Identifying the agent responsible for suspected cases of viral central nervous system (CNS) infection poses tremendous diagnostic challenges, and a specific organism is identified in only ∼30% of cases of suspected viral encephalitis.¹ Traditionally, definitive diagnosis has depended on: 1) culture of virus from cerebrospinal fluid (CSF) or brain tissue; 2) identification of viral particles, inclusions, antigen, or nucleic acid in brain tissue; or 3) demonstration of virus specific intrathecal antibody synthesis.

The ability to amplify small amounts of viral nucleic acid from CSF using the polymerase chain reaction (PCR) technique has revolutionised the diagnosis of viral CNS infections. CSF PCR is rapid, inexpensive, and only minimally invasive. Unfortunately, validation of the sensitivity and specificity of CSF PCR by comparison to a “gold standard” such as brain biopsy, is only rarely available.^2,3 False positive CSF PCR results are rare when tests are performed according to strict standards in experienced laboratories, with rigorous attention to procedures designed to avoid specimen contamination and to verify the specificity of amplification products. The sensitivity of CSF PCR varies with different viruses, and can be dramatically influenced by the timing of specimen collection in relation to onset of illness. For example, in herpes simplex virus (HSV) encephalitis false negative CSF PCR results may occur when specimens are collected either too early or too late.^2,4 In the study by Davies et al (this issue, pp 82–7)⁵ the prevalence of positive CSF PCR results was maximal when specimens were obtained at 3–14 days (16%–19% positive) after symptom onset and significantly lower when specimens were obtained earlier (6%) or later (2%).

Clinicians are still faced with the daunting task of ordering individual PCR tests for each virus of potential interest. Recently, “multiplex” CSF PCR assays have been developed that utilise multiple primers simultaneously in a single reaction mixture to amplify nucleic acid from a group of viruses. Davies and colleagues⁵ used this technology to evaluate 787 CSF samples from patients with suspected CNS infections for the presence of HSV 1 and 2, cytomegalovirus, Epstein-Barr virus (EBV), varicella-zoster virus, human herpes virus (HHV)-6, JC virus, and enteroviruses. CSF PCR was positive in 30% of patients with “likely” CNS viral infection—a result similar to other recent studies.^1,6,7 The 70% of cases in which a viral agent was suspected but never discovered may be due to: 1) unknown infectious agents; 2) unusual infectious agents not covered in the tests employed; 3) known agents missed because of false negative PCRs; or 4) non-infectious CNS diseases mimicking encephalitis.¹

There were several surprising results in the study by Davies et al.⁵ In 9 of 88 (10%) positive first CSF samples, nucleic acid from two or more viruses—including EBV in 6 cases—was detected. Four of the five patients with dual positive CSF PCRs for whom detailed clinical information was available were human immunodeficiency virus (HIV) positive. Multiple positive CSF PCRs on the same CSF specimen is fortunately uncommon, but may occur in immunocompromised patients. CSF PCR may detect latent lymphotrophic viruses such as EBV in CSF inflammatory cells, or such latent viruses may reactivate in the CNS producing “dual” infections.⁸ Another unexpected result was that CSF PCR was positive in 15 of 291 (5%) patients classified as “unlikely” to have CNS viral infection. 53% of those with a positive PCR had a normal CSF cell count and 34% had both a normal cell count and protein level. The clinical significance of these PCR positive tests is currently unclear. Clinical judgment must be used both in determining when to order diagnostic CSF PCR assays and in the interpretation of the findings.

Technical refinements of the basic PCR procedures—including use of real-time PCR and quantitative PCR—and PCR amplification to identify viral genes associated with resistance to antimicrobial chemotherapy have already entered clinical practice. An exciting research development is the availability of large scale microarrays that allow simultaneous detection of the expression of thousands of genes in single specimens. Microarrays could be used to quantify the expression of each gene in a viral genome to provide invaluable information about epidemiology, virulence determinants, and susceptibility to drugs.^9,10 Chips using multi-viral gene probe sets will facilitate future pathogen discovery and may lead to discovery of viral aetiologies in both established and novel CNS diseases.