Refractive outcomes of phacoemulsification after pars plana vitrectomy using traditional and new intraocular lens calculation formulas

Abstract

Purpose: To evaluate refractive outcomes of cataract surgery in previously vitrectomized eyes and compare the accuracy of intraocular lens (IOL) calculation formulas in this population.

Setting: Stony Brook University Hospital, Stony Brook, New York, USA.

Design: Retrospective case series.

Methods: All electronic medical records of patients who had phacoemulsification from 2013 to 2017 were reviewed. Patients who had previously had pars plana vitrectomy in the same eye were selected. Patients with a history of refractive surgery, silicone oil in the eye, or other factors preventing accurate IOL power calculation were excluded.

Results: The mean postoperative spherical equivalent was -0.16 diopter (D) ± 0.81 (SD) (n = 61). The mean prediction errors were 0.30 ± 0.82 D, -0.09 ± 0.76 D, 0.23 ± 0.76 D, 0.25 ± 0.81 D, 0.04 ± 0.85 D, 0.30 ± 0.82 D, 0.33 ± 0.79 D, and 0.45 ± 0.80 D with the Holladay 1, Wang/Koch (WKA) adjusted Holladay 1 (n = 14), Holladay 2 (n = 48), SRK/T, WKA SRK/T (n = 14), Hill-Radial Basis Function (n = 50), and Ladas formulas, respectively. The formula with the highest percentage of predictions within ±0.50 D of the postoperative outcome was the Holladay 2 (60.42%). Statistically significant differences between the predicted and actual refractive outcomes were found with all formulas (P < .05) except the WKA Holladay and WKA SRK/T. Intraclass correlation showed low repeatability (<0.50) for all formulas.

Conclusions: Refractive outcomes after cataract extraction in vitrectomized patients can be variable and more hyperopic than the predicted outcomes using traditional and newer IOL calculation methods. Thus, predicting refractive outcomes in this population is challenging, and patients should be counseled accordingly.