Keratitis caused by HSV, or herpes simplex keratitis (HSK), is the most common cause of corneal blindness in developed nations. It was previously thought that HSV-1 had a predilection for the trigeminal ganglion and HSV-2, for the sacral ganglion. However, an increasing number of cases of ocular herpes are caused by HSV-2, and anecdotal reports suggest that ocular HSV-2 infections may be more severe and cause more scarring.
Primary HSV infection occurs by direct contact with infected secretions. On contact, the virus enters epithelial cells and starts replicating. Within hours, it enters the sensory nerve endings and travels to the sensory ganglion, where it may remain in a dormant form called latency (Figure 2). Alternatively, it may replicate and travel back down the nerve to cause a primary infection that is clinically evident in 1% to 6% of infected patients. Once the primary infection resolves, the virus becomes latent and remains in this state until certain triggers, such as fever, sunlight exposure, stress, and menses, cause it to reactivate, multiply, and travel back down the nerve to cause recurrent infection. It is uncertain whether ocular recurrences are caused by virus that initially infected ocular tissues or by “back-door spread,” via the trigeminal ganglion, from an initial oral infection. HSV utilizes cellular enzymes for replication and the cell dies when it is released from the cell. The loss of ganglion cells with recurrent infections leads to decreased corneal sensation, one of the hallmarks of HSK.
Primary HSV ocular infection is frequently missed and rarely affects the cornea. The most common pattern of infection is blepharoconjunctivitis that heals without scarring. The associated follicular conjunctivitis is often mistaken for adenoviral conjunctivitis. However, unilateral, nonepidemic follicular conjunctivitis should always make one suspect HSV, as studies have shown at least 25% of such cases to be culture-positive for HSV.
In rare instances, especially in patients with severe eczema or other immunocompromised states, this usually innocuous infection can become life-threatening. Kaposi’s varicelliform eruption is characterized by extensive vesicular eruptions over the entire body surface and can lead to multisystem failure.
Recurrent HSV infection most frequently involves the cornea, although all other parts of the eye can be affected concurrently or independently. HSV can cause retinitis, trabeculitis, uveitis, and optic neuritis; discussion of these noncorneal infections is beyond the scope of this module.
HSK can be subdivided into epithelial and stromal keratitis, although both are often present to some degree. Understanding the anatomic basis for this classification may give the ophthalmologist a clearer picture of the pathophysiology and treatment of the disease.
These forms of keratitis are usually caused by actively replicating virus. Presentations of epithelial HSK include the following.
Dendritic ulcer is the classic herpetic corneal lesion, caused by replicating virus. The lesion is linear and dichotomously branching, with each branch terminating with a bulb (Figure 3). The borders of the lesion are slightly raised and grayish and consist of HSV-infected cells that stain with rose bengal (RB) dye. In Figure 3, these HSV-infected cells have undergone balloon degeneration. In contrast, the center of the lesion is devoid of cells and stains with fluorescein. The underlying stroma has minimal inflammation. After dendritic epithelial keratitis resolves, a dendritic scar, called a ghost dendrite, may remain in the superficial stroma.
Geographic ulcer is similar to the dendritic ulcer, also caused by replicating virus, but has a much larger epithelial defect. As it progresses, the ulcer loses its dendritic shape and takes on a form that often resembles the shape of a country—hence the termgeographic. This presentation usually occurs in persons with compromised immunity, especially patients taking topical corticosteroids. It can also occur in individuals who have untreated, long-standing originally dendritic ulcers, in which case it can be hard to distinguish from a metaherpetic ulcer. The dichotomous branching and terminal bulbs of the geographic ulcer, which are seen peripherally (Figure 4), however, often distinguish it from a metaherpetic ulcer.
In marginal keratitis, lesions are located near the limbus and can resemble staphylococcal catarrhal ulcers. They tend to have more underlying stromal inflammation and tend to be more resistant to treatment. Also, they are more likely to become a trophic ulcer.
Metaherpetic (trophic) ulcer is the only form of epithelial ulceration that does not have any live virus. The ulcer is called trophic if it arises de novo and metaherpetic if it follows a dendritic or geographic ulcer, although the terms are frequently used interchangeably. Metaherpetic ulcers result from the inability of the epithelium to heal. The causes of poor epithelial healing are multifactorial and include toxicity from antiviral medications, loss of innervation and neural-derived growth factors, poor tear surfacing, and underlying, low-grade stromal inflammation. Although metaherpetic ulcers are difficult to differentiate from geographic ulcers, they can be distinguished by their smooth, gray, elevated borders that do not stain with RB. The RB dye stains the unhealthy epithelial cells attempting to migrate across the base of the ulcer, whereas fluorescein leaks through these poorly adherent cells into the stroma and stains the periphery—so-called reverse staining (Figure 5).
These forms of keratitis are usually an immune-mediated response to nonreplicating viral particles, but more severe forms may be caused by live virus. The various forms of stromal keratitis cause a spectrum of disease, but they can be subdivided clinically based on the predominant site and type of involvement.
Immune stromal keratitis is an inflammatory response to viral antigen in the corneal stroma and can manifest as focal, multifocal, or diffuse stromal opacities. It is often accompanied by stromal edema and a mild anterior chamber reaction. Immune stromal keratitis may also occur as a partial or complete immune ring. If the inflammatory response is accompanied by corneal vascularization, it is called interstitial keratitis. These vessels can leak lipid, resulting in significant scarring. With quiescence, the vessels usually become ghost-like. One sensitive sign of recurrence is refilling of the vessels. HSV is now the most common cause of interstitial keratitis, especially unilateral, in the United States.
The significantly greater inflammation in necrotizing keratitis (Figure 6) is thought to be a reaction to live viral particles in the corneal stroma. This presentation is seen most commonly in patients with multiple recurrences of HSV infection, especially of HSV-2. Unless there is a high index of suspicion, necrotizing keratitis is difficult to distinguish from microbial forms of keratitis. It may cause corneal melting and perforation. There is frequently a significant associated uveitis, and/or trabeculitis leading to glaucoma.
Localized endothelial dysfunction from an inflammatory response to viral antigen results in a disc-shaped area of corneal edema called disciform keratitis (Figure 7). There is minimal inflammation in the stroma, although focal keratic precipitates underlying the edema are characteristic. Pseudoguttae and Descemet’s folds in the edematous area may cause disciform keratitis to be confused with Fuchs dystrophy. However, in disciform keratitis, the contralateral cornea is normal, as HSV endotheliitis is almost always unilateral; Fuchs dystrophy is always bilateral. Diffuse endotheliitis is rare and is usually accompanied by trabeculitis with elevated intraocular pressure.
In keratouveitis, uveitis predominates and is usually granulomatous, with large “mutton-fat” keratic precipitates on the endothelium (Figure 8). It is usually immune-mediated, but sectoral iritis, especially with focal endotheliitis, is thought to be a marker of live virus released into the aqueous from the sympathetic nerves. Keratouveitis can lead to significant morbidity from synechiae, cataracts, and glaucoma. Unilateral uveitis associated with high intraocular pressure should raise a high suspicion for HSV.
Diagnosis of HSV epithelial keratitis is usually made based on clinical findings, and laboratory tests are seldom needed. Laboratory tests are of no use in stromal keratitis. The following tests can be useful in challenging cases of epithelial keratitis.
Various viral transport media (eg, Richards viral transport, HH medium) can be used. The virus, once grown, can be typed to HSV-1 or HSV-2. A practical point is that RB is virucidal, and cultures done following RB corneal staining may be falsely negative for HSV.
The corneal swab can be smeared on a slide, or impression cytology using a nitrocellulose membrane can be performed. Use of fluorescein staining before FAB testing interferes with test results.
This test does not require a high number of viral particles, but it is more expensive than the other tests mentioned here.
Papanicolaou or Giemsa stains of corneal smears are examined for the presence of multinucleated giant cells and intranuclear eosinophilic inclusion bodies (Cowdry type A).
This is of limited use. The presence of anti-HSV IgM in a child may indicate active infection. However, false negatives are common and positive titers in adults merely indicate past infection, which is nearly universal.
Unfortunately, resolution of HSK does not protect the eye from future complications.
Recurrent disease is the major cause of the morbidity associated with HSK. Each recurrent episode causes increased inflammation and scarring and further decreased corneal sensation. In addition, the risk of stromal disease increases with multiple recurrences of HSV epithelial keratitis.
Even when HSK is treated promptly, scarring is common and can cause significant morbidity. A recurrent episode usually occurs adjacent to the site of a previous episode—thus the visual axis is invariably involved eventually.
Corneal hypesthesia is a sensitive sign of previous HSK. The lack of innervation causes poor tear production and decreased release of growth factors in response to injury, leading to persistent epithelial defects and neurotrophic ulcers that may melt and perforate.
Treatment of HSV epithelial keratitis differs dramatically from that of HSV stromal keratitis, reflecting the differing pathogeneses of these diseases. (See “Herpetic Eye Disease Study” for additional information.)
Epithelial keratitis often resolves spontaneously; however, the aim of treatment is to minimize scarring and stromal inflammation. The infected cells are poorly adherent, and gentle wiping debridement with a cotton-tipped applicator removes nearly all infected cells without damaging normal epithelium. Debridement results in much faster resolution and consequently less scarring. Debridement should be performed prior to initiating drug therapy.
The mainstay of treatment in the United States is topical trifluridine, which is very effective and should result in complete resolution. Topical acyclovir ointment, however, is the drug of choice in most other parts of the world because of low toxicity. It is not commercially available in the United States but can be formulated at compounding pharmacies. Topical acyclovir ointment may be used in children, in whom frequent drops would be impractical, and in patients with trifluridine allergy. Oral acyclovir may be another option for both pediatric and allergic patients. A good rule of thumb is that epithelial disease that is adequately treated but persists beyond 2 or 3 weeks is either metaherpetic (noninfectious), caused by virus that is resistant to the antiviral agent, or due to drug toxicity. If culture of persistent epithelial disease is negative, it should be treated as a metaherpetic ulcer.
The mainstay of treatment is topical corticosteroids. Although HEDS did not show any statistically significant difference in visual outcome with the use of corticosteroids, they do accelerate resolution of HSV ocular infection. Simultaneous oral antiviral prophylaxis decreases the risk of HSV reactivation at the ganglion level. Some authors advocate matching topical antivirals drop for drop with topical corticosteroids. Because doing so does not prevent reactivation at the ganglion level and may increase toxicity, this author does not advise this, except in the treatment of necrotizing keratitis. In that case, aggressive antivirals—both topical and systemic—and corticosteroid use are indicated.
The basic principle of therapy for this disease is to rapidly heal the epithelial defect. Methods to accomplish this include stopping use of toxic medications, performing punctal occlusion, instilling tear film supplements, fitting a bandage contact lens, amniotic membrane grafting, tarsorrhaphy, and, if there is significant underlying inflammation, cautiously using topical corticosteroids while watching carefully for corneal melts.
Oral antivirals are typically used as systemic prophylaxis against reactivation of HSV at the ganglion level. Topical medications are toxic with prolonged usage and are usually reserved for acute epithelial disease. Prophylaxis is useful in patients who have multiple recurrences (two or more in one year); those who have scarring close to the visual axis; those who are using topical corticosteroids for stromal disease; and those who are systemically immunocompromised. For prophylaxis of HSV infection, acyclovir is most commonly used at a dosage of 400 mg twice daily.
All the current antivirals used in the treatment of ocular HSV (see Table 1) are nucleoside analogues that inhibit viral replication by competitively inhibiting viral DNA polymerase. As they may also interfere with host DNA synthesis, they can cause significant toxicity. Acyclovir is the most specific for viral polymerase and thymidine kinase and therefore causes the least cellular toxicity of these antivirals; however, it is the most likely to induce viral resistance.
Topical antivirals are the drugs of choice for acute epithelial disease. Trifluridine is effective in the treatment of epithelial keratitis but can cause significant epithelial toxicity as mentioned above. It also contains thimerosal as a preservative and has a highly acidic pH that adds to its toxicity. Trifluridine has a short half-life and must be used every 2 hours. Systemic antivirals are used primarily for prophylaxis of recurrent disease or as an antiviral cover during corticosteroid therapy for stromal keratitis. The three main systemic antivirals for the treatment of HSK are acyclovir, valacyclovir, and famciclovir. They have a high therapeutic index but resistance is an issue, especially in immunocompromised patients. Unfortunately there is cross-resistance among these three agents.
Topical corticosteroids used for the treatment of HSV stromal keratitis and uveitis are always given under antiviral cover. Typically, 1% prednisolone acetate or 0.1% dexamethasone is used. The dosing frequency should be based on the severity of the inflammation. On resolution of the inflammation, the steroids need to be tapered gradually to prevent rebound inflammation.
The HEDS was a prospective, randomized, double-masked, placebo-controlled multicenter study to assess the effect of adding topical corticosteroids and oral acyclovir to conventional antiviral therapy (topical trifluridine, Viroptic) for HSV keratitis and iridocyclitis. There were six parts to the study: three therapeutic, two preventive, and one cohort. The main findings are summarized here.
Herpes Stromal Keratitis, Not on Steroid Trial (HEDS-SKN). Compared with the placebo group, the patients who had received prednisolone phosphate drops had faster resolution of the stromal keratitis and fewer treatment failures. However, delaying the initiation of corticosteroid treatment did not affect the eventual outcome of the disease, in that visual acuity was similar in the two groups at 26 weeks.
Herpes Stromal Keratitis, on Steroid Treatment (HEDS-SKS). There was no apparent benefit to adding oral acyclovir to the treatment regimen of a topical corticosteroid and trifluridine.
Herpes Simplex Virus Iridocyclitis, Receiving Topical Steroids (HEDS-IRT). The trial was stopped because of slow recruitment, but treatment failures occurred at a higher rate in the placebo group than in the acyclovir group, indicating a potential benefit to adding oral acyclovir to the regimen of a topical steroid and an antiviral.
Herpes Simplex Virus Epithelial Keratitis Trial (HEDS-EKT). This trial assessed the effect of adding oral acyclovir to topical antiviral therapy for acute HSV epithelial keratitis. The addition of oral acyclovir provided no benefit in preventing the development of stromal keratitis or iritis. Incidentally, the study found that in the control group (treatment with trifluridine alone), the risk of stromal keratitis or iridocyclitis was quite low in the year after an episode of epithelial keratitis; this risk was lower than that previously suggested in the literature.
Acyclovir Prevention Trial (HEDS-APT). In patients who had had an ocular HSV infection in the previous year, use of oral acyclovir reduced by 41% the probability that any form of HSK would return. The incidence of epithelial keratitis was reduced from 11% to 9% and of stromal keratitis from 13% to 8%. Importantly, the rate of recurrent HSV stromal keratitis decreased 50%. Unfortunately, the effect did not persist once the acyclovir was discontinued. More than one recurrence was reported in 4% of patients in the acyclovir group and in 9% of those in the placebo group.
Ocular HSV Recurrence Factor Study (HEDS-RFS). No association was found between psychological or other forms of stress and ocular HSV recurrences. Also, previous episodes of HSV epithelial keratitis were not a predictor of future occurrences, whereas previous, especially multiple, episodes of HSV stromal keratitis markedly increased the probability of subsequent stromal keratitis.
Limitations. Although HEDS was excellent in its conception and plan, it had some limitations. First, many of the trials had inadequate recruitment or a very high dropout rate. Second, in the prevention trials, oral acyclovir was used for only 3 weeks. Third, the corticosteroid regimen was standardized and not tailored to an individual patient’s inflammation. Finally, trifluridine was used in both the study group and the placebo group in the trials evaluating acyclovir for the treatment of stromal keratitis and iridocyclitis.
Reference: American Academy of Ophthalmology
