Brachial artery flow-mediated dilation and mean platelet volume in relation to diabetic retinopathy severity in type 2 diabetes: a retrospective cross-sectional study

Background To evaluate the cross-sectional associations of brachial artery flow-mediated dilation (FMD) and mean platelet volume (MPV) with diabetic retinopathy (DR) severity in adults with type 2 diabetes mellitus (T2DM), after adjustment for age, sex, diabetes duration, glycated hemoglobin (HbA1c), and systolic blood pressure; carotid ultrasound measures were examined secondarily. Methods This single-center, hospital-based cross-sectional study retrospectively analyzed clinical and imaging data obtained during routine inpatient care through a standardized multidisciplinary assessment pathway from 109 hospitalized adults with T2DM between June 2015 and December 2019: 46 with no DR (NDR), 36 with non-proliferative DR (NPDR), and 27 with proliferative DR (PDR). The main exploratory complete-case proportional-odds model included 86 participants (36 NDR, 30 NPDR, and 20 PDR) and simultaneously included age, sex, diabetes duration, HbA1c, systolic blood pressure, FMD, and MPV; a nephropathy-adjusted model and a ridge-penalized model were examined as sensitivity analyses. Results Median FMD was 6.9%, 6.1%, and 5.3% in the NDR, NPDR, and PDR groups (Kruskal-Wallis P = .025), and median MPV was 9.30, 9.60, and 10.00 fL (P = .012). In the main exploratory complete-case model (N = 86), lower FMD (adjusted odds ratio [OR] per 1-percentage-point increase, 0.831; 95% CI, 0.713-0.968; P = .018) and higher MPV (adjusted OR per 1-fL increase, 1.898; 95% CI, 1.238-2.912; P = .003) were associated with greater DR severity. After additional adjustment for recorded diabetic nephropathy, the FMD estimate was attenuated (OR, 0.855; 95% CI, 0.732-0.998), whereas the MPV estimate was preserved (OR, 2.166; 95% CI, 1.356-3.458). In a ridge-penalized sensitivity model, both estimates were attenuated (FMD OR, 0.899 and 0.919; MPV OR, 1.388 and 1.289 at the minimum-loss and one-standard-error penalties, respectively). Conclusions Lower brachial FMD and higher MPV were associated with greater DR severity in this selected hospital-based cross-sectional sample; the FMD association was sensitive to renal adjustment, and both associations, particularly that for MPV, were noticeably attenuated under penalization. These exploratory findings do not establish temporal sequence, causality, predictive performance, or clinical utility and require confirmation in larger prospective cohorts.