Review finds little evidence of missed H5N1 infections

Feb 6, 2013 (CIDRAP News) – The authors of a new review say there is little evidence of unrecognized human cases of H5N1 avian influenza, but the studies done so far have too many limitations to settle the controversial question.

Researchers from the Center for Biosecurity at the University of Pittsburgh Medical Center reviewed 29 serologic studies and found only four that identified anyone carrying antibodies to currently circulating H5N1 strains. But many of the studies reviewed had methodologic problems, such as lacking a comparison group with no exposure to the virus.

The authors say their findings suggest that mild or asymptomatic cases H5N1 cases are probably few, but the studies done so far are not capable of determining "the true prevalence or severity of H5N1 infections." Their findings were published online yesterday in Clinical Infectious Diseases.

The debate over the existence of mild or asymptomatic H5N1 cases has continued for years and was renewed last year amid the controversy over experiments in which lab-modified H5N1 viruses were found to have airborne transmissibility in ferrets.

The question centers on whether H5N1 is really as dangerous as it appears from the 59% case-fatality rate (CFR) in confirmed human cases recognized by the World Health Organization (WHO).

Some scientists argue that many mild or asymptomatic human cases may have gone unrecognized, which would mean the 59% CFR is too high and the virus is less fearsome than it seems. But others point to a number of serologic studies conducted in recent years, in which few people in H5N1-affected areas were found to have antibodies suggesting previous mild or unrecognized infections.

The Pittsburgh researchers say their aim was to conduct a more detailed review of all published H5N1 serologic surveys that tested for antibodies in asymptomatic people. They found 30 studies that met their criteria, but dropped one because all the subjects had received preventive antiviral (oseltamivir) treatment.

Discussing the studies' limitations, the authors say none of them included control groups that met their standard for "optimal unexposed controls," meaning individuals with essentially no chance of H5N1 exposure. Also, 9 of the 29 studies did not use the WHO's criteria for a positive test result for H5N1 antibodies, which calls for a titer of 1:80 or greater.

Seven of the 29 studies focused on H5N1 clades and genotypes that have not been detected in many years, such as the strain that caused the original human outbreak in Hong Kong in 1997, the report says. In general, these seven studies found the highest seropositivity rates of the 29 included in the review.

Among 16 studies that used the WHO's seropositivity criteria and involved current strains, 12 found no positives, while 4 found at least one positive. Among them were 2 studies conducted in Cambodian villages in 2006 after poultry H5N1 outbreaks, which found seropositive rates of 1% and 2.6%.

The third study found that 1 of 110 poultry sellers in Guangzhou province, China, who had slaughtered birds daily for 5 years, had H5N1 antibodies. In the fourth study, done in 2001, 3% of 200 Vietnamese poultry workers tested positive.

In summary, the report says, "There were only 4 studies using WHO criteria that show seropositive results to the genotype Z viruses that have spread throughout Asia, Europe, and Africa, and all 4 involved clade 1 viruses prior to 2007."

It adds that clade 1 was found only in Indochina and southern China and has now largely been replaced by clade 2.3 viruses. Therefore, there is no clear evidence of unrecognized infections with any of the clade 2/genotype Z now circulating in most affected countries.

On the other hand, antibody titers wane over time, which can cause cases to be missed, the authors caution. "It is essential that more specific tests for prior H5N1 infection be developed," they add, noting that measurement of T-cell responses is one possibility.

"The results of H5 serological studies that have been conducted to date have been useful but are not capable of establishing the true prevalence or severity of H5N1 human infections," the report concludes.

In an accompanying commentary, David M. Morens, MD, of the National Institute of Allergy and Infectious Diseases, notes that studies of household clusters of H5N1 cases suggest that such cases are most likely to occur in blood relatives, indicating the possibility of a specific genetic susceptibility to infection.

"If the confusing serosurveillance data can really be explained by such host susceptibility factors, it might suggest that H5N1 is less likely to emerge to cause a pandemic, and might also provide a mechanism to identify those at risk if it did emerge, as well as if it remains enzootic," Morens states.

He endorses the authors' appeal for the use of unexposed comparison groups in serologic studies. Also, he writes, it may be important to study "highly exposed non-cases" identified in connection with confirmed cases, along with possible host risk factors for infection and disease.

Toner ES, Adalja AA, Nuzzo J, et al. Assessment of serosurveys for H5N1. Clin Infect Dis 2013 Feb 5 (Early online publication) [Abstract]

Morens DM. Pandemic H5N1: receding risk or coming catastrophe? (Commentary) Clin Infect Dis 2013 Feb 5 (Early online publication) [Landing page]

See also:

Feb 9, 2012, CIDRAP News story "Undetected H5N1 cases seem few, but questions persist"

Newsletter Sign-up

Get CIDRAP news and other free newsletters.

Sign up now»


Unrestricted financial support provided by

Bentson Foundation Gilead 
Grant support for ASP provided by


  Become an underwriter»