Sir

News that Thai dogs have tested positive for antibodies to the influenza A H5N1 virus (“Thai dogs carry bird-flu virus, but will they spread it?” Nature 439, 773; 200610.1038/439773a) reinforces our notion that carnivore and scavenger species have the potential to act as important sentinel hosts for emerging human and livestock diseases, providing a valuable tool for surveillance and for determining spatial and temporal patterns of infection.

Domestic dogs may prove particularly useful as sentinel hosts, especially in developing countries. They are ubiquitous, with one dog for every 7 to 21 people in most parts of Africa and Asia. Dogs are known to be susceptible to a wide range of emerging human infections, and, as free-roaming scavengers in many parts of the world, they effectively ‘sample’ widely from a community environment. Despite appearances, domestic dogs in most developing countries are generally accessible for safe handling and sampling.

Our experience in Africa and Asia suggests that sampling dogs for disease surveillance would be particularly cost-effective if carried out in combination with rabies vaccination campaigns, as this provides owners with a strong incentive to participate. During these campaigns, several hundred dogs per day could be accessible for sampling at a cost of US$1–2 per dog vaccinated (K. Bögel and F. X. Meslin Bull. World Health Organ. 68, 281–291; 1990).

Domestic dogs, like other carnivore and scavenger species, may act as ‘bioaccumulators’ of pathogen exposure, with consumption of infected host material resulting in high rates of seroconversion. We suggest that they could therefore usefully be included as part of surveillance strategies to increase the efficiency of pathogen detection, particularly for pathogens that occur at low prevalences in animal reservoirs or are maintained in wild animal populations that are difficult to sample.

Age–seroprevalence data can also allow timing of outbreaks to be established retrospectively and with reasonable accuracy, for at least a number of years. This would be particularly valuable in areas where reporting and laboratory confirmation of human and animal disease outbreaks are limited, which may apply in many parts of the developing world.