Accuracy of intraocular lens power calculation in cataract surgery combined with trabeculectomy in open angle glaucoma

Background/Objectives: To assess the accuracy of several intraocular lens power calculation formulas in phacotrabeculectomy for open angle glaucoma. Methods: Patients who underwent phacotrabeculectomy for open angle glaucoma were included. Refraction and biometry measurements were repeated at 3, 6 and ≥12 months. Prediction error (PE) and absolute error (AE) were calculated using the SRK/T, Holladay 1, Hoffer Q, Haigis, Kane, Emmetropia Verifying Optical (EVO) and Barrett Universal II formulas at ≥12 months, and their accuracy was compared using linear mixed-effects models accounting for repeated measurements within the same eye and inter-eye correlation. Results: Sixty eyes from 40 patients were included. The linear mixed-effects model showed a significant overall effect of formula on PE (χ2(6) = 119.14, p < 0.001). Most formulas showed a tendency toward a hyperopic refractive shift, whereas Haigis showed a negative PE. Based on estimated marginal mean AE, the formulas were ranked as follows: EVO (0.548 D), Barrett Universal II (0.551 D), Holladay and SRK/T (0.561 D), Haigis (0.572 D), Kane (0.577 D) and Hoffer Q (0.617 D). However, the AE did not significantly differ among the formulas (χ2(6) = 3.75, p = 0.711). The percentage of eyes within ± 1.00D of PE ranged from 81.7% to 90% across the formulas (p > 0.05). Significant axial length shortening, anterior chamber deepening and mean keratometry reduction were detected postoperatively at ≥12 months (p < 0.05). Conclusions: Despite postoperative ocular anatomic changes, all formulas showed acceptable refractive accuracy after phacotrabeculectomy. Although no significant difference in the AE was detected among the formulas, the PE differed significantly, with most formulas showing a tendency toward a hyperopic shift and Haigis showing a myopic shift. This inter-formula difference should be considered when selecting the refractive target, particularly when using formulas that tend toward hyperopic PE.