Skip Over Navigation Links

Transcript of Dr. Jeffrey Shaman on Seasonal Flu and Humidity Podcast

February 23, 2010

Jeffrey Shaman talks about the role of absolute humidity in seasonal influenza outbreaks, as described in a paper published online in PLoS Biology on February 23, 2010.

The seasonality of influenza has been something that people have been puzzled over for a very long time. They've noticed that people get sick in temperate regions in the wintertime. Specifically for influenza, there's sort of three groupings of ideas as to what's responsible.

One has to do with behavior, basically. So people spend more time indoors in winter because it's colder. They're in more contact-there's more person-to-person contact, which facilitates the transmission of the virus.

Another camp has the idea that it's some host-mediated factor. So that something in the winter makes it so that your immune system is less-functioning, and as a result of that it's easier for the virus to set up camp in your system and it facilitates the transmission from person to person.

The third idea had to do with environmental factors controlling the survival and transmission of the virus. And previous studies of it (and this is going back to the '40s) had looked really almost exclusively at relative humidity and temperature.

Relative humidity is something that we hear about all the time on the local weather report. It doesn't tell you the actual amount of water vapor that's present in a volume of air. It tells you how close that volume of air is to saturation-the point at which a cloud would start to form. The amount of water vapor that you need in a volume of air to reach that point where you can start forming a cloud changes as the temperature changes.

So what I did looking at all these studies was I said, well, why don't I recalculate this. I can actually take the temperature and relative humidity and back out a measure of the amount of water vapor that's in the air, and that's called the absolute humidity. And when doing that, I found that the data people had previously looked at over the last 50-60 years actually was much better explained by this single variable, absolute humidity, than by either temperature or relative humidity. Absolute humidity has a very strong seasonal cycle. So in other words, in the winter it's very low, which is conducive for longer survival and more transmission of the virus, and in the summer it's very high, which is conducive for less survival of the virus and less transmission.

The initial study had looked at laboratory results. We decided to scale it up to the country scale and do it as an epidemiological study. So in an epidemiological study we're using human case data. And this case data was 31 years of the rates that people were dying as it was attributed to influenza and influenza-like illness. And we did it for the whole contiguous U.S.

And what we did is we applied a mathematical model—a dynamic model—that just describes the influenza transmission cycle. And we constructed it in a way so that absolute humidity would modulate the rates of transmission in a way that was similar to the laboratory analysis of the survival and the transmission of the virus. And then we used actual observed daily absolute humidity conditions for each of the states to drive the model forward. And what we found is that we could reproduce the seasonal cycle of influenza doing this very nicely in basically the whole country.

And we didn't have to use a different form of the model in New York State versus Florida versus Arizona. So because of that, it offered some very nice supporting evidence that in fact humidity could be this big seasonal forcing.

So the other thing we actually saw was that we could look at when flu outbreaks took place. We could go through each of those winters and say, well, when did the outbreak of influenza begin. And then we could look at that, and we already know that absolute humidity is already seasonally low at that time of year because it's the winter. What we found is that on average we found a significant drop in humidity in the weeks prior to when the outbreak took off.

What it seems to indicate is that-and this is related to weather. You have an average pattern, which is your climatology, and then you have weather for an individual year that comes through. If you had a real snap of, let's say, cold dry air, which is typically associated after a cold front comes through—you get clear skies, really cold temperatures, and really dry air—when you had one of these things come through, what was typically happening is it was increasing the likelihood that you would have an outbreak taking place. It would be something that actually could be used in the future to potentially make some sort of forecast.

Well the basic science implications is it explains something about how the environment is linked to a virus. This could certainly have implications for how people look at other pathogens out there.

Obviously the practical public health implications are is that it provides further evidence that absolute humidity is a factor that's related to the survival of the virus. It has implications for what potentially hospitals might want to do in the face of an outbreak of influenza. They may want to humidify the room. They may want to further control their wards so that they can control the humidity levels basically, the absolute humidity levels, and basically maybe reduce the suitability of the environment to the survival of the virus and its transmission. Now obviously there are tradeoffs because influenza is not the only pathogen out there. There are things like pathogenic molds that actually like it when it's very humid. But if we're in the face of an actual outbreak of influenza and that's the immediate concern, it may be worthwhile to consider raising humidity levels, for instance.

This podcast was produced by the National Institute of General Medical Sciences, part of the National Institutes of Health. NIGMS supported the research described here through its Models of Infectious Disease Agent Study.

This page last reviewed on October 14, 2011