Friday, 28 December 2012

Sipping glucose vs Bolus dose - Jenkins study

I got this oldish study from Donny who comments at hyperlipid.

In this study they got together 9 healthy males and asked them to drink a glucose drink containing 50g, but they varied the timecourse over which they could drink it. In the bolus dose test, the subjects drank the whole 50g in 5 minutes, in the sipping test, they sipped the glucose drink at a constant rate of 3.5g per 15 minutes. ( so the entire drink was consumed in 3.5 hrs )

Keep in mind, in both tests, the glycemic index is the same, because the form of the glucose is the same, but what changes is the glycemic load, I.E., in the amount of glucose entering the portal vein. ( technically its called the mass flow rate )

Here is the blood glucose graph.....

Not surprisingly, the bolus dose results in significantly higher maximum blood glucose concentrations. The area under curve ( AUC ) is a measure of the total number of glucose molecules entering circulation. Much to my shock, the total AUC for both the sipping and bolus are approximately the same indicating that in both cases, a similar amount of total glucose entered circulation. Remember, it is the job of the liver to extract glucose from the portal vein and convert it to glycogen/fat. Glucose that is harvested by the liver from the portal vein obviously does not hit circulation and so does not affect blood sugar.

So with the bolus test, we can say that unharvested glucose ended up accumulating in circulation. Once the glucose is in the circulation it needs to be disposed of the in peripheral, AKA, skeletal muscle and........ adipose tissue. To do that, we need insulin.

The higher the blood glucose concentration, the more insulin we need to "shove" that glucose into the peripheral.

So you would expect the total amount of insulin ( AUC ) secreted to be higher in the bolus test, wouldnt you? And indeed, thats exactly what we find......

No need for annotations here, we all know which curve belongs to which group. Calculated values for AUC appear in the following table.....

Now here comes the punchline, despite a similar amount of total glucose entering circulation in both the sipping and bolus tests, if you do the bolus test, youll secrete anywhere from 50% to 100% more insulin, just to deal with those higher blood glucose concentrations. All this may be quite obvious to some people, but this was actually news to me because I had thought that the amount of insulin secreted was dependent on the total amount of glucose molecules entering circulation, with some modifications due to insulin resistance ofcourse.

Although come to think of it, it does fit in with my personal anecdotes. For example ( without being technical anymore ) , I have long observed with my own body that the foods which most aggressively spike blood sugar are the exact same foods which are the most fattening. Fruit and honey do not cause nearly as much weight gain in me as potatoes/rice. And the latter spikes my blood sugar significantly more than the former.

Ever increasing blood glucose concentrations require ever increasing insulin concentrations to deal with them. You just have to ask yourself, at what point do those ever increasing insulin concentrations start to "permanently"  alter the structure and morphology of your adipose tissue, in a similar way that super concentrations of insulin from insulin injections cause localized lipohypertrophy.

Is it possible to cure that localized lipohypertrophy with diet and exercise?

If we could model obesity as simply small amounts of  "lipohypertrophy" occurring all over the body, then why are we assuming that obesity can also be resolved with diet and exercise?


  1. when was the last time you drank glucose? Sucrose, or HFCS, sure, but glucose?

    1. What do you mean mate?

      This study is not about drinking glucose, the point is that Jenkins is showing that maximum blood glucose concentrations strongly determines total insulin secretion and not total calories consumed or total carb grams.

      This is slightly counter-intuitive. ( in my opinion )

      With this in mind, counting carb grams to manage insulin is misleading, what should be focused on instead is looking at how particular sources of carbs spike blood sugar, because it is the spike in blood sugar which more strongly determines insulin response.

      For example, 50g of honey vs 50g of potato. Both are pure carbohydrate, and both contain same amount of calories, but the potato would *should in theory* result in much higher insulin secretion BECAUSE potato spikes blood sugar more intensely.

      This perfectly explains why low-GI diets are not fattening like high-GI diets are. High-GI diets result in more total insulin secretion, even if total calories are equal.

      Anyway like I said, this probably wont be news to many people. Like I said, I have always observed that some sources of carbs are way more fattening than others, now I have good evidence as to why that is.

    2. It looks like Dr.Emily Dean thinks it is time to drink glucose or at least eat starch rather than honey

      "Glucose: Increases insulin, increases GLP-1, decreases ghrelin: all of which increase satiety and decrease reward seeking behavior.

      Fructose: Barely increases insulin, doesn't increase GLP-1, and doesn't decrease ghrelin, so after ingestion you will presumably still be hungry and looking for the next bag of skittles."

    3. Sigh, then Emily is under-researched, in obesity glp-1 secretion in response to carbs is severely depressed that is exactly why obese people find carbohydrates not satiating.

  2. Very interesting. What we need is insulin meters instead of glucometers / ketone meters. I have noticed that honey is less fattening than starch, for me. The most fattening carb for me is wheat, hands down, especially wheat flour. Followed by potato/rice, followed by sugar. Of course sugar has more long-term deleterious effects but I'm talking in terms of immediate weight gain/lack of weight loss.

  3. I wonder what would be delta Gluc and delta Ins for a bolus of 20 gm glucose or starch (no fructose).

    I guess Bernstein might know, as he recommends no more than 20 gm (non fructose) CHO for Type 1's, and as an endo needs to titrate the meal dose of insulin.

    1. I'd like to see an array of very basic studies - postprandial levels of glucose, insulin, glucagon (!) etc after various meals, mixed and with just one or too macro nutrients. That would be very basic but important research, but it doesn't seem to be done. It would be easy too. There's especially a lack of studies about the glucagon response to various meals in healthy people...