We assessed parakeets and 3 species of songbirds for their susceptibility to avian-origin A(H7N9) (A/Anhui/1/2013) virus and found that they were highly susceptible to infection with this isolate. Shedding was limited to the oropharynx, which may have reduced contact transmission; however, at most times sampled, the water troughs contained large amounts of virus. Furthermore, low-pathogenicity influenza viruses (those lacking a multibasic cleavage site in the hemagglutinin protein), such as the A(H7N9) isolate used here, remain stable in water longer than their highly pathogenic counterparts (22) and could therefore serve as an inoculum for cage mates. Despite little shedding in the contacts, seroconversion of at least 1 contact animal in each of the finch and sparrow groups indicates exposure with antigenic epitopes of the subtype H7 hemagglutinin. The minimal transmission to direct contacts observed in our study is consistent with previous observations of avian influenza virus infection of songbirds (15,20,23–25). The parakeets in particular showed no contact animal shedding or seroconversion, a phenomenon we have previously observed with an A(H3N8) isolate from a songbird (R.G. Webster et al., unpub. data). This observation may have implications for risk assessment of this species in the pet bird trade; however, the lack of proper influenza transmission data with this species in the literature warrants additional studies to confirm the validity of this observation. One direct-contact society finch shed virus equivalent to titers in inoculated birds and shed for a longer period, which suggests that efficient transmission and replication in contact animals, although rare, is possible. The ability of these small birds to harbor and shed A(H7N9) viruses, usually with few signs of illness, creates a substantial potential for transmission to humans, as well as to poultry and wild birds.

Interspecies transmission has not yet been investigated, and the extent to which A(H7N9)-infected finches, sparrows, and parakeets may transmit virus to other species, including through shared water sources, is unknown. Two host groups have the greatest potential interaction with small birds. The first is domesticated poultry, primarily chickens but also a wide variety of gallinaceous and game birds. The peridomestic nature of songbirds facilitates an interaction with poultry in large production facilities and in backyard farms (14,15). In these cases, they may share common food and water sources (15). The interspecies transmission of influenza from songbirds to poultry is not without precedent. Nestorowicz et al. proposed that highly virulent A(H7N7) isolates from starlings and chickens were closely related and indicated that the virus had been transmitted between these 2 species (23). Forrest et al. experimentally demonstrated waterborne transmission of highly pathogenic A(H5N1) virus from chickens to starlings (15), although the high death rate of the inoculated birds might have limited the degree of interspecies transmission. In the case of the low-virulence A(H7N9) viruses, most inoculated birds shed virus while remaining clinically healthy. The absence of illness and death in A(H7N9)-infected birds has implications for a greater quantity and duration of virus shedding into the environment, as well as higher activity levels and likelihood of interaction with other susceptible hosts. Preliminary data suggest that chickens and quail are highly susceptible to infection with this human A(H7N9) isolate and that the virus is readily transmitted by direct contact (9,10). The high titers recovered from the shared water troughs of all species tested in our study suggest that the virus could be transmitted to poultry and other birds through this route. Therefore, reduced interaction of domestic poultry with wild passerine birds is advisable, although this precaution might not be feasible in developing countries where numerous backyard farms lack biosecurity. Comprehensive biosecurity also is often lacking at large poultry farms in industrialized and developing countries and should be enhanced to limit the access of wild songbirds to the poultry’s water and food sources.

A second host group with high potential to interact with songbirds and other small terrestrial birds are humans. Finches, sparrows, and parakeets are not only common in the wild but are popular pets worldwide (26). They are often sold in the live-bird markets of eastern Asia, where the risk for zoonotic influenza transmission of H7N9 and other influenza subtypes is already established (26,27). Such pet birds may be procured from the wild and may have been exposed to a variety of pathogens before entering the market chain (9). In China, the keeping of pet birds is associated with luck (28) and is common among elderly men, who often stroll through the parks with their caged birds (28,29). In caring for such pets, their owners could become infected by virus contaminated drinking water or from fomites on the feathers (deposited while bathing in water troughs or from saliva while preening) (30). This same demographic group (elderly men) experienced disproportionate rates of illness and death from A(H7N9) infection in China (31). A recent epidemiologic study by Rivers et al. concluded that the comparatively higher rates of infection among the elderly than among younger age groups cannot be entirely attributed to increased exposure to poultry. Furthermore, they assert that an “as-yet unknown epidemiological or immunological feature” may explain the high infection rates among older persons (27), which leaves open the contribution of alternate exposure sources, such as infected pet birds, as a possibility. Virus also could be transmitted to humans through religious ceremonies, such as the Buddhist practice of “merit release,” in which a songbird is purchased, held to the face, kissed, and released (25). Simply owning a pet bird increased a household’s rate of seroconversion during the 2003 A(H7N7) outbreak in the Netherlands (32). Although completely avoiding contact with pet birds during an avian influenza outbreak might not be feasible, surveillance of such species in the markets, and perhaps in the wild, would help to identify or rule out previously unsuspected hosts that might support or disseminate emerging viruses.

To expand on our findings, future studies should include the examination of genetic changes in the human A(H7N9) virus during replication and transmission in the songbirds and parakeets. The human isolate we used contained genetic markers (HA: S138A, G186V, and Q226L; PB2: E627K) indicating adaptation to mammals (33). Sequence analysis of virus shed by the small birds in our study might indicate which species would be most receptive to transmission. The loss of the markers of mammalian adaptation and reversion to an avian-like genotype might predispose the virus to transmission from the small birds to poultry and other birds, rather than to humans. Additionally, studies that assess the sharing of housing, water, and food by inoculated songbirds and poultry (particularly chickens or quail) would shed light on the interspecies transmission potential of A(H7N9) viruses.

Our demonstratration that parakeets and multiple species of songbirds are susceptible to influenza A(H7N9) virus isolated from humans during the recent outbreak in China further supports the possible contribution of songbirds and parakeets to the ecology, maintenance, and transmission of novel A(H7N9) viruses. Finally, they lead us to propose that finches, sparrows, and parakeets may be intermediate hosts and sources of A(H7N9) viruses and that their frequent interaction with wild birds, domestic poultry, and humans renders them a particular risk factor in the emergence and transmission of novel influenza strains.