Evaluating the Use of a Streptococcus pneumoniae Urinary Antigen Detection Kit for the Management of Community-Acquired Pneumonia in JapanKobashi Y. · Yoshida K. · Miyashita N. · Niki Y. · Matsushima T.
Division of Respiratory Diseases, Department of Medicine, Kawasaki Medical School, Kurashiki, 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.
Copyright © 2006 S. Karger AG, Basel
Streptococcus pneumoniae had been recognizedas the most common cause of community-acquired pneumonia (CAP) [1, 2]. However, there are many problems with respect to making a microbiological diagnosis of pneumococcal pneumonia since the causative microorganisms can only be isolated in 40–70% of cases, even though various types of specimens and several laboratorymethods have been investigated [3,4,5]. Since S. pneumoniae has an autolysin enzyme which makes it difficult to detect bacteria in the specimens that are obtained, many pneumococcal pneumonias in patients with CAP are diagnosed as being due to microorganisms of unknown origin.
The detection of S. pneumoniae antigens (usually capsular polysaccharides) in the urine of patients with pneumonia was first described in 1917 . It has been extensively studied using a variety of techniques, including counterimmunoelectrophoresis, latex agglutination, coagglutination, and for about the last 30 years, the enzyme-linked immunosorbent assay [7,8,9,10]. Recently, a new and rapid immunochromatographic membrane test (ICT), the Binax NOW S. pneumoniae urinary antigen assay, was developed for the detection of antigen of S. pneumoniae in urine samples . Unlike the other tests, this test is simple to perform. It detects the C-polysaccharide cell wall antigen common to all S. pneumoniae strains  and provides results within 15 min.
The purpose of this study was to evaluate the usefulness of the ICT by doing a large prospective study of CAP in Japan and to determine the relationship between positive ICT results and the clinical effects of treatment.
Patients and Methods
The subjects for this prospective study were recruited between October 2001 and September 2003 at Kawasaki Medical School Hospital, a 1,075-bed, university-affiliated teaching hospital that serves a population of 500,000 in Kurashiki, Japan. All patients over 15 years old with CAP admitted to our hospital were enrolled in this study. CAP was defined as an illness occurring in a patient presenting with new infiltrates on admission chest X-ray and at least two of the following: fever (axillary temperature measured with a clinical thermometer >38.0°C), production of purulent sputum, cough and leukocytosis (white blood cell count >10,000/μl). Patients were excluded from the study when abnormalities on chest X-ray were attributed to other causes, such as obstructive pneumonia due to lung cancer, congestive heart failure or pulmonary infarction. Clinical, radiological and laboratory data of all patients were recorded. To evaluate the severity of pneumonia, we used Fine’s score , which classifies patients with CAP into five stages according to outcome, with stage I including patients with the most favorable prognosis and stage V including those with the poorest prognosis. Informed consent was obtained from each patient at study entry approved by the Ethical Committee of Kawasaki Medical School.
Two sets of blood cultures (aerobic and anaerobic cultures) and sputum specimens for culture and Gram stain were collected at the discretion of the attending physician and were processed in the hospital microbiology laboratory. Urine samples for the detection of Legionella pneumophila urinary antigen and serum samples (obtained during the acute stage of illness and 2–4 weeks later) for serological testing were also collected at the discretion of the attending physician. All sputum samples obtained from patients with productive cough were transported to the microbiology laboratory in our hospital within 1 h. We evaluated only qualified samples, as defined by standard criteria: presence of >25 white blood cells and <10 squamous cells per low-power magnification field (×10). An immunochromatographic assay (Binax) was used to detect soluble L.pneumophila serogroup 1 antigen in the urine. A complement fixation test was performed to detect antibodies against Mycoplasma pneumoniae,Chlamydia pneumoniae,Coxiella burnetii, influenza viruses A and B, parainfluenza virus, respiratory syncytial virus and adenovirus. An indirect immunofluorescence test was used to detect antibodies against L. pneumophila, and a microimmunofluorescence test was used to detect antibodies against C. pneumoniae.
The following criteria were used to classify a case of pneumonia as being of known etiology: (1) for M. pneumoniae,Chlamydia psittaci, C. burnetii, influenza viruses A and B, parainfluenza virus, respiratory syncytial virus and adenovirus, a ≥4-fold increase in the antibody titer, as determined by the complement fixation test; (2) for C. pneumoniae, a 4-fold increase in immunoglobulin (Ig)G or the presence of IgM antibodies (≥20 fold); (3) for L. pneumophila, isolation of the organism from respiratory samples, Legionella antigen detected in the urine, or a ≥4-fold increase in the immunofluorescence antibody titer; (4) for S. pneumoniae,Haemophilus influenzae,Moraxella catarrhalis,Staphylococcus aureus and other bacteria, including Gram-negative enterobacteria, isolation of the definite pathogen from samples of blood or pleural fluid, or isolation of the probable pathogen when there was heavy growth (≥107 CFU/ml) on quantitative culture. Patients who did not fulfill the etiologic diagnostic criteria described above were considered to have pneumonia of unknown etiology.
The Binax NOW S. pneumoniae urinary antigen test (Binax; Portland, Me., USA) is a rapid diagnostic technique in a flow membrane format that relies on the detection of C-polysaccharide present in urine specimens. The ICT test was first performed on admission according to the manufacturer’s instructions, with a swab being dipped into the urine samples and then inserted into the test device in all patients. A citrate buffer was added to help in the flowing up of any antigen present in the samples. The conjugate rabbit anti-S. pneumoniae antibody binds pneumococcal antigen if it is present in the urine sample, and the resulting antigen/antibody-conjugated complexes are captured by immobilized rabbit anti-S. pneumoniae antibodies, forming the sample line. In addition, immobilized goat anti-rabbit IgG captures excess conjugated antibody, forming the control line. The results were read visually after 15 min. The test was interpreted by the presence or absence of visually detectable pink-to-purple colored lines. A result is positive if both the sample and control lines are detected and negative if only the control line is observed. The reading of the ICT results was blinded and performed by 2 independent persons.
The clinical results were judged by using the criteria for the effect of medications of CAP of the Japanese Society of Chemotherapy .
On the basis of the criteria described above, the final diagnosis was used as the standard for determining the diagnostic usefulness of the urine test in terms of sensitivity, specificity and positive and negative predictive values (since it is known that a negative ICT test result can convert to positive even during treatment, only the result of the first ICT test on admission was considered for the calculation of sensitivity). For a comparative study of two groups – those with S. pneumoniae-positive ICT results compared with all of the other patients with CAP – we used the χ2 or Fisher’s exact test to determine the proportions of qualitative variables and the Mann-Whitney U test or Student’s t test to compare quantitative variables. A p value <0.05 was considered to be statistically significant. The statistical precision of several test indices was determined by calculating the 95% confidence interval (CI).
Of the 168 consecutive patients who had symptoms and signs compatible with pneumonia, 12 were subsequently found not to have CAP (secondary infection with underlying respiratory disease 7, pulmonary nontuberculous mycobacterial disease 4, pulmonary tuberculosis 1), which left 156 patients. These 156 patients (110 males and 46 females) fulfilled the inclusion criteria and were studied in detail. The mean age (±SD) of the patients was 62.1 ± 11.4 years (ranging from 15 to 97 years). Seventeen patients had previously been bedridden.
Among the 156 patients, 125 (80%) had an underlying disease (table 1), 72 (46%) had a history of respiratory disease, with chronic obstructive pulmonary disease being the most common, 53 (34%) had a history of nonrespiratory disease, most commonly diabetes mellitus, and 31 had previously been in good health. Fifty-seven patients were current smokers, 49 were ex-smokers and 26 drank alcohol regularly. Eighteen patients (45%) had received corticosteroid treatment. Seventy patients (45%) had received antibiotics before admission. The distribution of patients with CAP according to Fine’s score was as follows: I, 36 (23%); II, 47 (30%); III, 34 (21%); IV, 32 (20%); V, 10 (6%). All 156 patients were admitted to the hospital, and 5 (3%) of them required admission to the intensive care unit.
|Table 1. Clinical characteristics of 156 patients with community-acquired pneumonia|
Causative microorganisms were isolated in 77 patients (49%); 5 of these patients showed evidence of polymicrobial agents, and in 5 patients, two infectious agents were present. The noninvasive microbiological tests that were used and the results obtained are listed in tables 2 and 3. Blood cultures were taken from all patients, and 7 samples yielded causative microorganisms. Satisfactory sputum samples were obtained from 114 patients, and 49 samples yielded causative microorganisms. The urinary antigen test for L. pneumophila was performed in 142 patients, and positive results for this test were obtained from 3 patients. Finally, serological examinations were performed in all patients, and in 22 patients, the results met the diagnostic criteria.
|Table 2. Results of diagnostic methods|
|Table 3. Causative microorganisms of community-acquired pneumonia|
The most common causative microorganisms, including the probable cases, were S. pneumoniae (containing penicillin-insusceptible S.pneumoniae or penicillin-resistant S. pneumoniae) found in 23 patients, followed by M. pneumoniae in 18, H. influenzae in 15, M.catarrhalis in 7, and C. pneumoniae in 5 patients (table 3).
The antigen detection kit yielded a positive reaction for 44 patients, and S.pneumoniae was isolated from 20 of these patients, 3 on blood culture and 17 on sputum culture. However, in the remaining 24 patients with a positive reaction, S. pneumoniae was not isolated from any of their specimens. On the other hand, 3 of the 23 patients whose cultures were positive for S. pneumoniae had negative results for S. pneumoniae with the ICT (table 4). Using positive blood or sputum culture results as the standard criteria for the diagnosis of pneumococcal pneumonia, the sensitivity of the urine test was 87% (95% CI 68–92), the specificity 82% (95% CI 63–88), the positive predictive value 67% (95% CI 45–82) and the negative predictive value 95% (95% CI 89–97).
|Table 4. Relationship between the results of the urinary antigen detection kit for S. pneumoniae and the blood and sputum culture in patients with community-acquired pneumonia|
Regarding the clinical characteristics of 44 patients with a positive urinary antigen test for S. pneumoniae, there were no significant differences between 44 patients with ICT positive results and 112 patients with ICT negative results.
With respect to the results of the other microbiological investigations of the 44 patients with a positive ICT reaction for S. pneumoniae, sputum samples were obtained from 34 patients; of these 34 samples, 25 samples met the standard criteria. Among these 25 patients, S. pneumoniae was isolated from 20 sputum cultures. In these 20 sputum-positive patients, blood cultures were positive in 3 patients. Six of these 20 patients (30%) had received antibiotics prior to admission. Among the 24 patients with a positive ICT test in whom S. pneumoniae was not isolated from any specimen, 16 patients (67%) had received antibiotics prior to admission. Concerning the relationship between the antimicrobial agents used and the clinical effects for all 44 patients with a positive ICT reaction for S. pneumoniae, 8 patients were treated with oral antimicrobial agents consisting of new quinolones (levofloxacin 6 and gatifloxacin 2), and all of these patients had a good clinical response. The remaining 36 patients received intravenous antimicrobial agents which have sensitivity to S. pneumoniae, consisting of penicillin (sulbactum/ampicillin 20, ampicillin 10) and carbapenem (panipenem/betamipron 3, meropenem 1, biapenem 1, imipenem/cilastatin 1). Most of these patients (89%) had a good clinical response. However, 3 patients died despite appropriate treatment; 2 of these patients had an underlying respiratory disease, such as lung fibrosis, old pulmonary tuberculosis or chronic obstructive pulmonary disease, and died of respiratory failure, and 1 had an underlying cardiovascular disease and died of heart failure. Furthermore, 3 patients (9%) needed mechanical ventilation and 2 of these patients died.
Recently, several reports have been published about the usefulness of ICT in increasing the rate of diagnosis of pneumococcal pneumonia in several western countries [11,15,16,17]. However, in Japan, there have only been a few reports on the usefulness of ICT [18,19,20]. In our study, testing for S. pneumoniae antigen in the urine from adults with CAP in many cases resulted in a rapid diagnosis of pneumococcal infection. When we compared the results of the ICT with those results attained via conventional diagnostic methods for pneumococcal pneumonia, the test had a sensitivity of 87% and a specificity of 82%.
Previous studies of ICT have shown variable results, with a sensitivity range of 52–91% [11, 15]. Murdoch et al.  detected S. pneumoniae antigen in 16 (80%) of 20 bacteremic patients admitted to hospital with CAP. In nonbacteremic patients, the sensitivity ranged between 52 and 78% [11, 15]. In our study, there were few bacteremic cases, and the sensitivity rates could be calculated as 100% (3 of 3 cases) in the bacteremic cases and 87% (20 of 23 cases) in the nonbacteremic cases. These are excellent results compared with those reported in previous studies. Smith et al.  reported that the ICT test is a valuable, sensitive and rapid test for the early diagnosis of bacteremia pneumococcal infections in adult patients, even after antibiotic treatment has commenced.
In this study, excluding the 20 cases in which S. pneumoniae was actually isolated, 24 cases (15%) with pneumonia of unknown etiology had a positive ICT result, with no other pathogen detected in an extensive, noninvasive microbiological workup which included optimal rapid transport of sputum specimens from the hospital to the microbiology laboratory. The reason for this lack of positive cultures appears to be the significant difference in the frequency of antibiotic use prior to admission between those patients in whom S. pneumoniae was not isolated (18/24, 67%) and those in whom S. pneumoniae was isolated (6/20, 30%). A comparison of the clinical characteristics of the 44 patients suspected of having pneumococcal pneumonia with the clinical characteristics of all patients with CAP during the treatment period showed no significant differences between the groups. However, the good clinical effect seen in 91% of the 44 patients with positive ICT results in whom appropriate antibiotics such as penicillin, carbapenem or new respiratory quinolone were then used demonstrates the clinical reliability of the ICT results for S. pneumoniae. The 3 remaining patients died of respiratory or heart failure due to severe underlying diseases despite the appropriate administration of highly effective antibiotics. We believe that there were many patients with positive ICT results who had no isolation of S. pneumoniae from various clinical specimens: (1) antibiotics such as cephem, macrolide and the new quinolones had already been administered prior to admission to half of the patients with CAP; (2) there were cross-reactions between Streptococcus mitis and S. pneumoniae, which have similar gene and biochemical characteristics, and this could have caused a false-positive ICT test result ; (3) there was a time delay between specimen collection and the performance of the microbiological studies which could have decreased the yield of the isolation. Nevertheless, with the ICT included as one of the supportive microbiological examinations for determining the causative microorganisms of CAP, the percentage of patients in whom an etiologic diagnosis could be made in our study increased from 15 to 28%.
There were 3 false-negative ICT results among patients with pneumococcal pneumonia diagnosed by isolation (≥107 CFU/ml) from qualified sputum specimens. In these cases, it may be suspected that S. pneumoniae was possibly present due to colonization of the oropharynx . Additionally, there is the opinion that a sequential approach of the ICT test is recommended for those for whom demonstrative results of sputum Gram staining are unavailable through comparison of the results of the ICT test with the results of sputum Gram staining. They emphasize that sputum Gram staining was initially used to demonstrate early diagnosis of CAP due to S. pneumoniae . In 1 case, the urine sample yielded a weekly positive reading, but we had labeled the ICT as a negative result. Stralin et al.  reported that while weak ICT positivity should be interpreted with caution, strong ICT positivity should be considered supportive of pneumococcal etiology in adult CAP.
Questions have been raised regarding the lack of specificity of the ICT among pediatric patients . Adegbola et al.  also mentioned that the ICT test is not useful for predicting etiology of disease in populations with a high rate of nasopharyngeal carriage of pneumococci. However, there have been few problems with adults. In fact, available data, including results from the present study, suggest that the ICT test has a high specificity in adults with CAP; in our study, only 1 (4%) of 28 urine samples obtained from patients with nonpneumococcal pneumonia yielded a positive result.
The major advantages of the ICT compared with conventional microbiological procedures in the diagnosis of pneumococcal pneumonia are its rapid results and simplicity of use. The ICT specimen is obtained noninvasively and can be read within 15 min. Furthermore, there is no need to prepare special instruments, and the test can be diagnostic even after the prior use of antibiotics.
However, there are several problems: (1) since the ICT examination is expensive (about USD 40 per examination), it is prohibitive to use it for all patients with CAP; (2) there have been false-positive results when the test was administered within 5 days after the administration of S. pneumoniae vaccine; (3) it is unknown for how long the ICT test shows positive results for S. pneumoniae. Yoshida et al.  reported that the mean duration of the urinary ICT kit was 7.3 weeks (4–12 weeks) in 11 of 21 cases with a positive urinary antigen reaction for S. pneumoniae. They noted that since the positive results are prolonged and are present even after recovery from pneumonia in some cases, the ICT kit is not suitable for the evaluation of therapeutic effects. In our study, we tested the urinary ICT kit continuously in 10 of 44 cases with ICT-positive results until the negative conversion was noted. We found that the mean duration of positive ICT kit results was 8.0 weeks (3–14 weeks), which is in agreement with the data of Yoshida et al. . Thus, the ICT kit should not be used in cases of recurrent pneumonia within an interval of <4 months after a pneumonia caused by S. pneumoniae.
In summary, we confirmed that the ICT kit is useful in detecting S. pneumoniae antigen in urine samples, and that it may increase the diagnostic yield for pneumococcal pneumonia in patients with CAP. This in turn may lead to shorter hospital stays and/or a reduction in medical costs. In this study, we were able to establish the clinical impact of the rapidity and simplicity of the ICT test for pneumococcal pneumonia. Thus, it is likely that after insurance licensure in January 2005 in Japan, the ICT test will become a popular and cost-effective method by which to positively determine the appropriate antibiotics that should be given to adult patients with CAP.
Yoshihiro Kobashi MD
Division of Respiratory Diseases, Department of Medicine
Kawasaki Medical School, 577 Matsushima
Kurashiki 701-0192 (Japan)
Tel. +81 86 462 1111, Fax +81 86 464 1041, E-Mail email@example.com
Received: August 15, 2005
Accepted after revision: December 19, 2005
Published online: March 31, 2006
Number of Print Pages : 7
Number of Figures : 0, Number of Tables : 4, Number of References : 27
Respiration (International Journal of Thoracic Medicine)
Vol. 74, No. 4, Year 2007 (Cover Date: July 2007)
Journal Editor: Bolliger, C.T. (Cape Town)
ISSN: 0025–7931 (print), 1423–0356 (Online)
For additional information: http://www.karger.com/RES