American Diabetes Association 78th Scientific Sessions:

Time to Rethink the Role of A1c in Patients with Diabetic Comorbidities

With Jordan E. Perlman, MD, and Irl B. Hirsch, MD

“In the study I presented today, we looked at the differences in hemoglobin A1c (HbA1c) and average glucose in patients with diabetes who also had one of three comorbid conditions: chronic kidney disease (CKD), anemia,” and nonalcoholic fatty liver disease (NAFLD), said Jordan E. Perlman, MD, a resident at the University of Washington (UW) Medical Center in Seattle, during a session at the American Diabetes Association 78th Scientific Sessions in Orlando, Florida. 


“What we showed is that average glucose levels [obtained by downloaded CMG data] for several common chronic conditions were greater than when converting an HbA1c for the basis of comparison,” Dr. Perlman told EndocrineWeb.

This was one outcome in the larger study, which examined the entire clinic population to assess the discordance in HbA1c in comparison to average blood glucose and how this information may inform our treatment decisions, she said. 

A1c No Match for CGM-Captured Average Glucose

A retrospective chart review of patients being managed at the UW Diabetes Center from 2011-2017, and who had two weeks worth of fingerstick glucose levels or glucose data recorded by a continuous glucose monitor.1 Average blood glucose (AG) was derived from HbA1c using an equation developed by the ADAG Study Group.2 The relationship between average blood glucose (AG) and HbA1c was compared and showed a linear relationship.

A summary of the study findings follow:1

  • Non-anemic vs anemic patients (n=220 and 123, respectively) had better ADAG agreement: 70% of AGs inside 15% of predicted (55% of AGs, P=0.006).
  • Patients with normal glomerular filtration rate (n=440) vs. those with anemia had better ADAG agreement: 71% of AGs inside 15% of ADAG estimation, which declined by 12% when limited to just anemic patients (n=83, P<0.001).
  • ADAG agreement decreased as a function of renal impairment but showed no effect with regard to abnormal level function tests. The impact of anemia had significance only at stage 3a CKD (n=49, P=0.02) but there was no correlation between AG and A1C in patients with stage 4 CDK/end-stage renal disease (n=14, R2=0.0004).
  • With regard to NAFLD, no correlation was found between AG and A1C (n=14, R2=0.03) but 64% of AGs were inside 15% of ADAG projection.

For many patients, glucose data, be it from both from fingerstick or CGM data, provided a necessary confirmation of the HbA1c assessment, said Dr. Perlman. “However, in certain conditions, we found significant discordance between AG and A1C. Specifically, there was no correlation between AG and HbA1c in advanced renal dysfunction and NAFLD, according to results from our lab.”

“The bottom-line is that the differences between measures of HbA1c and [CGM-derived] average glucose levels across all patients are discordant whether they have any comorbid conditions or not,” she said.

“Since [clinicians] rely on HbA1c as a reflection of the patient’s blood glucose control, and we learn that there is a large difference between the A1c and average glucose, it might not be so safe to make treatment decisions based solely on the HbA1c. Therefore, the next question for us to pursue is: how much discordance is too much,” said Dr. Perlman.

CGM Brings Precision Medicine to Diabetes Management


"The study that was presented by Dr.  Perlman reinforces to me that we may not have been treating patients with diabetes the right way for many years.1 Specifically, we’ve been using hemoglobin A1c (HbA1c) to make treatment decisions both in type 1 and type 2 diabetes for many years," said Irl. B. Hirsch, professor of medicine and chair of diabetes treatment and teaching at the Endocrine and Diabetes Care Center, University of Washington School of Medicine in Seattle.

We have guidelines for which there may be some disagreement, according to Dr. Hirsch; In effect, the guidelines are good for populations of patients but for not so for individuals.

"HbA1cs don’t tell you the granularity of what the number actually means in terms of average glucose, the amount of hypoglycemia the patient is experiencing, and a host of other factors," he said. 

Reliance On Blood Glucose Levels Necessary for Diabetes Treatment Planning

"Therefore, in order to make a sound treatment decision, I need to see the blood sugar levels because the data shows that the HbA1c doesn’t even give us what the average glucose is for a particular patient the way we think it does," Dr. Hirsch said "Yet, with continuous glucose monitoring (CGM), [blood glucose levels falls outside the desired range] about 18% of the time.

"The other big point which was shown 10 years ago but hasn’t been emphasized enough is that for any HbA1c level, we don’t know what the patient’s average glucose level is, and even if everything is being managed correctly. In effect, a HbA1c of 7% could mean that the patients average glucose is somewhere in the range of 146 -184 mg/dL," said Dr. Hirsch.

"You need the data from a CGM to show precisely where the patient's glucose falls [in-range or not], otherwise, you won’t know. Therefore, for the majority of patients who have type 2 diabetes, the drug therapy (ie, adding insulin) based on the HbA1c alone, is in my view, not very good medicine," he told EndocrineWeb.

None of the authors had any financial conflicts.

Last updated on

close X