Corneal topography as a predictor of refractive change in the prospective evaluation of radial keratotomy (PERK) study

The first operated eyes of 435 patients undergoing radial keratotomy in the Prospective Evaluation of Radial Keratotomy (PERK) study were evaluated by photokeratography to document the preoperative and postoperative corneal shape. We determined by regression analysis and analysis of variance that the corneal shape preoperatively improved the prediction of the corneal shape 6 months postoperatively in the 3-mm-clear-zone population. The reduction of myopia in all 435 eyes ranged from 1.25 to 9.75 diopters. We studied the effect of the preoperative corneal shape on this variability in the outcome of the surgery using rings 2 and 7 on photokeratography and corneal diameter. In the 3-mm-clear-zone group, eyes with flat prolate corneas had a greater reduction in myopia (4.65 D); those with steeper, more spherical corneas had less reduction in myopia (3.48 D). In addition, eyes with a 3-mm clear zone and flat central corneas alone (8.0 mm = 42.19 D) flattened approximately 0.75 D more than those with steep central corneas (7.0 mm = 48.21 D). In the 3.5-mm and 4.0-mm clear zone groups, the change in corneal curvature was not related to the preoperative curvature. A stepwise regression analysis of the 151 eyes in the 3.0-mm-clear-zone population demonstrated the following predictive equation for radial keratotomy: change in cycloplegic refraction = -14.55 + [-2.097 x average ring-2 radius] + [3.605 x average ring-7 radius] + [0.69 x horizontal corneal diameter] + [0.079 x age] + [-0.379 x spherical equivalent cycloplegic refraction]. There was a 1.17-D observed difference in the effect of radial keratotomy between those eyes with a steep/steep corneal topography (7.2% of the 3.0-mm-clear-zone population) and the flat/flat topography (29% of the 3.0-mm-clear-zone PERK population). A knowledge of corneal topography provides an additional tool for understanding the operative variability of radial keratotomy.