Prediction of postoperative astigmatism in cataract surgery.

BACKGROUND

The purpose of this study was to predict postoperative astigmatism using refraction, keratometry, and other preoperative information after phacoemulsification with posterior chamber intraocular lens (IOL) implantation.

METHODS

A retrospective study was conducted of 176 eyes of 161 patients undergoing phacoemulsification for visually significant cataract with capsular bag or sulcus IOL fixation at the Hunter Holmes McGuire Veterans Administration Medical Center, Richmond, Va. Eyes with complications and final-corrected visual acuity of less than 20/60 were excluded. Keratometric and refractive astigmatism were described by Jackson cross-cylinder with-the-rule (J0) and oblique (J45) components. Preoperative predictors of postoperative refractive J0 and J45 were determined by univariate and multivariate regression analysis.

RESULTS

The final multivariate model to predict postoperative with-the-rule astigmatism was J0Postoperative = 0.24 x J0Preoperative 0.46 x J0Keratometric (-0.08 )(coefficient of determination [r2] = 0.51, all p < 0.001). The multivariate model based on expected predictors of oblique astigmatism was J45Postoperative = 0.10 x J45Preoperative + 0.23 x J45Keratometric + 0.01, (r2 = 0.05, p = 0.09 for J45Preoperative and 0.03 for J45Keratometric). Temporal clear cornea (as opposed to superior scleral tunnel) incisions, and use of sutures in scleral tunnel incisions, were not predictive of postoperative astigmatism.

INTERPRETATION: Postoperative astigmatism can be estimated from preoperative astigmatism using the following weighted average: two-thirds keratometric and one-third refractive astigmatism. Preoperative refraction may predict postoperative astigmatism independent of keratometry because keratometry reflects only the paracentral anterior (not posterior) corneal curvature.