Evaluation of interface reflectivity and corneal aberrations following Descemet's stripping automated endothelial keratoplasty.

PURPOSE: This study aims to evaluate visual outcome after Descemet's membrane stripping automated endothelial keratoplasty (DSAEK) and relate to interface and corneal higher-order aberrations (HOAs).
MATERIALS AND METHODS: We enrolled 16 eyes of 16 patients (eight males and eight females) in this interventional case series and followed DSAEK operation for about two to 20 months. OCULUS Pentacam, as well as other ophthalmic evaluations in follow-up visits, examined interface reflectivity and HOAs. Statistical relations were analyzed.
RESULTS: There was statistically significant correlation between interface reflectivity and best corrected visual acuity (BCVA) (r = 0.56, P = 0.021). Pachymetry (central corneal thickness) and BCVA had a moderate correlation (r = 0.6, P = 0.013). There was no statistically significant correlation between pachymetry and follow-up time (r = -0.36, P = 0.16). Negative correlation between follow-up and interface reflectivity was also not statistically significant (r = -0.24, P = 0.35). Coma had a significant correlation with BCVA in cornea and cornea front maps (r = 0.74, P = 0.009 and r = 0.71, P = 0.013, respectively).
CONCLUSION: Significant correlation between interface reflectivity and BSCVA was found, and anterior corneal HOAs are significantly higher than posterior HOAs.

Introduction

Over the past decade, advances in surgical techniques and instruments in the field of corneal transplantation have led this impressive procedure to gain admirable results for the patients.[12345678] Descemet's membrane stripping automated endothelial keratoplasty (DSAEK) proposed as an alternative to penetrating keratoplasty (PK) in cases of dysfunctional endothelium involves mechanical stripping of diseased host endothelium and Descemet's membrane and replacement with a donor graft of endothelium, Descemet's membrane and a thin layer of posterior stroma.[9101112]

Performing DSAEK for corneal endothelial dysfunction offers many advantages over PK, including intraoperative closed-system conditions, minimal change of preoperative refraction, faster postoperative visual rehabilitation, and elimination of all suture-related late postoperative complications.[1314] However, transplantation of the posterior stroma leads to an increase in corneal thickness, the creation of host-donor interface that can influence postoperative visual performance.[151617]

The purpose of this study is to describe the visual outcome after DSAEK surgery with regard to host-donor interface and higher-order aberrations (HOAs).

Materials and Methods

We enrolled sixteen patients (16 eyes of 8 men and 8 women) with surgical indications such as bullous keratopathy, Fuchs endothelial dystrophy, and congenital hereditary endothelial dystrophy. All patients underwent DSAEK procedure at Navid Didegan Eye Clinic and University Hospital, Khatam Al Anbia, under-monitored general anesthesia. Donor tissue was supplied by a single eye bank (Iran Cornea Bank, Tehran, Iran). All grafts had an endothelial cell count of at least 2500 cells/mm2. The posterior lamellar grafts were either supplied as precut tissue from the Cornea Bank. All surgeries were performed by single surgeon (MN) and were uneventful in all patients. The entire procedure was performed with the surgeon sitting at the temporal position. An anterior chamber maintainer was used to prevent anterior chamber collapse during surgery. Descemet's membrane was stripped from the central 7.5–8.0 mm.

DSAEK grafts were punched to a diameter of 7.5–8.0 mm before surgery. The donor lamellae were folded and inserted into the anterior chamber with the so-called taco technique. No intraoperative complications, affecting the outcome, occurred. After the operation, all patients underwent a complete ophthalmologic examination, including best spectacle-corrected visual acuity (BSCVA), complete slit-lamp examination, measurement of intraocular pressure, fundus examination, and when necessary, B-scan ultrasonography at first, third, and 12 months' postoperative visits. We examined all of the eyes at different time-points, from the 2nd month after surgery to 20 months using Pentacam Scheimpflug imaging system (OCULUS Gmbh, Wetzlar, Germany). Reflectivity of the epithelium, endothelium and interface region (between graft and host cornea) in the center of optical axis was determined. HOA maps of cornea, back, and front were measured by means of the built-in software of the pentacam system using Zernike polynomials. Zernike coefficients of the central 6 mm and the root mean square value of HOAs (cornea, anterior, and posterior surfaces) were calculated. No postoperative complications including graft dislocation were recorded.

Data analysis was performed using available software SPSS for Windows (version 11.5, SPSS Inc., Chicago, Illinois, USA). The correlation was studied by means of Pearson correlation coefficient because the distribution of all values was normal. The difference was considered statistically significant if the P < 0.05. A correlation of 0.4–0.7 regarded as medium, 0.7–0.9 good, 0.9 high, and a value <0.4 indicated low.

Results

The mean age was 62.69 (ranged from 13 to 78 years). Table 1 provides demographic data of the study population. At the final visit, BSCVA had increased to 0.63 (range 0.1–1.6 log of the minimum angle of resolution).

Table 1

Demographic data of study population

Number of patients (eyes)	16
Male:female	8:8
OD:OS	6:10
Age (years)	13-78 (mean 62.69)
Preoperative visual acuity (logMAR)	0.8-3 (mean 1.97)
Donor cell count	2740-3937 (mean 3082.2)
Death to excision time (h)	19-399 (mean 30.75)
Death to operation time (day)	3-5 (mean 3.88)
Follow-up time (month)	1-20 (mean 7.44)
Central thickness	515-950 (mean 635.69)
BCVA (logMAR)	0.1-1.6 (mean 0.63)
Epithelial reflectivity	32-76.3 (mean 46.9)
Interface reflectivity	16.3-100 (mean 31.89)
Endothelial reflectivity	9.3-41.7 (mean 14.6)
Patient cell count at follow up	1053-2468 (mean 1879.8)
Posterior lenticule (mm)	50-160 (mean 118.1)
Cornea diameter (mm)	10.5-12.2 (mean 11.28)

BCVA: Best-corrected visual acuity, logMAR: Log of the minimum angle of resolution

Analyses on Pentacam Scheimpflug images [Figure 1] with three peaks (epithelium, interface, and endothelium) showed interface reflectivity ranged between 16.3 and 100 (mean 31.89) and the epithelium reflectivity was higher [Table 1]. There was statistically significant correlation between interface reflectivity and best-corrected visual acuity (BCVA) (r = 0.56, P = 0.021) (Pearson correlation coefficient analysis) [Figure 2]. There was also such a correlation with endothelial reflectivity (r = 0.59, P = 0.016). We recorded the highest intensity of reflectivity from epithelium (46.9); however, there was no statistically significant correlation between epithelial reflectivity and BCVA.

Figure 1

Iris image and interface reflectivity 12 months after Descemet's stripping automated endothelial keratoplasty captured by Pentacam

Figure 2

Relationship between interface reflectivity and best corrected visual acuity (P= 0.021)

Pachymetry (central corneal thickness) and BCVA had a moderate correlation (r = 0.6, P = 0.013) but a good positive correlation was found with interface reflectivity (r = 0.75, P = 0.001) [Figure 3]. There was no statistically significant correlation between pachymetry and follow up time (r = −0.36, P = 0.16). Negative correlation between follow-up and interface reflectivity was also not statistically significant (r = −0.24, P = 0.35) [Figure 4].

Figure 3

Correlation between pachymetry and interface reflectivity (P = 0.001)

Figure 4

Negative correlation between follow-up and interface reflectivity (P = 0.35)

The mean anterior and posterior corneal total HOAs of the central 6 mm were 1.276 ± 0.392 and 0.674 ± 0.160, respectively [Table 2]. There was no correlation between all of the Zernike terms from the total corneal HOAs and BCVA except coma that had a significant correlation with BCVA in total cornea and anterior cornea HOAs maps (r = 0.744, P = 0.009 and r = 0.716, P = 0.013, respectively).

Table 2

Anterior, posterior and total corneal aberrations in central 6 mm zones in Descemet’s stripping automated endothelial keratoplasty

6 mm	Mean±SD	Minimum	Maximum
RMS CF (total)	1.276±0.392	0.732	1.904
RMS coma	0.631±0.390	0.246	1.658
RMS trefoil	0.574±0.371	0.1	1.153
RMS spherical aberration	0.489±0.217	0.163	0.783
RMS astigmatism	0.278±0.254	0.061	0.708
RMS 4-foil	0.552±0.302	0.092	1.04
RMS CB (total)	0.674±0.160	0.482	0.981
RMS coma	0.191±0.108	0.044	0.377
RMS trefoil	0.368±0.218	0.105	0.897
RMS spherical aberration	0.220±0.107	0.087	0.371
RMS astigmatism	0.146±0.070	0.033	0.279
RMS 4-foil	0.356±0.206	0.145	0.752
RMS cornea (total)	1.310±0.367	0.797	1.958
RMS coma	0.620±375	0.284	1.678
RMS trefoil	0.544±0.305	0.227	1.105
RMS spherical aberration	0.527±0.231	0.208	0.868
RMS astigmatism	0.325±0.249	0.048	0.714
RMS 4-foil	0.632±0.249	0.059	0.914

Bold numbers indicate statistically significant relation with BCVA. RMS: Root mean square; CF: Cornea front; CB: Cornea back, BCVA: Best-corrected visual acuity, SD: Standard deviation

Discussion

Increasing advances of ophthalmology have led to surgical procedures with different capabilities. Descemet's stripping automated endothelial keratoplasty is a surgical method that is performed as an alternative to PK. A greater understanding of various aspects of this approach can be helpful to improve the outcomes. We wanted to observe the impact of this extra tissue on visual acuity, interface, and HOAs that are effective on patient's quality of vision.

Incomplete vision in some patients despite good postoperative results suggests other limiting factors. Some possible explanations are noted including, increased total corneal thickness, increased corneal aberrations of the posterior surface, subepithelial corneal haze, and interface creation with increased reflectivity, that may influence the visual outcome.

To evaluate the effect of recipient corneal thickness (RCT) on visual outcome after DSAEK, Ivarsen and Hjortdal evaluated RCT and corneal densitometry of 125 eyes with anterior-segment optical coherence tomography and Pentacam. In that survey, 3 years after operation, changes in RCT, densitometry, and CDVA were significantly correlated.[18] In contrast to Ivarsen and Hjortdal study, Ahmed et al. showed that thicker corneas and grafts are not associated with worse visual acuity.[19] They also revealed that stromal edema resolves by 3 months after DSEK for Fuchs dystrophy, whereas visual acuity continues to improve through 12 months. A study conducted by Pogorelovin Germany to quantify changes in donor and host corneal tissue after DSAEK using organ-cultured corneas and compared these findings with the visual acuity. Data showed a continuous decrease of thickness of the grafted lenticule during follow-up. The central corneal thickness decreased from 1057 μm at the first postoperative day to 661 μm after 6 months. Both central corneal thickness and the thickness of the posterior donor lamella correlated with the 6-month BSCVA.[20] In another study, Uchino et al. compared corneal thickness and haze in DSAEK and PK and showed that DSAEK grafts are significantly thicker with increased haze for up to 3 months after surgery. Graft evaluation with Scheimpflug system showed that haze in the DSAEK group was mainly observed in the subepithelial stroma.[21] In our study, central corneal thickness was 515–950 μm (mean 635.69). We found no statistically significant correlation between pachymetry (central corneal thickness) and follow-up time (r = −0.36, P = 0.16) but pachymetry and BCVA had a moderate correlation (r = 0.6, P = 0.013).

Another possible explanation for limited visual outcome after DSAEK is corneal HOA. In our study on Zernike polynomial decomposition, anterior corneal HOAs were significantly higher than posterior HOAs [Figure 5]. For anterior corneal HOAs, coma was the most important aberration component followed by trefoil, whereas for posterior corneal HOAs, trefoil was the most important aberration component followed by tetrafoil [Table 2].

Figure 5

Scheimpflug images by Pentacam system demonstrate corneal (a), anterior corneal (b), and posterior corneal higher-order aberrations in an eye underwent Descemet's stripping automated endothelial keratoplasty (c)

Conversely, Muftuoglu et al. showed that posterior corneal HOAs are significantly higher in eyes that underwent DSAEK. This difference may be related to different surgical indication that was mainly Fuchs' dystrophy in Muftuoglu et al. study and pseudophakic bullous keratopathy in our study. They also showed that coma and trefoil were the most important aberration component for anterior and posterior corneal HOAs, respectively. Their results of important aberration components are compatible with our study.[22]

Post-DSAEK visual outcome might also be influenced by corneal haze. Corneal haze may be located in subepithelial or interface. Weis et al. evaluated etiology of haze at the graft-host interface in a cat model. They showed that interface haze was not associated with myofibroblast differentiation at the graft-host interface, but rather with apoptosis and the development of a subsequent acellular zone.[23] Electron microscopy examination of interface opacity also demonstrated collagen fibrils of varying length that extended from the anterior donor corneal stromal surface. These collagen fibrils were absent on the anterior stromal surface of the control corneal disc.[24] Espana and Huang evaluated 25 eyes with in vivo confocal microscopic, and they showed a significant negative correlation between postoperative BSCVA and subepithelial corneal haze. This haze was located between the epithelial basement membrane and the anterior stroma below the Bowman's layer. However, in their study, donor-recipient interface haze did not correlate with BSCVA. In their study, the etiology for corneal edema requiring corneal transplantation was mainly Fuchs' dystrophy.[25] This finding is in contrast with what was found in a study from Italy that revealed a good quality interface is related to a better visual acuity and improves with time from surgery.[26]

The other statement that may cause limited vision after DSAEK is the presence of interface. To reveal the effect of donor-recipient interface, previous studies evaluated interface by confocal microscopy and pentacam. In a study from Germany, interface reflectivity was assessed with pentacam (mean 25.8) and showed statistically significant correlation between interface reflectivity and BSCVA as well as follow-up time. There was no statistically significant correlation between interface reflectivity and central corneal thickness.[27] As mentioned above, we also showed a significant correlation between interface reflectivity and BSCVA but not with follow-up time. In our study, mean interface reflectivity was higher than their result (31.89 vs. 25.8). We also measured epithelial and endothelial reflectivity (mean 46.9 and 14.6, respectively) and showed no statistically significant correlation between epithelial reflectivity and BCVA.

Conclusion

In the present study, we evaluated the interface quality and corneal HOAs following DSAEK surgery with pentacam. Similarly to others, we found significant correlation between interface reflectivity and BSCVA. In the evaluation of corneal aberration, our study revealed that anterior corneal HOAs are significantly higher than posterior corneal HOAs. There was no significant correlation between any of the Zernike terms from total and anterior corneal HOAs of the central 6-mm zones and BCVA except coma of the total cornea and anterior cornea HOAs maps that had a significant correlation with BCVA. Further studies with larger sample size are needed to evaluate the possible effects of interface reflectivity and HOAs on visual quality after DSAEK surgery.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.