Deep Anterior Lamellar Keratoplasty Versus Penetrating Keratoplasty in Keratoconus: A Retrospective Study Using Coarsened Exact Matching.

PURPOSE: To compare the graft survival rates and visual outcomes in patients with keratoconus (KCN) undergoing deep anterior lamellar keratoplasty (DALK) versus penetrating keratoplasty (PKP) and identify possible risk factors.

METHODS

This retrospective longitudinal study enrolled patients with KCN who underwent corneal transplantation at Labbafinejad Medical Center (Tehran, Iran) between 2006 and 2016. We utilized a one-to-one coarsened exact matching (CEM) approach. Matching considered the recipient's sex, donor age, and the presence and extent of corneal neovascularization in the recipient.

RESULTS

A total of 213 patients with KCN who underwent PKP (49.8%; n = 106) or DALK (50.2%; n = 107) were enrolled. After CEM was performed, 67 patients in the PKP group were well-matched with 67 patients in the DALK group. There was no significant difference in terms of baseline characteristics between the two groups. Kaplan-Meier survival analysis showed that the percentage of patients free of graft rejection in a one- and three-year follow-up was 78.7% and 74.5% in the PKP group, and 95.3% and 95.3% in the DALK group, respectively. On univariate analysis, the strongest risk factor predictive of graft rejection was corneal vascularization in both the PKP and DALK groups, followed by recipient age only in the PKP group and dry eye disease only in the DALK group. In a multivariate Cox regression analysis, PKP was identified as an independent predictor of graft rejection, but not of graft failure. The DALK and PKP groups were comparable in terms of postoperative best-corrected visual acuity (BCVA; P = 0.48) and suture complications (P = 0.87).

CONCLUSION

In patients with KCN, DALK demonstrated comparable outcomes to PKP in terms of BCVA, but showed superiority in graft rejection-free survival. Corneal vascularization was the major recipient risk factor for graft rejection and failure.