In a brief letter to Biosecurity and Bioterrorism, Alan Zelicoff notes a problem with serosurveys for influenza H5N1 infection:
…peak titers after H5N1 infection occur at about 4 to 6 weeks postinfection and may drop by as much as 32-fold over the course of a year, probably decreasing the sensitivity of serologic testing for past asymptomatic infections. Micro-neutralization testing may be more sensitive.
He cites a serological survey carried out on poultry workers in South Korea, in which 9 of 2,500 subjects were found to have antibodies to H5N1 virus, in the absence of illness. These seropositive individuals carried antibodies that neutralize H5N1 virus infectivity. Assays for antibodies that block infection may be more specific for infection than hemagglutination-based assays. His conclusion:
One can anticipate additional serological surveys that will better inform public health practitioners of the threat to humans from circulating H5N1 clades….morbidity from novel influenza strains does not equate with an impending pandemic, let alone one with high mortality. It would appear likely that a systematic, prospective cohort study is in order to adequately capture the frequency of asymptomatic infection.