Aqueous and Vitreous Concentrations Following Topical Administration of 1% Voriconazole in Humans

Objective  To determine the penetration of 1% voriconazole solution into the aqueous and vitreous following topical administration.

Methods  A prospective nonrandomized study of 13 patients scheduled for pars plana vitrectomy surgery. Aqueous and vitreous samples were obtained and analyzed after topical administration of 1% voriconazole every 2 hours for 24 hours before surgery. Drug concentration quantitation was performed using high-performance liquid chromatography.

Results  The mean (SD) sampling time after topical administration of the last voriconazole dose was 24 (14) minutes. The mean (SD) voriconazole concentrations in the aqueous and vitreous were 6.49 (3.04) μg/mL and 0.16 (0.08) μg/mL, respectively. Aqueous concentrations exceeded the minimum inhibitory concentration at which 90% of isolates are inhibited (MIC₉₀) for a wide spectrum of fungi and mold, including Aspergillus, Fusarium, and Candida species. Vitreous concentrations of voriconazole exceeded the MIC₉₀ for Candida albicans.

Conclusion  Topically administered voriconazole achieves therapeutic concentrations in the aqueous of the noninflamed human eye for many fungi and molds and achieves therapeutic levels in the vitreous for Candida species. Topical voriconazole may be a useful agent for the management of fungal keratitis and for prophylaxis against the development of fungal endophthalmitis.

Fungal endophthalmitis can be exogenous or endogenous in origin. Exogenous fungal endophthalmitis in healthy individuals has 3 principal causes: trauma, surgery, and contiguous spread from an external ocular infection (particularly fungal keratitis).¹ Progression of the infection is dependent on the size of the inoculum at the time of injury, the growth rate of the fungus, and the status of the host's immune system.^2,3 Exogenous fungal endophthalmitis can lead to notable vision loss; in a series of 19 cases from 1969 to 1986, only 8 patients recovered 20/400 or better visual acuity.¹

Fungal keratitis is a major cause of blindness in Asia, with filamentous fungi being the most frequently reported pathogen.^4-6 In South India, up to 44% of all central corneal ulcers are caused by fungi.⁷ Fungal keratitis is prevalent in other regions as well, including Asia (17% in Nepal and 36% in Bangladesh), Africa (37.6% in Ghana), and North America (35% in South Florida).^4-6,8,9 Recently, a minor epidemic of fungal keratitis was linked to specific contact lens disinfecting solutions.¹⁰

Treatment options for fungal keratitis are limited. Natamycin, a polyene compound, is the only commercially available antifungal agent approved for topical ocular use in the United States. It has limited efficacy because of poor corneal penetration and is typically used in nonsevere superficial keratitis.^11,12 Although amphotericin B has reasonable activity against Candida and Aspergillus species, it is not effective against Fusarium species,¹³ a common cause of posttraumatic fungal eye infection.^9,13-16

Voriconazole is a triazole antifungal agent. It is a second-generation synthetic derivative of fluconazole, with enhanced potency and spectrum of activity.¹⁷ Voriconazole has 96% oral bioavailability and reaches peak plasma concentrations 2 to 3 hours after oral dosing. In vitro studies have shown voriconazole to have a broad spectrum of action against Aspergillus species, Blastomyces dermatitidis, Candida species, Coccidioides immitis, Cryptococcus neoformans, Curvularia species, Fusarium species, Histoplasma capsulatum, Paecilomyces lilacinus, Penicillium species, Scedosporium species, and others.¹⁸ Voriconazole has excellent intraocular penetration in the noninflamed eye after oral administration and has been shown to be effective in treating patients with fungal endophthalmitis.¹⁹ Voriconazole has also been used to treat fungal endophthalmitis through intravitreal injection.²⁰

Investigations in animals have demonstrated that topically administered voriconazole results in high drug concentrations in the aqueous humor.²¹ We sought to determine the aqueous and vitreous concentrations of voriconazole following topical administration in noninflamed human eyes.

This study was performed with the approval of the Washington University School of Medicine Institutional Review Board, St Louis, Missouri. Thirteen adult patients (age range, 27-71 years; mean [SD] age, 55.6 [11.6] years) undergoing elective pars plana vitrectomy surgery between November 24, 2004, and March 1, 2005, at the Barnes Retina Institute, St Louis, were included in the study. Exclusion criteria were age younger than 18 years, active endophthalmitis, pregnancy or current breastfeeding, known sensitivity to triazole agents, the use of any other antibiotic or antifungal agents in the preceding 3 weeks, and fresh (<1-month-old) vitreous hemorrhage as an indication for vitrectomy.

All drug dilutions were performed by hospital pharmacists using a laminar flow hood and a clean room meeting United States Pharmacopeia chapter 797 sterile preparation risk level 2 standards. A 200-mg lyophilized powder vial of voriconazole (Vfend; Pfizer Inc, New York, New York) was diluted according to the package insert, substituting isotonic sodium chloride solution without preservatives instead of sterile water as the diluent to maintain appropriate osmolarity for corneal use. The effect of this substitution on ocular penetration was not assayed. The final concentration was 1% (10 mg/mL). The product was labeled with a 24-hour dating and refrigerated. Using a sterile 10-mL syringe, 5 mL was transferred into a sterile eyedropper.

After informed consent was obtained, patients were asked to topically self-administer 1% voriconazole ophthalmic solution during the 24 hours before surgery (1 drop every 2 hours throughout the night) in the eye scheduled for surgery. A final drop of 1% voriconazole was topically administered to all eyes 15 to 45 minutes before surgery.

In the operative suite, approximately 0.1 mL of aqueous fluid was aspirated using a 30-gauge needle attached to a syringe through a paracentesis site. Within 10 minutes, 0.2 to 0.3 mL of vitreous fluid was obtained using a vitreous cutting device attached to a syringe via a short length of tubing. Aqueous and vitreous samples were immediately frozen at −20°C. Voriconazole concentrations were determined in each of the samples using high-performance liquid chromatography.^22,23 Vitreous and aqueous voriconazole concentrations were compared with established data regarding the in vitro minimum inhibitory concentration at which 90% of isolates are inhibited (MIC₉₀).^24-26 The Mann-Whitney test was performed to determine if any statistically significant differences existed between various subsets of patients, including patients who had diabetes mellitus (DM) vs those who did not have DM and patients who had phakia vs those who had pseudophakia.

Indications for operation in the 13 patients were as follows (Table 1): epiretinal membrane (5 patients), macular hole (2 patients), tractional retinal detachment (2 patients), retinal detachment (1 patient), vitreous hemorrhage longer than 1 month (1 patient), macular edema associated with branch retinal vein occlusion (1 patient), and recalcitrant clinically significant diabetic macular edema (1 patient).

The mean (SD) voriconazole concentrations in the aqueous (n = 13) and vitreous (n = 13) were 6.49 (3.04) μg/mL and 0.16 (0.08) μg/mL, respectively. The mean (SD) sampling time after topical administration of the last voriconazole dose for the aqueous and vitreous was 24 (14) minutes (Table 1).

Eleven of 13 patients had DM. The mean (SD) voriconazole concentrations in the aqueous and vitreous for these patients were 6.14 (3.20) μg/mL and 0.14 (0.08) μg/mL, respectively. These levels were not statistically significantly different from those of the 2 patients with DM, whose mean (SD) aqueous and vitreous concentrations were 8.44 (0.13) μg/mL and 0.24 (0.02) μg/mL, respectively (P =.81 and P =.57, respectively).

Three of 13 patients had pseudophakia. The mean (SD) voriconazole concentrations in the aqueous and vitreous for these 3 patients were 6.66 (5.44) μg/mL and 0.20 (0.12) μg/mL, respectively (although 1 patient with pseudophakia had low levels [<1 μg/mL] in both aqueous and vitreous [Table 1]). The levels in the patients with pseudophakia were not statistically significantly different from those of the 10 patients with phakia, whose mean (SD) aqueous and vitreous concentrations were 6.44 (2.40) μg/mL and 0.14 (0.07) μg/mL, respectively (P =.41 and P =.23, respectively).

No adverse reactions were attributed to the use of voriconazole. Two patients reported mild transient stinging and burning sensation on topical instillation of 1% voriconazole. The use of topical voriconazole did not result in visible keratopathy. Visualization during surgery was excellent in all patients.

Fungal keratitis remains a potentially blinding condition. In parts of the world, it accounts for approximately half of all microbial keratitis cases. Treatment using traditional medications (eg, topical natamycin and systemic ketoconazole) is suboptimal. In a recent study²⁷ of 115 cases of fungal keratitis treated using topical natamycin (with or without systemic ketoconazole), primary treatment failure occurred in approximately 31% of cases, with more than half of these treatment failure cases ultimately perforating.

Topically administered voriconazole is a potentially useful antifungal agent for the treatment of fungal keratitis and perhaps endophthalmitis. Its clinical spectrum of action is favorable for the treatment of many causes of fungal eye disease. Marangon et al¹⁸ established the in vitro susceptibility profiles of 541 fungal keratitis and endophthalmitis isolates (collected between 1980 and 2002) to several antifungal agents. They found that all 541 organisms were sensitive to voriconazole. By comparison, sensitivities to amphotericin B were 50% for Aspergillus species and 45% for Fusarium species. Both agents showed activity against 100% of Candida isolates.

Several case reports describe patients initially treated using oral and topical antifungal agents without success and subsequently successfully treated using topical and oral voriconazole; these cases have included Alternaria keratitis, Scedosporium apiospermum keratitis, and a corneal abscess caused by Fusarium.^28-30 Successful treatment of intraocular fungal infection using systemic voriconazole has also been reported. One report involves a patient with exogenous endophthalmitis from a Fusarium solani infection secondary to severe keratitis.³¹ After a lack of response to conventional therapy, voriconazole was started systemically, topically, and intracamerally by injection. The patient's clinical status steadily improved, and voriconazole treatment was discontinued after 8 weeks, with complete resolution of the infection. Findings from these studies suggest that voriconazole may be a useful agent for the treatment of fungal eye disease.

The potential role of topical voriconazole as adjunctive treatment for intraocular fungal infection is unclear. Findings in animals suggest that the compound penetrates the cornea and results in therapeutic intraocular concentrations. Zhou et al²¹ studied albino rabbits treated using topical eyedrops of voriconazole at a low dose (50 μg/mL) and at a high dose (100 μg/mL) twice daily for 11 days. The following potentially therapeutic mean (SD) concentrations were found in the aqueous humor: 7.29 (5.84) μg/mL in the low-dose group and 14.56 (12.90) μg/mL in the high-dose group. In addition, the investigators determined that voriconazole penetrates the rabbit eye without metabolic modification.

The present study provides proof of principle that potentially therapeutic concentrations of topically administered voriconazole can be achieved in the noninflamed human eye. Table 2 compares the vitreous and aqueous concentrations of voriconazole obtained in this study with previously established in vitro MIC₉₀ data.^24-26 Included in the table are organisms that have been reported in the ophthalmic literature to cause fungal infections and for which in vitro MIC₉₀ data are available. Candida tropicalis is generally susceptible to voriconazole except for a single isolate that demonstrated an MIC₉₀ of greater than 16.0 μg/mL.¹⁷ Topically administered voriconazole achieved MIC₉₀ levels in the aqueous for almost all pathogenic fungi and achieved therapeutic levels in the vitreous for Candida species. The MIC₉₀ for Fusarium has been reported to range from 2.0 to 8.0 μg/mL; all aqueous isolates exceeded the lower value, but only 5 of 13 exceeded the upper value. Because the present study included only a single sampling following dosing every 2 hours for 24 hours, we do not know how rapidly topically administered voriconazole is cleared from the eye and how long the levels remain in the therapeutic range following dosing. Similarly, we do not know the extent to which dosing of the medication topically every 2 hours may have led to systemic absorption.

Systemically administered voriconazole has been shown to achieve therapeutic concentrations in the aqueous and vitreous.³² However, it may cause notable adverse effects. Voriconazole is metabolized primarily by the hepatic cytochrome P450 isoenzymes CYP2C19, CYP2C9, and CYP3A4. Hepatotoxicity has been reported with the use of systemic voriconazole. With systemic use, rash occurs at frequencies of 6% to 25%.²⁶ In addition, photosensitivity has been noted among patients taking systemic voriconazole, and the use of protective measures (ie, avoiding strong sunlight and covering exposed skin areas) is recommended. Triazole medications have been shown to be teratogenic in animal studies³³; they are considered class D drugs for use in pregnancy. Topical adverse effects in our small series seem limited. The primary adverse effect associated with topical treatment was mild transient stinging or burning sensation on instillation, which resolved within minutes. We do not know the extent to which topical administration of voriconazole affects plasma concentrations. However, the entire administered dose per drop is on the order of 500 μg (compared with a standard systemic dose of 200 mg twice daily for oral medication or 6-12 mg/kg/d for intravenous administration), so it is unlikely that systemic adverse effects would result from topical administration. To date, ocular adverse effects of voriconazole use have not been reported.²⁶ Toxic reaction studies in rats have shown no occurrence of electroretinographic or histological alterations in the retina following intravitreal voriconazole administration of up to 25 μg/mL.²⁰

Topically administered voriconazole achieves therapeutic aqueous concentrations in the noninflamed human eye, and the activity spectrum seems to appropriately include the most frequently encountered mycotic species causing keratitis and endophthalmitis. With its good tolerability, broad spectrum of coverage, bioavailability via topical administration, and low MIC₉₀ levels for the organisms of concern, 1% voriconazole solution may have notable usefulness in the treatment of fungal keratitis, particularly when associated with intraocular infection. The excellent aqueous penetration may make it a useful adjunct in the treatment of fungal endophthalmitis, and the potentially therapeutic vitreous concentrations against Candida species may render this agent useful for the treatment of Candida endophthalmitis (the most common form of fungal endophthalmitis), particularly in patients for whom drug interactions may preclude or disfavor the use of systemic voriconazole.

Correspondence: Russell N. Van Gelder, MD, PhD, Department of Ophthalmology and Visual Sciences, Washington University School of Medicine, Campus Box 8096, 660 S Euclid Ave, St Louis, MO 63110 (vangelder@vision.wustl.edu).

Submitted for Publication: February 5, 2007; final revision received April 8, 2007; accepted April 17, 2007.

Financial Disclosure: Dr Hariprasad participates on the speaker's bureau for Alcon and Genentech. Dr Mieler is on the advisory boards of Alcon, Genentech, and Pfizer. Dr Prince receives research support from Astellas, AstraZeneca, Cubist, Enzon, and Merck. Dr Van Gelder is a paid consultant for Alcon.

Funding/Support: This study was supported by National Institutes of Health core grant P30 EY02687 to the Department of Ophthalmology and Visual Sciences, Washington University School of Medicine. Dr Van Gelder is a recipient of the Culpeper Medical Scholar Award of the Rockefeller Brothers Fund.

Additional Contributions: Stephanie Porto, RPh, Holly Pryzgoda, PharmD, and the Barnes Hospital Research Pharmacy assisted with this study.