Welcome to the Science of Sport, where we bring you the second, third, and fourth level of analysis you will not find anywhere else.

Be it doping in sport, hot topics like Caster Semenya or Oscar Pistorius, or the dehydration myth, we try to translate the science behind sports and sports performance.

Consider a donation if you like what you see here!

Did you know?
We published The Runner's Body in May 2009. With an average 4.4/5 stars on Amazon.com, it has been receiving positive reviews from runners and non-runners alike.

Available for the Kindle and also in the traditional paper back. It will make a great gift for the runners you know, and helps support our work here on The Science of Sport.

Thursday, October 22, 2009

Deaths during running: Is exercise safe?

Sudden death during exercise:  The media, risk and running

For those who have not heard or read the news, three runners died during the Detroit Marathon/Half-marathon last weekend.  All three were running the half-marathon, and were aged 26, 36 and 65.  The three collapsed within 16 minutes of each other during the race.  The timing, the wide spread of ages, and the fact that three deaths occurred in one race (which was not hot, I might add) have given the story 'legs', and it was even covered on CNN.

Whenever this kind of incident happens, there is debate and discussion around the safety of running.   It's always bad press for running.  It is a topic that pops up fairly regularly, and we've actually covered it in a fair amount of detail here on The Science of Sport.  I'm not going to go into enormous detail on sudden deaths during exercise again, but for those who are interested, you may like to read the following posts:

A general perception of safety

So as I said, I wouldn't ordinarily post on this topic again and in the case of the Detroit runners, I don't know the cause or any details, and so it would be wildly speculative to discuss specifics.  However, this latest incident, and the media reporting around it, reminded me of a thought I'd had while watching the NBC coverage of the Chicago Marathon about 2 weeks ago.

You may recall that Jonathan and I were in the control area, doing real-time pacing for the race, and also spent some time in the medical tent.  But it was watching NBC's coverage, that I was struck by the fact that the person who was interviewed the most during a 3 hour broadcast was the race doctor, George Chiampas.  On no fewer than four occasions, Dr Chiampas was featured in a two -minute interview, giving his thoughts on race hydration, race safety, post-race safety, recovery, training and so forth.  And while he answered the questions very well, it was clear that the 'safety/danger' of running was of utmost importance to the broadcaster.

It struck me that there is a very real perception among mainstream media in the USA (remembering I'm from South Africa and so normally unaware of this message) that running is a risk.  That is, viewers who watched the broadcast of Chicago and who were NOT runners would be left under no illusions that attempting to run a marathon is a dangerous task.  The "shock and fear" coverage, which implies danger at every turn, sends a clear message that if you run a marathon, you are taking a chance with your life.

And this unfortunate, because it ignores the whole other side of the argument, and does so with little to back it up other than infrequent and over-hyped incidents.  A thoughtful, balanced approach would cover two additional aspects:
  1. It would consider whether the risk of death during running is in fact greater than during any other activities, and;
  2. It would look at whether the average  runner (from recreational to the marathon) was deriving a benefit from running, and whether this person was in fact less likely to die than someone who chose to stay on the couch because of all these "life-threatening" risks.

Millions of hours invested, but even stats don't tell the full story

Then there is the statistical approach, which many resort to in cases like this.  I read in a report from Fox News that a total of 425,000 runners completed marathons in the USA in 2008, and another 715,000 completed half-marathons.

If you convert that to time, assuming that the average marathon and half-marathon time is 4 hours and 2 hours, respectively, then you can work out that a total of 3.1 million hours of running time goes into those races.  And this does not include the training, or the 5km and 10km races done along the way.  If you assume that the average person trains 2.5 hours a week for 3 months to run a marathon or half-marathon race, then you get a grand total of 34.2 million hours of training time per year for those runners.  The total running time for marathon and half-marathon runners in the USA per year?  37.2 million hours of running (and this is an underestimate, I must point out - it does not take into account the millions who spend an hour a week jogging in the gyms, or those who train but don't race)

So what is the frequency of mortality for these runners?  Fox News reports that about 6 deaths per year occur during races.  How many during training?  We don't know, unfortunately.  But the point I'm trying to make is that these deaths are rather less common than they may seem - one per million hours, perhaps?  One per three million hours?  Until that is quantified, reports that marathon running is dangerous are simply irresponsible, the result of a classic 'media-led knee-jerk reaction', where news reporting makes us over-estimate the prevalence of such events.  A classic example is shark attacks - they are exceedingly rare, but when they happen, they're so dramatic that they receive hyped-up media exposure and so we think they're more common than they are.  I suspect the same is true for running-related deaths.  (What would be great is to compare this number with other activities - driving your car, flying, playing other sports.  If there are any economists or actuaries out there who know this, please speak up!)

The real story - the benefit that the media don't report

But these stats don't tell the whole story anyway.  What you really need to ask is whether exercise adds up to a longer, healthier life, even taking into account what I believe is a tiny risk.  In other words, you need to look at the overall benefit of being active, and ask whether those who run are less likely to die than those who do not?  There is no doubt, based on the evidence, that exercise reduces the risk of morbidity (disease) and mortality (death).  One of the most famous names in exercise science and health is Ralph Paffenbarger, and he demonstrated pretty clearly that increasing exercise was associated with decreased risk of disease and death.  The most famous study is perhaps this one, his Harvard Alumni paper.

Paffenbarger went so far as to show that people who exercised AND smoked, were less at risk than those who didn't smoke, but didn't exercise either.  So, if you want a debate about the benefits of exercise (and I include running here), the real issue is whether those three deaths, and then dozen or so that seem to happen each year during running, outweigh the fact that those same people, if inactive, would have a lower life expectancy and health status?  I doubt it does.

And I wish that the NBC, and all the other media covering running events here in the US and the rest of the world, would acknowledge that instead of focusing on the small risk of injury or death, there is a far bigger positive outcome to being active.  Maybe in future, doctors like George Chiampas will be explaining why those sitting on their couches SHOULD be getting up to run, rather than telling those who are running how not to hurt themselves doing it!

Fitness does not protect you, but nor does being under-trained increase your risk

One final point I have to make, in response to what I've seen is being discussed about this issue, is that people are not necessarily more likely to die from a cardiac event during exercise if they are untrained.  A lot of people have said that people who die during marathons are themselves responsible, because they're running when they are not fit enough.  This is not true, to the best of my knowledge.

The reality is that people who die during exercise have some underlying, probably undetected condition that predisposes them to a cardiac event during exercise.  Those who are simply unfit don't die - they just stop at the 10 mile mark (or sooner) and walk the rest of the way, because their brain does not allow them to continue running.  The fact of the matter is that there are conditions that predispose us to sudden cardiac death, and exercise can bring this out - but it could happen to the elite (Ryan Shay, a few soccer players in recent years) or to the average runner.  It's not that they're unfit or undertrained.

Of course, behaviours contribute to some deaths - overdrinking, for example, can lead to hyponatremia and death.  But even here, the criticism belongs with those who advocate excessive drinking, the dangers of "dehydration" and advertise sports drinks to unknowing consumers, not to the athlete who makes the mistake.

So in the light of the latest events, and until toxicology reports are in, deciding on the cause is premature.  Agreeing that it's sad for all involved, but recognizing that it's not running that killed them, is the way to go!



smackrabbit said...

I'm just speculating here, but I imagine that the reason that the doctor was interviewed so much at the Chicago Marathon is because of the race two years ago with the unusually high temperatures and the death that occurred during the race. I don't recall them talking to the doctor during the Boston Marathon coverage, or NYC Coverage, and so I believe that this was specific to that race for that very reason. In 2-3 years perhaps they won't interview the doctor at Chicago anymore, but next year I'm sure the race doctor at Detroit will get plenty of coverage as well locally.

Colenso said...

Spot on.

For further proof of just how hysterically worked up some of the Yanks can get about possible death by running, just read some of the anti-running & anti-marathon diatribes in the comments on TPP's latest post in her NYT Well blog at http://well.blogs.nytimes.com/2009/10/20/are-marathons-safe.

IMO, the only scientifically accurate term for a lot of the most rabid anti marathon posters is nutters - technically speaking of course, complete nutters.

kjn coach said...

It seems to me that Ross did a mistake while discussing two different questions in the same post.

Asking if running a marathon is risky for you life is something, but talking about the benefits of physical activity on general health is something else. We can't pass from the first subject to the other. One do not need to run a marathon to be physicaly active and reduce risk of health problems. He may even be a good idea to not run marathon if you want more chances to live longer.


a) Running a marathon may not be as dangerous as it seems when we listen to medias.

b) Physical activity on a regular basis is, for most of us, good for our health


-From a and b we cannot logicaly conclude that running a marathon may be GOOD for health. At best, it is not bad (but it still may be).

Ross did not say that running a marathon is good, so I am not trying to say he did. But when he moved the post from marathon's risk to health benefits of training, he let us think that these two topics are related when they are not. It is not because marathon is not as risky as shown in the press that we should start doing marathon to live longer. These are two different questions

I am a marathoner myself. I suspect there is a serious (while not statisticaly high, as shown) risk while running this event. It seems rather impossible to me, at the 35k point, that I am doing something good for myself while running this thing. I really want to do something great for my body and general health, I will start training 5 or 6 times a week with all kind of sports and sometimes doing a little 5, 10 or even 21.1k race.

Ross Tucker and Jonathan Dugas said...

Hi folks

Thanks for the comments.

To Smackrabbit:

Yes, true, the 2007 race was bad and they were on edge. I don't think it's a unique phenomenon, but without a doubt it contributes to the hype, which only re-inforces the general over-estimation of a small probability of risk when running, based on the media.

To kjn coach

While you are correct in your final paragraph that the physical activity benefits could be obtained simply by training, your argument is one that works in paper, not practice.

Therefore, I disagree. One can very easily pass from one to the other, because people who run marathons are physically active (through their training) and therefore receive the benefits of physical activity, just as a regular gym-goer would. Where there is a difference is that some marathon runners train exceedingly hard, and the volume of training they do puts them at risk of injury. Not sudden death, but injury, which this post was not about.

So while you're right that I did not say that running marathons was good (I believe the training one does for it is), the whole point is that people who run, with the intention of improving their fitness and running, are highly likely to take on a marathon at some point. I don't know how many - 1 in 2? 1 in 3? I don't think it really matters, because what is important is that the marathon is merely the goal, the incentive to train.

So many of the runners who train for the marathon are physically active only because that's their objective. Yes, they could be physically active training for 10km and so forth, but they aren't. And because the debate is around running marathons, the link is entirely logical.

Finally, just worth noting that the latest debate is NOT about marathons - it is about running, because all 3 deaths in Detroit were during a half-marathon.

Races are part of running - you do the training, you do the races. The link allows one to bridge the gap between health benefits of physical activity and racing.

Thanks for the thoughts!

Unknown said...

For those considering the statistical approach, the stats will probably show that more people die while sleeping than die while running, so the message is to stay away from your bed and rather go for a run!

My feeling is that if you have an underlying, undetected medical condition, the chances of you picking it up and being able to do something about it before it kills you are far greater if you run or exercise than if you sit on your couch. The reason being that you are more in tune with your body and are more likely to realise that something is wrong when you suddenly can't run as far or as fast or do as many repetitions as you usually can. If you sit on your couch all day with the same underlying condition, you will be blissfully unaware that anything is wrong until the slightest extra exertion kills you.

Another point. For those of you who watch the medical series "House", Dr. House is constantly telling people that the best place to have a heart attack is in a hospital. There is a similar scenario here. It unfortunately didn't help the three in Detroit, but where better to have a medical emergency than at a major athletic event where trained medical personnel are a lot closer to you than they ever are if you have an incident at home?

Finally, the media coverage / "hysteria" is not necessarily a bad thing. As with aeroplane crashes and the shark attacks you mentioned, we will be in a far worse situation if incidents are so common that they cease to be newsworthy.

adventurelisa said...

Ross, I totally agree. Media 'look' for a story, creating one if necessary. And the more drama you can create an fear you can strike into the audience's hearts, the better. Makes it seem tough and difficult and like the most amazing thing to do in the World. Just listen to local talk radio during the day; I often turn it off because it is nothing than overinflated drama and snot and tears stories. Snivel, snivel, snivel.
In South Africa we run back to back marathons and you can do one (or two, or three) every weekend and so they are, within our running community, not really considered a life changing accomplishment. In the US, on the other hand, is it a big deal. "Ohhh, you're a mar-a-thon runner," people say with amazement.
And if you want people to watch it on telly - up the hype, of course. I didnt watch the race so I can't comment on the coverage but I can assume that it is far easier and more convenient to interview the doctor multiple times and use words like 'death' and 'risk' thanit is to create a story about the racing and splits and records.

Yesterday there was a guy on local radio. He is going abroad to do a multiday race. On the radio he said how tough and difficult it would be and run at high altitude of 5000m - 6000m. I spoke to him on the phone on Monday and confrmed max altitude to be 3600m; he had bought Dimox! Also said it was run in the shadow of 4 major mountains; it is only nearish one, which is at least a hundred kilometres (probably quite a bit more) from the other three, which are fr, far away in the distance. We had about 60 runners of all ages, shapes and abilities when I did the race in 2007. All but one completed all 5 stages. Sure, there are ups and downs, but the race is one of the most pleasant and most beautiful that I've done. Super weather for running, quality terrain, decent accommodation and all meals provided. But, it wouldn't have sounded as dramatic on radio, would it?

Anonymous said...

After entertaining a bunch of media interviews about the safety or marathoning I was fairly disappointed by various interviewers who would not let the point drop regarding the relative safety….

The CDC has lots of statistics on annual causes of death… a few being:

Cigarette smoking ~350,000
Second hand smoke deaths ~38,000
Killed in car accidents 42,116
Killed by the common flu 20,000
Killed by murders 15,517

Basic “non-headline” stuff…

After digging deeper I found that 0.6% of all heart attacks are induced during sex and account for about 252 deaths per year in the US and the odds for dying during a marathon are something on the order of 0.000012%

I think I am ok with marathon running.... (sex too for that matter) ….though I
I would never belittle death under any circumstance

Mark B said...

Follow the link for a great blog on visualizing media distortion/risk assessment/just about everything!:

Unfortunately nothing sport related, but the HPV vaccine analysis is similar to quantifying the risk of running. It reminds me that I went to a talk by Dave Bedford once, who commented that deaths in the London marathon seem common, but every year dozens of spectators die at football matches and the like, but this goes unreported in the national press. Fact is, if you put a million people in the same place at the same time, one or two will become a medical emergency for reasons often completely unrelated to the event in question.

PreshawPete said...

One serious issue tends to get overlooked and that is OVERHYDRATION leading to WATER TOXICITY. In recent times there has developed a fetish for ensuring that athletes don't get dehydrated. The result is 'water, water everywhere' and this can be DANGEROUS. Overhydration can lead to brain swelling, confusion and if not diagnosed and treated in time, convulsions, coma and death. It is diagnosed on history, physical exam and confirmed by finding a low level of sodium in the blood (hyponatremia). I used to work at the hospital at the end of the London Marathon and have personal experience of dealing with this.

Anonymous said...

Why would you even listen to FOX News?

Unknown said...

The best estimate that I've seen of the death rate during marathons is 0.8 per 100,000 finishers, or 1.85 per million hours of marathon running (Redelmeier and Greenwald 2007). In 2008 in the U.S., there were 1.27 deaths per 100 million vehicle miles traveled. Assuming single passenger vehicles and an average speed of 40mph, that translates into about 0.5 per deaths million hours of driving. This makes running a marathon about as dangerous as driving for 15 hours. Not so bad.

Marco said...

Excelente página. Realmente muy interesante y constructivo el enfoque dado al deporte desde la ciencia.

Saludos de Buenos Aires.

Mike said...

Check this article out in the NY Times, maybe brought on by all the hype.
I tend to agree that just finishing a marathon is not where it is at. But I think that in SA we still have decent cut-off times for marathons of about 5hrs - 5h30. So it actually still means something.
What's funny in SA is that if you state you run marathons people always ask, oh are you doing Comrades, like running 42.2km is for wussies! Perhaps the yanks need not celebrate their mediocrity!

Amby Burfoot said...

Ross: Was good to see you in Chicago. Hope the rest of your trip is going well. I'm not sure I agree with all you say here, but not totally clear on what you're saying either. Seems clear that exercise raises death risk DURING exercise by 5-7x, and this risk during exercise is greatest for those who are infrequent exercisers. Thus marathons increase risks, especially for low, infrequent trainers. Of course, overall mortality is lower among exercisers, and that's the important stat. I have looked long and hard for honest comparative data, and it's hard to come by, as we measure exercise by precise hours and most other studies are "annual" or "miles driven" or some such. But pingpongpaul did a nice job above, and his numbers look good to me: they put exercise death risk at 3.7x driving car per hour.

Anonymous said...

Per Amby's suggestion the issue of hours is answerable....though someone would have to dig in to all the events to do it...

If you had (a) the number of participants racing in a year, multiplied by how long each participant took to complete the course, you would have (c) total person hours of running for marathon running....

Once you have this number "all you need" is the number of fatal events and with any basic statistical software you can calculate the odds ratio for mortality. It's basic epidemiology.

Using a very broad and simplistic example, if you have 500,000 marathon runners who complete a marathon in 4 hours you have 2,000,000 hours of "marathon running exposure.” Of course, this example is inaccurate b/c of the obvious running times involved.

Theoretically, however, it is possible and if someone had the time, money, patience, etc. they could collect race stats from a variety of marathons, say, for the last 5 years, and come up with the answer...

Then, at this point you could compare...

I wonder if Nike or some other group would be willing to provide a grant for such an undertaking? I’m inclined to explore it but don’t have the inclination to put that much “sweat equity” into the project.

Amby Burfoot said...

Conrad: The studies you're proposing have been done by Bill Roberts et al from the Twin Cities Marathon and Dan Pedoe at London Marathon, and others. Their studies line up fairly well at about 1 death per 90,000 marathon runners. And they're counting dead bodies, a fairly accurate method, if you ask me. Redelmeier counted newspaper reports, so he missed some, and came out somewhere around 1 per 125,000. Interestingly there's little or no research at shorter distances, though Roberts is beginning to accumulate half-marathon data. Way back when--1970s--Paul Thompson found one death per 400,000 hours jogging in Rhode Island. Divide by 4+ hours/per average marathon, and you come out roughly the same as the marathon studies. What I've found more difficult to obtain is equivalent "per hour" death rates from other activities, ie, sleeping, driving, bicycling, swimming, sex, etc. They're almost always reported with different denominators than /hour.

Anonymous said...

Amby. Probably the major problem is examining "time spent" in an activity vs. the frequency of the event.

For example, someone may spend X hours in car or X hours sitting....which, by the way, is evolving as a powerful risk factor in and of itself, X hours sleeping, or X hours exercising. For these factors, the NHANES data base has a lot information and there are papers attesting to some of these risks.

Frequency events are potentially more tenuous. Most of the epi- databases I know record events like "sex" as event (i.e, X times per week) rather than as time.... Frequency is also in many data bases for exercise. For example, rather than "how many hours per week," the recording is "how many times per week." So, it is a tough call.

Thanks for the great information on marathon stats, though.... good stuff

Ross Tucker and Jonathan Dugas said...

Hi All

Thanks so much for the great discussion. I'm sorry I've been uninvolved - I've been tied into a symposium at the US Olympic Centre since yesterday and just haven't had the time to do justice to your comments! What I'm going to do is a follow-up post where I take some of your great stats and comments and put them out for others to see - so thanks and I have read them!

Just one to respond to, and that's Amby's:

I think maybe the misunderstanding is that this post was not a full essay on sudden death and so forth - that has been done before and I wrote much more on that back when I tried to summarize the whole feeling after Ryan Shay's death - that's the link I posted in this post. I wrote four or five articles at the time where your points from this comment are definitely agreed with! So I think we're very much in agreement! At the time that I wrote this post, it crossed my mind that I should explain what you've raised, and copy and paste from this historical posts, but I decided that what was more interesting for now (to me, anyway) was the media coverage and I wanted to point out that there's more to it.

So this post was more a call to runners to not be swayed by the media's habit of over-hyping small probabilities. I'm sure there is a name for this, but I called it the "shark attack" phenomenon, or at least drawn parallels. So my intention was only to say that there is more to this issue than the simple "marathons can kill you" line that we're getting from most media. I believe that the probability of death during running is incredibly small, and the purpose was to emphasize this.

And here, I extend it to all running, including training and shorter distance runs, because the media has done the same when they issue warnings. The big danger is that your average runner who does NOT run marathons says "it's dangerous, I shouldn't run". And what is interesting is that even during marathons, very few athlete die after the finish of the marathon. They normally collapse early on, so it's not marathons, but 8km, 16km, 25km, general running, that becomes seen as dangerous. In the case of Detroit, it's 21km, or less. 2 hours of running, perhaps.

So this post was to point out, as you've said, that the risk is small, that those who run are healthy, and that anyone who gets anxious about the safety of running (regardless of their level) should be assured by the fact that:
a) They are seeing the result of a media error (shark attack)
b) They are safer overall than maybe they realise!

The great readers of this site and provided a lot of great info to compare it to, so I'll have to post on it!

Thanks for the comments!

P.S. To Amby: Yes, great seeing you in Chicago, I'll let you know how I'm going in New York, for a follow up!


Anonymous said...

I'm from Chicago and a little known story about the death in 2007 was it could have been avoided.

The marathon hired a bunch of medical and ambulance teams stationed throughout the course. The team that assisted the fallen runner was from a suburb of Chicago and not familar with the city. They got lost taking him to the hospital. What should have taken no more than 10 minutes took almost 45 minutes.

It was speculated and many doctors agreed that had the fallen runner made it to the hospital 35 minutes earlier he might have lived.

The story of the 2007 death at the Chicago marathon is more about administration and preparation of hired medical teams than about the safety of running.

John said...

I would like to see data looking at diet as a variable. NBC's "running is bad" pre-event motive is pretty transparent.

From what I read, anyone who eats an animal-based diet, particularly one that is high in fat and sugar, has cardiac disease to a greater or lesser degree.

butterflyblue said...

Have we covered the stress one may feel before doing such an event, pushing an underlying condition over the threshold point? Nerves, higher blood pressure, all of the anticipation for a race? We can be physically ready for an event, but it is a pretty safe bet that our HR's are higher than normal while waiting to start, blood pressure may be higher, cortisol and other hormones spiking.

It does seem interesting that for the most part, I have heard of more heart problems when someone is running than when doing the first two events in a triathlon. This year at IMC an athlete had a heart attack during the swim and passed on. More often I hear of this happening during the run. Is there a connection to how the body deals with stress when you are upright, vs, swimming or on a bike?

To add, we are taught to stick with routine when preparing for an event, follow the same plan you do as when training: pre race meal, rest, etc. It could be possible that some athletes drift from routine. It also brings to mind that most of us are sleep deprived going into the event as the excitement and nerves from the night before keep may keep one awake most of the night.

All of these things might be worth considering (one or as a whole) as they can be stressors on the system in addition to the one we are undertaking.

Donny said...

For what it's worth I'll add a comment from the other side. I was a competitive swimmer and then national water polo player into my late 30's, and then went on to running as a means of fitness. I am the opposite of a couch potato. I ran my second Comrades Marathon in June 2002 at the age of 47 (first was in 2000). Six weeks later I ran a 21k and experienced chest pains. On the 9th September I had had an angiogram and two stents fitted. After the procedure I mentioned to a 67 year old doctor (who had undergone the same procedures) how shocked I had been to have this happen to me rather than my overweight smoking colleagues. His reply was that this was genetics at work and that my colleagues won't get that early warning as they were likely to have their first heart attack in front of the TV set. I have subsequently run my best 15k and swum my best Midmar Mile times. I ran the Comrades again this year and had an unforgettable day. By continuing to exercise the way I do I believe I will get another early warning should things go wrong again. Until that day comes, long live the euphoria that comes with exercise.

r4 firmware said...

Thanx for the valuable information. This was just the thing I was looking for, If done scientifically... and properly I think exercise is always safe...
keep posting. Will be visiting back soon.

middle aged mid packer said...

I could be mis-remembering, but wasn't one of the main points of "Bowling for Columbine" that the US media is very geared up to producing constant scare stories?

That's part of the explanation right there. I'm from the UK (and on top of that, tend to watch the fairly unemotional Euronews channel) and can confirm you do really notice this on arrival in the US. (The irony is that in the UK the US has the image of being all gun-toting muggers and nutjobs, when actually a lot of it is very very safe and quiet indeed!).

Anonymous said...

Would anyone care to comment on the following article I just read? The basic idea is that a recent study showed that marathon runners showed far more evidence of subclinical heart disease than controls:


Amby Burfoot said...

Ross: Was good to see you in Chicago. Hope the rest of your trip is going well. I'm not sure I agree with all you say here, but not totally clear on what you're saying either. Seems clear that exercise raises death risk DURING exercise by 5-7x, and this risk during exercise is greatest for those who are infrequent exercisers. Thus marathons increase risks, especially for low, infrequent trainers. Of course, overall mortality is lower among exercisers, and that's the important stat. I have looked long and hard for honest comparative data, and it's hard to come by, as we measure exercise by precise hours and most other studies are "annual" or "miles driven" or some such. But pingpongpaul did a nice job above, and his numbers look good to me: they put exercise death risk at 3.7x driving car per hour.