The big debates of the past decade: 5) Is running good for your health?

Is running good for your health? Once the answer seemed simple.  However, the sadly premature deaths of several charismatic advocates of endurance running, Jim Fixx, John (Hadd) Walsh and Caballero Blanco, have provided grounds for questioning the claim that a large volume of running is healthy.

The past decade has seen a vigorous debate driven by enthusiasts who claim that the evidence indicates that any more than about 2.5 hours of moderate intensity exercise per week is harmful.   The high priest preaching warnings about the risks is James O’Keefe, a cardiologist from Kansas City, who summarised his views in a review in the respected Mayo Clinic Proceedings and also popularised them in an evangelistic TED talk.   I have discussed O’Keefe’s views previously in a post in Jan 2013. Though I consider that he himself comes across as something of an evangelist, the evidence that he assembles does indeed confirm beyond reasonable doubt that endurance athletes are not immune from either coronary heart disease or from potentially fatal disturbances of cardiac rhythm.

My own belief is that the overall balance between health benefit and harm of running is inclined towards greater benefit than harm, at least up to upper limit of running done by the majority of recreational runners. I reviewed the evidence in my post in Jan 2013, and will not present the details again here. Instead, I will provide an update on the evidence that has been assembled since Jan 2013.

However, my main interest is not on the statistical evidence for or against harm, but rather on what research of the past decade has revealed about likely mechanisms by which running might be expected to cause harm, with the ultimate goal of developing strategies for minimising the risk of harm. Even if the balance of evidence suggests that running is of greater benefit than harm to health, the undeniable evidence that at least some runners do suffer harm suggests to me that the sensible approach is to take what steps we can to minimise the risk of unnecessary harm.

Life expectancy

The picture is moderately clear: life expectancy increases with increasing amount of exercise, but the rate of increase levels off at higher levels. For example, in the study by Wen and colleagues, the reduction in mortality rate (that is the number of deaths that occurred compared with the number expected during the follow-up period) was observed to increase with increasing amount of exercise, but levelled off at a mortality reduction of around 45%  for 50 minutes or more of vigorous exercise per day.   Although there was little evidence of an actual decrease in life expectancy with very large amounts of exercise, the number of individuals exercising at extreme levels is small, so it is difficult to draw statically robust conclusions.

Perhaps the most thought provoking evidence comes from the Copenhagen Heart Risk study, a longitudinal study of nearly 18000 people followed over a period of up to 35 years. As in other relevant studies, the main conclusion regarding the effects of exercise is that jogging reduces mortality. The age-adjusted increase in survival with jogging was 6.2 years in men and 5.6 years in women. However the investigators reported a U-shaped relationship, with best outcome in those who jogged less than 2.5 hours a week at a slow pace. But the numbers in the relevant groups were very small. In those reporting that they jogged slowly, there were 3 deaths in 178 people, whereas among the fast group there were 5 deaths among 201.   These numbers are too small to justify robust conclusions.

In contrast , a study of mortality in a cohort of 49 219 men and 24 403 women who participated in any of the Vasaloppet long-distance ski races in Sweden (90Km for men; 90 and 30Km for women) during 1989–1998,revealed substantially reduced standardised mortality rate (SMR) from all major causes of death including heart disease and cancer. Overall, 410 deaths occurred up to Dec 1999, compared with 851 expected, yielding an SMR of 0.48.   It is reasonable to assume that the majority of competitors in these races had undertaken extensive training.

More recently, a study by Dr Jodi Zilinki and colleagues at Massachusetts General Hospital, reported at this year’s American College of Cardiology meeting in Washington, found evidence of decreased cardiac risk factors following marathon training in 45 recreational runners aged 35-65 who had not achieved Boston Qualifying time for the 2013 event but had nonetheless obtained places to raise money for charity. Half had run at least three marathons in their lifetime. Prior to an 18 week training program, over half had at least one risk factor for heart disease, such as high cholesterol, high blood pressure or a family history of heart disease. During training, potentially harmful low density lipoprotein (LDL) cholesterol decreased by 5% while triglycerides decreased by 15%. Thus for the majority of recreational runners, it appears that marathon training is likely to be good for their health

Specific cardiovascular risks


With regard to cardiovascular risks, the most compelling evidence indicates increased risk of disturbance of heart rhythm, such as atrial fibrillation, in middle-aged male endurance athletes. This is revealed by many studies (which I reviewed in a post in Jan 2012), though the question of whether or not this results in greater mortality is not clearly established. It is possible that other health benefits of running might outweigh the risks associated with arrhythmia. For example, the reduced blood pressure, lower levels of harmful low density lipoprotein cholesterol associated with endurance training would be expected to confer protection.

Coronary artery disease

Perhaps the most worrying issue is the possibility of increased coronary artery disease.   There have been a many of reports of clogged coronary arteries revealed by coronary angiography in endurance athletes. The most substantial of these was a study of 50 men who had competed in the Twin Cities (Minneapolis-St Paul) Marathon for twenty five consecutive years. The proportion of individuals with atheroma was similar in the two groups (atheroma in 60% on marathon runners, and 52% percent of controls), while the extent of the atheroma was significantly greater in the runners. This finding was initially reported at the American College of Cardiology meeting in Atlanta, Georgia in 2010. However, only an abstract was published at that time.

The definitive publication by Robert Schwartz and colleagues appeared in 2014 in a relatively obscure journal, Missouri Medicine. The list of co-authors includes James O’Keefe, who is an Editorial Board Member of Missouri Medicine Preventive Medicine.  I mention the details about publication because when a paper as potentially important as this appears belatedly in a relatively low profile journal, the first question that occurs to me is: why was this not published in a more authoritative journal.   Was it was rejected by other journals?

I am not a cardiologist, though I do frequently provide peer review of scientific and medical manuscripts submitted to high-profile journals. If I had been asked to review this manuscript, my greatest concern would have been for the procedure used to match runners and controls. In any study in which participants of interest are compared with a control group it is important that the two groups are matched on factors that might have an independent effect on outcome. For example, it is usually necessary to match for age, sex and social class. It is often desirable to match for other factors relevant to the condition of interest, such as family history of heart disease. However it is usually undesirable to match for factors that reflect the mechanism by which the ‘treatment’ (in this case, running multiple marathons) might achieve its benefit or harm. For example, because exercise tends to decrease the levels of harmful lipids in the blood, it might be misleading to match the groups on these variables. Such matching would be expected to produce a control group who happen to exhibit below-average risk of cardiovascular disease for incidental reasons related to their genes or environment.   But this is precisely what Schwartz and colleagues did. In effect, they determined the effect of marathon running after allowing for some of the anticipated benefits of marathon running.  Thus the statistical comparison of the two groups is biased and must be interpreted with extreme caution.

Nonetheless, whatever one makes of the statistical comparison, the number of marathon runners with atheroma in their coronary arteries and the extent of their atheroma was alarmingly high.  If I had been the referee reviewing this manuscript, I would have recommended publication subject to a critical discussion of the possible bias introduced by the matching procedure. I think it would be unwise to brush aside these findings simply because of some flaws in the science – there are very few medical studies that are totally free of all possible bias.  In evaluating medical evidence it is necessary to weigh up the whole picture, including the plausibility of the findings on the basis of what we know of human physiology.

Plausible mechanism for cardiovascular damage

I consider that there are at least three plausible mechanisms by which prolonged or intense endurance exercise might lead to cardiovascular damage. These are elevation of cortisol, chronic inflammation and acute release of potassium due to breakdown of muscle cell membranes.

Cortisol is a key hormone that mediates the body’s acute response to stress, but if elevation of cortisol is sustained, it damages most tissues of the body.   Skoluda and colleagues have presented evidence that endurance athletes tend to have sustained elevation of cortisol.

Inflammation is the process by which the body repairs itself following acute damage. It is probable that acute inflammation plays a central role in the repair and strengthening of the body after vigorous training. Thus, it is a key mediator of training effects. However inflammation that becomes chronic is harmful and probably plays a large part in the over-training syndrome. Furthermore, chronic inflammation promotes the formation of coronary atheroma.

The imbalance of concentration of potassium ions across cell membranes plays a central role in nerve conduction and muscle contraction, in both skeletal muscle and in the heart. Potassium concentration within cells is normally high while levels in extra-cellular tissues including blood are low. Damage to muscle cell membranes during vigorous exercise releases potassium into the extracellular tissues. The mechanism for pumping potassium back into muscle cells and also the action of the kidneys can normally cope with this tendency towards excess extra-cellular potassium. But under some rare circumstances, potassium can rise to high levels causing fatal cardiac arrest.

The important thing about all three of these mechanisms is that there are things we can adjust about our training and our lifestyle, including diet, that have the potential to ameliorate all of these risks. Thus, even though for most of us, the health risks of endurance training and racing are small and likely to be outweighed by the benefits, we can shift the balance even further in the direction of benefit by adjusting our training and lifestyle. In future posts I will present my conclusions regarding the best strategies for achieving this.

7 Responses to “The big debates of the past decade: 5) Is running good for your health?”

  1. Robert Osfield Says:

    Hi Canute,

    Do any of the studies look at what specific issues that a likely to have caused the heart attacks in endurance athletes? Like you I suspect compounding factors such as chronically elevated cortisol and inflammation. I wasn’t aware of the Potassium issue.

    I do wonder whether diet and lifestyle choices of some endurance athlete might compound the issues. Getting enough sleep and not eating healthfully will impact health in us all, but with endurance athletes you can hide the outward signs of ill health quite well – you look slim, athletic and “healthy” but inside the body can be under stress.

    I do keep wondering about just when we might be able have devices to measure our various hormone and various blood levels outside of the lab to provide some early warning that the body is being pushed outside healthy levels. Right now it’s all guess work when it comes to judging just where I might be w.r.t Cotrisol, Insulin, Potassium, Blood sugar etc.

  2. canute1 Says:

    Thanks for your comment. I am not aware of any studies that have demonstrated a direct relationship between coronary atheroma and either inflammation of cortisol in endurance athletes. However, many studies of the general population demonstrate that atherosclerisois is largely an inflammatory processes. ( e.g. ) However, the relationship between endurance exercise and chronic inflammation is complex. Acute Exercise produces acute inflammation. Moderate exercise is associated with reduced chronic inflammation but overtraining is associated with increased inflammation.

    It would be expected that relationships would be complex because moderate exercise tends to generate beneficial homeostatic responses, while extreme exercise would be expected to produce opposite effects. Hence, a direct answer to the questions you pose would require systematic measurements of hormones, inflammatory markers and markers of cardiac damage over a long a period of quantified heavy training. So at this stage the evidence is indirect.

    While it is tempting to think that monitoring relevant stress hormones and inflammatory markers would be helpful for adjusting training load, there is a problem with individual variation in the relationship between the level of such adverse markers and the extent of damage to relevant body tissues. On the other hand, measurements of stress such as HR, HRV and even mood state, might be a more sensitive measure of the actual tissue damage sustained, but such measurements would also be expected to be subject to idiosyncratic variation.

    My present approach is to perform an array of measures of stress and performance and draw conclusions from the overall picture. This amounts to ‘listening to the body’ in a structured way. However such an approach would only be practical if one could devise a reliable procedure for a hierarchy of tests starting with a few simple screening observations followed by further testing when the initial screen indicated abnormality.

  3. Ewen Says:

    Canute, thanks for an intersting post. Looking forward to your next on the strategies to employ to minimise the risk of endurance training/racing.

    I’m wondering if the training/racing I like doing is helping to minimise risk – usually daily running at MAF effort (“goldilocks training – not too easy, not too hard, just right) for 40 minutes to an hour; most usually short races (5k) – perhaps less cortisol is produced and inflammation caused? Very few marathon-like races requiring many days to recover from or races that produce undue muscle soreness (an indication of damage).

    • canute1 Says:

      Ewen, Thanks for your comment.
      I think that your training and racing is sensible. Goldilocks training paces are probably best, though it is possible that very high volume training might be as damaging as HIIT. Perhaps good recovery is even more important than the type of training. I understand that you also aim to eat healthily. I am not sure about your job – unfortunately, most people have little control over the demands of their job, so I think that adopting practices that promote relaxation and recovery from work stress may as important as recovery from training stress. Sleep is clearly important. I am intrigued by the possible value of psychological strategies such as mindfulness, which is similar to the detachment of Eastern meditative practices, but with a rationale that is more consistent with neuroscience.

  4. Seth Leon Says:

    Hi Canute,

    Thanks for your excellent review. Your readers might also be interested in Alex Hucthinsons excellent breakdown of the available evidence which can be found here:

    I also agree that integrating a western and eastern approach may be most beneficial. The idea of mindfully balancing activity with recovery is entirely consistent with an eastern approach. I think the focus on optimal recovery is the key.

    Thanks again 🙂

  5. canute1 Says:


    Thanks for your comment and for posting the link to Alex Hutchinson’s excellent review of the topic.

    In my planned post on cortisol I will discuss recovery in more detail

  6. Does training induce long term muscle damage? | Canute's Efficient Running Site Says:

    […] of the evidence that heavy training can produce a long lasting over-training syndrome, and also the controversial evidence regarding the reversed-J shaped relationship between training load and health outcomes, such that […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: