BACKGROUND:

This study examined whether annual variation in glycosylated hemoglobin A1c (HbA1c) and fasting plasma glucose (FPG), as represented by the coefficient of variation (CV), can predict diabetic nephropathy independently of mean FPG, mean HbA1c, and other risk factors in patients with type 2 diabetes.

METHODS:

A computerized database of patients with type 2 diabetes aged ≥30 years and free of diabetic nephropathy (n = 3220) who were enrolled in the Diabetes Care Management Program of China Medical University Hospital before 2007 was used in a time-dependent Cox proportional hazards regression model.

RESULTS:

The incidence rates of diabetic nephropathy were 16.11, 22.95, and 28.86 per 1000 person-years in the first, second, and third tertiles of baseline HbA1c-CV, respectively; the corresponding incidence rates for FPG-CV were 9.46, 21.23, and 37.51 per 1000 person-years, respectively. After multivariate adjustment, the corresponding hazard ratios for the second and third tertiles versus the first tertile of annual HbA1c-CV were 1.18 (95% confidence interval [CI], 0.88-1.58) and 1.58 (95% CI, 1.19-2.11), respectively, and 1.55 (95% CI, 0.99-2.41) and 4.75 (95% CI, 3.22-7.01) for FPG-CV, respectively. Therisks of diabetic nephropathy for HbA1c-CV and FPG-CV stratified according to age, gender, renal function, and hypertension status were provided.

CONCLUSIONS:

Annual FPG and HbA1c variations have a strong association with diabetic nephropathy inpatients with type 2 diabetes. Whether intervention for reducing glucose variation should be administered needs to be examined in a future study.