Some people in Papua New Guinea who once feasted on their own relatives did not succumb to the prion disease kuru until 50 years later, say researchers who have laboriously tracked down the last sufferers in remote villages. The discovery renews concern that another human prion disease, variant Creutzfeldt–Jakob disease (vCJD), might be incubating silently in some populations and could rear its head decades from now.

Neurologists have long been fascinated by kuru, which caused an unprecedented epidemic of neurodegenerative disease in the Fore people of Papua New Guinea that peaked in the 1950s and early 1960s. In death rituals, families steamed and ate the bodies of their relatives — along with, it was later discovered, infectious prion proteins that caused the debilitating and fatal disease. The ritual was prohibited in the mid-1950s by the Australian authorities who governed that part of Papua New Guinea, and the disease eventually became less frequent.

Interest in kuru reawakened with the realization that vCJD, transmitted from cows infected with bovine spongiform encephalopathy, might cause a similar epidemic among those who ate infected meat in the 1980s and 1990s. So far, only 156 deaths from vCJD have been reported in Britain, the worst-affected country, and the number of new cases peaked in 2000, suggesting that vCJD takes about ten years to incubate. But debate continues about whether another wave of cases could yet appear.

John Collinge at University College London and his colleagues went to Papua New Guinea to find out. Most people with kuru have already died, but the team ramped up existing disease monitoring to find the last of the epidemic. Working with local communities, they scoured isolated villages that are typically more than 2,000 metres above sea level, lost in dense rainforest and connected only by tracks. “It's arduous trekking,” says Collinge.

Between 1996 and 2004, the researchers found what they believe are the last 11 cases of kuru. Patients' histories were collected, to piece together when they were probably infected. The longest incubation time was calculated to be at least 56 years, and perhaps seven years longer — although the average incubation time seems to be 12 years (J. Collinge et al. Lancet 367, 2068–2074; 2006). “For the first time we can see the extraordinary incubation period in human prion disease,” says Collinge. “It's sobering that, half a century on, this disease has not disappeared.”

Collinge says that vCJD could have a much longer average incubation time, of 30 years or more, because the prions are passing from cows to humans rather than between humans. A species barrier extends incubation times in animal tests. People who have already succumbed to vCJD might have been particularly genetically susceptible, as other evidence has suggested.

Mathematical models used to predict the size of a vCJD epidemic could now include these findings. “Most people seem to think we're over the worst,” Collinge says. “We have to be cautious about assuming this disease is going away.”