My mother always tells me to bundle up before I go outside during the winter, because otherwise I’ll “catch a cold.” When I first learned about the germ theory of disease, I thought she was wrong. Cold doesn’t make you sick, I thought. Germs make you sick.
Recently, it’s become obvious that I was wrong and she was right. Germs are responsible for disease, but it’s crucial to remember that germs are everywhere. Our immune systems are constantly fighting these microscopic invaders. In fact, we’re probably infected with the common cold to some extent even when we don’t show any symptoms.
During the flu season of 2005, an experiment was performed to test the idea that being cold can make you sick. 90 people kept their feet in a bowl of ice water for 20 minutes, while a control group of 90 people put their feet in an empty bowl for 20 minutes. Over the next 5 days, 29% of the group with chilled feet developed cold symptoms, compared to only 9% of the control group.
Professor Eccles explained this effect by saying that our bodies restrict blood flow to the extremities when we get cold to help conserve body heat for the torso and brain, which really need to be warm. Cutting off the blood flow reduces the supply of white blood cells which are the immune system’s primary weapon against germs.
While his explanation makes sense, there may be a more general effect at work. The human body is a machine that accepts fuel in the form of food, and uses that fuel’s energy to keep us warm and to power our immune systems, muscles and brains. However, in frigid conditions our bodies have probably evolved to say “who cares if I might get sick a week later when I’m going to die of hypothermia in half an hour?”
In other words, the optimal survival strategy during bitterly cold conditions is probably to divert all the energy normally used by the immune system into keeping our bodies warm. Mammals whose bodies didn’t make this sacrifice weren’t as susceptible to disease in the long run, but that didn’t matter because they dropped dead of hypothermia before they could enjoy their good health. So that survival strategy would be eliminated by natural selection.
Anyway, the point is I was wrong, and my mother was right. Sorry, Mom!
Last modified April 7th, 2009
I was always pretty much with your mother on that. It’s pretty obvious that there’s no more reason to get a cold in the winter than the summer, if all you’ve got to go on is Pasteur.
But some seasonal diseases are still hard to explain. For example, why do flu epidemics come mostly in the winter? Moreover, why are there successive waves of mutated flu viruses that sweep around the world each winter?
My personal favorite explanation has always been the Pigs in the House theory. (Ok, maybe it’s my favorite because I made it up.) That is, when it gets cold, people bring farm animals into their houses, or barns immediately adjacent to their houses. This allows the viruses that like to migrate between humans and domestic animals to start an infection cycle that then sweeps the world, thanks to modern methods of rapid travel. Of course, people in the U.S. and Europe don’t bring their animals into their houses when it gets cold any more…but people in poorer, more backward countries–like China–do. That’s why flu epidemics come from “Asia”. Well, it makes as much sense to me as any other explanation I’ve ever heard.
To reply to my own comment…do flu epidemics come in the “summer” (i.e. same time as our winter) in the Southern Hemisphere…or do they coincide with colder weather. Facts are urgently needed here…
According to the Wikipedia article on the flu season, there are actually two flu seasons each year for the Northern and Southern winters. In fact, the World Health Organization produces two different vaccines every year– one for each flu season.
I thought the winter flu season could be explained by a combination of my “energy diverted from immune system to body heat” idea and the fact that people stay indoors more in the winter, allowing viruses to spread more easily. Lower humidity during the winter may be a factor too. Now that you mention it, Pigs in the House probably plays a role in less developed countries. But your question got me thinking, so I did some googling that led me to this article which reports on research performed in March 2008.
Apparently, the influenza virus which causes the flu has a special protein coat (a protective “shell” around the infectious DNA “payload” of the virus). This special protein coat is stable even at low temperatures which allows the virus to survive even in frigid temperatures. Once the virus is inhaled, the protein coat gently melts in the throat and releases the virus.
Just because a culture is different does not mean they are “backward”.
I have not read the details of the experimental method used by Professor Eccles, but it does not seem very rigorous.
Immersing the feet in ice water would only reduce the temperature of the feet (and perhaps the lower leg), but not the core body temperature. The site of infection is in the throat and sinuses (extending to the chest if severe), but the temperature here is not going to be affected by cold feet. On a cold day however, the chilly air would cool these parts of the body – though how much is significant is unknown – not the feet.
“Cutting off the blood flow reduces the supply of white blood cells which are the immune system’s primary weapon against germs” – this is fundamentally wrong – the blood flow is not “cut off” – it is redirected to the torso – the number of white cells remains unchanged, as does the ability to fight infection.
While it is true that metabolic rate decreases with decreased temperature (just like all chemical reactions) this decline is not significant at the range of body temperatures viable with life – normal body temperature is 37C – hypothermia is set at 35C and very few survive a fall to 32C (except those on bypass having heart surgery)
The fact that transmission of the cold and flu viruses is higher in cold seasons, is thought to be mainly due to the increased crowding of people in buildings and public transport (unwilling to go out). In addition, rather than being “stable even at low temperatures” it is more accurate to say that they are LESS stable at HIGHER temperatures
By decreasing the temperature of any extremity for a period of time you would decrease the core temperature.
While an infection may manifest in a particular area, you will find that the infectious agent is present in other areas of the body. It just happens that the majority of the infection is localized to a particular area.
So you’re saying that a significant diversion of blood flow does not mean “cut off”. I’m fine with that, but a a significant diversion of blood flow directly impacts the population of white blood cells in an area in the body. Semantics.
Perhaps there are multiple factors that can contribute to the increase in flu infections? How does your theory contradict the stated theory? Also, the author cited a study to back up their theory. Please cite yours for evaluation.
Here’s a more comprehensive overview of Eccles’ study. He says “It’s well known that chilling the feet causes vasoconstriction in the nose.”
This wasn’t well known to me; in fact I was impressed by the apparent lack of connection between the temperatures of the feet and the nose, given the evidence he presented. The risk of developing symptoms of a cold more than doubled when peoples’ feet were immersed in ice water, which seems to suggest that there’s a causal relationship at work.
Granted, the p-value of this result is a little large for my tastes at 0.047. (A p-value is the chance that the result would’ve happened even if the tested hypothesis were false. Lower p-values mean an experiment is more convincing.) This p-value is only just below the customary “p < 0.05″ limit for clinical studies, and I tend to agree with researchers who want to place a more stringent limit like “p < 0.01″ in order to prevent the publication of false positives. But I think Eccles’s findings are strong enough to take this hypothesis seriously, at least until more evidence is gathered.
I’m not talking about a decrease in metabolic rate. Instead, I’m proposing that as the body temperature drops below 37° C, metabolic activities are focused on a specialized set of activities that keep the body warm. For example, blood flow to the extremities is restricted by vasoconstriction, which requires energy. When that energy runs out, the blood vessels open up again even if the body is still cold, which results in a feeling of warmth. As a result, it’s common for hypothermia victims to shed their clothes right before dying.
I’m guessing that there are other mechanisms that perform the same trade-off. Energy that would otherwise be sent into the immune system is re-routed to tasks that provide/retain the most body heat for the fewest calories of fuel. I doubt this phenomenon would result in a lowered metabolic rate. Quite the opposite, in fact.
You’re right to say that the influenza coating melts at higher temperatures, which results in an alternative mechanism: “warm weather melts the coating, slowing the infection rate of influenza and other similarly coated pathogens.” This may act in concert with the other hypotheses I’ve discussed; Anonymous is correct to say that multiple factors may account for the cold weather/sickness correlation.
What caught my attention wasn’t the stability of the coating at lower temperatures, but rather the fact that the coating melts gently as it’s warmed. There’s no phase transition that could harm the virus; it’s completely continuous. That seems like a very effective cold-weather adaptation.
I just noticed an inconsistency. The first link to Dr. Eccles’s research says that 29% of the group with chilled feet developed cold symptoms, compared to only 9% of the control group. Each group had 90 people, so that’s 26 and 8 people sick respectively.
But, the second link explicitly says that 13 people with chilled feet got sick compared to 5 of the control subjects.
Both articles agree on the number of people in each group, so there’s a problem here. Which article, if either, is correct?
I emailed Dr. Eccles, and he said that both statements are correct, but they’re referring to different parts of the experiment. The first statement refers to colds diagnosed via symptom severity changes and the second relates to self-diagnosed colds.
He was also kind enough to give me a copy of the paper. The first link’s data (29% chilled colds, 9% control colds) was based on a repeated survey that ranked various symptoms from 0 (not present) to 3 (severe). This result has a p-value of 0.001.
The second link is referring to a simpler repeated question “are you suffering from a cold?” This result has the p-value of 0.047 which I previously quoted.
Something’s been bothering me about Dr. Eccles’s study. It couldn’t possibly have been double-blind because there’s a perceptible difference between ice water and an empty bowl. I wonder if the increased reporting of cold symptoms could have been enhanced by the subjects’ expectations of catching a cold.
I’d like to see another independent study reproducing these results at some point, but the only way to create a truly double-blind study would be to sedate the subjects, immerse their feet, dry them off and warm their feet back up, then wake them up. Hopefully the subjects wouldn’t know if they were in the control or experimental group if these precautions were taken.
Dr. Eccles quickly responded to these concerns:
Cold Weather Really Can Make You Sick is a statement which is backed up by a considerable amount of epidemiological data. Studies have shown very convincingly that there is a very direct correlation between sudden drops in temperature and morbidity and even mortality.
On the topic of the common cold one very important factor that is a likely contributor to the common cold is the drying of the upper airways. Airways are instantly challenged when exposed to cold air, which happens with the first breath you may take on a cold day. Cold air is inherently dry and the inspiration of this dry air was first studied by Dr McFadden who clearly established the penetration of cold air deep into the thoracic region. Prior to Dr McFadden’s research it was commonly believed that all inspired air way preconditioned extrathoracicly. Many health care professionals still believe this.
The result of cold air penetrating deep into the airways is the potential to dry the mucus membranes lining the airways. The mucus membrane is the airways front line of defense. When the airways dry they loss this line of defense and become exposed to any pathogen which may be present in the airways awaiting the opportunity to attack. Of course the body has methods to compensate for this challenge but it cannot compensate immediately and hence just a few breaths of cold air may be enough to trigger that cold or sore throat.
All of this is very consistent with Dr Eccles’s cold feet study since the hypotheses are centered around the occurrence of vasoconstriction and blood shunting to the core. When this occurs the blood flow particularly in the nose is reduced and as a result the airway mucosa may dry and leave the door open to virus and or infection.
Best advice is never doubt your mother.
Heh. Agreed. Thanks for the info!
Another interesting idea about the relationship between the increase in cold and flu incidence during cold weather is that there appears to be an inversely proportional relationship between a person’s vitamin D levels and their likelihood of contracting a viral respiratory illness. Since one of the ways a person can increase his vitamin D stores is exposure to sunlight, it makes sense that with colder weather (and greater coverage of the body) and the less direct sunlight that occur in the winter could lead to lower vitamin D levels and greater chance of illness.
Hmm… my mom also tells me to get outside more. Maybe this is why.
I’m curious to see if this relationship is spurious or not. They say that vitamin D levels are positively correlated with exercise prior to entering the military, and inversely correlated with smoking. Perhaps these factors make vitamin D appear to be related to immunity, when it’s really just a matter of exercising and not smoking. We’ll have to wait until their supplement studies are published to find out.
Anyway, thanks for the interesting link!
I just noticed these sentences in the paper:
So my objections were already considered and disproven.
I live in Los Angeles where it rarely gets below 50 during the cold/ winter season. The amount of energy used to keep my body warm on a 55 degree day in a coat has to be negligible at best, yet I get sick every winter. More so the energy spent has to be significantly less than if I were to jog a mile in 90 degree weather no? And people aren’t getting sick more in summer. And what about athletes in general wouldn’t they have insanely high infection rates? So it can’t simply be the depletion of an energy payload diverted from the immune system can it? The air and airway theories seem to make more sense if it’s cold weather related at all.
Humans shed excess heat by sweating, which results in dehydration but the energy required probably isn’t comparable to shivering and blood vessel contraction in the extremities, etc.
Conscious exercise also consumes energy, but perhaps involuntary shivering has a higher “priority” in the body’s metabolism than exercise unaccompanied by massive doses of stress hormones. In other words, maybe natural selection favors metabolisms that don’t divert crucial immune system energy to exercise unless the creature is in fight-or-flight mode.
No. That’s why I didn’t suggest that all cold-correlated illnesses are due to a single cause. Multiple factors are certainly involved, as discussed here and here and here.
Yes, I’ve already agreed with Bruce that this factor seems significant. What I’m suggesting is that “vasoconstriction in airways leads to a cold” is just one example of a more general pattern where our bodies sacrifice long-term health for immediate warmth.
What experiment was this? This experiment described above doesn’t prove anything. First of all, immersing someones feet in ice water for 20 minutes will not do anything but cause their blood vessels to constrict at their feet – called peripheral vasoconstriction. Next, the experiment states within 5 days, 23% of the experimental group developed cold symptoms. Nowhere in the experiment does it state if either of these groups were exposed to a known virus or bacteria. If both groups for example were exposed to the same virus at the same time for those twenty minutes and then observed over the next 5 days with no external contact, then maybe the experiment could hold some value. But we have no idea what this experimental group did in the 5 days after, who and what they were exposed to, versus the control group.
In summary, I do believe that cold weather can lower host defenses. But hypothermia is very different from going out without your jacket for 30 minutes in the 35 degree weather. Putting feet in ice water for 20 minutes is not considered a form of hypothermia.
Aside from the fact that the word “experiment” in that excerpt links to an article describing this research, you must have missed the discussions I had with one of the authors. Judging by the rest of your comment, you missed every other comment in this article, too.
I apologize but I didn’t read every single comment in this discussion. Just a few. However, from what I read, I still feel the study has some flaws.
I looked up the study and found this.
According to the conclusion:
“Acute chilling of the feet causes the onset of common cold symptoms in around 10% of subjects who are chilled. Further studies are needed to determine the relationship of symptom generation to any respiratory infection.”
However in the results section it states: “The subjects who reported that they developed a cold (n = 18) reported that they suffered from significantly more colds each year (P = 0.007) compared to those subjects who did not develop a cold (n = 162).” I apologize if I am incorrect but I did not see any mention of this is the discussion above. Could it be that the patients that reported these changes were more prone to infection due to factors other than chilling of the feet? Or could these patients just report more symptoms because they generally feel they get more sick than others? This brings up another point.
Symptoms are subjective because they are experienced by the patient. As per the last line of the conclusion, “Further studies are needed to determine the relationship of symptom generation to any respiratory infection.” The study did determine any actual infection, just subjective symptoms.
Dr. Ronald Turner of the University of Virginia states “The researchers didn’t check to see if a virus was ever present. They measured symptoms… They didn’t do any virology, so that study has nothing to do with becoming infected.”
There were two studies done in the late 50′s and 60′s and both concluded that cold weather and infection rates by the cold virus were not related.
I don’t think the points discussed above were brought up prior to this, but if they have, I apologize.
Discussed here.
Dr. Eccles already said this study examines how cold activates latent infections.
That’s the point of having 90 people in each group. Larger sample sizes reduce the chances of the experimental group getting exposed to another infection vector “significantly” more than the control group.
Asking people to stick their feet in frigid water then answer some surveys is probably easier than asking them to stick their feet in frigid water, then sit in solitary confinement for 5 days. Good luck getting that study funded.
As Dr. Eccles noted, “It’s well known that chilling the feet causes vasoconstriction in the nose.” As Bruce noted, vasoconstriction of the upper airways dries them out, removing the first line of defense against airborne pathogens. Of course, all this is on the first page of the paper too.
Not surprising, given that Dr. Eccles’ hypothesis concerns activation of latent sub-critical infections. Here’s the last paragraph in results:
There was no difference in colds incidence between the two test groups at baseline as illustrated in Table 1. However, when looking at both test groups combined, those subjects who believed there were suffering from a cold had a history of more colds each year (median 2.00, range 1–10) compared to those who did not develop a cold (median 3.00, range 2–8, P = 0.007).
This last sentence has me confused. It seems backwards. But their conclusion seems reasonable:
An interesting finding in the present study was that the subjects who reported they developed a cold after the chill or control procedures also reported that they suffered from significantly more colds each year, than the subjects who did not report a cold after the procedures. This finding may indicate that there is a sub population in the general population who are more susceptible to developing common cold symptoms each year and that they may have a ‘common cold constitution’.
The real question is, why would these people be more common in the experimental group rather than the control group? As the paper says:
“The differences between the chilled and control groups could have occurred as a chance finding, as it was expected that some subjects would develop cold symptoms due to natural exposure to common cold viruses. The probability value for the different reporting of colds was just below P = 0.05 (P = 0.047) but the P-value for the difference in total symptom scores was more convincing with P = 0.013, and for the dichotomous analysis was P = 0.001. With two different measures of the incidence of common cold providing significant differences between the two test groups it is unlikely that the results are solely due to chance.”
Already discussed, and I’ve already said that I want to see future research.
Yes, of course. The paper’s third paragraph specifically states that Dr. Eccles’ hypothesis regards the activation of latent, sub-clinical infections.
That’s why the paper’s second paragraph includes references to papers (6-8) which make exactly that point.
Also, the citations list two relevant papers.
“Not surprising, given that Dr. Eccles’ hypothesis concerns activation of latent sub-critical infections…. ”
Virus latency or viral latency is the ability of a pathogenic virus to lie dormant within a cell, denoted as the lysogenic part of the viral life cycle. A latent viral infection is a type of persistent viral infection which is distinguished from a chronic viral infection. Simply a latent virus is a virus that is never totally cleared, is not replicating, and not producing any symptoms of infections. HSV, HPV, & VZV are some examples of latent viruses. However, I have never read any documentation stating latency with viruses that cause the common cold. However, this may exist. If you have any literature, please let me know as I’d be interested in seeing it.
“As Dr. Eccles noted, “It’s well known that chilling the feet causes vasoconstriction in the nose.” As Bruce noted, vasoconstriction of the upper airways dries them out, removing the first line of defense against airborne pathogens. Of course, all this is on the first page of the paper too.”
Drying out upper airways does remove the first line of defense against airborne pathogens. But then viruses that cause the common cold are not part of the normal flora of the upper airways, to my knowledge (Again if you have any info that proves otherwise, I’d love to see it). The re-activation of a latent infection would mean they would have to be there to begin with. So not quite sure what you mean by latent sub-critical infections.
“The real question is, why would these people be more common in the experimental group rather than the control group? …”
Not really sure what you mean by this. The people that admitted to getting symptoms of a cold admitted that they suffered from significantly more colds each year. These individuals were both in the experimental and the control group. It wasn’t that only the experimental group stated they suffered from colds more each year. Couldn’t this just be that the experimental group had more people who are more susceptible to colds than the control group?
So I’m a bit confused to what this study proved other than, according to what you said, vasocontriction and drying up of upper airways due to immersing feet in ice water for 20 minutes.
I’m referring to paragraph 3, sentence 3: Some years later Sir Christopher Andrewes suggested that exposure to a cold environment may trigger the development of a cold but only in people who are carrying the latent cold virus. (Reference: Andrewes C. The Common Cold. New York: Norton; 1965.)
I’m not a professional biologist, so I won’t claim to know whether your definition of “latent” is the only correct use for that word in this context. But perhaps you should ask Dr. Johnson and Dr. Eccles that question?
Again, I’m not a biologist, so these aren’t my words. Paragraph 3, sentence 4: Eccles developed these early observations by proposing a hypothesis that acute cooling of the body surface causes a reflex vasoconstriction in the nose and upper airways, and this vasoconstrictor response may inhibit respiratory defence and cause the onset of common cold symptoms by converting an asymptomatic viral infection (sub-clinical infection) into a symptomatic viral infection (clinical infection). The novel idea in this hypothesis was that when common cold viruses are circulating in the community a proportion of those infected will have sub-clinical infections, and that when any of this sub-group are exposed to chilling of the body surface this could aid conversion of a sub-clinical infection to a clinical infection. This study was aimed at testing this hypothesis, by studying the onset of common cold symptoms after acute chilling of healthy asymptomatic subjects, during the winter, when common cold viruses are circulating in the community. (Reference: Eccles R. Acute cooling of the body surface and the common cold. Rhinol 2002; 40(3): 109–114.)
Table 1 states that the median number of colds per year is the same (2.0) for both groups. In fact, the experimental group’s range only goes up to 8 colds per year, while the control group had at least one person who caught 10 colds per year. If anything, the experimental group was slightly less susceptible to catching cold.
Those are just suggested causal mechanisms. The real point is the observed correlation between exposure to ice water and development of cold symptoms. As I’ve said, I’d like to see more studies with larger sample sizes and better controls to make the P value smaller. But the only two studies I’ve found citing this paper seem to confirm its findings, so for the moment I remain impressed with this study by Dr. Johnson and Dr. Eccles.
A third paper.
Well thanks for clearing up the confusion about median number of colds in both groups. I must have misinterpreted the passage.
As far as the correlation to exposure to ice water and development of cold symptoms, I don’t buy it. But I agree with you that further studies should be done to check for actual infections, not just subjective symptoms as it would take out some of the questionable variables.
Thanks for all the links. Very interesting stuff.
No problem. To be fair, I’m a little confused about one of their median cold statements too. Sorry for being rude earlier.
No worries. I apologize as well. I came off a bit arrogant myself.