Detecting early microvascular disease in type 1 diabetes: wavelet transform analysis of Doppler blood velocity waveforms

Abstract

Diabetic retinopathy is associated with markedly increased risk of cardiovascular events. Analysis of retrobulbar blood velocity waveforms should help characterise microvessels since waveform morphology is partly determined by wave reflection. This approach could potentially allow the detection of abnormalities earlier than is currently possible. Ultrasound examinations of the common carotid, ophthalmic (OA) and central retinal (CRA) arteries were undertaken on 39 subjects with type 1 diabetes (no retinopathy) and 39 controls. Waveforms were characterised using the discrete wavelet transform, and frequency data categorised into 11 bands. After correcting for possible confounders, in the OA, the mean amplitude was higher for subjects with diabetes in band 4 (p=0.022). In the CRA, the mean amplitude was higher for subjects with diabetes in bands 2 (p=0.012), 3 (p<0.001) and 5 (p=0.028). Abnormalities were most pronounced when signals were captured close to the microvasculature, providing evidence that they may be due to microvascular alterations.