Sacha E.M. Teunissen MSc^1,2, Gozewijn D. Laverman^1,2

¹ Division of Nephrology, Department of Internal Medicine, Ziekenhuisgroep Twente, 7609PP Almelo, The Netherlands ² Department of Biomedical Signals and Systems, University of Twente, 7522NB Enschede, The Netherlands

Email: s.teunissen@zgt.nl

Background: Long-term glucose dysregulation is a risk factor for various complications in type 2 diabetes patients. The average HbA1c level over time is a widely used measure of glucose regulation and a predictor of complications and mortality. The same applies for glycaemic variability, which is evaluated using the coefficient of variation and/or the standard deviation of HbA1c over time, but these measures are hard to interpret in clinical practice. Therefore, the HbA1c Variability Score (HVS) has been introduced, which calculates the percentage of HbA1c changes greater than 5.5 mmol/mol. However, little is known about whether HVS, like the coefficient of variation and standard deviation, is a predictor of complications and mortality. This study focuses on HVS as predictor of mortality and the development of cardiovascular complications (CVC), acute kidney injury (AKI) and malignancies in type 2 diabetes patients.

Methods: HbA1c values of 724 type 2 diabetes patients from the follow-up of the DIALECT cohort were analysed over a median follow-up period of 9.1 years. After excluding the patients with fewer than 5 HbA1c values and missing data on endpoints, the association between HVS and the four endpoints was evaluated using two Cox proportional hazard models per endpoint: one without adjustments and one adjusted for various risk factors per endpoint.

Results: An increased HVS was associated with an increased risk of mortality (Model 1: HR 1.01 [1.001 – 1.02], Model 2: HR 1.01 [1.004 – 1.023]) and an increased risk of developing AKI (Model 2: HR 1.011 [1.001 – 1.021]). Other associations were not significant.

Conclusion: HVS appears to be a predictor of mortality and AKI. Although an increased HVS also seems to be associated with malignancies and CVC, these results were not significant. Further research in larger populations is needed to confirm the predictive value of HVS.