Type 2 diabetes occurs when not enough insulin is produced by the body or when the insulin that is produced fails to work properly. Most patients with type 2 diabetes are treated with drugs to lower blood sugar (glucose) levels, with the aim of reducing or delaying the damaging effects of diabetes.
It is well known that high blood glucose has damaging effects on the body, but lower-than-normal blood glucose (hypoglycaemia) can also be harmful and this may explain the increased risk of death associated with strict glucose control. Some diabetes drugs are associated with a higher risk of hypoglycaemia than others. The objective of this study is to compare the association between glucose levels and risk of death in patients with type 2 diabetes who are treated with the higher-risk drugs and in those treated with the low-risk drugs.
People with type 2 diabetes will be identified from CPRD and classified into treatment groups according to whether they are being treated with higher-risk or low-risk therapies. The association between glucose levels and risk of death within each treatment group will be examined using statistical modelling to adjust for other potential causes of death in type 2 diabetes.
A U-shaped association between glycated haemoglobin (HbA1c) and all-cause mortality has been demonstrated previously. An increased risk of hypoglycaemia could explain the increased risk of all-cause mortality associated with low HbA1c values. We will use a retrospective cohort study design. The objective of this study is to explore the association between HbA1c and all-cause mortality in patients with type 2 diabetes treated with glucose-lowering therapies associated with a higher risk of causing hypoglycaemia (insulin, meglitinides and sulfonylureas as monotherapy or in combination with other glucose-lowering therapies) and those treated with glucose-lowering therapies associated with a lower risk of causing hypoglycaemia (acarbose, metformin, DPP-4 inhibitors, SGLT-2 inhibitors, GLP-1 agonists and thiazolidinediones as monotherapy or in combination with other glucose-lowering therapies excluding insulin, sulfonylureas or meglitinides).
Regimens initiated between 1 January 2004 and 31 December 2014 will be selected. HbA1c will be modelled as 12 monthly updated time-dependent mean and cumulative mean values. Cox proportional hazards models will be used to explore the association between HbA1c and all-cause mortality within each regimen of interest.
Craig Currie - Chief Investigator - Cardiff University
Sarah Holden - Corresponding Applicant - Pharmatelligence Limited t/a Human Data Sciences
Sara Jenkins-Jones - Collaborator - Pharmatelligence Limited t/a Human Data Sciences
HES Admitted Patient Care;ONS Death Registration Data