I've wondered about that. In general ketones are more reactive then aldehydes and alcohols. This explains some of the increased AGE production from fructose, a ketone sugar, vs glucose, an aldehyde sugar.
I'm curious if there are repair mechanisms that ameliorate the effects...
Now here is a nice piece form wikipedia
. It appears the main source of methylglyoxal formation, according to that article at least, is GLYCOLYSIS gone wrong.
then this article:
that describes decreased rates of methylglyoxal formation in diabetics who are placed on metformin. Metformin increases insulin sensitivity and tends to mitigate some of the metabolic derailing common in glycolysis, especially in the diabetic.
Then we look here
and find that fasting is recommended as a means of decreasing methylglyoxal production, mainly due to its action on glycolysis...but fasting is a a state of ketosis!
So it looks to me that glycolysis is the main culprit in MG formation. Regarding the paper you cited it would be interesting to know the insulin sensitivity of the folks on the atkins protocol...and it would be super interesting to know EXACTLY how they implemented the protocol. Did the researchers introduce any non-glycemic load items that inhibit normal insulin sensitivity (artificial sweeteners, whey protein).
If ketosis was REALLY the issue we should see sky-high MG levels in fasting, starvation and to some degree in caloric restriction...but instead we see the opposite. This might be a good topic to email the lead author of that study and ask that question.