Objectives: We evaluated the effects of various graft-marking techniques on surgical results in patients undergoing Descemet membrane endothelial keratoplasty.
Materials and Methods: In this single-center retrospective study, 65 eyes from 55 patients that had received various types of marking or no marking and that had been used for Descemet membrane endothelial keratoplasty endothelial graft preparation were included. Patients were divided into 3 groups according to the marking technique used: group I (F-marked graft; 17 eyes), group II (asymmetric triangle-marked graft; 12 eyes), and group III (unmarked graft; 36 eyes). The main outcome measurements were best-corrected visual acuity, endothelial cell density, central corneal thickness, postoperative complications, rebubbling, and secondary keratoplasty rates.
Results: In groups I, II, and III, rates of patients with 6-month best-corrected visual acuity ≥ 20/32 were 35.7%, 77.8%, and 71.9%, respectively (P = .04). The mean 6-month endothelial cell density decrease for each group was 43.3%, 48.8%, and 46.4%, respectively (P = .589), whereas the mean 6-month central corneal thickness decrease for each group was 7.7%, 15.8%, and 34.0%, respectively (P = .001). Rates of primary graft failure for groups I, II, and III were 35.3%, 8.3%, and 13.9%, respectively. Rebubbling was performed in 21.5% of eyes, and secondary keratoplasty was required in 29.2% of eyes.
Conclusions: Although graft-marking techniques for Descemet membrane endothelial keratoplasty greatly facilitate graft positioning during surgery, both the potential toxic effects of alcohol on the endothelium when marking with gentian violet dye and the risk of graft detachment with asymmetric marking must be considered.
Key words : Corneal graft detachment, Corneal transplantation, Gentian violet dye, Graft orientation
Descemet membrane endothelial keratoplasty (DMEK) was introduced by Melles and colleagues in 20061 and has become popular because of the better visual outcomes with more rapid visual rehabilitation. In addition, DMEK has lower endothelial rejection and general complication rates than Descemet stripping automated endothelial keratoplasty and penetrating keratoplasty (PKP).1-5 However, the learning curve for performing DMEK can be steep for surgeons new to the technique because of the more difficult donor tissue preparation, implantation, orientation, and unscrolling.6
One of the most difficult stages of DMEK is positioning the graft correctly in the anterior chamber and then unscrolling it. However, positioning is the most important factor determining surgical success; accidental upside-down placement of the graft results in iatrogenic primary graft failure.7 Various techniques have been defined to ensure placement of the graft in the correct position, and instruments and techniques that help orientation include the handheld slit beam,8 endoilluminator-assisted transcorneal illumination,9 and intraoperative optical coherence tomography.6,10 Marking the stromal endothelial graft aspect with a surgical skin marker (S-stamp, F-mark) and marking the graft periphery with a semicircular punch or with the single triangular technique have also been reported as methods to facilitate graft positioning.11-15
The aim of this study was to determine the effects of 2 endothelial graft-marking techniques (F-mark and asymmetric triangular mark) on best-corrected visual acuity (BCVA), endothelial cell density (ECD) decrease, central corneal thickness (CCT), postoperative complications, and need for rebubbling and secondary keratoplasty. We are not aware of any other published study that has reported effects of multiple graft-marking techniques on DMEK success.
Materials and Methods
The Institutional Review Board and Ethics Committee of the Dr. Lutfi Kirdar Kartal Education and Research Hospital approved the initiation of the study (IRB No: 2018/514/143/1). The study was conducted according to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all individual study participants.
This study included 65 eyes of 55 patients who met all inclusion criteria and had undergone the DMEK procedure between May 2014 and February 2018 as a result of diagnosis of Fuchs endothelial dystrophy, pseudophakic bullous keratopathy, or failed DMEK. All procedures were performed at a single center. Patients were divided into 3 groups according to the marking technique used during the endothelial graft preparation stage: group I (F-marked endothelial graft; n = 17 eyes), group II (asymmetric triangle-marked endothelial graft; n = 12 eyes), and group III (endothelial graft prepared without any marking technique; n = 36 eyes). Patients who were aphakic or had a dislocated intraocular lens, a history of glaucoma surgery, iris or pupil disorders, uveitis, or microcornea and those on whom an F-mark or asymmetric triangle could not be observed on the intraoperative graft were excluded from the study. There were 20 male (36.4%) and 35 female (63.6%) patients with a mean age of 69.78 ± 9.82 years (range, 49-84 y). The 65 surgeries were performed by a single experienced surgeon (BK) and consisted of 48 DMEK cases (73.8%), 13 triple DMEK cases (20%), and 4 repeat DMEK (re-DMEK) cases (6.2%). The mean follow-up was 19.2 ± 13.8 months (range, 6-48 mo). Demographic data, DMEK indications, and surgeries performed are presented in Table 1.
Donor tissue preparation and marking technique
The corneoscleral button was affixed onto a suction holder (Barron; Katena Products, Inc., Parsippany-Troy Hills, NJ, USA) while the donor Descemet-endothelium complex was being prepared. The peripheral endothelium was marked with a Y-shaped hook or a 9.0- to 9.5-mm trephine. The mark was then made visible using the 60-second 0.06% trypan blue staining procedure (Vision Blue; DORC, Zuidland, the Netherlands). The Descemet-endothelium complex was partially detached by pulling with 2 forceps in a parallel and centripetal manner. The graft was then placed in the same position and cut with 8.0-, 8.25-, or 8.50-mm trephine.
After the graft cut with the punch trephine was stripped halfway, viscoelastic substance (sodium hyaluronate 1.4%; Bio-hyalur EV; Biotech, Gujarat, India) was inserted between the endothelial surfaces in the region to be marked to prevent contact in the F-mark group. An F-mark was placed directly onto the stromal aspect of the Descemet-endothelium complex with a surgical skin marker, and the graft was brought into its original position using balanced salt solution. The marked graft edge was then grasped using non-toothed forceps to enable stripping from the stroma (Figure 1).
The asymmetric triangular marking group went through the same initial procedures, after which the graft was cut with 8.0-, 8.25-, or 8.50-mm trephine. An asymmetric triangle was created on the Descemet-endothelium complex at a single region and from the endothelial surface, using an ophthalmic stab knife at 45° in a counterclockwise manner (Figure 2).15 The prepared Descemet-endothelium complex was stained with 0.06% trypan blue and then placed into a standard lens cartridge (Zaraccom cartridge and injection system; Sivas, Turkey).
Retrobulbar anesthesia was administered to all patients with the modified Van Lint facial block technique using a 4-mL mixture of 2% lidocaine hydrochloride and 0.5% bupivacaine hydrochloride. Intraocular pressure was then decreased with 5 minutes of ocular massage and 15 minutes of Honan balloon use. After routine surgical preparations were completed, the corneal epithelium was denuded to obtain a clear image when needed. The 2 side ports and the corneal surface were marked at an approximately 9.0-mm diameter using a surgical skin marker. The anterior chamber was filled with air after the side ports were opened at the 10:30 and 1:30 clock positions. A reverse Sinskey hook was used to perform a Descemetorhexis (removal of the Descemet membrane) with a diameter of 9 mm by scoring a circular Descemet membrane area and then stripping it from the posterior stroma.
Graft insertion was then performed through a
3.0-mm tunnel incision created at the limbus 12-o’clock position. Miosis was created by administering carbachol (Miostat Single 0.01%, Novartis Pharma GmbH, Nürnberg, Germany) into the anterior chamber. A peripheral iridectomy was performed at the 6-o’clock position using a 23-gauge vitrectomy probe. A repeat iridectomy was not performed as one was already present in cases with repeat DMEK. The cartridge was then used to insert the graft into the anterior chamber. A 30-gauge cannula was used for corneal anterior surface tapping while unfolding the graft. After confirmation that the graft had opened in the correct position and the marks were in the appropriate places, adhesion of the graft to the receiver posterior stroma was ensured by filling the anterior chamber with air.
Triple Descemet membrane endothelial keratoplasty
The triple procedure consisted of endothelial keratoplasty after standard phacoemulsification surgery and intraocular lens implantation.16 For these surgical cases, we performed a Descemetorhexis of approximately 9.0 mm in diameter under viscoelastic material (sodium hyaluronate, Bio-hyalur EV, Biotech, India) after cataract surgery and intraocular lens implantation. The viscoelastic material and Descemet membrane remnants in the anterior chamber and behind the intraocular lens were then removed with irrigation-aspiration. Miosis was ensured by administering carbachol (Miostat Single, Novartis Pharma GmbH) to the anterior chamber. Peripheral iridectomy was performed at the 6-o’clock position using a 23-gauge vitrectomy probe. The graft was implanted with a standard lens cartridge and then fully opened, with air then administered between the graft and the iris. We confirmed that the DMEK graft was transplanted on the receiver graft stroma after it was completely unfolded.
The corneal entry sites were sutured, and gentamycin (Gentamicin, DEVA Holding, Istanbul, Turkey) and dexamethasone (Dekort, DEVA Holding) were injected subconjunctivally at the end of the surgery. The patients were then instructed to remain supine for 48 hours to ensure adhesion of the graft to the stroma.
All eyes received 0.5% moxifloxacin hydrochloride (Vigamox, Alcon, Fort Worth, TX, USA) and 0.1% dexamethasone (Maxidex, Alcon) 5 times per day postoperatively. The corneal sutures were removed within 15 days. Topical antibiotic treatment was stopped on postoperative day 10 in all patients. Topical steroid treatment was decreased 1 drop per day every month and changed to 0.5% loteprednol etabonate (Lotemax, Bausch & Lomb, Bridgewater, NJ, USA) 3 times per day at the end of month 3. The local steroid treatment was then gradually decreased to a once daily maintenance dose.
Patients with corneal graft detachment on microscopy were evaluated together with anterior segment optical coherence tomography results to determine whether the rebubbling procedure was necessary. Rebubbling was performed in the first month for peripheral detachments that were large or showed scroll formation or interfered with the visual axis. This procedure consisted of unrolling the graft and keeping the patient in a supine position afterward as necessary. Peripheral detachments that were straight or did not interfere with the visual axis were followed up without intervention.
Patients whose cornea did not clear postoperatively were considered to have primary or secondary graft failure. Primary graft failure was defined as grafts that actually attached but did not clear at any time, whereas secondary graft failure was defined as attached grafts that had shown signs of becoming clear but that had then decompensated.
We used PASW Statistics software (SPSS: An IBM Company, version 17.0, IBM Corporation, Armonk, NY, USA) for statistical analyses. Results are shown as means ± standard deviation. Before comparisons, the Shapiro-Wilks test was used to determine whether the group data had a normal distribution. The chi-square test, Wilcoxon test, paired sample t test, Fisher exact test, and independent sample t test were used for evaluation. Statistical significance level was accepted as P < .05.
After exclusion of patients with low visual acuity potential (diabetic macular edema, age-related macular degeneration) and those who had undergone unsuccessful DMEK surgery (repeat DMEK or secondary PKP) before 6 months, the overall preoperative and postoperative 6-month mean BCVA values were 2.02 ± 0.87 and 0.50 ± 0.69 logMAR, respectively (P < .001). Preoperative mean BCVA values for groups I, II, and III were 2.11 ± 0.99, 2.04 ± 0.98, and 1.98 ± 0.81 logMAR, respectively, whereas postoperative mean BCVA values were 0.86 ± 0.90, 0.33 ± 0.59, and 0.39 ± 0.58 logMAR, respectively (group I vs II: P = .871; group I vs III: P = .648; group II vs III: P = .855). Overall, 63.6% of eyes attained a BCVA ≥ 20/32 (≥ 0.6 Snellen fraction [decimal]) and 27.3% attained ≥ 20/20 (≥ 1.0).
Distributions of ocular comorbidities within the groups were as follows. In group I, 2 eyes had age-related macular degeneration and 1 eye had diabetic macular edema. In groups II and III, 1 eye in each group had age-related macular degeneration. Patients who underwent unsuccessful DMEK surgery before 6 months were distributed to the groups as follows: secondary PKP before 6 months in 2 eyes in group II and secondary PKP before 6 months in 1 eye and re-DMEK before 6 months in 2 eyes in group III.
When the 6-month BCVA results were evaluated in the 3 groups after ocular comorbidities and graft failure instances were omitted, the percentage of patients with a BCVA of ≥ 20/32 (≥ 0.6) in groups I, II, and III was 35.7%, 77.8%, and 71.9%, respectively (P = .04). A 6-month BCVA result of ≥ 20/32 (≥ 0.6) was statistically significantly lower in group I than in groups II and III (group I vs II: P = .04; group I vs III: P = .02). There was no significant difference in BCVA between group II and group III (P = .72). In groups I, II, and III, a 6-month BCVA value of ≥ 20/20 (≥ 1.0) was seen in 7.1%, 33.3%, and 34.4%, respectively. The percentage of patients with a BCVA of ≥ 20/20 (≥ 1.0) was higher in groups II and III than in group I, but there was no statistically significant difference between the groups (group I vs II: P = .26; group I vs III: P = .073; group II vs III: P = 1.0).
Table 2 presents the visual acuity percentages for each group.
Endothelial cell density
We excluded 14 of the 65 eyes (21.5%) from the ECD analysis (6 had a secondary reoperation before the 6-month follow-up and 8 had corneal edema that prevented proper evaluation). In the remaining 51 eyes, the mean donor ECD was 2817.07 ± 310.13 cells/mm2 before and 1496.45 ± 336.82 cells/mm2 6 months after surgery, indicating a decrease of 46.01 ± 11.96%. In groups I, II, and III, the percent ECD decrease at 6 months was 43.3%, 48.8% and 46.4%, respectively, with no significant difference between groups (group I vs II: P = .176; group I vs III: P = .468; group II vs III: P = .647).
Central corneal thickness
In all groups, the mean CCT decreased 23.29% ± 17.06%, from 761.87 ± 148.40 μm preoperatively to 569.52 ± 70.88 µm at postoperative month 6 (P < .001). The decrease in CCT at month 6 was 7.7%, 15.8%, and 34.0% in groups I, II, and III, respectively. The percent CCT decrease was significantly higher in group III patients than in the other groups (group I vs III: P < .001; group II vs III: P = .002). The percent CCT decrease was also significantly higher in group II patients than in group I patients (group I vs II: P = .006).
Postoperatively, 14 eyes (21.5%) required rebubbling (air reinjection), 5 eyes (7.7%) required secondary PKP, and 14 eyes (21.5%) required repeat DMEK. We encountered primary graft failure in 12 eyes (18.4%), secondary graft failure in 6 eyes (9.2%), and graft separation in 16 eyes (24.6%).
Rates of primary graft failure in groups I, II, and III were 35.3%, 8.3%, and 13.9%, respectively. Primary graft failure rate was higher in group I than in the other groups, although not significantly (group I vs II: P = .187; group I vs III: P = .143; group II vs IIIL P = 1.000). Rates of secondary graft failure rate were 0% in group I, 41.7% in group II, and 2.8% in group III. The secondary graft failure rate was significantly higher in group II than in the other groups (group I vs II: P = .007; group I vs III: P = 1.0; group II vs III: P = .002).
Rates of graft detachment were 17.6%, 50.0%, and 19.4%, respectively, in groups I, II, and III, that is, higher in group II than in the other groups (group I vs II: P = .106; group I vs III: P = 1.00; group II vs III: P = .039). The rebubbling rate was 17.6% in group I, 50% in group II, and 13.9% in group III (group I vs II: P = .106; group I vs III: P = .701; group II vs III: P = .01). Although rebubbling rate was not significantly different between groups I and II and between groups I and III, the rate in group II was significantly higher than the rate in group III.
A secondary corneal transplant was required in 29.2% of the eyes overall. The secondary corneal transplant rate was 35.3%, 50%, and 19.4%, respectively, in groups I, II, and III (group I vs II: P = .428; group I vs III: P = .211; group II vs III: P = .039). Although secondary corneal transplant rate was not significantly different between groups I and II, group II rate was significantly higher than the rate in group III.
Glaucoma was present in 4.6% of eyes (n = 3, with 1 in group I and 2 in group II) in the postoperative period, and treatment was started with 0.1% topical combined timolol and dorzolamide (Cosopt, Merk & Co, Inc., Whitehouse Station, NJ, USA) twice daily. Allograft rejection developed in 1 eye (1.5%) from group III, which was treated with topical steroid eye drops containing 0.1% dexamethasone sodium phosphate (Maxidex, Alcon). Table 3 presents resurgery data and postoperative complications for the groups.
We compared the 6-month clinical outcomes of patients who had DMEK performed after graft marking with the F-mark or an asymmetric triangle and after those performed without a graft mark. A total of 65 eyes that had undergone DMEK surgery by a single surgeon were included in our study.
We evaluated the BCVA, ECD, CCT, secondary keratoplasty, and complication rates in DMEK cases with and without preliminary graft marking, considering that the marking techniques used on grafts could influence surgical success because of the possible toxic or traumatic effects on the endothelium.
There are limited numbers of studies on graft-marking techniques in the literature.7,11-15 Marking techniques used for graft orientation include the F-mark,11 S-stamp,12 3 circular marks,13 4 asymmetric marks,14 and a single peripheral triangular mark.15 The reason that such marking techniques are needed is because correct anatomic orientation is the most important factor for DMEK success. The DMEK tissues tend to create a scroll, leading to a rotation potential during the injection or unfolding phases. A device to ensure continuous correct orientation during insertion and manipulation so far unfortunately does not exist. Other causes of difficulty in orienting the graft include suboptimum intraoperative anterior chamber view, for example, due to edematous and scarred corneas, and poor graft scrolling tendency as with older donor grafts.15
Several in vitro and in vivo studies have evaluated the effects of the F-mark and S-stamp techniques using gentian violet dye on the graft and the relevant surgical results.7,12,17 In an in vitro study on graft marking with the S-stamp technique using pixel calculation with special software after the graft was stained with calcein, associated endothelial cell loss was shown to be 0.6%. This loss is reported to make up 4.2% of all the endothelial loss at the endothelial preparation stage.17 In a study that compared the 6-month clinical results of 133 DMEK cases that had the S-stamp graft-marking technique and 31 DMEK cases without marking, there were no significant differences between the 2 groups with regard to BCVA, ECD decrease, and rebubbling rates.12 The upside-down graft implantation rate was 0% in the DMEK S-stamp group and 9.4% in the DMEK group without marking. In our study, we found that the primary graft failure rate was higher (35.3%), although not significantly, in patients who had undergone DMEK after the F-mark was placed directly with the surgical skin marker compared with the other groups. The percentage of patients with a6-month BCVA of ≥ 20/32 (≥ 0.6) was 35.7%, 77.8%, and 71.9%, respectively, in the F-mark, asymmetric triangular mark, and the unmarked groups, with levels significantly lower in the F-mark group than the other groups.
The 6-month ECD decrease was 43.3%, 48.8%, and 46.4%, respectively, in groups I, II, and III in our study. Because it was not possible to obtain postoperative corneal measurements, we could not include 29.4%, 33.3%, and 13.8% of these eyes, respectively, in calculations for groups I, II, and III. We believe the reason for the lack of a significant difference regarding ECD measurements between the groups could be because only data of measurable eyes were evaluated and because of the low number of patients in the groups. The 6-month CCT decrease was 7.7%, 15.8%, and 34% in groups I, II and III, respectively, in our study, with this decrease significantly lower in group I patients versus the other groups. The BCVA levels were lower in the F-mark group as the CCT measurement could be performed in all patients in the groups and because the CCT decrease is directly related to the BCVA level.
Stoeger and associates stated that gentian violet dye ink can be safely used in the S-stamp technique for Descemet stripping automated endothelial keratoplasty grafts.18 However, the authors emphasized that one needs to wait for 10 seconds before stromal surface marking is performed so that the alcohol can dry after ink is applied to the stamp.12,18 The ink was applied to the stamp, followed by waiting 10 seconds for the alcohol carrier to dry, and then the dried dye was applied to the stromal surface. In our study, we believe that the alcohol in the ‘‘wet’’ gentian violet dye used directly on the graft stromal surface without a stamp in our study created a toxic effect on the endothelium and decreased the surgical success rates in the F-mark group.
Making asymmetric cuts in the endothelium is another option for correct orientation of the graft. Three techniques have been defined in the literature: 3 circular marks (Bachmann technique)13, 4 asymmetric marks (Matsuzawa technique),14 and a single peripheral triangular mark15 (Bhogal technique). The marking is created in the graft periphery using dermatologic biopsy punches with the Bachmann13 and Matsuzawa14 techniques and a 30-degree incision knife with the Bhogal technique. The theoretical endothelial cell loss due to marking has been reported as 2.5% with the Bachmann technique, 5.8% with the Matsuzawa technique, and 0.7% with the Bhogal technique. The authors stated that these loss numbers are acceptable compared with the results that occur with incorrect graft positioning. The Bachmann technique was reported to cause a 1-month ECD loss of 37% and primary graft failure rate of 12%; the Matsuzawa technique resulted in a 6-month ECD loss of 44% with no primary graft failure. In our study, the 6-month ECD decrease was 48.8% and the primary graft failure rate 8.3% in the DMEK cases performed with an asymmetric triangle. The ECD loss rates were similar in the cases that underwent the procedure with the Bachmann or Matsuzawa technique and our asymmetric triangle group. However, we found the secondary keratoplasty rates in the asymmetric triangle group to be higher than with the Bachmann or Matsuzawa technique. We believe this could be due to the high graft separation and rebubbling rates in the asymmetric triangle group.
The rebubbling, primary graft failure, and secondary keratoplasty rates with the F-mark and asymmetric triangle marking techniques in our study group were markedly higher than the rates in other DMEK series in the literature. However, the results for our group without marking were similar to those reported in other studies.19,20
Our study is the first to report the clinical results of multiple graft-marking techniques to the best of our knowledge. These techniques greatly facilitate graft orientation in cases where the intraoperative anterior segment view is suboptimum and the graft has poor scrolling tendency. However, we encountered primary graft failure due to alcohol toxicity in our cases where gentian violet dye was used for marking as we had directly used a surgical marker for the marking procedure. We also observed graft separation starting from the marking area with asymmetric triangle marking that resulted in similar increases in rebubbling, secondary keratoplasty, and secondary graft failure rates. Although the graft-marking technique greatly facilitates intraoperative graft position orientation during DMEK surgery, one must consider the toxic effects of alcohol on the endothelium when marking with the gentian violet dye and the increased graft detachment risk with the asymmetric graft-marking technique.
DOI : 10.6002/ect.2019.0424
From the 1University of Health Sciences, Kartal Dr. Lütfi Kırdar Training and
Research Hospital, Department of Ophthalmology, Istanbul, Turkey; and the
2Boyabat 75th Year State Hospital, Department of Ophthalmology, Sinop, Turkey
Acknowledgements: The authors report no conflicts of interest and received no financial support for the study. The authors alone are responsible for the content and writing of the paper. Datasets used in this study are available from the corresponding author on reasonable request.
Corresponding author: Burak Tanyıldız, University of Health Sciences, Kartal Dr. Lütfi Kırdar Training and Research Hospital, Department of Ophthalmology, Semsi Denizer Street, E-5, 34890 Kartal, Istanbul, Turkey
Phone: +090 554 277 57 17
Table 1. Demographic Data, Indication for Descemet Membrane Endothelial Keratoplasty and the Surgeries Performed
Table 2. Postoperative Visual Acuity Levels at 6 Months After Descemet Membrane Endothelial Keratoplasty
Table 3. Repeat Surgeries and Postoperative Complications
Figure 1. Graft Preparation With the F-Mark Technique
Figure 2. Graft Preparation With the Asymmetric Triangle Mark Technique