I think we have to be very careful with the term fasting insulin.
If we read, in a clinical paper, that fasting insulin level was X iu/ml it is perfectly reasonable to assume that this level simply reflects the carbohydrate content of the diet over the two or three days in the lead up to the blood draw. You only have to look at Grey and Kipnis' paper to see that, independent of weight change, fasting insulin can be simply dialed by adjusting the macronutrient ratio. It can be dropped from 40 microIU/ml to 10microIU/ml and cranked back up to just over 50microIU/ml, each shift occurring over a few days:
If we go to a rather better conducted study we can look at the effect of starvation on fasting insulin levels. What happens if you live on water for 5-6 weeks? Well, I guess it's obvious that body weight drops. Here are the clinical data for the eleven volunteers:
I worked out the average weights at the start and end of the study. The drop was from 135.8kg to 115.6kg, something in the region of 20kg of body weight. Obviously some of this would be glycogen, glycogen-water and muscle, but a big chunk must be fat.
What happens to fasting insulin?
Well, there are three different "fasting" insulins on this graph. The first is 45microIU/ml. This is the fasting insulin on the normal diet of an obese person. Second is about 38microIU/ml, after restriction of carbohydrate to 300g/d with caloric intake at 2500kcal/d. The third is between 20 and 14microIU/ml, achieved after three days total fasting and this level basically doesn't budge over the following 6 weeks, even though bodyweight drops by 20kg.
This later value is a great deal higher than a non obese person would have under prolonged fasting conditions and remember that the people in this study are preselected as having failed on every diet they have ever tried and they are willing to undergo the risks of a prolonged water fast. They do not appear to be hyperinsulinaemic as a consequence of their excess weight if a 20kg acute weight loss has no effect on blood insulin levels.
The blood glucose normalises within three days of the start of fasting. At this point physiology's role is to control hepatic glucose output. All tissues which use glucose via insulin should have stopped accepting glucose to spare it for the brain.
In these people the level of insulin required to do this in the region of 10 times that of a spontaneously slim person.
Obviously, if you perform a cross sectional observational study of fasting insulin vs bodyweight there will be a positive correlation between the two variables. It is a perfectly valid hypothesis to propose that obesity CAUSES hyperinsulinaemia. Equally, if you are as stuck in the rut of fasting insulin inhibiting inter-meal lipolysis as I am, it would be perfectly reasonable to hypothesise that people with the highest fasting insulin are the fattest because hyperinsulinaemia CAUSES obesity. Both are potentially valid explanations of the observation.
Who would lose weight fastest on a water fast?
Calories in, calories out... Obviously calories-in during starvation is solely supplied by lipolysis and protein breakdown, once glycogen is depleted. With a BMI of 50kg/m2 "calories-in" from fat breakdown are essentially unlimited, if they happen to be metabolically available. So weight loss should be determined by basal metabolic rate plus exercise/spontaneous movement. A fat person should have a slightly higher basal metabolic rate just to run the support tissue for moving their fat around, even if the fat itself has a relatively low metabolic rate. You must also remember that an overweight person is like me doing a squat with 60kg on my back every time they sit down and get back up again from a chair. So on both of these counts you would expect the fattest people to have highest "calories-out" and so lose weight more rapidly than less obese people.
I data trawled and carefully selected choice points from table 1, discarding the half which don't fit the line. I used the blokes only. All is forgiven Dr Keys. Plotting weight loss against starting weight gives a crude (negative) correlation for men. Let me be the first to admit that the relationship does not hold if you include the female subjects. Life would have been easier if we had been given individual starvation insulin levels, rather than having to take bodyweight as a rather crude surrogate. The three women outliers who ruin the plot are, interestingly, short stature.
Here's the plot for the men:
On a water fast the higher your starting weight (surrogate for "fed" fasting insulin, remote surrogate for "starvation" fasting insulin), the less weight you lose over 5-6 weeks.
Elevated insulin is associated with obesity BECAUSE it inhibits lipolysis.
Maybe there are other explanations. I just can't see them. None as blind as...
Of course addressing what causes elevated fasting insulin and why it doesn't normalise on prolonged fasting is a whole new ball game. People should look in to it. Carbohydrate restriction obviously gets you part way to sorting the problem. It side steps it rather than curing it. I have said this before.