Article Text
PDF
Abstract
Background To investigate the difference in clinical outcomes between large-diameter deep anterior lamellar keratoplasty (L-DALK) and standard DALK (S-DALK) for the treatment of keratoconus.
Methods 132 patients (132 eyes) from the Zhongshan Ophthalmic Center with a clinical diagnosis of keratoconus were enrolled. The participants were featured by the intolerance to rigid gas-permeable contact lenses or unsuccessful fitting of contact lenses. Using stratified blocked randomisation, eligible eyes were allocated into the L-DALK group or the S-DALK group (66 eyes, respectively). Postoperative uncorrected visual acuity (UCVA), best spectacle-corrected visual acuity (BSCVA), refractive sphere, manifest cylinder and spherical equivalent refractive error were tested at 6, 12, 18 and 24 months after surgery.
Results After surgery, the L-DALK group had better UCVA and BSCVA than the S-DALK group (p=0.000 and 0.021, respectively). At 24 months, mean BSCVA was 0.17±0.10 logarithm of the minimum angle of resolution (logMAR) (Snellen equivalent, 20/25) in the L-DALK group vs 0.22±0.10 logMAR (Snellen equivalent, 20/32) in the S-DALK group. Differences were observed between the L-DALK group and the S-DALK group in terms of refractive sphere (p=0.015), manifest cylinder (p=0.014) and spherical equivalent refractive error (p=0.034) at any time interval postoperatively. At 24 months, the mean spherical equivalent refractive error was −3.5±3.2 D and −5.2±2.6 D in the L-DALK and S-DALK groups, respectively.
Conclusions L-DALK can reduce the degree of postoperative myopia and manifest astigmatism and improve visual acuity outcomes in keratoconus compared with S-DALK.
Trial registration number Chinese Clinical Trial Registry (TRC-13003122).
- Cornea
- Clinical Trial
- Treatment Surgery
Statistics from Altmetric.com
Introduction
Deep anterior lamellar keratoplasty (DALK) is an alternative surgical procedure to penetrating keratoplasty (PK) to treat various corneal pathologies without the involvement of the endothelium (including keratoconus).1–3 Its principal advantage is that it preserves the endothelium, thereby decreasing the prevalence of graft rejection4 and continuous loss of corneal endothelium,5–7 which are common after PK treatment.
The cornea extends forward progressively in patients diagnosed with keratoconus, consequently presenting as high myopia and astigmatism.4 Hence, the principal aim of DALK for the treatment of keratoconus is to regain corneal refractive regularities. However, a series of studies have reported that DALK produced significantly higher myopia than PK, which had a great influence on the improvement of uncorrected visual acuity (UCVA) in patients with keratoconus.5 ,7–12
Conversely, large-graft diameters are associated with reduced astigmatism and improved visual acuity in PK.13–15 Hence, we propose that large-diameter DALK (L-DALK) will reduce spherical errors and astigmatism in patients with keratoconus. Here, we describe the L-DALK for the treatment of keratoconus and evaluation of clinical outcomes in comparison with conventional DALK.
Materials and methods
Patients
Between September 2006 and August 2010, 132 patients (132 eyes) with keratoconus were enrolled in this prospective comparative clinical trial. The participants were stratified into two stratums based on preoperative keratometry (≥60 D or <60 D). Using block enrolment, the patients were assigned either to the L-DALK group or to the standard DALK (S-DALK) group in each stratum. Each group had 66 patients. Table 1 summarises the characteristics and preoperative baseline data of patients. The patients with a history of hydrops or scarring down to Descemet's membrane were excluded from the study. The eyes without Descemet's membrane exposure or with intraoperative and postoperative complications were excluded from visual outcome and manifest refraction analysis.
Characteristics and operative data of patients in the large-diameter DALK and standard DALK groups
Main outcome measures
Main eye examination, including UCVA, best spectacle-corrected visual acuity (BSCVA), refractive sphere, manifest cylinder and spherical equivalent refractive error, was carried out at scheduled intervals by one of the two investigators (HT, ZH). UCVA and BSCVA were examined at 3, 6, 9, 12, 18 and 24 months' follow-up; manifest refraction was tested at 6, 12, 18 and 24 months' follow-up. The outcomes at 24 months' follow-up were defined as the primary endpoints (figure 1).
The Consolidated Standards of Reporting Trials flow diagram.
Surgical procedure
We limited the indication for surgery to intolerance to rigid gas-permeable contact lenses or unsuccessful fitting of the contact lens and poor BSCVA. In the L-DALK group, corneal buttons of 8.50 mm size were designated as large grafts, used for 8.25 mm recipients, whereas corneal buttons of 8.0 mm size were designated as standard grafts, used for 7.75 mm recipients in the S-DALK group.
All patients underwent surgery by a single surgeon (HT) at the Zhongshan Ophthalmic Center. The big bubble method with some modifications16 was used. Retrobulbar anaesthesia was administrated in all surgeries.
In brief, a partial-depth incision of the recipient cornea was carried out using Barron trephine (Katena, Denville, New Jersey, USA). The depth of trephination was set to reach 375 μm into the recipient stroma. Superficial corneal stroma was excised step by step to reduce the thickness of recipient bed using a crescent knife. A 30-gauge needle was connected to a 5 mL syringe filled with air. The needle was bent 60° at 5 mm from the tip and inserted into the deep stroma and air was injected into the stroma in order to form a big air bubble between the corneal stroma and the Descemet's membrane. The posterior stroma was removed with a blunt-tipped scissor, thereby exposing the Descemet's membrane. If a large air bubble could not be formed using the above technique, a layer-by-layer stromal dissection was carried out. Using this method, we undertook very deep lamellar keratoplasty in which a very thin stromal layer was left behind.
Barron donor punch was used to punch out the donor buttons from the endothelial side. The Descemet's membrane with endothelial layer was removed with a wet sponge. A 10-0 nylon suture was used to suture the button with 16 interrupted sutures or one running, 20-bite suture.
Postoperative management and follow-up
When complete corneal reepithelialisation was achieved, topical dexamethasone 0.1% was administered four times daily for 1 month. Topical artificial tears were applied four times daily for 6 months. Topical steroid was tapered gradually within three months.
Complete suture removal was completed by 18 months. When there was loose suture or vascularisation in the host cornea, early suture removal would be administrated immediately, and this often occurred at 16.2±2.9 (range, 9–18) and 16.6±2.6 (range, 9–18) months in the L-DALK and S-DALK groups, respectively (p=0.460).
Statistical analyses
Statistical analyses were carried out using SAS for Windows V.8.1 (SAS Institute, Cary, North Carolina, USA). Comparing means using repeated-measures analysis of variance was used to compare the treatment groups for visual outcome and manifest refraction analysis.
Data on complications (including perforation of the deeper stroma, graft–host interface opacity, stromal rejection) and eyes that had a hyperopic spherical equivalent errors were analysed with Pearson's χ2 or Fisher's exact test; p<0.05 was considered significant.
Results
Frequency of exposing true Descemet's membrane
Using the big bubble method, we successfully exposed the Descemet's membrane in 59 eyes (59/66, 89.4%) in the L-DALK group and in 56 eyes (56/66, 84.8%) in the S-DALK group (p=0.604). In the remaining 7 (L-DALK) and 10 eyes (S-DALK) in which a big bubble formation could not be achieved by air injection, layer-by-layer stromal dissection was carried out. Using this method, we undertook very deep lamellar keratoplasty in which a very thin stromal layer was left behind.
Complications
In total, 3 of 7 (42.9%) and 3 of 10 (30.0%) eyes presented a small perforation of the recipient bed in the L-DALK and S-DALK groups, respectively, all of which presented in the patients that underwent layer-by-layer stromal dissection. A double-anterior chamber was noted in two eyes in the L-DALK group immediately after surgery, yet it regressed spontaneously at 6 and 7 days after surgery.
The mild opacities presented in the early stage after surgery in 4 of 7 eyes (57.1%), in which predescemetic DALK was carried out in the L-DALK group, and in 5 of 10 eyes (50.0%), in which predescemetic DALK was carried out in the S-DALK group. Interface opacities gradually cleared or reduced within three months after surgery in all the involved eyes.
Stromal rejection occurred in 2 of 59 eyes (3.4%) in the L-DALK group at 8 and 12 months, and in 2 of 56 eyes (3.6%) in the S-DALK group at 11 and 14 months after surgery (p=1.000). Stromal rejection was characterised by stromal oedema with decreased vision. All patients with rejection episodes responded well to topical corticosteroids, and grafts recleared after intensive corticosteroid treatment for 2–3 weeks.
Suture loosening was noted in 8 of 59 eyes (13.6%) in the L-DALK group at 4–8 months and in 7 of 56 eyes (12.5%) in the S-DALK group at 5–8 months (p=0.866).
Vascularisation of the host cornea occurred in 12 of 59 eyes (20.3%) in the L-DALK group and in 10 of 56 eyes (17.9%) in the S-DALK group (p=0.735).
Elevation of intraocular pressure (IOP) was found in 6 of 59 eyes (10.2%) in the L-DALK group and in 4 of 56 eyes (7.1%) in the S-DALK group (p=0.807). All patients with IOP spikes were applied topical timolol 0.1% and alphagan 0.2% two times daily, along with the termination of the use of topical steroid, and these treatments effectively reduced the IOP to normal ranges for all patients.
Visual acuity
Figure 2 shows the UCVA and eyeglasses-corrected changes (BSCVA) after DALK treatment. Both L-DALK and S-DALK were comparable in terms of UCVA and BSCVA through 1 year after surgery, but the former had both better UCVA (p=0.000) and BSCVA levels (p=0.021) than the latter after 1-year follow-up. At 24 months, the L-DALK group had a mean BSCVA value of 0.17±0.1 logarithm of the minimum angle of resolution (logMAR) (Snellen equivalent, 20/25), whereas this was 0.22±0.1 logMAR (Snellen equivalent, 20/32) for the S-DALK group. In the L-DALK group, 86.3% of the eyes reached a vision of 20/40 or better, whereas this was 76.6% in the S-DALK group.
Change in uncorrected visual acuity (UCVA) and best spectacle-corrected visual acuity (BSCVA) after large-diameter deep anterior lamellar keratoplasty (L-DALK) and standard DALK (S-DALK). No significant differences were noted within 12 months, but the differences occurred after 1-year follow-up in terms of UCVA (p=0.000) and BSCVA (p=0.021) between the two groups. logMAR, logarithm of the minimum angle of resolution.
Continuous and interrupted sutures were applied in 47 and 12 eyes in the L-DALK group and 47 and 9 eyes in the S-DALK group, respectively. In the L-DALK group, continuous sutures appeared to yield a poorer final BSCVA than interrupted sutures, although this was not found to be statistically significant (0.17±0.1 logMAR vs 0.15±0.1 logMAR; p=0.685), but the final UCVA levels were similar (0.35±0.1 logMAR vs 0.35±0.1 logMAR; p=0.893). In the S-DALK group, continuous sutures appeared to yield similar final BSCVA and UCVA levels compared with interrupted sutures (0.23±0.1 logMAR vs 0.20±0.1 logMAR, p=0.535 for BSCVA; 0.51±0.2 logMAR vs 0.51±0.2 logMAR, p=0.987 for UCVA).
Manifest refraction
Postoperatively, there were significantly lower levels of myopia (refractive sphere) (p=0.015) and astigmatism (manifest cylinder) (p=0.014) in the L-DALK group compared with the S-DALK group at any follow-up appointments. At 24 months, the mean refractive sphere was −2.9±2.8 D in the L-DALK group, whereas it was −3.8±2.1 D in the S-DALK group. The mean absolute value of manifest cylinder was 3.8±1.2 D in the L-DALK group, whereas it was 4.5±1.7 D in the S-DALK group. The mean postoperative spherical equivalent was −3.5±3.2 D and −5.2±2.6 D in the L-DALK and S-DALK groups, respectively (p=0.034). In total, 13.7% of the eyes had a hyperopic spherical equivalent in the L-DALK group, whereas it was 4.3% in the S-DALK group (p=0.162). Figure 3 demonstrates the fluctuations of manifest refraction in different treatment groups during follow-up.
Change in manifest refraction after large-diameter deep anterior lamellar keratoplasty (L-DALK) and standard DALK (S-DALK). Differences were observed in terms of refractive sphere (p=0.015), refractive cylinder (manifest astigmatism) (p=0.014) and spherical equivalent refractive error (p=0.034) between the two groups at all time intervals postoperatively.
In the L-DALK group, the mean spherical equivalent was 3.78±3.2 D vs −2.55±3.6 D (p=0.262) at 6 months, −3.64±2.9 D vs −3.41±3.7 D (p=0.824) at 12 months and −3.44±2.9 D vs −3.83±2.8 D (p=0.585) at 24 months for the eyes with continuous sutures versus interrupted sutures, respectively. In the S-DALK group, the mean spherical equivalent was −5.03±2.7 D vs −5.39±2.7 D (p=0.735) at 6 months, −4.53±2.8 D vs −3.31±3.0 D (p=0.270) at 12 months and −5.22±2.6 D vs −5.39±2.5 D (p=0.851) at 24 months for the eyes with continuous sutures versus interrupted sutures, respectively.
Discussion
DALK is an alternative surgical procedure to PK to treat various corneal pathologies without the involvement of the endothelium. Recently, certain studies have revealed that DALK had comparable visual outcomes with PK. However, a review of published literature on DALK in comparison to PK revealed no advantage of DALK for refractive error outcome.17 Several studies have demonstrated less favourable results after DALK.8 ,9 In those studies, the limited improvement of vision could be attributed to the residual posterior corneal stroma after manual lamellar dissection.5 ,8–9 ,11 ,18 Few studies have recognised that postoperative refractive shifts might be the additional factors interfering with visual acuity.
This study evaluated cases in which the true Descemet's membrane was exposed using the big bubble method, and no residual stroma was left behind on the recipient bed. Therefore, postoperative refractive shifts were the principal factors influencing visual acuity. Previous studies have demonstrated that PK with larger diameters of grafts can reduce postoperative manifest cylinder and improve visual acuity results in corneal ectasia diseases (including keratoconus).13 ,19 ,20 Those studies suggested that graft sizes could be closely related to the refractive shifts in cases with keratoconus after PK. On this basis, it is rational to attempt to alleviate postoperative refractive shifts using L-DALK.
The pressure of grafting buttons on the recipient bed is an important factor for the correction of keratoconus in DALK. The pressure reduced significantly by increasing the thickness of the recipient bed, which could induce insufficient correction. Therefore, some authors suggest it might be the residual stroma that attributes to higher degree of myopia after DALK compared with PK.5 ,7–12 We observed a tendency towards a lower myopia in the L-DALK group than in the S-DALK group. However, the difference in myopic refractive outcomes between the two treatment groups might have resulted from donors with different graft diameters because no residual stroma was left behind on the recipient bed. We have reason to accept the assumption that a larger graft diameter of the donor can result in a stronger pressure on the recipient bed.
Considering that Asian populations have a smaller corneal diameter, many previous studies used the donor grafts with a diameter of 7.75–8.0 mm for patients with keratoconus.10 ,16 In this study, corneal buttons of 8.50 mm size were designated as large grafts, whereas corneal buttons of 8.0 mm size were designated as standard grafts. Moreover, the identical size of buttons was used in the treatment groups, which eliminated the effects of the size of buttons on clinical outcomes. Hence, our outcomes were more convincing compared with previous studies,5 ,7–12 in which donor buttons with a wide range of sizes were used; consequently, there was much more variation on refractive outcomes.2–3 ,6–9 ,11
In this study, the L-DALK group had less manifest astigmatism than the S-DALK group (absolute value 3.8±1.2 D vs 4.5.±1.7 D at 24 months postoperatively) and both positive and negative cylinders were present in the two groups. However, the average postoperative manifest cylinder was +2.9 D (range, 0 to +4.75 D) in the study reported by Skeens and Holland,19 in which larger grafts were reported to range from 8.75 to 10.0 mm. It might appear that large grafts not only lead to less manifest astigmatism but also change the nature of the astigmatism. We suppose that corneal grafts ranging from 8.75 to 10.0 mm in diameter may result in more sufficient correction for corneal ectasia, and even induce the reversal of astigmatism, than grafts of standard diameter. Table 2 shows the comparison of our outcomes with the outcomes of other studies.
Comparison of this study with other studies for DALK for the treatment of keratoconus
The principal shortcoming of DALK is that it is technically challenging, which is also true of L-DALK. If L-DALK is converted intraoperatively to PK due to the perforation of deep stroma, a huge concern is whether large-diameter PK induces a higher rate of immunological rejection. Previous studies have stated that large-diameter PK is an important risk factor for graft rejection due to limbal proximity.19 ,23–25 This is why we carefully monitor the prevalence of graft rejection when L-DALK is converted to PK in patients with keratoconus. In the present study, we provided predescemetic DALK instead of PK if a perforation or a tear in the deep stroma occurred intraoperatively. To achieve a very thin recipient bed with constant thickness of stromal layer left behind, a layer-by-layer stromal dissection was required because of corneal ectasia and apical stromal thinning.
The main limitation of the study was that we compared only two sizes of grafts diameter (8.50 vs 8.0 mm) for the evaluation of outcomes. Grafts >8.5 mm in diameter were expected to further reduce postoperative refractive irregularities, but they might lead to a higher prevalence of stromal rejection (even though endothelial rejection was avoided in DALK). Further studies are required to confirm the cut-off size of graft diameter whereby refractive irregularities are controlled without additional complications.
Acknowledgments
We thank all the patients participating in the study for their cooperation. We also thank all the doctors and nurses for their excellent assistance in performing surgeries. Thanks also to the professors for their fruitful and constructive advice. Finally, we thank all the researchers involved in the study for their hard work.
References
Footnotes
Contributors TH conceived and designed the experiment; performed the experiment; wrote the paper; thoroughly revised and finalised the manuscript. CH, MG and YH analysed the data. TH, HZ and YW contributed reagents/materials/analysis tools.
Competing interests None.
Patient consent Obtained.
Ethics approval The Ethics Committee of Zhongshan Ophthalmic Center (Guangzhou, China) agreed the study protocol, which abided by the tenets of the Declaration of Helsinki (approval no. 2012KYNL027).
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- At a glance