Purpose: To evaluate the mechanism(s) producing refractive instability and corneal haze development after photorefractive keratectomy (PRK).
Design: Prospective, nonrandomized, comparative case series, self-controlled.
Participants: Seventeen eyes of 17 patients with low- to moderate-grade myopia (-2.88 to -9.13 diopters [D]) were included.
Methods: Surgical intervention was a standardized, 6-mm diameter PRK procedure using the Meditec MEL 60 excimer laser (Aesculap-Meditec, Heroldsberg, Germany). The photoablation center was evaluated before surgery and at 1, 3, 6, 9, and 12 months after PRK using rapid, continuous z-scans of confocal images, termed confocal microscopy through focusing (CMTF).
Main outcome measures: Simultaneous epithelial and stromal thickness analysis and objective assessment of corneal light backscattering were obtained from digital image analysis of the CMTF scans. Corneal reinnervation and anterior stromal keratocyte density and wound healing morphologic features were evaluated on high resolution, in vivo confocal images. Manifest refraction was measured and corneal clarity was graded by slit-lamp biomicroscopy.
Results: Epithelial thickness averaged 45+/-10 microm at 1 month, 50+/-8 microm at 3 months, and 52+/-6 microm at 12 months after PRK, as compared with 51+/-4 microm before surgery, demonstrating complete restoration of the preoperative thickness without compensatory hyperplasia. Interestingly, epithelial rethickening had no significant correlation with refractive regression. By contrast, stromal regrowth (from 1-12 months) averaged 6+/-12 microm (range, 27 microm thinning-22 microm rethickening) and correlated closely (r = 0.84, P<0.001) with changes in refraction that averaged 0.84+/-1.23 D, ranging from -1.63 D (hyperopic shift) to +3.38 D (myopic regression). Stromal rethickening increased proportionally with the actual photoablation depth (r = 0.63, P<0.01); linear regression analysis suggested an average regrowth rate of 8% per year for the entire study group. Stromal rethickening was not associated with CMTF haze development over time, suggesting that haze and regression were caused by two independent wound healing mechanisms. In agreement with these findings, all "hazy" corneas showed increased numbers of anterior stromal wound healing keratocytes with increased reflectivity of both nuclei and cell bodies, suggesting that cellular-based reflections, as opposed to extracellular matrix deposition, are the major origin of increased corneal light scattering after PRK.
Conclusions: Taken together, these data indicate that keratocyte-mediated regrowth of the photoablated stroma appears to be the main cause of myopic regression in humans treated with a 6-mm diameter PRK, whereas hyperopic shifts appear to be a direct consequence of stromal thinning. By contrast, the corneal epithelium appeared to restore its preoperative thickness without contributing significantly to the refractive changes after PRK. Finally, this study also provides strong evidence that the development of haze after PRK is directly associated with increased cellular reflectivity from high numbers of wound healing keratocytes.