Please stop saying hot drinks cool you down

Everyone loves a counterintuitive scientific belief. Is this one actually true?

Brrrrr!

Hot enough for ya?! There’s currently a heat-wave in the UK—apparently temperatures in the South of England might rise as high as 40°C (104°F) in the next few days—so there’s been a rash of articles giving advice on how to stay cool. A lot of them seem to have one piece of advice in common:

Heatwave UK: The surprising drink found to help cool you down - study (Daily Express)

This very unexpected drink will actually cool you down during a heatwave (Cosmopolitan)

Do hot drinks cool you down in hot weather? (Manchester Evening News)

Do hot drinks cool you down in a heatwave? (Metro)

How to beat the heat: why lukewarm showers, hot drinks and closed curtains can help you cool down (i News)

Can hot drinks cool you down? What the science and experts say (Birmingham Mail)

How to sleep in a heatwave: science says to drink this hot drink before bed in hot weather (Stylist)

How to sleep in the heat: Cool down by sipping HOT drink before bed - expert (Daily Express, again)

Dr Amir's surprising hack to stay cool today - drink the nation's favourite hot drink (Daily Express, a third time(!))

There are tons of examples from previous years too, and there’s also one individual on Twitter who has been regularly tweeting about this, bamboozling his (likely fairly gullible) followers.

Where does this mad-sounding idea come from? Apparently it’s an “old wives’ tale”, and “what your grandmother would’ve told you” (which is news to me). But the more recent popularity of the idea stems from one scientific study, published in 2012 and publicised in the Smithsonian magazine. I know this because almost all the above-linked studies refer back to the Smithsonian article and its interview with the (at the time) University of Ottawa heat researcher Ollie Jay:

…a hot drink can cool you down, but only in specific circumstances. “If you drink a hot drink, it does result in a lower amount of heat stored inside your body, provided the additional sweat that’s produced when you drink the hot drink can evaporate,” Jay says.

How does this work? “What we found is that when you ingest a hot drink, you actually have a disproportionate increase in the amount that you sweat,” Jay says. “Yes, the hot drink is hotter than your body temperature, so you are adding heat to the body, but the amount that you increase your sweating by—if that can all evaporate—more than compensates for the the added heat to the body from the fluid.”

So, already we have an “if” here: if all the sweat you produce can evaporate, then it might cool you down. But if humidity is high, or you’re wearing too many clothes, and sweat can’t evaporate so easily, having a nice cup of tea won’t be much use, temperature-wise.

Let’s look at the study in question. Nine “semi-nude” men cycled non-strenuously on a stationary bike for 75 minutes in a warm-ish room (24°C), and had their temperatures measured in their oesophagus (via a probe in their nose) and rectally. They were linked up to a calorimeter, which calculated their heat production from the amount of oxygen they were using and the carbon dioxide they were breathing out. They were also weighed with a very sensitive scale to measure sweat loss.

Just before, and then a few times during, the exercise they were given a controlled amount of water, adding up to just under a litre across the whole experiment, at various temperatures: 1.5°C, 10°C, 37°C, and 50°C. Each participant did four sessions, one at each of the water temperatures.

To note a few things already: n = 9; only males; doing exercise for quite a long time; in a not-particularly-hot room; rapidly drinking water, not sipping a cup of tea. They also had a big fan pointed at them to help their sweat to evaporate away. These are just a few of the differences between this experiment and perhaps the average person’s situation in a heatwave - and I haven’t even mentioned the fact they were doing all this while they had a thermometer “inserted to a minimum of 12 cm past [their] anal sphincter”.

The most important result reported in the paper is that heat storage—measured by the calorimeter I mentioned above—was lower when the participants drank 50°C water, compared to water at cooler temperatures. As Jay explained in the quotation above, he thinks this is caused by additional sweating - and indeed they found more whole-body sweat loss with the hotter drinks.

Here’s the frustrating part: for both these results—the heat storage and the sweat loss—all we’re shown is a graph with some asterisks, and all we’re told is that they’re statistically significant, “p < 0.05”. We don’t get the exact p-value; no standardized effect sizes are given; they don’t even give the means and standard deviations (if you want to, you can use a useful online tool called WebPlotDigitizer to retrieve the means and standard deviations from the graph, but this only gets you so far - since it’s a repeated-measures study where each participant did each condition, we’d need the raw data if we wanted to do any stats or work out the exact p-values).

Why do researchers do this? If I was to be charitable, it’s just sloppiness. They wrote “p < 0.05” and forgot to add in the exact p-values later, maybe. But the annoying thing is that for many of the other, less important results in the paper, we’re given means, standard deviations, and exact p-values. So, less charitably, we might wonder whether the authors are being vague about the p-values because they’re all very close to the borderline for a “significant” result - technically they might be lower than the magic threshold of 0.05, but not very impressively so (not only that, but there’s also no correction for multiple comparisons).

There’s also another complication: although they found the above results in the calorimetry data, they found the opposite for the rectal thermometer data: hotter drinks meant more heat storage. They argue that the thermometer method isn’t as good as the calorimetry one and doesn’t represent total heat storage across the whole body as well.

My teeth are absolutely chattering just looking at this image

All that said, and the ambiguity of the statistics aside, I don’t think this is a bad study: it’s just a small and non-conclusive one. A few other studies on this have been done, and were reviewed in 2018 by Jay and one of his colleagues. The others tend to be very small too - see for example this 2016 one, which also had only 9 participants, but which reported really large differences in heat storage and loss between drinking ice slurry and drinking 37°C water (consistent with the earlier study, drinking the ice slurry meant for less sweating and thus more heat retention overall).

That study didn’t test anything hotter than body-temperature though - indeed, according to the review, only the original 2012 paper and one other study from 2007 have ever used hotter-than-37°C water in this kind of experiment (that latter study, n = 9, found that body temperature didn’t change as much as expected after ingestion of cold or hot water, probably due to the body’s own thermoregulation systems - but it used only thermometers, and not calorimeters, so might be considered fundamentally flawed by the other researchers). That’s it: all the “drink hot tea” stuff is based on a total of 18 men in two studies, drinking 50°C water while cycling.

That’s because these studies aren’t aimed at testing whether drinking a cup of tea will cool the average person down on a hot day. They’re sports-science studies, aimed at understanding what athletes should do to maintain or improve their performance, or get the edge over rivals. Here’s what Jay says at the end of the review article:

…cold water and ice slurry ingestion can be recommended for cooling an athlete during exercise in hot, humid and still environments; but not warm, dry and windy environments. In other conditions, athletes should ingest fluids at the temperature they find most palatable and therefore best maintain hydration status.

Leave my grumbling about the statistics aside and assume that the general idea is correct. This research is not saying that athletes should be running along chugging Earl Grey from their Thermos flask. It’s just that drinking icy water isn’t necessarily optimal for cooling under all conditions. It’s also not clear whether the average person—or indeed an athlete—would even notice the effect: there’s no measure of subjective feelings of coolness or comfort in the studies (though of course such a measure would be vulnerable to the lack of blinding - you know whether you’ve just had ice versus a hot drink).

Jay and his colleagues have, more recently, published a paper in The Lancet on what the average person should be told about staying cool in a heatwave. Conspicuous by its absence is “put the kettle on”. Yes, the stuff about colder drinks leading to less sweating is briefly mentioned, but they say this:

The temperature of drinks if ingested when a person has already started sweating does not affect body temperature.

Not only that, but if you look at Jay’s list of top tips for staying cool, the final one is: “cold water ingestion”.

So, even if you take them at their word—and you should remain extremely wary of the fact all this advice comes from 18 people—“the scientists” aren’t saying that hot drinks are the best thing in a heatwave. This seems like an interesting, under-researched, aspect of human physiology. What it doesn’t seem like is a solid and generalisable enough result that it should be splashed across dozens of news articles every year.

What’s going on here isn’t just people writing about a scientific result. It’s people getting off on being contrarian and counterintuitive, relishing the boggle reaction they get when they tell people that “actually, drinking hot drinks is the best thing to do in hot weather”. We saw this kind of bias towards counterintuition many times during the pandemic, where it confused people about wearing masks (actually, it makes you more likely to get COVID!), and border closures (actually, it has no effect on the number of infected people coming into your country!) and much else. Very often the counterintuitive belief—in this case, “drinking hot drinks cools you down”—is either totally wrong or very context-specific; the more boring, prosaic alternative is often the more realistic one.

Now, if you’ll excuse me, I’ve got a bit of a headache, so I’m off to bash my forehead off a brick wall for 10 minutes straight. That should sort me right out.

Image credits: Getty

Edit 17 July: changed “Canadian” to “(at the time) University of Ottawa”. Changed “nasally” to “in their oesophagus (via a probe in their nose)”, with regard to the non-rectal measurement of temperature. Thanks to Aaron Caldwell for noting these errors!