That might well be, Steven, thanks for your input. The authors stratified the participants by pre-existing CVD, and the correlation held for the CVD positive participants. But there is no mention about T2DM, for example, so yes, your caveat holds.