Abstract
Purpose :
Clinical calculators generating outcome-based risk stratification have been developed in various medical and surgical specialties. The purpose of this retrospective study is to identify clinical and anatomical characteristics of epiretinal membrane (ERM) predictive of improved visual acuity (VA) following surgery to determine the feasibility of developing a web-based predictive calculator.
Methods :
A retrospective chart review of adults who underwent their first epiretinal membrane removal surgery at the Duke Eye Center from January 1, 2018 to December 31, 2020 and had ≥ one-year post-surgical follow-up was performed. Adults with idiopathic as well as secondary causes of ERM were included. Data on demographics, medical history, and preoperative ocular exam and OCT findings were collected. Change in VA was calculated at the one year post-surgical visit. Using the median of the differences in VA, patients with a change in VA greater or less than the median were considered to be good or poor responders, respectively. Data for preoperative characteristics were noted for each group (Figure 1). Univariable analysis followed by a stepwise multivariable logistic regression (p<0.1 cutoff) was performed to identify characteristics associated with good response to surgery.
Results :
Seventy-eight patients were enrolled with an average age of 68 +/- 8 years. The average VA (logMar) at the 1-year follow-up (0.33 +/- 0.28) was improved compared to the pre-op visit (0.61 +/- 0.37, p<0.001). On univariable analysis the good response group had a higher presence of vitreomacular traction or adhesion (55% vs 31%, p=0.03), worse preoperative VA (logMAR 0.76 vs 0.47, p<0.001), reduced duration of symptoms (8 vs 14 months, p=0.02), and thinner center point thickness (CPT) on OCT (326 vs 461 μM, p=0.005). On multivariable analysis, significant indicators of good response were worse preoperative VA (OR=12.9, p=0.006) and thinner CPT (OR=0.997, p=0.013).
Conclusions :
Results showing preoperative VA and CPT stratify risk for more improvement in VA following ERM peeling suggests continued study of risk prediction in adults undergoing ERM surgery is warranted. Additionally, by including idiopathic and secondary etiologies of ERM our results suggest a clinical risk calculator may apply to a real-world population of individuals with ERM.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.