Author: Maria Vittoria Cicinelli
Co-authors: Giulia Gregori, Rosangela Lattanzio, Francesco Bandello
Abstract
Purpose: To longitudinally investigate the factors associated with diabetic retinopathy (DR) severity fluctuations and to explore their prognostic implications in terms of progression to proliferative DR (PDR).Setting: Longitudinal retrospective cohort study on patients with DR seen at San Raffaele Scientific Institute (Milano, Italy)
Methods: Patients with DR having at least 2 ultra-widefield (UWF) fundus photography examinations and no history of PDR-related complications at inclusion visit were recruited. UWF fundus images were graded using the Early Treatment Diabetic Retinopathy Study severity scale (DRSS). Multivariable linear mixed models were used to assess factors associated with DRSS standard deviation (SD) as a proxy of fluctuations. Risk factors for DR improvement or progression were computed with mixed-effect Cox hazard models. The mean DRSS area-under-the-curve (AUC) was included as a covariate in all statistical analyses.
Results: 111 eyes with a median follow-up of 44 months and a median of 3 (IQR 2-5) UWF fundus photography exams per study eye were included. Type 1 diabetes (estimate[95% confidence interval(CI)]= 4.36[0.77-7.91] SD DRSS, p=0.02) and macular non-perfusion (estimate[95% CI]= 5.30[1.70-8.72] SD DRSS, p=0.004) were associated with wider fluctuations in DR severity. During the follow-up, 89 eyes (73%) underwent intravitreal injections for macular edema; higher number of anti-vascular endothelial growth factor injections administered (hazard ratio(HR)[95% CI]=1.15[1.01-1.33] for injection, p=0.03) was associated higher chance of ≥1-step improvement in DRSS. Higher HbA1c (HR[95% CI] =1.83[1.26-2.69] for 1% increase, p=0.01), higher mean DRSS-AUC (HR[95% CI]=1.25[1.10-1.43] for DRSS/month increase, p
Conclusions: Patients with larger DR variability may need attentive follow-up to early identify DR worsening. DRSS-AUC may be a useful prognostic factor for future studies.
