ECG–gated retinal vessel calibre as a novel measure of aberrant pulsatile retinal flow in diabetes mellitus: a cross-sectional study

Abstract

Abstract Purpose To evaluate ECG-gated retinal vessel calibre as a novel measure of aberrant pulsatile retinal flow in diabetes mellitus. Methods A novel mydriatic ECG-gated fundoscope was used to acquire retinal vessel calibre measurements at four cardiac cycle time points from end diastole to early diastole. An inhouse software recorded the exact time when an image was captured to corroborate ECG-synchronisation. Arterial applanation tonometry, an alternative method of assessing aberrant blood flow, was used to measure carotid-femoral pulse wave velocity (cPWV) and augmentation index (AIx). The intraclass correlation (ICC) was used to perform intra- and inter-observer reliability analyses. Two reviewers measured the retinal vessel calibre in single retinal arterioles and venules. A receiver operating characteristic curve determined associations with diabetes mellitus. Results In this study 119 controls and 120 participants with diabetes mellitus were recruited. Mean peak change in retinal arteriolar calibre from baseline was higher in diabetes mellitus compared with controls (controls: 0.92%, IQR 0.63 vs diabetes mellitus: 2.05%, IQR 1.25, p<0.0001). In a subset of 9 controls and 11 participants, the intra-and inter-observer reliability was high (ICC 0.87-0.97) in mean peak changes in retinal vascular responses from baseline. In a subset of 36 controls and 95 participants with diabetes mellitus, diabetes mellitus was more strongly associated with retinal arteriolar pulsatility (AUC 0.85, 95%CI 0.76-0.93) than applanation tonometry (cfPWV AUC 0.72, 95%CI 0.62-0.82 vs AIx AUC 0.56, 95%CI 0.45-0.68). Conclusion Higher retinal arteriolar pulsatility appears to be more strongly correlated with diabetes mellitus than arterial applanation tonometry.