Drought may bring monkeys, mosquitoes, and yellow fever closer to Brazil’s cities

Wet weather usually means more mosquito-borne disease – more rain, more standing water, more mosquitoes. Yet Brazil’s first urban yellow fever epidemic in nearly a century broke out in the middle of a historic drought. A new model suggests the drought itself may have driven the outbreak by pushing thirsty monkeys and forest mosquitoes into the cities.

Jamie Caldwell, Ph.D., a disease ecologist at Princeton University‘s High Meadows Environmental Institute, led the research. Her co-authors were Princeton’s Bryan Grenfell and Gabriel Vecchi, along with Joelle Rosser of Stanford University School of Medicine.

The team can’t rerun 2017 without a drought, so the model doesn’t prove the connection. What it shows is that the epidemic’s strange shape only makes sense when the drought’s effects are included.

Yellow fever returned to Brazilian cities

Yellow fever usually starts with fever and flu-like symptoms. In severe cases it causes internal bleeding and organ failure, and the jaundice that can yellow a patient’s skin gave the disease its name.

By the 1930s, aggressive mosquito control and an effective vaccine had stopped the virus from spreading in Brazilian cities. The virus never disappeared, though. It kept circulating in the forest, passing between forest mosquitoes and monkeys and occasionally infecting people at the forest edge.

Starting in late 2016, the virus moved back into the cities. Cases spread through the state of Minas Gerais and into neighboring states, including Rio de Janeiro and São Paulo. More than 2,000 people fell ill, and more than 700 of them died.

Drought pushed animals toward cities

The epidemic wasn’t the region’s only once-in-a-century event. By a standard drought index, the dry spell leading up to the outbreak was 9.5% more extreme than a once-in-a-century drought. The extreme dry conditions were in place four months before the cases began.

Caldwell’s team suspected the timing was no accident. As water disappeared from the forest, howler monkeys and marmosets, two primates quick to adapt to city edges, would have moved toward town looking for it. Haemagogus, the forest mosquito that carries the virus, can fly long distances and would have followed.

Dry air may have changed the insects’ habits too. Earlier research has suggested that mosquitoes bite more often in dry conditions, taking in fluid from blood to keep from drying out.

And during the outbreak, far more monkey carcasses turned up in and around the city than in a typical year.

Animal movement made the model work

To test the idea, the team built a mathematical model that simulates how the virus passes among two mosquito species, howler monkeys, marmosets, and people. They fed it data from Minas Gerais, where the outbreak began: how long each animal lives, how likely a bite is to pass the virus, and how fast hosts recover.

Then they ran four versions of it. The full version let the forest animals move toward the city with the seasons and allowed mosquito biting frequency to change as well. The other versions switched off one or both of those behaviors.

Only the full version reproduced the epidemic. Its predictions matched the recorded cases in monkeys and people alike, scoring 0.90 on a scale where 1.0 is a perfect match. The stripped-down versions mistimed the first wave of infections or missed the second wave entirely.

Caldwell answered Earth.com’s questions by email, and said what surprised her was that the model needed both behaviors at once.

“At first, it looked like increased mosquito biting frequency alone wouldn’t explain the pattern,” Caldwell told Earth.com. “But when we incorporated it into the full model, it was essential to reproducing both the timing and magnitude of the outbreak.”

Forest mosquitoes drove the epidemic

The species usually blamed for urban outbreaks barely figured in this one. Aedes aegypti, the city-dwelling mosquito that spreads dengue and Zika, is also the mosquito historically associated with urban yellow fever transmission.

This time, mosquito surveys kept finding infected Haemagogus. Despite extensive sampling, they never found one infected Aedes aegypti.

Part of the reason was numbers. Spraying during the 2015 and 2016 Zika outbreaks had already cut the city mosquito’s numbers to between a third and a half of the previous decade’s. Feeding habits likely mattered more: Aedes aegypti feeds almost entirely on people, while Haemagogus will bite monkeys and humans alike.

When the researchers removed Aedes aegypti from the model, the epidemic’s course barely changed.

Vaccination had the strongest effect

The team also replayed the epidemic to see how different interventions would have changed its course. One simulation sprayed insecticide every rainy season, while another limited the number of forest animals entering the city as a stand-in for conservation measures.

A third simulation began vaccination three months earlier than the real campaign and continued with routine childhood vaccination.

None of the interventions prevented the second wave, even when the researchers combined all three.

Vaccination still had the strongest effect, and its benefit kept growing over time. Even with 90% of people vaccinated in the simulation, the second wave arrived anyway. Monkeys carry the virus whether or not people are immune.

Combined defenses limited transmission

Asked by Earth.com what other findings stood out, Caldwell pointed to that second wave. The defenses still did real work: they kept the virus from settling in for good.

“One finding that surprised me was that interventions could not prevent the second wave of the outbreak. However, the interventions still had an important impact: they reduced longer-term transmission and helped prevent the virus from establishing an endemic cycle,” she said.

She added, “Even when one intervention is the dominant control measure, additional interventions can provide an important buffer and reduce the risk of longer-term transmission.”

Brazil, in reality, used all three measures to different extents, and no new cases were reported after 2018.

Drought could give months of warning

A drought builds slowly, and that slowness is useful. The extreme dry conditions arrived four months ahead of the outbreak. In principle, that’s time enough to start vaccinating before the first case.

Caldwell said the data for that kind of warning already exists. Meteorologists track droughts with a range of standard tools, she explained. But public health teams rarely use that data to watch for disease risk, because solid evidence tying drought to outbreaks has been sparse.

In the Earth.com interview, she named the countries she’s watching.

“Bolivia, Colombia, Guyana, and Peru face several of the same conditions we saw in Brazil, our study region, that could make them vulnerable to a drought-influenced outbreak,” she said. “All four countries have also recently seen an uptick in yellow fever virus reports, making them particularly important places to watch.”

Animal tracking could improve forecasts

The model leaves one big question open: the researchers had no fine-scale record of where the monkeys and mosquitoes went as the water disappeared. That movement enters the model as a seasonal pattern, not a measurement. Tracking the animals through a real drought would turn this explanation into a forecasting tool.

The study projects the region’s harshest dry spells will be 47% drier by century’s end, under a moderate emissions path, than they were between 1970 and 2000.

Until the tracking data exists, the four countries on Caldwell’s watch list are where the idea will be tested next.

The full study was published in the journal Science Advances.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–