Long-term comparison of full-bed deep anterior lamellar keratoplasty and penetrating keratoplasty in treating keratoconus

Abstract

Objective: To compare postoperative outcomes of full-bed deep anterior lamellar keratoplasty (DALK) with penetrating keratoplasty (PK) in treating keratoconus.

Methods: Seventy-five eyes of 64 patients who received full-bed DALK and 52 eyes of 51 patients who received PK between June 2000 and August 2010 were included in this retrospective study. Full-bed DALK was performed using Yao's hooking-detaching technique. PK was performed using a standard technique. Intraoperative and postoperative complications, visual acuity, rejection, graft survival, endothelial cell density, corneal sensation recovery, and re-innervation were compared between the two groups.

Results: A best correct visual acuity of 0.5 or better was achieved in 90.7% of eyes after full-bed DALK and in 92.3% of eyes after PK (P=0.75). By the fifth postoperative year, graft endothelial cell loss reached 34.6% in the PK group vs. 13.9% in the full-bed DALK group (P<0.001). There were no statistical differences in corneal sensitivity recovery or corneal re-innervation between the groups (P>0.05). Intraoperative microperforation occurred in seven out of 75 (9.3%) eyes with a temporally postoperative double anterior chamber in two eyes in the full-bed DALK group. Postoperative complications in the PK vs. the full-bed DALK groups respectively were: rejection (7.7% vs. 0%, P=0.015), high intraocular pressure (IOP) (46.2% vs. 1.3%, P<0.001), secondary glaucoma (9.6% vs. 0%, P=0.006), complicated cataract (19.2% vs. 0%, P<0.001), and wound dehiscence (9.6% vs. 0%, P=0.006).

Conclusions: Both full-bed DALK and PK can offer long-term satisfactory visual outcomes for keratoconus. Graft rejection, secondary glaucoma, complicated cataracts, and constant endothelial cell loss were observed in eyes only after PK.

Fig. 1

Demonstration of one-step full-bed deep anterior lamellar keratoplasty (DALK) The Descemet’s membrane (DM) was directly detached in full bed by viscoelastic injection through the initial DM pocket. (a) The DM was exposed as an initial pocket around the trephined margin at 12 o’clock. (b) Viscoelastic injection exactly between the stroma and DM creates a cyst-like elevation of the stroma without edema. (c) The stroma is removed from the DM exposing the full-bed DM. (d) Grafting a cryopreserved donor corneal button

Fig. 2

Demonstration of two-step full-bed deep anterior lamellar keratoplasty (DALK) Viscoelastic injection through the initial pocket does not go exactly between the stroma and the Descemet’s membrane (DM) but runs into the deep stroma layer. Complete exposure of the DM needs a second round of stromal hooking and viscoelastic injection. (a) Viscoelastic injection through the initial pocket creates stromal layer elevation with significant edema. (b) A second pocketing exposure of the DM was performed by hooking on the bed retaining tiny stroma after the first round of viscoelastic injection, visco-delamination, and removal of the stroma. (c) Viscoelastic injection through the second pocket creating thorough detachment of the retained stroma from the DM with a clear detached margin. (d) Removal of the retained stroma sheet around the trephined margin

Fig. 3

Overall rejection-free survival rates after full-bed deep anterior lamellar keratoplasty (DALK) and penetrating keratoplasty (PK) Graph shows overall rejection-free survival rate (Kaplan-Meier method) of 75 consecutive full-bed DALKs and 52 PKs. Rejection-free survival in the full-bed DALK group was significantly better than that in the PK group (P=0.02, log-rank test)

Fig. 4

Overall graft survival rates (Kaplan-Meier method) of 75 consecutive full-bed deep anterior lamellar keratoplasties (DALK) and 52 penetrating keratoplasties (PK) P=0.27, log-rank test

Fig. 5

Periodic changes in corneal endothelial cell density after full-bed deep anterior lamellar keratoplasty (DALK) and penetrating keratoplasty (PK) Six months after surgery, corneal endothelial cell density was significantly higher in the PK group than in the full-bed DALK group. However, from one to five years after surgery, cell density in the full-bed DALK group remained almost stable, while in the PK group it showed a continuous decline. Values are expressed as mean±standard deviation (SD)

Fig. 6

Best corrected visual acuity (BCVA) in the two groups at all postoperative follow-up visits Graph shows BCVA expressed as the logarithm of the minimum angle of resolution (logMAR) before and after surgery in 75 eyes with full-bed deep anterior lamellar keratoplasty (DALK) and 52 eyes with penetrating keratoplasty (PK). Values are expressed as mean±standard deviation (SD). ^* Statistical significance (P<0.01, Mann-Whitney U test)

Fig. 7

Recovery of central corneal sensitivity after full-bed deep anterior lamellar keratoplasty (DALK) and penetrating keratoplasty (PK) Values are expressed as mean±standard deviation (SD). No significant difference was found between the two groups at any time point

Fig. 8

Sub-basal nerve regeneration in representative cases after full-bed deep anterior lamellar keratoplasty (DALK) or penetrating keratoplasty (PK) Graph shows regeneration of sub-basal nerves after PK (a, b, c) or full-bed DALK (d, e, f). Similar corneal re-innervation was observed in the two groups