HbA1c Variability in Type II Diabetes

One of the last unanswered question in relation to the influence of glycemic control on diabetes complications is whether increased month-to-month changes in blood glucose (as measured by variability in glycated hemoglobin (HbA1c)) compounds the complication rate and if this can be altered with intervention. Qatar has a high prevalence of diabetes, affecting approximately 23% of the adult population (International Diabetes Federation 2014) that is going to lead to the development of both microvascular and macrovascular complications resulting in the increased morbidity and mortality associated with the disease. The fact that improved glucose control in type 2 as well as type 1 diabetes reduces the risk of microvascular complications is well established. However, more recently it has been demonstrated that the month-to-month variability (the 'rises and falls') in glucose control are also associated with an increased risk of developing these diabetes-associated problems. An individual's long term measure of blood glucose control is represented by the amount of HbA1c measured in the blood. The HbA1c level changes slowly over a much longer period than the constantly fluctuating glucose levels, giving a good indication of overall glucose control in the preceding 2-3 months. What is not known is whether interventions to reduce variability in HbA1c could, in turn, lead to a reduction in diabetes complications. For example even when HbA1c mean is the optimal 7%, there can be high variability in the HbA1c measures (large standard deviation) that may still lead to complications.

This study proposes to gather data to determine whether different treatment thresholds for diabetes in Qatar people have inherently different effects on the variability of HbA1c on a month-to-month basis. By establishing an understanding of how different treatment regimens for hyperglycemia may affect HbA1c variability, this study would then inform on a long term study designed to determine whether interventions to reduce HbA1c variability can reduce micro- or macrovascular complication risk independently of mean HbA1c. If proven, this concept would allow patients to help avoid glycaemia-related vascular complications without having the high potential risk of hypoglycemia that is associated with the current gold standard of diabetes care.

The investigators plan to recruit 150 patients on any glucose lowering medication (HbA1c 7.5-9%), randomize them into one of two treatment threshold groups and test their HbA1c every 6 weeks for 20 visits (visit 1 baseline; therefore 114 weeks) to assess the HbA1c variability of each group. Self-monitored fasting plasma glucose (FPG) measurement will be undertaken 3 times weekly and reported back to the medical team as part of the safety monitoring. Patients will be randomly divided into 2 treatment thresholds. In one the investigators will intensify treatment in those with FPG\>140mg/dl until their FPG is \<90mg/dl, using whatever treatment is clinically appropriate for them, and only intensify it further if their FPG rises to \>140mg/dl again. In the other group Investigators will intensify if their FPG is \>115 mg/dl until it is \<=115 mg/dl and intensify further if \>115 mg/dl again. As such the study will be treatment threshold dependent and therapy independent. This will help circumvent any concern that the drug regimen could complicate the analysis or present a confounder. In practical terms it means the investigators give both groups of patients the same therapy that is intensified according to the treatment threshold with the addition of the same hypoglycemic agents as used in routine clinical practice. Intensifying treatment dose would be undertaken if three consecutive FPG were above the target of 140 or 115 mg/dl. This will be advised by the patient ringing the study coordinator and/or the study coordinator ringing the patient weekly and advising the consultant what the FPG values are for action. It is anticipated that the mean HbA1c will be comparable but the variability of the HbA1c will differ between the 2 populations.