Comparison of refractive prediction for intraoperative aberrometry and Barrett True K no history formula in cataract surgery patients with prior radial keratotomy

PURPOSE To compare prediction errors of the Barrett True K No History (Barrett TKNH) formula and intraoperative aberrometry (IA) in eyes with prior radial keratotomy (RK). METHODS A retrospective, non-randomized study of all patients with RK who underwent cataract surgery using IA at the UCHealth Sue Anschutz-Rodgers Eye Center from 2014 to 2019 was conducted. Refraction prediction error (RPE) for IA and Barrett TKNH was compared. General linear modelling accounting for the correlation between eyes was used to determine whether absolute RPE differed significantly between Barrett TKNH and IA. Outcome by number of RK cuts was also compared between the two methods. RESULTS Forty-seven eyes (31 patients) were included. The mean RPEs for Barrett TKNH and IA were 0.04 ± 0.92D and 0.01 ± 0.92D, respectively, neither was significantly different than zero (p = 0.77, p = 0.91). The median absolute RPEs were 0.50D and 0.48D, respectively (p = 0.70). The refractive outcome fell within ± 0.50D of prediction for 51.1% of eyes with Barrett TKNH and 55.3% with IA, and 80.8% were within ± 1.00D for both techniques. Mean absolute RPE increased with a higher number of RK cuts (grouped into < 8 cuts and ≥ 8 cuts) for both Barrett TKNH (0.35D and 0.74D, p = 0.008) and IA (0.30D and 0.80D, p = 0.0001). CONCLUSIONS There is no statistically significant difference between Barrett TKNH and IA in predicting postoperative refractive error in eyes with prior RK. Both are reasonable methods for choosing intraocular lens power. Eyes with more RK cuts have higher prediction errors.