Nefeli M. Dimitropoulou¹, Simone J.P.M Eussen², Casper G. Schalkwijk¹, Bastiaan E. de Galan¹

¹Department of Internal Medicine, Maastricht University Medical Center+, the Netherlands, ²Department of Epidemiology and Care and Public Health Research Institute (CAPHRI), Maastricht University, the Netherlands.

Email: nefeli.dimitropoulou@maastrichtuniversity.nl

Background: Globally, the incidence of type 2 diabetes (T2D) is increasing, leading to an increased risk of cardiovascular disease (CVD). Skin autofluorescence (SAF), a reflection of Advanced Glycation Endproducts (AGEs), is a strong predictor of CVD. SAF has been found to be increased in people with T2D, and is thought to result directly from elevated glucose in the interstitium. It is unknown whether glucose variability (GV) may also drive the formation of AGEs, as measured by SAF. We therefore aimed to assess the association of SAF with GV in people with prediabetes, T2D and normal glucose metabolism (NGM) using data from continuous glucose monitoring (CGM) downloads.

Methods: Data were derived from The Maastricht Study, a population-based cohort study. The population was stratified by NGM, prediabetes and T2D individuals. Coefficient of variance (CV) and standard deviation (SD) were used to investigate the correlation (Spearman) of SAF with GV in the total population and between the three groups. Multiple linear regression analyses were conducted to adjust for potential confounders, including age, sex and cardiovascular risk factors.

Results: We included 459 people with NGM, 174 people with prediabetes and 162 people with T2D. SAF was greater in T2D compared to prediabetes and NGM (2.26±0.59 vs 2.12±0.435, p=0.014 vs 2.00±0.401, p=0.007). SAF was significantly correlated to SD (R=0.24, p<0.001) and CV (R=0.20, p<0.001) in the total population. These correlations were driven by the effects in NGM (R=0.13, p=0.004) and prediabetes (R=0.15, p=0.046) subgroups for SD and in the NGM subgroup for CV (R=0.11, p=0.016), whereas these correlations were non-significant in the T2D subgroup (R=0.12, p=0.12 and R=0.11, p=0.16, respectively). The associations between SD and CV with SAF in the total population did not change after adjustments.

Discussion/Conclusion: SAF is independently associated with GV in people with NGM and prediabetes, but not or less so in people with T2D. Further research to explain this difference is needed.