Laboratory Testing Practices for West Nile Virus in the United States

Introduction

W est Nile virus (WNV), a mosquito-borne flavivirus, is the leading cause of arboviral disease in the United States (Reimann et al. 2008). Following its initial detection in 1999, almost all of the WNV diagnostic testing in the United States was performed at the Centers for Disease Control and Prevention (CDC) using an in-house enzyme-linked immunosorbent assay (ELISA) and confirmatory plaque reduction neutralization tests (PRNT). By 2002, many state public health laboratories began performing WNV testing using the CDC-developed assays. Starting in 2004, several Food and Drug Administration (FDA)–cleared WNV diagnostic kits became commercially available, and have since been used by commercial diagnostic reference laboratories to offer testing.

There are four FDA-cleared WNV serological kits available in the United States. These kits are labeled for use on serum to aid in a presumptive diagnosis of WNV infection in patients with clinical symptoms consistent with neuroinvasive disease. According to the product inserts (Focus Diagnostics, Inc., 2009, InBios International, Inc., 2004, Inverness Medical Innovations Australia Pty Ltd., 2005, and Spectral Diagnostics, Inc., 2006), all positive results obtained with these assays should be confirmed by PRNT as false-positive can occur due to nonspecific reactivity or cross-reactivity to other flaviviruses (e.g., St. Louis encephalitis, Japanese encephalitis, dengue, or yellow fever virus) (CDC 2003). The sensitivity and specificity of these kits were established using the CDC serologic assays as the gold standard (Focus Diagnostics, Inc., InBios International, Inc., Inverness Medical Innovations Australia Pty Ltd., and Spectral Diagnostics, Inc.).

In September 2008, a nationwide investigation was performed after identifying an increased rate of false-positive WNV immunoglobulin (Ig)M results that was linked to one lot of a commercially available FDA-cleared WNV ELISA test kit (CDC 2009). At the time of the investigation, it was unknown how many tests had been performed using the affected kit and how many laboratories were using the kit. To address these questions and gain knowledge about WNV testing practices, we surveyed state public health and commercial diagnostic reference laboratories to determine the types and quantities of WNV tests being performed in the United States.

Materials and Methods

In late 2008, a telephone survey was conducted with all 50 state public health laboratories and 6 commercial diagnostic reference laboratories (i.e., ARUP Laboratories, Focus Diagnostics, LabCorp, Mayo Medical Laboratories, Specialty Laboratories, and Quest Diagnostics). The survey included questions regarding the types of WNV testing that each laboratory performed in 2008 and the specific WNV IgM assays or kits that they used. In addition, since full year data were not available for 2008, the total number of WNV IgM tests performed by each laboratory in 2007 was obtained as an indicator for the quantity of WNV tests being performed. Analyses were performed with SAS 9.2 software.

Results

Overall, 46 (92%) of the 50 state public health laboratories and all 6 commercial diagnostic reference laboratories performed WNV testing in 2008. For both state public health and commercial diagnostic reference laboratories, WNV IgM assays were the most common test performed, followed by IgG assays, and reverse transcriptase-polymerase chain reaction (Fig. 1). Only 12 (24%) of the state public health laboratories and 2 (33%) of the commercial diagnostic reference laboratories performed confirmatory PRNT. Viral culture was performed only at 9 (18%) state public health laboratories.

OPEN IN VIEWER

FIG. 1.

Proportion of state public health and commercial laboratories performing specific West Nile virus tests, 2008.

Thirty-three (66%) public health laboratories reported using a WNV IgM assay developed at CDC, including microsphere-based immunoassay (21, 42%) or ELISA (12, 24%). However, 17 (34%) of the state public health laboratories were using a commercially available FDA-cleared WNV IgM kit in addition to (n = 6), or instead of (n = 11), the CDC-developed assays. All six commercial diagnostic reference laboratories reported using a commercially available FDA-cleared IgM ELISA test kit. At the time of the survey, all six laboratories reported using a test kit produced by the same manufacturer. However before the survey, one of the commercial diagnostic reference laboratories had been using the recalled kit but subsequently switched to a different kit following the recall.

Overall, the surveyed laboratories reported performing 94,194 WNV IgM tests in 2007. Of these, only 22,137 (24%) were performed in a state public health laboratory compared to 72,057 (76%) WNV IgM tests performed at the six commercial diagnostic reference laboratories (p < 0.01).

Discussion

These findings document a change in WNV testing practices since the virus was first detected in the United States in 1999. During the first 5 years, most of the WNV testing in the United States was performed at CDC or state public health laboratories using assays developed at CDC. Ten years after the introduction of the virus, the majority of testing is being done in commercial diagnostic reference laboratories utilizing commercially available FDA-cleared kits. Further, an increasing number of state public health laboratories are using these same kits.

According to the package inserts, any positive result from an FDA-cleared WNV IgM kit should be confirmed by PRNT (CDC 2003). While specific practices with regard to confirmatory testing were not queried during the survey, our recent investigation that identified a 72% false-positive rate with a commercially available FDA-cleared test kit found that <10% of over 500 presumptive positive samples had been sent for confirmation (CDC 2009). In addition, we found that only 25% of surveyed state public health or commercial diagnostic reference laboratories have the capacity to perform confirmatory testing. While two commercial diagnostic reference laboratories surveyed did offer PRNT, the test would need to be ordered by a clinician. Clinicians should be aware that the commercially available serologic kits are screening tests with the possibility of false-positive results due to nonspecific reactivity, cross-reactivity to other flaviviruses, or errors in testing; therefore, they are not adequate to determine the agent causing the disease. Clinicians can mitigate the risk of obtaining false-positive results by ordering confirmatory testing and gathering information on possible reasons for cross-reactivity (e.g., the patient's country of origin, travel history, or vaccination history with yellow fever or Japanese encephalitis vaccine). In the event that confirmatory PRNT testing is not available, IgM testing of cerebrospinal fluid (CSF) or acute and convalescent serum specimens may improve the accuracy of the diagnostic testing. Further, laboratories should consider performing confirmatory testing on any IgM-positive specimens.

While commercially available test kits are cleared by the FDA upon first release, continued monitoring is the responsibility of the test kit manufacturers. Clinical Laboratory Improvement Amendments regulations also mandate that laboratories monitor the performance of any assays or test kit it utilizes (CDC 2004, Centers for Medicare & Medicaid Services 2009). State public health laboratories that perform WNV testing currently undergo annual proficiency testing (Johnson et al. 2007); however, no mechanism exists for commercial diagnostic reference laboratories.

This survey documents the extensive use of commercially available FDA-cleared kits for the diagnosis of WNV disease in the United States. These findings indicate the need for both manufacturers and laboratories to more closely monitor the performance of these kits. Clinicians and public health officials should be aware of the limitations of these kits and the need for additional testing to confirm a diagnosis of WNV disease.