If peak influenza activity is an accurate measure, flu viruses prefer the weather either cold and dry or humid and rainy, according to an analysis of climate variables and flu patterns around the world.

Researchers from several US universities and the National Institutes of Health (NIH) reached this conclusion primarily by studying flu and climate data from 78 sites in 40 countries. Sponsored in part by the NIH, the study was led by geographer James D. Tamerius, PhD, of Columbia University as first author.

"The paper, published in PLoS Pathogens, presents a simple climate-based model that maps influenza activity globally and accounts for the diverse range of seasonal patterns observed across temperate, subtropical and tropical regions," the NIH said in a press release yesterday.

Flu epidemics strike during the winter in temperate regions, but the seasonality of flu is less clear in the tropics, where outbreaks tend to occur during rainy seasons or year-round. The findings suggest that rain and humidity are key factors in tropical regions, while adding further evidence that cold and dryness are important in temperate areas. They fit with previous research showing that flu viruses survive longer in dry air.

The researchers used a fairly new global database that contains information on flu peaks at the 78 sites around the world, the report says. The locations range from 1 to 60 degrees latitude, and 39% are in the tropics. The flu data were gleaned from 85 studies conducted from 1975 to 2008, with a median duration of 2 years. In most cases they focused on a city rather than a province or region.

To provide an independent check on the findings, the team also used epidemiologic data from nine countries participating in the World Health Organization´s global flu surveillance program, FluNet. The countries—Spain, Tunisia, Senegal, Philippines, Vietnam, Colombia, Paraguay, South Africa, and Argentina—each provided several years´ worth of data. .

The scientists gathered data on temperature, relative humidity, and precipitation for each of the sites. They used the temperature, relative humidity, and atmospheric pressure to calculate the average monthly specific (absolute) humidity, meaning the ratio of water vapor to air.

By comparing the flu patterns and climate data, the researchers found that temperature and specific humidity were the best individual predictors of peak flu activity, the NIH release said. They determined that in temperate regions, flu was more common 1 month after periods of the lowest specific humidity, which coincided with the coldest months.

"In contrast, sites that maintained high levels of specific humidity and temperature were generally characterized by influenza epidemics during the most humid and rainy months of the year," the NIH said.

"The models we used predicted the timing of peak influenza activity with 75 to 87 percent accuracy," senior author Cecile Viboud, PhD, of the NIH´s Fogarty International Center, said in the release.

The team identified a specific climatic threshold that separated sites with "cold-dry" and "humid-rainy" flu peaks, according to their report. In places where monthly specific humidity dropped below about 11 to 12 grams of water per kilogram of air and temperatures sank below 18o to 21oC for at least a month, flu activity peaked in winter.

"For sites where specific humidity and temperature do not decrease below these thresholds, seasonal influenza activity is more likely to peak in months when average precipitation totals are maximal and greater than 150 mm [about 6 inches] per month," the report says.

Although specific humidity was found to be a good predictor of flu peaks overall, this was mainly due to its performance in higher latitudes, the authors write, adding, "In low latitudes, precipitation was a stronger predictor of the timing of influenza activity." They also note that peak flu activity was harder to predict in low-latitude places, some of which do not have well-defined flu seasons.

The report also says that the predictions of the climate-based models compared favorably with the epidemiologic information collected from the nine countries involved in FluNet.

Viboud said the findings could be used to refine flu transmission models, target surveillance efforts, and improve the timing of seasonal flu vaccine delivery, according to the release.

But the statement cautioned that climate is just one of several factors that influence flu seasonality. "Further work should focus on examining the role of population travel and other factors in influenza transmission," said Mark Miller, MD, director of the Fogarty Center´s Division of International Epidemiology and Population Studies, who wasn´t involved in the study.