The macular capillary morphology in diabetic patients by optical coherence tomography angiography

Abstract

Objective To observe the macular capillary morphology in diabetic patients. Methods A total of 61 patients (104 eyes) with diabetes mellitus (DM group) and 31 healthy controls (41 eyes) were enrolled in the study. According to the degree of diabetic retinopathy (DR), the DM group was divided into non-DR (NDR) group, non-proliferative DR (NPDR) group, and proliferative DR (PDR) group. There were 13 patients (23 eyes), 21 patients (34 eyes) and 27 patients (47 eyes) in each group, respectively. According to whether there was diabetic macular edema (DME), the DM patients were divided into DME group and non-DME group, each had 20 patients (28 eyes) and 41 patients (76 eyes), respectively. The age (F=2.045) and sex (χ2=2.589) between the control group, the NDR group, the NPDR group and PDR group were not statistically significant (P=0.908, 0.374). The 3 mm × 3 mm region in macula was scanned by optical coherence tomography angiography (OCTA), and the retinal capillary morphological changes of superficial capillary layer (SCL) and deep capillary layer (DCL) were observed. Chi-square test and t test were used to compare data among different groups. Results There was no abnormal change of retinal capillary morphology in control group. Microaneurysms and foveal avascular zone (FAZ) integrity erosion can be found in NDR group. There were microaneurysms, FAZ integrity erosion, vascular tortuosity bending, capillary non-perfusion and venous beading in NPDR and PDR groups. The microaneurysms of DCL were significantly more than that of the SCL (t=4.759, P<0.001). The eyes with microaneurysms in NDR group, NPDR group, and PDR group showed significant differences (χ2=44.071, P<0.001), and the eyes with FAZ integrity erosion among these three groups also showed significant differences (χ2=30.759, P<0.001). Compared with NPDR group and PDR group, there were significant differences in vascular tortuosity bending and capillary non-perfusion (vascular tortuosity bending: OR=0.213, 95%CI 0.070−0.648, P=0.004; capillary non-perfusion: OR=0.073, 95%CI 0.022−0.251, P<0.001), and there was no significant difference in venous beading (OR=0.415, 95%CI 0.143−1.208, P=0.102). SCL blood flow density in the 4 groups (control, NDR, NPDR and PDR group) was 49.233±1.694, 48.453±2.581, 45.020±4.685 and 40.667±4.516, respectively. While the difference between the control and NDR group was not significant, the differences between other pairs (control vs NPDR/PDR, NDR vs NPDR/PDR, NPDR vs PDR) were significant. The ratio of FAZ integrity erosion and non-perfusion of DME group was significantly higher than those of non-DME group (vascular tortuosity bending: OR=7.719, 95%CI 1.645−36.228, P=0.004; capillary non-perfusion: OR=14.560, 95%CI 3.134−67.646, P<0.001). Conclusions OCTA can distinctively detect the abnormal retinal capillary changes of SCL and DCL in diabetic patients. Even in DM patients without diabetic retinopathy, OCTA can detect abnormal blood vessels. Key words: Diabetic retinopathy; Regional blood flow; Tomography, optical coherence; Foveal avascular zone