While seroprevalence of HPAI H5N1 virus antibodies was low, we were able to detect evidence of human infections with HPAI H5N1 virus in Hanoi, Vietnam during 2001, after the 1997 outbreak of clade 0 HPAI H5N1 cases in Hong Kong and before the emergence of clade 1 HPAI H5N1 cases in 2003 
. Overall, 8 Hanoi MPWs had antibodies against one or more HPAI H5N1 viruses tested. HPAI H5N1 virus antibodies were detected at a higher frequency among MPWs than among controls, although this was not significantly different. Two non-poultry worker controls were seropositive for antibodies to both HPAI H5N1 viruses isolated in Hong Kong in 1997 and 2003, highlighting the importance of non-occupational exposure to poultry as a risk factor for HPAI H5N1 virus infection 
. Serologic evidence of human infection with an HPAI H5N1 virus isolated from healthy geese in two Hanoi poultry markets at the time of the study was found in 2 MPWs, but not in controls.
The 4% seroprevalence of clade 0 HPAI H5N1 virus antibodies in 2001 among MPWs was slightly higher than reported for clade 0 or clade 1 HPAI H5N1 virus antibodies among a similar population in southern China during 2007–08 (0.8%) 
, but less than the estimated 10% seroprevalence of clade 0 HPAI H5N1 antibodies among Hong Kong poultry workers during 1997–98 
. Since the kinetics of the HPAI H5N1 virus antibody response in infected humans with asymptomatic or clinically mild illness suggest that antibody titers decline below the WHO seropositive threshold 
over 6–12 months following exposure 
, the seropositives we identified most likely indicate relatively recent HPAI H5N1 virus infection in relationship to sera collection in October 2001. It is possible that other HPAI H5N1 virus infections could have been missed in the study participants if a limited antibody titer response occurred or if the titer had declined below the threshold for a seropositive. Nevertheless, these studies and others indicate the potential for HPAI H5N1 virus to cause asymptomatic infection or mild illness in adults. We also identified a similar low seroprevalence of antibodies to LPAI H9N2 virus among MPWs and non-poultry worker controls in Hanoi, suggesting that occupational exposure to poultry was not a major risk factor for LPAI H9N2 virus infection among the study population in 2001.
These findings are limited by the cross-sectional study design, and the low seroprevalence of H5 or H9 antibodies, for which the study was underpowered to detect significant differences to identify risk factors for human infection with HPAI H5N1 or LPAI H9N2 viruses. Since both study populations were selected by convenience sampling, and the MPWs were older than controls, our findings may not be applicable to other populations in Vietnam. A particular strength of this study is the extensive laboratory methods performed to ensure reliability of the serological results, including performing adsorption with H2N7 and H3N2 viruses to minimize detection of H9 seropositivity due to cross-reactivity with antibodies to human influenza A viruses. In one H9N2 vaccine study, individuals born before 1969 exhibited pre-vaccination neutralizing titers against H9N2 (G1) 
. In another H9N2 vaccine study that was conducted in the U.S., baseline pre-vaccination H9N2 (G1 and G9 lineages) virus neutralizing antibody titers were higher in persons aged 44–59 years compared to 18–38 years, whereas titers detected by hemagglutination-inhibition assay were higher to the G9 lineage virus compared with the G1 lineage virus in both age groups 
. Following adsorption with H2N7 virus, we observed a reduction in the prevalence of individuals with neutralizing antibodies to H9N2 viruses in persons aged ≥33 years. We also observed a reduction in the prevalence of neutralizing antibodies to H9N2 viruses in persons aged <33 years after adsorption with H3N2 virus. Taken together with the previous studies, these findings suggest that there may be age-, virus-, and assay-specific differences in cross-reactivity when detecting antibodies to H9N2 virus, and that additional serologic investigations are needed. Of note, other published H9N2 virus antibody seroprevalence studies either did not address or did not perform additional laboratory tests to reduce or eliminate potential cross-reactivity with other human influenza A virus subtype antibodies which may limit the interpretation of their results 
The findings from this study can serve as a baseline for additional serologic studies to assess avian-to-human transmission of HPAI H5N1 and LPAI H9N2 viruses as the prevalence of these evolving viruses has increased substantially among poultry in northern Vietnam since 2001.