Here’s the story:
A regional division of the South African traffic department held a recruitment drive, looking for 90 potential traffic officers. Inundated by the response – over 35,000 applicants for the 90 jobs – the department used a 4km fitness test as a “filter” of sorts to trim down the numbers and arrive at their best 90 candidates.
In late December, in the small town of Pietermaritzburg, over 30,000 applicants, aged around 18 to 22, took to a 4km time-trial, effectively racing for employment. South African summers mean heat, and the temperatures were above 30 degrees Celsius (about 90F). The end result of the catastrophe was six deaths, attributed in the media to “dehydration” (more on this below), plus a suicide after the race in frustration at not qualifying.
Whether a fitness test 4km time-trial is a suitable way to squeeze 30,000 into 90 can be debated, as can the obvious implications of this for South Africa’s employment problems. It is a story that has political, management, administrative, socio-economic and even health implications. But let’s stick with the physiology, and discuss the risk of dying during exercise, because it’s a topic that unfortunately comes up often, and the lessons that can be learned are important.
Just last year, in the London Marathon, a 30-year old woman died within sight of the finish line, making news headlines. At a hot Chicago marathon in 2007, a man died amidst accusations of the danger of running in the heat – his death was attributed first to dehydration and heatstroke, later to an existing heart condition. Even elite athletes are not immune – Ryan Shay during an Olympic-qualifying marathon in New York, and Fabrice Muamba during a Premier League Football match. It’s a recurring event, often, but not always linked to high temperatures.
Dehydration – the ever-present scapegoat, for everything
Unfortunately, the media continue to propagate a temperature and dehydration myth, which helps nobody because it obscures the more likely causes. In the case of the six traffic officer deaths in South Africa last week, the first line in most media accounts was “six people have died from suspected dehydration”.
Let’s be clear – you cannot die from dehydration within the first four kilometers of any endurance event. It’s just not possible – the body has too much water to reach a critical level of dehydration, whatever that even means. Dehydration is the easiest diagnosis to make, because we have all been so ‘drowned’ by marketing messages that tell us that fluid loss is a potentially catastrophic risk during exercise and that if we do lose fluid, we will be in mortal danger of dropping down dead.
Physiology says that the body is well able to withstand quite large fluid losses with no detrimental effects on performance or health. It had to be this way, because hunting for our survival didn’t benefit from an “-ade” station every 2km, and those who have heard of persistence hunting will also know that a common tactic was to hunt larger animals in the hottest part of the idea, exposing the animal (and the hunter) to many hours of prolonged exercise, without fluid, in the heat. It worked, because we are adapted for this.
Supposedly, as little as 2% dehydration impairs performance by 10%, which is amusing because when the world’s elite marathon runners finish in 2:05, they have lost at least 2% body weight, which means they’re running two minutes slower than they would’ve done had they listened to many Gatorade advertisements and scientists sponsored to tell this “truth”. The problem is not dehydration, it’s thirst – the discomfort created by feeling thirsty is without doubt detrimental, which is why drinking ad libitium, in response to thirst, is both good enough to ensure our health and to optimize performance.
But we’re not talking performance here, we’re talking mortality, and again, dehydration in an event lasting at most 25 min, is just not on the table.
Dehydration is also blamed for other heat-related afflictions. On Monday night, I watched the BCS Championship Football match, and within about 30 minutes of play, the Alabama quarterback AJ McCarron was shown receiving treatment to his calf muscle on the sidelines. The commentators (who part of me can excuse for not knowing better, but part can’t), speculated that the high humidity in Florida, where the match was played, was to blame for muscle cramp. This after probably only 20 min of playing time for the player. It was later revealed that the calf was injured by contact, not cramp. But again, it highlights the dogma that says that dehydration, among many other risks, also causes cramp, which simply does not make physiological sense, and has recently been discredited by laboratory studies and theoretical flaws (I’ll dedicate a separate article to this in the near future)
Similarly, there is no link between fluid loss and heatstroke. Human beings can safely lose big volumes of fluid without their body temperature shooting through the roof. Typically, in a marathon on a reasonably warm day, we lose about 2 to 3 L of fluid over many hours. Faster runners lose more – Haile Gebrselassie is reported to have finished his Berlin World Record 5kg lighter than at the start. We have a race in South Africa, the Comrades Ultra-Marathon, run over 90km, from morning to evening, with temperatures typically in the mid- to high-20s for about six of those hourse, and controlled research has found that most of the field finish with around 2 to 4% body weight loss, a proxy for fluid loss.
These people are not ill. They may be thirsty, and they sure are tired after 11 to 12 hours of exercise, but there is nothing medically wrong with them. Their body temperature is normal for exercise – that is, elevated to perhaps 39-40 degrees, but this is expected. I can all but guarantee that none of the six men who tragically died in South Africa, or any of the other high profile deaths, which tend to happen in people who are running relatively slowly and in cool conditions, have lost anything like 4% of their body weight. So when next this happens (and it will), let’s immediately disregard the diagnosis of dehydration being the killer, because it simply isn’t true.
Heatstroke – an abnormal physiology in most instances
Heatstroke is a viable candidate for the tragic deaths that sometimes happen, but it’s a grossly overstated risk and those who diagnose any athlete’s collapse or medical condition on a hot day as ‘heatstroke’ are also taking a lazy and possibly very wrong option. The reality is that heatstroke is a pretty complex phenomenon, and is likely to involve some kind of pathology. Once again, I’d draw attention to the difference between the perception of being hot and actually getting to the kind of dangerous temperatures that characterize heatstroke. We’re not talking about feeling hot, uncomfortable and slowing down or stopping here.
Five years ago, I was a co-author of a paper that was written to investigate five hospitalizations during mass-participation events here in SA . Four people died during a 109-km cycle race, and one was hospitalized during a 56-km Ultra-marathon. As is the media way, all were blamed on dehydration and heatstroke.
However, once the specific cases were investigated, the interesting discovery is that none of these athletes was exercising at the kind of exercise intensity that would be needed to raise their body temperatures to the levels measured. An important point is that these cases actually were CONFIRMED as heatstrokes, based on the symptoms observed in hospital, and their highly elevated body temperatures – all were well above 41 degrees celsius. This is unusual, because many times, the temperature is not measured, but the death is attributed to heatstroke anyway because of “lazy” diagnosis.
The principle here is that body temperature rises during exercise as a result of heat produced by muscle contraction, and the harder we exercise, the higher it goes. We lose much of the heat through convection (wind cooling) and evaporation (sweat), but we “settle” on a temperature up around 39 degrees. That’s homeostasis in action.
In these athletes, that clearly hadn’t happened. They’d overshot, gained too much heat and ended up critically ill. Now, there are only two ways for this to happen. The normal control of body temperature is a balance between heat production and heat loss. So to overshoot the normal homeostatic control of body temperature, they have either produced excessive amounts of heat, or their heat loss mechanisms have failed (of course, a combination of both is possible too).
But “normal” heat production cannot explain most cases of heatstroke. If you are running a 4-hour marathon, or cycling along at 15 km/hour, you are not producing enough heat to raise your body temperature to critical levels. This is what the athletes were doing in the study. It’s different for elite athletes who are doing shorter, high-intensity exercise. Running a 5km or 10km time-trial, with extreme levels of motivation, can put an athlete right on the boundary of what one would call “uncompensable heat production”. Closing the final 10km of a marathon at world record pace can push the rate of heat production high enough that if the environment is too warm, it becomes potentially limiting and the athlete must slow down. This is why the world record for the marathon will become more and more difficult to break – it is now close to a thermal limit and so requires absolutely perfect conditions for it to happen. Even a degree too warm over the final 10km is too much. In our lab studies, the highest body temperatures we measure are at the end of 10km time-trials in hot conditions.
But recreational athletes don’t produce enough heat to develop heatstroke through normal muscle activity. Therefore, we look at alternative theories – either these individuals are failing to lose heat, or they produced excessive heat from unnatural means. We called that “excessive endothermy” in the paper, and considered it more likely, because convective cooling on a bicycle is large enough that even a loss of sweating can’t explain how people overheat so quickly in these events.
I’ll never forget being in the medical tent for one of the cases – the runner was brought in, his temperature measured and found to be elevated – above 40. He was placed in a large tub of ice-water for rapid cooling. Over the next half and hour, he got even hotter. Sitting passively in ice, with the most enormous cooling method you can imagine, this athlete was still producing enough heat to push his already high body temperature above 42 degrees celsius.
Also of interest is that many of the documented cases of heatstroke (that is, published in the literature, complete with diagnosis and description), have occurred in cool or moderate conditions, very early on during events, and with low intensities. Here are two examples:
- A 17-year old army trainee develops heatstroke (40C) after only 15 min of fast walking at only 8min/km. The air temperature? Only 17C. One hour after admission, his body temperature has climbed to 42.8C, while he remains unconscious 
- Runner collapses with a body temperature of 42C only 45 min into a 10km fun run at a moderate temperature of 24C. This study documented what were described as 15 cases of heat problems, ranging from mild to serious, out of a field of 13,000, and it wasn’t even particularly warm. Only one was true heatstroke, however, the others were just feeling hotter than usual because they were unacclimatized to the conditions, and this is often confused for ‘heat illness’ – there’s a big difference between feeling hot, and being hot, and training status affects that more than anything 
When you work out the rate of heat production and compare it to the potential rate of heat loss given the documented environmental conditions in these events, you discover that there is no normal way for any of these athletes to overheat unless something goes very wrong (see the Endothermy paper for more).
So the key points from those case studies – there are 18 documented cases, I’ve only discussed three – is that the athletes who suffer REAL heatstroke most often are not exercising very hard, they’re not in impossibly hot conditions, and they show ‘abnormal’ heat gain even after they’ve finished exercise, sitting out of the heat (in a bucket of ice, in one case). Clearly, there’s something else going on, and heatstroke does not happen just because we run hard on a hot day.
Having said all this, in the case of the six traffic officers, you did see a perfect combination of factors for some of these athletes to develop genuine heatstroke. That’s because they were highly motivated (90 jobs available, 1 in 300 chance), untrained (our ability to tolerate and lose heat is poorer when untrained) and running for only 4km, which means a very high relative running intensity, and thus higher rates of heat production.
Were they heatstrokes? Only by measuring body temperatures at the time would this ever be confirmed, and I don’t know if this was done. Autopsies may shed further light, if done, because they reveal changes in the muscle that point to excessively high temperatures and pathological conditions such as rhabdomyolysis, which is one of the likelier candidates for the ‘abnormal’ heat production that I described above.
A colleague of mine, Dr Tertius Kohn, is studying the muscles of animals that are captured or hunted in the wild, because there is evidence of heatstroke in these animals. I once accompanied him on a muscle-obtaining trip, and remember cutting muscle out of an antelope that had been hunted. The muscle was, quite literally, cooked. It resembled a menu item at a restaurant, and his working theory is that under extreme stress, with the right pathology or underlying muscle condition, excessive heat production can overwhelm homeostasis. Is this what happens in humans? Possibly, though too little is known at this point to make any conclusions.
Exercise and sudden death
So if not heatstroke, and if not dehydration (highly, highly unlikely), then what is the most likely cause of death during exercise? Again, this is a topic we’ve discussed a great deal here on The Science of Sport, and I’d point you to these two articles – one written to discuss potential causes of death after Ryan Shay’s death in New York, and another giving some perspective to the issue.
But if you really want to learn a bit more about the prevention of sudden cardiac death, then listen to this podcast, by BJSM with Prof Jon Drezner. In it, he talks about the prevalence, the accuracy and sensitivity of screening, the treatment, the prevention and the education.
The problem is this – there are conditions, underlying and dormant, that increase the risk of sudden cardiac events. A precipitating event can take the form of endurance exercise, the result of which is that the athlete, for all intents and purposes healthy and fit (the London Marathon death, Claire Squires, had just climbed Mt Kilimanjaro), can suffer a cardiac event. Drezner describes a prevalence as high as 1 in 40,000, which means that every major city marathon has ‘candidates’ for this kind of tragic event. So too, the 4-km fitness trial of the traffic officers is likely to expose at least one person to the kind of “precipitating event” to trigger sudden cardiac arrest. It’s no guarantee that it happens, and nor is it guaranteed to be limited to only one.
It’s quite possible, too, that the prevalence is higher in some populations, either randomly or determined by other existing medical conditions, and that the addition of heat as a stress makes it even more likely that the event will occur. Now, in the case of most marathons, the runners who line up on the starting line are to an extent “self-selected”. The 30,000 traffic officers were not, and so in their untrained states, exposed to the stresses of a maximal time-trial, on a very hot day, without screening for those conditions, you have the ingredients for a potential disaster, which is what transpired.
Also, the media coverage is disproportionate in the case of these events occurring during public events. Consider how many cases of cardiac arrest go unnoticed because the person with the risk condition leads a sedentary life. But, when it happens on a football field during a televised match, or during a US-Olympic marathon trial, then the world takes notice, because our expectation is that it shouldn’t happen.
Quite what to do about it is difficult. For professional athletes, the screening debate starts up every time there is such an event. Drezner talks about this in detail in the podcast, and I also wrote some thoughts on it here. Treatment is clearer – the risk of dying as a result of a cardiac arrest decreases from about 50% without an emergency defibrillator, to between 5 and 8% with an emergency defibrillator, so the presence of equipment and personnel to administer treatment is crucial.
Ultimately, there will be deaths during exercise that are neither predicted or preventable, at least for now. Proper training, adaptation to the environment, screening and treatment greatly reduce the risk, but don’t eliminate it altogether.
When trained runners line up to run a marathon, then even in very difficult environmental conditions, the risks are small and probably unrelated to the conditions. But when untrained individuals, be it fun-run athletes, or aspirant traffic officers, try to run at maximal levels, then even short runs or moderate conditions suddenly start to pose great challenges to the physiology. It’s a lesson to heed because it emphasizes the obvious value of training, as well as the importance of staying aware that as much as trained athletes do things that seem mundane, the physiology can be, under the right (or wrong) circumstances, more ‘fragile’ than we think.
- D.E. Rae, G.J. Knobel, T. Mann, J. Swart, R. Tucker, and T.D. Noakes, “Heatstroke during endurance exercise: is there evidence for excessive endothermy?”, Medicine and science in sports and exercise, 2008. http://www.ncbi.nlm.nih.gov/pubmed/18580397
- C.J. Parnell, and J. Restall, “Heat-stroke: a fatal case.”, Archives of emergency medicine, 1986. http://www.ncbi.nlm.nih.gov/pubmed/3730075
- R.L. Hughson, and J.R. Sutton, “Heat stroke in a “run for fun”.”, British medical journal, 1978. http://www.ncbi.nlm.nih.gov/pubmed/709275