The Metabolic Assessment of Patients With Diabetes Mellitus

This thesis critically evaluates two developments which have both brought about major advances in the assessment of glycaemic control in diabetic patients. First is the measurement of glycated proteins in the form of glycated haemoglobin and serum fructosamine and the second is the use of portable test strips and meters for the self-monitoring of blood glucose. Both methods of assessment have gained widespread acceptance but there remains a number of clinical and methodological problems associated with their use. This thesis reviews the literature relating to the use of these tests and describes a number of studies demonstrating new benefits and difficulties which exist. The majority of clinicians now use glycated haemoglobin measurement as their principal objective indicator of glycaemic control in diabetic patients. We have performed several detailed studies elucidating clinically relevant aspects of glycated haemoglobin measurement which can affect the assessment of glycaemia in these patients. The setting of target values for glycated haemoglobin measurement is severely hindered by the lack of standardisation in methodology. In an attempt to account for this, European guidelines define categories of glycaemic control as an HbA1 or HbA1c concentration so many standard deviations from a particular method's non-diabetic population mean. Using these guidelines, we found that HbA1 could classify the same diabetic patients differently to HbA1c, even when using the same instrument and reference range individuals. This new finding led, in part, to a change in the European guidelines for glycated haemoglobin targets. Many commonly used glycated haemoglobin methods include both glycated and non-glycated fetal haemoglobin (HbF) in their result. This means patients with elevated HbF concentrations (>0.5%) can have spuriously high glycated haemoglobin values. We presented the first evidence that insulin treated adult patients had a significantly greater prevalence of elevated HbF concentrations than either non-insulin treated or non-diabetic controls, thereby compounding the problem in this group of patients. In addition, we found that HbF increased the apparent imprecision of these assays. Discrepancies exist when comparing glycated haemoglobin and serum fructosamine as indicators of glycaemic control. We found that in newly diagnosed Type II diabetes, fructosamine only showed a correlation with HbA1 after glucose control had stabilised and not during the period of changing glycaemia. However, because changes in fructosamine preceded those of HbA1, the ratio of fructosamine/ HbA1 was able to predict the change in HbA1 in the forthcoming month. Thus, parallel measurement of fructosamine with HbA1 provided additional information on the future trend of a HbA1concentration. A further reason for disparity between fructosamine and glycated haemoglobin was established when investigating the effects of ageing on glycation in non-diabetic subjects. Mean HbA1 values rose with increasing subject age while fructosamine and fasting plasma glucose values did not. Consequently, when Type II diabetic patient samples were classified according to European HbA1 guidelines, significantly fewer patients were in good control and more in poor control when a young reference population was used compared to an age matched one. Thus, age related reference intervals may be required for glycated haemoglobin measurement. Using test strips and meters to measure blood glucose has become widespread amongst diabetic patients in the community and in the monitoring of acutely ill hospital patients. We have documented studies which describe the effect on glucose measurement of variations in several physiological parameters. We also describe how new technology is improving glucose meter measurement. In vitro variations of sample haematocrit have been shown to be a source of error in several glucose meter systems. Red cell count and mean cell volume are intimately related to haematocrit, so it makes it impossible to distinguish if the test strip error is due to haematocrit variations per se or to changes in the number of red cells. We found that in groups of individuals with different mean cell volumes, meter error was still related to the haematocrit of the sample rather than its red cell count. Having established this, 10 patients undergoing cardiopulmonary bypass (whose haematocrits routinely fall to 20% intraoperatively) were investigated to determine the effect of in vivo variations in haematocrit. Changes in test strip glucose accuracy were found to be consistent with those shown previously in vitro. While the use of whole blood to measure glucose is convenient, it is not possible, unlike plasma or serum, to tell if a sample is in any way haemolysed. We showed that extreme sample haemolysis can affect several glucose meter systems, while even modest degrees of red cell lysis were found to give clinically inaccurate results when using the Accutrend instrument. (Abstract shortened by ProQuest.)