Web Exclusives: Diseases
Q&A: Jeffrey Shaman on Seasonal Flu and Humidity
For decades, scientists have wondered what drives seasonal influenza outbreaks in the wintertime, with theories ranging from staying cooped up to having weaker immune systems. Now, a study led by Oregon State University climatologist Jeffrey Shaman and Harvard School of Public Health epidemiologist Marc Lipsitch suggests that the main culprit is absolute humidity—the amount of water vapor in the air.
By analyzing 30 years of U.S. flu and weather data and then building a computer model that reproduces historical patterns of seasonal flu spread, Shaman's team found that seasonal changes in absolute humidity predicted when and how easily people caught the flu and infected others. The results could improve flu research, help manage contagion and even forecast seasonal outbreaks.
How does absolute humidity affect flu spread?
We found that low absolute humidity is conducive for longer survival of the virus and more transmission, and high absolute humidity is conducive for less survival of the virus and less transmission. In our computer model, we used humidity to control influenza's basic reproductive number. That's the average number of other people a single sick person infects with the flu, which depends on factors such as how much contact people have with each other, how infectious a person is, and how long a person remains infectious. When the humidity is high, a person may be expelling virus into the environment, but if it's not surviving as long, that's going to diminish their ability to create secondary infections, which decreases transmission rates.
That doesn't mean people can avoid the flu by using humidifiers, right?
Right. The best defense against influenza remains vaccination. There's no reason anybody should give up vaccination and instead rely on humidification.
If low humidity helps the flu virus survive longer, why did we see H1N1 outbreaks in the summer?
We modeled epidemic influenza—that's the seasonal flu we typically experience each winter. This is in contrast to pandemic influenza like the recent H1N1 "swine flu." The dynamics of pandemic and epidemic influenza differ because the susceptibility of the general population to these pathogens differs. Only 30-50 percent of the population is susceptible to epidemic flu in a given winter season, whereas almost the entire population is susceptible to pandemic flu. Pandemics develop following a major shift in the structure of a flu virus. We saw outbreaks of H1N1 in the summer partly because more of the population was susceptible to pandemic flu.
Most likely the survival of pandemic influenza is also affected by humidity. However, due to the difference in population susceptibility, the transmission dynamics of pandemic flu differ from those of epidemic flu.
How do you know that something like the school calendar isn't primarily driving seasonal flu outbreaks?
We did look at the school calendar within a select number of states. When we did that and tried to model flu seasonality, we weren't able to get as good an explanation, or as good a "model fit." Also, when you go into the tropics, there really isn't a strong seasonal [flu] cycle, yet they do have a school calendar there. What they lack there is a strong seasonal cycle of absolute humidity, it turns out.
I do think the school calendar has an effect. I do think it's very important when you mix populations and you bring people together. But I just don't think it's the dominant force at work here.
What do you want to study next?
We want to see how well we can forecast influenza risk—how well we can say whether an environment is suitable for the transmission of influenza.
We also need to look at the geographic spread of influenza. Each year, influenza moves around the globe. It's not an exact pattern, but there are consistent ways it moves from the tropics to temperate regions. If we can fold humidity into an understanding of that, it might give us a better sense of how and why influenza moves around as it does.
It also would be good if we could apply our model to the deep tropics and see what, if anything, it says about transmission there, where there's not a strong cycle of absolute humidity.
How does computer modeling help you in your research?
My background is in atmospheric science. I study climate and hydrology, and I look at large-scale motions in the atmosphere, the waves that move through the atmosphere that communicate information across the globe. The work that I do there involves a lot of modeling. I constantly am using models to explore physical phenomena.
For influenza, models are really integral to figuring out what's going on. They give you the ability to actually translate a hypothesis about how the physical affects the biological into a system that you can control and explore.
And not only that, it then ultimately forms the basis of what can be a predictive model.
What's it been like to work on this project?
It's very exciting. Having something that not only has importance in terms of basic science and inquiry but also has such obvious public health implications is a very strong motivation for me.