Text

Purpose: To investigate vectorial astigmatism time course and identify its predictors following combined phacoemulsification and pars plana vitrectomy.

Methods: This multicenter retrospective cohort study included 1,041 eyes from 1,041 patients undergoing phacovitrectomy at four centers. Eyes with macula-off retinal detachment, prior vitreoretinal surgery, silicone oil endotamponade, or postoperative CDVA worse than 0.4 (decimal) were excluded. One eye per patient was analyzed. Preoperative astigmatism was derived from keratometry and corrected using the Abulafia–Koch method, while postoperative astigmatism was obtained from refraction. Astigmatism was vectorially decomposed into J0 and J45 components. Changes over time were assessed using multivariate repeated-measures ANOVA, and predictors of ΔJ0 and ΔJ45 were identified using multivariate linear regression.

Results: The orthogonal astigmatism component (J0) decreased from +0.62 ± 0.38 D (95% CI: 0.60 to 0.64) preoperatively to +0.05 ± 0.45 D (95% CI: 0.02 to 0.08) postoperatively, while the oblique component (J45) increased from 0.003 ± 0.25 D (95% CI: -0.01 to 0.02) to +0.04 ± 0.30 D (95% CI: 0.02 to 0.05), indicating a systematic vectorial shift (p <0.001, ηp² = 0.349). The final regression models demonstrated high predictive accuracy, explaining 88.5% of the variance in ΔJ0 (adjusted R² = 0.885, SEE = 0.267, p <0.001) and 83.5% of the variance in ΔJ45 (adjusted R² = 0.835, SEE = 0.205, p <0.001). Baseline astigmatism was the dominant predictor in both models. Axial length, incision site, and incision meridian significantly predicted ΔJ0, while laterality, incision size, and incision meridian significantly predicted ΔJ45 (all p <0.001). Endotamponade type and gauge size showed no relevant effect.

Conclusions: Phacovitrectomy induces minimal but statistically significant changes in astigmatism, characterized by a reduction in with-the-rule component and a small induction of oblique astigma-tism. These changes are highly systematic and predominantly determined by incision characteristics and baseline astigmatism rather than intraocular factors.