What is on the title is absurd. Somebody linked to the actual paper, that isn't open, but from the abstract:
- 2 cans a week is the amount it will start to be harmful. Effect size is proportional to the extra consumption.
- It's a correlation detected over a large, but very homogenous population. There is space for all kinds of spurious results. That's reason to look for other studies, not to simply dismiss this one.
- What it actually finds is that exercise benefits do not interfere at all with harm from excessive sugar. You can have either combination of them.
- Non-sugared beverages effects (if any) were indistinguishable from noise. (IMO, that one is very surprising.)