Community-acquired pneumonia is a significant cause of morbidity and mortality, with the most common aetiological agent in virtually all studies being Streptococcus pneumoniae[1,2]. A definite diagnosis of pneumococcal pneumonia requires the isolation of S. pneumoniae from blood, pleural fluid effusions, trans-tracheal aspirate, trans-thoracic aspirate or samples obtained by protected brush. However, blood and pleural fluid cultures are positive in only 30% of cases, and invasive samples are not routinely or indiscriminately performed because they require trained personnel and may have adverse effects.
According to the guidelines of the Infectious Diseases Society of America  and the American Thoracic Society , any significant pleural effusion should be sampled to rule out the possibility of empyema or complicated parapneumonic effusion. Pleural fluid examination should include Gram's stain and bacterial culture. However, the diagnostic accuracy of these techniques may be reduced by various factors, including preceding antibiotic therapy. A rapid and accurate detection method for pneumococcal antigen in pleural fluid effusions would therefore be most useful. Polysaccharide capsular antigen detection methods, i.e., agglutination tests and counter-immunoelectrophoresis (CIE), have been used for antigen detection in respiratory and other samples , including pleural effusions . An immunochromatographic test (ICT) (NOW Streptococcus pneumoniae Antigen Test; Binax, Portland, ME, USA) has been developed to detect C polysaccharide antigen (PnC) in urine samples [5,6]. The aim of the present study was to determine the utility of the ICT for the detection of S. pneumoniae-specific antigens in pleural effusions.
In total, 91 pleural fluid samples were included in the study. Patients and samples were classified into four groups: group 1 included 19 parapneumonic pleural fluids from 16 patients (nine male and seven female; mean age 54.26 years, range 18–91 years) who were diagnosed with pneumococcal pneumonia; group 2 included seven parapneumonic pleural effusions from six patients (one male and five female; mean age 45.83 years, range 31–69 years) with non-pneumococcal pneumonia aetiologies (one Pseudomonas aeruginosa, one Morganella morganii and four pleural tuberculosis); group 3 included 24 pleural fluids from 21 patients (18 male and three female; mean age 57.33 years, range 30–79 years) with pneumonia, but with no identified pathogen (four patients subsequently developed empyema, yielding two Eschericha coli and two Peptostreptococcus spp.); and group 4 included 41 pleural effusions from 40 patients (25 male and 15 female; mean age 66.62 years, range 34–88 years) with a non-infectious aetiology (11 neoplasia, 17 secondary to heart failure, six idiopathic pleural effusions, one cirrhosis, two thoracic ascitis, and four secondary to thoracic trauma).
S. pneumoniae was isolated from blood culture with the BactAlert system (bioMérieux, Marcy-L'Etoile, France). Microbiological examination of the pleural effusions was performed by Gram's stain and bacterial culture, with identification based on standard microbiological criteria . Pneumococcal antigen in pleural effusion was detected by CIE using a pneumococcal omniserum (Staten Serum Institut, Copenhagen, Denmark) that includes antibodies to 84 serotypes. The technique was performed using a modification of a procedure described previously . Urinary pneumococcal antigen was detected by the ICT assay using concentrated urine according to the instructions of the manufacturer. Pleural fluid specimens were collected and frozen at − 20°C until required, and were then thawed immediately before being tested by the ICT assay. Pleural fluids were diluted 1 : 4 in EDTA 4% w/v, boiled for 5 min, and then centrifuged at 3000 g for 15 min in order to minimise possible non-specific reactions. The ICT was then used to detect PnC in the supernatant of pleural fluid specimens. The results were read visually after 15 min by two independent trained investigators.
Table 1 summarises the results obtained for the group 1 samples (i.e., patients who had been diagnosed with pneumococcal pneumonia by microbiological methods).
|No. of patients||Blood culture||Urinary antigen detection by ICT||Pleural effusion|
detection by CIE
Overall results for the four groups are summarised in Table 2. Among the 19 patients with pneumococcal pneumonia, S. pneumoniae antigen was detected in 15 cases by ICT, and was positive for the only pleural effusion from which S. pneumoniae was isolated. The ICT was always negative in patients with non-pneumococcal pneumonia, but was positive in three cases with a non-infectious aetiology. The ICT was also positive for one of 24 pleural fluids from patients with pneumonia for which no pathogen had been isolated. Therefore, the overall sensitivity was 79% and the specificity was 93.6%.
|Group of patients||Number of positive results/total patients evaluated by ICT (%)|
|Pneumococcal pneumonia||15/19 (79)|
|Pneumonia with another aetiology||0/7 (0)|
|Pneumonia with no pathogen identified||1/24 (4.2)|
|Non-infectious aetiology||3/41 (7.3)|
Detection of pneumococcal antigen in urine by the ICT test has proven useful in increasing the number of diagnoses of pneumonia in adults [5,6]. Nevertheless, there are several situations in which the detection of urinary antigen is difficult. Obtaining large volumes of urine is not possible for some patients, such as those who have oliguria secondary to pneumonia, acute or chronic kidney failure or altered mental status, and psychiatric patients. In addition, up to 20% of results obtained with urine specimens may be false-negative .
Although the ICT was developed to detect pneumococcal antigen in urine samples, it has been adapted to other specimens, such as nasopharyngeal samples , otitis media effusions , and cerebrospinal fluid samples for the diagnosis of pneumococcal meningitis , with high sensitivity and specificity. Several studies have examined the possibility of using other methods to detect pneumococcal antigen in pleural effusions. Lampe et al. used CIE with pleural fluid, and obtained favourable results in comparison with conventional methods, while Boersma et al. obtained positive results in eight of nine patients diagnosed with pneumococcal pneumonia using a latex agglutination method, with a specificity of 92%.
The present study demonstrates that the ICT is a sensitive and specific test for detecting pneumococcal antigen in pleural fluid samples. Antigen detection by ICT in pleural fluid is advantageous because results are obtained more rapidly than with culture. Furthermore, the test remains positive even if effective therapy has already been started. Although the ICT was developed to detect PnC from S. pneumoniae, it has been shown previously that it is capable of detecting capsular polysaccharide . Therefore, its sensitivity may vary according to the serotype responsible for the pneumonia. In the present study, only one serotype 3 isolate from a blood culture was available for testing, and this was detected efficiently.
The positive antigen result obtained in the group of patients in whom no pathogen was identified could be interpreted as a true-positive result, based on the specificity of the test. It was not possible in this case to obtain urine to perform antigen detection; therefore, detection of pneumococcal antigen in pleural fluid may increase the rate of diagnosis of pneumococcal pneumonia. The three false-positive results in the group of patients with a non-infectious aetiology are more difficult to explain. Two of these patients suffered heart failure and one had neoplasia. However, a concomitant lower respiratory tract infection was documented clinically in all three cases, and the possibility of a pneumococcal infection cannot be eliminated. It was therefore concluded, based on its sensitivity and speed, that the ICT may be a valuable tool for detecting pneumococcal antigen in pleural effusion samples and for the management of pneumonia.