Case in point: I just read a five year randomized but not blinded study of type 2 diabetics over a five year period. They were randomized to take either NPH twice a day or Lantus once a day. The people doing the study had (have) a hope that Lantus acts to reduce risk of retinopathy development. Towards the beginning of their paper, they claim that Lantus might prevent retinopathy by lowering overall blood sugar.
Their Lantus group started out with an average A1c of 8.41%, and the NPH group started with an A1c of 8.31%. Their Lantus group, five years later, had an average A1c of 7.80%, while the NPH group had improved a little bit MORE on average, to 7.55%. That is, they'd started out 0.1 better and ended up 0.25 better. The data analysis says with greater than 95% certainty that it wasn't a coincidence- that something about the NPH group or the NPH resulted in lowered blood sugar. Whether it was the NPH or something else, who knows.
The progression of retinopathy was similar between the two groups (it was slightly lower on the measure they chose in the Lantus group at five years, but not much, and it was worse on other retinopathy measures in the study), and so in their conclusion the authors claim that the A1cs back their case because the Lantus users didn't have more retinopathy despite having higher A1cs. I dunno... would you really expect a difference in retinopathy rates that would show up in their study based on a difference in A1c of .25%? I also thought it was noteworthy that the retinopathy rate that was seen in the study groups was higher than the researchers expected- in a five year period, a quarter of participants had retinopathy that got a lot worse.
The other studies making me think about this were about insulin degludec, which is currently in clinical trials and which also shows no real difference in A1c compared to Lantus or NPH or anything else, but which claims to lower rates of severe hypoglycemia. And a review of studies on CGMs, finding that most find no significant difference in A1c (although some do and I think that overall the studies show a very small average decrease in A1c) but less hypoglycemia.
So it seems to me that if we could trust research conclusions, the new stuff should reduce hypoglycemia and offer minimal or negative benefit to reducing hyperglycemia and long term complications.
On a slightly more jolly note, the other day I got to thinking about how my sweat smells different after a particularly bad hypoglycemic episode, or when I wake up sometimes, if I've been hypo in the night. I woke up Friday morning and thought I smelled funny, sniffed my armpit and knew that I smelled hypo, so I'd either been hypo earlier in the night or was currently hypo (as it happened, bg was 54 and Dex showed that I'd gone LOW four hours earlier but had been gradually rising). I know there are dogs (and cats) that alert people to hypoglycemia, but I wonder if we could train humans to sniff hypoglycemia? Even better, could we be trained to wake up upon smelling it?
Using the search term "Hypoglycemic odor," I found a number of interesting articles informing me that insulin decreases the rate of death of cells responsible for the sense of smell, and that giving rats insulin changes their sense of smell in unpredictable ways (fun fact, amiright?), but I did not find anything about how hypoglycemic people smell. Maybe I'll look more later.
I have to say, my sense of smell has only changed in a way big enough for me to notice once in my life, and that was about two months after I went on testosterone, when suddenly I started smelling smells I'd never smelled before. That was seventeen months after I went on insulin, so I don't think the two are related.