Abstract
PURPOSE: To assess the accuracy of refraction prediction after intraocular lens (IOL) exchange using a software integrated in a Scheimpflug-based biometer. SETTING: 2 tertiary referral hospitals. DESIGN: Case series. METHODS: Eyes that underwent uneventful bag-to-bag IOL exchange between 2017 and 2024 and complete preoperative and postoperative measurements were included. IOL exchange was performed by 2 experienced surgeons. IOL power was calculated using the Barrett RX formula integrated in Pentacam software, incorporating both phakic and pseudophakic biometric data and subjective refraction. Prediction errors were computed by comparing actual postoperative and predicted refractions. Moreover, an unpublished estimation formula was evaluated. RESULTS: 25 eyes of 25 patients (mean age 60.2 ± 11.4 years) were included. Mean interval between primary IOL implantation and exchange was 25.3 ± 31.5 weeks. In 11 of 25 eyes, a different IOL model was implanted during the exchange. Using the Barrett RX formula, the mean prediction error (MPE) was 0.02 ± 0.32 diopters (D), mean absolute error (MAE) 0.27 ± 0.17 D, and median absolute error (MedAE) 0.22 D. A total of 56% and 92% of eyes were within ±0.25 D and ±0.50 D of target refraction, respectively. The conventional formula yielded a comparable MPE of -0.12 ± 0.68 D ( P = .479), but higher MAE and MedAE (0.53 ± 0.42 D and 0.40 D, P < .001), with 32% and 56% of eyes within ±0.25 D and ±0.50 D, respectively. CONCLUSIONS: The Barrett RX formula was demonstrated to be an efficacious method for calculating IOL power in eyes undergoing IOL exchange, with more than half of cases within ±0.25 D of the predicted postexchange refraction.