CORVALLIS, Ore. - The seasonal increase of influenza has long baffled scientists but a new study has found that seasonal changes of absolute humidity are the apparent underlying cause of these wintertime peaks.

The study also found that the onset of outbreaks might be encouraged by anomalously dry weather conditions - at least in temperate regions.

The study, published this week in the science journal PLoS Biology, builds on laboratory research that found influenza virus survival rates increased greatly as absolute humidity decreased. In this new study, the researchers used 31 years of observed absolute humidity conditions to drive a mathematical model of influenza and found that the model simulations reproduced the observed seasonal cycle of influenza throughout the United States.

They also discovered that the start of many influenza outbreaks during the winter was directly preceded by a period of weather that was drier than usual.

"This dry period is not a requirement for triggering an influenza outbreak, but it was present in 55 to 60 percent of the outbreaks we analyzed so it appears to increase the likelihood of an outbreak," said Jeffrey Shaman, an Oregon State University atmospheric scientist and lead author on both studies. "The virus response is almost immediate; transmission and survival rates increase and about 10 days later, the observed influenza mortality rates follow."

Scientists have long suspected a link between humidity and seasonal (epidemic) flu outbreaks, but most of the research has focused on relative humidity - the ratio of water vapor content in the air to the saturating level, which varies with temperature.

Absolute humidity quantifies the actual amount of water in the air, irrespective of temperature. Though somewhat counter-intuitive, absolute humidity is much higher in the summer, said Shaman, an assistant professor in the College of Oceanic and Atmospheric Sciences at Oregon State University.

"In some areas of the country, a typical summer day can have four times as much water vapor as a typical winter day - a difference that exists both indoors and outdoors," Shaman said. "Consequently, outbreaks of influenza typically occur in winter when low absolute humidity conditions strongly favor influenza survival and transmission."

Shaman, who specializes in the relationship between climate and infectious disease transmission, said anomalously low absolute humidity in much of the continental United States is often associated with excursions of colder air masses from the north. These air masses, which can follow a cold front, bring cloud-free skies, reduced surface temperature and lower humidity levels.

Though the findings by Shaman and his colleagues build a strong case for absolute humidity's role in influenza outbreaks, they do not constitute a predictive model, he emphasized.

"Certainly absolute humidity may affect the survival of the influenza virus, but the severity of outbreaks also is dependent upon other variables, including the type of virus and its virulence, as well as host-mediated factors such as the susceptibility of a population and rates of population mixing and person-to-person interactions," Shaman said.

"But what we can do is identify when environmental conditions are rife for transmission, which is the first step toward a predictive model," he added.

In the modeling portion of their study, the researchers examined influenza in New York, Washington, Illinois, Arizona and Florida, and found that the absolute humidity conditions in those state all produced model-simulated seasonal outbreaks of influenza that correlated well with the observed seasonal cycle of influenza within each state.

Shaman and colleagues then extended their model to the rest of the continental U.S. and were able to reproduce the seasonal cycle of influenza elsewhere. Other potential "drivers" of outbreaks - such as temperature, relative humidity, solar radiation and even school days - were examined, but no variable worked as well as absolute humidity.

The association between the onset of wintertime influenza outbreaks and unusually dry weather conditions was based on state-by-state analysis of nearly 1,500 winters throughout the United States.

"In a few winters, there was little influenza transmission and thus no definable onset," Shaman pointed out. "In the other winters, outbreaks may have begun as early as November or as late as March. In more than half of these cases, the onset was preceded by a period of drier-than-normal weather."

"Conditions are already seasonally dry during winter and favorable for influenza virus survival and transmission," Shaman said. "Drier-than-normal weather conditions appear to additionally amplify the survival rate of the virus."

Marc Lipsitch, a professor of epidemiology at the Harvard School of Public Health and senior author on the new study, said the new analysis may have implications for other diseases.

"Seasonality of infectious diseases is one of the oldest observations in human health, but the mechanisms - especially for respiratory diseases like flu - have been unclear," Lipsitch said. "This study, in combination with Shaman and (Melvin) Kohn's earlier analysis of laboratory experiments on flu transmission, points to variation in humidity as a major cause of seasonal cycles in flu."

"Seasonal variation in flu, in turn, helps to explain variation in other infectious diseases - such as pneumococcal and meningococcal disease - as well as seasonal variation in heart attacks, strokes and other important health outcomes." 

Lipsitch directs the Center for Communicable Disease Dynamics, of which Shaman is a member. This study and the center are supported by the Models of Infectious Disease Agent Study, or "MIDAS Program," of the U.S. National Institute of General Medical Sciences - one of the National Institutes of Health.

"The discovery of a link between influenza outbreaks and absolute humidity could have a major impact on the development of strategies for limiting the spread of infection," said Irene Eckstrand, who oversees the MIDAS program. "Understanding why outbreaks arise is an important first step toward containing or even preventing them, so it is essential for scientists to follow up on this intriguing connection."

Additional collaborators on the study published in PLoS Biology were Virginia Pitzer and Bryan Grenfell, Princeton University; and Cecile Viboud, National Institutes of Health Fogarty International Center.


Jeff Shaman, 541-737-4915

Click photos to see a full-size version. Right click and save image to download.