With a lot of experimentation! Flat fasting line during daylight hours, no problem when I’m on split Levemir. When I was taking Lantus in the morning, dip before lunch. when I took Lantus at lunchtime, dip late afternoon.
It’s night time when the fun really starts, though. Imagine a slack washing line, higher at both ends and curved downwards towards the middle. That’s what I achieve with x units of basal. Increase basal, x + 1 and the curve gets much steeper on the way down, then flattens out and remains flat for the second half of the night, (as if you’ve tilted the curve one way). Decrease basal, x-1, and the overnight trace stays flat for the first half of the night, then steeply rises for the second half, as if you’ve tilted the whole curve the other way. Conclusion. It’s my body that produces the dip overnight, not the basal.
I need to start off in the 7s at bedtime to avoid going below 4 at 3am, and I need to whack in 2 units of Novorapid when I wake up, irrespective of breakfast bolus, to counteract the rise. If I was on a pump, this would be built in with my basal pattern).
I used to really struggle working out patterns before I started using the Libre. Once I could see the whole 24hrs, it became obvious. I assume my own hormones (cortisol?) must reduce glucose output as soon as my head hits the pillow, and then tell my liver to fire up again in the wee small hours. My consultant also assumes I'm still producing a bit of endogenous insulin, because my basal needs are fairly low, but I need an average amount of bolus. (There’s some research, which has been mentioned on the forum from time to time, that found most Type 1s even if they developed diabetes quite young, still produce a bit of their own).
Sorry, that ended up more long winded than I was expecting!