Reference Values for the SERION classic ELISA for Detecting Legionella pneumophila Antibodies

Abstract

The first serological test to identify antibodies against Legionella pneumophila was the indirect immunofluorescence antibody test (IFAT) developed in 1977 at the Centers for Disease Control, Atlanta, Ga., USA [1]. Several other tests have since been developed, among them rapid microagglutination tests (RMAT) [2] and numerous (mostly experimental) enzyme-linked immunosorbent assays (ELISA). ELISAs have the advantages of simplicity and rapidity as well as the potential for quantitative and automated performance Nevertheless, due to a lack of general standardization, ELISA assays have not been introduced into clinical practice on a large scale until recently. The current availability of commercial ELISA kits has resulted in the increasing use of these products, despite the fact that few studies determining their sensitivity and specificity are available. In order to fill this knowledge gap, we decided to evaluate the specificity of one commercial ELISA test (SERION classic ELISA; VIRION-SERION, Germany) that is being increasingly used in Dutch laboratories to detect antibodies against Legionella pneumophila. For comparison, we also evaluated an older RMAT test used in the Netherlands. Previously, the ELISA we tested had only been evaluated in one small patient group [3] with an immunofluorescence assay used as the reference method, and in one other small control series [4]. To determine the specificity of the ELISA and RMAT tests in our study, we used a series of control samples taken from a serum bank that had been established in 1995/1996 during a cross-sectional population-based nationwide sero-surveillance study carried out in 40 randomly selected municipalities [5]. In the present study, one sample was taken randomly per municipality (n=40) from persons aged 15–24 years, 10 samples were taken from persons aged 25–64 years, and one sample was taken from persons aged 65–79 years, yielding a total of 480 samples representing an age structure roughly similar to that of a population studied in an outbreak investigation [6]. The commercially available indirect ELISA (SERION classic ELISA) was used to detect immunoglobulin (Ig)Mand IgG-antibodies against Legionella pneumophila (serogroups 1–7). According to the manufacturer, the interserial coefficient of variation is maximally 16%, while the intraserial coefficient of variation is maximally 10% [7]. Among the control samples tested, 5.4% were seropositive for IgG according to the criteria provided by the manufacturer (IgG>70 U/ml), and another 4.4% had a borderline increased IgG titer (50–70 U/ml) (Table 1). None of the samples were seropositive for IgM according to the manufacturer’s criteria (>140 U/ml), and only two (0.4%) samples had borderline increased IgM titers (120– 140 U/ml). Logarithmic transformed ELISA IgM and IgG titers were much more weakly correlated (r=0.09), and of borderline statistical significance (P=0.05), in this control population than in a previously studied population that was partly sub-clinically infected [6] (r=0.26). This could suggest that differences between individuals are mostly due to random variation and not caused by previous infection. H. C. Boshuizen ()) IMA, National Institute for Public Health and the Environment, P.O. Box 1, 3720 BA Bilthoven, The Netherlands e-mail: Hendriek.Boshuizen@rivm.nl Tel.: +31-30-2742944 Fax: +31-30-2744456