Hong Kong media were full of lurid accounts Monday of pigs staggering around with blood pouring from their bodies in Gaoyao and neighboring Yunfu, both in Guangdong Province. Apple, a daily newspaper here, said that up to 80 percent of the pigs had died in the area, that peasants were engaged in panic selling of ailing animals at deep discounts and that pig carcasses were floating down the river.

Medical experts said that the extent of the reported bleeding from the pigs, including bloody skin lesions, did not sound like common symptoms of bird flu, but added that the pig deaths needed to be investigated. Because pigs can be infected with many avian and human influenza viruses, the most popular scientific model for how avian influenza viruses cause pandemics in humans is that human and avian influenza viruses exchange genetic material when they infect a pig at the same time.

The above comments on the swine deaths in Guangdong province are cause for concern. Bleeding under the skin was common in the 1918 pandemic, and bleeding from the mouth has been frequently reported in bird and human H5N1 infections, including clusters in Thailand, Turkey, and Iraq.

Moreover, the only country to report H5N1 to a public database has been China (see list below). Several different strains of H5N1 have been reported in Guangdong province including Fujian, Qinghai, and the sequences from the Shantou geese.

The co-circulation of distinct strains also offers the opportunity for dual infections and recombination and the recent spread of the disease in swine coincides with increased migration.

It was almost exactly two years ago that the Qinghai stain was reported at Qinghai Lake in central China. The sequences in the Qinghai isolates had links to Shantou isolates.

The large number of dead swine should be investigated aggressively. The outbreak at Qinghai Lake in 2005 signaled a major extension of the geographical reach of H5N1, and intensive testing will be required to rule out H5N1 in these swine deaths. Involvement of H5N1 in these deaths would be a significant event in the evolution of H5N1, which is becoming increasingly complex as its geographical reach continues to spread.

A detailed investigation into the etiology of these swine deaths would be useful.