Multiple regression and vector analyses of laser in situ keratomileusis for myopia and astigmatism

To construct a quantitative model relating refractive results to laser settings and other factors.

A prospective clinical trial was performed, including 14 surgeons and 523 eyes of 278 patients who desired correction of myopia from -1.00 to -16.00 D (mean, -6.20 D) and astigmatism up to 6.00 D (mean, 1.10 D). Myopia and astigmatism were corrected by laser in situ keratomileusis (LASIK) with sequential spherical and cylindrical ablations using a 5.5-mm ablation zone and a transition zone to 7.0 mm. The Nidek EC-5000 excimer laser and the Chiron Automated Corneal Shaper were used.

Surgically induced refractive changes 3 months after surgery were measured. Spherical ablation by LASIK produced a 19% greater refractive change than that predicted by the Nidek PRK algorithm. The laser setting needed to achieve each 1.00 D of cylinder correction induced 1.30 D change in spherical equivalent refraction (0.80 D more than expected). Patient age and nonlinear effects had small but statistically significant influences on refractive outcome. Sex, left/right eye, and surgeon were not significant factors. Residual variations in spherical equivalent refraction included a 0.50-D constant plus 9% of the predicted spherical equivlanet change. Residual variations in the parallel and orthogonal components of cylinder correction were, respectively, 28% and 13% of the predicted cylinder change plus a 0.30-D constant component. The standard deviation of axis alignment error for cylinder ablation was 3.7 degrees.

Cylindrical ablation produced a spherical change that was larger than expected. To compensate for this, spherical ablation should be reduced in eyes with astigmatism. Variability in the correction of astigmatism was proportionally larger than that for spherical correction and was primarily due to magnitude rather than axis error.