Here's an interesting semi-relevant experiment apparently showing that insulin sensitivity is circadian, irrespective of blood glucose levels (at least in the short term).
http://www.pubmedcentral.nih.gov/pic...p;blobtype=pdf
My concern was that the production/conversion of .6-.7 (gram?) glucose from each gram of protein (I remember protein efficiency was .6 or so, I think carbs were 1:1) might raise glucose enough to "satisfy" the debt caused by an overnight fast, thereby decreasing the morning sensitivity. Apparently that is not the case (thank goodness).