Deep anterior laminar keratoplasty (DALK) has inherent advantages over penetrating keratoplasty (PK), such as closed-eyeball surgery or respecting the endothelium of the recipient, therefore without the possibility of endothelial rejection and with minimal cell loss due to the surgery itself^1-6. This chapter attempts to summarize the extensive information available on the functional results, both overall and depending on the pathology and the different surgical techniques of DALK, as well as its comparison with PK.

RESULTS IN KERATOCONUS: COMPARISON WITH PENETRATING KERATOPLASTY

Keratoconus is perhaps the most frequent indication of DALK, so the study of its results is especially relevant. A meta-analysis that included 16 works with a total of 6625 eyes operated on for keratoconus⁷ (1185 DALK and 5440 PK) concluded that the best corrected visual acuity (BCVA) in the PK group was significantly better than in the DALK group – odds ratio (OR) = 0.48; confidence interval (95% CI) = 0.39-0.60; p <0.001. After 12 months, however, the differences in BCVA between both techniques were no longer statistically significant.

There were fewer cases of rejection in the DALK group than in the PK group (OR = 0.28, 95% CI = 0.15-0.50, p <0.001). However, the frequency of graft failure was similar in both (OR = 1.05, 95% CI = 0.81-1.36, p = 0.73). No significant differences were found in terms of spherical equivalent (SE) (p = 0.70), astigmatism (p = 0.14) and central corneal thickness (p = 0.58) between both groups. The endothelial cell count in the DALK group was better than in the PK group (p <0.001). In a 5-year follow-up study, it was demonstrated that the mean endothelial loss was 22.3% in the DALK, while it was 50.1% in the PK⁸. An extrapolation of these results in terms of viability of the transplanted tissue predicts a mean graft survival of 40 years in the DALK group versus 17 years in the PK group⁹.

Regarding postoperative complications such as high intraocular pressure (IOP) or cataract formation, fewer cases were found in the DALK group than in the PK group (high IOP, OR = 0.22, 95% CI = 0.11-0.44, p < 0.001, cataract: OR = 0.22, CI 95% = 0.08-0.61, p = 0.004). In this meta-analysis, no cases of expulsive hemorrhage or endophthalmitis were found⁷.

The most common cause of failure in DALK are opacities and scarring at the donor-recipient interface, while in PK it is endothelial rejection¹⁰. With DALK, an open-sky procedure should be avoided, so that glaucoma and cataract complications are reduced^11-12, and possibly other rarer, such as expulsive hemorrhage, although this is more difficult to prove. Although the risk of endothelial rejection is eliminated with DALK, stromal and epithelial rejections are similar in frequency to PK^13-14. These types of rejection are, however, milder and easier to control with topical corticosteroids^14-15. For this reason, they are generally used at lower doses in the postoperative period of the DALK, which would partly explain the lower risk of cataract and glaucoma. In any case, it should be kept in mind that these comparisons do not distinguish between different DALK techniques.

DESCEMETIC DALK (BB) COMPARED WITH PREDESCEMETIC (MANUAL) AND PK

For performing a DALK there are two main types of techniques: descemetic, such as the "big bubble" of Anwar (BB), among others, and predescemetic, such as Melles’ and others of manual dissection (see chapter 5.4). With the first one, the sheet formed by the Descemet membrane (DM) and the predescemetic layer of Dua (PDL) is dissected, while with the second one a thin layer of deep stroma remains on the bed. The latter can condition the visual results.

Borderie et al compare the results and complications of DALK by means of BB with the technique by manual dissection and with PK⁹. They studied 142 consecutive cases scheduled for PK and another 142 consecutive cases scheduled for DALK. These two groups were comparable (without significant differences) in terms of preoperative diagnosis – keratoconus, stromal dystrophies, scars after infectious keratitis and trauma – risk of rejection, IOP and glaucoma, size of the graft and the recipient window, and ages of the recipient and the donor. In all cases of DALK the BB was tried first, but it was only achieved in 46% (65 eyes); in the remaining 54% (77 eyes) a slit lamp-assisted manual dissection technique was used.

No differences were found in postoperative IOP between the PK group and the DALK group, nor between the subgroups of BB, manual and PK (p > 0.05). No differences were found in the refractive cylinder between PK and the overall group of DALK, nor between the 3 subgroups (p > 0.05). The endothelial cell loss was greater in the PK, as we already mentioned in the previous section.

The BCVA was clearly worse in the manual subgroup, both when compared with the BB (2.2 to 2.5 lines) and with the PK (1 to 1.8 lines). However, no differences were found when comparing the subgroup of BB with that of PK (p > 0.05). The mean thickness of the residual stroma – measured by OCT – was 87 ± 26 μm in the manual subgroup; no correlation was found between this data and the BCVA. There were also differences in final BCVA according to the preoperative diagnosis (p ≤ 0.001); in general, it was better in cases of low risk of rejection than in those at high risk. Contrast sensitivity was better in the BB group compared to the manual technique.

Complications such as episodes of rejection (epithelial, subepithelial, stromal, endothelial or mixed), ocular hypertension, infectious keratitis and other complications (miscellaneous) were less frequent in the DALK group than in the PK groups (Table 1).

PREDESCEMETIC DALK (MANUAL) COMPARED WITH PENETRATING KERATOPLASTY

In relation to the predescemetic technique of DALK, the results of the study by Borderie et al previously discussed may be biased by the fact that manual dissection was used only when the BB failed. Although it can be deduced from the literature that visual results are generally better in PK than in manual DALK, in experienced hands and with certain manual techniques it is possible to obtain visions similar to those of BB and PK – with the other advantages of the DALK –. This is indicated by some studies^2,3, and our experience with the Melles’ technique in 5 cases of ectasia after LASIK surgery, in which everyone improved their BCVA and only one patient did not reach the maximum BCVA prior to LASIK – by one line – (table 2)¹⁶.

In keratoconus, satisfactory results can also be obtained in comparison with PK, as indicated by a retrospective study of 41 eyes of 36 patients operated between 2002 and 2011 – 20 PK and 21 manual DALK (Melles) – in La Arruzafa Hospital of Córdoba¹⁷. In the PK group, 12 were made by conversion after failure of BB attempt; the manual technique did not require conversion under any circumstances. All of them presented an advanced keratoconus not susceptible to other treatments to improve their vision and the minimum follow-up was longer than one year, except in 2 PK eyes and 3 DALK eyes that were excluded because they did not reach it. Both groups were comparable in terms of age, sex and time of follow-up, which was on average 4.85 years for the PK group (range 1-10.5) and 4.21 years for the DALK group (range 1-7). The BCVA, the spherical refractive defect, the refractive cylinder and the SE were analyzed, measured once all the sutures were removed, as well as the complications (rejection, dehiscence of sutures, etc.) in each group.

The BCVA (Snellen decimal, mean ± ET) was 0.77 ± 0.32 in the PK group and 0.62 ± 0.29 for the DALK group. If we excluded from the second group 3 eyes that had a maximum potential VA of 0.3 for amblyopia, the average BCVA would be 0.67. The mean spherical refractive error in the PK group was -1.73 diopters (D) and the mean spherical equivalent of -3.92 D. The DALK group presented values ​​of -2.67 D and -4.55 D respectively. The residual cylinder, once the sutures were removed, was 4.47 ± 2.47 D in the PK group and 3.77 ± 1.63 D in the DALK group. No significant differences were found for any of these variables when comparing both groups by means of the Student's T test for independent samples, even when converting the BCVA to a logarithmic scale (logMAR).

There were three episodes of rejection in PK, one of which caused graft failure, while in the DALK group there was none. There were also three cases of suture dehiscence in the PK group and none in the DALK group. These differences did not reach significance (p = 0.06 assuming equal variances, and p = 0.08 not assuming equal variances), although the sample is small in this respect¹⁷.

In conclusion, the functional results in this series seem slightly better in the PK group (approximately 1 more line of BCVA and 1 D less of residual sphere, with 0.75 D more cylinder on average) without reaching statistical significance. Given this discrete difference, the general advantages of DALK and those specific to the predescemetic technique make it preferable, in our opinion, as the first option in cases of high risk of intraoperative perforation with BB technique: keratoconus with very high curvature, requiring large diameter grafts, with significant scars or a history of hydrops, as well as in cases where intra- and postoperative safety prevail, such as single eyes or patients with limited capacities (children, Down’s syndrome, etc.).

Although the heterogeneity in the DALK techniques makes it difficult to compare the functional results with those of the PK, it does not seem to be any differences in the case of the descemetic DALK (BB) and these are modest in the case of the predescemetic DALK, at least with some techniques like that of Melles. The essential advantages of DALK, such as greater safety due to closed eyeball surgery and no endothelium transplant, make it clearly preferable in terms of visual results and seem to compensate for the slight inferiority of manual techniques.

The main weakness of BB, its risk of rupture and therefore of conversion to PK, defines a field of indications for manual techniques in situations of greatest risk. Since the functional limitation of the latter is related to the presence of residual stroma in the recipient bed in front of the DM/PDL, it is expected that the incorporation of new technologies such as intraoperative OCT will allow to minimize the thickness of said residual stroma and thus approximate the results to those of the descemetic techniques and PK. In any case, every cornea surgeon must master each of the techniques available to offer the most appropriate option to each patient.