We report these cases to provide evidence for the safe and effective use of foscarnet eyedrops for herpetic keratitis refractory to conventional antiviral therapy. No randomized control trial has evaluated the use of foscarnet eyedrops, and we do not know the necessary duration of treatment or potential toxicity.

Systemic antiviral therapy is frequently utilized as a first-line treatment for herpetic keratitis. This is typically achieved with acyclovir, a guanosine analog, or its prodrug valacyclovir, followed by long-term lower-dose prophylactic treatment with the same antiviral [1, 2]. All 6 of our patients had persistent herpetic keratitis despite a therapeutic dose of oral guanosine analog therapy. Notably, none of our patients had a known immunosuppressive condition. Once thought to be a complication only for the immunocompromised, resistance to antiviral drugs is increasingly recognized in immunocompetent individuals with herpetic keratitis [2].

Multiple mechanisms play a role in antiviral drug resistance in herpetic keratitis. One mechanism is due to the cornea’s immune privilege status. In non-ocular herpetic infections, such as herpes labialis and genital herpes, viral replication is largely inhibited by local immune responses, and antivirals play more of a supportive role. In the cornea—an immunoprivileged site with reduced local immune response—viral control is much more dependent on antiviral sensitivity. However, prolonged use, such as long-term prophylactic use, can lead to resistant viral strains [3, 4].

At the enzymatic level, herpetic resistance to antiviral therapy is most commonly caused by reduced thymidine kinase (TK) activity [5]. Acyclovir is a guanosine analogue that requires phosphorylation by the viral thymidine kinase (TK) prior to conversion to its active form. The same is required of acyclovir’s prodrugs including valacylovir and famciclovir. Resistance to these medications can occur with viral mutations in TK. Foscarnet is a direct viral DNA polymerase inhibitor and does not require activation by TK. Thus, it may be an alternate antiviral therapy when resistance is due to a TK mutation [5, 6].

Foscarnet has low oral bioavailability and is most commonly administered intravenously. It has been reported to be a successful treatment for acyclovir-resistant herpetic keratitis [6]. However, systemic foscarnet is associated with significant systemic adverse effects including renal damage, nausea, and electrolyte derangements [7]. Additionally, it is administered intravenously three times daily, a significant burden to patients, who would require either admission to a hospital or a peripherally inserted central catheter (PICC) line. Five of our six patients tolerated topical foscarnet well without any reported adverse effect. One of our patients had resolution of presumed VZV pseudodendrites after 2.5 weeks of foscarnet eyedrops; however, the lesions recurred after discontinuation of the drops. Upon restarting the drops, the patient reported new irritation, and the drops were discontinued.

The safe and effective use of topical foscarnet for HSV-related corneal disease has been reported in the literature in a small group of case series and case reports. In 1992, Behrens-Bauman reported that topical foscarnet 3% was as effective as trifluridine 1% in the treatment of HSV dendritic keratitis, without any reported side effects [8]. In 1998, Fabricius studied varying concentrations of topical foscarnet in 10 patients with HSV keratitis and found that corneal toxicity developed in 3 of 6 patients using foscarnet 1.9% after 26–40 days of treatment, but the medication was well tolerated by all 7 patients using foscarnet 1.2% and 1.4% [9]. In 2001, Cao et al. found topical foscarnet 3% to be both safe and more effective than topical acyclovir 0.1% for HSV keratitis (65% cure rate versus 41.27% cure rate, respectively) [10]. In 2012, Yu et al. reported that while both topical foscarnet and topical ganciclovir led to significantly decreased corneal signs and symptoms of HSV epithelial keratitis, treatment with topical foscarnet led to significantly decreased recurrence rate [11]. Most recently, in 2022, De Clerck et al. published a case of a patient with genotypically proven acyclovir-resistant HSV keratitis successfully treated with topical foscarnet 1.2% five times daily [12].

Given the small number of patients in our group, the true incidence of toxicity to foscarnet 2.4% eyedrops remains unknown. While one of six patients (20%) is a relatively high percentage of adverse events, we are reassured that the one patient with toxicity had mild symptoms that resolved upon drug cessation. She complained of ocular irritation but did not present with any exam findings of allergic keratoconjunctivitis or corneal toxicity. We believe that the tolerance exhibited by our patients and those in prior reports is reassuring to continue investigating this treatment option for patients suffering from disease that is refractory to established treatments. Further studies investigating response at different dosages, such as the one conducted by Fabricius, should be conducted. As with any medication, all patients starting foscarnet eyedrops must be counseled on the risks and benefits of treatment.

The role of guanosine analogues and their derivatives are unclear when used in combination with foscarnet. All 6 of our patients are maintained on treatment regimens that include valacyclovir or famciclovir, despite the initial lack of response to these medications at therapeutic doses. A case from the dermatology literature in 1998 reported a patient with culture-proven acyclovir-resistant HSV-2 genital ulceration who was successfully treated by stopping oral acyclovir and starting topical foscarnet 2.4% cream. One month after treatment, sensitivity cultures revealed that the virus recovered sensitivity to acyclovir [13]. It is possible that in some of our patients, treatment of the resistant virus with the use of topical foscarnet allowed a guanosine analogue-sensitive strain to appear.

Another possible mechanism by which foscarnet eyedrops treat refractory herpetic disease is through synergy with other antivirals. All six of our patients were maintained on oral antiviral therapy with a guanosine analogue throughout their treatment course. In vitro and in vivo studies have shown guanosine analogues to be synergistic with other drugs that inhibit HSV and VZV replication through alternative mechanisms of action, including cidofovir, amenamevir, vidarabine, and docosanol [14,15,16]. Such studies have not been conducted on foscarnet, and understanding the interactions between foscarnet and the guanosine analogues would require extensive experimental investigations.

Patient 3 and 6 had HSV endotheliitis without epithelial disease, both of whom had resolution of their corneal haze after the addition of foscarnet eyedrops. Similarly, Fabricius treated 3 patients with disciform endotheliitis without epithelial disease with foscarnet 1.9% eyedrops, all of whom had resolution of their keratitis. Given that endotheliitis is due to active HSV infection in the anterior chamber [1], this suggest that foscarnet eyedrops may be able to penetrate the cornea and achieve adequate aqueous humor concentration. Topical antivirals vary in their ability to penetrate the cornea—while topical trifluridine can only achieve therapeutic concentrations in the anterior chamber when the cornea has been debrided or damaged, topical acyclovir has been shown to achieve therapeutic levels with an intact epithelium [17]. Ganciclovir 0.15% levels in aqueous humor have not been studied. Further studies are required to investigate the pharmacokinetics of foscarnet and its levels in aqueous humor following topical application.

A limitation of this case series is that no patient had pheno- or genotypically proven guanosine analogue resistance to HSV or VZV, and we assumed some level of resistance given the lack of response to therapeutic doses of oral antiviral medication. Of note, patient 3 and 6, both with persistent endotheliitis, were on prednisolone 1% drops in addition to oral antivirals at the time of initiation of foscarnet treatment. It is possible that in the absence of adequate antiviral penetration of the oral medications, the virus was able to continue proliferation in the anterior chamber in part due to this steroid use.

Another limitation is that prior to the initiation of foscarnet eyedrops, only 1 of the 3 patients with presumed VZV pseudodendrites had been trialed on topical ganciclovir, which has been reported to be an effective treatment for VZV pseudodendrites that are resistant to other oral or topical antivirals [17]. As a commercially available product in the United States, ganciclovir 0.15% ophthalmic gel may be a more accessible treatment option for patients with refractory herpetic keratitis.

Another limitation is that our 3 patients with VZV pseudodendrites did not have positive VZV ocular cultures, and thus the diagnosis of VZV was presumed. However, patient 2 had a remote case of VZV affecting V2 on the ipsilateral side, and patient 4 and 5—who did not have any known VZV history—both had classic VZV pseudodendritic features and responded to antiviral therapy. These cases suggest that a classic pseudodendritic lesion should raise suspicion for occult VZV, even in the absence of a clinical history of shingles.

In conclusion, supplemental treatment with foscarnet eyedrops led to the near or complete resolution of active keratitis in all 6 of our patients who had previously been resistant to other antiviral medications. This is an off-label use of foscarnet. We hope that the favorable outcomes of our patients lead to increased awareness of an additional treatment option for ophthalmologists when faced with managing these difficult cases.