Sir,

We report a case of cytomegalovirus (CMV) corneal endotheliitis following fluocinolone implant for uveitis.

Case report

A 21-year-old HIV-negative Chinese woman presented to Singapore National Eye Centre with panuveitis and Snellen acuity 6/45 bilaterally. Behcet's disease was diagnosed and systemic immunosuppression commenced. Vision improved to 6/7.5 bilaterally, with relapsing inflammation. Her left eye became quiescent after insertion of fluocinolone acetonide 0.59 mg implant. The non-implanted eye required continued systemic therapy for control. Left phacoemulsification was required 4 months postimplant.

After 14 months, pigmented keratic precipitates and corneal oedema developed inferiorly. Trabeculectomy was performed for elevated intraocular pressure. Corneal oedema increased over the next 9 months (Figure 1). There was no evidence of retinitis (although fundoscopy was impossible once diffused corneal swelling occurred). At 2-year postimplantation, aqueous humour was positive for CMV DNA (3.6 × 104 copies viral DNA per ml) and mRNA by polymerase chain reaction. CMV endotheliitis was diagnosed. Serology for CMV IgG was positive, but negative for IgM and CMV antigen.

Figure 1
figure 1

Inferior corneal oedema and keratic precipitates consistent with endotheliitis.

Full size image

Fluocinolone was explanted to reverse local immunosuppression, and systemic valganciclovir commenced. Vitreous was positive for CMV DNA and mRNA. Eight months later, she remained CMV mRNA positive. Corneal oedema persisted despite subsequent insertion of a ganciclovir implant. This eye has only light perception vision.

Comment

Cytomegalovirus has been identified as a cause for anterior uveitis and endotheliitis in immunocompetent patients.1, 2 CMV retinitis can complicate intravitreal triamcinolone administration, and one case of post fluocinolone implant has been reported.3

Our patient had CMV endotheliitis, and this is the first report of this complication following fluocinolone implantation. We cannot explain why the cornea was the predominant site of infection in our patient, considering vitreous was positive for CMV. However, histopathological studies of CMV-infected eyes in immunodeficiency showed the presence of virus in iris, ciliary body, and cornea.4 We suspect that CMV remains latent in the anterior segment in some eyes and reactivates if local immunity is altered. This is supported by the patient's serological findings.

Anterior segment CMV infection can develop in immunocompromised eyes, and clinicians should be aware of their presentation. Aqueous sampling for CMV should be considered when introducing intravitreal depots of steroid.