Person-to-person transmission of influenza A (H5N1).

In February, 2004, in response to suggestions that limited person-to-person transmission of strain H5N1 of the influenza A virus might have occurred, the then Prime Minister of Thailand, said “the possibility of person-to-person transmission is 0·00001%”. Temperatures were running high, and any mention of person-to-person transmission of H5N1 was thought by some to be reckless. Although we have moved on, an air of tension still surrounds this disease, particularly in the corridors of power within the international health and political communities, and apprehension remains about what the continued pandemic in poultry means for human health. Human H5N1 infections thankfully remain rare, but evidence has accumulated that in some circumstances H5N1 viruses are capable of limited person-to-person transmission.

Person-to-person transmission of H5N1 was first mooted after the 1997 Hong Kong outbreak, in which family members and at least two health workers might have been infected by contact with patients.2, 3 Since then, one report of a family cluster concluded that person-to-person transmission was probable, and an additional four reports stated that it could not be ruled out in at least six families.5, 6, 7, 8 In today's Lancet, another convincing report of probable person-to-person transmission is published by researchers from China and the USA.

Transmission of avian influenza virus between mammals is not, however, restricted to H5N1 in human beings: H7N7 can also be transmitted from person to person, and there is evidence of transmission of H5N1 among other mammalian species. Given that the species barrier can be breached, the intriguing question is why the transmissibility of H5N1 among people remains so low?

Successfully crossing the species barrier is itself a step that might select viruses partly adapted to the new host and, once infected, the new host might select subpopulations of the virus adapted to the new environment. Apparent host-induced selection of a human-adapted H5N1 virus has been reported, but most sequence analyses of viruses isolated from sporadic and clustered cases show no substantial differences from the strains found in poultry. However, most of these analyses used viruses isolated from human specimens in cell cultures of animal origin, thus grown under different selective pressures from those in human respiratory tracts. Whether better adapted viral subpopulations exist during human infection is unknown, as is the diversity in virus populations in individual human beings or poultry. A possible lack of host-induced evolutionary pressure, disseminated viral replication and high viral loads in infected people, and the rarity of person-to-person transmission suggest that the infecting avian viruses might be already well adapted to the individual in which they find themselves, but not to the wider human population.

With the exception of occasional infection in health workers, all published incidents of possible or probable person-to-person transmission report transmission between genetically related individuals. Although this finding could be related to the intensity and intimacy of contact between family members, host genetic factors might also play a part in susceptibility to H5N1, and when we see limited person-to-person transmission we might be observing an interaction between two well matched subpopulations. Studying both within-host virus diversity and host genetic diversity might help to clarify the nature of the species barrier and the conditions necessary for widespread transmission between people. These studies are hard to do and need sustained, coordinated, and collaborative efforts, as in Wang and colleagues' study, coupled with acceptance that person-to-person transmission of H5N1 can and does happen.

Whatever the underlying determinants, if we continue to experience widespread, uncontrolled outbreaks of H5N1 in poultry, the appearance of strains well adapted to human beings might be just a matter of time. In the meantime, all family contacts of a patient with probable or confirmed H5N1 should be given chemoprophylaxis and placed under surveillance. Personal protection and advice must be extended to the family members and health workers visiting and looking after patients in hospital. These steps rely on having systems capable of detecting cases and clusters early enough to start treatment and isolation, to identify and protect contacts, and to assess the extent of transmission.

Tension around emerging infections has led to often acrimonious and dispiriting debates about sharing of samples, international collaboration, and the roles of academic institutions, governments, and international agencies in fighting this common threat. Today's study is a superb piece of epidemiological work showing the benefit of a longstanding and trusting international collaboration that began during the severe acute respiratory syndrome epidemic. Such collaborations sustained over several years, centred in affected countries, and closely linked with WHO are our best chance of combating current and future threats to international health and ensuring that benefits are shared worldwide.