The effect of fasting plasma glucose variability on the risk of retinopathy in type 2 diabetic patients: retrospective long-term follow-up.

Aims: This study aimed to determine whether fasting plasma glucose (FPG) variability can predict diabetic retinopathy development and progression independently of glycemic control.

Methods: Subjects consisted of 170 type 2 diabetes mellitus patients not showing diabetic retinopathy on their first visit to our hospital between 1966 and 1979, and continuously visited thereafter for 27-40 (mean, 33) years. Plasma glucose and HbA1c data obtained on every hospital visit were collected. As FPG variability parameter, standard deviations (SD) were used.

Results: The investigation in which patients were classified based on the median value of the mean HbA1c or FPG and that of FPG SD have revealed that the risk of retinopathy development and progression were similar between the group showing good control and high variability and that with poor control and low variability. In multivariate analysis models with the mean FPG or HbA1c as a covariate and models using the mean FPG or HbA1c as a time-dependent covariate after adjustment for age, sex, diabetes duration, BMI, hypertension, and hypoglycemia, FPG SD was a significant independent risk factor for retinopathy.

Conclusion: In type 2 diabetic patients, FPG variability is a risk factor for diabetic retinopathy independent of the mean FPG or HbA1c.