Objectives: We compared long-term endothelial cell survival after penetrating versus after deep anterior lamellar keratoplasty for keratoconus.
Materials and Methods: We retrospectively compared 64 eyes of 55 patients who had penetrating keratoplasty and 40 eyes of 37 patients who had deep anterior lamellar keratoplasty for keratoconus (October 2003-February 2021). Best-corrected visual acuity, Goldmann applanation tonometry, fundus examination with 90D lens, and specular microscopy with CEM-530 (Nidek) were performed preoperatively and every 6 months postoperatively. Main outcomes were endothelial cell density, central corneal thickness, and visual acuity. Secondary outcomes were coefficient of variation, hexagonality, graft rejection episodes, and graft clarity.
Results: We found no significant differences between the 2 treatment groups regarding patient age, donor age, preoperative vision, central corneal thickness, and recipient-donor trephine diameters. Mean follow-up was 92.5 months. In deep anterior lamellar keratoplasty, the endothelium was preserved significantly better for 10 years versus for penetrating keratoplasty. Mean endothelial density in penetrating versus deep anterior lamellar keratoplasty was 2006.7 versus 2354.7 cells/mm² at 1 year (P = .010), 1170.5 versus 2048.2 at 5 years (P < .001), and 972.5 versus 1831.6 at 10 years (P < .001). Cumulative endothelial cell loss was 43% and 19.7% at 10 years for penetrating and anterior lamellar keratoplasty, respectively. Significantly more thickening of central cornea was shown in penetrating keratoplasty after 7 years. Corneal thickness was 583.0 µm in penetrating and 545.1 µm in deep anterior lamellar keratoplasty (P = .002) at 10 years. Vision gain and coefficient of variation were similar. Hexagonality decreased significantly in both groups at 10 years. Rates of rejection were 12.5% in penetrating and 7.5% in deep anterior lamellar keratoplasty. Graft survival rates were 97.5% and 96.9%, respectively.
Conclusions: In keratoconus, endothelial vitality is better preserved with deep anterior lamellar keratoplasty than with penetrating keratoplasty over a 10-year follow-up.

Key words : Corneal endothelium, Corneal surgery, Corneal transplantation, Endothelial cell survival, Lamellar keratoplasty

Introduction

Keratoconus is an ectatic corneal disease that mostly affects young people. Mild cases are typically treated with spectacles or contact lenses, whereas advanced cases that cannot be managed with contact lenses may require corneal surgery.¹ Penetrating keratoplasty (PK) has been the only surgical method for the treatment of keratoconus for several decades.² However, after Anwar and Teichmann described the big-bubble technique in 2002, deep anterior lamellar keratoplasty (DALK) has become more popular, since this technique is fast, safe, and easier to perform than previous lamellar keratoplasty methods.^3,4 In DALK surgery, the recipient’s endothelium is left intact together with the Descemet membrane and the Dua layer if a type I bubble occurs. Because endothelial cells are protected together with their microen-vironment in DALK, the lifespan of cells is longer than in PK. Clinical studies revealed a higher endothelial cell survival 2 to 4 years after DALK compared with after PK.^5,6 In addition, DALK eliminates the risk of endothelial rejection and consequently decreases the failure rate of corneal graft.² The median predicted graft survival has been reported as 49.0 years in DALK and 17.³ years in PK.⁷ Longer graft survival is an important advantage in young patients with keratoconus who often have surgery under 30 years of age and have a longer life expectancy. In this study, we investigated the long-term endothelial cell viability after DALK and PK surgeries performed on patients with keratoconus.

Materials and Methods

This retrospective, comparative interventional study included 64 eyes of 55 patients who had PK and 40 eyes of 37 patients who had DALK for keratoconus, between October 2003 and February 2021. This study was conducted under the tenets of the Declaration of Helsinki. Ethical approval for the study was obtained from the Institutional Board, Research Ethics Committee of the University of Health Sciences, Izmir Bozyaka Education and Research Hospital (February 22, 2023, reference 2023/25). Written informed consent was obtained from all participants included in the study.

Best-corrected visual acuity (BCVA), Goldmann applanation tonometry, fundus examination with a 90D lens, and noncontact specular microscopy with CEM-530(Nidek) were performed preoperatively and every 6 months for the first 3 postoperative years and yearly thereafter. If the number of endothelial cells evaluated by the specular microscope was under 10, the measurement was not considered valid.

Donor cornea endothelial data, including preoperative endothelial cell density (ECD), coefficient of variation of the cell area, and hexagonality, for the PK group were measured with an eye bank specular microscope; recipient endothelial data were measured with a clinical specular microscopy for the DALK group; ECD, coefficient of variation of the cell area, hexagonality, and central corneal thickness (CCT) values were also noted.

Endothelial photographs were also checked for quality and abnormal cell morphology. Donor corneas were stored in a corneal storage medium at 4 °C in our eye bank facility. Donor endothelia were photographed and analyzed by a trained eye bank technician with an eye bank specular microscope (Eye Bank Keratoanalyzer, EKA-04, Konan Medical).

Inclusion criteria were an uneventful surgery except for minor complications, such as microper-foration in DALK, and at least 24 months of follow up. In both groups, patients who underwent arcuate keratotomy to correct postoperative astigmatism were included in the study. Because arcuate keratotomy was applied to the periphery of the graft, it has no effect on the central endothelial cell density. Patients from both groups missed follow-up appointments during the full shutdown periods during the COVID-19 pandemic. These patients were called by phone to allow routine checks to continue, and these patients were also included in the study. Exclusion criteria were cataract extraction during or after keratoplasty, serious penetrating ocular trauma after surgery, severe endothelial or stromal rejection episode causing significant endothelial loss leading to graft failure or regrafting and manual DALK cases with interface opacity that prevented measurement. Two patients with mental retardation were also excluded from the study because no measurements could be made.

All surgeries were performed by a single experienced surgeon (BY). The PK surgeries were performed with the use of disposable vacuum trephine. Donor corneas were cut from the endothelial side by using a disposable donor punch. For PK surgery, the surgeon used 8-mm recipient vacuum trephine in 38eyes (59%) and 7.5-mm trephine in 26eyes (41%). For DALK surgeries, these numbers were 24 (60%) and 16 (40%), respectively. A 0.25-mm larger donor trephine was used for both of the keratoplasty techniques. Sixteen to 20 bite 10/0 nylon running suture was used for donor cornea fixation.

An Anwar big-bubble technique was used for DALK, with big bubble achieved in 37 of 40 eyes. Manual DALK was performed in the remaining 3 eyes. Microperforation occurred in 3 eyes. Cases of DALK that were converted to PK due to macroperforation were included in the PK group. In PK, prednisolone acetate drops were tapered after 2 months and discontinued at 6 months; afterward, fluorometholone drops twice per day were used until completion of 1 year. Loteprednol acetate drops were continued for several years in certain cases. Steroids were discontinued earlier in DALK cases. Discontinuation of steroids was decided according to the condition of the eye for both methods.

The main outcome measures were ECD, CCT and BCVA. Secondary outcome measures were coefficient of variation of the cell area, hexagonality, graft rejection episodes, and graft survival. In both groups, the final outcome analysis was performed after complete suture removal.

Statistical analyses

We used SPSS version 21.0 (SPSS Inc) for statistical analyses. Data obtained by taking the average of all measurements were recorded as means ±SD. We used repeated-measures analysis of variance for multiple comparisons and paired t test for pairwise comparisons. We used independent t test for comparisons between groups. P<.05 was considered significant.

Results

In the PK group, 34 patients (62%) were male and 21 (38%) were female. In the DALK group, 11 patients(30%)were male and 26 (70%) were female. Mean follow-up period for all eyes included in the study was 92.5±58.2 months (range, 24-294 mo). The number of eyes with 5-year follow-up was 59 (92%) in PK group and 37 (93%) in the DALK group. The number of eyes with 10-year follow-up was 42 (66%) and 26 (65%), respectively. Characteristics of patients in the PK and DALK groups are summarized in Table 1. As preoperative ECD, donor ECDs were given in the PK group and recipient ECDs were given in the DALK group. We observed no significant differences between the 2 groups in terms of patient age, donor age, preoperative BCVA,CCT, coefficient of variation of the cell area, intraocular pressure, and recipient/donor trephine diameters.

Preoperative specular microscopy revealed an abnormal endothelial morphology in 2 patients in the DALK group. Abnormally large cells in the form of cobblestones that were 4 to 5 times larger than the other cells were observed in an area surrounded by healthy cells. One of the eyes had endothelial count of 1569cells/mm² and 48% coefficient of variation of the cell area; the other had an endothelial count of 2047 cells/mm² and 50% coefficient of variation. Because these 2 eyes had sufficient endothelial density for corneal transparency, they were included in the study and statistical analysis. In these 2 eyes, the cornea was transparent during follow-up, and no graft failure developed.

Sutures were removed significantly earlier in patients in the DALK group (P < .001). Mean suture removal time was 12.4 ± 1.5 months (range, 7-19 mo) in the PK group and 9.1±2.9 months (range, 3-13 mo) in the DALK group. The duration of steroid use was also significantly shorter in the DALK group compared with the PK group, with duration of 37.9±36.2 months (range, 10-180 mo) in the PK group and 22.7 ± 21.2 months (range, 3-90 mo) in the DALK group (P = .013).

Mean ECDs by year are shown in Table 2. In the DALK group, the endothelium was preserved signi-ficantly better for 10 years, starting from the first year, compared with the PK group. Of note, differences between the 2 methods gradually widened during the first 3 years. When year 1 ECD was compared with preoperative value, a significant loss was observed in both the PK (P < .001) and DALK (P = .001) groups due to operative trauma. At 1 year, ECD was >2000 in 76.9% of PK group eyes and 90.1% of DALK group eyes. When year 2 ECD was compared with year 1 ECD, we observed a significant decrease in PK again (P < .001) but not in DALK (P = .386). A comparison of year 3 versus year 1 also revealed significant endot-helial loss in the PK group (P < .001) but not in the DALK group (P = .242). In the PK group, significant decreases in ECD continued through years 2 to 10 compared with year 1 ECD (P < .001); however, these decreases only became significant as late as year 4 in the DALK group (Figure 1).Cumulative endothelial cell loss was 43% in the PK group and 19.7% in the DALK group at 10 years. Few patients completed late follow-ups extending 16 years. Mean ECD at 12 years was 868.4 cells/mm² in the PK group and 1754.0 cells/mm² in the DALK group, with 846.5 at 14 years and 821 cells/mm2 at 16 years in PK. These numbers were not considered in our statistical analyses.

Significant CCT increases results were shown in both groups at year 1 because of the transplanted tissue (P < .001) (Table 3). When we compared CCT for each year versus the previous year, annual changes in CCT over 10 years were not significant in both groups. When CCT values of PK and DALK were compared year by year, no significant differences were found between groups over 6 postoperative years. However, starting with year 7, corneas in the PK group were significantly thicker than in the DALK group (Figure 2). At 5 years, increase in CCT compared with year 1 was 18.5 μm (+3.4%) in the PK group and 13.1 μm (+2.4%) in the DALK group. At 10 years, it was 40.2 μm (+7.4%) in PK group and 4.4 μm (+0.8%) in the DALK group. At year 5 compared with year 1, CCT significantly increased in the PK group (P = .040) but not in the DALK group (P = .627). At year 10 compared with year 1, CCT also significantly increased in the PK group (P < .001) but not in the DALK group (P = .187).

Significant visual improvement, as shown by BCVA, was achieved at year 1 compared with preoperative BCVA in both groups (P < .001). At year 1, BCVA ≥20/100 was achieved by 88.2% in the PK and 100% in the DALK group, with BCVA ≥20/40 achieved by 40.1% and 51.3%, respectively. However, annual comparisons of each year with the previous year revealed no significant changes in the groups. When logMAR BCVA values between 1 and 10 years were compared year by year, no significant differences were observed between the PK and DALK groups. LogMAR BCVA at postoperative year 1 was 0.48 ± 0.24 (range, 0.1-1.0) in the PK group and 0.43 ± 0.25 (range, 0.1-1.0) in the DALK group (P =.464), with results of 0.38 ± 0.18 (0.1-0.7) and 0.33 ± 0.26 (0.1-1.0) at 10 years, respectively (P = .374).

In comparisons between the PK and DALK groups, no significant differences were found in terms of coefficient of variation of the cell area in any year between years 1 and 10. In terms of polymegethism, the 2 groups were thus similar. When each year’s coefficient of variation of the cell area was compared with the previous year, annual changes over 10 years were also not significant between groups. Mean coefficient of variation of the cell area at 1 year was 29.9 ± 10.1 (range, 17-56) in the PK group and 28.9 ± 7.5 (range, 21-48) in the DALK group (P = .686), with results of 31.5 ± 12.2 (range, 17-70) versus 30.7 ± 8.8 (range, 20-54) at 5 years (P = .786) and 35.2 ± 10.8 (range, 22-71) versus 32.2 ± 9.7 (range, 19-54) at 10 years (P = .297). Compared with year 1, increases in coefficient of variation of the cell area at 5 and 10 years were not significant in both groups.

Preoperative hexagonality in the DALK group was significantly better than in the PK group because it was untouched recipient endothelium. Preoperative hexagonality was 55.2 ± 10.7 (range, 35-85) in the PK group and 65.9 ± 9.4 (range, 45-80) in the DALK group (P = .001). However, differences were no longer shown at 1 year. Hexagonality was 69.5 ± 4.5 (range, 60-78) in the PK group and 69.5 ± 8.0 (range, 54-80) in the DALK group (P = .983) at 1 year. Mean hexagonality was 66.9 ± 7.6 (range, 53-80) in the PK group and 67.9 ± 9.9 (range, 48-80) in the DALK group at 5 years (P = .491). No further statistically significant difference was seen between the groups in the following years. Hexagonality values at 10 years were 61.3 ± 7.0 (range, 46-80) in the PK group and 64.7 ± 5.7 (range, 55-77) in the DALK group (P = .061). When these values were compared with year 1 values, a significant decrease was found in both the PK group (P < .001) and the DALK group (P = .013), indicating a decrease in the number of hexagonal cells and an increase in polymorphism at 10 years.

Three stromal rejection episodes (3/40; 7.5%) occurred in the DALK group. Type of rejection episodes included stromal, stromal and subepithelial infiltrates, and epithelial. One of the stromal rejections resulted in graft failure (1/40; 2.5%) at year 1 because of patient stopping medication. The patient did not accept regrafting. The other 2 eyes were reversed with medication. Eight rejection episodes (8/64; 12.5%) occurred in the PK group. Types of rejection episodes included endothelial, stromal, subepithelial infiltrates, and epithelial. Two of them resulted in graft failure (2/64; 3%) at postoperative years 5 and 23. Regrafting (3%) was performed on these patients. Graft survival rate within the study period was 97.5% in the DALK group and 96.9% in the PK group.

Early suture loosening was observed in 7 of 40 (17.5%) of eyes in the DALK group and 2 of 64 eyes (3%) in the PK group. Increased intraocular pressure was observed in 4 eyes (6.3%) in the PK group and 2 eyes (5%) in the DALK group, which were successfully treated with topical medication. There was no significant difference between the groups in terms of increased intraocular pressure. Mean intraocular pressure at last visit was 15.1 ± 2.7 mm Hg (range, 10-22) in the PK group and 15.2 ± 4.5 mm Hg (range, 10-38) in the DALK group (P = .857). High astigmatism over 6D developed in 18 eyes in the PK group (28.1%) and 5 eyes (12.5%) in the DALK group.

Discussion

Our study showed that, in a period of >10 years, endothelial densities in both DALK and PK surgeries were above the minimum value required for corneal transparency.⁸ However, in DALK, healthy recipient corneal endothelium, which has higher ECD and hexagonality values preoperatively than the allograft corneal endothelium in PK, was better preserved for 10 years after surgery. Preoperative ECD was higher in DALK. DALK does not have the disadvantages associated with the use of donor endothelium in PK, such as donor endothelial cell loss during waiting period from death to retrieval and during corneal excision and storage. In addition, as in our study, the recipient endothelium was younger in DALK (32.0 years) than in donor corneas in PK (46.8 years).

In the postoperative period, we showed a prog-ressive loss of endothelium in PK, especially starting from year 2. Likewise, Khattak and colleagues reported significantly lower ECD in PK (1795 cells/mm²) compared with DALK (2250 cells/mm2) at 2 years.⁵ These values were 1702 and 2289 cells/mm², respectively, in our study. Cumulative endothelial loss was -29% in PK and -13.5% in DALK at 2 years. Kubalo?lu and colleagues reported that the difference in endothelial losses became significant starting from year 1 and continued until year 4, with reported mean ECD at 4 years of 1476 cells/mm2 in PK and 2250 cells/mm2 in DALK and loss rates of -47.8% and -21.6%, respectively.⁶ In our study, the corresponding results were 1372.8 (-43%) and 2012.8 (-24%), respectively. The rate of loss slowed after 5 years in both groups.

Van Dooren and colleagues reported endothelial cell loss after DALK: -283 at 6 months, -335 at 1 year, and -421 cells/mm² at 2 years. When each time was compared with its previous measurement point, loss rates were 11%, 2%, and 1.2%, respectively.⁹ In their 6-year study, Romano and colleagues reported that postoperative ECD did not significantly vary from preoperative values in DALK.¹⁰ In a series that included keratoconus and stromal dystrophies, the average 5-year endothelial cell loss was -50.1% in PK and -22.3% in DALK. The early-phase annual rate of loss was -15.2% and -8.3%, and the late-phase loss was -7.8% and -3.9% per year, respectively.⁷ Similarly in our study, the average 5-year postoperative endothelial cell loss was -51.1% in PK and -22.6% in DALK. Early-phase annual loss in the first 5 years was -13.2% in PK and -3.7% in DALK. Late-phase loss between year 6 and year 10 was -3.6% and -2.1%, respectively. These data demonstrate that DALK is significantly superior to PK in terms of long-term endothelial viability.^10,11

Central corneal thickness is a good indicator of endothelial density and pump function. We observed no significant differences in CCT between the PK and DALK eyes over a 6-year postoperative period. During this period, ECD remained sufficient for the pump function in both groups. Borderie and colleagues reported significantly thinner corneas in DALK at 1 year. Mean CCT at 12 months was 536 μm in PK and 523 μm in DALK.⁷ However, in our study, CCT was similar at 1 year (542.8 μm in PK and 540.7 μm in DALK). However, after year 6, the corneas in our PK group became thicker, with ECD down to 1000 cells/mm² in the PK group and above 1800 cells/mm2 in the DALK group. In fact, in some eyes in the PK group, it was as low as 500 to 600 cells/mm². These data suggest that subclinical edema started after year 6 in the PK group. Kubaloglu reported even earlier corneal thickening (at 2-4 years) after PK (+15%) compared with DALK (+8%).⁶ In our study, the difference becomes more pronounced at 10 years, with PK corneas significantly thicker (ie, edematous). At 10 years, CCT increased 7.4% (+40.2 μm) in the PK group compared with year 1, but only by 0.8% (+4.4 μm) in the DALK group.

In parallel with that shown in previous reports,^5,12,13 both groups in our study had equal visual improvement, as shown by similar graft transparency rate. At postoperative year 1, logMAR BCVA was 0.48 in PK and 0.43 in DALK. This trend did not change until year 10. Likewise, Khattak and colleagues reported similar visual results with PK and DALK (0.28 vs 0.25 logMAR).⁵ Donoso and colleagues showed a result of 0.17 logMAR for both methods.¹² Irregular astigmatism was the most important factor reducing vision. Late ectasia was another problem in some cases. Borderie and colleagues reported better visual acuity with big-bubble DALK compared with PK.⁷ However, Feizi and colleagues reported 0.18 for PK and 0.26 for DALK, showing thus a significantly higher vision gain with PK.¹⁴ Abdelaal and colleagues reported a similar rate of BCVA ≥20/100 with PK and DALK for keratoconus.¹³ In our study, these rates at year 1 were 88.2% in PK and 100% in DALK.

Polymegethism and pleomorphism are early indicators of corneal endothelial stress compared withECD.¹⁵ By definition, the coefficient of variation of the cell area is the ratio of the standard deviation of cell area to the mean cell area. It is an indicator of endothelial cell polymegethism. Its normal value is ≤30%, with the value increasing after any trauma or toxicity to the corneal endothelium. In our study, the coefficient of variation of the cell area showed a continuous increase in both groups for 10 years, reaching 45 in PK and 34 in DALK with no significant difference between the 2 groups. Kubalo?lu and colleagues reported a significantly higher coefficient of variation of the cell area in PK than DALK at 4 years (0.27 vs 0.24). Mean increase compared with preoperative value was +41.6% in PK and +20.6% in DALK.⁶ In our study, the mean coefficient of variation of the cell area was 31.2 in PK and 30.5 in DALK at 4 years, with rates of increase compared with preoperative values of 10% in PK and 9% in DALK. These results show that the size differences of endothelial cells in both groups gradually increased over time.

Hexagonality is the percentage of endothelial cells with 6 sides. Its normal value is 60% to 70%. In our study, preoperative endothelial hexagonality was higher in DALK eyes because they are in their natural environment at body temperature, whereas PK donor endothelium is in storage medium at 4 °C. In addition, factors mentioned earlier for donor corneas in PK play a role. In the first year, the difference disappeared since the endothelium of both groups were in their natural environment. However, when year 1 and year 10 hexagonality values were compared, a significant decrease was observed in both groups. These findings indicate that the number of hexagonal cells has decreased and the endothelial morphology has deteriorated in both groups over time.

Generally, histopathological changes occur in the epithelium, Bowman, and stroma in keratoconus, whereas the Descemet membrane and endothelium remain intact.¹⁶ However, in our study, endothelial cell anomalies were observed in 2 of 40 DALK eyes (5%) in parallel with previous reports.^17-19 The most common endothelial dystrophy associated with keratoconus is Fuchs dystrophy.¹⁸ However, the endothelial abnormality seen in our cases did not resemble Fuchs or posterior polymorphous dystrophy. Abnormality was associated with low ECD and high coefficient of variation values. Thus, specular microscopy is recommended to detect such endothelial abnormalities before DALK surgery. If necessary, shots of the other eye can also be used for this purpose.

Earlier suture removal provides faster visual rehabilitation in DALK.¹ In addition, shorter duration of steroid use can result in less cataracts and glaucoma. The mean duration of topical steroid use has been reported as 11.9 months in PK and 7.8 months in DALK.⁶ Our study had longer periods since our center uses weak steroids like fluoro-metholone and loteprednol for longer period. Suture-related complications have been reported more frequently in DALK than in PK.^5,14 Likewise, in our study, early suture loosening was more common in the DALK group (17.5%) than in the PK group (3%). In DALK, there is more superficial and longer placement of sutures to avoid Descemet perforation and to press the donor stroma downward against the Descemet, which may cause premature loosening of sutures.

In our study, rejection episodes occurred more frequently in PK (12.5%) than in DALK (7.5%). Similarly, Kubalo?lu and colleagues reported rates of graft rejection of 18% in PK and 4% in DALK,6 and Feizi and colleagues reported these rates as 33.5% and 19.7%.¹⁴ The absence of endothelial rejection in DALK plays a role in these outcomes. These episodes were detected early and quickly treated with intravenous and topical steroids to preserve the transparency of the graft. As a result, in our study, the graft survival rate was 96.9% in the PK group and 97.5% in the DALK group. Feizi and colleagues reported graft survival rates of 98.2% for PK and 94.8% for DALK at 6 years.¹⁴ Romano and colleagues reported that DALK provided a lower risk of graft rejection and longer endothelial survival compared with PK.¹⁰ For DALK, 1-, 3-, and 5-year graft survival rates were reported to be 100%, 93%, and 90%.20 Although less graft rejection episodes occur in DALK, graft failure rate was reported to be the same as in PK.^14,21 Kubalo?lu and colleagues reported a regrafting rate of 6.7% in PK,6 which was 3% in our study. Abdelaal and colleagues reported that less rigorous postoperative steroid use and absence of endothelial rejection render DALK to be the procedure of choice for corneal pathologies with intact endothelium.¹³

Guan and colleagues suggested that it was difficult to conduct a long-term follow-up with a big sample size since some patients were lost to follow-up.²⁰ Comparative 2-, 4-, and 5-year studies revealed higher endothelial cell survival in DALK compared with PK in keratoconus.^5-7 Given the scarcity of long-term studies comparing the endothelial survival of PK and DALK in keratoconus, our 10-year study may be considered a contribution to the literature. However, limitations included a relatively small number of patients and no randomization of case selection. There were disconnections in patient follow-up because of the COVID-19 pandemic, restricting data collection in long-term follow-up.

In conclusion, our study, together with the literature data,^{6,10,12,13,22} showed that DALK is superior to PK in terms of long-term endothelial cell viability. These results suggest DALK to be a superior option in young keratoconus patients with a longer life expectancy.

References:

1. Santodomingo-Rubido J, Carracedo G, Suzaki A, Villa-Collar C, Vincent SJ, Wolffsohn JS. Keratoconus: an updated review. Cont Lens Anterior Eye. 2022;45(3):101559. doi:10.1016/j.clae.2021.101559
   CrossRef - PubMed
   
2. Gonzalez A, Price MO, Feng MT, Lee C, Arbelaez JG, Price FW, Jr. Immunologic rejection episodes after deep anterior lamellar keratoplasty: incidence and risk factors. Cornea. 2017;36(9):1076-1082. doi:10.1097/ICO.0000000000001223
   CrossRef - PubMed
   
3. Anwar M, Teichmann KD. Big-bubble technique to bare Descemet's membrane in anterior lamellar keratoplasty. J Cataract Refract Surg. 2002;28(3):398-403. doi:10.1016/s0886-3350(01)01181-6
   CrossRef - PubMed
   
4. Sarnicola V, Toro P, Gentile D, Hannush SB. Descemetic DALK and predescemetic DALK: outcomes in 236 cases of keratoconus. Cornea. 2010;29(1):53-59. doi:10.1097/ICO.0b013e3181a31aea
   CrossRef - PubMed
   
5. Khattak A, Nakhli FR, Al-Arfaj KM, Cheema AA. Comparison of outcomes and complications of deep anterior lamellar keratoplasty and penetrating keratoplasty performed in a large group of patients with keratoconus. Int Ophthalmol. 2018;38(3):985-992. doi:10.1007/s10792-017-0548-9
   CrossRef - PubMed
   
6. Kubaloglu A, Koytak A, Sari ES, Akyol S, Kurnaz E, Ozerturk Y. Corneal endothelium after deep anterior lamellar keratoplasty and penetrating keratoplasty for keratoconus: a four-year comparative study. Indian J Ophthalmol. 2012;60(1):35-40. doi:10.4103/0301-4738.90490
   CrossRef - PubMed
   
7. Borderie VM, Sandali O, Bullet J, Gaujoux T, Touzeau O, Laroche L. Long-term results of deep anterior lamellar versus penetrating keratoplasty. Ophthalmology. 2012;119(2):249-255. doi:10.1016/j.ophtha.2011.07.057
   CrossRef - PubMed
   
8. Lass JH, Sugar A, Benetz BA, et al. Endothelial cell density to predict endothelial graft failure after penetrating keratoplasty. Arch Ophthalmol. 2010;128(1):63-69. doi:10.1001/archophthalmol.2010.128.63
   CrossRef - PubMed
   
9. van Dooren BT, Mulder PG, Nieuwendaal CP, Beekhuis WH, Melles GR. Endothelial cell density after deep anterior lamellar keratoplasty (Melles technique). Am J Ophthalmol. 2004;137(3):397-400. doi:10.1016/j.ajo.2003.09.053
   CrossRef - PubMed
   
10. Romano V, Iovieno A, Parente G, Soldani AM, Fontana L. Long-term clinical outcomes of deep anterior lamellar keratoplasty in patients with keratoconus. Am J Ophthalmol. 2015;159(3):505-511. doi:10.1016/j.ajo.2014.11.033
    CrossRef - PubMed
    
11. Kubaloglu A, Sari ES, Unal M, et al. Long-term results of deep anterior lamellar keratoplasty for the treatment of keratoconus. Am J Ophthalmol. 2011;151(5):760-767.e1. doi:10.1016/j.ajo.2010.11.020
    CrossRef - PubMed
    
12. Donoso R, Diaz C, Villavicencio P. Comparative study of keratoconus between Anwar's deep anterior lamellar keratoplasty versus converted penetrating keratoplasty. Arch Soc Esp Oftalmol. 2015;90(6):257-263. doi:10.1016/j.oftal.2014.11.008
    CrossRef - PubMed
    
13. Abdelaal AM, Alqassimi AH, Malak M, Hijazi HT, Hadrawi M, Khan MA. Indications of keratoplasty and outcomes of deep anterior lamellar keratoplasty compared to penetrating keratoplasty. Cureus. 2021;13(3):e13825. doi:10.7759/cureus.13825
    CrossRef - PubMed
    
14. Feizi S, Javadi MA, Karimian F, Bayat K, Bineshfar N, Esfandiari H. Penetrating keratoplasty versus deep anterior lamellar keratoplasty for advanced stage of keratoconus. Am J Ophthalmol. 2023;248:107-115. doi:10.1016/j.ajo.2022.11.019
    CrossRef - PubMed
    
15. Benetz BA, Lass JH. Specular Microscopy. Cornea. 2018;37 Suppl 1:S7-S8. doi:10.1097/ICO.0000000000001642
    CrossRef - PubMed
    
16. Bulirsch LM, Weber C, Sassmannshausen M, et al. [Histological changes in keratoconus and wound healing after corneal cross-linking]. Ophthalmologe. 2022;119(4):342-349. doi:10.1007/s00347-021-01537-8
    CrossRef - PubMed
    
17. Orlin SE, Raber IM, Eagle RC, Jr., Scheie HG. Keratoconus associated with corneal endothelial dystrophy. Cornea. 1990;9(4):299-304.
    CrossRef - PubMed
    
18. Cremona FA, Ghosheh FR, Rapuano CJ, et al. Keratoconus associated with other corneal dystrophies. Cornea. 2009;28(2):127-135. doi:10.1097/ICO.0b013e3181859935
    CrossRef - PubMed
    
19. Vincent AL, Jordan C, Sheck L, Niederer R, Patel DV, McGhee CN. Screening the visual system homeobox 1 gene in keratoconus and posterior polymorphous dystrophy cohorts identifies a novel variant. Mol Vis. 2013;19:852-860.
    CrossRef - PubMed
    
20. Guan M, Zhao W, Zhang Y, et al. Graft survival rate of deep anterior lamellar keratoplasty for keratoconus: A meta-analysis. Medicine (Baltimore). 2018;97(28):e11404. doi:10.1097/MD.0000000000011404
    CrossRef - PubMed
    
21. Liu H, Chen Y, Wang P, et al. Efficacy and safety of deep anterior lamellar keratoplasty vs. penetrating keratoplasty for keratoconus: a meta-analysis. PLoS One. 2015;10(1):e0113332. doi:10.1371/journal.pone.0113332
    CrossRef - PubMed
    
22. Vabres B, Bosnjakowski M, Bekri L, Weber M, Pechereau A. [Deep lamellar keratoplasty versus penetrating keratoplasty for keratoconus]. J Fr Ophtalmol. 2006;29(4):361-371. doi:10.1016/s0181-5512(06)77694-x
    CrossRef - PubMed