OBJECTIVE To assess the clinical characteristics, incidence, and pathologic correlation of corneal epithelial adhesion abnormalities encountered during LASIK. METHODS Prospective noncomparative interventional case series. METHODS Five hundred consecutive eyes of 268 patients undergoing primary LASIK procedures by one surgeon utilizing the Moria LSK One microkeratome and VISX Star S-2 excimer laser. METHODS Corneal epithelial adhesion was assessed immediately preoperatively using a cellulose surgical sponge (adhesion test), and the incidence, extent, and location of epithelial defects occurring during the microkeratome incision of the corneal flap were recorded. Epithelial specimens from 7 corneas requiring debridement of dysadhesive epithelium were examined by transmission electron microscopy. METHODS The characteristics of the study population (age, gender, contact lens use, relevant ocular surface or systemic disease, refractive error, keratometry, pachymetry) and the microkeratome-related variables (head selection and vacuum level) were compared with the results of the preoperative epithelial adhesion test, the development of intraoperative epithelial effects (size and location), and the postoperative outcome. RESULTS Epithelial defects were sustained by 51 corneas (10.2%), and among these, 31 (6.2%) were microdefects and 20 (4.0%) were macrodefects. The adhesion test was positive (indicative of compromised adhesion of epithelium to stroma) in 20 (64.5%) corneas with microdefects, but false negatives (epithelial defect despite negative adhesion test) occurred in 11 cases (35.4%). The adhesion test was positive in 16 (80%) of corneas having macrodefects, with 4 (20%) false negatives. Thus, the overall positive predictive value of the adhesion test was 59%, and the percentage of positive prediction was 92% (Bayes' theorem). Among all other outcome measures assessed, only corneal flap thickness seemed a potential risk factor, as 40 (78.4%) epithelial defects were associated with the creation of a 180- micro m-thick flap, whereas 10 (19.6%) were associated with a 160- micro m-thick flap and only 1 (2%) occurred with a 130- micro m flap. These trends were not, however, statistically significant (P = 0.15, Fisher exact test). Transmission electron microscopy of all epithelial debridement specimens disclosed consistent abnormality of the basement membrane adhesion complex, as thickened and multilaminated basement membrane remained adherent to the intact epithelial sheet. CONCLUSIONS : Corneal epithelial dysadhesion and defects occurring in the course of LASIK surgery may be associated with an intrinsic compromise of the basement membrane adhesion complex, as evidenced clinically by the adhesion test and demonstrated pathologically by duplicated basement membrane.