Association between the extent of urinary albumin excretion and glycaemic variability indices measured by continuous glucose monitoring

Aims The contribution of glycaemic variability to the microvascular complication of diabetes has not been established. We examined whether there is an independent association between indices of glycaemic variability in continuous glucose monitoring and extent of albuminuria. Methods A total of 173 patients with Type 2 diabetes (without insulin therapy, n = 96; with insulin therapy, n = 77) who had unexplained large fluctuations in blood glucose values underwent three-day continuous glucose monitoring. We used a multinomial logistic regression model to determine whether the indices of glycaemic variability independently affected the odds of having a spot urine albumin/creatinine ratio of 30–299 mg/g and ≥ 300 mg/g. Results Higher standard deviation (P = 0.002), mean of daily differences (P = 0.023) and mean amplitude of glycaemic excursion (P = 0.043) significantly increased the odds of having a urine albumin/creatinine ratio of ≥ 300 mg/g. In multivariable analysis, only higher standard deviation, but not mean amplitude of glycaemic excursion and mean of daily differences, independently increased the odds of having a urine albumin/creatinine ratio of ≥ 300 mg/g (P = 0.025). Coefficient of variation (sd/mean) was not associated with the odds of having a urine albumin/creatinine ratio of 30–299 or ≥ 300 mg/g. Conclusions The independent association between standard deviation and the extent of albuminuria was lost when the measures were normalized by mean glucose level. At least in terms of relative measures of glycaemic variability, we failed to demonstrate an independent association between glycaemic variability and albuminuria extent in patients with inadequately controlled Type 2 diabetes.