Abstract:AIM: To investigate the relationship between the coefficient of variation(CV)of blood glucose and microvascular changes in the macular area and visual function in patients with non-proliferative diabetic retinopathy(NPDR), and to provide a new perspective for optimizing blood glucose management strategies and delaying the progression of diabetic retinopathy(DR).
METHODS:This was a cross-sectional study. Patients with NPDR admitted to our hospital from January 2025 to March 2026 were enrolled and divided into mild NPDR group, moderate NPDR group, and severe NPDR group according to NPDR severity, and further divided into normal visual function group and abnormal visual function group based on visual function status. Multiple linear regression was used to analyze the correlation between CV and macular microvascular parameters. Receiver operating characteristic(ROC)curves were constructed to evaluate the value of CV in assessing the progression of NPDR severity. Multivariate logistic regression was performed to identify independent influencing factors of visual function. Generalized linear models were employed to explore the interaction between CV and NPDR severity and visual function. Restricted cubic spline(RCS)models were used to analyze the dose-response relationship between CV and visual function across different severity levels.
RESULTS:A total of 160 patients(160 eyes)with non-proliferative diabetic retinopathy(NPDR)were enrolled in this study. Only one eye per patient was included as the study eye. If both eyes met the diagnostic criteria for NPDR, one eye was randomly selected using a random number table. According to the severity of NPDR, the patients were divided into a mild group(62 patients, 62 eyes), a moderate group(58 patients, 58 eyes), and a severe group(40 patients, 40 eyes). The mild group included 33 male and 29 female patients, with a mean age of 58.26±6.53 y; the moderate group included 30 male and 28 female patients, with a mean age of 59.14±6.64 y; and the severe group included 21 male and 19 female patients, with a mean age of 60.33±6.71 y. Statistically significant differences were observed among the three NPDR severity groups in disease duration, fasting blood glucose(FBG), glycated hemoglobin(HbA1c), CV, foveal avascular zone(FAZ)area, mean superficial capillary plexus(SCP), upper SCP, lower SCP, mean deep capillary plexus(DCP), upper DCP, and lower DCP(all P<0.05). Multiple linear regression analysis demonstrated that CV was positively correlated with FAZ area(β>0)and negatively correlated with mean SCP, upper SCP, lower SCP, mean DCP, upper DCP, and lower DCP(β<0, all P<0.05). ROC analysis revealed that when CV assessed NPDR progression to moderate and severe stages, the area under the curve(AUC)was 0.851, with a sensitivity of 0.828, a specificity of 0.810, and an optimal cut-off value of 29.82%; when assessing progression to severe NPDR, the AUC was 0.896, with a sensitivity of 0.900, a specificity of 0.838, and an optimal cut-off value of 36.21%. According to visual function status, the patients were divided into a normal visual function group(88 patients, 88 eyes)and an abnormal visual function group(72 patients, 72 eyes). The normal visual function group included 46 male and 42 female patients, with a mean age of 58.51±6.81 y; the abnormal visual function group included 38 male and 34 female patients, with a mean age of 59.82±6.93 y. The abnormal visual function group had significantly longer disease duration and higher FBG, HbA1c, CV, and FAZ area than the normal visual function group, while mean SCP, upper SCP, lower SCP, mean DCP, upper DCP, and lower DCP were significantly lower(all P<0.05). Predictive model 1 for abnormal visual function included disease duration, FBG, HbA1c, FAZ, mean SCP, and mean DCP; predictive model 2 included FBG, HbA1c, FAZ, mean SCP, mean DCP, and CV. Compared with model 1, model 2 showed significantly improved AUC(0.817, 95%CI: 0.785-0.849), integrated discrimination improvement(IDI; 0.082, 95%CI: 0.064-0.096), and net reclassification improvement(NRI; 0.618, 95%CI: 0.479-0.819)(all P<0.05), and the Hosmer-Lemeshow test indicated good calibration for both models. Interaction analysis showed that the positive association between NPDR severity and abnormal visual function gradually strengthened with increasing CV levels. RCS analysis revealed a non-linear dose-response relationship between CV level and abnormal visual function across different severity levels(nonlinear test, χ2=16.124, P<0.001).
CONCLUSION: CV is a risk factor for macular microvascular injury and visual function abnormalities in patients with NPDR. As CV increases, FAZ area expands, SCP and DCP vessel density decreases, and the risk of NPDR progression and visual function impairment increases stepwise. In the glycemic management of NPDR patients, in addition to achieving HbA1c targets, attention should be paid to the control of blood glucose fluctuations. It is recommended to maintain CV below 29.82% to reduce the risk of NPDR progression, and to use CV≥36.21% as a high-risk warning threshold for severe NPDR and visual function impairment. CV should be considered as a supplementary indicator for assessing DR risk and guiding treatment decisions.