ABSTRACT
Focus Technologies has developed an immunoglobulin M (IgM) capture enzyme-linked immunosorbent assay (ELISA) kit that utilizes recombinant West Nile virus (WNV) antigens to detect WNV IgM in serum. We evaluate here the utility of the kit for detecting WNV IgM in cerebrospinal fluid (CSF). The sensitivity was evaluated by using 52 CSF specimens from the 2002 WNV season that were positive in both the Public Health Service Laboratories WNV IgM ELISA and an in-house WNV IgM ELISA with native WNV antigen. The specificity was evaluated with two groups of specimens: (i) 73 CSF specimens submitted for in-house WNV IgM ELISA testing from February through April 2003 and yielding a negative WNV IgM result and (ii) 60 CSF specimens determined to be positive for another virus by PCR testing. Using these 185 CSF specimens at a screening dilution of 1:2, the kit was determined to be 100% sensitive and 100% specific. Endpoint titers were determined for 20 IgM-positive CSF specimens by testing serial twofold dilutions and ranged from 1:8 to 1:512. Index values (specimen absorbance value/calibrator absorbance value) for the screening dilution (1:2) showed no correlation with IgM titers, whereas index values for higher dilutions showed significant correlation with IgM titers. CSF screening dilutions of greater than 1:2 are not recommended, however, due to the risk of obtaining false-negative results. These findings show that the Focus Technologies WNV IgM capture ELISA, when utilized as recommended, offers accurate qualitative detection of WNV IgM in CSF specimens.
Laboratory assays designed to detect West Nile virus (WNV)-specific immunoglobulin M (IgM) have proven to be very useful for the diagnosis of West Nile virus (WNV) infection (1, 2, 4, 5, 7, 8, 10-13). The vast majority of WNV-infected patients have detectable WNV IgM in serum specimens collected at presentation (6, 9). Involvement of the central nervous system in the infectious process is demonstrated by the detection of WNV-specific IgM in cerebrospinal fluid (CSF) (6, 9, 11, 12). Serum and CSF are typically screened for WNV IgM by using an IgM capture enzyme-linked immunosorbent assay (ELISA). Due to the cross-reactivity observed among flaviviruses in the Japanese encephalitis virus seroocomplex, ELISA-positive samples may be tested by using a confirmatory assay, such as the plaque reduction neutralization test, to show that the IgM detected is specific for WNV (6, 9, 10, 12).
Focus Technologies has developed a WNV IgM capture ELISA kit for the detection of WNV IgM in serum. This kit, which utilizes recombinant WNV premembrane and envelope proteins (3), demonstrates excellent sensitivity and specificity compared to the assays used during the 2002 WNV season in the United States (W. R. Hogrefe, R. J. Moore, and H. E. Prince, unpublished observations). The purpose of the studies presented here was to determine the performance characteristics of the Focus Technologies WNV IgM capture ELISA kit for detecting WNV IgM in CSF.
MATERIALS AND METHODS
CSF specimen panels.The three panels used to assess assay sensitivity and specificity are described in Table 1. The sensitivity panel consisted of frozen (−70°C) aliquots of 52 CSF specimens giving a positive result in an in-house WNV IgM capture ELISA (utilizing native WNV antigens) performed by the Focus Technologies reference laboratory Immunology section during the 2002 WNV season (10, 11). This assay was performed by using CSF at a 1:5 dilution as described previously (11). Duplicate aliquots of these specimens were forwarded to PHSL (13 different State Public Health Laboratories) for further evaluation during 2002, and all 52 specimens were also positive in the Public Health Service Laboratories (PHSL) WNV IgM capture ELISA (7, 11). Only five of these CSF specimens were tested by the plaque reduction neutralization test, and all were positive for WNV-specific antibodies. The specificity was evaluated by using two different panels. The first specificity panel consisted of 73 CSF specimens submitted for testing by the in-house WNV IgM capture ELISA (native antigen) during a period of low WNV endemicity (February through April 2003) and giving a negative WNV IgM result. The second specificity panel consisted of 60 CSF specimens positive for another virus by PCR testing performed by the Focus Technologies reference laboratory Molecular Diagnostics section; samples were frozen (−70C) after PCR testing until utilized for the present study.
CSF specimens evaluated
WNV IgM capture ELISA kit.The Focus Technologies (Cypress, Calif.) WNV IgM capture ELISA kit utilizing recombinant antigens was used according to the manufacturer's instructions, with the exception that CSF was diluted 1:2. The kit contains a calibrator serum, which is included in all runs; the reactivity of each kit control and unknown specimen is expressed relative to the reactivity of the calibrator serum, thus allowing for the interassay comparison of results. Calibrator serum (in triplicate), negative control serum, and positive control serum were diluted 1:101 in sample diluent (phosphate-buffered saline containing 0.1% bovine serum albumin and 0.1% Tween 20), and CSF specimens were diluted 1:2 in sample diluent. Diluted specimens (0.1 ml) were added to microtiter wells (Nunc PolySorp) coated with rabbit anti-human IgM (heavy-chain specific). After an hour at room temperature, the wells were washed three times with 0.25-ml volumes of wash buffer (phosphate-buffered saline plus 0.1% Tween 20); the wells then received 0.1 ml of WNV recombinant antigen (lyophilized material reconstituted with sample diluent per the package insert) for 2 h at room temperature. After three washes as described above, 0.1 ml of horseradish peroxidase-conjugated anti-flavivirus monoclonal antibody (6B6C-1), supplied as a ready-to-use reagent, was added to all wells. After incubation at room temperature for 30 min and three washes as described above, the wells received 0.1 ml of enzyme substrate (tetramethylbenzidine), supplied as a ready-to-use reagent, for 10 min; the reaction was then stopped by the addition of 0.1 ml of 2 N sulfuric acid. The absorbance at 450 nm was measured by using an ELISA reader. For each specimen, including serum controls and patient CSF specimens, the index value was calculated by dividing the specimen absorbance value by the mean absorbance value obtained for the calibrator serum. Index values of <0.90 were considered negative, values of 0.90 to 1.10 were considered equivocal, and values of >1.10 were considered positive for WNV IgM.
Titration of positive CSF specimens.To further characterize CSF specimens giving a positive result with the WNV IgM capture ELISA kit (recombinant antigen), serial twofold dilutions of positive specimens with sufficient volume (n = 20) were tested as described in the previous section. The endpoint titer was defined as the highest dilution giving an index value of ≥0.90.
Statistical analyses.The relationship between index values observed at various CSF dilutions and endpoint titers was evaluated by using linear regression anal-ysis. A correlation coefficient (R2) with a P value < 0.01 was considered statistically significant.
RESULTS
Sensitivity and specificity.The results for the sensitivity and specificity panels are shown in Fig. 1. All CSF specimens in the sensitivity panel exhibited WNV IgM capture ELISA (recombinant antigen) index values of >1.10, whereas all CSF specimens in the two specificity panels exhibited index values of <0.90. Thus, when used to test CSF specimens at a 1:2 dilution, the Focus Technologies WNV IgM capture ELISA kit was 100% sensitive and 100% specific.
Distribution of index values for CSF specimens tested in the Focus Technologies WNV IgM Capture ELISA kit. Each point represents the result for one specimen.
Titration studies.WNV IgM-positive CSF specimens with sufficient volume were serially diluted and tested to determine the endpoint titer. Index values are presented as a function of dilutions in Fig. 2. For 16 of 20 CSF specimens, the titer was measurable using dilutions between 1:2 and 1:256; four specimens required retesting with dilutions expanded to cover the range 1:2 to 1:1,024. The distribution of endpoint titers was as follows: 1:8 (n = 1), 1:16 (n = 4), 1:32 (n = 1), 1:64 (n = 5), 1:128 (n = 4), 1:256 (n = 2), and 1:512 (n = 3).
Titration curves for 20 WNV IgM-positive CSF specimens. Each curve represents the results for one CSF specimen.
Relationship of index values at various dilutions to endpoint titers.To determine whether index values could be used as a semiquantitative indicator of endpoint titers, index values obtained for 1:2 to 1:16 dilutions were plotted as a function of endpoint titers for the 20 WNV IgM-positive CSF specimens tested by serial dilution (Fig. 3). No significant correlation was observed between endpoint titers and index values at a 1:2 dilution. In contrast, endpoint titers were significantly correlated with index values at 1:4, 1:8, and 1:16 dilutions. Higher dilutions were associated with higher correlation coefficients. However, the strong correlation between endpoint titers and index values at the 1:16 dilution was offset by a decrease in assay sensitivity; the specimen with a titer of 1:8 exhibited a false-negative result at the 1:16 dilution, and two of three specimens with titers of 1:16 were equivocal.
Relationship of index values at various dilutions to endpoint titers for 20 WNV IgM-positive CSF specimens. Correlation coefficient (R2) values at 1:4, 1:8, and 1:16 dilutions (B, C, and D, respectively) were statistically significant (P < 0.01), whereas the correlation coefficient for the 1:2 dilution (A) did not reach statistical significance.
DISCUSSION
The findings presented in this report clearly demonstrate that the Focus Technologies WNV IgM Capture ELISA kit, initially designed for testing serum, accurately detects WNV IgM in CSF. Using a panel of well-characterized CSF specimens known to contain WNV IgM, the kit exhibited 100% sensitivity. Likewise, the kit was 100% specific; all 73 CSF specimens collected during a period of low endemicity and negative for WNV IgM by using an in-house ELISA based on reactivity with native WNV antigens were also WNV IgM negative with the kit. Studies conducted with an additional specificity panel demonstrated that CSF specimens from patients with infections caused by other viruses, including enteroviruses and herpesviruses, did not give false-positive results with the WNV IgM kit.
Due to the unavailability of CSF specimens from patients with other flavivirus infections, we were not able to perform CSF cross-reactivity studies with the Focus Technologies IgM kit. However, studies using similar IgM capture assays to evaluate serum specimens have shown that cross-reactivity may be extensive, depending on the infecting flavivirus. Martin et al. (8) found that 63% of sera from patients infected with St. Louis encephalitis virus gave a positive signal in a WNV IgM capture ELISA; in contrast, only 31% of patients with Japanese encephalitis virus infection were reactive. These findings emphasize the need to perform additional testing for patients whose CSF and/or serum are positive in a flavivirus IgM capture ELISA.
Our choice of a screening dilution of 1:2 for CSF was a compromise between precedent and practicality. Although CSF specimens are tested undiluted in the PHSL WNV IgM ELISA (7, 8), our experience in a reference laboratory setting demonstrates that the volume of CSF submitted is often too low to permit testing of undiluted sample. Thus, some degree of dilution is required to facilitate testing of all submitted CSF specimens. Further, subjecting CSF specimens to dilution serves to minimize any adverse matrix effects that may characterize utilization of undiluted CSF.
Detailed characterization of WNV IgM detection in CSF at various dilutions revealed that, at the screening dilution of 1:2, the WNV IgM capture ELISA kit functions as a qualitative assay. There was no significant correlation between the index observed at 1:2 and the endpoint titer determined by testing serial twofold dilutions. Although the results indicated that the assay exhibited quantitative characteristics at higher dilutions, the routine utilization of higher dilutions for CSF testing is not recommended. This position is based mainly on the lack of data regarding the endpoint titers of WNV IgM-positive CSF specimens exhibiting index values in the range of 1.1 to 3.0 when tested at a 1:2 dilution. Although three WNV IgM-positive CSF specimens exhibited index values in this range (Fig. 1), none had sufficient volume to allow titration studies. Along these same lines, the population of WNV IgM-positive CSF specimens used for these experiments was biased toward strongly positive specimens, since they were positive when tested at a 1:5 dilution in the 2002 WNV IgM ELISA utilizing native WNV antigen (11). Thus, it remains unclear what proportion of CSF specimens truly positive for WNV IgM, as assessed at a 1:2 dilution with the new WNV IgM ELISA kit, may give false-negative results at higher dilutions. Until the findings of such studies are available, we recommend that the Focus Technologies WNV IgM capture ELISA kit be used as a qualitative assay for the detection of WNV IgM in CSF specimens.
ACKNOWLEDGMENTS
We thank all PHSL staff members who generated and supplied PHSL WNV IgM ELISA results. Jane Filamor, Maryam Saber- Tehrani, and Susan Vogeli provided expert technical assistance.
FOOTNOTES
- Received 2 August 2003.
- Returned for modification 11 September 2002.
- Accepted 11 October 2003.
- Copyright © 2004 American Society for Microbiology
