Errors in clinical assessment of glycemic control in patients with non-insulin-dependent diabetes mellitus.

Errors in the assessment of glycemic control by clinicians, using the conventional method, based on random blood glucose and clinical parameters other than HbA1, were assessed in 125 non-insulin-dependent diabetic (NIDDM) patients (Group B). The expected mean post-prandial plasma glucose concentrations (y) for these patients were obtained from their HbA1 values (x) using the regression equation y = 39.6 X x - 157.6. This equation was derived using the data from 35 NIDDM patients (Group A), in whom mean post-prandial plasma glucose concentrations (measured twice daily), correlated excellently (r = 0.82; n = 100; p less than 0.001) with mean HbA1 values (measured monthly) determined over a period of 3 months. A comparison of the observed and expected post-prandial plasma glucose concentrations in Group B indicated that only an overall 42% of the total (n = 455) observed values (as against 60% in Group A) fell within the 99% confidence limits of the expected mean value. Fifty-eight % of the observed plasma glucose values fell outside the expected range of which 36% were underestimates and 22% were overestimates. Thus, we concluded that even in patients with stable diabetes the physician's assessment of the overall degree of glycemic control should be based on both plasma glucose and HbA1 concentrations for optimizing therapeutic measures.