David Mendosa looked at this and had a few comments. Population studies are inherently studies of association. It is really hard to know whether people with impaired cognitive function or macrovascular problems have a lower A1cs or vice versa. And if those with low A1cs are truly experience severe hypos then it is entirely believable that having a lows could result in these problems. But that doesn’t mean that an A1c of 5% without severe hypos leads to either cognitive impairment or macrovascular problems. And I’ve never seen credible studies that suggest A1cs of 5-5.5% represent issues. And I don’t consider things like ACCORD to be credible.
I would note that if you look at the source of this number it is from a 1999 report on diabetic patients younger than 40 years old and represents “all cause mortality.” This is not mortality from hypos. And these patients were from the “pre DCCT” era when people didn’t believe that blood sugar control affected complications and the standard of care was conventional insulin therapy and there weren’t any blood sugar meters.
Philip Cryer is considered an expert in this area and has a book “Hypoglycemia in Diabetes: Pathophysiology, Prevalence, and Prevention.” But I consider Cryer and bit of a wolf cryer. For instance, in his chapter on hypoglycemia in the 2016 book “Textbook of Diabetes” he writes:
Three early reports indicated that 2–4% of people with diabetes die from hypoglycemia [87–89]. More recent reports indicated that 6% [83], 7% [90], and 10% [91] of deaths of people with T1DMwere the result of hypoglycemia. In T2DM, mortality rates of up to 10% during episodes of severe sulfonylurea-induced hypoglycemia have been reported [92]. In one trial of T2DM, between 1 and 9% of evaluable deaths were attributed to hypoglycemia [93].
But if you look up the references:
83, 87-89 and 90 were all in the time era I talked about above and confounded all cause mortality with hypos.
90 was a study entitled “Acute Complications and Drug Misuse Are Important Causes of Death for Children and Young Adults With Type 1 Diabetes”
91 was a study of children born between 1972 and 1982 and were under 15 years when diagnosed. They looked at causes of mortality, none of which included hypos.
92 A 2003 paper entitled “Risk of Hypoglycaemia with Oral Antidiabetic Agents in Patients with Type 2 Diabetes” which didn’t have a mortality finding as a major result.
93 was a study of severe hypoglycemia in the ACCORD study which while it found that severe hypos increased risk of death, tight control did NOT increase severe hypos or death
What is uniformly dissappointing about Cryers work is that as a scientist he is critically flawed. He takes these studies, misinterprets the results, fails to understand the difference between association, doesn’t understand confounding factors and then spins number that are likely wildly inflated. And these numbers are repeated and repeated. Even the JDRF repeats them.
And I personally think he is just crying wolf. Not that severe hypos can’t kill you. But that the number suggesting that upwards of 10% of us will die from a hypo is just not even in the ballpark of real. It is an alternative fact.