Abstract
PURPOSE: To compare the accuracy of different formulas for calculating toric intraocular lens (IOL) power. METHODS: PubMed, EBSCO, Web of Science, and the Cochrane Library were systematically searched for studies published from 2015 to 2025. Numbers of eyes with a prediction error (PE) within ±0.50 and ±1.00 diopters (D) were synthesized for the meta-analysis. Besides, continuous vector analysis of mean centroid magnitudes was quantitatively analyzed to assess directional bias in refractive outcomes. RESULTS: Nine retrospective clinical studies, including 1863 patients, 1959 eyes, and 19 calculation formulas, were identified. According to the ranking based on the surface under the cumulative ranking curve by the Bayesian method, the three highest ranking formulas are EVO MPCA (86.9%), Kane (81.8%), and Hoffer QST (75.9%) in PE within ±0.5D. In the range of 1.0D, the three highest ranking formulas are "EVO MPCA" (83.7%), "Haigis" (70.6%), and "Kane" (70.1%). As for the comparison of mean centroid, the highest three are "Z CALC2 PPCA" (87.0%), "Holladay 2" (79.3%), and "ATCTCRP" (78.8%). CONCLUSION: This network meta-analysis indicated that EVO MPCA formulas performed best among the 19 toric IOL power formulas.