Validation of hemoglobin glycation models using glycemia monitoring in vivo and culturing of erythrocytes in vitro.

Glycated hemoglobin A1c (HbA1c) concentration in blood is an index of the glycemic control widely used in diabetology. The aim of the work was to validate two mathematical models of HbA1c formation (assuming irreversible or reversible glycation, respectively) and select a model, which was able to predict changes of HbA1c concentration in response to varying glycemia courses with higher accuracy. The experimental procedure applied consisted of an original combination of: in vivo continuous glucose concentration monitoring, long-term in vitro culturing of the human erythrocytes and mathematical modeling of HbA1c formation in vivo and in vitro with HbA1c values scaled according to the most specific analytical methods. Sixteen experiments were conducted in vitro using blood samples collected from healthy volunteer and stable type 1 diabetic patients whose glycemia was estimated beforehand based on long-term monitoring. The mean absolute difference of the measured and predicted HbA1c concentrations for the in vitro experiments were equal to 0.64 +/- 0.29% and 1.42 +/- 0.16% (p = 0.0007) for irreversible and for reversible model, respectively, meaning that the irreversible model was able to predict the glycation kinetics with a higher accuracy. This model was also more sensitive to a deviation of the erythrocytes life span.