Graphic structures of voriconazoleand fluconazole.
Voriconazole levels obtained inplasma and vitreous (R = 0.64).
Voriconazole levels obtained inplasma and aqueous (R = 0.61).
Hariprasad SM, Mieler WF, Holz ER, Gao H, Kim JE, Chi J, Prince RA. Determination of Vitreous, Aqueous, and Plasma Concentration of OrallyAdministered Voriconazole in Humans. Arch Ophthalmol. 2004;122(1):42-47. doi:10.1001/archopht.122.1.42
To investigate the penetration of voriconazole, a new-generation triazoleantifungal agent, into the vitreous and aqueous humor after oral administration.
A prospective, nonrandomized clinical study included 14 patients scheduledfor elective pars plana vitrectomy surgery between December 1, 2002, and February28, 2003, at the Cullen Eye Institute, Houston, Tex. Aqueous, vitreous, andplasma samples were obtained and analyzed from 14 patients after oral administrationof two 400-mg doses of voriconazole taken 12 hours apart before surgery. Assayswere performed by means of high-performance liquid chromatography.
Mean ± SD voriconazole concentrations in plasma (n = 14), vitreous(n = 14), and aqueous (n = 11) were 2.13 ± 0.93 µg/mL, 0.81 ±0.31 µg/mL, and 1.13 ± 0.57 µg/mL, respectively. Mean ±SD sampling times after oral administration of the second voriconazole dosefor plasma, vitreous, and aqueous were 2.4 ± 0.6 hours, 3.0 ±0.5 hours, and 2.9 ± 0.5 hours, respectively. The percentages of plasmavoriconazole concentration achieved in the vitreous and aqueous were 38.1%and 53.0%, respectively. Mean vitreous and aqueous minimum inhibitory concentrationsfor 90% of isolates (MIC90) were achieved against a wide spectrumof yeasts and molds, including Aspergillus speciesand Candida species, along with many other organisms.
Orally administered voriconazole achieves therapeutic aqueous and vitreouslevels in the noninflamed human eye, and the activity spectrum appears toappropriately encompass the most frequently encountered mycotic species involvedin the various causes of fungal endophthalmitis. Because of its broad spectrumof coverage, low MIC90 levels for the organisms of concern, goodtolerability, and excellent bioavailability with oral administration, it mayrepresent a major advance in the prophylaxis or management of exogenous orendogenous fungal endophthalmitis.
Fungal endophthalmitis can be either exogenous or endogenous in origin.Exogenous fungal endophthalmitis occurs in healthy individuals by 3 principalcauses: surgery, contiguous spread from an external ocular infection, andtrauma. In all cases, fungus gains entry into the anterior chamber, the vitreous,or both. Progression of the infection is dependent on the size of the inoculumat the time of injury, the growth rate of the fungus, and the status of thehost's immunologic system.1,2
Endogenous fungal endophthalmitis occurs primarily in patients who arein a compromised state of health. Risk factors include the use of long-termand broad-spectrum antibiotics, corticosteroids and cytotoxic agents, indwellingintravenous catheters and hyperalimentation, and intravenous narcotic drugs.The increased survival of patients with debilitating diseases and AIDS hasalso contributed to the increasing incidence of endogenous fungal endophthalmitis.Spread of fungus to the eye characteristically occurs after fungemia, usuallycaused by a yeast. The most commonly encountered yeast is Candida species (primarily Candida albicans).The major filamentous fungus seen in endogenous endophthalmitis is Aspergillus species.1 The incidenceof endogenous fungal endophthalmitis has increased substantially during thepast few decades and has been reported to vary between 2% and 45% among patientswith systemic fungal infection.3- 5
Although fungal endophthalmitis is rare in the grand scheme of intraocularinfection, it remains an important clinical problem in ophthalmology becauseof the potentially devastating consequences resulting from these infections.In addition, ocular fungal infections have traditionally been very difficultto treat because of limited therapeutic options both systemically and intravitreally.While postoperative fungal endophthalmitis may be rare in the Western world,it is more common in developing countries.6 Forexample, in recent studies from India, fungi accounted for up to 21.8% ofall culture-positive postoperative endophthalmitis cases.7 Theliterature is replete with case series and reports of exogenous postoperativefungal endophthalmitis.8- 13 Intraocularspread from fungal keratitis and inoculation by penetrating ocular traumaare both very important causes of exogenous fungal endophthalmitis.8,14 In fact, one of the earliest descriptionsof fungal endophthalmitis from an exogenous source was described by Römer15 in 1902, who cultured Aspergillusfumigatus from the vitreous of an eye enucleated for severe inflammationthat developed 9 days after penetrating trauma. There have been suggestionsin the literature that the recent widespread use of topical corticosteroidsand broad-spectrum antibiotics in eye treatment have contributed to the increasingnumber of fungal infections.16
The recognition that fungal infections were on the rise prompted severalinvestigations to assess the intraocular penetration of systemically administeredantifungal agents. The intravitreal penetration of amphotericin B was firstinvestigated by Green et al17 in 1965, whodiscovered that intravitreal penetration was limited after intravenous administration.This was confirmed in 1985 by O'Day and colleagues,18 whofound that the intravitreal concentration of amphotericin B achieved afterintravenous administration barely reached the minimum inhibitory concentration(MIC) against Candida parapsilosis Since amphotericinB is associated with numerous adverse effects related to both the drug (nephrotoxicity)and its administration (fever, rigors, and hypotension), it was not believedthat the marginal intravitreal levels achieved after intravenous administrationwarranted its routine use for treating fungal endophthalmitis. Although flucytosineachieved seemingly high intravitreal levels after systemic administration,the spectrum of activity was so narrow that it did not achieve the MIC formany commonly encountered organisms in fungal endophthalmitis.18,19 In1987, Malecaze and colleagues20 found thatvitreous levels of ketoconazole were undetectable in noninflamed rabbit eyesafter oral administration. The first promising data regarding systemic antifungalagents for use in fungal endophthalmitis came in 1990, when O'Day and colleagues21 discovered that the triazole agent fluconazole wasable to achieve significant levels in the vitreous of rabbit eyes. The improvedocular penetration of fluconazole compared with older-generation antifungalagents was attributed to its lower protein binding and improved water solubilitycharacteristics. Although these data were very promising, fluconazole lackeda broad spectrum of coverage against many of the most commonly encounteredorganisms found in fungal endophthalmitis.
In the past few years, there have been major strides in the developmentof antifungal agents, and their potential use in the treatment of fungal endophthalmitisneeds to be explored. The new-generation triazoles, such as voriconazole,posaconazole, and ravuconazole, represent advances in the evolution of thetriazole antifungal class and have been developed to address the increasingincidence of fungal infections and the limitations of the currently availableagents.
Voriconazole is a triazole antifungal agent and is a second-generationsynthetic derivative of fluconazole. It was developed by Pfizer PharmaceuticalsGroup (Pfizer Inc, New York, NY) as part of a program designed to enhancethe potency and spectrum of activity of fluconazole. Voriconazole differsfrom fluconazole by the addition of a methyl group to the propyl backboneand by the substitution of a triazole moiety with a fluoropyrimidine group,resulting in a marked change in activity (Figure 1).22 Voriconazole has 96%oral bioavailability and reaches peak plasma concentrations 2 to 3 hours afteroral dosing. Protein binding is moderate at 58%, with wide distribution ofthe agent throughout the body into many tissues and fluids. Previous in vitrostudies have shown voriconazole to have a broad spectrum of action against Aspergillus species, Blastomyces dermatitidis, Candida species, Paecilomyceslilacinus, Coccidioides immitis, Cryptococcus neoformans, Histoplasma capsulatum, Penicillium species, Scedosporium species, Curvularia species, and others.
We chose to study the intraocular penetration of orally administeredvoriconazole in humans for 3 reasons. First, older-generation triazoles suchas fluconazole have been shown to achieve significant levels in the vitreousafter oral administration in the noninflamed eye.21 Second,whole-body autoradiography studies in rats showed that voriconazole is veryhighly concentrated in the retina, second only to levels observed in the liver.23 Third, the MICs for 90% of isolates (MIC90)of voriconazole against the pathogens most commonly responsible for exogenousand endogenous fungal endophthalmitis were generally lower than those forthe other antimycotic agents we surveyed (Table 1).19,23,24
This study was carried out with the approval of the Baylor College ofMedicine Institutional Review Board. Fourteen adult patients, aged 30 to 87years (mean, 58 years), undergoing elective pars plana vitrectomy surgerybetween December 1, 2002, and February 28, 2003, at the Cullen Eye Institute,Houston, Tex, were included in the study. Exclusion criteria included thefollowing: age younger than 18 years, known sensitivity to triazole agents,use of any other antibiotic(s) or antifungal(s) in the preceding 3 weeks,pregnancy or current breastfeeding, fresh vitreous hemorrhage as an indicationfor vitrectomy (less than 1 month old), and active endophthalmitis. In addition,patients with significant drug-drug interactions with voriconazole were excludedfrom the study (Table 2).23
After informed consent was obtained, patients were asked to take a totalof four 200-mg voriconazole tablets orally (2 tablets one-half day beforesurgery and 2 tablets just before surgery). Prospectively completed data formswere designed to include medical history, collection times of various samples,and concentrations of voriconazole in plasma, aqueous, and vitreous. Patientswere asked to record the exact time of oral administration, and this was reportedto the surgeon on the day of the operation. Aqueous, vitreous, and blood sampleswere obtained before infusion of any intravenous or intraocular irrigatingsolution to obtain pure samples. Approximately 8 to 10 mL of venous bloodwas collected less than 1 hour before surgery in the preoperative holdingarea. In the operative suite, approximately 0.1 mL of aqueous fluid was aspiratedwith a 30-gauge needle attached to a syringe through a paracentesis site inpatients in whom it was believed safe to do so (ie, patients with pseudophakiceyes or patients with phakic eyes with deep anterior chambers). Within 10minutes, 0.2 to 0.3 mL of vitreous fluid was obtained by using a vitreouscutting device attached to a syringe via a short length of tubing. Aqueousand vitreous samples were immediately frozen at −20°C. The bloodsample was centrifuged and the plasma collected from this was frozen as well.Voriconazole concentrations were determined in each of the samples by meansof a previously described high-performance liquid chromatography technique.25,26 Vitreous, aqueous, and plasma voriconazoleconcentrations were compared with already established in vitro MIC90 data (Table 1).19,23,24 A paired t test was performed to determine whether any significant differencesexisted between various subsets of patients, including diabetic vs nondiabeticpatients, and those whose eyes have phakic status. Mean values are mean ±SD.
Indications for operation in the 14 patients were as follows (Table 3): epiretinal membrane (9 patients),macular hole (2), traction retinal detachment (2), and dense white cataract(1).
Mean voriconazole concentrations in plasma (n = 14), vitreous (n = 14),and aqueous (n = 11) were 2.13 ± 0.93 µg/mL, 0.81 ± 0.31µg/mL, and 1.13 ± 0.57 µg/mL, respectively. Mean samplingtimes after oral administration of the second voriconazole tablets for plasma,vitreous, and aqueous were 2.4 ± 0.6 hours, 3.0 ± 0.5 hours,and 2.9 ± 0.5 hours, respectively (Table 3). The percentages of plasma voriconazole concentration achievedin the vitreous and aqueous were 38.1% and 53.0%, respectively. Positive correlationswere observed between plasma and vitreous concentrations of voriconazole (Figure 2; R = 0.64).A similar correlation was observed between plasma and aqueous concentrationsof voriconazole as well (Figure 3; R = 0.61).
Two of the 14 patients were diabetic. The mean voricionazole concentrationsin the plasma and vitreous for these 2 patients were 2.80 ± 0.59 µg/mLand 0.77 ± 0.24 µg/mL, respectively. These levels were not significantlydifferent from those of the 12 nondiabetic patients, whose plasma and vitreousconcentrations were 2.01 ± 0.94 µg/mL and 0.82 ± 0.33µg/mL, respectively (P = .28 and P = .86, respectively).
Five of the 14 patients had pseudophakic eyes. The mean voriconazoleconcentrations in the plasma, vitreous, and aqueous for these 5 patients were2.13 ± 0.68 µg/mL, 0.90 ± 0.38 µg/mL, and 1.25 ±0.71 µg/mL, respectively. These levels were not significantly differentfrom those of the 9 patients whose eyes had phakic status, whose plasma, vitreous,and aqueous concentrations were 2.12 ± 1.08 µg/mL, 0.76 ±0.28 µg/mL, and 1.03 ± 0.48 µg/mL, respectively (P = .99, P = .44, and P = .56, respectively).
No serious adverse reactions were attributed to voriconazole. Two patientscomplained of transient visual blurring occurring approximately 30 minutesafter taking the first voriconazole dose (patients 6 and 8; Table 3).
Therapy for fungal infections can be difficult and prolonged. The difficultyin treatment is due to a combination of the growth characteristics of fungi,the limited availability of effective antifungal agents, and the poor tissuepenetration of previously investigated agents. The most important therapeuticprinciple in endophthalmitis is early diagnosis and correct identificationof the fungus, as early treatment is more likely to yield a better visualoutcome.27
One of the most common current treatment regimens for fungal endophthalmitisinvolves the use of intravenous amphotericin B. While this antimycotic agentis effective in treating disseminated fungal infection, it has very limitedintraocular penetration.17,18 Therefore,vitrectomy with intravitreal amphotericin B has been the most current treatmentfor fungal endophthalmitis.1 The value of intravitrealamphotericin B has not been proved, and toxicity questions do remain.28 In addition, if mistakes are made in preparing dilutionsor if the preparation is injected into an air-filled eye, there is the potentialfor serious toxic adverse effects to the retina. For these reasons, Christmasand Smiddy29 investigated alternate managementtechniques for fungal endophthalmitis to avoid the risk of toxic adverse effectsto the retina. They found that pars plana vitrectomy with systemic fluconazolesuccessfully treated Candida endophthalmitis in severalpatients. Since their report in 1996, there have been major advances in thedevelopment of antifungal agents. If the use of systemic antifungal agentsis to be considered for the prophylaxis, or as an adjunct in the management,of fungal endophthalmitis, a systemic agent must be found with the highestpossible intraocular penetration, as well as the lowest MIC90 forthe organisms of concern. Given our findings, we believe that voriconazolemay represent a major advance in this regard (ie, in vitro potency of voriconazoleagainst yeasts was 60-fold higher than that of fluconazole24).
We are aware of 4 published reports of successful treatment of fungalendophthalmitis with the new-generation triazoles. The first report involvesa patient with exogenous endophthalmitis from Fusarium solani infection secondary to a severe keratitis. Because of a lack of responseto conventional therapy, voriconazole treatment was started systemically,topically, and by intracameral injection. The patient's clinical situationsteadily improved, and voriconazole treatment was stopped after 8 weeks, withcomplete resolution of infection.30 More recently,Garbino and colleagues12 described a patientwith exogenous P lilacinus endophthalmitis secondaryto complicated cataract surgery. The patient's clinical situation had deteriorateddespite treatment with several agents in addition to oral itraconazole andintracameral fluconazole. The patient's treatment regimen was switched tovoriconazole, 400 mg every 12 hours, when cultures showed that the organismwas highly susceptible to this new triazole. During the course of 3 months,the patient's condition steadily improved, with complete resolution of infection.Sponsel and colleagues31 described an immunocompetentwoman who developed F solani keratitis secondaryto contact lens wear. She was treated with topical antibiotics for concurrentbacterial keratitis, as well as topical and intravenous amphotericin B, natamycin,and ketoconazole. Despite this intensive regimen, the infection spread intothe anterior chamber. Therapy was switched to the new triazole posaconazole(oral and topical) after cultures demonstrated amphotericin B–resistant Fusarium species. Within 1 week of this new therapy therewas significant clearing of the cornea, and 3 months later there was no evidenceof infection. Finally, Kim and colleagues32 recentlyreported their experiences treating a 65-year-old woman with refractory A fumigatus scleritis and a nodular epibulbar abscess causedby scleral buckle infection. Oral and topical itraconazole, ketoconazole,and amphotericin B treatment only resulted in progression of infection duringthe course of 4 months. All therapy was discontinued and treatment with oralvoriconazole, 200 mg twice a day, was begun. After 1 week of treatment, oculartenderness disappeared and the infection resolved steadily thereafter.
Because of the study design, it is not known whether sampling occurredduring voriconazole peak or trough intraocular levels. This investigation,however, provides proof of principle that therapeutic intraocular levels oforally administered voriconazole can be achieved in the noninflamed humaneye. Table 1 compares the vitreousand aqueous concentrations of voriconazole obtained in this study with previouslyestablished in vitro MIC90 data.19,23,24 Includedin this table are organisms that have been reported in the ophthalmic literatureto cause fungal infections and for which in vitro MIC90 data wereavailable. Candida tropicalis is generally susceptibleto voriconazole, with the exception of a single isolate that demonstratedan MIC of more than 16.0 µg/mL.22 Inaddition, voriconazole was unable to achieve intraocular levels effectiveagainst Fusarium species. This is not to suggestthat voriconazole may not be of therapeutic value for fungal endophthalmitisbecause of these organisms; a previous study suggests that intraocular penetrationof systemic agents may be higher in an eye that has sustained trauma, is infected,or is inflamed.33 This may be owing to disruptionof the blood-ocular barrier and could help explain the beneficial effectsof voriconazole and posaconazole against Fusarium speciesin the previously mentioned case reports.
Voriconazole is very well tolerated, with most adverse reactions describedas mild. The primary adverse effects observed include transient visual disturbance,hepatotoxicity, and skin reactions. Visual disturbance is the most frequentdrug reaction, observed in 23% to 35% of individuals, and has been describedas enhanced light perception, color vision change, and blurring. These symptomstypically occur 30 minutes after administration and during the first weekof therapy. Resolution of these symptoms occurs within 30 minutes after onset,and symptoms generally resolve even with continued therapy.23 Twopatients in our study reported these symptoms after taking their first voriconazoledose (patients 6 and 8; Table 3).Decreased-amplitude waveforms on electroretinograms in humans and dogs suggestthat the retina is the tissue causing these effects.23 Inaddition, studies in dogs have shown that no structural alterations in theretina or visual pathways occur as a result of voriconazole administration.Fortunately, there are no known long-term ocular side effects of voriconazoleuse.23 Voriconazole is metabolized primarilyby the hepatic cytochrome P450 isozymes CYP2C19, CYP2C9, and CYP3A4.23 Therefore, there are several drug-drug interactions,as described in Table 2. Hepatotoxicityhas been known to occur with voriconazole use23;therefore, monitoring of hepatic functions is recommended. Rash occurs witha frequency of about 6% to 25%.23 In addition,photosensitivity has been noted during voriconazole treatment, and protectivemeasures are recommended (avoiding strong sunlight and covering exposed skinareas).23 Finally, triazoles in general havebeen shown to be teratogenic in animal studies.23 Therefore,until data on voriconazole from humans are available, it is recommended thatwomen of childbearing potential use contraception. Voriconazole use in pregnantwomen should be considered only if the benefits outweigh the risks.23 There are certain dosing modifications required inspecial clinical circumstances (ie, renal impairment), and we advise reviewingthe voriconazole package insert before initiating systemic voriconazole therapy.
The recommended oral dosage of voriconazole is one 200-mg tablet every12 hours. In certain clinical situations, a loading dose of 400 mg every 12hours for 1 day may be considered. In our study design, we chose to use theloading dose regimen before sample collection to achieve peak plasma concentrationsmore rapidly (when oral loading dose regimens are administered to healthysubjects, peak plasma concentrations close to steady state are achieved withinthe first 24 hours of dosing). In severe ocular mycotic infection, use ofthis loading dose regimen may be considered to help achieve peak plasma concentrationsof voriconazole more rapidly.
The realization that voriconazole may have important ocular clinicalapplication is not novel. In a recent study by Zhou and colleagues,34 albino rabbits were treated with topical eyedropsof voriconazole at a "low dose" (50 µg/mL) and a "high dose" (100 µg/mL)twice a day for 11 days. Aqueous humor samples were then obtained though aparacentesis site with a 30-gauge needle attached to a syringe and later analyzedby means of liquid chromatography–mass spectrometry with electrosprayionization. Impressive voriconazole concentrations were found in the aqueoushumor: 7.29 ± 5.84 µg/mL in the low-dose group (n = 4) and 14.56± 12.90 µg/mL (n = 7) in the high-dose group. In addition, Zhouet al discovered that voriconazole penetrates the normal eye without metabolicmodification.34 That study demonstrated aqueousconcentration of voriconazole several-fold higher than the levels observedin our study. There is an explanation for this large discrepancy; Zhou andassociates' investigation was an animal study and the route and duration oftherapy were very different from those in our study. In addition, penetrationof topical voriconazole into the vitreous was not investigated; therefore,no conclusions can be drawn regarding its use in open-globe trauma or fungalendophthalmitis involving the posterior segment. Given Zhou and coworkers'findings, however, further investigation of topical voriconazole is warrantedto determine whether the human eye can tolerate such a formulation.
In summary, orally administered voriconazole achieves therapeutic aqueousand vitreous levels in the noninflamed human eye, and the activity spectrumappears to appropriately encompass the most frequently encountered fungalspecies involved in the various causes of exogenous and endogenous fungalendophthalmitis. In addition, oral voriconazole may present an alternate managementtechnique for fungal endophthalmitis by which the risk of toxic adverse effectsto the retina associated with intravitreal amphotericin B injection can beavoided. Because of its broad spectrum of coverage, low MIC90 levelsfor the organisms of concern, good tolerability, and excellent bioavailabilitywith oral administration, it may represent a major advance in the prophylaxisor management of fungal endophthalmitis.
Corresponding author: William F. Mieler, MD, Cullen Eye Institute,6565 Fannin St, NC-205, Houston, TX 77030 (e-mail: firstname.lastname@example.org).
Submitted for publication March 5, 2003; final revision received July25, 2003; accepted August 21, 2003.
This study was supported in part by an unrestricted grant from Researchto Prevent Blindness Inc, New York, NY.
We gratefully acknowledge the following individuals for their contributionsto this study: Stephen C. Pflugfelder, MD, Douglas D. Koch, MD, and ShannaNoel. The authors also would like to thank Dan J. Sheehan, PhD, and PfizerPharmaceuticals Group for their generous supply of voriconazole tablets usedin this study.