Descemet-stripping automated endothelial keratoplasty (DSAEK) has recently replaced penetrating keratoplasty as the treatment of choice in the treatment of corneal endothelial dystrophies because of its lower rate of complications and shorter recovery time.1–3 Reported interface complications of DSAEK include epithelial ingrowth, infection, persistent fluid, and retained Descemet membrane.4–11
We report 14 cases of reticular haze of unclear etiology at the donor–recipient interface without stromal infiltration. This complication has been termed textural interface opacity.12 We further classify textural interface opacity as total if it involves the whole interface area or partial if it involves only a portion of the interface. In this retrospective case series, we describe the clinical course, possible etiologies, and treatment strategies for total and partial textural interface opacity and the visual and graft outcomes.
Patients and methods
This retrospective study, approved by the Duke University Health System Institutional Review Board, reviewed charts of patients seen by the cornea service from September 2009 to May 2012. Clinical information collected included corrected distance visual acuity (CDVA) preoperatively and at the last clinic visit, donor tissue information (graft endothelial cell count [ECC], graft thickness, graft bed size, graft size), surgical technique, pathology reports (when available), and anterior optical coherence tomography (AS-OCT) (Visante, Carl Zeiss Meditec Inc.) and handheld Fourier-domain OCT (Bioptigen) (when available).
Descemet-stripping automated endothelial keratoplasty was performed in the usual way. An 8.0 mm ring marker was used to mark the cornea, and superior and inferior paracenteses were created with a side-port blade. Preservative-free lidocaine diluted with a balanced salt solution and epinephrine was injected into the anterior chamber for anesthesia; this was followed by injection of sodium hyaluronate 1.0% (Provisc). If the patient had concurrent phacoemulsification with intraocular lens (IOL) implantation, sodium hyaluronate 3.0%–chondroitin sulfate 4.0% (Duovisc) with sodium hyaluronate 1.0% was used instead. A Gorovoy irrigating stripper cannula (Harvey Precision Instruments) was used to score and strip Descemet membrane. In 1 eye, trypan blue was injected into the anterior chamber to visualize remaining Descemet membrane. The irrigation/aspiration unit was used to aspirate the ophthalmic viscosurgical device (OVD) and residual Descemet membrane. All donor corneal tissues were obtained from Ocular System, Inc., and precut with a Moria microkeratome within 5 days of the DSAEK procedure. The tissue was preserved in a chondroitin sulfate–based corneal storage medium (Optisol GS, Bausch & Lomb). A 7.5 to 8.5 mm Ultrafit trephine (Angiotech) was used to cut the donor cornea. The donor cornea was then irrigated with a balanced salt solution and placed in the moisture chamber for protection. The graft was inserted with an Endoserter device (Ocular Systems, Inc.), insertion forceps, or straight cystotome. Full air tamponade was kept for 8 minutes before partial removal of the air bubble. Intraocular pressure was then measured with a Tono-Pen (Reichert Technologies) 1 hour after the procedure and adjusted to normal by burping the paracentesis incision; it was also measured at all postoperative visits.
The patient was placed on antibiotics and prednisolone acetate 1.0%. The prednisolone doses ranged from 1 drop every 1 to 2 hours for 1 to 2 weeks followed by a slow taper over several months.
The patients who had repeat DSAEK procedures had the graft specimen retrieved for histopathologic analysis. Hematoxylin–eosin (H&E), Masson trichrome, periodic acid-Schiff (PAS), and colloidal iron stain were performed.
Four cornea fellowship–trained physicians identified 14 eyes of 14 patients with total or partial textural interface opacity. The mean age of the 4 men and 10 women was 74 years (range 61 to 88 years). The indication for DSAEK included Fuchs endothelial dystrophy, pseudophakic bullous keratopathy, and both diseases (Table 1). Seven eyes had visually significant cataract with Fuchs endothelial dystrophy. Six eyes had other ocular disease (Table 1).
Of the 14 cases, 11 eyes developed total textural interface opacity (Figure 1) and 3 eyes developed partial textural interface opacity. Partial textural interface opacity (ie, reticular interface texture through a portion of the interface) was found in the superior half of 1 cornea, temporal half of 1 cornea, and inferior half of 1 cornea (Figure 2). No interface fluid or space was noted on slitlamp examination. The mean time of onset of textural interface opacity after DSAEK was 2.2 weeks (range 1 day to 7 weeks). All cases of partial textural interface opacity were observed 1 week postoperatively. The mean follow-up after DSAEK was 6.4 months (range 1 week to 2 years).
The mean donor tissue ECC was 2911.7 cells/mm2 (range 2754 to 3115 cells/mm2), the mean donor total corneal thickness was 508.6 μm (range 414 to 566 μm), the mean endothelial graft thickness was 106.3 μm (range 90 to 117 μm), and the mean bed size was 9.75 mm (range 9.0 to 10.5 mm). The mean graft size was 8.25 mm (range 7.5 to 8.5 mm). No correlation was found between the time of donor tissue being cut and DSAEK surgery with the occurrence of textural interface opacity. The Endoserter device was used to deliver the donor cornea in 11 eyes, including all eyes with partial textural interface opacity; an insertion forceps was used in 2 eyes; and a cystotome needle was used in 1 eye.
Seven eyes had concurrent phacoemulsification with IOL implantation and DSAEK, and 7 eyes were pseudophakic at the time of DSAEK. Four of the 14 patients with textural interface opacity had DSAEK in both eyes. Of those 4 patients, total textural interface opacity was observed in 3 eyes and temporal partial textural interface opacity in 1 eye.
The CDVA ranged between 20/30 and 20/400 preoperatively and between 20/25 and 20/80 at the last follow-up clinic visit. Nine eyes had regression of textural interface opacity, with a CDVA of better than 20/50. In 3 eyes with total textural interface opacity, persistent haze, and a CDVA worse than 20/60, a decision was made to repeat the DSAEK procedure (at 2 months, 10 months, and 6 months, respectively). The 3 eyes with repeat DSAEK eyes had good outcomes, with a CDVA of 20/25 in 2 eyes and of 20/60 in 1 eye 7 months after the repeat procedure. There was no evidence of recurrence in any of these patients at the last follow-up (Table 2 and Figure 3).
Histopathology of the removed grafts in 3 eyes did not show fibrosis, deposits, epithelial ingrowth, or inflammatory cells. Staining with Masson trichrome, PAS, and colloidal iron also showed no abnormalities.
Anterior Segment Optical Coherence Tomography
Anterior segment OCT was performed in 2 eyes with total textural interface opacity (Table 3). The images showed mild linear reflectivity (Figure 4), which appeared to be more prominent than in eyes that did not develop textural interface opacity after DSAEK. No interface fluid or space was seen on AS-OCT imaging.
We describe a different type of interface complication after DSAEK; the complication is also known as textural interface opacity. Of our 14 patients, 1 was diagnosed with total textural interface opacity 1 day postoperatively; the others were diagnosed with textural interface opacity later, with a mean onset of 2.2 weeks. Given the marked corneal edema that occurs after DSAEK, it is often difficult to visualize the interface until the edema resolves. Textural interface opacity was likely present in these cases soon after the surgery; however, the interface haze was visualized only after the corneal edema resolved.
Most eyes improved and had a CDVA of better than 20/50 during the follow-up. Three eyes with total textural interface opacity had a CDVA worse than 20/60 and had repeat DSAEK with no recurrence of the opacity. One patient had a CDVA of 20/60, which we believe was secondary to other factors, such as graft thickness and shape, as well as to graft–host interface alignment. Our experience suggests that if the CDVA has not improved by 6 months because of textural interface opacity, repeat DSAEK is a viable treatment option. One patient had repeat DSAEK at 2 months because the patient was unhappy with the quality of vision and did not want to wait 6 months for repeat surgery. Although the etiology of this condition is unknown, our study results indicate that the graft ECC, graft thickness, graft bed size, and graft size do not seem to be associated with the occurrence of textural interface opacity. Interface epithelial ingrowth and downgrowth, interface fungal keratitis, interface blood, interface vascularization, persistent lamellar interface fluid, retention of host embryonic Descemet membrane, and calcareous degeneration as causes of interface haze or opacity have been reported as possible DSAEK complications.4–15 However, these causes have been localized to focal regions of the cornea. This is in contrast to textural interface opacity, which was found diffusely over the total or partial interface.
Histologic sections of removed grafts of 3 patients with textural interface opacity who had repeat DSAEK did not show epithelial ingrowth, plastic deposits, fibrosis, retained OVD, or retained Descemet membrane. Furthermore, no PAS-positive material, abnormal collagen, or glycosaminoglycan accumulation was identified. Polarized microscopy of the H&E-stained sections did not show irregular patterns at the interface that would correlate with reticular texture. The accumulated materials at the interface could have been destroyed during histologic tissue processing.
Interface haze after DSAEK has been reported recently in the literature.12,16–18 Kymionis et al.16 and deSanctis et al.17 report interface wavelike deposits after DSAEK in 2 patients each. In every case, a trial of topical steroids was attempted but the deposits persisted. Visual acuity was excellent in 2 of the cases (20/20 and 20/25) but remained poor in the other 2 cases (20/70 and 20/200). In another series of 5 eyes, it was suggested that the interface haze resulted from retained OVD.18 Three of the 5 eyes had eventual resolution of the haze with a CDVA of 20/40 or better, while 2 had surgery to irrigate the remaining OVD from the interface.
A larger case series by Vira et al.,12 who termed the condition “textural interface opacity,” identified 2 subtypes. The elongated interface haze may be secondary to an irregular cut of the donor tissue by the microkeratome, causing a disruption in stromal fibers. The punctate type may be from retained OVD. Vira et al. report excellent CDVA outcomes (20/40 or better except in 1 patient, whose vision was limited by posterior pole pathology). One eye had repeat DSAEK. No patient in our study had the punctate variety. As in our patients, histopathologic examination did not show inflammatory cells. Anterior segment OCT imaging in these studies were consistent with our findings, showing an area of increased hyperreflectivity in the graft–host interface.12,17,18
Retained OVD or a chemical reaction of the interface with the OVD material may cause textural interface opacity. We do not believe that such a reaction is inflammatory. On AS-OCT imaging, there was no obvious collection of OVD or inflammation in the interface. Histopathology of the interface did not show inflammatory cells, and retained material may have been destroyed by processing. Since the appearance of textural interface opacity in the patients in this study, we have changed our cohesive OVD to sodium hyaluronate 1.4% (Healon GV). Whether the more cohesive nature of this OVD material allows more complete removal from the eye or whether as a result of the properties of the material itself, we have not had further cases of textural interface opacity. Another technique would be to strip Descemet membrane with air or a balanced salt solution in the anterior chamber. Although technically more difficult, this change would avoid the issue of retained or reacting OVD.
Another possible etiology of the interface texture is retained microkeratome-generated plastic particles during preparation of the corneal tissue. Microkeratome-generated plastic particles have been associated with interface haze after laser in situ keratomileusis.19 A previous study20 found that these particles have a characteristic hyperreflectivity on slitlamp examination and confocal microscopy. However, these hyperreflective particles, although clinically visible, did not affect the final visual acuity. We, however, did not find plastic particles on the histopathology. Also, this etiology is less likely given that the condition regressed in our patients and would not likely regress if it were related to plastic particles.
This study has a small sample of 14 eyes, thus limiting statistical analysis. A prospective randomized control study using different OVDs and with a larger number of patients is necessary to determine the underlying cause of textural interface opacity. Other ancillary imaging studies, such as confocal microscopy and scanning electron microscopy of the graft–host interface, may provide a better understanding of this complication.
In conclusion, textural interface opacity is a rare complication that can occur in the early period after DSAEK and can affect the CDVA. In 9 eyes with total or partial textural interface opacity, the CDVA was better than 20/50 and the opacity spontaneously resolved without increasing the steroid dose. Total textural interface opacity in 3 eyes was associated with a CDVA of worse than 20/60, with the eyes requiring repeat DSAEK for persistent poor visual acuity and textural interface opacity that did not resolve. After repeat DSAEK, the CDVA improved to better than 20/25. Although the etiology of textural interface opacity is unclear, our study suggests that repeat DSAEK can improve the final visual outcome and may be an effective treatment.
What Was Known
- Patients having DSAEK can present with interface complications. Previously reported complications include epithelial downgrowth, infection, lint and plastic deposits, and retained Descemet membrane.
- Recent results suggest that interface haze does not seem to be caused by the aforementioned complications.
What This Paper Adds
- Anterior segment OCT imaging and pathology confirmed that this interface haze, termed textural interface opacity, is not caused by the foreign bodies listed above.
- Textural interface opacity is likely related to retained OVD or a property of the OVD used during the surgery.
- Textural interface opacity may be self-limiting; however, repeat DSAEK was an effective treatment for refractory cases.
1. Bahar I, Kaiserman I, McAllum P, Slomovic A, Rootman D. Comparison of posterior lamellar keratoplasty techniques to penetrating keratoplasty. Ophthalmology
2. Koenig SB, Covert DJ. Early results of small-incision Descemet’s stripping and automated endothelial keratoplasty. Ophthalmology
3. Gorovoy MS. Descemet-stripping automated endothelial keratoplasty. Cornea
4. Shih CY, Ritterband DC, Rubino S, Palmiero P-M, Jangi A, Liebmann J, Ritch R, Seedor JA. Visually significant and nonsignificant complications arising from Descemet stripping automated endothelial keratoplasty. Am J Ophthalmol
5. Suh LH, Shousha MA, Ventura RU, Kieval JZ, Perez VL, Wang J, Dubovy SR, Rosenfeld SI, Culbertson WW, Alfonso E, Forster RK. Epithelial ingrowth after Descemet stripping automated endothelial keratoplasty: description of cases and assessment with anterior segment optical coherence tomography. Cornea
6. Suh LH, Dawson DG, Mutapcic L, Rosenfeld SI, Culbertson WW, Yoo SH, O’Brien TP, Dubovy SR. Histopathologic examination of failed grafts in Descemet’s stripping with automated endothelial keratoplasty. Ophthalmology
7. Suh LH, Yoo SH, Deobhakta A, Donaldson KE, Alfonso AC, Culbertson WW, O’Brien TP. Complications of Descemet’s stripping with automated endothelial keratoplasty; survey of 118 eyes at one Institute. Ophthalmology
8. Gorovoy MS, Ratanasit A. Epithelial downgrowth after Descemet stripping automated endothelial keratoplasty. Cornea
9. Lee WB, Foster JB, Kozarsky AM, Zhang Q, Grossniklaus HE. Interface fungal keratitis after endothelial keratoplasty: a clinicopahological report. Ophthalmic Surg Lasers Imaging
10. Hurmeric V, Wang J, Kymionis GD, Yu SH. Persistent lamellar interface fluid with clear cornea after Descemet stripping automated endothelial keratoplasty. Cornea
11. Chen ES, Shamie N, Terry MA, Phillips PM, Wilson DJ. Retention of host embryonic Descemet membrane in endothelial keratoplasty. Cornea
12. Vira S, Shih CY, Ragusa N, Sheyman A, Feder R, Weisenthal RW, Rosenwasser GO, Hannush SB, Udell IJ, Bouchard CS. Textural interface opacity after Descemet stripping automated endothelial keratoplasty: a report of 30 cases and possible etiology. Cornea
13. Ebrahimi KB, Oster SF, Green WR, et al. Donor corneal stroma and host–donor interface vascularization after Descemet’s membrane stripping with automated endothelial keratoplasty [letter]. Acta Ophthalmol. 88, 2010, p. e7-e8, Available at: http://onlinelibrary.wiley.com/doi/10.1111/j.1755-3768.2009.01508.x/pdf
. Accessed April 30, 2014.
14. Schmitt AJ, Feilmeier MR, Piccoli FV, Ide T, Yoo SH. Interface blood after Descemet stripping automated endothelial keratoplasty. Cornea
15. Ebrahimi KB, Oster SF, Green WR, Grebe R, Schein OD, Jun AS. Calcareous degeneration of host–donor interface after Descemet membrane stripping automated endothelial keratoplasty. Cornea
16. Kymionis GD, Ide T, Yoo SH. Interface wavelike deposits after Descemet stripping automated endothelial keratoplasty. Arch Ophthalmol. 127, 2009, p. 1389-1390, Available at: http://archopht.jamanetwork.com/data/Journals/OPHTH/10122/esc1009_1389_1390.pdf
. Accessed April 30, 2014.
17. de Sanctis U, Brusasco L, Grignolo F. Wave-like opacities at the interface after Descemet stripping automated endothelial keratoplasty. Cornea
18. Anshu A, Planchard B, Price MO, da R Pereira C, Price FW Jr. A cause of reticular interface haze and its management after Descemet stripping endothelial keratoplasty. Cornea
19. Ivarsen A, Thφgersen J, Keiding SR, Hjortdal J.ϕ, Mφller-Pedersen T. Plastic particles at the LASIK interface. Ophthalmology
20. Dawson DG, Holley GP, Geroski DH, Waring GO III, Grossniklaus HE, Edelhauser HF. Ex vivo confocal microscopy of human LASIK corneas with histologic and ultrastructural correlation. Ophthalmology