Blood selectin levels as a predictive factor for diabetic retinopathy and diabetic macular edema in type 2 diabetes

Abstract

Background: Diabetes mellitus (DM) is still a noninfectious global pandemic. Diabetic retinopathy (DR) is one of the most common and socially significant complications of both type 1 and type 2 diabetes. Predicting the probability of DR progression and potential diabetic macular edema (DME) development is still important. Currently available predictive models include a wide number of predictors, and such predictors as blood selectin levels seem to be promising. Purpose: To determine relationships between blood levels of selectins and DR progression and DME development in patients with T2DM, and to develop particular predictive models. Material and Methods: Of the 124 patients (124 eyes) involved into this study, 95 (95 eyes) had T2DM and retinopathy (group 1 of 29 eyes with mild non-proliferative DR (NPDR); group 2 of 35 eyes with moderate or severe non-proliferative NPDR; and group 3 of 31 eyes with proliferative DR (PDR)), and 29 (29 eyes) had no diabetes (controls). Patients underwent a routine eye examination and spectral domain optical coherence tomography (SD-OCT) to determine central retinal thickness (CRT). The presence of DME was based on an increased macular thickness in the Early Treatment Diabetic Retinopathy Study (ETDRS) subfields compared to the upper limit of normal for patient's age and gender. Enzyme-linked immunosorbent assay (ELISA) kits from Invitrogen Thermo Fisher Scientific (USA) were used to determine selectin levels in blood. Statistical analyses were conducted using MedStat and MedCalc v.15.1 (MedCalc Software bvba) software. Results: There was a statistically significant tendency for an increase in progression of DR to be accompanied by an increase in blood selectin levels. In the presence of DME, blood LS levels were statistically significantly increased in mild NPDR; blood PS levels, in PDR; and blood ES levels, in all stages of DR. A regression model for predicting the progression of DR was built. In the model developed, diabetes duration and blood selectin levels were positively correlated, while HbA1c level was inversely correlated with DR progression. The model was found to be adequate R2adjust = 0.84; F = 97.9, p less 0.001, demonstrating high correlation of the selected independent variables with the stage of DR. The calculated Y index may be considered a quantitative reflection of the severity of DR. A patient is predicted to have mild NPDR (with a prediction accuracy of 86.2%) if Y less 1.5, moderate of severe NPDR (with a prediction accuracy of 88.6%) if 1.5 less equal Y less 2.35, and PDR (with a prediction accuracy of 100%) if Y more equal 2.35. A regression model for predicting the development of DME was built, with CRT and blood ES level found to increase the risk of the development of DME. A high AUC value for the ROC curve (AUC = 0.97; 95% CI, 0.92-0.99) indicated a very strong correlation of the risk of DME development with the levels of selected independent variables. In addition, sensitivity and specificity values at the optimal cut-off point were 93.2% (95% CI, 83.5%-98.1%) and 92.3% (95% CI, 83.0%-97.5%), respectively. Conclusion: Our findings confirmed the current view that high blood selectin levels are important in the presence of T2DM and related to major pathogenetic mechanisms of microvascular complications, and blood selectin levels can be considered predictors of the progression of DR as well as the development of DME.