Inter-eye asymmetry of ocular surface parameters after penetrating and deep anterior lamellar keratoplasty versus endothelial keratoplasty

To compare postoperative ocular surface parameters between penetrating keratoplasty/deep anterior lamellar keratoplasty (PK/DALK) and endothelial keratoplasty (EK), with separate evaluation of operated and fellow eyes and assessment of inter-eye asymmetry. This single-center retrospective cross-sectional study included unilateral keratoplasty patients who underwent standardized ocular surface evaluation during routine follow-up at least 1 year after surgery. Patients were clinically stable, had discontinued topical corticosteroids for at least 3 months, and had no graft rejection episode or suspected inflammatory reaction within the preceding 3 months. Ocular surface evaluation included Schirmer testing, non-invasive tear break-up time (NIBUT), tear meniscus height (TMH), ocular surface staining, tear matrix metalloproteinase-9 (MMP-9), meibomian gland loss, and corneal sensitivity. Between-group comparisons were performed separately for operated eyes and fellow eyes, followed by within-group paired comparisons, delta-based inter-eye asymmetry analysis, and mixed-effects modeling. Each group comprised 26 patients, each contributing one operated eye and one fellow eye. The EK group was older and had a shorter postoperative interval than the PK/DALK group. In operated eyes, corneal sensitivity was lower in the PK/DALK group than in the EK group after adjustment for age and log-transformed postoperative duration, whereas TMH showed only a borderline difference. In fellow eyes, only TMH differed significantly between groups. Within-group analyses showed that the EK group demonstrated a significant operated-to-fellow eye difference only in corneal sensitivity, whereas the PK/DALK group showed higher tear MMP-9 grade, shorter NIBUT, and lower corneal sensitivity in operated eyes. Delta analysis showed greater inter-eye asymmetry in the PK/DALK group for NIBUT ( P = 0.049) and corneal sensitivity ( P = 0.001). Mixed-effects models showed significant keratoplasty category-by-eye interactions for NIBUT ( P = 0.028) and corneal sensitivity ( P < 0.001), indicating greater operated-to-fellow eye disparity in the PK/DALK group. Compared with EK, PK/DALK was associated with greater inter-eye asymmetry in corneal sensitivity and tear film stability in the late postoperative period. These findings suggest that postoperative ocular surface assessment after keratoplasty may benefit from evaluation of both the operated and fellow eyes, with corneal sensitivity serving as a particularly informative marker.