NAATs for detection of enterovirus RNA have replaced viral culture and become the method of choice for the diagnosis of enteroviral meningitis at many centers for a number of compelling reasons. Currently, a wide variety of user-developed and two FDA-cleared NAATs with different performance characteristics are used by diagnostic laboratories. The NucliSens EasyQ enterovirus assay, which uses nucleic acid sequence based amplification to detect enterovirus RNA in the CSF, was approved by the FDA in June 2008. The technical complexity of most NAATs requires that they are run in daily batches, and this restricts their availability to laboratories with appropriate expertise in nucleic acid diagnostics. In facilities lacking the appropriate expertise, the tests are often sent to central or regional laboratories, leading to further delays imposed by specimen transport, thereby limiting the clinical utility of the results. In addition, user-developed tests lack standardization, and poor interlaboratory reproducibility of enterovirus NAAT results were observed in a multicenter proficiency testing study (22
). The availability and acceptance by clinical laboratories of an FDA-cleared assay with well-defined performance characteristics will contribute to optimization and standardization of molecular diagnostics for enteroviral meningitis.
The analytical sensitivity and specificity of the Xpert EV assay relative to other NAATs for detection of enterovirus RNA in CSF have been previously published (4
). The sensitivity ranged from 97.1 to 100%, and the specificity was 100% in these studies. In this multicenter study we established the clinical performance characteristics of the Xpert EV assay relative to a definition of enteroviral meningitis that included clinical evidence of meningitis, the absence of another detectable pathogen in CSF, and either detection of enterovirus in CSF by two reference NAATs or by viral culture. Using this definition of clinical truth we found the Xpert EV assay to have a sensitivity of 94.69% and a specificity of 100%.
We also compared the results of the various laboratory-developed NAATs used at the clinical trial sites as the standards of care with the Xpert EV results and found an overall agreement in 95.9% of cases. However, the Xpert EV Assay detected ca. 10% more cases of EV meningitis than the individual laboratory-developed tests.
In the present study only 35.5% of samples from patients with enteroviral meningitis were positive by culture. Although a variety of conventional viral culture methods were used at the clinical trial sites, the majority (82.2%) of the cultures were performed at a reference laboratory that used state-of-the art shell vial cultures using a mixture of sensitive cell lines and pan-enterovirus antibody for detection and identification of enteroviruses (1
False-positive results due to target and amplicon cross-contamination are a significant concern with many NAATs. Since the Xpert EV assay cartridge is completely self-contained and performs all assay steps including sample preparation, false-positive results due to cross-contamination should not occur. As anticipated, no false-positive Xpert EV assay results occurred in the present study.
The on-board positive control in the Xpert EV assay ensures that the sample is adequately prepared, the critical reagents are functioning properly, and the sample is free of interfering substances, thus reducing the risk of false-negative test results due to operator error or interfering substances. In the present study the on-board control failed on the first attempt in only 13 of 434 (3%) tests performed on evaluable patients, indicating that the assay is robust and not prone to operator error. All but one gave a valid result on the second attempt. The actual rate of invalid results may be higher in routine practice. Sefers et al. (17
) reported an 8.2% invalid rate over a 2-year period of routine use. These authors found that either a 1:5 sample dilution of CSF in saline, or freezing and thawing the undiluted CSF markedly reduced the invalid rate with minimal impact on the sensitivity of the assay. All samples in the present study were subjected to one freeze-thaw cycle before testing with the Xpert EV assay.
The primers and probes used in the Xpert EV assay detect all of the recognized enterovirus serotypes, including polioviruses, and do not cross-react with other microorganisms known to cause meningitis-like symptoms (4
). The present study included patients infected with 16 different serotypes, which confirmed the inclusivity of the assay. However, the Xpert EV assay will not detect the closely related human parechoviruses, which like enteroviruses can also cause aseptic meningitis, acute flaccid paralysis, and encephalitis (20
). Of note, three samples were found to contain enterovirus 71 that has the unique ability among the nonpolio enteroviruses to cause brain stem encephalitis and paralysis. Enterovirus 71 has caused widespread outbreaks of disease in Eastern Europe and Southeast Asia and smaller outbreaks in the North America, Europe, and Australia (7
). NAATs are particularly important for detection of this agent since it is difficult to recover from CSF by culture. In the present study all three patients with enterovirus 71 detected in their CSF had a clinical course more consistent with meningitis rather than encephalitis. Only one of the 113 patients with enterovirus detected in the CSF had seizures and/or altered consciousness, suggestive of encephalitis. Coxsackievirus B5, one of the limited number of enterovirus serotypes associated with encephalitis, was detected in this patient.
A child with a positive EV NAAT result could potentially have bacterial meningitis due to either a false-positive NAAT result or a bacterial coinfection. Although rare coinfections of the CSF by bacteria and EVs have been reported (23
), a recent, retrospective, cohort study of children presenting to emergency departments in 20 participating hospitals showed that among 735 children with a positive EV NAAT result none had bacterial meningitis (0%; 95% CI = 0 to 0.5%) (8
). In our study population we found six cases of bacterial meningitis, two caused by Escherichia coli
, two caused by Streptococcus agalactiae
, and one each caused by Neisseria meningitidis
and S. pneumoniae
. All six cases were negative for EV RNA by the NAATs. In mixed infections the clinical presentations are severe enough that the early identification of EV from the CSF is unlikely to deter clinicians from use of antibiotics. Consequently, the detection of EV RNA in CSF from a patient with a clinically compatible illness is sufficient to establish the EV as the cause of the CNS disease.
The benefits of the Xpert EV assay are the ability to obtain reliable results rapidly, within several hours of spinal fluid collection, and to decentralize the testing process so that rapid turnaround time for results can be realized within the medical decision making process. The Xpert EV assay is a fully automated NAAT that can deliver reliable molecular diagnostic results for enteroviral meningitis on demand without the need of specially trained laboratory staff, batching, or dedicated molecular diagnostic facilities. The results of rapid, on-demand testing for enteroviral meningitis could affect the decision to admit patients with suspected meningitis to hospital. Theoretically, patients with positive tests could be managed more conservatively than those with negative tests, without the need for hospitalization, intravenous antibiotics, and attendant nosocomial risks. Since enteroviral meningitis accounts for a large number of hospitalizations of children and adults, this strategy has the potential to simultaneously save significant healthcare resources and provide the most appropriate patient management.