That the concept was brought up in discussions. I think that's why they suggest it's a correlation and lacks sufficient data to be termed a causation. OTOH I further suspect that the size of the study (60,000 women) probably waters down the likelihood of outside factors significantly skewing the results. Once again, there may be some effect, but since it's highly unlikely that all 60,000 participants shared exactly the same outside factors, it limited their impact on the eventual outcome.

Another idea is the opposite...that's it is contrary to your idea: namely that many of the so-called health nuts I know (think marathon or triathlon athletes) are avid diet Coke drinkers. If your assumption was consistent they wouldn't drink diet soda or would have a net positive skew on the rates of cardiovascular health.