There is little doubt that the balance of benefits and risks to the well-being of the heart arising from regular running favours the benefits. The risk of heart attack following a marathon is small but real, though it is generally accepted that this risk often arises from underlying heart abnormality. However there is one heart problem that is relatively common and might arise largely from factors related to fitness itself. That is atrial fibrillation.
Atrial fibrillation is a disturbance of heart rhythm in which the electrical events that initiate contraction do not arise in a regular manner from the sinoatrial node embedded in the muscle of the right atrium of the heart, but instead arises in a chaotic manner from a variety of different sites in the walls of the atria. Instead of a well organized contraction of the atria followed by a regular contraction of the ventricles, the walls of the atria contract in a poorly coordinated manner and blood is not transferred to the ventricles in a way that ensures efficient filling of the ventricles before ventricular contraction ejects the blood into the pulmonary artery (that connects the right ventricle to the lungs) and the aorta (that carries blood from the left ventricle to the rest of the body). Usually a person with atrial fibrillation feels fatigued and may be aware of palpitations – an irregular heart beat.
The evidence for increased risk in endurance athletes
Many studies report that atrial fibrillation is more common in endurance athletes. In a 5 year follow-up of 146 young elite athletes from a larger sample 1772 elite athletes who had been found to have heart rhythm disturbances, 13 (9%) had atrial fibrillation (Furlanello et al Atrial fibrillation in elite athletes. J Cardiovasc Electrophysiol. 1998;9(8 Suppl):S63-8.) In 5 of the these 13 young athletes, an underlying heart abnormality was found (most commonly abnormality of the electrical conduction path from atria to ventricles known as the Wolf-Parkinson-White syndrome, but in 8 athletes, no underlying cause was detectable.
In a meta-analysis that examined pooled data from 6 previously published studies comparing a total of 655 athletes with 985 age and sex matched non-athletes (with all members of both groups drawn from clinic attenders), there were 147 cases of atrial fibrillation among athletes and 116 among the non-athletes (Abdulla J, Nielsen JR Is the risk of atrial fibrillation higher in athletes than in the general population? A systematic review and meta-analysis. Europace. 2009;11(9):1156-9.). Mean age was 51+/-9 years and 93% were men. The authors concluded that the risk of atrial fibrillation is significantly higher in athletes compared with non- athletes, but that this finding should be confirmed further in large-scale prospective longitudinal studies.
In a prospective longitudinal study, individuals representative of the group of interest (eg male endurance athletes) are recruited before they are known to have any abnormality and followed over a period of years. This provides an estimate of the risk that members of that group will develop the problem over the number of years of the follow-up. Such a prospective longitudinal study, of long distance skiers, has just been published this month (Grimsmo J, Grundvold I, Maehlum S, Arnesen H. High prevalence of atrial fibrillation in long-term endurance cross-country skiers: echocardiographic findings and possible predictors–a 28-30 years follow-up study. Eur J Cardiovasc Prev Rehabil. 2010 Feb ;17(1):100-5.) Grimsmo and colleagues studied cross-country skiers with a long history of training, from three different age groups, identified on the basis of participating in the Birkebeiner ski marathon, an annual 54 Km cross-county race from Rena to Lillehammer, in Norway. 122 individuals entered the study in 1976 and of these 78 were available for the follow-up assessment in 2004-2006. 12.8% of those followed up over the period of 28-30 years had ‘lone’ atrial fibrillation that was not accounted for by any other underlying heart abnormality. The mean age of onset of the atrial fibrillation was 58 years. Although atrial fibrillation is quite common among the elderly, occurring in about 5% of men over the age of 65, 12.8% is a very high rate to find among relatively young middle aged men. There seems little doubt that male endurance athletes have a substantially increased risk of atrial fibrillation.
Is endurance training the cause of the problem?
In the study by Grimsmo and colleagues, the factors that were associated with the development of atrial fibrillation were a long interval between the p wave and the q wave in the electrocardiogram (this indicates a long conduction time between atria and ventricles); a slow heart rate; and a large left atrium. Slow heart rate and enlarged left atrium are features produced by endurance training. So it is plausible that the endurance training plays a causal role in the problem. It is well known that both excess parasympathetic activity (which is responsible for low resting heart rate) and excess of sympathetic activity can predispose to atrial fibrillation, with the parasympathetic form more common in individuals without other evidence of heart disease (Coumel P. Autonomic influences in atrial tachyarrhythmias. J Cardiovasc Electrophysiol. 1996;7(10):999-1007.) While there is a widespread belief that the increased parasympathetic activity associated with endurance training is a good thing, the link to atrial fibrillation should give us pause for thought.
It is also possible that the mechanism involves atrial scarring. In general, training effects arise as a result of transient damage to muscle that leads to compensation – this is well established for skeletal muscle and is likely to occur in heart muscle as well. If the transient damage results in scars (residual fibrous tissue) then these scars might be expected to interfere with the coordinated conduction of electrical signals through the muscle.
What are the consequences of atrial fibrillation?
In the short term, atrial fibrillation usually results in fatigue and limitation of athletic performance. The evidence for serious long term adverse effects is less straightforward. The chaotic contraction of the atria can result in clotting of blood within the atria creating a risk that the clots might break free and be transported via the large arteries and produce blockage in the smaller diameter peripheral arteries. One possible serious consequence is a stroke due to blockage of arteries in the brain. In a 9 year follow-up study between 1993 and 2002, of 30 athletes with atrial fibrillation, Hoogsteen and colleagues found that by 2002, 3 (10%) were dead; paroxysmal atrial fibrillation continued in 15 (50%) ; permanent atrial fibrillation emerged in 5 (17%) ; and 7 (23%) of them showed no further atrial fibrillation. Thus at least in this group of athletes, the outcome was worrying. However it should be noted that these were not specifically described as cases of ‘lone atrial fibrillation’ and hence in at least some cases it is possible that there were other underlying heart abnormalities.
A personal note
Not only do atrial hypertrophy and over-activity of either the parasympathetic or sympathetic divisions of the autonomic nervous system appear to increase the risk of atrial fibrillation but there is also evidence that inflammation can increase the risk. In the presence of inflammation, the body produces a protein known as C-reactive protein. A meta-analysis of the relationship between C-reactive protein and recurrence of atrial fibrillation in patients under medical treatment found evidence suggesting that baseline C-reactive protein levels predict the long-term risk of recurrence of atrial fibrillation (Korantzopoulos P, Kalantzi K, Siogas K, Goudevenos JA Long-term prognostic value of baseline C-reactive protein in predicting recurrence of atrial fibrillation after electrical cardioversion. Pacing Clin Electrophysiol. 2008 ;31(10):1272-6.).
Recently I suffered an episode of inflammatory arthritis and as a precaution I cut back on training. On 13th February, about two weeks after the worst of the joint pain, I did my regular test of aerobic fitness on the elliptical cross-trainer. In this test, I record heart rate during 2 minutes on each of a series of steps of increasing power output in the range 30 to 232 watts. In recent months, my heart rate at 232 watts is usually around 146-148 which I regard as just a little above my anaerobic threshold. I can comfortably maintain a breathing rate of 40 breaths per minute, at this heart rate but if I increase further, then my breathing rate increases rapidly up to around 80 per minute. On 13th Feb, I noted that my heart rate was typically around 5 beats per minute higher than usual on each of the lower steps and by the final step at 232 watts, it settled at 150 beats per minute. I did not feel stressed, so I simply assumed that I had lost a little fitness in the preceding three weeks of reduced training.
I examined the Poincare plot which provides an indication of function of the autonomic nervous system. The plot for the final 1 minute at power output of 232 watts is shown in figure 1.
As I have discussed in previous postings the spread of points across the 45 degree line denotes high frequency heart rate variability and is an index of parasympathetic activity, though there is some controversy about the details of the interpretation of the Poincare plot during upper-aerobic exercise. The measurement that represents the amount of spread at right angles to the 45 degree line (sd1) is 3.9ms, while the quantity presenting spread along the 45 degree line (sd2) is 3.3 ms. This is fairly typical of what I usually observe when maintaining a constant power output at the upper end of the aerobic range.
Last summer, when I developed parasympathetc overactivity as I increased training during recovery from my episode of illness, the spread of points across the line was far greater (up to 12 ms). The only minor feature of note in figure 1 is that there is one point well away from the cluster above the 45 degree line and a matching aberrant point below the line. These two points indicate that one beat was initiated about 10 milliseconds earlier than might have been expected and was followed by a corresponding lengthening of the next inter-beat interval before normal sinus rhythm resumed. This 10 ms advance is a trivial advance of the usual beat rate though it might indicate that a location outside of the sinoatrial node had captured the role of initiating that beat. However I would regard one slightly early beat among around 150 beats as utterly trivial, and therefore I had no qualms about setting off for an easy 8Km run in the lower aerobic zone. I maintained an easy pace of around 5:30 minute per Km, which would normally be associated with a heart rate of around 118 bpm. Because I did not have any concerns, and I felt quite comfortable running, I did not take much notice of my heart rate as I ran. I was therefore amazed when I examined the beat by beat variation afterwards (figure 2.).
As expected, my heart rate rose steadily over the first two minutes to about 117 bpm but then within the next 6 minutes there were 7 strikingly premature beasts, typically occurring 150-200 milliseconds early. If this is not an artifact due to monitor malfunction, it suggests that a rogue site was capturing the role of initiating the beat. Furthermore during that period the heart rate fluctuated a lot more than usual, though it tended mostly to be in the range 100-120 bpm. However, those fluctuations were nothing compared with what happened after around 9 minutes. The rhythm went haywire, and did not settle until I got home, at around 45 minutes. The heart rate then settled rapidly reaching around 75 bpm with only modest (and potentially healthy) amount of beat by beat variability, 4 minutes after I stopped running. What might this mean? My first conclusion is that my monitor had malfunctioned. I still do not have any firm evidence to rule out this possibility. The other possibility is that I experienced about 30 minutes of atrial fibrillation. However the amazing thing is that I felt quite comfortable running. Perhaps the run felt slightly more effortful than usual for a lower aerobic run, but after a few weeks of reduced training, that would not have been surprising.
I have not yet resumed full training but have done a number of 8-10Km lower aerobic runs and about 10 sessions on the elliptical cross trainer, since mid February. There has been no further sign of this crazy rhythm. Typically, when running my heart rate has been around 620-650 beats/Km, which indicates scarcely any loss of aerobic fitness since mid January. My observed heart rate during elliptical sessions has been back to normal, and my resting heart rate is around 46 bpm which I regard as healthy.
So at this stage I plan to build up my training cautiously. As the unusual rhythm occurred when running at any easy pace, I think that long easy runs during which parasympathetic activity would be expected to be higher, might actually be riskier than shorter faster runs, but nonetheless, I will be very cautious about any sessions that might be stressful.
With regard to the longer term, I will simply have to wait and see. I still have no clear evidence that it was not simply a malfunction of my monitor, and I would be interested to know if others have observed similar patterns with the Polar RS800CX. Although I do not entirely trust the monitor, I am nonetheless very glad that I have it, as it gives me confidence that if in fact I am developing atrial fibrillation, I will be able to detect it. The chance of detecting it during a planned clinical ECG recording would be very low at present. Whatever the eventual outcome, it has confirmed that running when exhibiting any signs of inflammation is probably not a good idea. On the other hand, I suspect that my life expectancy will be greater if I remain fit, so I am quite happy to keep on running while I monitor the situation. As for my plan to run a good marathon in 2012, only time will tell.