Necrotizing Herpetic Retinitis

Necrotizing herpetic retinitis (NHR) is an uncommon, but serious infection of the retina caused by one of the human herpesviruses (HHV).1,2 In immunocompetent patients, most cases of NHR result from infection by either varicella zoster virus (VZV) or herpes simplex virus type-1 or type-2 (HSV-1 or HSV-2). Whereas VZV retinitis often occurs in older patients as acquired cell-mediated immunity directed against VZV wanes, HSV retinitis can occur at any age. Moreover, HSV retinitis is more likely to be associated with prior or concurrent encephalitis or meningitis, which can be life-threatening. Cytomegalovirus (CMV) retinitis, in contrast, typically occurs in immunocompromised patients.3–5 While infection by the human immunodeficiency virus (HIV) is by far the most common cause of immune compromise in patients with CMV retinitis, other factors that can contribute to relative weakening of the immune system include advanced age, diabetes mellitus, cancer, or use of corticosteroids and/or non-corticosteroid immunosuppressive medications to either treat autoimmune disease or malignancy, or to prevent solid organ or bone marrow rejection. CMV retinitis can also occur in the setting of inherited or acquired immune deficit disorders, such as Good syndrome, but such cases are rare. Other HHVs, most notably HHV-4, or Epstein Barr virus (EBV),6,7 and HHV-6,8 have been identified in eyes with NHR, but determining causality is often complicated by the fact that infiltrating leukocytes are known to harbor these particular HHVs.9 The degree of vitreous inflammation in eyes with NHR depends in part upon the underlying level of immune suppression, with trace to mild inflammation typically occurring in severely immunocompromised patients with acquired immune deficit syndrome (AIDS) and moderate to severe inflammation occurring more often in patients with no or limited immune deficits. Such moderate to severe inflammation is a defining feature of acute retinal necrosis (ARN) syndrome.1,10 Other clinical characteristics of ARN include a predilection for the peripheral retina and rapid progression, which often produces an occlusive vasculitis. Non-CMV NHR that occurs in patients with AIDS can also progress rapidly, but unlike ARN tends to have little or no vitreous inflammation. This syndrome has been termed progressive outer retinal necrosis (PORN) and is caused most often by VZV.11,12 Two research articles13,14 and two letters15,16 in this issue of Ocular Immunology & Inflammation present important findings relevant to the clinical presentation, progression, and management of NHR.

Roy and associates described their 10 year experience with ARN managed at a tertiary referral center in Chennai, South India.13 The cohort included 62 eyes of 53 patients, 38 (71.7%) of whom were men and 15 (28.3%) of whom were women. Age ranged from 6 to 70 years, with a mean of 36.0 years. Two patients were HIV positive. Varicella zoster virus was identified in 28 eyes (45.2%) and HSV in 19 eyes (30.6%). No cases of CMV retinitis were diagnosed. Follow-up ranged from six months to 10 years, with a mean of just over two years. Nine patients (17.0%) developed bilateral retinitis, with an average time to fellow eye involvement of seven months. Mean vision at presentation was approximately 20/2000 (logMAR 2.02), which on average was about four weeks from onset of symptoms. All patients were initially treated with a week or more of intravenous acyclovir. An oral anti-viral agent, either acyclovir or valacyclovir, was then continued for an additional six to eight weeks. All patients also received high-dose systemic corticosteroids within two to three days after institution of anti-viral therapy. The corticosteroids were then tapered over four to six weeks. Two-thirds of eyes had 12 or more clock hours of involvement, and in three eyes the entire retina was involved. Rhegmatogenous retinal detachment (RRD) occurred in approximately two-thirds of eyes, with slightly less than half having RRD at presentation. The mean time from onset of symptoms to development of RRD was 10 weeks. Nineteen eyes were treated with prophylactic laser photocoagulation and 12 of these (63.2%) progressed to RRD despite laser treatment. Among the 43 eyes that did not receive prophylactic laser, 29 (67.4%) developed RRD. Among all eyes that developed RRD, just over one-third progressed to proliferative vitreoretinopathy (PVR). The mean vision at final visit was about 20/1250 (logMAR 1.78), with 28 eyes (45.1%) achieving 3/60 or better vision. The most common contributors to poor visual outcome were RRD (47.1%) and optic atrophy (29.4%). Logistic regression analysis identified both a longer time from onset of symptoms to the initiation of anti-viral therapy and the extent of retinal involvement to be predictors of poor outcome. A higher proportion of eyes without RRD achieved 3/60 or better vision (57.1% vs 39.0%), but this difference failed to achieve statistical significance. Meghpara and colleagues17 reported a similar correlation between the extent of retinal involvement and final vision and noted specifically that eyes with less than four clock hours of involvement tended to have the best prognosis. Prophylactic laser treatment appeared to offer no therapeutic benefit in the cohort from South India, a finding reported previously by others.18,19

Takakura and associates described five patients who developed unilateral NHR following intraocular or periocular corticosteroid administration.14 The authors added to this series a comprehensive review of the literature, which identified 25 similar patients. Unlike ARN in general, the most common cause of NHR following periocular or intraocular corticosteroid injection was CMV, identified in 76.7% of eyes. Intraocular as opposed to periocular corticosteroid use was associated with the development of retinitis in most patients (25 of 30 eyes; 83.3%), with an average time to diagnosis of between three and four months. Diabetes mellitus was a frequent co-morbidity, observed in 40.0% of cases. Nine patients (30.0%) had active uveitis and so were using other immunosuppressive agents. Two patients were HIV positive with inactive CMV retinitis prior to injection. In both instances, local corticosteroid injections were used to treat complications of immune recovery uveitis. Local and/or systemic anti-viral agents were effective at controlling the retinitis in all patients. Five of the nine patients with active uveitis also underwent variable modifications in their systemic immunosuppression. The reported rate of RRD was relatively low at 13.3%.

Papageorgiou and colleagues described a 22-year-old man recently diagnosed with EBV infectious mononucleosis who then developed encephalitis, bilateral ARN, a right afferent pupillary defect, a left fourth nerve palsy, and a right incongruent, incomplete homonymous hemianopia.15 A contrast-enhanced MRI of the brain and spinal cord revealed numerous areas of enhancement, including the prechiasmal portion of the right optic nerve and the intracranial portion of the right third nerve. HHV-6 DNA was identified in the cerebrospinal fluid using a real-time, quantitative polymerase chain reaction (PCR)-based test, leading the authors to conclude that HHV-6 caused many, if not all, of the complications and that the recent EBV infection may have facilitated reactivation of latent HHV-6. The patient was treated with intravenous acyclovir for 10 days followed by oral valacyclovir and systemic corticosteroids for one month and improved. Vision at last visit was 20/20 in each eye, with resolution of visual field defects and improvement in diplopia. While the scenario as described is possible, as both HHV-6 associated retinitis8 and fourth nerve palsy20 have been described previously, it should be noted that ocular complications of HHV-6 infection are quite rare and that interpretation of such PCR results is complicated by both the high rate of EBV and HHV-6 infection in the adult population at large and the fact that these particular HHVs are known to remain latent within leukocytes following infection.9 Histopathological confirmation, using either immunohistochemistry or in situ hybridization directed against the HHV of interest, offers the most conclusive evidence in such cases, but is not always feasible or indicated. Neither EBV nor HHV-6 would be expected to show a therapeutic response to currently available anti-viral therapies.

Last, Guney and associates described a 45-year-old man with HSV-2 ARN who was treated with systemic corticosteroids and intravenous acyclovir, which caused nephrotoxicity, followed by oral valacyclovir, which produced a toxic hepatitis.16 Both drugs were discontinued and the patient was given two intravitreal ganciclovir injections over a one-week period. He then developed RRD, which was treated with pars plana vitrectomy followed by silicone oil placement. Retinal reattachment surgery was successful, the retinitis resolved, and vision stabilized at 20/50.

Together, these four studies highlight several important points related to NHR. First, while clinical context contributes to the likelihood of identifying a given HHV in an eye with retinitis, with VZV and HSV more likely in immunocompetent patients and CMV more likely in patients who are either immunocompromised or have received either periocular or intraocular corticosteroids, the correlation is imperfect and as such PCR-based testing remains an important tool for distinguishing VZV, HSV, and CMV retinitis. It should be remembered, however, that the sensitivity of PCR-based testing for these particular HHVs drops dramatically if done following initiation of anti-viral therapy. Hence, negative PCR results should be interpreted with caution if performed after initiating anti-viral therapy. Second, when considering EBV and HHV-6 as a possible rare cause of retinitis or uveitis, positive PCR testing of ocular fluids needs to be weighed with the knowledge that infiltrating leukocytes may also carry these viruses. Hence, when feasible and indicated, PCR testing for EBV and HHV-6 in particular is best used in conjunction with more specific in situ analyses directed against the HHV of interest, such as immunohistochemistry or in situ hybridization. And third, while use of local and systemic anti-viral agents is generally well tolerated and controls VZV, HSV, and CMV retinitis in the vast majority of cases, drug-induced adverse event can occur. It is important, therefore, to monitor blood counts, as well as liver and renal function, in all patients receiving systemic treatment.

Acknowledgements

Supported in part by The Pacific Vision Foundation (ETC) and The San Francisco Retina Foundation (ETC). We thank Dr. Jyotirmay Biswas and Dr. Eleni Papageorgiou for thoughtfully commenting on an early version of this editorial.

Financial Conflicts: The authors have no financial conflicts.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.