To the Editor: The Japanese encephalitis flavivirus is the most common cause of encephalitis in Asia. Death occurs in 25% of clinical cases, and permanent neurological deficits occur in up to 50% of survivors.1 Infection is transmitted from amplifying hosts (primarily waterbirds and pigs) by Culex mosquitoes. Although the virus has been isolated in the Western Province of Papua New Guinea,2 and clinical cases have been described in the Western Province and suspected in the Milne Bay region,3 to our knowledge cases have not been reported from around Port Moresby.

In January 2004, a 66-year-old man of European background was evacuated to our hospital with a 7-day history of fever and confusion. On examination, he had generalised upper motor neurone signs and a Glasgow coma score fluctuating between 6 and 10. Computed tomography and magnetic resonance imaging showed multiple non-specific white-matter lesions bilaterally. An electroencephalogram (EEG) demonstrated diffuse slowing in the delta to theta range in both hemispheres, with preserved response to painful stimulation. Lumbar puncture showed clear cerebrospinal fluid (CSF), with a leukocyte count of 65 × 106 cells/L (81% mononuclear) (reference range [RR], < 5 × 106 cells/L), normal erythrocyte count, raised protein level of 0.79 g/L (RR, 150–500 mg/L); glucose level of 4.3 mmol/L (RR, 2.8–4.0 mmol/L) and negative bacterial and fungal cultures. The CSF was also negative for cryptococcal antigen and by polymerase chain reaction (PCR) testing for enterovirus and herpes simplex, Japanese encephalitis, Murray Valley encephalitis and Kunjin viruses.

Serological tests were negative for syphilis and human immunodeficiency virus infection. Paired sera from Days 2 and 19 of admission were tested in parallel against a panel of flaviviruses using a haemagglutination inhibition assay.4 This showed fourfold rises in antibody titre against dengue virus serotypes 1, 3 and 4, and Japanese encephalitis, Murray Valley encephalitis, Kunjin, Alfuy and Kokobera viruses, and twofold rises in titre against dengue virus serotype 2 and Stratford virus. Overall, these results were diagnostic of recent flavivirus infection but were non-specific. IgM antibody responses to the same flaviviruses were measured in sera and CSF using an in-house enzyme-linked immunosorbent assay (ELISA), with strongest reactivity demonstrated to Japanese encephalitis virus (Box).

The patient had lived in Papua New Guinea since 1970, predominantly on a church-run farm at Bootless Bay, about 20 km from Port Moresby. He had not travelled outside this region in the month before his illness, and had no history of Japanese encephalitis vaccination or of dengue fever. The farm was situated about 150 metres from a piggery. The patient had no direct contact with this piggery. His accommodation was poorly screened against mosquitoes.

After 3 weeks of primarily supportive intensive care, the patient was discharged to a general ward. His neurological recovery was slow. After 5 months, he was able to walk with assistance and required a tracheostomy to protect his airway. He was judged likely to experience permanent neurological deficits.

The clinical, epidemiological, radiological, EEG and serological features of this case strongly support a diagnosis of Japanese encephalitis. Japanese encephalitis virus is difficult to detect in CSF by isolation or PCR because of neutralising antibodies and the limited duration of viraemia, which may have accounted for the negative PCR result in this case, despite the use of a highly sensitive method.5

This case highlights the desirability of further defining the epidemiology of Japanese encephalitis in the Port Moresby region, as well as reconsidering the current recommendation to vaccinate Australians only if they intend travelling to the Western Province of Papua New Guinea.1

IgM antibody levels, measured against a panel of flaviviruses by enzyme-linked immunosorbent assay (ELISA)

JE = Japanese encephalitis. MVE = Murray Valley encephalitis. * IgM levels were measured as the P/N (positive/negative) ratio (ratio of the absorbance of the test sample to the absorbance of a negative control sample tested against the same antigen).