I'm sorry no one responded yesterday. I saw your post. I was a work and it's been more than mayhem there these days.
One of my rules of thumb is if I tested and got a number that seemed completely out of sync, I would retest. It's entirely possible to get a reading from your meter but you had a bad strip, not quite enough blood, or it was a glitch.
It also possible that when number drop into a range your cat is no longer used to spending time in, they "bounce" back to a higher range. This happens if numbers drop low, drop fast, or as I said, your kitty isn't used to being in a lower range. We refer to this as a bounce when the numbers spike back up. This results from the liver and pancreas releasing a stored form of glucose along with counterregulatory hormones. It can take 3 days for a bounce to clear.
One thing that doesn't work is to reduce the dose. Bounces are normal -- then annoy the heck out of us but are a normal, protective mechanism.
If you are following SLGS, you want to hold the dose for 7 days. At that point, you should run a curve (i.e., test every 2 hours for a cycle) and if numbers are above 90, increase the dose.
I'm not sure how you measured a dose of 2.58u. With SLGS, dose changes are made in 0.25u increments. You can microcode but Nova's numbers aren't at a place yet where fine adjustments are needed.
Likewise, Lantus doesn't do well with abrupt changes. Lantus likes consistency. This is due to the depot. Any time you do anything that influences the dose, it disrupts the depot. o
Let's see what the cycle looks like today. If needed, ti would probably make sense to put the dose back to 2.5u.