Materials and Methods

Animals—Six healthy adult horses were used in the study. The horses (4 mares and 2 geldings) included 2 Arabians, 1 Thoroughbred, 1 Quarter Horse, 1 Standardbred, and 1 Dutch warmblood. Body weight of the horses ranged from 430 to 631 kg (mean, 526 kg). All horses were considered to be healthy on the basis of results of a CBC and physical examination conducted prior to inclusion in the study.

Horses were housed separately in box stalls beginning 18 hours before drug administration and for the duration of sample collections. They were fed their typical ration of grass hay and a 10% pelleted feed, except that food was withheld for 12 hours before to 4 hours after drug administration. All horses had access to fresh water at all times during the study. This study was approved by the North Carolina State University Institutional Animal Care and Use Committee.

Testing of drug stability—Before voriconazole was administered to the horses, stability and potency of the voriconazole powder in the vehicle were tested. To accomplish this, voriconazole powder was mixed with corn syrup to form a suspension with a final concentration of 33 mg/mL. This concentration was determined on the basis of a dosage of 4 mg/kg for a 500-kg horse administered in a total volume of 60 mL. After voriconazole was mixed with the vehicle, the suspension was stored at 8°C and protected from light. For drug analysis, the original suspension was diluted 1:1,000 by the addition of a mixture of 50% water and 50% acetonitrile to achieve a final concentration of 33 μg/mL. Voriconazole concentrations were determined immediately after mixing and 24 and 48 hours after mixing.

Drug administration—Voriconazole was administered IV or orally in accordance with a randomized crossover design. There was a minimum washout period of 3 weeks between drug administrations. A commercially available injectable formulation of voriconazole^b was used for IV administration. The drug was packaged in vials containing 200 mg of lyophilized voriconazole powder complexed with cyclodextrins to improve solubility. Each vial was reconstituted for injection by the addition of 19 mL of sterile water and then further diluted to a final volume of 250 mL by the addition of saline (0.9% NaCl) solution. The final concentrations of voriconazole ranged from 1.72 to 2.52 mg/mL (mean, 2.10 mg/mL) to comply with the manufacturer's dosing recommendations of 0.5 to 5.0 mg/mL. Voriconazole was then administered IV at a dose of 1 mg/kg via a 16-gauge, 3.25-inch catheter inserted in a jugular vein. The drug was infused during a period of 5 to 6 minutes.

For oral administration, voriconazole powder (99.4% pure) was provided by the manufacturer.^c The dose for each horse (4 mg/kg) was weighed, mixed with 60 mL of corn syrup, and administered as a suspension via an oral dosing syringe. Corn syrup was selected because it is commonly used in our clinic as a vehicle for oral administration of drugs and as a flavoring agent. The doses were chosen to mimic the range of doses used in other species.

Collection of samples—Blood samples (6 mL) were collected via a 14-gauge, 5.25-inch catheter inserted in a jugular vein. For IV administration, blood samples were collected from the jugular vein contralateral to the jugular vein used for drug administration. For both IV and oral administration, blood samples were collected before drug administration (time 0) and 10, 20, and 40 minutes and 1, 1.5, 2, 3, 4, 6, 8, 12, 24, 36, and 48 hours after drug administration. For IV administration, time points for sample collection related to the end of the actual drug infusion.

Each blood sample was transferred to lithium heparin tubes; tubes were immediately centrifuged and the plasma harvested. Plasma samples were frozen and stored at −70°C until analyzed.

Assessment of protein binding—Plasma from 6 healthy untreated adult horses was harvested and pooled for use in analysis of plasma protein binding. Three replicates of plasma were fortified with 2 concentrations of voriconazole (2.5 and 1 μg/mL). Samples were prepared for analysis and added to an in vitro microcentrifugation system.^d Samples were incubated for 30 minutes in a water bath at 37°C and then centrifuged at 1,000 × g for 20 minutes. A protein-free ultrafiltrate was obtained in the reservoir of the system. This ultrafiltrate, as well as unprocessed plasma, was then analyzed to determine voriconazole concentrations. Plasma protein binding was determined by use of the following equation:

View Expanded

Assessment of lipophilicity—The octanol:water partition coefficient was measured by use of established methods^7–9 and a procedure used by our laboratory group.^10,11 The aqueous phase consisted of a dibasic sodium phosphate buffer (0.1M) with the pH adjusted to 7.4 by use of 85% phosphoric acid. The buffer was then fortified by the addition of voriconazole (final concentration in 2 solutions, 5 and 2 μg/mL) and added to an equal volume of octanol^e in a screw-top tube. The tube was rotated for 1 hour at 25°C to allow equilibration of the drug between the lipid and aqueous phases and then centrifuged for 10 minutes at 2,000 × g. The aqueous layer was analyzed to measure the voriconazole concentration before and after incubation. Drug concentrations were determined with a calibration curve prepared by use of the phosphate buffer. The partition coefficient was determined by use of the following equation:

View Expanded

Measurement of drug concentrations—Voriconazole concentrations in plasma samples, as well as those for evaluations of plasma protein binding and assessment of lipophilicity, were determined by use of HPLC with UV detection at 263 nm. All plasma samples were subjected to solid-phase extraction by use of cyano-bonded cartridges^f before injection onto the HPLC system. Cartridges were attached to a vacuum manifold and initially conditioned with 1 mL of methanol and 1 mL of water. Each plasma sample (1 mL) was then extracted, and the cartridge was washed with 1 mL of a 95:5 (vol:vol) mixture of water:methanol. Samples were then eluted by the use of 1 mL of 100% methanol into clean glass tubes. The resulting eluate was evaporated under a stream of compressed room air at 40°C for 25 minutes. Evaporated samples were reconstituted by the addition of 200 μL of mobile phase, which consisted of a 50:50 (vol:vol) solution of water:acetonitrile with 0.02% trifluoroacetic acid added to adjust the pH to 2.38. A C8 reverse-phase column^g was used for separation. Flow rate was 1.0 mL/min, and the injection volume was 25 μL.

Calibration curves were prepared daily by use of pooled plasma obtained from untreated horses. A blank sample was processed and analyzed at the beginning of each assay to check for interfering peaks. Calibration curves were linear between the concentrations of 0.025 and 10 μg/mL (r² > 0.99), and all values were within 15% of the expected range. The lower limit of quantification was defined as the lowest concentration that was consistently linear, as determined on the basis of regression analysis of the calibration curve. For these conditions, this value was 0.025 μg/mL for voriconazole in plasma. At concentrations of 5.0, 1.0, and 0.5 μg/mL, mean ± SD intraday accuracy of the HPLC assay was within 3.9 ± 1.8%, 1.6 ± 1.5%, and 2.7 ± 1.7% of the true value, respectively, and mean intraday precision was within 4.3%, 1.5%, and 1.7% of the mean, respectively. Mean recovery of voriconazole from plasma was 103.11 ± 3.77%.

Pharmacokinetic analysis—Plasma concentrations of voriconazole were analyzed with compartmental pharmacokinetic methods by use of a specialized computer program.^h Appropriate compartmental models were chosen on the basis of visual inspection of plasma concentration-versus-time curves plotted on semilogarithmic paper as well as by use of Akaike information criterion. Weighting adjustments were chosen separately for each compartmental model on the basis of the best fit of the data at the terminal portion of the elimination curve. Our laboratory personnel used standard methods for analysis, as described elsewhere.¹²

For IV administration, a 2-compartment open model with elimination from the central compartment was used. The model was described by the following equation:

View Expanded

For oral administration, a 1-compartment open model with first-order input was used. The model was described by the following equation:

View Expanded

View Expanded

Results

Voriconazole was tolerated well by all horses after oral and IV administration. No adverse effects were detected.

Voriconazole was stable in the corn syrup mixture for a minimum of 48 hours when stored at 8°C. Potency was within 5% of the predicted values. Mean ± SD in vitro plasma protein binding of voriconazole in horses was 31.68 ± 1.92%. Mean octanol:water partition coefficient of voriconazole (pH, 7.4) was high (64.69 ± 0.38). This value is equivalent to a logarithm of the partition coefficient of 1.81.

Plasma concentrations over time after IV and oral administration of voriconazole were plotted (Figure 1). After IV administration, mean ± SD value for clearance of voriconazole was low (1.89 ± 0.46 mL/kg/min) and the mean Vd at steady state was high (1.35 ± 0.06 L/kg). Mean terminal half-life was 8.89 ± 2.31 hours. Other relevant pharmacokinetic variables after IV and oral administration of voriconazole were summarized (Tables 1 and 2).

View Full Size

Mean ± SD plasma concentrations of voriconazole at various time points after oral (4 mg/kg; circles) and IV (1 mg/kg; triangles) administration of a single dose of voriconazole to 6 horses. There was a minimum wash-out period of 3 weeks between successive administrations to each horse.

Citation: American Journal of Veterinary Research 67, 6; 10.2460/ajvr.67.6.1070

• Download Figure
• Download figure as PowerPoint slide

Drug loss during oral administration to all 6 horses was minimal. After oral administration, mean ± SD systemic bioavailability was 135.75 ± 18.41%. The C_max was 2.43 ± 0.4 [.mu]g/mL at 2.92 ± 1.2 hours after administration. Mean terminal half-life was 13.11 ± 2.85 hours.

Discussion

Analysis of the results of the study reported here indicated that voriconazole was absorbed well after oral administration of a single dose (4 mg/kg), reaching a mean ± SD C_max of 2.43 ± 0.4 μg/mL approximately 3 hours after administration. Absorption of many drugs after oral administration has been correlated with water solubility as well as the permeability of the drug through a lipid membrane (eg, lipid membranes found in the cells of the gastrointestinal tract).¹³ Drugs that have high solubility and permeability are more readily absorbed and often have high bioavailability after oral administration. These principles can be applied to absorption of antifungal drugs in horses. Ketoconazole and itraconazole have solubility of < 0.01 mg/mL, and this solubility is a highly pHdependent event.¹⁴ Both of these drugs have poor bioavailability after oral administration to horses, unless they are administered in highly acidic formulations designed to increase solubility.^1,2

In contrast, fluconazole and voriconazole are soluble in water at concentrations of 1 to 10 mg/mL, and the solubility is less dependent on pH in the gastrointestinal tract.¹⁴ The improved solubility also makes these drugs more amenable to compounding with aqueous vehicles. This was supported for voriconazole in the study reported here, in which the oral formulation was stable when stored in a refrigerator and did not diminish absorption after oral administration.

Both fluconazole and voriconazole have systemic availability > 100% when administered to horses.³ Drug permeability has been correlated with the logarithm of the partition coefficient of the drug.¹³ Of the azole antifungal drugs, itraconazole, ketoconazole, and voriconazole are considered highly permeable with a logarithm of the partition coefficient of 5.16,ⁱ 3.78,^j and 1.81, respectively. Therefore, both solubility and permeability of voriconazole are favorable physicochemical properties that support our observation of excellent bioavailability in horses.

For the study reported here, a pure voriconazole powder was used for oral administration. When horses are orally administered tablets, the tablets are most commonly crushed or ground into a powder before administration, which makes them similar to the product used here. Because the commercially available tablets are designed for immediate release and do not contain specialized enteric coatings,^a absorption should be comparable to the powder used in this study.

The bioavailability of voriconazole in horses was > 100%. This would suggest a nonlinear disposition of the drug, although it may have been attributable to other factors, such as a more complete characterization of the terminal elimination phase because the drug was detected for 12 hours longer after oral administration, compared with the duration of detection after IV administration. Other commonly used signs of nonlinearity, such as a convex shape of the plasma concentration-versus-time curves and divergent (nonparallel) terminal phases, were not evident. We did not administer a range of doses for each route to examine whether there are saturable metabolisms in horses, nor did we administer multiple doses; therefore, the linearity or nonlinearity of voriconazole still needs to be investigated.

Mean ± SD apparent volume of distribution after IV administration of voriconazole was large (1.29 ± 0.08 L/kg). This may have been related to the low plasma protein binding measured in the study (31.68 ± 1.92%) because only free, unbound drug is available to diffuse into the tissues and interstitial fluid. Protein binding in horses is lower than that reported for other species. Values range from 45% in guinea pigs⁵ to 67% to 78% in mice.¹⁵ Values are 51% and 58% in dogs and humans, respectively.⁵ Compared with plasma protein binding of other triazole antifungals in horses, plasma protein binding of voriconazole is significantly lower than itraconazole (> 98%)² and, in our experience, higher than fluconazole (12.3%). In other studies^2,10,11 conducted by our laboratory group, protein binding was the major determinant of drug distribution into the interstitial fluid. Therefore, protein binding of the magnitude described in the study reported here should not impede penetration of voriconazole into the interstitial space, which is the site of drug action when combating a fungal infection.

The mean ± SD half-life of voriconazole reported here was long after both oral and IV administration (13.11 ± 2.85 hours and 8.89 ± 2.31 hours, respectively), which is convenient for a proposed once-daily dosing regimen. Differences in the half-life values between oral and IV administration may be caused by nonlinear pharmacokinetics with the higher orally administered dose or by prolonged absorption as a result of stomach emptying and transit through the gastrointestinal tract. The half-life of voriconazole is longer in horses than in other species. Because the Vd is similar among species (1.35 L/kg in horses, 1.3 L/kg in dogs, and 2.1 L/kg in rats),⁵ this is most likely related to lower systemic clearance in horses (1.89 mL/min/kg). Clearance of voriconazole in other species is mainly dependent on metabolism.⁵ The major metabolizing enzyme identified in humans is the cytochrome P450 isoenzyme 2C19.¹⁶ Additional studies are necessary to determine the amounts and activities of metabolizing enzymes in horses.

The effects of multiple doses of voriconazole have been studied^5,17 in several animals, including humans, rats, mice, dogs, rabbits, and guinea pigs. Substantial differences among species are evident with regard to the effects of voriconazole on the hepatic enzymes responsible for its own metabolism. In rats, mice, and, to a lesser extent, dogs, autoinduction of hepatic cytochrome P450 metabolizing enzymes has been reported in repeat-dosing experiments.⁵ This results in a decrease in the AUC and C_max over time, and steadystate concentrations were not reached at any point during the sample collection period in that study. Because of the autoinduction seen in these species, regimens that involve escalation of doses are often required to maintain therapeutic drug concentrations. However, autoinduction of metabolism was not seen in humans, rabbits, and guinea pigs, and the drug reached steadystate concentrations after approximately 5 days when the drug was administered in accordance with a twicedaily dosing regimen.^5,16 Unless studies are conducted to determine pharmacokinetics and concentrations after administration of multiple doses to horses, it is not possible to predict the effect of voriconazole on hepatic metabolizing enzymes in horses.

Dosing regimens for antibacterial drugs are often based on the PK-PD markers predictive of clinical outcome, including the amount of time for which plasma concentrations are greater than the MIC, the ratio of AUC_0–24 to MIC, or the ratio of C_max to MIC.¹⁸ Unfortunately, these variables are not as clearly defined for antifungal drugs. A limited number of in vivo and in vitro studies have been conducted to examine the PK-PD of voriconazole against various species of fungi. Voriconazole has fungistatic activity against most Candida isolates, with maximum activity attained at 3 times the MIC in vitro.^19,20 In an in vivo study¹⁵ that involved neutropenic rats with systemic candidiasis, the PK-PD variable most closely associated with efficacy was a ratio of AUC_0–24 to MIC of 20 to 25. This is similar to results obtained with other triazole antifungals, including fluconazole and the newer agents, posaconazole and ravuconazole,^21–23 which suggests that PK-PD variables are similar among the classes of antifungals. Reported MICs for Candida spp are usually ≤ 0.39 μg/mL.^3,16,24 In the study reported here, the ratio of AUC_0–24 to MIC for horses at a dose of 1 mg/kg, IV, and 4 mg/kg, PO, was 21 and 91, respectively. Therefore, either of these administrations would be expected to attain plasma concentrations consistent with an effective outcome in the treatment of horses with systemic candidiasis.

In contrast to the effects of voriconazole against Candida spp, voriconazole has fungicidal activity against Aspergillus isolates in vitro.²⁵ Maximum fungicidal activity was detected at 2 to 4 times the MIC in in vitro experiments.¹⁹ The reported MICs for Aspergillus spp are usually ≤ 0.5 μg/mL. Therefore, by use of 0.5 μg/mL as the target MIC, and on the basis that the mean C_max calculated for horses in the study was 2.43 μg/mL, we achieved a ratio of C_max to MIC of 4.86, which should be adequate for the treatment of horses with aspergillosis when voriconazole is administered orally at the rate of 4 mg/kg. Furthermore, plasma concentrations remained higher than the target MIC for > 24 hours in all 6 horses. Alternatively, the ratio of AUC_0–24 to MIC for Aspergillus spp was approximately 16 and 71 for the IV and oral administrations, respectively, which may be low for the IV administration but is higher than necessary for the oral administration.

Voriconazole may also be useful for the treatment of horses with other fungal diseases. For example, Histoplasma spp, Blastomyces spp, and Coccidioides spp have MICs ≤ 0.03 μg/mL²⁶; Cryptococcus neoformans has an MIC < 1 μg/mL^27–29; and Fusarium spp have MICs in the range of 1 to 4 μg/mL.²⁴

We conclude that voriconazole has high systemic availability after oral administration in horses. Because it also has low protein binding, a high lipophilic partition coefficient, and high antifungal activity, voriconazole may be a useful drug for the treatment of horses with fungal infections. Long-term dosing studies are needed to determine the effects of multiple doses on metabolizing enzymes as well as to determine whether the drug accumulates. On the basis of analysis of results of the study reported here, a dose of 4 mg/kg administered orally once daily would be more than adequate for the treatment of horses with fungal infections for which the fungi have an MIC ≤ 1 μg/mL. Lower doses of voriconazole may be possible for use in treating horses with some infections; however, given the potential for nonlinearity of this drug, it will be necessary to conduct dose escalation studies and clinical trials to determine the optimal dosing regimen.