Post #894: Why is flu seasonal? Humidity and COVID-19 spread, final version

Posted on November 18, 2020

Source:  Graph from blog.sas.com, original data from CDC Pneumonia and Influenza Mortality Surveillance.

Source:  Potential impact of seasonal forcing on a SARS-CoV-2 pandemic DOI: https://doi.org/10.4414/smw.2020.20224 Publication Date: 16.03.2020 Swiss Med Wkly. 2020;150:w20224 Neher Richard A., Dyrdak Robert, Druelle Valentin, Hodcroft Emma B. Albert J.

This is a followup to Post #879 and Post #880.  It is, in fact, little more than an expansion of the literature review for Post #879.


Executive summary

This post is a review of the evidence on why flu is seasonal.  To cut to the chase, many lines of evidence suggest that flu season in temperate climates is primarily driven by dry indoor air.

Why does that matter?  The same factors that make flu seasonal (top graph) apply equally to coronaviruses (bottom graph above).  As discussed below, dry indoor air results in impairment of the lung’s defenses against pathogens, and improvement of virus survival times.  And that unfortunate combination applies just as much to COVID-19 as  it does to the flu.

In short, it’s a good bet that the upsurge we’re currently seeing in coronavirus cases is being driven by the same factors that create “flu season”.  Foremost among those is low indoor humidity.  And that has a lot of implications, both for the extreme outbreak going on now in the upper Midwest, and for the country more generally.

There’s one huge caveat in this:  What we actually observe is the impact of humidity as modified by human behavior.  This year, we’ve had two stark examples of good sanitation overriding environmental factors such as humidity.  First, this year, for the first time ever, there was effectively no flu season in the Southern Hemisphere.  That is widely attributed to COVID-19 sanitation measures such as masks and distancing.  Second, within the northern hemisphere, as winter sets in, there is an unmistakable contrast now between the mask-wearing Asian nations and the West.  Cases are surging in the North America, Europe, and Russia.  But not in China, Japan, and South Korea. 

So I’m not trying to say that our weather is our fate.  Far from it.  But I am trying to say that our weather, combined with our current behavior, is our fate.  And that the combination of cold, dry winter air and an aversion to wearing masks is what’s creating the astounding surge in cases in the U.S. Midwest.

In pictures, here’s the whole of this argument.

NY Times map of mask use.  White = lowest use rate.

 

 

 

NOAA map of October 2020 temperatures

 

 

 

NY Times map of current incidence of COVID-19.

 

 

 

Overlay the first, on the second, and you get some crude version of the third.  And the argument here is, that’s not coincidence.

Links to specific topics are given below.  Use your browser back button to return.

The rest of the executive summary follows.   This is the argument I am trying to make, in words this time, not pictures.

The relationship between dry indoor air and spread of respiratory illness is hardly news.  If I had to boil it down to one diagram, from one reputable source, to try to bring this message home, it would be this:

Source: Humidification of indoor air for preventing or reducing dryness symptoms or upper respiratory infections in educational settings and at the workplace.  Cochrane Systematic Review – Intervention – Protocol Version published: 01 June 2016 https://doi.org/10.1002/14651858.CD012219

That’s from a “Cochrane review”, generally accepted as the gold-standard method for reviewing the medical literature. Note the “optimum zone” of 40% to 60% relative humidity.  That range comes up over and over again in the research literature.  Note the line labeled viruses.  Note the line labeled respiratory infections.  Note that they are depicted as far more of a threat in dry air than in air humidified to that optimum zone.  This chart doesn’t quantify those effects, but it shows you the right picture.

I believe that a review of the medical and epidemiological literature supports the following observations:

  • Many independent lines of evidence show that flu spreads more readily with low indoor humidity.  That’s based on direct experiments with mammals, a pre-post-with-control experiment using children in school classrooms, and copious epidemiological evidence, including a very good correlation between short-term changes in humidity and subsequent changes in flu incidence and deaths.
  • There are reasons for this directly related to human physiology.   Dry air inhibits the lung’s main defense against pathogens, “mucociliary transport”, in which tiny hairs in the respiratory tract sweep mucous up and out of the lungs.  Dry air also induces other unfavorable changes in the immune system within the respiratory tract.  The end result is that dry air makes humans much more susceptible to respiratory illness.
  • There are further reasons related to the behavior of virus and droplets in dry air.  Flu virus in droplets remains viable much, much longer in dry air than in humid (40% relative humidity) air.  Similarly, COVID-19 droplets remain viable on surfaces longer at low relative humidity.
  • There is good epidemiological evidence that dry air is, in fact, the primary determinant of onset and duration of flu season.  That is, good correlation between dry air and flu in temperate climates, accounting for as many other factors as possible.   By some measures, half the variation in flu deaths can be explained by humidity alone.

This has policy implications that don’t (yet) seem to be on anybody’s radar screen.  So just humor me for a bit, and assume I’m right about this.  What are the implications, if you accept that, in the type of climate we have in the U.S.A., respiratory viruses including COVID-19 spread much more effectively in dry air?

First, there’s the obvious implication of this, for public health policy:  Humdify indoor spaces.  In addition to doing all the other things you’re supposed to be doing to prevent COVID-19 spread.  Make sure to keep the relative humidity in your home and work spaces at 40% or so, because that seems to be the sweet spot for suppressing spread of viral respiratory diseases.

Second, behavior that was safe in summer may no longer be safe.  In particular, we’re going to have to find a way to help people in the U.S. Midwest un-learn behavior that had been successful prior to the change of seasons.  And that’s going to be hard.  If you’ve been going around maskless for nine months, hearing about problems elsewhere, and not experiencing them yourself, it’s going to be a hard to believe that you need to wear one now.

Third, focusing on seasonality could provide the best way to explain this to mask-averse Midwest populations.  Getting people to change would be a whole lot easier if you could tell them exactly what is different now, relative to two months ago.

And it would make flip-flopping mask-averse Midwest governors appear a little bit more rational.  They could say “winter is here, you know that winter is serious in the Midwest, and here’s why that changes things.”  Surely, people who live in the upper Midwest understand that you have to take winter seriously.  And in some sense, these new mask mandates in Midwest states could be couched as just an extension of that idea.

Fourth, if low indoor humidity leading to enhanced virus transmission is the root cause of the “third wave” in the U.S., that needs to trickle up into the new mask mandates themselves.  Both North Dakota and Iowa passed mask mandates that last just a few weeks.  As if this will all have blown over by then.  But if the real culprit here isn’t some sort of mysterious one-time outbreak, but instead is the weather — and low indoor humidity — then a three-week mask mandate is just the worst sort of wishful thinking.  Those mandates will be expiring just as we are going into the heart of the period when they are most needed.

Finally, I keep harping on this because I’m afraid that it’s soon going to be Virginia’s turn.  We have good levels of mask use here in NoVA.  And that’s going to help.  But we are just now entered the period for low indoor humidity.  Arguably, we had our first frost in Vienna, VA last night.  Ss the temperatures have dropped, I’ve seen the humidity level in my house drop from mid-60-percents last week, to mid-40-percents now.

I’ve prepped.  I have two large evaporative humidifiers that I put into service this last night.  Even if it turns out that I’m wrong about humidity and viral spread (but I’m not — see literature review below), I’d feel remiss taking this precaution myself and not making the rationale for it public.

And so, that’s why, in this post, I’m going to present my review of the scholarly literature, and argue that the main cause of a “flu season” is low indoor humidity.

Some final sections repeat the policy context, and give some obvious caveats.

 


A quick review of the literature on indoor air humidity and flu in temperate climates.

The “story” linking all this together is now the executive summary.  This is just individual pieces of evidence, separated by horizontal lines.


Seasonal variation in vitamin D:  A plausible minor contributor to flu season, but cannot explain the timing of flu season.

You will read that low wintertime vitamin D levels might be a root cause of the wintertime flu season.  This is plausible because vitamin D plays some key roles in the immune system, and population vitamin D levels drop in wintertime.

But if you read through a recent literature review on this topic, you’ll see that the experimental (randomized trial) data on vitamin D and flu is mixed.  In some cases, provision of vitamin D supplements appeared to reduce respiratory illness greatly.  In other cases not.  Most studies were very small-scale.

There seems to be general agreement that providing vitamin D to truly vitamin-deficient individuals does indeed help them ward off respiratory illnesses (reference).  And, conversely, that individuals who show up with respiratory illnesses are more likely to be vitamin D deficient.

But given the mixed results, it’s no particularly surprise that a recent meta-analysis (summary of other research) found a noticeable-but-modest impact.  On average, individuals given weekly or daily vitamin D supplements had about a 12 percent lower chance of getting a respiratory infection.

The implication there is that natural declines in vitamin D might lead to perhaps 12 percent more flu cases in a season.  That is not a big enough effect to cause the sharp spike in cases that occurs in flu season.  Most tellingly, it’s difficult-to-impossible to reproduce the actual timing of flu epidemics based on vitamin D levels.  Vitamin D slowly declines over months.  Flu tends to spike (as in the graph at the top of this posting). There’s no simple arithmetic that will get you from the shallow smooth curve below to the series of spikes at the top of this posting.

Source:  vitamindwiki.com

And so, while a good supply of vitamin D might help ward off the flu, population vitamin D levels appear unlikely to be the main driver of flu season.


More time indoors in winter: Another apparently minor factor.

You’ll also see the hypothesis that we have flu in winter because we spend more time indoors, around other people.   That one is more-or-less impossible to test empirically, and in any case the magnitude is wrong.  US adults spend almost all their time indoors regardless of the season, about 22 hours a day (summertime) to 23 hours a day (wintertime), based on this source.

 

Source:  “It’s about time: A comparison of Canadian and American
time–activity patterns”, JUDITH A. LEECH,WILLIAM C. NELSON, RICHARD T. BURNETT, SHAWN AARON, AND MARK E. RAIZENNE, Journal of Exposure Analysis and Environmental Epidemiology (2002) 12, 427 – 432

The modest increase in indoor time during winter lacks face validity as being the prime driver of flu season.  The difference in indoor time, on average, appears far to small to account for the vast increase in flu cases.


Cochrane review of humidifying indoor air as a strategy for reducing respiratory infections; other reviews of building environments and COVID-19.

Those of you “in the know” in terms of health care research know that that the best place to start is a Cochrane review.  These review are large-scale analyses of the scientific literature, adhering to a strict set of standard regarding methods and allow-ability of evidence.  They aren’t foolproof, but in my line of work, they’re about the best thing we’ve got in terms of separating fact from fiction.

Luckily, there is a 2016 review that is more-or-lesson target.  So let’s start with the summary findings from that, which focused, in part, on humidify indoor air as a strategy for reducing respiratory infections.

Source: Humidification of indoor air for preventing or reducing dryness symptoms or upper respiratory infections in educational settings and at the workplace.  Cochrane Systematic Review – Intervention – Protocol Version published: 01 June 2016 https://doi.org/10.1002/14651858.CD012219

Note the “optimum zone” of 40% to 60% relative humidity.  Note the line labeled viruses.  Note the line labeled respiratory infections.

Note the generally U-shaped relationship between humidity and suppression of viruses and respiratory infections.  It’s not that humidity is either good or bad.  It’s that too little or too much indoor humidity is bad.  And that what you should shoot for is that optimum zone of 40% to 60% indoor relative humidity.

Obviously, having condensed a large amount of detail into that handful of lines, there are going to be a lot of details that have been omitted.  But at first glance, based on the professional opinions of the experts who write that Cochrane review, there’s a lot to be said for 40% indoor relative humidity.

—————-

Separately, when building engineers were asked to evaluate ways to minimize spread of COVID-19 in the “built environment”, they too came up with a 40% to 60% relative humidity range as being optimal for limiting spread of the virus.  This was in addition to using better air filters, introducing more outdoor air, and so on.

They went so far as to suggest that current industry standard for allowable indoor humidity levels be revised.  The current AHSRAE standard is 20% to 60%, they suggested that 40% to 60% would be more appropriate in light of the pandemic.

This paragraph summary of the humidity issues pretty much hits all the high points, and is worth reproducing in its entirety.  Emphasis mine:

Increasing evidence indicates that humidity can play a role in the survival of membrane-bound viruses, such as SARS-CoV-2 (6365). Previous research has found that, at typical indoor temperatures, relative humidity (RH) above 40% is detrimental to the survival of many viruses, including CoVs in general (63, 66, 67), and higher indoor RH has been shown to reduce infectious influenza virus in simulated coughs (67). Based upon studies of other viruses, including CoVs, higher RH also decreases airborne dispersal by maintaining larger droplets that contain viral particles, thus causing them to deposit onto room surfaces more quickly (63, 68, 69). Higher humidity likely negatively impacts lipid-enveloped viruses, like CoVs, through interactions with the polar membrane heads that lead to conformational changes of the membrane, causing disruption and inactivation of the virus (70, 71). Furthermore, changes in humidity can impact how susceptible an individual is to infection by viral particles (72) and how far into the respiratory tract viral particles are likely to deposit (68). Decreased RH has been demonstrated to decrease mucociliary clearance of invading pathogens and weakened innate immune response (7274). However, RH above 80% may begin to promote mold growth, inducing potentially detrimental health effects (75). Although the current ventilation standard adopted by health care and residential care facilities, ASHRAE 170-2017, permits a wider range of RH from 20% to 60%, maintaining a RH between 40% and 60% indoors may help to limit the spread and survival of SARS-CoV-2 within the BE, while minimizing the risk of mold growth and maintaining hydrated and intact mucosal barriers of human occupants (50, 67). Indoor humidification is not common in most HVAC system designs, largely due to equipment cost and maintenance concerns related to the risk of overhumidification increasing the potential of mold growth. While administrators and building operators should consider the costs, merits, and risks of implementing central humidification, especially during new construction or as a retrofit, it may be too time intensive to implement in response to a specific viral outbreak or episode. In addition, increased RH may lead to increased buildup on filters, decreasing airflow. However, in pandemic situations, this practice likely increases the effectiveness of capturing viral particles, and this benefit outweighs the increased filter maintenance required. Therefore, targeted in-room humidification is another option to consider, and this may reduce the likelihood of a maintenance oversight causing overhumidification.

Among the things I learned, from that single paragraph, is that most buildings don’t have centralized humidifiers built into their HVAC systems.  That’s the sentence in orange above.  And so, in many places, indoor humidity is going to be the passive byproduct of the absolute humidity of the outdoor air.


Impact of dry air on the body:  Mucociliary clearance and other immune system effects.

This is one of those COVID-related terms that I wish I’d never had to learn.  Your entire upper respiratory tract is lined with mucous membranes, and in addition much of the surface is lined with little hairs (cilia).  Mucous itself has substances that fight pathogens, and the cilia sweep the mucous toward the top of your throat, where you (ahem) eliminate that mucous in some fashion.

This is the primary mechanism by which your lungs protect and clean themselves.  Of anything that lands on the surface of the lungs.  Mucous traps things before they can actually get to your lung cells.  And then your lungs continuously sweep the mucous lining up toward your throat, where it gets disposed of.

Dry air inhibits mucociliary clearance, and humid air increases it.  (Also referenced halfway through this review article.)  And it’s not exactly rocket science to understand it:  Dry air dries out your mucous.  That slows down the rate of transport.  And so your entire upper respiratory tract functions less well at cleaning itself, and protecting itself from pathogens.

And sure enough, even something as simple as drinking lots of fluids helps to reduce that.  I had always assumed that the old “drink plenty of fluids” advice was somehow related to your kidneys or something.  But, nope, in fact, in cases where you are not losing fluid via (e.g.) vomiting or diarrhea, that advice is intended to address your mucociliary clearance.  The whole point of “drink plenty of fluids” is to keep your mucous loose, per WebMD.

The direct effect of “drinking lots of fluid”, on nasal mucocilliary clearance in dry air has been demonstrated empirically (per this scholarly reference).  But the direct effect of drinking fluids, on colds/flu, had not ever been tested empirically.  So we know that being well hydrated helps fight the effects of dry air on slowing mucociliary clearance, but there are no controlled clinical trials of fluid consumption and colds/flu.

In mice, dry air has been directly demonstrated to have several additional immunity-compromising effects in addition to reduced mucociliary clearance.  “Mice housed in dry air had impaired mucociliary clearance, innate antiviral defense, and tissue repair function.”  You can see a popular-press writeup here that makes the key point:  This research largely explains why people are more frequently sickened by flu in winter months in temperate climates.  The plausible mechanism is the direct effect of low relative humidity on the respiratory tract.


Studies of flu transmission in children and guinea pigs.

A study of guinea pigs from the mid-2000s demonstrated the direct link between dry air and higher rates of respiratory virus transmission.  As I understand this one, at one point, they literally put infected and uninfected animals in side-by-side cages, modified the humidity levels, and counted the number of animals that got sick.  You can see the scholarly citation here, or a popular-press writeup here.  

One little quote to get the gist of the results, emphasis mine

Five different humidity levels were tested at 20°C: 20%, 35%, 50%, 65%, and 80%. The researchers found that transmission was most efficient at the two lowest humidity levels, at which three or four of the exposed guinea pigs became infected. At 50% relative humidity, only one of the four animals contracted influenza.

This is not the only experimental evidence.  Back in the 1970s, they literally performed a pre-post-with-control experiment on school children, testing the impact of humidifying classrooms on the transmission of winter colds and flu.

Fully realizing that I’m citing a humidifier manufacturer, this is nonetheless the clearest writeup of that research, and well worth reading.  They found that humidifying classrooms of schoolchildren  (from about 15% up to about 50% relative humidity) cut the rate of respiratory illness in half.  And when the removed the humidifiers?  The illness rate went right back to the average.

I note that they didn’t arrive at this experiment out-of-the-blue.  In fact, they had already measured a strong statistical association between low relative humidity and rates of student illness in Canadian schools.   The experimental analysis just took that one step further, and demonstrated that in in this case, the correlation was in fact signaling causation.


Impact of dry air on virus survival times.

Here, I’m just going to cite a few examples that are directly relevant.  The bottom line is that both common flu and COVID-19, in droplets and on surfaces, survive much better at low relative humidity.

Here’s a writeup of a 2013 study that used realistic manikins to “cough” some flu virus, and then to breathe in the results of the cough.  The idea was to simulate sitting in a waiting room with sick people.  Key quote:

At humidity levels of 23 percent, 70 to 77 percent of the flu virus particles were still able to cause an infection an hour after the coughing simulation. But when humidity levels were raised to 43 percent, just 14 percent of the virus particles had the ability to infect.

That study is apparently the source of this little bit of advice on the humidifier I just bought.  It cites that 43% relative humidity number as if it were somehow unique.  When, in fact, it’s just one of the humidity levels that they happened to have studied in the research above.

The finding the viruses simply die faster higher relative humidity, for the range of humidities you would find in a typical living space, has been found for many (but not all) viruses.  In particular, for viruses with a lipid envelope (like coronaviruses), survival times in the air are higher at low relative humidity.  This has been tested and found true for a wide range of such viruses.  (See Table 1 of this study.)

 

As far as I know, the only direct measurement of the link between humidity and COVID-19 viral lifetime is for viral droplets deposited on surfaces.  As in this table developed by the US Department of Homeland Security:

 

 

Some of the scholarly research underlying that table can be found in this analysis.  They used viruses similar to COVID-19 and looked at survival times on surfaces at a variety of temperatures and relative humidities.  The virus survived longer at low temperatures, and longer at either very low or very high relative humidities, similar to the U-shaped blocks in the Cochrane analysis above.

It’s probably worth reproducing their key figure.  This is virus survival on room-temperature surfaces, at three different humidities.  The top and middle charts contrast 20% relative humidity and 40% relative humidity.

I’m including this to get across how large the effect of relative humidity is in destroying virus particles.   The vertical scale is a base-10 log scale showing the amount of virus that was intact.  That is, each step on the vertical scale is a factor of ten.  After seven days, comparing the first two charts, the amount of virus remaining at 40% relative humidity was about 1% of the virus remaining at 20% relative humidity.

 


No relevant impact on droplet size and travel distances.

You will see citations suggesting that dry air leads to smaller droplets and aerosols.  The implication being that these smaller particles would hang in the air longer, and travel further.

After a more detailed look at the literature, I don’t think this matters in the ranges of relative humidities found in the home.  At very high humidity, moisture in the air prevents droplets from shrinking (evaporating), and thus forces them to fall out of the air faster.  But that effect really only matters at very high relative humidity.

Here, I’m relying on one figure from this excellent scholarly summary of the literature on flu, humidity, and temperature.

Source:  Marr LC, Tang JW, VanMullekom J, Lakdawala SS. 2019 Mechanistic insights into the effect of humidity on airborne influenza virus survival, transmission and incidence. J. R. Soc. Interface 16: 20180298. http://dx.doi.org/10.1098/rsif.2018.0298

 

 

This diagram shows the final droplet size, relative to the initial droplet size, on the vertical scale, and relative humidity in the horizontal scale.  At anything below roughly 60% relative humidity, the droplets dry out completely, and there’s no difference in size.  At relative humidities commonly found in the home, there would be no difference in droplet sizes due to changes in relative humidity.

 


The statistical association between dry air and flu/COVID

Let me start this with a study of flu mortality, using monthly flu and weather data from more than 300 US urbanized counties, for about three decades.  They found that humidity was the key predictor of flu mortality.  more than 300 US urban areas, and more than a decade of time.

Key quote, emphasis mine:

Model predictions suggested that approximately half of the average seasonal differences in US influenza mortality can be explained by seasonal differences in absolute humidity alone. 

While they studied weather data and looked at absolute humidity, that would translate directly to differences in indoor relative humidity, assuming a constant temperature for indoor air. (That point about absolute and relative humidity can be found in this reference.)

Most interestingly, they identified the absolute humidity cutoff below which mortality rates rose as 6 grams of water per kilogram of air.  If that air were heated to room temperature (68F), that would result in a relative humidity of 40%.  In effect, their statistical analysis of flu death rates identified the same 40% indoor relative humidity threshold as the Cochrane review above.


Nor is the US the only place where that association can be found.  A similar statistical analysis of flu and pneumonia deaths in a semi-tropical climate yielded similar results:  Roughly three weeks after any cold, dry days, deaths from pneumonia and flu increased.  So there’s nothing unique to the US climate in this finding.

The results were reliable enough in that location (Auckland, NZ) that they called for a sort of flu forecasting.

“Given the ability to forecast cold and dry periods in advance, such information could potentially be used for ‘real time’ respiratory infection forecasting,” the researchers write. “This could help reduce the impact of pneumonia and influenza in a number of ways: Communication to vulnerable groups (particularly the elderly and those with established respiratory disease), deployment of preventive measures (eg, vaccine and antivirals), and planning for increased service demands in hospitals.

———————-

The evidence of the impact of humidity on airborne spread of COVID-19 is based mainly on epidemiological data.  That is, unlike flu, there have been no actual controlled trials in animals (I think).  Instead, researchers observed where the virus did and did not spread well, accounted for other factors (e.g., mask-wearing) as best they could, and arrived at some summary judgment.  Their conclusion:

Furthermore, the spread of types of diseases caused by betacoronavirus, such as SARS-CoV-1 [11] and MERS-CoV [36], ... hot and humid climates demonstrated the ability to decrease the viability of these viruses, while in places with low temperature and humidity there was greater viral stability.

I think this statistical analysis provides an excellent complement to the laboratory-based survival time studies above.  In the lab, it sure looks like the virus should survive better in cold and dry climates.  And in the real world, as best you can tell from observational data, that is the case.


Policy analysis 1:  Why should we care why flu is seasonal?

First, the same factors that cause flu to be seasonal appear to be at work with COVID-19.  In other words, it looks like we’re entering our expected winter outbreak of COVID-19.  Nobody could be sure that such an outbreak would occur, but it would have been a pretty good guess.  Dr. Fauci has been warning of this for months now.

Second, popular press coverage has been hugely misleading.  The press largely treat this U.S. “third wave” of COVID cases as if they were the result of some mass moral failure.  As if the entire US Midwest has suddenly gotten “COVID fatigue”.   And that bit of widely-accepted misdirection is unhelpful, and keeps us from focusing on some potentially achievable policy goals.

As a result, we see a lot of tsk-tsking going on, a lot of (well-deserved) finger-pointing aimed at anti-mask Republicans. But not much in the way of sensible discussion that might lead to policies to mitigate this US “third wave”.  That press coverage loses sight of the fact that the entire northern hemisphere (outside of Asia) is suffering from the same problem.  And, as importantly, the entire southern hemisphere is largely NOT suffering from this problem.

Just let that sink in for a second.  Can we really blame this on “COVID fatigue”?  Is it really plausible that half of the earth just decided to throw in the towel, regarding COVID-19 precautions.   But the other half didn’t.  All at the same time? 

Source:  Wikipedia.

Or, maybe, people were being sloppy about COVID precautions all along.  As identified, say, in Wisconsin, back in June (Post #709, Perhaps Cheese Protects Against Coronavirus).   But now, with the change of seasons, that sloppiness matters, and so the press is now discovering “COVID fatigue” that has been occurring all along.

You can believe that the current upsurge is due to mass “COVID fatigure”, if you want.  But you can also look at the weather.  And if you’re a believer in using Occam’s Razor, the change in the weather provides a much simpler explanation of this world-wide phenomenon.

I’ll put the weather-based explanation this way:

(Low humidity) * (poor COVID hygiene) = recipe for disaster.

Assuming that’s true, there are some obvious implications for health care policy and for your own safety that are going to be missed if nobody cottons onto the fact that humidity is a major factor in suppressing COVID-19 spread.


Policy analysis 2:  Implications.

I’ve been beating this drum about humidity now for a couple of weeks.  And I’ll continue to beat it, because I have several sound policy rationales for doing so.  Mainly, humidifying indoor air is cheap to do.  And, based on my reading of the literature, should be an effective way to reduce spread of COVID-19.  Or, more to the point, failure to humidify indoor area will likely contribute to enhances spread of COVID-19.

So just humor me for a bit, and assume I’m right about this.  What are the implications, if you accept that, in the type of climate we have in the U.S.A., respiratory viruses spread more effectively in dry air?

First, there’s the obvious implication of this, for public health policy:  Humdify indoor spaces.  In addition to doing all the other things you’re supposed to be doing to prevent COVID-19 spread.  Make sure to keep the relative humidity in your home and work spaces at 40% or so, because that seems to be the sweet spot for suppressing spread of viral respiratory diseases.

A second policy implication is that we’re going to have to find a way to help people in the U.S. Midwest un-learn behavior that had been successful prior to the change of seasons.  And that’s going to be hard.  If you’ve been going around maskless for nine months, hearing about problems elsewhere, and not experiencing them yourself, it’s going to be a hard to believe that you need to wear one now.

Getting people to change would be a whole lot easier if you could tell them exactly what is different now, relative to two months ago.  So, instead of Republican governors who have U-turned on masks doing this:

I know what I said. Listen to what I'm saying now!

Source:  Yarn

They could say “winter is here, and here’s why that changes things.”  That the behavior that (in hindsight) had held little-or-no danger to them in summer is now a prime driver of COVID-19 deaths in winter.  Surely, people who live in the upper Midwest understand that you have to take winter seriously.  And in some sense, these new mask mandates in Midwest states could be couched as just an extension of that idea.

Third, if low indoor humidity leading to enhanced virus transmission is the root cause, then that needs to trickle up into the new mask mandates themselves.  Both North Dakota and Iowa passed mask mandates that last just a few weeks.  As if this will all have blown over by then.  But if the real culprit here isn’t some sort of mysterious one-time outbreak, but instead is the weather — and low indoor humidity — then a three-week mask mandate is just the worst sort of wishful thinking.  Those mandates will be expiring just as we are going into the heart of the period when they are most needed.

Finally, I keep harping on this because I’m afraid that it’s soon going to be Virginia’s turn.  Sure, we have good levels of mask use here in NoVA.  And that’s going to help.  But it hasn’t really gotten cold around here yet.  And that’s finally starting to happen.  Arguably, we had our first frost in Vienna, VA last night (or maybe we’ll have it tomorrow night.)  And as the temperatures have dropped, I’ve seen the humidity level in my house drop from 65% last week, to just over 50% (and falling) now.

I’ve prepped.  I have two large evaporative humidifiers that I am putting into service this evening.  Even if it turns out that I’m wrong about humidity and viral spread (but I’m not — see literature review below), I’d feel remiss taking this precaution myself and not making the rational for it public.

And so, that’s why, in this post, I’m going to present my review of the scholarly literature on why flu is seasonal.


Background 1:  We’re only talking about temperate climates.

Everybody knows that flu is seasonal.  Except where it isn’t.

First, in temperate (non-tropical, non-arctic) climates in the northern hemisphere, flu season is northern hemisphere winter — December, January, February. As illustrated in the top chart in this posting.  Everybody in the US knows that.

Second, for temperate climates in the southern hemisphere, flu season is  southern hemisphere winter, except when COVID-19 protections essentially eliminate the flu season.  (I am going to come back to this point later, I think.)

Third, tropical countries may or may not have a distinct flu season, may have two different flu seasons per year, and so on.  The whole concept of a distinct “flu season” in the tropics is debatable.   You can see that summarized on this page from the US CDC.

Even within the US, Hawaii has a flue season that can be completely different from the US mainland.  Mostly, that’s due to the high tourist volume — tourists bring their own strains of flu from all over the world.  But also due to a climate that’s radically different from the rest of the US.

Fourth, I could not find reliable information on flu season in arctic climates.  Except to note that Alaska has a highly variable flu season, with a peak that can occur any time from November to March (per the Alaska Department of Health), or possibly as late as May (per this source).

As an odd footnote, there are occasional unexplained colds and flu among groups isolated in Antarctica.

Finally, flu is seasonal, but it never goes away.  The virus itself remains present in the population year-round.  Again, back to that top chart, some people die of flu in the middle of the summer. So it’s not about flu virus being absent from the population in summer.  It’s about the conditions that lead to wide and rapid spread of flu virus, versus conditions that suppress the spread.

One reason to stress “temperate climate” is that the relationship between flu spread and air humidity is, to some degree, U-shaped.  We in the U.S.A. just happen to live on one arm of that U.  And on that arm — the lower-humidity arm — the relationship between typical indoor air humidity and viral transmission shows that drier air leads to more transmission.

Arguably, the “sweet spot” for indoor air humidity is somewhere in the 40%-60% relative humidity range.  If I had to put it in a nutshell, you’re looking for a level of winter indoor air humidity that keeps your sinuses happy, keeps you from waking up with a dry sore throat, and so on.