Researchers in the U.K. suggest that a hemoglobin A1c of 7.5% may be optimal in terms of longevity for type 2 diabetics treated with drugs, according to a study published recently in The Lancet.
Hemoglobin A1c (HgbA1c) is a blood test widely used as a gauge of blood sugar control, reflecting average blood sugars over the previous three months. The American Diabetes Association recommends a HgbA1c goal of 7% or less. The American Association of Clinical Endocrinologists recommends 6.5% or less. Dr. Richard K. Bernstein, a diabetologist and himself a type 1 diabetic, recommends HgbA1c’s as near normal as possible (about 5%).
Many physicians believe that keeping blood sugar levels as close to normal as possible—often referred to as “tight control”— will help prevent certain diabetes complications such as blindness, kidney failure, and nerve damage. We have good supportive evidence.
We assume tight control would also help prevent premature death from heart attacks and strokes, too. Several recent studies—the ACCORD and ADVANCE trials—call this into question, however. The ACCORD trial, for example, achieved near-normal glucose control with multiple medication options, yet found that the effort was linked to increased death rates from cardiovascular disease and from any cause (all-cause mortality).
The scariest thing about tight control is hypoglycemia, which can kill you quickly, for example, if you’re operating dangerous machinery (e.g., driving), scuba-diving, or rock-climbing.
U.K. researchers recently reviewed records of diabetics treated either with 1) two oral medications (usually metformin and a sulfonylurea), or 2) a regimen containing insulin. Each group had over 20,000 subjects. They found that risk of death for those with an average HgbA1c of 6.4% (the lowest blood sugar levels in this study) was 52% higher than those with HgbA1c of 7.5%. Those with the highest blood sugar levels over time—HgbA1c over 10% if I recall correctly—had the highest risk of death. In general, those taking insulin had higher rates of death than those on pills.
It’s extremely difficult to interpret studies like this. There are myriad ways to treat diabetes. We have 10 classes of drugs for treatment of diabetes: this study looked at three. There are at least three types of “diabetic diet” in common use: low-fat/high-carb, low-carb, and just regular eating, which depends on where you live. Exercise, too, plays a role in treatment and longevity.
With all these variables, should we put much stock in a study that looks at longevity from the perspective of just two therapeutic regimens? How well would a football team do with just two plays in its play-book?
You’d think we would have a definite answer to the “tight versus loose control” issue by 2010. We don’t. It’s still very appealing to me to think that, if done right, tight control would yield the better outcomes. Problem is, we don’t always know what’s right.
One thing is clear: Having a HgbA1c of 7.5% is better than 10% in terms of health and longevity.
But is 7.5% really better than 6.5 or 5.5 or 5.0% for a particular individual on a particular treatment program? Probably not. That’s why the ADA and AACE emphasize that treatment programs be tailored to the individual patient.
Maybe controlling blood sugar levels is like controlling high blood pressure. The ideal may be 120/80, but you gain very little, if any, by reducing high blood pressure below 140/90 (130/85 for diabetics). HgbA1c of 5.0% may be ideal, but not necessary.
Currie, Craig, et al. Survival as a function of HgA1c in people with type 2 diabetes: a retrospective cohort study. The Lancet, January 27, 2010. Early online publication doi: 10.1016/S0140-6736(09)61969-3
Dluhy, Robert and McMahon, Graham. Intensive glycemic control in the ACCORD and ADVANCE trials. New England Journal of Medicine, 358 (2008):2630-2633.
One response to “How Well Should Diabetes Be Controlled?”
“Maybe controlling blood sugar levels is like controlling high blood pressure.”
I would argue the opposite for a couple of reasons.
#1 – while it is probably pragmatically impossible to show protection below 140/90 because the trial would be cost prohibitive with respect to size, I’m not aware of any data that says that going from 140 to 120-130mmHg is harmful. With diabetes, however, there is data to say that it can be harmful to take people with a long history of poorly controlled diabetes and aggressively treat them pharmacologically to more “normal” levels.
#2 – with the exception of beta blockers and co-morbidities (ie, ACEi for hypertension with renal disease), the “means to the end” don’t seem to matter much to bp. If you lower it through ACE inhibition, an ARB, calcium channel blocker, or a diuretic, it’s more important to get the bp down and the patient to tolerate the drug than it is to worry about which drug it is. With respect to diabetes, though, the end does not seem to justify the means with respect to drugs, at least at our current knowledge (or level of ignorance). Metformin and insulin seem to have some evidence of macrovascular disease protection like heart attacks whereas the TZD class doesn’t. At the end of the day, it seems more important to control bp and LDL cholesterol in diabetics than it is to get A1c down below 7.5% by any means necessary.
Until the medical community starts demanding hard endpoints from the diabetes trials, we’re going to continue to be plagued with drugs in which we are unclear as to their net long term benefits with respect to lessening the burden of microvascular disease (blindness) or macrovascular disease (heart attacks). We’ve focused on treating glucose numbers for too long and ignored the sum total of the disease.