Article Text
PDF
Abstract
Aim To assess the influence of graft thickness and regularity on visual recovery and postoperative wavefront aberrations after endothelial keratoplasty (EK).
Methods 150 EKs performed in eyes with corneal endothelial disorders and no other ocular comorbidities, preoperative and postoperative assessment with spectral domain optical coherence tomography and postoperative assessment with whole eye wavefront aberrometry were retrospectively analysed. Eyes were classified into five groups: Descemet Membrane Endothelial Keratoplasty (DMEK), nanothin Descemet Stripping Automated Endothelial Keratoplasty (DSAEK) (15–49 µm), ultrathin DSAEK (50–99 µm), thin DSAEK (100–149 µm) and conventional DSAEK (150–250 µm).
Results The preoperative diagnosis was Fuchs dystrophy in 139 eyes (92.7%). The graft thickness measured after graft deswelling was in average 74 µm with a mean coefficient of variation of 17%. The average follow-up time was 32 months. The mean spectacle-corrected logarithm of minimum angle of resolution visual acuity improved from 0.76 (20/116) before surgery to 0.14 (20/27) at last follow-up visit. No significant differences in final visual acuity were found between the five groups. The time to reach 20/40 vision was significantly shorter in the DMEK and nanothin DSAEK groups compared with the remaining three DSAEK groups. No significant differences in postoperative aberrometry measurements were found between the five groups. Shorter time to reach 20/40 visual acuity was associated with better preoperative visual acuity and thinner graft. Higher final vision improvement was associated with poorer preoperative visual acuity. Higher postoperative high-order aberrations were associated with poorer preoperative visual acuity.
Conclusion The main advantage of DMEK and nanothin DSAEK over thicker DSAEKs was the rapidity of visual recovery. Final quality of vision was not influenced by graft thickness and regularity.
- aberrometry
- cornea
- DMEK
- DSAEK
- endothelial keratoplasty
Statistics from Altmetric.com
Introduction
Surgical treatment of corneal endothelial disorders has evolved from penetrating (PKP) to endothelial (EK) keratoplasty with a current trend to use very thin grafts.1 Implementation of EK has led to faster visual recovery, fewer complications and better visual acuity outcomes compared with PKP.2 3 Two main surgical techniques are currently routinely used to treat corneal endothelial disorders, that is Descemet Stripping Automated Endothelial Keratoplasty (DSAEK) and Descemet Membrane Endothelial Keratoplasty (DMEK), with the latter providing optimum visual outcomes and decreased rejection rates.4–6 Limitations of DMEK include a longer learning curve and a higher frequency of rebubbling.7 8 The current evolution of DSAEK surgery is ultrathin DSAEK (UT-DSAEK), or even nanothin DSAEK, where the donor lamella is thinner than in conventional DSAEK and improved visual outcome with minimal surgery-induced damage to the donor tissue is expected.9 A randomised multicenter clinical trial has shown that compared with DSAEK, UT-DSAEK results in faster and better visual recovery with similar refractive outcomes, endothelial cell loss and incidence of complications.10 A prospective series showed similar visual outcomes and rejection rate between UT-DSAEK and DMEK.11 However, UT-DSAEK has not been clearly defined in the literature regarding the maximal graft thickness corresponding to ultrathin grafts.12
The origin of suboptimal visual outcomes after EK is multifactorial. High-order aberrations (HOAs) and light scattering are likely to be synergistic in reducing visual quality in the early postoperative period.13–16 Few studies have currently assessed wavefront aberrations after transplantation, while corneal aberrations derived from corneal topographic elevation data have been widely studied.14 On the one hand, better visual outcomes with DMEK compared with DSAEK could be explained by the stroma-stroma interface that would be inferior to the Descemet membrane-stroma interface. However, similar visual outcomes between UT-DSEK and DMEK suggest the interface may not be the main limiting factor. On the other hand, the DSAEK graft stromal thickness and regularity could induce HOAs that would limit visual recovery.
A few studies have shown that DMEK led to faster visual recovery compared with DSAEK.17 18 In addition, thinner graft may be better for vision recovery after EK.19 However, we still miss comprehensive comparison of UT-DSAEK with DMEK regarding not only graft thickness but also graft regularity.
The aim of the present study was to assess the influence of graft thickness and regularity (ie, variability of graft thickness in the central zone) on visual recovery and postoperative aberrations after EK using grafts with different thicknesses ranging from pure Descemet membrane to conventional DSAEK.
Patients and methods
Study design
This hospital-based retrospective study was conducted at the French National Eye Hospital.
Data from all consecutive patients with EK performed between April 2007 and January 2019 were reviewed. Inclusion criteria were the following: EK in eyes with corneal endothelial disorders, preoperative and postoperative assessment with spectral domain optical coherence tomography (SD-OCT), postoperative assessment with corneal topography and wavefront aberrometry. All eyes with pre-existing ocular comorbidities including age-related macular degeneration, advanced glaucoma, optic neuropathy, retinal detachment and retinal dystrophies were excluded. For each patient, demographic and clinical data, SD-OCT, corneal topography and wavefront aberrometry were collected.
Eyes were classified into five groups according to the graft thickness as follows: DMEK, nanothin DSAEK (15–49 µm), UT-DSAEK (50–99 µm), thin DSAEK (100–149 µm) and conventional DSAEK (150–250 µm). Eyes with very deep anterior chamber, a history of vitrectomy, poor iris condition or low visibility of the anterior chamber due to advanced corneal oedema and patients with contraindications to general anaesthesia were usually selected for DSAEK.
Surgical procedures
Donor tissue preparation for DSAEK
Donor tissue was prepared with a microkeratome device (Moria SA, Antony, France) under an operative microscope. The artificial anterior chamber was filled with deswelling medium and the graft was carefully centred. The height of the infusion bottle was set at 120 cm to get an artificial anterior chamber pressure over 65 mm Hg. The pressure was checked with an applanation tonometer, and the infusion was not closed during graft cut in order to maintain a constant pressure in the anterior chamber. The donor corneal epithelium was completely removed, and Bowman layer was carefully rinsed with balanced salt solution to make sure no epithelial fragments were remaining. Central corneal thickness was measured with ultrasound pachymetry, and peripheral corneal thickness was controlled using the operative microscope slit lamp with adjustable slit width (Carl Zeiss Meditec, Jena, Germany) or high-resolution anterior segment optical coherence tomography (Rescan, Carl Zeiss Meditec). The operative microscope slit lamp width was adjusted to get a narrow slit allowing peripheral corneal thinning to be detected.20 In our experience, both techniques allow major peripheral corneal thinning associated with advanced keratoconus, pellucid marginal degeneration or Terrien marginal degeneration as well as clear corneal incisions to be detected. In such cases, corneal cutting with the microkeratome is started from the opposite corneal zone. The microkeratome head was chosen according to central corneal thickness as follows: CCT <515 µm, 350 µm head; CCT 515–590 µm, 400 µm head; CCT 595–670 µm, 450 µm head; CCT 675–735 µm, 500 µm head; CCT >740 µm, 550 µm head. The donor cornea was cut with the microkeratome under the operative microscope as slowly as possible. Attention was paid to get a regular movement, and cutting time was slower if the CCT was close to the upper limit of the interval (ie, a 730 µm thick donor cornea was cut very slowly with the 500 µm head and a 680 µm thick donor cornea was cut more quickly with the same head). After cap creation, the peripheral corneal thickness was controlled a second time. If the donor tissue appeared to be thick in the peripheral zone, additional dissection was performed with a crescent blade to get a DSAEK graft as regular as possible. Cut donor tissue was trephined with an 8.25 mm Hessburg-Barron donor cornea punch (BPI, Grand Blanc, Michigan, USA).
Donor tissue preparation for DMEK
Donor tissue was prepared with the scuba technique under an operative microscope. A peripheral tunnel between Descemet membrane and stroma was first created with a crescent blade and forceps, and Descemet membrane was trephined with a 9.5 mm Hessburg-Barron donor punch. The peripheral rim of Descemet membrane was removed with forceps, and the central part was carefully partially detached. An 8 mm Hessburg-Barron donor punch was used for graft trephination. The graft was stained with trypan blue, completely detached with forceps and inserted in a Geuder injector (Geuder AG, Heidelberg, Germany).
Graft insertion
A 2.2 mm corneal incision was used for DMEK, and a 4 mm scleral or 2.2 mm corneal incision was used for DSAEK. The recipient Descemet membrane was stripped with a reverse Sinskey hook after an 8 mm mark was made on the corneal surface. An inferior peripheral iridectomy was made. DSAEK grafts were inserted with the Endosaver injector (Emmetrop, Courbevoie, France), and DMEK grafts were inserted with the Geuder injector. Intraoperative OCT was used to control graft positioning and graft was centred according to the epithelial mark. The anterior chamber was then filled with air 80% and SF6 20%.21 Postoperative treatment included topical dexamethasone and ofloxacin four times per day and 1% topical apraclonidine three times per day. Apraclonidine was discontinued after 5 days, ofloxacin after 1 month and the steroid eye drop regimen was progressively diminished. Dexamethasone was replaced by fluorometholone eyedrops after 3 months. The latter were maintained in the long term. Patients were hospitalised for 1 or 2 days. They were then examined prospectively at 1 week; 1, 3, 6, 9, 12, 18, 24, 30 and 36 months; and 5, 7 and 10 years after surgery.
At each postoperative visit, SD-OCT and corneal topography were assessed with the RTVue device (Optovue, Fremont, California, USA) and Orbscan IIz device (Bausch & Lomb Surgical, Rochester, New York, USA), respectively. Graft thickness was measured with OCT at least 3 months after surgery in order to allow the graft to deswell. Apex-centred horizontal corneal scans were used to measure graft thickness every 0.5 mm in the 5 mm central zone (figure 1). These 11 measurements were used to calculate the coefficient of variation of graft thickness.
Spectral domain optical coherence tomography of a nanothin Descemet Stripping Automated Endothelial Keratoplasty. The graft thickness is measured every 0.5 mm in the 5 mm central zone on an apex-centred horizontal corneal scan.
Ocular aberrations were measured using the iTrace Wavefront Aberrometer (Tracey Technologies, Houston, Texas, USA) at a pupil size of 4 mm or more, without pharmacologic dilatation. Ocular aberrations after EK were compared with those of a group of 15 normal patients candidate to refractive surgery. These patients were younger than 40 years (mean age, 29 years), and they featured normal slit lamp examination, normal corneal topography and normal SD-OCT assessment.
Statistical analysis
Snellen best spectacle-corrected visual acuity was converted into the logarithm of minimum angle of resolution (LogMAR) for statistical analysis. The χ2 test and analysis of variance (ANOVA) were used to compare the five groups of eyes. The relationship between recipient and graft characteristics and visual recovery was assessed with the Pearson correlation coefficient and ANOVA. Variables were first selected in univariate analysis and then analysed in multivariate linear regression.
Results
One hundred fifty eyes with EK were included. The average recipient age was 71.2±8.2 years (mean±SD). The preoperative diagnosis was Fuchs endothelial dystrophy in 139 out of 150 eyes (92.7%). Phacoemulsification was combined with EK in 81 eyes (54%). All eyes were pseudophakic after surgery. The graft thickness measured after graft deswelling was in average 74±49 µm with a mean coefficient of variation of 17%±15%. The average follow-up time was 32±29 months. The mean central corneal thickness decreased from 659±80 µm before surgery to 589±67 µm after surgery. The mean spectacle-corrected LogMAR visual acuity improved from 0.76 (20/116)±4.4 lines before surgery to 0.14 (20/27)±1.2 lines at last follow-up visit. The average time to reach 20/40 vision was 6.3±8.8 months. The average postoperative aberrations were as follows: root mean square, 0.90±0.58 µm; low-order aberrations, 0.88±0.71 µm; HOAs, 0.31±0.22 µm; HO coma, 0.15±0.13 µm; HO spherical aberrations, −0.01±0.10 µm; HO trefoil, 0.15±0.13 µm. The average postoperative corneal topography figures were as follows: simulated keratometry cylinder, 1.44±1.06 D; 3 mm irregularity, 1.94±1.12 D; 3 mm mean corneal power, 43.3±2.3 D.
Comparison of EK eyes with normal eyes showed no significant differences for root mean square (RMS), low order aberrations (LOAs) and HO spherical aberrations, whereas HOAs (0.31±0.22 µm vs 0.13±0.08 µm), HO coma (0.15±0.13 µm vs 0;08±0.06 µm) and HO trefoil (0.15±0.13 µm vs 0.06+0.02 µm) were significantly higher in EK eyes (p<0.001).
Table 1 shows the characteristics of recipients, grafts and visual outcomes in the five EK groups defined by graft thickness. The preoperative visual acuity was significantly better in the DMEK group compared with all four DSAEK groups, whereas no significant differences in this figure were found between the four DSAEK groups. The preoperative central corneal thickness was significantly lower in the DMEK group compared with all DSAEK groups except the thin DSAEK group. It was significantly higher in the conventional DSAEK group than in the thin and ultrathin DSAEK groups. No significant differences in final visual acuity were found between the five groups. ANOVA including only eyes with follow-up >6 months (n=136) did not reveal significant differences in this figure between the five groups. The time to reach 20/40 vision was significantly shorter in the DMEK and nanothin DSAEK groups compared with the remaining 3 DSAEK groups, whereas no significant differences in this figure were found between the DMEK group and the nanothin DSAEK group. ANOVA including only eyes with follow-up >6 months demonstrated the same differences. As vision improvement and recovery timing can be influenced by cataract surgery combined with EK, we stratified the analysis. Stratification did not modify the analysis results except for vision improvement in eyes with no combined phaco where the differences between groups did not reach significance. No significant differences in postoperative aberrometry and corneal topography measurements were found between the five groups except for the 3 mm irregularity that was significantly lower in the DMEK group than in the ultrathin DSAEK group.
Characteristics of grafts, recipients and visual outcomes in the five endothelial keratoplasty groups defined by graft thickness
Table 2 shows factors associated with vision recovery after EK. Poorer final visual acuity was associated with older recipients. In univariate analysis, higher final vision improvement was associated with poorer initial visual acuity, higher preoperative central corneal thickness, preoperative diagnosis other than Fuchs dystrophy, thicker graft and higher variability of graft thickness. However, in multivariate analysis only the preoperative visual acuity significantly correlated with final vision improvement (p<0.00001). Higher HOAs were associated with poorer initial visual acuity. Incision site and combined cataract surgery were not significantly associated with vision recovery.
Recipient and graft-related factors associated with vision recovery after endothelial keratoplasty
Postoperative aberrations did not significantly correlate with postoperative OCT measurements including central corneal thickness, graft thickness, coefficient of variation of graft thickness, minimal corneal thickness, minimum-maximum corneal thickness, central epithelial thickness and SD of epithelial thickness, except for low order defocus that decreased with central corneal thickness (r=−0.24, p=0.008) and graft thickness (r=−0.26, p=0.003).
Discussion
In our series, DMEK eyes featured better preoperative visual acuity and less oedematous cornea compared with DSAEK eyes showing the former were operated at an earlier stage of the disease. Previous studies reported preoperative mean LogMAR visual acuity ranging from 0.27 to 0.70 for DMEK and from 0.35 to 0.70 for DSAEK.10 18 22 23 DMEK in eyes with preoperative visual acuity >20/40 (LogMAR <0.37) is a common practice that has been reported to be as frequent as 40%.24
Final visual recovery was good in our series (LogMAR 0.14, 20/27, in average) with no significant differences between groups showing that the remaining stroma in DSAEK grafts did not significantly limit final visual recovery. Visual recovery was reported to range from 0.04 (20/22) to 0.39 (20/49) after EK.4 9 10 25 26 Compared with DSAEK, DMEK may have higher clinical potential with 75% of cases reaching 20/25 or better within 1–3 months.19 A prospective randomised clinical trial comparing DMEK with ultrathin DSAEK showed better 12-month visual acuity with DMEK but similar vision-related quality of life.26 27 We also found quicker visual recovery in DMEK eyes than in ultrathin DSAEK eyes, but final visual acuity was similar in both groups.
Wavefront HOAs, astigmatism and corneal irregularity were low in all groups showing the quality of vision was good after EK. Compared with normal eyes from young patients candidate to refractive surgery, EK eyes did not feature higher RMS, LOAs and HO spherical aberrations. As expected, HOAs, coma and trefoil were higher in EK eyes but the differences remained moderate. In fact post-EK HOAs, coma and trefoil were two to three times those of normal eyes, which can be considered as quite acceptable. In a prospective randomised study, DMEK resulted in lower 12-month posterior corneal HOAs measured with Scheimpflug imaging associated with better visual acuity compared with ultrathin DSAEK.28 However, when the combined corneal surfaces were taken into account, differences were no longer significant. DMEK was also shown to be associated with less corneal aberrations measured with Scheimpflug imaging than DSAEK and PKP.29 Whole eye wavefront HOAs may be more closely linked to quality of vision than posterior corneal HOAs as wavefront aberrometry assesses the quality of the image formed on the retina. This could explain why DMEK was not found to be associated with better vision-related quality of life despite lower 12-month posterior corneal HOAs.28 29
In 54% of cases, the patients underwent concomitant phacoemulsification in combination with EK. Improvement in wavefront data from these patients can result from cataract surgery and EK. However, as all eyes were pseudophakic after EK, potential differences in aberrations between groups could originate from differences in EK techniques and not from lens status. Further studies are needed to analyse the role of cataract surgery and EK in vision improvement after EK combined with cataract surgery.
Interestingly, graft thickness and its variability were not significantly associated with visual recovery and postoperative HOAs. The mean variability of graft thickness in DSAEK grafts ranged from 15% to 33%. This figure was less than 1% in DMEK grafts. The low difference in refractive index between cornea and aqueous humour might explain the absence of whole eye wavefront HOAs such as coma, trefoil or irregular astigmatism related to irregular posterior corneal surface.
Visual recovery was faster in eyes with DMEK grafts or nanothin DSAEK grafts compared with DSAEK grafts thicker than 50 µm. In addition, visual recovery time significantly increased with graft thickness among our 150 EK eyes. However, shorter visual recovery in DMEK eyes can be explained, at least partly, by lower preoperative corneal central thickness. The former two surgical techniques might be the best options for treating corneal endothelial disorders.9 However, a nanothin graft cannot be obtained in all cases despite improvements in microkeratome device. In the present series, only 18 out of 112 (16%) DSAEK grafts were found to be thinner than 50 µm after postoperative graft deswelling. Fast visual recovery is then more likely with the DMEK technique. The choice between DMEK and DSAEK is also influenced by the eye condition (ie, DSAEK might be more likely to be successful in eyes with previous vitrectomy, glaucoma surgery or very deep anterior chamber or anterior chamber intraocular lens).30 DSAEK may also be preferred in elderly patients with high comorbidity because the rate of rebubbling is lower compared with DMEK. However, this may not be true for the current status of DMEK, especially when 20% SF6 tamponade is used.31
The benefit:risk ratio of EK includes the risk of postoperative complications including primary graft failure, rebubbling, glaucoma, endothelial cell loss and failure, and infectious keratitis.32–34 Precise assessment of these risks and the long-term graft survival in PKP, DSAEK and DMEK eyes is needed before precise indication for EK can be determined. A recent study reported better graft survival, lower rate of graft rejection and elevated intraocular pressure after DMEK compared with DSAEK and PKP.35
The association between lower preoperative visual acuity and higher postoperative HOAs is an argument for early surgery in corneal endothelial disorders. In a retrospective study, preoperative visual acuity values below 20/100 were associated with poorer 12-month visual recovery after DMEK.36 Furthermore, >60% of eyes with preoperative visual acuity of 20/40 reached postoperative visual acuity above 20/25 compared with 40% of eyes with preoperative visual acuity of 20/200. In our series, better preoperative visual acuity did not correlate with final visual acuity but it was associated with rapid visual recovery and lower HOAs. In a prospective observational study, preoperative corneal central thickness <625 µm was associated with better 12-month visual acuity after DMEK. In our series, lower preoperative corneal central thickness did not correlate with final visual acuity but it was associated with rapid visual recovery. Finally, preoperative corneal backscatter was shown to correlate with postoperative visual recovery after DMEK. All these findings are relevant arguments for early EK in eyes with endothelial disorders.
Older recipients had poorer final visual acuity despite absence of pre-existing visual comorbidities. This finding is likely to be explained by loss of photoreceptors with ageing, which makes the visual system less efficient in elderly patients.37 38
Study limitations include the size of the study population and the retrospective design with no randomisation of treatments. In addition, a substantial selection bias is associated with DMEK patients who had less degrees of corneal oedema preoperatively. However, our study corresponds to real life data with extended follow-up of patients and precise assessment of graft thickness and regularity and quality of vision.
In conclusion, the main advantage of DMEK and nanothin DSAEK over DSAEK thicker than 50 µm was the rapidity of visual recovery. Final quality of vision was not influenced by graft thickness and regularity.
References
Footnotes
Contributors MY and YC contributed to collection and/or assembly of data, data analysis and interpretation, manuscript writing and final approval of manuscript. JP contributed to collection and/or assembly of data and final approval of manuscript. CG contributed to collection and/or assembly of data, data analysis and interpretation and final approval of manuscript. SK and CT contributed to provision of study material or patients and final approval of manuscript. NB contributed to provision of study material or patients, collection and/or assembly of data, data analysis and interpretation, manuscript writing and final approval of manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests VB, Chiesi (Parma, Italy; consulting), Dompe (Milan, Italy; consulting), Optovue (Fremont, California, USA; consulting).
Patient consent for publication Obtained.
Ethics approval Approved by the Ethics Committee of the French Society of Ophthalmology and adhered to the tenets of the Declaration of Helsinki.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article.