Objective: To report 4 cases of undiagnosed cytomegalovirus (CMV) endotheliitis in patients who underwent Descemet's stripping automated endothelial keratoplasty (DSAEK).
Design: Retrospective interventional case series.
Participants: Four eyes of 4 patients diagnosed with active CMV endotheliitis after DSAEK.
Methods: Retrospective review of the medical records of 4 patients with DSAEK who had an aqueous tap that was positive for CMV DNA but negative for herpes simplex virus (HSV) and varicella zoster virus.
Main outcome measure: Clinical features and management.
Results: Four immunocompetent Chinese male patients with a mean age of 67 years underwent DSAEK for posterior polymorphous dystrophy (1), Fuchs' heterochromic cyclitis (1), pseudophakic bullous keratopathy (1), and herpetic keratouveitis (1). Clinical findings seen in all patients were localized corneal edema, increased intraocular pressure, pigmented keratic precipitates (KPs), and no/minimal anterior chamber (AC) activity. An unexplained sudden decrease in endothelial cell count (ECC) in the absence of rejection or significant inflammation was seen in 3 patients, whereas 1 patient also developed concomitant retinitis. CMV DNA was positive in all aqueous specimens and from the vitreous of the patient with retinitis. All patients were treated with oral valganciclovir with resolution of inflammation; 2 patients had recurrences; 1 patient developed recurrent retinitis; and 1 patient developed recurrent CMV endotheliitis and is currently receiving maintenance therapy with oral valganciclovir.
Conclusions: CMV endotheliitis with corneal edema masqueraded as a variety of other endothelial conditions, which resulted in DSAEK surgery being performed in these patients who may have responded to antiviral treatment without the need for endothelial transplantation. A heightened awareness is required to exclude CMV endotheliitis as the cause for endothelial decompensation or unexplained, sudden reduction in ECCs post-DSAEK in the absence of other complications, and it should be differentiated from allograft rejection in view of the critical difference in treatment.