The Real Case Fatality Rate of H5N1 Avian Influenza
Recombinomics Commentary
February 21, 2005
>> Although this 76 per cent human fatality rate looked terrifyingly high, Dr Cox said it might be exaggerated by under-reporting of less serious cases of H5N1, which might not be recognised as avian flu. "Some studies are going on to get a better handle on what the real case fatality case is," she said. For example, poultry workers exposed to the virus would be checked for any H5N1 antibodies in their blood. <<
Although the issue of the "real case fatality rate" has been brought up many times, measuring "any H5N1 antibodies in their blood" may not provide much information on the case fatality rate for the H5N1 in Vietnam, Thailand, and Cambodia that has resulted in reported human fatalities. It is this version of H5N1 that is the most likely precursor to a bird flu pandemic strain.
H5N1 was originally isolated in 1959 in Scotland, but the first H5N1 isolated in Asia was in Guangdong Province in 1996. H5N1 gained attention in 1997 when human cases were identified in Hong Kong. The 1997 H5N1 had a case fatality rate of 33% in patients (6 of 18 infected patients died). However, culling eliminated that virus in Hong Kong, and 2004 H5N1 in Vietnam and Thailand is quite distinct, even though both are H5N1.
These differences can be measured in many ways. One is simply by looking at antibodies generated by the isolates. A pandemic vaccine was created against the 1997 H5N1, but the 2004 is so different that the vaccine was considered largely ineffective, and new vaccines against the 2004 version are being created.
The 2004 H5N1 has a 20 aa deletion in NA as well as a 5 aa deletion in NS. These two mutations are found in most of the H5N1's isolated in Asia in 2004, but there were subtler differences that distinguished the 2004 isolates. There were a number of regional specific polymorphisms found in most of the genes. Some of the polymorphisms were just in the isolates from Vietnam, others specific for Thailand, and a large number just found in Vietnam and Thailand. Since the only two countries reporting human fatalities in 2004 were Vietnam and Thailand, these polymorphisms unique to those two countries defines H5N1 strains that produce the 76 per cent human fatality rate mentioned above. One example is the two mutations in the M2 gene that lead to resistance to Amantadine and Rimantadine. These two mutations in the same gene are found only in isolates from Vietnam and Thailand
Thus, a true fatality rate for 2004 H5N1 in Vietnam and Thailand would be limited to people infected with that particular version of H5N1. However, H5N1 has been detected in live markets in Hanoi as early as 2001. The virus has also been isolated from nearby Guangxi Province in China in 2001. There were also many H5N1 isolates from Hong Kong in every year since 2000, but no reports of Hong Kong human fatalities caused by the 2004 H5N1 found in Vietnam and Thailand (which are somewhat different than the two 2003 human isolates from Hong Kong which probably originated in Fujian Province). Because these earlier non-lethal virus were circulating in southeast Asia, merely measuring antibody levels to H5N1 in poultry workers would not address the true case fatality rate of the 2004 version.
However, more direct analysis points toward a high case fatality rate. The rate in Vietnam and Thailand at the beginning of last year, as well as over the summer, yielded rates in the range of 70-80% for each country in each time period. A similar rate has been generated for this season, although there may be some differences between cases in southern and northern Vietnam.
An expanded look at antibody levels was done in Thailand and described in a recent report. 33 confirmed or suspected cases were identified. 8 of the 12 confirmed cases died, yielding a case fatality rate in the same range as the official cases in Vietnam and Thailand. More importantly however were the more than 600 cases who had milder respiratory disease and were negative for H5N1. If milder cases were common and these cases had antibody to H5N1, they would have been placed in the confirmed category and sharply reduced the case fatality rate. Thus, the study almost doubled the number of H5N1 lab confirmed cases or deaths, but did not identify large numbers of H5N1 patients with lower case fatality rates.
Indirect evidence for the lethality of the virus in mammals also comes from the H5N1 outbreak at the tiger zoo in Thailand. Although many of the 441 tigers were probably not exposed to H5N1 because of quarantine away from the sick tigers, 147 tigers died, which would give a case fatality rate between 33% and 100%, depending on how many of the 441 tigers were actually infected with H5N1.
Thus, all of the hard data for H5N1 in Vietnam and Thailand for 2004 indicates the virus is quite lethal in children, young adults, and tigers.
These data provide little support for large numbers of individuals infected with H5N1 in 2004 that did not die.
Media link