The evaluation and the effect of surgically induced astigmatism in phacoemulsification surgery

Abstract

Three issues regarding surgically induced astigmatism (SIA) include the choice of SIA calculators, the relationship between SIA consistency with phacoemulsification technique variation, and the effect of SIA prediction error on toric IOL calculation. Phase I compared the individual SIA value and analysis report from SIAC, SIA2.1, SIA3.1, OC6.0 and VVC calculators. Mean difference and agreement between calculators were assessed. No significant differences in mean SIA between the calculators (p = 0.71) were found. The 95% limit of agreement between calculators ranged from -0.006 to 0.005D. OC6.0 and VVC reported astigmatism analysis according to Alpins method, whereas SIAC, SIA2.1 and SIA3.1 were based on Holladay method. Holladay method calculators provided surgeons' SIA centroid which is the actual SIA. SIAC and SIA3.1 provided subset analysis of SIA. SIA2.1 and SIA3.1 provided coherence analysis which is a clinically relevant value to evaluate SIA consistency. Thus, SIA2.1 were used in Phase II and III of this study for its suitability to the study objective. In Phase II study, the relationship between SIA consistency and phacoemulsification technique variation were evaluated. The actual SIA and coherence of four surgeons were calculated. A questionnaire (PTechSIA) was developed, validated and used to collect information on the surgeons’ surgical technique. Input from the PTechSIA were used to quantify the surgeons' phacoemulsification technique variation using technique variation score (TVS). Strong negative correlation between SIA coherence and TVS (Spearman's r=-0.95, p=0.05) were found, indicating that a consistent phacoemulsification technique contributed to a better SIA consistency. In Phase III, we evaluated the surgeons' SIA prediction error and its effect on toric IOL selection. SIA prediction error was observed in all four surgeons. Two surgeons had a statistically significant SIA prediction error in X and Y value (p <0.01). This error resulted in different toric IOL selection in 85% of subjects. Furthermore, underestimation of SIA had resulted in higher IOL toricity selection and vice versa. From the three-phase study, all issues were addressed accordingly. SIA calculator with centroid and coherence analysis are suggested to determine surgeons' actual SIA and monitor the SIA consistency. A consistent surgical technique contributes to a high SIA consistency, and SIA prediction error significantly affects toric IOL selection. These findings justified the clinical significance and relevance of SIA value.