Normal blood glucose as defined by sampling does indeed show that people spike to 110,120,130 and even 140 and beyond. The problem that I have with these samples is that the population being sampled are not that healthy. For example, it is estimated that around half the population is prediabetic and that most people will need treatment for one of the vascular diseases as they age. We also know that high percentage of that population is also overweight and that many are obese. So how healthy are those numbers?
In my opinion, they are not. That's one reason I accept Dr. Bernstien's definition of 83 mg/dl (for pragmatic reasons I just target under 100, but my average is in the mid- 80s).
The second reason I accept his solution is the biochemistry of ketosis. Ketosis is a state where the body has adapted to burning fat rather than glucose. And it is that state that minimizes both circulating glucose levels and insulin levels. While I have not seen a study of blood glucose levels in people in ketosis, I strongly suspect that their levels would be close to Dr. Bernstine's. Note that glucose levels do not spike while maintaining ketosis and if they do it will kick the body out of ketosis. I know that when I am in ketosis (almost all the time) my levels are close to Dr. Bernstien's.
Third since I have 'diabetic' complications, I not only want to prevent future complications, I also want to reverse them to the extent possible. So far after almost of two years being on a very low carb diet I have seen improvement, so I plan to continue to target his definition of normoglycemia for glucose.
Fourth, is the realization that it is increased levels of insulin - likely caused by excessive consumption of carbohydrates - that cause vascular damage long before type 2 diabetes becomes diagnosed. In other words increased the risk of heart disease is not a direct risk of type 2, it is the metabolic disease that is at the root of both. It is this increase in consumption of carbohydrates that explains the obesity, heart disease, diabetes, and other manifestations of metabolic disease when native people started consuming white flour and sugar (see Taubes for a description of the diseases of civilization and his carbohydrate-insulin hypothesis).
So I do not think that the general observational studies are the way to define normoglycemia, and I accept Dr. Bernstien's definition.