T2 and insulin pumping strategies -- accounting for endogenous insulin

NOTE: I had originally thought to post this in the T2 pumping group, but the activity is so low there I suspect there may be some who could benefit from this discussion who would not see it. So I'm posting it in general. Apologies in advance to anyone that annoys.

Insulin-using T2s usually face an additional complication to figuring insulin dosing that most T1s do not: The presence of native insulin production from a flagging pancreas. By the time a T2 is forced on to insulin by progression of the disease, endogenous insulin production is pretty weak.

This presents little issue when figuring and delivering a "spike" bolus -- the same IC ratio calculations and dosing that work for a T1 work well for the vast majority of insulin-using T2s. This is because it basically replaces pancreatic function, and the pancreas just doesn't put out to cover those carbs, due to the insulin it senses already in the bloodstream, especially if the bolus worked well and kept BG under control.

However, this same phenomena messes with longer duration bolus strategies, most often used for TAG. If you're low-carbing, you'll need to account for TAG in administering insulin to keep BG in line.

Problem is, when you program an extended bolus over several hours based on an IC ratio for carb-equivalent protein or fat, part of that bolus amount doesn't serve to clear the protein/fat-derived carbs (from gluconeogenesis), but rather simply replaces basal requirements because the pancreas stops secreting due to the exogenous insulin in the bloodstream. The IC calcs don't work.

I solved this by doing basal testing to find out what my basal requirements were. Then, experimented some more to see what my pancreas was capable of... Could it handle the basal needs on its own? If not, how much could it handle?

The answer to that question provides a new parameter for insulin calcs for T2s, one I call "basal substitution factor", or BSF. It's the amount of insulin your pancreas can produce to cover basal requirements.

The BSF can be less than your pancreas is capable of actually producing. It can also be less than your body needs to maintain stable BG (in which case you must continuously supplement with a basal setting on your pump).

Regardless, once you know your BSF, you can adjust your bolus programming for TAG to get much tighter control. I follow this method now, and have refined the numbers to the point that it is working amazingly precise for me: When I follow this approach, I

  1. Stay under 140 peaking after the meal much much more frequently
  2. Return to close to my target fasting BG within 3 hours almost every time
Before I started doing this, those two issues were quite variable, and I'd be doing a lot of corrections, especially when I had a low carb, high protein (not fat as much, but still an effect) meal.

Now, it's so reliable.

The Method
Use a dual-bolus program, with an initial impulse delivery of all the carb insulin, and an extended square wave bolus configured as follows:
D = duration (I generally use 3 hours, but will vary this depending on the content of the meal; mega-fatty, I may go 4 hours; low fat low carb, high protein I may drop it to 2.5)
C = carb content in grams
P = protein content in grams
I = insulin dose in units
BSF = bolus substitution factor, U/hr
IC = insulin/carb ratio

So, a meal with 40g carb and about 40g protein (Chick-fil-A Asian Chicken Salad, my lunch today) would be bolused for me this way:

Total I = (C + P/2) * IC + D * BSF
= (40 + 40/2) * 0.25 + 3 * 1
= 18U

The carb insulin is 10U. So, I program a bolus for 18U, to deliver 10U immediately, the other 8U over 3 hours.

This covers the carbs, and the glucose from the protein over the next 3 hours, plus the 1U/hr that my pancreas usually contributes to basal needs but would be seen by the pancreas as "replaced" by a part of the 8/3 = 2 2/3 units/hr from the extended bolus.

I'll follow up with a subsequent reply with details about how I determined BSF.