ABSTRACT
Legionnaires’ disease (LD) refers to a serious form of acute pneumonia caused by Legionella species. LD can be difficult to diagnose because the signs and symptoms are nonspecific, and therefore a rapid laboratory diagnosis is of paramount importance. In this study, a recently introduced immunochromatographic test (Immunocatch Legionella; Eiken Chemical Co., Ltd.) for Legionella pneumophila (serogroup 1) urinary antigen detection was compared with the Sofia Legionella fluorescent immunoassay (FIA) (Quidel) (routinely used in our laboratory) and with the widely used BinaxNOW Legionella assay (Alere). A total of 248 urine samples (60 frozen and 188 fresh) were evaluated. All of the samples were collected from patients with high clinical suspicion of Legionnaires’ disease. The three assays were performed simultaneously according to the manufacturers’ instructions. A total of 180 concordant negative and 66 concordant positive results were obtained. Only 2 discrepant results were registered. The sensitivity and specificity of Immunocatch compared with Sofia were, respectively, 98.5% and 99.4%. Cohen's kappa coefficient and overall percent agreement between Immunocatch and Sofia were also calculated and resulted in, respectively, 0.97 and 99.2%. These performances suggest that the Immunocatch test is a useful tool for Legionella pneumophila (serogroup 1) urinary antigen detection.
INTRODUCTION
Legionnaires’ disease (LD) is an important acute pneumonia commonly associated with exposure to water devices contaminated with Legionella bacteria species. Legionella pneumophila is a cause of community-acquired, nosocomially acquired, and travel-associated pneumonia. According to a U.S. summary report of LD surveillance (2014 to 2015), approximately 10% to 15% of all reported cases of LD occur in people who traveled during the 10 days before onset of symptoms, 19% to 20% of cases are in patients who reported health care exposure, and the remaining percentage of cases includes those that were acquired in assisted or senior living facilities, workplace environments, and the community (1).
Clinical manifestations and radiological features suggestive of pneumonia cannot discriminate LD from other bacterial pneumonias; therefore, a rapid laboratory diagnosis is of paramount importance to provide critical opportunities for appropriate antibiotic therapy in suspected LD patients (2). Legionella pneumophila serogroup 1 antigen detection (Legionella urine antigen test [LUAT]) in urine is recommended to evaluate a possible case of legionellosis, even if other serogroups of L. pneumophila and species of the genus Legionella could be involved in human disease. Legionella pneumophila is responsible for about 90% of the cases of legionellosis; with 70 to 80% of the total caused by serogroup 1 (3). The duration of antigenuria varies from a few days to as long as a year depending on the patient’s characteristics, such as immunosuppression condition (4). The urinary antigen test for L. pneumophila detection is an easy-to-use and rapid test, and the urine samples can be collected easily from most patients, in contrast to sputum, which is generally scarce in LD patients (5). The aim of this study was to compare a new immunochromatographic test (ICT) (Immunocatch Legionella; Eiken Chemical, Tokyo, Japan) with the Sofia Legionella assay (Quidel, San Diego, CA, USA). The second method is routinely used in our laboratory. In addition, the comparative evaluation was also performed with one of the most widely used LUATs, the BinaxNOW Legionella assay (Alere-Abbott, Scarborough, ME, USA).
MATERIALS AND METHODS
A total of 248 (188 freshly harvested and 60 kept frozen at −80°C and never thawed before being used for the study) urine samples (US) submitted for LUAT between December 2016 and August 2018 were evaluated. All of the samples were collected from patients with high clinical suspicion of pneumonia. Our clinical colleagues, for the management of community-acquired pneumonia (CAP), have followed the 2009 British Thoracic Society guidelines (6). The Immunocatch Legionella is a rapid immunochromatographic test (ICT) for qualitative detection of L. pneumophila serogroup 1 urinary antigen. When the urine sample migrates by capillary action through the conjugate pad, Legionella antigen in the specimen binds to anti-Legionella antibodies conjugated to gold colloidal particles to form and highlight the antigen-antibody complex. The antigen-antibody complex migrates up the test strip and specifically binds to anti-Legionella antibody fixed on the membrane to generate a red line. The presence of Legionella antigen is then determined by visually observing the presence of the red line. The test is validated by the presence of a deep purple control line, indicating that the reaction on the test strip was successful. BinaxNOW Legionella, as well as Immunocatch Legionella, is a lateral flow ICT with manual or automatic reading, whereas the Sofia Legionella fluorescent immunoassay (FIA) is an ICT that uses immunofluorescence technology with automated reading. In our study, BinaxNOW was interpreted visually. All of the three assays were performed simultaneously on individual samples according to each test manufacturers’ instructions. The three tests were performed simultaneously by a first technician. The same technician carried out the interpretation of the Sofia assay, routinely used in our laboratory. Second and third well-trained technicians, not informed about the Sofia result, were involved in the study in order to visually interpret the BinaxNOW and Immunocatch assays. Faint and very faint lines were interpreted as positive as recommended by the manufacturers. Rheumatoid-like factors have been associated with the occurrence of false positives in the Sofia immunoassay. In order to increase the specificity of the Sophia test, all of the positive urine samples were boiled and tested again as per the manufacturer’s recommendation. The US positive by the Sofia assay were boiled (100°C for 5 min), centrifuged (15 min at 1,000 × g), and tested repeatedly in order to eliminate nonspecific and potential interference (7, 8). The concentration of individual US was achieved by centrifuging 4 ml of sample for 15 min at 1,500 × g in an Amicon Ultra-4 Ultracel 10K device (Millipore Corporation, Bedford, MS, USA) (9). All of the concordant positive results obtained with the three tests evaluated have been further confirmed as positive by detection of immune seroconversion from two subsequent serum samples collected not earlier than 2 weeks one from another and by identification of Legionella-specific DNA sequences in biological specimens from the respiratory tract by PCR.
RESULTS
A total of 180 concordant negative and 66 concordant positive results were obtained. Only 2 discrepant results were registered (Table 1). The results obtained are summarized in Table 1 and in Table 2. A total of two discrepant samples were found. A first discrepant US was positive with Sofia (after boiling) and negative when tested by the BinaxNOW and Immunocatch assays, but after urine concentration, the three assays yielded concordant positive results. The second discrepancy refers to a US that resulted negative with Sofia (after heat treatment), negative with BinaxNOW, and positive (with a very weak test line) when tested with Immunocatch. The result of the Immunocatch remained positive when retested after urine heat treatment, and unfortunately, it was not possible to collect additional US or sputum from this patient; as a result, the final evaluation could be a false positive for the Immunocatch result. In order to measure the overall diagnostic performance of the Immunocatch test, parameters like sensitivity, specificity, and concordance rate (Cohen’s Kappa value) were determined. The sensitivity and specificity of Immunocatch compared with Sofia were, respectively, 98.5% and 99.4%. The overall percent agreement (OPA) between Sofia and Immunocatch assays was determined by using two-by-two contingency tables. This measure represents the proportion of samples identically classified by the Sofia and Immunocatch methods. The OPA value was calculated and was 99.2%. Cohen’s Kappa value was also calculated as a further measure of the agreement between the Sofia and Immunocatch assays. The agreement between the Sofia and the Immunocatch methods showed an excellent Cohen’s kappa coefficient of 0.97.
Two by two contingency table that summarizes the concordant and discrepant results between the Sofia and Immunocatch assaysa
Two by two contingency table that summarizes the concordant and discrepant results between the BinaxNOW and Immunocatch assays
DISCUSSION
The diagnosis of pneumonia caused by Legionella spp. is based on a multidisciplinary approach that includes many laboratory data, such as the following: detection of the germ in respiratory specimens by culture or nucleic acid amplification, identification of the specific antibody response, and results of LUATs. The effect that the wide use of LUATs has generated in the last several decades is a dramatic increase in the number of LD cases diagnosed worldwide (10). Overall, the sensitivity and specificity of LUATs range from 75% to 80% and about 100%, respectively. One of the major variables that can influence the sensitivity of a LUAT is the severity of the disease, with a positive correlation between severe cases and higher sensitivity (11). The novel Immunocatch Legionella shares a basic diagnostic principle with most of the first-generation LUATs, as follows: the urinary antigen is bound by a specific antibody, and this immune complex is highlighted by a secondary antibody (anti-primary antibody) bound to colloidal gold particles in a later flow device. This device is very easy and rapid to use, thus allowing a faster and simpler laboratory workflow based on LUATs. A possible weakness of the novel Immunocatch test could be the visual reading for the interpretation of the results, whereas both of the competitor tests investigated offer the possibility to score the results by an automated reading instrument. It has been recently underlined that there is a necessity to perform additional clinical and laboratory studies to better understand the impact of LUAT results on the management of patients suffering from suspected Legionella-related pneumonia (12). In the present study, all of the concordant positive results obtained with the three tests evaluated have been further confirmed as positive by additional laboratory testing (i.e., detection of immune seroconversion on two subsequent serum samples collected not earlier than 2 weeks one from another and/or identification of Legionella-specific DNA sequences in biological specimens from the respiratory tract by PCR). The first US that gave negative results with the Immunocatch and Binax assays (positive with the Sofia test) turned positive after concentration in both of the previously negative tests, thus, suggesting lower sensitivity of the Binax and Immunocatch methods in comparison with that of Sofia. The remaining discrepant results were not confirmed due to the impossibility of getting additional biological samples or serum. The final conclusion is a possible false negative by Immunocatch. Overall, these performances suggest that the Immunocatch test is a useful tool for Legionella pneumophila (serogroup 1) urinary qualitative antigen detection, as it is simple to perform and does not need any instrumentation. The Immunocatch assay was financially profitable with respect to the other two assays, and the entire workflow in urine sample processing was simpler.
ACKNOWLEDGMENTS
The authors thank Eiken Chemical Co. Ltd. (Tokyo, Japan), which kindly provided the reagents for the Immunocatch Legionella and BinaxNOW Legionella assays.
We declare no conflicts of interest.
FOOTNOTES
- Received 25 February 2019.
- Returned for modification 19 March 2019.
- Accepted 10 May 2019.
- Accepted manuscript posted online 15 May 2019.
- Copyright © 2019 American Society for Microbiology.
