Thoughts about the epidemiology of exercise volume and health effects

Considerable data exist that  allow reasonable estimates of the volume of exercise required to : 1) decrease cardiovascular disease (CVD) risk 2) decrease the risk of heart failure.

There are considerable coarse grain epidemiologic data that support the notion that regular aerobic exercise will decrease heart attack risk, reduce heart failure risk and reduce all-cause mortality.

The 2018 Physical activity guidelines for Americans recommended at a minimum of  approximately 8 metabolic-equivalent hours per week ( 8 met-hours or 500 met minutes per week.)

The recommendations translate to 150 minutes of moderate exercise (less than 7 Mets) or 75 minutes of vigorous  exercise per week.

 Arem et al ( 1) pooled data from 6 studies (661,137 men and women) and  demonstrated a 20% reduction in mortality among those who exercised at the recommended level with an HR of 0.8 (CI 0.78-0.82) and a 37% lower risk for those who exercised at 2-3 times the minimum level and the maximal benefit at 3-5 times the minimum (  0.61, CI 0.59-0,62)

Further no excess risk was evident even at exercise levels of 10 times the minimum.Here more did seem to be better and no definitive upper limit was apparent.

Wen et al (2) published all cause mortality  data similar  to Arem and  also found that the maximum  CVD mortality risk reduction (45%) ( HR 0.55 CI 0.46-0.66)  at an exercise level of 523 minutes per week, a level roughly 3.5  times the minimum recommendation.


These 2 articles are part of the evidence  regarding the safety of exercise levels higher than the 2018 recommendation. Drs Carl Lavie and James O'keefe have spoken ( TED) and warned  in print about the purported hazards of long bouts of  or too many years of aerobic exercise. In 2015,  perhaps in part reaction to the Arem article and commentary by Dr. Ben Levine,  Lavie and O'Keefe toned down their emphasis on the risk of long distant running.See here for a Runner's World article concerning that issue.

The conclusion to the Arem publication is refreshingly non-ambiguous and prescriptive ,quoting
Arem:

"In regard to mortality health care professionals should encourage inactive adults to perform leisure time physical time activity and do not need to discourage adults who already participate in high-activity levels."


Coarse grain epidemiological evidence strongly support the generally accepted notion that people who exercise none or very little have the highest risk and those who exercise the most have the lowest risk.

In a nutshell-the large coarse grain epidemiologic studies demonstrated that relativity low levels of exercise are better than none and further these studies were not able to demonstrate a upper level of exercise that is more risky than no exercise  though some data indicted that risk reduction at the highest level studied may have reached a plateau.

It is possible that there could  be some harmful or potentially harmful cardiac effects in a small number of long time and/or high volume endurance athletes that would not be detected by large data analysis such as that of Arem and Wen.

 That appears to be the case.

There are  three  conditions that  appear to be more common in high level exercisers. These are 1) atrial fibrillation 2) myocardial fibrosis and 3)coronary calcification.

While the current consensus view is that the relationship between exercise level and all-cause and cardiovascular mortality is curvilinear and not U shaped, the same cannot be said in regard to atrial fibrillation (AF) .

In a  2018 review (4) of the "extreme exercise hypothesis", which states that there is some level of exercise   that results in a harmful effects. Three conditions have been considered in that regard: 1) atrial fibrillation (AF), 2) myocardial fibrosis (detected by MR gadolenium scaning) and 3) acceleration of coronary artery calcification (CAC).


Eijsvogels,Thompson and Franklin (4) concluded in regard to AF :

 "that the relationship between physical activity and incident AF is best summarized by a reverse J shaped curve.Light to moderate  amounts of exercise decrease but large volumes of exercise potentially increasing the risk of AF."

Two of the studies mentioned are their review are  the Henry Ford study and Anderson's cross country ski study .

Even though a study from the Henry Ford Exercise testing project (5) demonstrated that higher cardiorespiratory fitness was associated with a graded reduction in AF (the higher the fitness level the lower the risk of AF),a large cohort study of cross-country skiers showed that those who finished more races and those who raced faster had higher risk of AF than those who did only one race and those who raced at a slower pace. In that study by Anderson  (6) of  52,755 long distance skiers those who finished five or more races versus those who finished only one race had a Hazard ratio for AF  of 1.29 95% CI 1.04-1.61.

The Henry Ford study looked at the relationship between fitness and incident AF and the Anderson study looked at volume of exercise and intensity of exercise and while  fitness and exercise level are correlated they are not the same. After the entry into the study when fitness level was determined there are no data on  the exercise history of the participants at Henry Ford.The Anderson study used number of races and speed of racing to give some broad measure of amount exercise which relates to the questions of exercise volume versus AF risk which is not directly addressed by the Henry Ford study.


The large mega data studies,such as those by Arem and by Wen , lack the statistical power to detect any mortality effect that might occur from AF,CAC and myocardial fibrosis in endurance athletes whose exercise volumes fall at the  extreme right end of the volume axis since there are relatively few subjects exercising at mega volumes.

Even though the relationship between exercise volume and/or intensity and CVD and all cause mortality is not U shaped, the relationship between exercise volume and AF is. 

The issue of the relationship between exercise volume and coronary calcification is less straight foward. Defina's(7) study presents data that high levels of exercise  (3,000 Met min  per week which is about 6 times the minimal 2018 recommendation) are associated with prevalent coronary artery calcification but not  increase in cardiovascular mortality. This apparent paradox might be resolved if it turns out that the increased in runner's calcium scores represent density and not extent of calcium and that the increased density is related to heavy calcified plaques which are made less vulnerable to rupture. 

1)  Arem H   et al. Leisure time activity and mortality: a detailed pooled analysis of the dose-response relationship. JAMA Int Med 2015 Jun 175 (6) 959  (full text is on line)

2)Wen,CP Minimum amount of physical activity for reduced mortality and extended life expectancy .A prospective cohort study. Lancet 2011,378 , 144.

3)Franklin, BA Exercise related acute cardiovascular events and potentially deleterious adaptations
following long term exercise training. Placing the risks into perspective-An update A scientific statement from the American Heart Association. Circ 2020 Feb 26 PMID 32100573

4) Eijsvogels, TMH et al The "extreme exercise hypothesis":Recent findings and cardiovascular health implications." Curr Treat Options cardio med 2018 20  84 

5)Querishi,WT Cardiorespiratory fitness and risk of incident atrial fibrillation:results from the Henry Ford Exercise (FIT) project Circ 2015: 131 ;1827

6) Anderson K et al Risk of arrhythmias in 52,. 755 long distance cross country skiers: a cohort study
 Eur Heart J 2013 Dec 34 (47)36

7) Defina,LF Association of all-cause mortality and cardiovascular mortality with high levels of physical activity and concurrent coronary artery calcification . JAMA Cardiolog 2019, 42 (2) p 174

addendum 7/27/2020 Several changes made in the et to clarify meaning .8/14/20 and again minor changes to clarify meaning. also on 11/19/2020.3/21/21 Minor editing and typo corrections. 5/26/21 still more typos corrected.9/27/21 Comments regarding Defina's data added .