Is that a problem for vegans ?
I'm surprised I can bleed at all given the amount of greens I eat these days.
I was more just thinking that if you have a deficiency of vitamin K*, then your blood clotting will not be working so well, a major clotting chemical pathway is entirely dependent on vitamin K.
The deficiency will make you ill in other ways, but your risk of blood clots (which is closely tied to indicators collated on this trial) will be lower.
I think vitamin K deficiency would be less likely in veg<>ns than meat eaters (some of whom may barely ever touch a vegetable), and their levels would certainly be higher in general (I've been a bit crap with the veggies recently, and the amount of warfarin** I need to take to reach the optimal level for my personal health conditions has markedly dropped).
So one area of speculation might be that enough meat eaters are just slightly deficient enough in vitamin K to have increased their general clotting risk to a level above the evolutionarily optimal*** level, which is probably a little on the 'sticky' side for modern life (I go around with a clotting rate about a third of a normal person and haven't noticeably suffered from it when having the odd bump and scrape), and hence have the slightly higher clotting risk.
On the other hand, the veg<>ns are likely to be benefitting from the vitamin in ways that more than offset this effect and are not measured by this particular study.****
And might also benefit in the case of a serious accident.
* - which is in a lot of things but green leafy veg is really packed
** - its anticoagulant effect is modulated by working as a vitamin K antagonist
*** - as well as mitigating agaist blood clots, the rate of clotting has a major effect on risk of wound infection, which was a major threat during pretty much our entire evolutionary history - we should also bear in mind that strokes and ischaemic heart disease tend to occur after reproductive age, and so the negative selection effect would be limited
**** - we should bear in mind that a study that tries to measure too many things is generally a bad one