In this ted talk, cardiologist James O'Keefe argues that intensive running is bad for your heart. This conclusion is pretty alarming to me, and counter to what I have always thought. Is there a consensus among the medical community that too much running can be detrimental to your health? If so, under what circumstances? I can understand that extreme distance runners could be doing too much, but what about short distance runners? Do I need to be concerned about high intensity, short distance (say 2-8 miles) runs?
The consensus view is given here. So, the views of James O'Keefe are not widely accepted in the medical community. Suppose, however, that the adverse effects of strenuous long endurance exercise he argues for, will be rigorously established. Even then, there is still an issue with the diet as a confounding variable.
According to O'Keefe, the U-shaped curve is supposed to go up again somewhere around 40 minutes of running at 12 km/h, But at that level of exercise, you need to eat significantly more than people who just run for a few miles at a slower pace, as he is suggesting is much healthier. A rule of thumb is that an hour of running at a pace of 12 km/h burns 1000 Kcal, but this will also depend on the weight of the person, whether you run of flat terrain etc.
Burning 1000 Kcal means that you must be eating 1000 Kcal more, which raises two issues. The first is whether you are going to eat that extra 1000 Kcal in the form of unhealthy high energy density foods. We can address this first issue by investigating the diets. Based on the results we can then correct for e.g. an average drift toward unhealthier diets by people who burn more calories.
But there exists another problem with the diet that is then not corrected for. Suppose that the diet of most people, regardless of exercise intensity, is actually a cause of cardiovascular disease. This is a real possibility, given e.g. the results of this recent study. One can find a similar conclusion also in this much older study (note that the linked article is to a recent reprint, the original author comments here about this old research). Then it's entirely natural to find a U-shaped curve if exercise by itself is beneficial well above the limits O'Keefe is arguing for, because you'll obviously get some level of diminished returns from exercise more and more, while the adverse effects of the diet will increase with the increased calorie intake, without flattening off.