Outcomes of Corneal Transplantation for Irreversible Corneal Decompensation Secondary to Corneal Endotheliitis in Asian Eyes


To describe outcomes of corneal transplantation for irreversible corneal decompensation from corneal endotheliitis in Asian eyes.


Retrospective, observational case series.


We reviewed consecutive patients with corneal endotheliitis (32 eyes of 31 subjects) who underwent keratoplasty (January 1, 2008-December 1, 2009). All eyes had preoperative aqueous polymerase chain reaction (PCR) analysis for viruses, including cytomegalovirus (CMV). CMV-positive patients were treated preoperatively with topical corticosteroids and anti-CMV treatment (oral valganciclovir 900 mg twice daily, topical ganciclovir 0.15% 5 applications per day, for 6 weeks) with complete resolution of ocular inflammation, and quiescence for at least 6 months before corneal transplantation. Our main outcome measure was recurrence of endotheliitis within 1 year after corneal transplantation.


Five eyes were CMV positive; the remaining 27 eyes were negative for all viruses on PCR analysis. CMV-positive patients had a higher rate of recurrence of endotheliitis within 1 year after corneal transplantation, compared with CMV-negative eyes (60% vs 7.4%, P = .01). The CMV-positive eyes had recurrent endotheliitis at a median of 10 months (range 3-11 months) after corneal transplantation. After successful anti-CMV treatment, all 5 CMV-positive eyes then continued to have clear grafts for a median duration of 21 months (range 13-44 months).


Our study suggests that Asian patients with corneal endotheliitis may benefit from preoperative aqueous PCR analysis before corneal transplantation. Such patients were more likely to have a recurrence of endothelial inflammation if they were CMV positive preoperatively, despite successful anti-CMV treatment before surgery.

Corneal endotheliitis is an entity characterized by localized corneal edema with the presence of mild iritis and keratic precipitates (KPs). Although corneal endotheliitis was once considered “autoimmune” or “idiopathic,” there have been increasing reports that viruses such as cytomegalovirus (CMV) and herpes simplex virus (HSV) may play a role as etiologic agents. Polymerase chain reaction (PCR) analysis of aqueous humor has been shown to be capable of detecting the virus ; however, real-time PCR and quantification of viral DNA may be useful in distinguishing an etiologic virus, as opposed to a “bystander” virus (the latter does not play a causative role in inciting inflammation or endotheliitis). A viral etiology in corneal endotheliitis is further supported by reports that endothelial lesions and inflammation resolve rapidly with the application of topically applied and systemic antiviral treatment.

Cytomegalovirus is increasingly recognized as the most common virus implicated in corneal endotheliitis, especially in Asian patients. Although the exact pathogenesis of CMV corneal endotheliitis is not yet understood, it is postulated that anterior chamber–associated immune deviation (ACAID) normally prevents the proliferation of CMV within the anterior chamber (AC), leading to viral infection of the endothelium. However, it remains unclear whether CMV endotheliitis is attributable to a primary CMV infection or a local reactivation of CMV in the anterior segment secondary to dendritic cell activation provoked by a separate trigger, seen in other parts of the body.

Although medical therapy remains the mainstay of treatment, persistent corneal edema or irreversible corneal decompensation often ensues because of the cytopathic effect on corneal endothelial cells, leading to the need for a corneal graft. In this study, we describe the outcome of various forms of corneal transplantation, but mainly Descemet stripping automated endothelial keratoplasty (DSAEK) performed in eyes that presented with endotheliitis and corneal decompensation. We also compared the outcomes between eyes that were “CMV positive” and “CMV negative,” as assessed using real-time PCR preoperatively.


Study Cohort

We conducted a retrospective review of consecutive patients who presented with corneal endotheliitis (ie, localized corneal edema with mild AC inflammation and KPs ) who subsequently underwent corneal transplantation. All patients underwent corneal transplantation at the Singapore National Eye Centre (SNEC) between January 1, 2008 and December 1, 2009. We obtained clinical data from our ongoing Singapore Corneal Transplant Study cohort, an audited longitudinal prospective study that contains preoperative, intraoperative, and yearly postoperative follow-up clinical data on all corneal transplants performed by corneal surgeons at the SNEC. We conducted this study with approval from our institution’s Singapore Health Services Institutional Review Board and followed the principles of the Declaration of Helsinki.


At presentation, all eyes with suspected corneal endotheliitis with localized corneal edema, KPs, and mild AC inflammation had an AC tap performed preoperatively at the slit lamp using an aseptic technique. A minimum of 100 μL of aqueous was obtained and the samples were either delivered to the laboratory at 18 C-25 C within 1 hour, or stored and transported at 2 C-4 C within 24 hours. The samples were aliquoted immediately upon arrival at the laboratory and were either used for DNA extraction immediately or kept at −20 C and extracted for DNA within 1 week. The DNA was extracted using Qiagen QIAmp DNA Extraction Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions and tested for CMV DNA by PCR. The aqueous samples were sent for a validated, nested multiplex PCR for HSV, varicella zoster virus (VZV), CMV, and Toxoplasma gondii DNA as previously described. The lower limits of detection for HSV-1, HSV-2, CMV, VZV, and Toxoplasma gondii are 360, 120, 50, 40, and 3 copies per PCR reaction, respectively, with a sensitivity exceeding 95% for all organisms. Patients who were CMV positive were also analyzed for CMV DNA using real-time PCR with a sensitivity of 7 copies/mL with a 95% confidence interval (CI). All patients with a positive aqueous CMV DNA also underwent tests for CMV antigen (cytospin), CMV antibodies, and human immunodeficiency virus (HIV) antibodies in the serum.


All patients with endotheliitis were co-managed by the SNEC Cornea Service and the SNEC Ocular Inflammation and Immunology Service. Patients included in this study underwent either a penetrating keratoplasty (PK) (4 eyes), DSAEK (27 eyes), or Descemet membrane endothelial keratoplasty (DMEK) (1 eye) using techniques as previously described. Preoperatively, patients were treated with topical corticosteroids and topical antiglaucoma medications if intraocular pressure (IOP) was raised. Patients who were CMV positive were co-managed with the infectious disease physician at the Singapore General Hospital and treated with oral valganciclovir (900 mg) twice daily for 6 weeks followed by 450 mg twice daily for a further 6 weeks. Topical ganciclovir ophthalmic gel 0.15% (Virgan; Laboratoires Théa, Clermont-Ferrand, France) was also prescribed 5 times a day. A week before completion of therapy, a repeat aqueous analysis for CMV using real-time PCR was done. Quiescent ocular inflammation for at least 6 months and undetectable CMV DNA titer with repeat aqueous PCR analysis for CMV-positive eyes were conditions required before corneal transplantation was performed.

Our main outcome measure was a recurrence of endotheliitis, presenting clinically as localized corneal edema with AC inflammation and KPs. Recurrence was also suspected when patients presented with an unexplained sudden decrease in endothelial cell count (performed routinely at 6 months, then yearly postoperatively) in an uninflamed eye and normal graft appearance or intraocular inflammation that failed to clear or worsened with conventional high-dose corticosteroid treatment regimens.

Postoperatively, visual acuity (VA), IOP, graft clarity, and anterior segment inflammation were monitored, with a high index of suspicion for any recurrence of endotheliitis. We used similar postoperative topical corticosteroid medication regimens in all patients: a single drop of prednisolone acetate 1% every 3 hours for 1 week, 3 times a day for 6 months, twice daily for 3 months, then once a day for up to 1 year. All patients received topical levofloxacin 0.5% every 3 hours for 1 week, followed by 3 times a day for 6 months. Patients who suffered from recurrent endotheliitis postoperatively underwent a repeat aqueous tap, and the aqueous sample was reanalyzed with the tetraplex PCR as described above, with real-time PCR analysis for CMV-positive eyes. Patients who were CMV positive were restarted on topical ganciclovir ophthalmic gel 0.15% every 3 hours and were watched closely for clinical resolution of endotheliitis. Patients who were negative for all viruses tested on PCR analysis were presumed to have immunologic rejection and were treated with only topical prednisolone acetate 1% every 3 hours.

Statistical Analysis

Statistical analysis included descriptive statistics, where the mean and standard deviation (mean ± SD) were calculated for the continuous variables, while frequency distribution and percentages were used for categorical variables. Comparisons between categorical variables were conducted using χ 2 /Fisher exact tests as appropriate. The independent t test and 1-way analysis of variance test were used for continuous measures group comparisons. P value <.05 was considered statistically significant. All analyses were performed using STATA version 11 (StataCorp LP, College Station, Texas, USA).


A total of 32 eyes of 31 patients underwent corneal grafts for irreversible corneal decompensation secondary to endotheliitis. Five eyes were found to have evidence of CMV from aqueous PCR analysis preoperatively. The remaining 27 eyes did not have any evidence of CMV or any other viruses tested by the nested multiplex PCR—and all of these eyes were not found to be CMV positive during the study period. All patients were found to be immunocompetent and negative for CMV antigen and antibody tests as described above. Table 1 compares the clinical characteristics between these CMV-positive and CMV-negative eyes.

Table 1

Baseline Demographics of Subjects With Corneal Endotheliitis Comparing Cytomegalovirus-Positive and Cytomegalovirus-Negative Eyes

Cytomegalovirus Negative (n = 27) Cytomegalovirus Positive (n = 5) P Value a
Mean age (years ± SD) 65 ± 12 62 ± 5 .53
Sex (male) 15 5 .13
Mean follow-up (months ± SD) 18 ± 11 25 ± 6 .16
Chinese 18 5 .78
Malay 5 0
Indian 1 0
Other 3 0
Previous surgery
Trabeculectomy 4 2 .23
Tube 0 1 .16
Pseudophakic 15 4 .63
Type of graft
DSAEK 23 4 .86
PK 3 1
DMEK 1 0

DMEK = Descemet membrane endothelial keratoplasty; DSAEK = Descemet stripping automated endothelial keratoplasty; PK = penetrating keratoplasty.

a χ 2 /Fisher exact tests as appropriate.

Cytomegalovirus-positive Eyes

Patients who were CMV positive were all successfully treated with anti-CMV treatment as described above, until ocular inflammation was quiescent for at least 6 months with an undetectable CMV DNA titer, with repeat aqueous PCR analysis before corneal transplantation. Postoperatively, a significantly higher proportion of patients who were CMV positive had recurrence of endotheliitis after corneal transplantation within 1 year postoperatively (3/5, 60%), as compared with those who were CMV negative (2/27, 7.4%) ( P = .01). However, there was no significant difference in terms of graft failure between CMV-negative and CMV-positive eyes (3.7% vs 0%, respectively, P = .65). Visual outcomes at 1 year (best-corrected visual acuity 20/40 or better) were also comparable between CMV-positive (3/5) and CMV-negative eyes (13/27) ( P = .84).

The 3 CMV-positive eyes presenting with recurrence, at a median of 10 months (range 3-11 months) after corneal transplantation, were seen to develop low-grade AC inflammation, development of new keratic precipitates on the graft endothelium, and reduction in graft clarity. A repeat aqueous sample was obtained in these eyes, which demonstrated the presence of CMV DNA on real-time PCR analysis ( Table 2 ). Figure 1 illustrates the clinical presentation of Patient 2. Preoperatively, this patient presented with clinical signs of corneal endotheliitis, with fine to medium KPs, corneal edema, and low-grade AC inflammation, and was also noted to be CMV positive ( Figure 1 , Top left). He underwent an uncomplicated DSAEK ( Figure 1 , Top right) but developed recurrent corneal endotheliitis 3 months after the surgery, leading to graft edema, and the presence of KPs and mild AC inflammation ( Figure 1 , Bottom left). After institution of topical ganciclovir and oral valganciclovir, the endotheliitis and inflammation resolved and the graft remained clear ( Figure 1 , Bottom right). Similarly, Patient 4 ( Figure 2 ) was CMV positive preoperatively ( Figure 2 , Top left) and developed significant corneal decompensation ( Figure 2 , Top right) despite anti-CMV therapy. DSAEK was successfully performed, but the patient suffered recurrence of corneal endotheliitis and inflammation postoperatively ( Figure 2 , Bottom left) and required reinstitution of anti-CMV therapy ( Figure 2 , Bottom right). The graft remained clear with resolution of inflammation after treatment.

Table 2

Outcomes of Patients With Cytomegalovirus-Positive Corneal Endotheliitis who Underwent Corneal Transplantation

Patient Demographics (Age/Ethnicity/Sex) Preoperative BCVA Preoperative Real-Time CMV PCR Postoperative Treatment Time to Corneal Endotheliitis Recurrence (Months) Postoperative Real-Time CMV PCR (Copies) ECC (1 Year) Follow-up Duration (Months) Final Outcome/Type of Graft Final BCVA
1 58/Chinese/male 20/2000 7.7 × 10 3 Topical ganciclovir gel 0.15% NA Nondetectable 2817 21 Clear graft/PK 20/25
2 60/Chinese/male (right eye) 20/60 770 Topical ganciclovir gel 0.15% 3 1.8 × 10 6 3067 16 Clear graft/DSAEK 20/30
3 60/Chinese/male (left eye) 20/400 1.8 × 10 6 Topical ganciclovir gel 0.15% NA Nondetectable 2538 13 Clear graft/DSAEK 20/25
4 61/Chinese/male 20/400 1.1 × 10 3 Topical ganciclovir gel 0.15%
Oral valganciclovir 900 mg once daily
10 1.0 × 10 3 1318 44 Clear graft/DSAEK 20/40
5 71/Chinese/male 20/160 1.6 × 10 6 Topical ganciclovir gel 0.15% 11 6.0 × 10 6 923 32 Graft failure/DSAEK HM

Only gold members can continue reading. Log In or Register to continue

Jan 9, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on Outcomes of Corneal Transplantation for Irreversible Corneal Decompensation Secondary to Corneal Endotheliitis in Asian Eyes

Full access? Get Clinical Tree

Get Clinical Tree app for offline access