Thanks, heatherisone, for this link, it was helpful. There is a charge of $20 to read this 25-page medical article at leisure. A five minute preview is available free of charge but you have to be quick. I skipped directly to the charts. One graph shows the level of "plasma acetone" (ketones) in the blood in a "starvation" state (considered as such in those days, actually being only about eight days into a fast) leveling off at 6 mmol/L. Converting this to mg/dL by the factor of 18 gives 108 mg/dL, near the top of my reagent strip and more than my own measured level of 40 mg/dL in urine. So this seems to answer my question if 40 mg/dL is too high. It is not too high, assuming that urine levels equal blood levels. There could be some dilution. I did not have time to find out if ketone levels drop before actual starvation begins but I doubt if the studies would have pushed the subjects that far.