Preoperative structural and electrophysiological markers associated with visual outcome after idiopathic epiretinal membrane surgery.

purposeVisual recovery after idiopathic epiretinal membrane (iERM) surgery varies widely and is incompletely predicted by anatomy alone. We investigated whether preoperative structural (optical coherence tomography [OCT]) and functional (pattern electroretinography [PERG], photopic negative response [PhNR]) parameters were associated with final postoperative best-corrected visual acuity (BCVA) after combined phacovitrectomy for idiopathic epiretinal membrane and explored their associations with the magnitude of visual gain.

methodsThis prospective study included 35 eyes of 35 patients undergoing 25-gauge PPV with ERM and ILM peeling. Best-corrected visual acuity (BCVA), spectral-domain OCT (central macular thickness [CMT], ganglion cell complex [GCC]), PERG (P50, N95 amplitudes), and PhNR amplitude were assessed preoperatively and 6 months postoperatively. Multiple linear regression assessed preoperative factors associated with final postoperative BCVA, while exploratory logistic regression and ROC analyses evaluated favorable final acuity (BCVA ≤ 0.3 LogMAR); visual gain was analyzed separately as a continuous change score.

resultsMean age was 69.1 ± 7.7 years. Following combined phacovitrectomy, BCVA improved significantly (0.79 ± 0.26 to 0.31 ± 0.24 LogMAR, p < 0.001); CMT and GCC decreased, and PERG P50 and N95 amplitudes increased (all p < 0.001). Among PhNR-derived measures, baseline-to-trough amplitude and the PT/b ratio increased (both p < 0.001), whereas PT amplitude decreased (p = 0.013). Preoperative CMT (β = 0.349, p = 0.027) and P50 amplitude (β=-0.440, p = 0.004) were independently associated with visual outcome in the prespecified model; lower CMT and higher P50 were associated with better postoperative BCVA (lower LogMAR). Preoperative P50 was associated with final postoperative BCVA but not with the magnitude of visual gain (ΔLogMAR; p = 0.211). An exploratory logistic model including CMT, P50, and GCC was associated with favorable final postoperative BCVA, but its classification performance should be interpreted cautiously because of the small sample and lack of external validation. Exploratory ROC analyses showed modest discrimination for favorable final postoperative BCVA with preoperative P50 amplitude (AUC 0.706) and CMT (AUC 0.703).

conclusionIn this prospective cohort undergoing combined phacovitrectomy, higher preoperative PERG P50 amplitude was independently associated with better final postoperative BCVA, suggesting that preoperative electrophysiological assessment may provide adjunctive prognostic information beyond OCT-based anatomical measures. These preliminary findings suggest that preoperative PERG may provide adjunctive prognostic information in iERM; however, any P50 threshold should be considered exploratory until validated in larger independent cohorts.