A Simple High-Performance Liquid Chromatography Method for Simultaneous Determination of Three Triazole Antifungals in Human Plasma

Abstract

A rapid and simple high-performance liquid chromatography (HPLC) assay was developed for the simultaneous determination of three triazole antifungals (voriconazole, posaconazole, and itraconazole and the metabolite of itraconazole, hydroxyitraconazole) in human plasma. Sample preparation involved a simple one-step protein precipitation with 1.0 M perchloric acid and methanol. After centrifugation, the supernatant was injected directly into the HPLC system. Voriconazole, posaconazole, itraconazole, its metabolite hydroxyitraconazole, and the internal standard naproxen were resolved on a C₆-phenyl column using gradient elution of 0.01 M phosphate buffer, pH 3.5, and acetonitrile and detected with UV detection at 262 nm. Standard curves were linear over the concentration range of 0.05 to 10 mg/liter (r² > 0.99). Bias was <8.0% from 0.05 to 10 mg/liter, intra- and interday coefficients of variation (imprecision) were <10%, and the limit of quantification was 0.05 mg/liter.

INTRODUCTION

Invasive fungal infections are a serious and challenging clinical problem and remain a frequent cause of significant morbidity and mortality (1, 2). Voriconazole, posaconazole, and itraconazole are triazole antifungals that are widely used for the management of fungal infections as a result of their broad-spectrum antifungal activity. Hydroxyitraconazole, the major metabolite of itraconazole, has antifungal potency comparable to that of the parent drug (3). Triazole antifungals display clinically relevant concentration-effect relationships for both desired effect and toxicity. They also exhibit marked variability in patient plasma drug concentrations as a result of inconsistent absorption, metabolism, elimination, or interaction with concomitant medications. Therapeutic drug monitoring (TDM) of these antifungals can therefore be useful to both optimize efficacy and reduce drug toxicity (1–3).

TDM of triazole antifungals requires a rapid and simple analytical method to analyze the concentrations in plasma. Various analytical methods for simultaneously measuring triazole antifungals, including voriconazole, posaconazole, and itraconazole, in human plasma have been developed. These methods have used liquid chromatography-mass spectrometry (LC-MS) (4) or LC-tandem mass spectrometry (LC-MS/MS) (5–7). It appears that only one high-performance liquid chromatography (HPLC) method has been reported for simultaneously measuring multiple triazole antifungals (without the metabolite of itraconazole, hydroxyitraconazole), which requires 300 μl of plasma and solid-phase extraction for sample preparation (8).

In recent years, LC-MS/MS has been demonstrated to be a powerful technique for the quantitative determination of drugs and metabolites in biological fluids. It provides high selectivity, as well as simplification of both sample extraction procedures and chromatography. However, LC-MS/MS systems are not always available in ordinary hospital laboratories, because they are expensive compared with a simple HPLC system. The aim of the current work was to develop and validate a rapid, simple, robust, and reliable HPLC method for the simultaneous determination of voriconazole, posaconazole, itraconazole, and the metabolite of itraconazole, hydroxyitraconazole, in human plasma, suitable for routine use in ordinary, busy hospital laboratories.

For sample preparation, HPLC methods normally require a complex extraction procedure, such as liquid-liquid extraction or solid-phase extraction, to clean up plasma samples. These extraction procedures are reagent and time consuming, and therefore, an aim of the current work was also to utilize a simple one-step protein precipitation for sample preparation, which would reduce the turnaround time for results.

MATERIALS AND METHODS

Materials.

Voriconazole, posaconazole, itraconazole, and hydroxyitraconazole in analytical-grade pure powder forms were used as standards for the HPLC assay. Voriconazole was kindly donated by Pfizer (New York, NY). Posaconazole was purchased from Merck (Rahway, NJ). Itraconazole and hydroxyitraconazole were kindly donated by Janssen Research Foundation (Beerse, Belgium). Naproxen (the internal standard, an analytical-grade pure powder) was purchased from Sigma Co. (Australia). HPLC-grade acetonitrile, methanol, and perchloric acid were purchased from BDH (Poole, United Kingdom). Distilled, deionized water was produced by using a Milli-Q reagent water system (Millipore, MA). The human plasma used as the assay blank and for the preparation of standards and quality control samples was obtained from New Zealand Blood Services (Christchurch, New Zealand).

Chromatography.

HPLC analysis was performed on an Agilent 1200 series system equipped with a quaternary pump, a variable wavelength detector set at 262 nm, and a refrigerated autosampler set at 4°C (Hewlett-Packard, Waldbronn, Germany). Voriconazole, posaconazole, itraconazole, hydroxyitraconazole, and the internal standard naproxen were separated under gradient elution using a Gemini C₆-phenyl, 5-μm, 150- by 4.6-mm-internal-diameter analytical column equipped with a Gemini C₆-phenyl, 4- by 3.0-mm-internal-diameter guard column (Phenomenex, Torrance, CA). The mobile phase consisted of solvent A (0.01 M phosphate buffer, pH 3.5, containing 5% acetonitrile) and solvent B (95% acetonitrile in water). The flow rate was set at 1.0 ml/min. The time program for gradient elution is shown in Table 1. The total analysis time for each sample was 17 min.

Table 1.

Time program for gradient elution

Time (min)	% Solventa:
	A	B
0.00	70	30
0.01	50	50
5.00	40	60
6.00	40	60
10.0	30	70
12.0	30	70
12.1	70	30
17.0	Stop

^aThe mobile phase consisted of solvent A (0.01 M phosphate buffer, pH 3.5, containing 5% acetonitrile) and solvent B (95% acetonitrile in water). The flow rate was set at 1.0 ml/min.

Standards.

The standard stock solutions of voriconazole and posaconazole (both 1.0 mg/ml as free base) were prepared by dissolving 10 mg of the respective standards in 10 ml of methanol and stored at 4°C. A standard stock solution of itraconazole (0.4 mg/ml as free base) was prepared by dissolving 10 mg of itraconazole in 25 ml of methanol and stored at 4°C. A standard stock solution of hydroxyitraconazole (0.1 mg/ml as free base) was prepared by dissolving 10 mg of hydroxyitraconazole in 100 ml of methanol and stored at 4°C. Two sets of the same standard stock solutions were prepared for standard curves and for quality control (QC) samples, respectively. The standard curves of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole were constructed by spiking drug-free human plasma with the standard stock solution of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole, giving a calibration range of 0.05 to 10 mg/liter for all compounds. Voriconazole, posaconazole, itraconazole, and hydroxyitraconazole plasma quality control (QC) standards were prepared in single 10-ml aliquots in concentrations of 0.05, 0.2, 2.0, and 10 mg/liter and stored at −30°C until analyzed. The stock internal standard naproxen solution (0.6 mg/ml) was prepared by dissolving 15 mg of naproxen in 25 ml of methanol. A working solution of the internal standard (10 μg/ml) was prepared by diluting the stock solution with water.

Sample preparation.

The internal standard naproxen, 100 μl of 10 μg/ml, was added to 100 μl each of blank, standard, QC, or patient samples. The mixture was vortexed briefly. After adding 25 μl of 1.0 M perchloric acid to acidify the sample, 400 μl of methanol was added to precipitate the proteins. After centrifugation at 15,000 × g for 5 min, 30 μl of clear supernatant was injected into the HPLC system.

Validation.

The standard curves were the plots of the peak area ratios (analyte/internal standard) of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole versus the corresponding concentrations of these four compounds. The linearity of the standard curves was evaluated using linear regression analysis. To determine the levels of recovery of voriconazole, posaconazole, itraconazole, hydroxyitraconazole, and the internal standard naproxen from plasma, two sets of standards were prepared for voriconazole, posaconazole, itraconazole, and hydroxyitraconazole at concentrations of 0.05, 0.2, 2.0, and 10 mg/liter and of naproxen at 10 μg/ml, the concentration used in the assay. The first set was prepared in plasma from six different sources, and the second set in the extracts of plasma after protein precipitation of blank plasma from the same six different sources. Every set included six samples at each concentration. Absolute recoveries at each concentration were measured by comparing the peak areas of voriconazole, posaconazole, itraconazole, hydroxyitraconazole, and the internal standard naproxen in plasma to those in the spiked blank plasma extracts after protein precipitation at the corresponding concentrations (n = 6) as follows: [absolute recovery = (peak area of analyte from the spiked plasma sample)/(peak area of analyte from the spiked blank plasma extract sample after protein precipitation) × 100%]. Quality control was assessed by analysis of six samples at each concentration on the same day (intraday) and of one sample at each concentration on six different days (interday). Bias was determined as the measured minus the actual concentration, expressed as a percentage of the actual concentration. Imprecision was measured as intra- and interday coefficients of variation. The lowest concentrations for the standard curves of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole were considered to be the lower limit of quantitation (LLOQ) at which the concentrations of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole could be determined with acceptable accuracy and precision. According to the U.S. Food and Drug Administration “Guidance for Bioanalytical Method Validation” (9), the mean value determined at LLOQ should not deviate by more than 20% of the actual value, and the precision determined at LLOQ should not exceed 20% of the coefficient of variation (CV).

The effects of freezing and thawing on the concentrations of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole were studied using QC samples at 0.05, 0.2, 2.0, and 20 mg/liter, which were subjected to four freeze-thaw cycles before analysis. The stability of QC samples at −30°C was evaluated by concentration analysis at 2-week intervals for 3 months. The stability of stock standard solutions at 4°C for 6 months was evaluated by comparing the response with that of the freshly prepared standard solution. The stability of the processed samples at 4°C (the temperature of the autosampler) for 24 h was evaluated by comparing the results with the original results. In all cases, the compounds were considered to be stable as long as degradation was <10% of the concentration at time zero.

RESULTS AND DISCUSSION

Chromatography.

To achieve optimized chromatographic resolution, peak sharpness, and signal intensity, a variety of LC analytical columns, including the Phenomenex Gemini C₆-phenyl, Aqua C₁₈, and Luna C₈ columns, and mobile phase compositions including 0.01 M phosphate buffer, pH 3.5, and various concentrations of acetonitrile and methanol were evaluated. The optimized LC condition was a mobile phase consisting of 0.01 M phosphate buffer, pH 3.5, containing 5% acetonitrile and 95% acetonitrile in water with gradient elution on a Phenomenex Gemini C₆-phenyl column. Under these conditions, the retention times were approximately 4.3, 5.5, 6.8, 7.2, and 11.3 min for voriconazole, internal standard naproxen, posaconazole, hydroxyitraconazole, and itraconazole, respectively (Fig. 1). Blank plasma samples from six different sources of the same matrix were tested for interference, and the peaks of all the compounds of interest were free of interference from any other peaks present in the blanks (Fig. 1).

Fig 1.

Representative HPLC chromatograms of blank plasma (top), plasma sample spiked with voriconazole, posaconazole, itraconazole, and hydroxyitraconazole at 5.0 mg/liter (middle), and the commercial antifungal QC sample with voriconazole, posaconazole, itraconazole, and hydroxyitraconazole at 1.5 mg/liter (bottom). I.S., internal standard.

Sample preparation.

Protein precipitation is the simplest and most rapid method of plasma sample preparation for the measurement of drug concentrations. To find the most efficient precipitant for sample preparation, three widely used precipitating agents (acetonitrile, methanol, and perchloric acid) were compared. Precipitation with acetonitrile or perchloric acid gave better sample cleanup, but the levels of recovery of posaconazole and the internal standard naproxen were lower. Precipitation with methanol gave high extraction efficiency, but the sample cleanup was less efficient and the supernatant was not very clear. Acidifying plasma samples with a small volume of 1.0 M perchloric acid and then extracting with methanol provided the cleanest samples and highest levels of recovery of voriconazole, posaconazole, itraconazole, hydroxyitraconazole, and the internal standard naproxen.

Method validation.

The standard curves of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole were linear (r² > 0.99) over the concentration range of 0.05 to 10 mg/liter. The intercept with the y axis of each standard curve was not significantly different from zero. The slopes of standard curves in the 6 different preparations for each compound were practically the same. The mean linear regression equations of plasma standard curves were as follows: y = 0.086 (±0.0055)x − 0.00068 (±0.0029) for voriconazole, y = 0.203 (±0.0.012)x − 0.0019 (±0.0054) for posaconazole, y = 0.204 (±0.0188)x − 0.0016 (±0.0016) for itraconazole, and y = 0.194 (±0.0155)x − 0.000017 (±0.0011) for hydroxyitraconazole, where y represents the ratio of analyte peak area to that of the internal standard and x represents the plasma concentration of analyte. The LLOQ was 0.05 mg/liter. The accuracy and imprecision were assessed at the LLOQ and low-, medium-, and high-level QC concentrations. The amounts of bias for LLOQ and QCs were <8.0% (Tables 2 and 3). Imprecision was small, as indicated by both intra- and interday coefficients of variation of <10% at the LLOQ and QC concentrations (Tables 2 and 3). The absolute recoveries of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole from plasma at concentrations of 0.05, 0.2, 2.0, and 20 mg/liter were similar and consistent, with mean values of around 100%. The absolute recovery of the internal standard naproxen at the concentration employed was 103% ± 0.60% (n = 6).

Table 2.

Intraday coefficients of variation and accuracy of determination of various antifungals or metabolite in plasma^a

Antifungal or metabolite	Sample	Concn (mg/liter)		% Bias	% Imprecision (CV)
		Spiked	Found (mean ± SD)
Voriconazole	LLOQ	0.05	0.052 ± 0.0015	3.3	3.0
	QC1	0.20	0.207 ± 0.0042	3.6	2.0
	QC2	2.00	2.08 ± 0.026	3.9	1.3
	QC3	10.0	9.98 ± 0.080	−0.2	0.8
Posaconazole	LLOQ	0.05	0.050 ± 0.0007	0.1	1.4
	QC1	0.20	0.203 ± 0.0057	1.6	2.8
	QC2	2.00	2.15 ± 0.030	7.7	1.4
	QC3	10.0	9.99 ± 0.088	−0.1	0.9
Itraconazole	LLOQ	0.05	0.052 ± 0.0013	3.0	2.6
	QC1	0.20	0.202 ± 0.0032	0.9	1.6
	QC2	2.00	2.00 ± 0.018	0.1	0.9
	QC3	10.0	10.0 ± 0.064	0.0	0.6
Hydroxyitraconazole	LLOQ	0.05	0.051 ± 0.0017	2.5	3.3
	QC1	0.20	0.202 ± 0.0042	0.9	2.1
	QC2	2.00	2.06 ± 0.020	3.0	1.0
	QC3	10.0	10.0 ± 0.072	0.0	0.7

^aIntraday coefficients of variation (% CV) and accuracy (% bias) of the determination of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole in plasma (n = 6).

Table 3.

Interday coefficients of variation and accuracy of determination of various antifungals or metabolite in plasma^a

Antifungal or metabolite	Sample	Concn (mg/liter)		% Bias	% Imprecision (CV)
		Spiked	Found (mean ± SD)
Voriconazole	LLOQ	0.05	0.052 ± 0.0028	3.2	5.4
	QC1	0.20	0.204 ± 0.069	1.8	3.4
	QC2	2.00	2.02 ± 0.109	1.1	5.4
	QC3	10.0	10.0 ± 0.239	0.4	2.4
Posaconazole	LLOQ	0.05	0.051 ± 0.0026	2.2	5.0
	QC1	0.20	0.201 ± 0.0076	0.3	3.8
	QC2	2.00	2.04 ± 0.102	1.9	5.0
	QC3	10.0	10.4 ± 0.227	4.1	2.2
Itraconazole	LLOQ	0.05	0.052 ± 0.0023	3.0	4.5
	QC1	0.20	0.202 ± 0.0068	0.9	3.4
	QC2	2.00	2.02 ± 0.074	1.2	3.7
	QC3	10.0	10.1 ± 0.223	1.3	2.2
Hydroxyitraconazole	LLOQ	0.05	0.051 ± 0.0041	2.2	8.1
	QC1	0.20	0.196 ± 0.0082	−2.0	4.2
	QC2	2.00	1.95 ± 0.051	−2.7	2.6
	QC3	10.0	10.0 ± 0.268	0.5	2.7

^aInterday coefficients of variation (% CV) and accuracy (% bias) of the determination of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole in plasma (n = 6).

Voriconazole, posaconazole, itraconazole, and hydroxyitraconazole were found to be stable in plasma for at least four freeze-thaw cycles when stored at −30°C. The QC samples at concentrations of 0.05, 0.2, 2.0, and 20 mg/liter were stable for at least 3 months at −30°C. Stock standard solutions of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole remained stable for at least 6 months at 4°C. The processed samples were stable for at least 24 h at 4°C. In all cases, the concentrations of stored samples of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole deviated <5% from the freshly prepared samples.

Application of the assay.

The method presented is currently being used in our laboratory service to monitor the concentrations of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole in plasma for patients on antifungal therapy (Fig. 2). There were no samples with concentrations above the upper limit of quantitation (ULOQ). Although the concentrations of patient samples varied widely, they were well encompassed by the concentration ranges of the standard curves. To ensure the accuracy and reproducibility of the method, the commercial antifungal QC samples at two levels (1.5 mg/liter and 5.0 mg/liter) and some plasma samples from patients on voriconazole, posaconazole, or itraconazole (over the concentration range of 0.05 to 5.0 mg/liter) were reanalyzed. All the repeat values were <10% different from the target values of QC samples or the initial value of the patient samples. Requests for antifungal assays are received almost daily in clinical practice. We have analyzed a total of 304 samples obtained from 126 patients receiving voriconazole, posaconazole, or itraconazole, including 152 samples from 53 patients on voriconazole, 128 samples from 69 patients on itraconazole, and 24 samples from 14 patients on posaconazole. Requests for posaconazole analysis have been increasing steadily. Blood samples were taken from patients before the next dose at steady state to analyze trough concentrations. Our in-house recommended therapeutic ranges are 1.0 to 5.0 mg/liter for voriconazole, >1.0 mg/liter for the combined concentration of itraconazole and its metabolite hydroxyitraconazole, and >0.7 mg/liter for posaconazole. During our routine clinical service, a high number of subtherapeutic plasma concentrations were observed from the first patient samples analyzed for antifungal concentrations, especially for posaconazole concentrations, including 25 of 53 patients (38%) with voriconazole concentrations of <1.0 mg/liter, 28 of 69 patients (41%) with combined concentrations of itraconazole and its metabolite hydroxyitraconazole of <1.0 mg/liter, and 12 of 14 patients (86%) with posaconazole concentrations of <0.7 mg/liter. There are many causes of subtherapeutic antifungal concentrations, and therefore, concentration monitoring of these antifungals can be useful to identify patients who need increased doses so that therapeutic plasma concentrations are achieved.

Fig 2.

Representative HPLC chromatograms of plasma sample from a patient on voriconazole (voriconazole concentration = 4.2 mg/liter) (top), plasma sample from a patient on posaconazole (posaconazole concentration = 2.2 mg/liter) (middle), and plasma sample from a patient on itraconazole (itraconazole concentration = 0.87 mg/liter and hydroxyitraconazole = 1.51 mg/liter) (bottom).

Conclusions.

A validated HPLC method for the determination of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole has been described. In comparison with the published HPLC method, this method required a smaller volume of plasma and used simpler sample preparation without losing specificity and sensitivity. The method has proven to be rapid, simple, robust, and reliable and is currently being used in routine clinical service to monitor the plasma concentrations of voriconazole, posaconazole, itraconazole, and hydroxyitraconazole in patients on antifungal therapy.