Can "excessive " exercise increase the risk of atrial fibrillation in men but somehow spare women hyper-endurance exercisers from that risk? That seems to be what the data show in a 2016 Meta analysis by Mohanty et al published in the June issue of the Journal of Cardiovascular Electrophysiology . See here.
The really fun thing about meta analyses is the frequency with which one can cite conflicting results from various studies. In another meta analysis Kwok et al found no association with increasing exercise and risk of AF in either sex. In spite of these conflicting data sets a quasi consensus seems to be that at some level "too much" exercise predisposes to atrial fibrillation.- determining the level of too much is another matter. As in many biological and medical issues it is easier to get a pretty good idea of the direction of the vector than is ascertaining the magnitude but sometimes the direction is in dispute as well.