Serological diagnosis of TBE can cause problems. Cross reactivity due to previous flavivirus vaccination or infection or a tests with low sensitivity or specificity may affect diagnostic precision. Using standardised enzyme-linked immunosorbent assay (EIA) with appropriate controls, at least 96% of TBE cases in the second meningoencephalitic phase of the disease are IgM positive [2]. Old indirect EIA tests are considered less specific compared to analysis based on microcapture techniques, and generate more false positives. However, more recently developed indirect EIA techniques and immunoblots for TBE diagnosis have both high sensitivity and specificity [2, 3, 4]. In a Swedish prospective evaluation, we found that all TBE cases with specific IgM reactivity on hospital admission could be verified by presence of increased IgG antibody activity in convalescent sera and by intrathecal IgM antibody production [2, 5]. Complement binding reaction with four-fold titre increase in paired sera is an outdated technique that has been replaced by modern EIA technology. TBE antigen detection by virus isolation or polymerase chain reaction (PCR) in the IgM positive phase of the disease is, except for rare positive cases usually post-mortem, negative, and not a useful tool in the diagnosis of TBE [6, 7].

The criteria for a case definition proposed by Stefanoff et al [1] are reasonable. The results and the revision of Polish national surveillance data using the proposed case definition are probably relevant for many TBE endemic countries in Europe. If the discussion is limited to TBE CNS disease, possible cases of TBE will include all cases presenting with meningoencephalomyelitis in a TBE endemic area during the tick season, extended with the longest possible incubation period for CNS symptoms to occur (about four weeks). Consumption of unpasteurised milk products originating from endemic areas should be included in the case definition. Whether cerebrospinal fluid (CSF) pleocytosis is also required in all cases could be debated. In several large consecutive studies on TBE meningoencephalomyelitis, all patients presented with CSF pleocytosis [5, 8, 9,10]. Although not clearly stated, pleocytosis is such an inherent part of the diagnostic process that it almost becomes a compulsory inclusion criteria in these studies. A selection bias with regard to the presence of CSF pleocytosis can therefore not be fully excluded. Nevertheless, TBE associated CNS disease without CSF pleocytosis must be rare, probably even more than in herpes simplex encephalitis. If such cases are encountered, false positive serological diagnosis must be ruled out. Apart from the epidemiological criteria, a possible case could be defined by the presence of specific serum IgM antibodies. Preceding flavivirus disease (visit abroad) or vaccination (TBE, yellow fever and Japanese encephalitis) must, of course, be excluded. TBE IgM antibodies may persist for at least one year [2] and a previous asymptomatic or less apparent TBE virus infection might cause diagnostic problems in a case of non-TBE meningoencephalitis. Based on an estimated maximum yearly TBE seroconversion rate of 1.2-2.4% [11] and a fairly low incidence of non-TBE viral meningoencephalitis, the risk of false positive diagnosis of TBE is of little importance. Diagnosis based on detection of TBE IgM antibodies is, in our opinion, sufficient in routine clinical practice and additional confirmatory tests are not necessary. According to a description of a large consecutive sample of TBE cases, the risk of false negative IgM test in early meningoencephalitic phase was 3 /656 [8]. To overcome this low risk for missed diagnosis of TBE, an additional serum sample could be taken later in the acute phase or during convalescence. An alternative simplified approach could be to analyse acute and convalescent sera for TBE in IgM negative patients not fully recovered at three months follow up in order to establish the diagnosis in the fairly high percentage of TBE cases with long lasting sequelae [2, 10]. Confirmatory tests, which include IgG seroconversion in acute and convalescent sera or detection of intrathecal antibody production could be limited to special cases. The increasing problem of TBE vaccinated patients with possible TBE requires methods for detection of intrathecal antibody production and is an important task for qualified virological laboratories, to detect vaccine failure. Detection of TBE neutralising antibodies is rarely required: only in the few patients where interference with other flaviviruses including vaccines is suspected.

With such a TBE case definition and a reporting system including only cases with TBE meningoencephalomyelitis with, as a minimum requirement, the presence of TBE serum IgM antibodies, reliable and comparable surveillance data between countries and over time will be ensured. Introduction of national systems to detect vaccine failures will further add to quality of the TBE surveillance in Europe.