Comparison of three kinds of commercially available kits for detecting antigen for Legionella in the urine from patients with legionellosis.
M. Koide,T. Shinzato,F. Higa,M. Tateyama,H. Sakugawa,A. Saito
DOI: https://doi.org/10.2169/INTERNALMEDICINE.44.673
2005-06-01
Internal Medicine
Abstract:Bacteria of the genus Legionella are the cause of both community-acquired and nosocomial pneumonia. Because pneumonia caused by Legionella spp. cannot be differentiated from other types of pneumonia using clinical, radiographic, or laboratory findings, the specific etiologic diagnosis must be based on microbiological tests. Soon after Legionnaires’ disease was first identified, the advantages of diagnosing legionellosis by urinary antigen detection were recognized, including early detection, rapidity of testing, and ease of specimen collection. These tests for urinary antigen have proven to be more than 95% accurate. At present in Japan, three commercially available kits can be used for detecting Legionella urinary antigen, including the Binax Legionella Urinary Antigen EIA kit (Binax, Portland, ME, USA), the Binax Now Legionella Urinary Antigen kit (Binax), and the Biotest Legionella Urine Antigen EIA kit (Biotest, Dreieich, Germany). The Binax Now test is an immunochromatographic assay using a hinged, book-shaped test device. Binax kits detect excreted lipopolysaccharides of Legionella pneumophila serogroup 1. In contrast, the Biotest kit detects excreted heat-stable and heat-labile antigens of L. pneumophila serogroup 1 as well as some other serogroups and some Legionella spp. antigens (1–3). In the current studies, 90 urine samples were obtained from 38 patients with suspected legionellosis between 1993 and 2002 from throughout Japan. The urine samples were distributed between 2 and 81 days after the onset of symptoms. Between one and six samples were gathered from each patient. For 27 of the patients, both respiratory and serum samples were available in addition to urine samples. Of these patients, the causative species was detected as L. pneumophila serogroup 1 in 17 of the patients and L. pneumophila serogroup 4 in 1 patient. The causative species of the remaining 9 patients from this group of 27 could not be determined. For another 8 of the total 38 patients, serum samples were available in addition to the urine samples, but respiratory samples were not available. The causative species was Legionella dumoffii for one of these 8 patients. The causative species of the remaining seven patients could not be determined. Finally, for the remaining 3 of the total 38 patients, only urine samples were available so that the causative species could not be identified (Table 1). Urine samples were tested using the three Legionella urinary antigen test kits. All 38 patient’s urine samples (72 total) showed positive results with the Binax Legionella Urinary Antigen EIA kit, 37 patient’s urine samples (70 total) were positive with the Binax Now Legionella Urinary Antigen kit, and 33 patient’s urine samples (63 total) were positive with the Biotest Legionella Urine Antigen EIA kit. Of the three patients from whom only urine samples were collected, one patient had two samples that were negative with all three kits and one sample that was positive with all three kits. Another one of these patients had two urine samples, both of which showed positive results with the Binax EIA and Biotest EIA kits but negative results with the Binax NOW kit. The final one of these three patients had only one urine sample, which showed a positive result with the Binax EIA and Binax Now kits but negative results with the Biotest EIA kit. Finally, of 14 patients that were positive for L. pneumophila serogroup 1 in culture, 3 had urine samples that were negative with the Biotest EIA kit. Because some patients excrete urinary antigen intermittently, care must be taken not to overlook the patients that are negative in their first urine samples. Negative results with all three kits were obtained using control urine samples from 15 patients with pneumonia caused by pathogens other than Legionella and from 25 patients with urinary tract infection caused by various etiologic bacteria. Helbig et al (4), reported that the lowest limit of detection was 3×10 sonicated bacteria/ml for the L. pneumophila serogroup 1 monoclonal subtype Philadelphia antigen in all three assays. For strains subtype OLDA or Bellingham, the sensitivity of the two Binax assays was slightly but insignificantly lower, whereas the Biotest EIA had a limit of detection of 6×10 cells/ml (4). In the patients infected by L. pneumophila serogroup 1 monoclonal subtypes other than Philadelphia, the quantitative zone that showed a positive result by Binax assays only, might exist in their urine. Benson et al (5), using patient’s urine, reported the sensitivity of Binax EIA and Biotest EIA to Legionella spp. other than L. pneumophila and L. pneumophila non-serogroup 1. They detected urinary antigen in only 1 of 19 patients with Legionella spp. and in 13 of 26 patients with L. pneumophila non-serogroup 1. Biotest EIA showed excellent sensitivity especially in cases of L. pneumophila serogroup 6 (5). But as they did not contain urine from L. dumoffii patients, the sensitivity of kits against L. dumoffii were unknown. It remains unclear, if whether patients infected by Legionella strains not belonging to L. pneumophila serogroup 1 excrete smaller amounts of urinary antigen or if the immunoassay has a lower sensitivity (1, 4). The other Legionella urinary antigen kit which we did not use in this study, Bartels ELISA ( Trinity Biotech, Wicklow,