On 21 May 1997, the fifth day of his hospitalization, a three-year-old boy from Hong Kong died in an intensive-care unit in Hong Kong, with a final diagnosis of Reye syndrome, acute influenza pneumonia and respiratory distress syndrome (ARDS). No indications of other underlying disease, including immunodeficiency or cardiopulmonary disease, were observed. From a tracheal aspirate, we isolated an influenza virus in MDCK and LLC cells. We were unable to grow any pathogenic bacteria from respiratory specimens. In haemagglutination inhibition assays, the virus did not react with ferret antisera to recent isolates of human and swine subtypes. By complement fixation tests we found it to be an influenza A virus, confirmed by an indirect immunofluorescence assay on cells from the original tracheal aspirate.

Haemagglutination inhibition assays using antisera to 14 H subtypes showed that the isolate was an H5 influenza A virus. Biochemical neuraminidase inhibition tests, using antisera to nine N subtypes, indicated that the neuraminidase was of the N1 subtype. Nucleotide sequence analyses of parts of the H and N genes of the virus allowed a phylogenetic comparison with other influenza viruses. Our analyses clearly confirmed that the virus is of the H5N1 subtype (Fig. 1). The H5N1 virus was indeed isolated from the sample of the child, and could not be attributed to a laboratory cross-contamination, as there was no H5N1 virus present in the human diagnostic laboratory. Furthermore, we reisolated the same virus from the original tracheal aspirate specimen.

Figure 1: Phylogenetic relationship of the H and N genes from the human virus isolate A/Hongkong/ 156/97 (HK97) and the corresponding genes from selected influenza A viruses.
figure 1

Japan57, A/Japan/305/ 57 (H2N2); CkPen83, A/chick/Pennsylvania (H5N2); TyEng91, A/turkey/England/5092/91 (H5N1); SwNed-96, A/swine/Netherlands/609/96 (H1N1); Texas91, A/Texas/36/91 (H1N1); Wuhan95, A/Wuhan/359/95 (H3N2); USSR77, A/USSR/90/77 (H1N1); ParUl73, A/parrot/Ulster/73 (H7N1); Bangk79, A/Bangkok/1/79 (H3N2). Sequences were from GenBank or nucleotide sequence analysis using the Dye Deoxy Terminator kit (Applied Biosystems, Foster City, CA). Alignment was done with 827 nucleotides (38-865) of the H genes and 200 nucleotides (1-200) of the N genes by PILEUP (GCG, Madison, WI). Phylogenetic trees were constructed in Phylip (Felsenstein, Seattle, WA) using DNAPARS and DRAWGRAM.

The contribution of the influenza A H5N1 virus infection to the child's disease, eventually leading to his death, is not yet clear. But the virus identification is important as it is the first documented isolation of an influenza A virus of this subtype from humans. Subtype H5 influenza A viruses can cause lethal avian influenza (fowl plague), a disease that may decimate flocks of domestic poultry. These viruses have so far not been identified in pigs. We feel that the identification of the H5N1 influenza A virus and its presently unknown pandemic potential, should be the basis for an intensive monitoring of the epidemiology and the clinical manifestation of infection with this virus by the international World Health Organization influenza surveillance network.