For as long as they can remember, the people of the high Andean river valleys of Peru have endured recurring waves of a deadly and disfiguring disease called bartonellosis.
It is always present in their villages. But every four to six years, they suffer a "bad year" - a mysterious surge in the number of cases and severity of the illness.
Victims develop a fever and flu-like symptoms, followed by anemia that kills 40 percent to 60 percent of its victims if untreated. Months later, those who survive develop a terrible rash they call verruga - bleeding warts, all over the body, that last two to five months.
It is an ancient scourge. Anthropologists have found evidence of verruga in pre-Columbian mummies they've examined in the region. Spanish conquistadors described it in their diaries. And an outbreak in the 1870s killed 7,000 people during railroad construction in northern Peru.
But its source, and the reasons for its cyclical nature, remained obscure.
In an unusual collaboration, scientists from the NASA Goddard Space Flight Center in Greenbelt and the Uniformed Services University of Health Sciences have finally begun to unravel the mystery.
In findings reported last month to the American Meteorological Society, team members say they have found evidence linking the periodic surges of bartonellosis in the Peruvian Andes to the mating habits of a lowly sand fly and to the El Nino phenomenon in the waters of the eastern Pacific.
More research is needed to confirm the results. But the findings add to a lengthening list of apparent cause-and-effect links between El Nino and outbreaks of human disease. Those include surges of hantavirus infections in the U.S. Southwest, cholera in South Asia and Peru, dengue fever in Vietnam and malaria in Africa.
The scientific teams working in Peru hope that by improving their forecasts of developing El Nino events, they can provide the advance warnings that Peruvian authorities need to mount cost-effective preventive campaigns against the worst outbreaks, which seem to occur in El Nino years.
"If you could do it at the heaviest periods of time, it would probably have the greatest impact," says Dr. Larry Laughlin, chairman of preventative medicine and biometrics at the Uniformed Services University - a federal medical school and research center in Bethesda.
Laughlin says the university has long had a research interest in emerging infectious diseases in poor countries. But there is a military interest, as well.
"The military rationale was that if for global reasons, and in particular perhaps for drug interdiction reasons, we would have troops in this part of the world, we would have to know something about the disease and help construct control programs," he says.
Bartonellosis appears to be limited to Andean mountain valleys, but its true geography is uncertain. "People haven't looked in a lot of places," Laughlin says.
It is caused by a bacterium that invades the bloodstream, enters the oxygen-carrying red blood cells and destroys them, causing an acute anemia.
"It's a wicked disease if unattended and you don't know what you're dealing with," he says.
It is a seasonal illness, appearing each year, six to eight weeks after the start of the annual rainy season, in November or December.
There was tentative evidence that a sand fly called lutzomyia was involved in transmitting the bacteria. But that hadn't been proven, either. And the episodic nature of the worst outbreaks was simply a mystery.
Laughlin and his team launched their study among 1,600 people living in four villages in a steep mountain valley along the Rio Santos, in Peru's northern Caraz district.
The investigators found that at least 10 percent of the people in the 5-mile-long study area became sick with bartonellosis in a normal year. Most were younger than 15.
"Old folks seem to be protected, probably because they got it when they were kids," Laughlin says.
Fortunately, the study team was able to treat victims early with antibiotics and cut the mortality rate to less than 1 percent. But even after treatment and recovery, they discovered, a quarter of the survivors continue to be reservoirs - carriers - of the bacteria.
Entomologists, meanwhile, collected hundreds of sand flies, analyzed the bugs' meals and confirmed that they prefer human blood.
About 1 percent of the flies also carried the bartonellosis bacteria. But tests of local fleas and mosquitoes found none. "So, we could make a case with supporting evidence that the sand fly is the vector," Laughlin says.
While the team was in the region in 1997, northern Peru began to experience the effects of a strong El Nino, with heavy rains and warmer temperatures. It was followed by a doubling of expected bartonellosis cases - to about 20 percent of the population - in Caraz and an outbreak in Cuzco, in southeastern Peru, where it had not been reported before.
That's when the group began to wonder whether El Nino might be linked to the disease.