Variability in the relationship between mean plasma glucose and HbA1c: implications for the assessment of glycemic control.

BACKGROUND Previous studies have shown a single linear relationship between mean plasma glucose (MPG) and hemoglobin A(1c) (HbA(1c)). We examined the relationship in different treatment groups of patients with type 1 diabetes participating in the Diabetes Control and Complications Trial (DCCT). METHODS Seven-point glucose profiles (premeal, postmeal, and bedtime) and HbA(1c) were measured quarterly during the DCCT. We studied measurements from (a) intensively treated patients at study commencement, (b) intensively treated patients after stabilization of their glycemia (from 6 months onward), and (c) conventionally treated patients from 6 months onward. Only complete glucose profile and HbA(1c) pairings were considered (n = 589, 11 483, and 11 855, respectively). RESULTS From 6 months into the trial, conventionally treated patients had consistently higher MPG concentrations than intensively treated patients at any given HbA(1c) value (mean difference, 1.6 mmol/L at 7% HbA(1c), increasing to 2.8 mmol/L at 11% HbA(1c)). Similarly, at the same HbA(1c), the MPG of intensively treated patients at baseline was higher than in the same individuals after 6 months of intensive treatment (1.2 mmol/L difference at 7% HbA(1c), increasing to 4.6 mmol/L at 11% HbA(1c)). CONCLUSIONS The relationship between MPG and HbA(1c) is not constant but differs depending on the glycemic control of the population being studied. Having lower mean glucose at the same HbA(1c) may help explain why intensive DCCT treatment appeared intrinsically linked to both increased hypoglycemia and decreased microvascular complications compared with conventional treatment. These findings may also have implications for expressing HbA(1c) as mean blood glucose equivalent.