Sarah, this can get very confusing, so make sure you run through it over and over to get it down well!
One thing to keep in mind that I mentioned in that other discussion is that splitting doses can leave you with periods where you are well below your nominal basal insulin needs, and this will cause you to see a BG spike. Not because of DP, liver dumping, or anything else. Because you don't have enough basal insulin.
Here's a graphical picture to help understand. For this example, assume this person has a nominal basal requirement of 1U/hr. Lantus works in him for 19 hours, at which time its action falls off over the next hours or so. It is gone at the 20 hour mark.
Standard dosing would be to prescribe 24* 1U = 24U of Lantus, 1x a day. If administered that way @ 6AM every day, this person would expect to see their BG start to rise about 2AM, getting quite a ways up there by 6AM, when the next Lantus injection occurs (as well as a big correction bolus).
Now, let's split that dose into two 12U injections, taken at 6AM and 6PM. We'll ignore the bootstrapping* problem for the moment, and assume Lantus has been taken in this manner steady-state (i.e., there was not starting point -- there's already Lantus in the person from a previous injection with each subsequent injection).
At 6AM when the morning 12U is taken, this adds 0.5U/hr to the existing 0.5U/hr from the injection 12 hours prior, for a total of 1U/hr. That's good -- that's what we need.
However, 8 hours later, at 2PM, the evening injection from the previous day wears off. Now only the most recent injection is active. So the basal rate drops to 0.5U/hr for the next 4 hours -- from 2PM to 6PM -- when there's another injection. This person is going to see their BG rising all afternoon, and be correcting correcting correcting, until they inject Lantus again at 6PM.
The same dip will occur from 2AM to 6AM, for the same reasons -- which could certainly look like DP. Here's a graphical view of what basal levels look like in this theoretical individual for single, 2, and 3 evenly split doses:
![]()
Now this is a very crude approximation of what's going on, but it serves to illustrate the point: Basal levels are not constant when using long-acting and splitting doses. You will have several points in the day where "inadequate" basal is available (in fact, with evenly split doses, the number of dips is the same as the number of injections).
Also, we can clearly see that the more doses you split the total into, the less severe the dips.
This offers some options for someone willing to MDI on their basal. It may be easier, from a convenience/timing standpoint, to split into 3 or 4 doses, spread evenly. This will have the added advantage of raising the low points of insulin availability, reducing high BG excursions. The 3-dose approach works well with human sleeping patterns -- dose right before bed, again in 8 hours when you get up. Sleep through the night. Dose again in another 8 hours, and then another 8. Repeat.
The math behind all this is too much for most people (including doctors), so no one bothers. However, it's not snake-oil -- you can split your lantus into 10 doses a day (if they aren't less than 1U), and everything works out fine -- if you don't make a mistake.
Anywho, a little more food for thought. I'd be happy to help out with the math, etc. in dialing this in for you once you know what your own pharmacokinetics (how long Lantus works for you) and your nominal basal needs.
