Clinical and genetic determinants of progression of type 2 diabetes: a DIRECT study

Abstract

Objective: To identify the clinical and genetic factors that explain why the rate of diabetes progression is highly variable between individuals following diagnosis of type 2 diabetes.

Research design and methods: We studied 5,250 patients with type 2 diabetes using comprehensive electronic medical records in Tayside, Scotland, from 1992 onward. We investigated the association of clinical, biochemical, and genetic factors with the risk of progression of type 2 diabetes from diagnosis to the requirement of insulin treatment (defined as insulin treatment or HbA1c ≥8.5% [69 mmol/mol] treated with two or more noninsulin therapies).

Results: Risk of progression was associated with both low and high BMI. In an analysis stratified by BMI and HbA1c at diagnosis, faster progression was independently associated with younger age at diagnosis, higher log triacylglyceride (TG) concentrations (hazard ratio [HR] 1.28 per mmol/L [95% CI 1.15-1.42]) and lower HDL concentrations (HR 0.70 per mmol/L [95% CI 0.55-0.87]). A high Genetic Risk Score derived from 61 diabetes risk variants was associated with a younger age at diagnosis and a younger age when starting insulin but was not associated with the progression rate from diabetes to the requirement of insulin treatment.

Conclusions: Increased TG and low HDL levels are independently associated with increased rate of progression of diabetes. The genetic factors that predispose to diabetes are different from those that cause rapid progression of diabetes, suggesting a difference in biological process that needs further investigation.

Figures

Figure 1

BMI distribution and its effect on progression to insulin requirement. Distribution of BMI in the population studied is shown on primary axis. For each BMI band the Hazard Rate ratio is shown relative to the BMI band 24-26 kg/m2, error bars are the 95% confidence intervals for the hazard rate ratio.

Figure 2

Type 2 diabetes GRS association with different time spans in diabetes progression. The linear regressions were adjusted for BMI at diagnosis.