Discrepancy Between Preoperative and Intraoperative Central Corneal Epithelial Thickness During Alcohol-Assisted PRK

To evaluate the agreement between preoperative optical coherence tomography (OCT)-derived central corneal epithelial thickness (CET) and an intraoperative epithelial thickness estimate derived from optical coherence pachymetry (OCP) during alcohol-assisted photorefractive keratectomy (PRK).

A retrospective analysis was performed on 116 eyes of 58 consecutive patients who underwent bilateral myopic alcohol-assisted PRK between June 2020 and October 2021. Preoperative CET was assessed using anterior-segment OCT. The OCP-derived intraoperative epithelial thickness estimate was calculated as the difference between total corneal pachymetry readings obtained immediately before and after epithelial removal. Demographic, preoperative clinical, and refractive data, as well as intraoperative laser parameters, were collected.

Mean age was 26.31 ± 7.44 years (range, 18-45), and 63.8% of patients were female. Mean sphere and spherical equivalent were -3.49 ± 1.63D and -3.16 ± 1.63D, respectively. Mean preoperative OCT-derived CET was 53.2 ± 3.5  μm (range, 42-69  μm), whereas the mean OCP-derived intraoperative CET was significantly higher at 59.8 ± 14.8  μm (range, 21-99  μm; p < 0.0001). This discrepancy was consistent in both eyes of each patient. The intraoperative estimate exceeded the preoperative OCT-derived CET in 73.3% of eyes. Bland-Altman analysis demonstrated poor agreement between the two approaches, with wide limits of agreement from -15.7  μm to +34.8  μm. Preoperatively, 98.3% of eyes had CET values within the 40-60-μm range, compared with 44.8% of the OCP-derived intraoperative estimates. The difference between measurements significantly correlated with preoperative corneal thickness (r = 0.45, p < 0.001) and intraoperative CET (r = 0.97, p < 0.0001) but not refractive error or intraoperative and environmental parameters.

Preoperative OCT-derived CET and OCP-derived intraoperative CET demonstrate a systematic discrepancy and poor agreement during alcohol-assisted PRK. The intraoperative estimates were higher on average and substantially more variable than preoperative OCT measurements. These differences may have implications for surface ablation procedures, particularly transepithelial PRK, in which an assumed epithelial thickness profile is incorporated into treatment planning.