Featured Post

Is the new professionalism and ACP's new ethics really just about following guidelines?

The Charter ( Medical Professionalism in the New Millennium.A Physician's Charter) did not deal with just the important relationship of ...

Thursday, May 29, 2014

Swedish study provides More data but not much definitive useful information regarding levels of exercise and atrial fibrillation risk

A recent article in BMJ has stirred  more comments regards the possibility that there is a "j shaped" curve in regard to the relationship between level of exercise and development of atrial fibrillation (AF).

There have been number of publications addressing this issue and to my eye there is good evidence that there is an increased incidence of AF in long time endurance athletes. The magnitude of this increased risk and what this correlates with is less clear-duration of exercise,intensity, height?, genetic profile, confounding factors, etc etc.

The BMJ article is from Sweden by Nikola Drca and is a long term followup of over 44000 men who completed exercise questionnaires and provided  in part retrospective estimates  of hours per week exercised at ages 15,30, and 50.These questionnaires were then linked with data indicating whether they had developed AF.The AF numbers are relatively hard data the historical data much less so.

Of those men who said they exercised more than 5 hours per week at age 30 there was a relative risk (RR) of 1.19 (CI 1.05-1.36) this increased risk in the greater  than 5 hour per week exercise category was only found for age 30, not at age 15 nor age 50 and the RR was higher for the  high exercise at age 30 group  who then stopped exercising  (RR 1.49).

Several points come to mind

When the number in the study group is very large, very small differences in the measured outcomes become statistically significant.  Relative risks less than 2-3 are generally not considered very convincing evidence that there may be causation.My favorite quote in this regard is from Michale Thur , epidemiologist at the American Cancer Society:,

 With epidemiology you can tell a little thing from a big thing.What's very hard to do is to tell a little thing from nothing at all.

and a RR of 1.19 is pretty little 

In trying to assess significance of RRs  from observational epidemiology studies. it is sometimes useful to consider what is the prior evidence and look at biological plausibility (which old time internists like to think of sometimes as pathophysiology or disease mechanisms.)

There are a number of studies that suggest long term endurance athletes have a increased risk of AF but  that is not what the data here suggest at all.The 30 year old heavy exercisers had  increased risk while the 50 year old exercisers did not and there was even  greater risk  in those 30 year old heavy exercisers  who quit. That does not seem like a dose response relationship, i.e more exercise more AF.

What would be the pathophysiology evoked to explain heavy exercise at 30 but not at 50 being related to increased risk of AF.

So what would be the take home advice? Don't begin heavy exercise until age 50 ? But if you are 30 and exercising a lot , don't quit.. None of that makes sense if we believe the is a j shaped curve regarding duration of exercise  and risk of AF or if there is a positive relationship between  duration of exercise and AF.I n spite of headlines emphasizing the 30 year old heavy exercisers risk I think overall the study is more reassuring to long time endurance athletes than it is concerning.The RR for the 30 year old group could easily be just statistical noise and the lack of increased risk for the others a more reliable finding.

 For a more detailed and less biased discussion of this general topic I suggest the excellent blog written by Dr. Larry Creswell.

Disclosure of conflict of interest: I have been doing long distance running for almost 40 years .(Fortunately I did not begin distance running until after age 30,so there should be no problem). So, my mind set  is to be critical of studies that purport to show a problem with too much running.

No comments: