When researchers from Nike were plotting the details of their Breaking2 marathon project in 2017, one of the variables they considered was start time. The usual early-morning starts give you cool air that gradually heats up—but an evening start could give you cool air that gradually gets even cooler as the runners heat up. They eventually stuck with the morning start, mostly to avoid practical problems like figuring out what runners should eat all day before an evening marathon. But the discussion made me realize that there’s more to race temperatures than what the thermometer reads when the starting gun fires.
A pair of recent papers in Medicine & Science in Sports & Exercise explores the topic of heat stress in the Boston Marathon. Boston is a major outlier among marathons, with a traditional start time of noon that was changed in 2007 to 10 A.M. for the first wave of the mass start—still much later than most races. One of the papers, from a team led by sports science consultant Samuel Cheuvront, analyzes weather data from 1995 to 2016 to conclude that runners were 1.4 times more likely to face conditions associated with exertional heat illness—a spectrum that includes cramping, heat exhaustion, and heat stroke—with the old start time compared to the new one.
That seems entirely logical. But the other paper, from a team led by sports medicine physician Rebecca Breslow, digs into the actual heat stroke data from Boston Marathon medical records and ends up with a more complicated picture, both in terms of who gets heat stroke and what factors contribute to it. Breslow and her colleagues looked at records between 2015 and 2019 (full records from earlier years apparently aren’t available) and identified a total 51 cases of heat stroke out of 11,001 runners who were treated in medical tents along the course or at the finish line.
The runners most likely to end up with heat stroke tended to be younger and faster than the rest of the field. This isn’t as surprising as it might appear. We often think of heat stroke as a consequence of being out in the sun for too long and not drinking enough. But in the context of endurance sports, the biggest factor is the heat you generate yourself—and faster runners generate more heat. In fact, some researchers argue that there’s a greater risk of heat stroke in shorter races like 10Ks than in marathons, because the faster running speeds allow you to generate more heat.
The role of dehydration in heat stroke remains very controversial, and has been since South African scientist Tim Noakes began questioning the links between the two in the 1990s. The most recent American College of Sports Medicine guidelines, from 2007, list dehydration as a risk factor for heat stroke, but also note that “hyperthermia [i.e. overheating] may occur in the absence of significant dehydration when a fast pace or high-intensity exercise generates more metabolic heat than the body can remove.”
That’s what happens to a few high school football players during summer practice every year, no matter how much they drink. And it’s also what happens to some marathoners. The Boston data doesn’t tell us anything about how much the heat stroke patients had drunk, but the medical records do tell us how they were treated. About a third of the patients—18 of 51—were given intravenous fluids. Another nine were simply given something to drink, and 24 weren’t given anything at all (at least according to the medical charts). All of them recovered. This doesn’t prove anything one way or another, but it does contrast with the popular image of heat stroke victims as parched stragglers who collapse because they didn’t drink enough.
The final key point was the weather. Marathon conditions are often expressed on a scale called wet-bulb globe temperature, which reads like an ordinary temperature but incorporates other factors that also affect heat stress like solar radiation, humidity, and wind. Road race directors use WBGT to assess medical risks for their events: for marathons, a value above 70 degrees Fahrenheit signals an elevated risk of heat stroke.
For the five races studied, here are the starting WBGT values and the peak values in the four hours following the start:
2015: 43.0 F / 45.1 F
2016: 70.0 F / 70.0 F
2017: 63.0 F / 70.0 F
2018: 41.0 F / 45.0 F
2019: 58.0 F / 69.1 F
It’s not hard to guess that heat stroke wasn’t an issue in 2015 and (famously) 2018. But which year do you think had the most cases of heat stroke? There’s one obvious answer—and like many obvious answers, it’s wrong. There were just four cases in 2016, the year with the highest WGBT values, compared to 21 in 2017 and 26 in 2019.
What explains the difference? It’s impossible to know for sure. Factors like wind speed and cloud cover were generally similar in the three warm years, and are accounted for in the WBGT values anyway. But there is one thing that jumps out. In 2016, it was hot at the start but didn’t get any hotter. “It cooled off toward the end actually,” one runner told the Boston Globe afterward. “You could tell a change, but it started off, and it was hot.” In contrast, the starts in 2017 and 2019 were a bit cooler but then warmed up during the race.
I asked Breslow why she thought this made a difference. “One possibility is that runners start more slowly if it’s already hot,” she said in an email. Beyond the behavioral explanation, it’s also possible that a rise or fall in temperature during the late stages of a race is what matters most physiologically, because that’s where runners are pushing hardest—and generating the most heat—as they approach the finish line. Breslow also pointed out previous data from Marine Corps recruits that found heat stroke to be most common between 7 and 9 A.M., when WBGT is increasing most rapidly, rather than later in the day when it’s steadily high.
Is it possible, then, that Boston’s midmorning start is actually better, from a heat stroke perspective, than an early morning start would be? Frankly, I doubt it. Given the choice between a rising or falling temperature in otherwise comparable conditions, it seems like falling temperatures might have some underappreciated benefits. But if it’s cool enough, as in 2015 and 2018, then heat stroke is a non-issue. If you’re playing the odds year after year as a race director, then your best bet for cool conditions is probably an early start, as Cheuvront’s study suggested. Still, when you’re checking that pre-race hourly forecast, it might be worth keeping trajectory in mind. If the mercury is rising, be a bit more cautious than the conditions seem to call for. If it’s dropping, go nuts.
For more Sweat Science, join me on Twitter and Facebook, sign up for the email newsletter, and check out my book Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance.