Acute interface fluid syndrome after laser in situ keratomileusis in a case of cytomegalovirus (CMV) endotheliitis and secondary glaucoma

Abstract

Cytomegalovirus (CMV) can cause recalcitrant recurrent keratouveitis and secondary glaucoma. We report a case of chronic recurrent anterior uveitis with secondary glaucoma presenting with acute visual loss and interface fluid 9 years after laser in situ keratomileusis. Based on clinical presentation, a viral aetiology was suspected. Aqueous tap was positive for CMV-DNA by real-time quantitative PCR of the aqueous humour. The patient was treated with systemic antivirals, topical corticosteroids and antiglaucoma medications. The interface fluid resorbed rapidly. The intraocular pressure (IOP) was controlled by trabeculectomy. There was no further corneal deterioration at 7-month follow-up and the IOP had also stabilised. We believe this is only the third reported case of CMV-related interface fluid syndrome. This case highlights the role of quantitative PCR analysis for establishing viral aetiology in recurrent unilateral hypertensive anterior uveitis and reports the unusual finding of interface fluid which resolved after starting systemic antiviral therapy.

Background

Cytomegalovirus (CMV)-related anterior segment inflammation can present as unilateral endotheliitis with diffusely distributed keratic precipitates, elevated intraocular pressure (IOP) and coexisting anterior uveitis in immunocompetent individuals.¹ Clinically, it may initially manifest as recurrent episodic iritis with raised IOP, with low-grade anterior uveitis.² If a patient has undergone laser in situ keratomileusis (LASIK) (as in the present case), IOP elevation and compromised endothelial function can lead to interface fluid syndrome (IFS), which is characterised by accumulation of fluid in the stroma.³ CMV infection can be diagnosed by testing the aqueous sample for viruses using real-time PCR test. Oral antiviral treatment is essential for management. IFS usually resolves once the IOP is under control provided the endothelial function is still intact. In the case of refractory glaucoma (as in this case), surgical intervention is indicated.

Case presentation

A 35-year-old immunocompetent woman presented with complaints of acute-onset blurred vision in her left eye (LE) for 2 months. She was diagnosed elsewhere with Posner-Schlossman syndrome, with onset 16 years ago and with recurrent episodes of at least three a year. She had undergone LASIK 9 years ago for high myopia (for refractive error of −9 dioptre and −11 dioptre in her right and left eye, respectively) and cataract surgery a year ago in LE following which she had satisfactory vision.

Uncorrected visual acuities were logarithm of the minimum angle of resolution (logMAR) 0.5 and logMAR 1.7 in the right and left eye, respectively. The right eye (RE) was unremarkable except for the faint scarring of the LASIK flap. LE showed ciliary congestion and the cornea showed large, granulomatous, diffusely distributed keratic precipitates, an oedematous LASIK flap, and fluid gap at the stromal interface (figure 1A, B). The pupil was mid-dilated and sluggish, and a posterior chamber intraocular lens was noted in the capsular bag. Fundus examination revealed advanced glaucomatous cupping with bipolar notch.

Figure 1.

(A) Slit lamp photograph at presentation shows oedematous cornea with large, diffusely distributed, granulomatous keratic precipitates. (B) Optical section with fluid at the interface. (C) Anterior segment optical coherence tomography at presentation shows hyporeflective area corresponding to the area of fluid accumulation at the interface, increased stromal hyper-reflectively and hyper-reflective dots in the endothelium. The flap, fluid space and residual stromal bed showed thickness of 121 micrometers, 117 μm and 386 μm, respectively.

The IOP was 16 mm Hg and 14 mm Hg in the right and left eye, respectively, over the central cornea and 42 mm Hg in the peripheral cornea superiorly, out of the flap region, on Goldmann applanation tonometry.

Investigation

Anterior segment optical coherence tomography (AS-OCT) of the LE at presentation showed cleft-like hyporeflective area within the anterior stroma (corresponding to interface fluid accumulation), high reflectivity of the entire stroma (corresponding to oedema), and discrete, small, rounded hyper-reflective projections on the Descemet’s membrane corresponding to keratic precipitates (figure 1C). The flap, fluid space and residual stromal bed showed thickness of 121 μm, 117 μm and 386 μm, respectively, at presentation. On follow-up 1 week later, AS-OCT showed no interface fluid, reduction in corneal thickness and decreased posterior corneal hyper-reflectivity correlating with decreased cornea oedema noted clinically (figure 2A–D).

Figure 2.

(A) At 1 week, follow-up slit lamp image shows reduction in oedema and keratic precipitates. (B) AS-OCT at 1 week shows resolution of the interface fluid and decreased posterior corneal hyper-reflectively. (C) Slit lamp examination at 7-month follow-up shows a clear cornea. (D) AS-OCT at 7-month follow-up shows a normal LASIK flap with normal interface. AS-OCT, anterior segment optical coherence tomography; LASIK, laser in situ keratomileusis.

Specular microscopy was performed at 1-week, 4-month and 7-month follow-up (figure 3A–C). Initially the central corneal endothelial cell density (ECD) could not be measured in LE (RE count was 2611/mm²). Abnormal endothelial cell morphology and reduced counts in LE (1048/mm²) were noted at 4 months, with no further worsening at 7 months (ECD 1369/mm²).

Figure 3.

(A) Specular image at 1 week: endothelial mosaic cannot be made out and count cannot be estimated due to corneal oedema. (B) Specular image at 4 months: endothelial counts are reduced compared with the fellow eye (endothelial cell density: 1048/mm²), with abnormal morphology of cells. (C) Specular image at 7 months: counts have stabilised (endothelial cell density: 1369/mm²). CCT, Central Corneal Thickness.

Differential diagnosis

Fuchs uveitis syndrome was ruled out in view of recurrent episodes of redness and absence of characteristic stellate keratic precipitates or iris changes or vitritis. The differential diagnosis also included other viral aetiologies of corneal endotheliitis, such as herpes simplex virus (HSV) and varicella zoster virus (VZV). Other causes of recurrent anterior uveitis were not applicable to this case in view of the pattern of recurrent hypertensive uveitis.

Treatment

Diagnostic aqueous tap was done and real-time PCR testing using the CMV R-GENE kit (bioMérieux) revealed CMV-DNA (26×10⁹copies/mL). The fluid was negative for VZV and HSV DNA. The patient was started on oral valganciclovir 900 mg two times per day for 6 weeks, followed by 900 mg once daily for 6 weeks, prednisolone acetate 1% eye-drops in tapering doses, and antiglaucoma agents (acetazolamide 250 mg tablet two times per day and brinzolamide 1% eye-drops thrice a day). At 1-month follow-up, corneal oedema had reduced, the interface fluid had cleared and there was a marked decrease in anterior segment inflammation (figure 2A, B), with improvement in visual acuity to logMAR 0.8. The IOP was still high at 32 mm Hg despite maximum medical treatment (travoprost and timolol maleate had been added to the therapy). In view of advanced disc damage and uncontrolled IOP, the patient underwent trabeculectomy with mitomycin C 6 weeks after presentation. She was continued on oral valganciclovir and tapering dose of topical steroids.

Outcome and follow-up

Follow-up at 4 months revealed complete resolution of corneal oedema and anterior segment inflammation, with the IOP under control. Subsequently she was maintained on low-dose topical steroids and the oral antivirals were stopped. At 7 months, her visual acuity was maintained at logMAR 0.8 and and IOP at 15 mm Hg (figure 2C, D). She was off all topical and oral medications.

Discussion

CMV infection has been well recognised as an important cause of endotheliitis, uveitis and secondary glaucoma. The presence of CMV-DNA in the aqueous tap is a significant risk factor for raised IOP.⁴ To date only two cases of IFS after LASIK with CMV uveitis have been published.^{5 6} The putative mechanisms of IFS include corneal endothelial dysfunction and high IOP, which lead to abnormal stromal pump and accumulation of fluid at the interface, as was noted in our case. The resolution of IFS within a week of starting antivirals, even without control of IOP, suggests improvement in endotheliitis due to control of the infection.

The importance of noting IFS in a case of pre-existing glaucoma cannot be overemphasised. It has been widely reported that IOP measurement with Goldmann applanation tonometer (GAT) post-LASIK can give fallacious readings when tested over the central cornea due to the interface fluid.⁷ In such situations, IOP measurement using a device such as Tonopen is highly recommended. In our case, GAT underestimated the central IOP, but readings in the peripheral cornea were higher. Therefore, if GAT is used, the IOP should be measured external to the flap area. Once the fluid resolves, the IOP readings from the central cornea can then be considered accurate (after accounting for the thinner corneas post-LASIK).

In our case, AS-OCT was extremely useful in visualising the location and height of the fluid pocket at the initial visit and also to note its disappearance on follow-up. In addition, OCT also delineated the large keratic precipitates which resolved with treatment. Therefore, in cases of IFS, this modality has been found to be extremely useful in gaining additional information about the aetiology.⁸ Although there have been several large series of CMV infection in recent times,⁹ the diagnosis still remains elusive in several cases. In our case too, the patient had the disease undiagnosed for over 15 years, probably due to lack of widespread awareness of this entity. The role of PCR, especially quantitative and real-time PCR, is invaluable in the detection and quantification of the infection. Repeating the aqueous tap for repeat PCR after completion of treatment has also been recommended, although we did not plan the same in our case. Optimising treatment in CMV endotheliitis has been challenging, as there seems to be no clear consensus. However, most reports have shown beneficial results with the use of oral valganciclovir therapy. Our case showed immediate response to oral valganciclovir in terms of resolution of interface fluid, probably related to improvement in endothelial function and decrease in anterior segment inflammation. In our case, although the patient had significant decrease in the ECD as compared with the normal eye, once we could initiate therapy there was no further decrease (figure 3A–C), suggesting stabilisation of the condition.

However more rigorous IOP control was warranted in the setting of advanced glaucomatous disc damage and known risk of poor visual prognosis if left untreated in such cases.⁹ Trabeculectomy with Mitomicin-C (MMC) was the preferred surgical modality in this case rather than a drainage shunt device. In comparison with standard trabeculectomy (or trabeculectomy with mitomycin C), a glaucoma drainage device has an added risk of corneal decompensation, and as the patient already had endothelial cell loss it was decided against the same. In addition, erosion or extrusion of a glaucoma drainage device is a complication that can happen over a period of time.¹⁰ At our institute, the primary procedure for most patients with healthy conjunctiva is trabeculectomy with or without MMC over primary drainage shunt device, hence the selection of this procedure.

This report highlights CMV as an important cause of recurrent hypertensive uveitis. Chronic disease leads to endothelial dysfunction and secondary glaucoma, which in this case presented as IFS. Quantitative real-time PCR of the aqueous tap helped clinch the diagnosis. Initiating appropriate therapy helped in preserving residual vision and stabilising the corneal condition.

Learning points.

• Unilateral, chronic recurrent hypertensive uveitis with granulomatous diffuse or discrete keratic precipitates is likely to be due to viral aetiology and merits aqueous tap with quantitative real-time PCR analysis.

• Post-laser in situ keratomileusis interface fluid syndrome (IFS), when it occurs in cytomegalovirus endotheliitis, can be reversed by antiviral therapy and not just intraocular pressure (IOP) control alone.

• IOP measurement in the central cornea with applanation tonometry is fallacious in IFS and readings should be taken from the peripheral cornea.

• Repeated inflammation may lead to permanent damage to the trabecular meshwork and surgical management of glaucoma should be considered under antiviral cover.