquired pneumonia in 148 hospitalized adult patients

In a previous prospective study, Streptococcus pneumoniae was identified as the causative agent in 148 (42.8%) of 346 adult patients hospitalized over the course of one year with community-acquired pneumonia (CAP) in the Soroka Medical Center, Beer-Sheva, Israel. The present study characterizes those cases in which Streptococcus pneumoniae was the only pathogen and those in which additional etiological agents were identified. Pneumococcal CAP was diagnosed by standard blood cultures or positive serological tests by one of two laboratory methods. In 100 (67.6%) patients, at least one other etiological agent of CAP was identified in addition to Streptococcus pneumoniae. Compared with patients who were not infected by Streptococcus pneumoniae, patients with Streptococcus pneumoniae CAP were older and had a higher rate of comorbidity (39.5% vs. 29.8%). Streptococcus pneumoniae CAP had a more severe clinical course and a higher mortality rate, especially when Streptococcus pneumoniae was the only pathogen. Community-acquired pneumonia due to Streptococcus pneumoniae only was more similar in its clinical manifestations to classic typical pneumococcal pneumonia. When an additional etiological agent was identified, the clinical characteristics could not be distinguished from those of atypical pneumonia. It is concluded that Streptococcus pneumoniae remains the principal cause of CAP in this region. The frequency of additional etiological agents of CAP and the difficulty in differentiating clinically between cases due to Streptococcus pneumoniae only and those due to Streptococcus pneumoniae plus other organisms necessitates initial empirical treatment that covers Streptococcus pneumoniae as well as other causative agents of atypical pneumonia.

Rapid detection of pneumococcal antigen in lung aspirates: comparison with culture and PCR technique

Detection of pneumococcal antigen has been used to increase the rate of diagnosis of pneumococcal pneumonia. The present study was designed to determine the value of rapid detection of pneumococcal antigen in samples obtained by transthoracic needle aspiration (TNA) from patients with community-acquired pneumonia (CAP) in a comparative analysis with culture and polymerase chain reaction (PCR). Pneumococcal antigen was detected by latex agglutination. One hundred and ten consecutive patients diagnosed with CAP underwent TNA. Patients were grouped, according to PCR, culture and serological results, into pneumococcal pneumonia (n = 18), other known aetiology (n = 67) and unknown aetiology (n = 25). In patients with pneumococcal pneumonia, antigen was detected in 17 (94.4%) cases. Antigen was detected in one and nine patients with pneumonia of other known or unknown aetiologies, respectively, yielding a specificity of 89.1%. In conclusion, detection of pneumococcal antigen on samples obtained by TNA from patients with CAP provides a sensitive and specific diagnosis of Streptococcus pneumoniae infection. Furthermore, its rapid results would reduce the dependence on empirical treatments.

Pneumococcal antigen detection in bronchoalveolar lavage fluid from patients with pneumonia

BACKGROUND

Pneumococcal pneumonia can be diagnosed by the detection of capsular antigen in sputum, serum, pleural fluid, or urine using countercurrent immunoelectrophoresis and latex agglutination. In addition, quantitative cultures of bronchoalveolar lavage (BAL) fluid are also reliable for establishing the aetiology of pneumonia. This study investigated the value of rapid detection of pneumococcal antigen in BAL fluid from patients with pneumonia.

METHODS

Pneumococcal antigen was detected by countercurrent immunoelectrophoresis and latex agglutination. Patients were grouped according to BAL quantitative culture results into pneumococcal pneumonia (n = 24), other known aetiology (n = 18), and unknown aetiology (n = 17). Thirteen patients with interstitial lung disease and without pneumonia served as a control group.

RESULTS

In patients with pneumococcal pneumonia, antigen was detected by countercurrent immunoelectrophoresis in 50% and by latex agglutination in 54% of cases. In patients with pneumonia of unknown aetiology pneumococcal antigen was detected by latex agglutination in 53% of cases. Antigen was not detected in patients with pneumonia of other known aetiology or in control patients, yielding a specificity of 100%.

CONCLUSIONS

In patients with pneumococcal pneumonia requiring fibreoptic bronchoscopy detection of pneumococcal antigen in BAL fluid may rapidly and accurately confirm the aetiology. Furthermore, in nearly half the cases of pneumonia of unknown aetiology antigen can be detected, suggesting that Streptococcus pneumoniae is a major causative agent in such patients.

Demonstration of circulating pneumococcal immunoglobulin G immune complexes in patients with community-acquired pneumonia by means of an enzyme-linked immunosorbent assay

An enzyme-linked immunosorbent assay was developed for quantitation of circulating immune complexes (CICs) containing specific antipneumococcal immunoglobulin G (IgG). These CICs were detected in 17 (85%) of 20 patients with bacteremic pneumococcal pneumonia, 4 (36.4%) of 11 patients with probable pneumococcal pneumonia, 3 (16.7%) of 18 patients with pneumonia of other (nonpneumococcal) etiology, and 13 (41.9%) of 31 patients with pneumonia of unknown etiology. There was no correlation between CICs and serum IgG antibody levels. Pneumococcal capsular antigen was demonstrated in dissociated CICs by latex agglutination.

Antigen detection in oropharyngeal secretions for rapid diagnosis of pneumococcal pneumonia

To determine the value of detection of antigen in the oropharynx in the diagnosis of pneumococcal pneumonia, oropharyngeal secretions were cultured for the presence of Streptococcus pneumoniae and tested for the presence of pneumococcal antigen. Sputum (if available) collected on the same day was also investigated for the presence of antigen. Detection of pneumococcal antigen was found to be directly related to the severity of pneumococcal carriership or infection (p < 0.0001) and was not related to culture results. Patients with pneumococcal pneumonia had the highest antigen detection rate (38%), followed by patients with pneumonia of unknown etiology (32%) and patients with an acute lower respiratory tract infection due to Streptococcus pneumoniae (20%). Pneumococcal carriers had a detection rate of only 9%. Antigen could be detected in only one patient of the control groups. Although antigen detection in sputum was superior to that in oropharyngeal secretions, concordant results were obtained in 8 (40%) and 6 (36%) patients with pneumococcal pneumonia and pneumonia of unknown etiology respectively. The results strongly suggest that pneumococcal carriage seldom leads to a detectable level of antigen, and that antigen detection in the oropharynx appears to be of additive value in the diagnosis of pneumococcal pneumonia.

Rapid detection of pneumococcal antigen in pleural fluid of patients with community acquired pneumonia

BACKGROUND

Detection of pneumococcal antigen may help to increase the rate of diagnosis of pneumococcal pneumonia. This study was designed to determine the value of rapid detection of pneumococcal antigen in pleural fluid from patients with community acquired pneumonia.

METHODS

Thoracentesis was performed in patients suspected of having empyema and in patients with pneumonia of unknown aetiology. Pneumococcal capsular antigen was detected by latex agglutination and this method was compared with Gram staining and culture, specimens of pleural fluid being examined in parallel by the three methods.

RESULTS

Pleural fluid was radiographically identified in 63 of 135 patients with community acquired pneumonia. In nine of 45 patients with pneumococcal pneumonia and pleural fluid pneumococci were identified by Gram stain in two and by culture in one specimen of pleural fluid, whereas antigen was detected in eight of these specimens. In 12 of 33 patients with pneumonia of other known aetiology only one pleural fluid specimen was antigen positive, providing a specificity of 92% for this test. Pleural fluid obtained from 12 of 58 patients with pneumonia of unknown aetiology yielded detectable antigen in seven cases.

CONCLUSIONS

Detection of pneumococcal antigen by latex agglutination in pleural fluid may yield important and rapid information in patients with community acquired pneumonia.

Detection of pneumococcal capsular antigen in the presence of penicillin in vitro

Eight strains of Streptococcus pneumoniae were tested in vitro for their ability to produce capsular antigen in the presence of penicillin. It was found that, provided 10(6) to 10(7) pneumococci/ml were present, capsular antigen could be detected during the 72 h in which the experiment was conducted, irrespective of whether penicillin was added at 0 h or 8 h, and even when no viable pneumococci remained. When fewer pneumococci were present, capsular antigen could not be detected at any time in the presence of penicillin. Control cultures, without penicillin, yielded detectable capsular antigen only when the threshold value of 10(6)-10(7) pneumococci/ml was reached. It is concluded that the presence of penicillin does not influence the detection of pneumococcal capsular antigen, but demonstration of this antigen is totally dependent on the number of pneumococci present.

Pneumococcal antigen persistence in sputum from patients with community-acquired pneumonia

The purpose of this study was to establish the diagnostic value of pneumococcal capsular antigen by comparing this with the results of Gram stain and culture in representative and nonrepresentative sputa during follow-up in patients with community-acquired pneumonia. Antigen was detected by a latex particle agglutination test. At the time of hospital admission, antigen was detected in 17 representative sputum specimens from 30 patients with pneumococcal pneumonia, which was comparable to the results of Gram stain and culture. In five additional patients, antigen was demonstrated in nonrepresentative specimens. During follow-up under antibiotic treatment, this number increased by six: three patients with representative and three patients with nonrepresentative sputum, respectively. Two of the 22 patients with pneumonia of other known cause had an antigen-positive sputum on admission and in another two patients, sputum antigen was detected during follow-up. Ten of 34 patients with pneumonia of unknown cause had detectable antigen in representative or nonrepresentative sputum on admission. During follow-up, antigen was detected in sputa of an additional seven patients. There was no difference in duration of antigen persistence between patients with pneumococcal pneumonia and pneumonia of unknown cause. It was observed that the first antigen-positive sputum specimen was always detected within the first five days of the hospital stay. We conclude that antigen detection in both representative and nonrepresentative sputum specimens at the time of hospital admission and during follow-up is of additional value for the diagnosis of pneumococcal pneumonia. It markedly increases the number of patients with pneumococcal pneumonia detected, who would otherwise be considered to have pneumonia of unknown cause. However, antigen-positive results should be interpreted carefully, especially in those pneumonia patients with chronic bronchitis, because detectable antigen may be caused by pneumococcal carriership of the lower respiratory tract.