Ribavirin Quantification in Combination Treatment of Chronic Hepatitis C

Ribavirin in combination with alpha 2 interferon is the consensustreatment for chronic hepatitis C. However, recent preliminarypharmacological studies have suggested that the bioavailabilityof ribavirin displays great interindividual variability. Inorder to monitor serum ribavirin levels during combination treatment,we developed and validated a quantitative assay using an approachadaptable for routine hospital laboratories. The method involvedsolid-phase extraction on phenyl boronic acid cartridges followedby high-performance liquid chromatography with a C₁₈-bondedsilica column and a mobile phase containing 10 mM ammonium phosphatebuffer (with the pH adjusted to 2.5) and UV detection (207 nm).The sensitivity, recovery, linearity of the calibration curve,intra- and interassay accuracies, precision, and stability at4°C were consistent with its use in the laboratory routine.In addition, other nucleoside analogues sometimes used withribavirin in patients coinfected with hepatitis C virus (HCV)and human immunodeficiency virus did not interfere with thequantification of ribavirin levels. The ribavirin concentrationwas quantified in 24 serum samples from patients with chronichepatitis C treated with a combination of ribavirin and alpha2 interferon. The mean serum ribavirin concentration was 2.67± 1.06 µg/ml (n = 24) at week 12 of treatment (W12)and 3.24 ± 1.35 µg/ml (n = 24) at week 24 of treatment(W24). In addition, ribavirin concentrations displayed highinterindividual variabilities: the coefficients of variationof the serum ribavirin concentrations adjusted to the administereddose were 44 and 48% at W12 and W24, respectively. Moreover,the ribavirin concentration was higher in patients with a sustainedvirological response (n = 11) than in patients with treatmentfailure (n = 13), with significant intergroup differences atW12 (P = 0.030) and W24 (P = 0.049). The present study describesa simple analytical method for the quantification of ribavirinin human serum that could be a useful tool for the monitoringof ribavirin concentrations in HCV-infected patients in orderto improve the efficacy and safety of therapy with ribavirinplus interferon.

INTRODUCTION

Top

Introduction

In patients chronically infected with hepatitis C virus (HCV),combination therapy with ribavirin plus alpha 2 interferon hasgreatly improved the rates of sustained biochemical and virologicalresponse compared to those achieved with interferon monotherapy(3, 6, 12, 14). Unfortunately, more than 40% of treated patientseither remain nonresponsive or have relapses with this treatment.Although some epidemiological factors support the predictivevalue of the treatment in terms of the long-term outcome, noclear explanation for the lack of efficacy in some patient subgroupshas been provided. Resistance to interferon is the hypothesisthat has been the most investigated, but other mechanisms includingthose related to ribavirin should also be closely examined.

The pharmacokinetics of ribavirin were initially investigatedin studies with healthy volunteers and patients after the administrationof either a single dose or multiple doses (4). Although intermediatebioavailability and a high level of interindividual variabilitywere described, no further evaluation was performed with patientswith chronic HCV infection receiving combination therapy. However,the extensive use of this drug, which induces numerous sideeffects, requires complementary pharmacological investigationsin order to adjust the dose of the drug according to individualpatient characteristics. For instance, no data on the bioavailabilityof ribavirin from large-scale clinical trials of the efficacyof combination treatment have been published. One reason couldbe that no standard assay for plasma ribavirin concentrationdetermination was available for routine laboratories, althoughdescriptions of some analytical methods have been published,including radioimmunoassay (1), high-performance liquid chromatography(HPLC) with UV (5, 7, 10, 13, 19, 20) or mass spectrometric(15) detection, and gas chromatography-mass spectrometry (16).However, these methods differ according to their sensitivities,specificities, and difficulty (time and cost).

The main purpose of the present study was to adapt and validatea routine assay for the quantification of ribavirin in serumfor the management of therapy in patients with chronic HCV infection.In this regard, an HPLC method with commercially available reagentswas chosen since its use is feasible in most hospital laboratories.In order to illustrate this technique, we retrospectively testeda small number of treated patients and compared their serumribavirin concentrations to their response to treatment.

MATERIALS AND METHODS

Top

Materials and Methods

Chemicals and instruments. All chemicals were of HPLC or reagent grade and were obtainedfrom Prolabo (Paris, France). Ribavirin (1-beta-D-ribofuranosyl-1H-1,2,4triazole-3-carboxamide) (purity, 99%) and 3-methylcytidine methosulfate(purity, 98%), which was used as an internal standard (IS),were purchased from Sigma (Saint-Quentin Fallavier, France).Zidovudine, lamivudine, and abacavir were from Glaxo Wellcome(London, United Kingdom); acyclovir was from Glaxo Wellcome(Marly le Roi, France); and zalcitabine and ganciclovir werefrom Roche (Neuilly-sur-Seine, France). Didanosine and stavudinewere from Bristol-Myers Squibb (Princeton, N.J.).

Phenyl boronic acid cartridges (Bond Elute phenyl boronic acid[100 mg/10 ml]), which were used for solid-phase extraction(SPE), were obtained from Varian (Harbor City, Calif.) SamplePreparation Product. The cartridges were positioned on a VacElut vacuum manifold from the Varian Sample Preparation Product.Stock solutions of ribavirin (10 mg/ml) and 3-methylcytidinemethosulfate (10 mg/ml) were prepared in distilled water andwere stored in 1.0-ml aliquots at -20°C until use. Workingsolutions of ribavirin (100 and 10 µg/ml) and 3-methylcytidinemethosulfate (100 µg/ml) were prepared weekly by dilutionof stock solutions in methanol. Stock solutions of anti-humanimmunodeficiency virus (anti-HIV) nucleoside analogues (0.1mg/ml) were prepared in methanol, stored at -20°C untiluse, and then diluted in distilled water to 30 µg/ml forabacavir and to 10 µg/ml for all the other nucleosideanalogues.

HPLC apparatus and conditions. The HPLC system used consisted of the following: an M-510 HPLCpump (Waters Associates, Milford, Mass.) set to deliver 0.7ml/min, which led to a typical back pressure of 1,500 lb/in²;a Rheodyne 7125 injection system with an injection volume of20 µl; a C₁₈-bonded silica column with 5-µm beads(3.9 by 300 mm; Novapak); an L4000 UV detector (Merck, Darmstadt,Germany); and a Shimadzu model C-R3A integrator.

The wavelength was fixed at 207 nm, and the sensitivity was0.005 absorbance units. The mobile-phase solvent consisted of10 mM ammonium phosphate buffer (the pH was adjusted to 2.5with orthophosphoric acid). After the run analysis, the HPLCcolumn was stored in an acetonitrile-water (50:50; vol/vol)mixture.

Solid-phase extraction. The extraction procedure that follows was performed by a methodologyderived from a previous report (7). The assay was calibratedby using five nonzero calibrators containing 0.2, 0.5, 1, 3,and 5 µg of ribavirin per ml and the IS (2.5 µg/ml).The calibrators were obtained by diluting the working solutionswith blank bovine serum (Biotrol, Biotrol Diagnosis, Chennevrières,France). Previous experiments demonstrated that double blankswere similar in human serum and bovine serum (data not shown).

The samples (1 ml) were diluted (1:4) in 250 mM ammonium acetatebuffer (the pH was adjusted to 8.5 with ammoniac). The dilutedsamples were loaded onto phenyl boronic acid cartridges thathad been pretreated with 1 ml of 3% formic acid in methanol,followed by 5 ml of 250 mM ammonium acetate buffer (pH 8.5).The samples were drained away and the cartridges were washedfive times with 1-ml aliquots of 250 mM ammonium acetate buffer.Ribavirin and the IS were subsequently eluted with 1 ml of 3%formic acid into glass tubes. The effluents were dried undernitrogen and reconstituted with 200 µl of the mobile phase.Twenty microliters of the reconstituted samples was injectedonto the HPLC column.

Validation of the method. The validation of the method was performed in accordance withFood and Drug Administration guidelines (17, 18). The linearitywas tested by assaying a set of five calibrators (0.2, 0.5,1, 3, and 5 µg/ml) six times and by calculating the measuredconcentration of each calibrator from the calibration curve.A plot of the percent deviation from the calculated concentrationagainst the expected concentration was drawn and inspected fortrends.

Within-day precision was calculated by repeated analysis ofspiked serum samples (0.2, 1, and 3 µg/ml) during 1 workingday by the same operator. Between-day precision was calculatedby analysis of serum samples spiked with the same concentrationof ribavirin, with one analysis being performed a day.

The stability of ribavirin in serum was assessed by five replicateanalyses of three samples spiked with ribavirin (0.2, 1, and3 µg/ml). The samples were analyzed immediately or afterstorage at 4°C for 24, 48, 72, or 96 h, respectively.

To determine the stability of ribavirin after freezing-thawing,control plasma samples spiked with 2 µg of ribavirin perml were compared with samples of the same batch that were frozenat -20°C and thawed at ambient temperature. This freeze-thawcycle was repeated three times. In addition, the stability ofribavirin in dried residues after SPE was tested with four replicatesspiked with 0.2, 1, and 3 µg of ribavirin per ml. Eachsample was extracted, dried under nitrogen, and analyzed immediatelyor after storage at -20°C for 24, 48, or 72 h, respectively.

In order to investigate the influence of the matrix on the concentrationof ribavirin, the specificity of the assay was investigatedby analyzing blank plasma from six different human plasma samplesspiked with 2 µg of ribavirin per ml.

Lastly, some nucleoside inhibitors most likely to be encounteredin the plasma of HCV-positive patients as a result of concomitanttherapy were screened by the HPLC assay. Drugs were tested aspure solutions diluted in methanol so that the concentrationswere 30 µg/ml for abacavir and 10 µg/ml for theother nucleoside analogue tested; the drugs were then injectedinto the system, and their peak retention times were comparedwith the peak retention time of ribavirin. An interfering drugwas defined as a molecule that exhibited a retention time within0.3 min of that of ribavirin or the IS.

Demographic and clinical characteristics of the patients tested. Twenty-four patients with chronic HCV infection previously treatedwith combination therapy were selected for the analysis, providedthat blood samples recovered on week 12 of treatment (W12) andweek 24 of treatment (W24) were available from a stored serumbank. The patients' characteristics are summarized in Table1. All patients were infected with a genotype 1 isolate of HCVand were drug-naive patients receiving their first anti-HCVtreatment, consisting of interferon plus ribavirin at the dosesusually prescribed at that time, as follows: subcutaneous injectionof standard alpha 2 interferon at 3 million IU three times perweek and 1,000- or 1,200-mg ribavirin tablets per day accordingto their weight (less than or more than 75 kg, respectively).Because of side effects caused by the drugs, two patients werereceiving a reduced ribavirin dose (600 mg/day) at the timeof assay for ribavirin quantification. All the patients werenegative for HBsAg and HIV infection. The clinical outcome analysisdefined among the 24 patients two subgroups of patients withequilibrated responses: 11 patients had sustained virologicalresponses and 13 patients had treatment failures (the patientswere either nonresponders or relapsers). Responders were definedas patients with a negative test result for HCV RNA at the endof therapy (48 weeks, which is the period of treatment for theresponders at present), and nonresponders were defined as patientswith a positive test result for HCV RNA at the end of a courseof 24 weeks of therapy.

View this table:
[in this window]
[in a new window]
TABLE 1. Clinical characteristics of patients with chronic HCV infection and treated with ribavirin-interferon combination therapy^a

Blood samples were collected from October 1998 to August 2000in Vacutainer tubes without additives. After centrifugationthe serum was transferred to another tube and stored at -80°Cuntil analysis. Ribavirin quantification was carried out withserum samples drawn 2 to 4 h after the morning administrationof ribavirin at W12 and W24 of a 48-week combination treatment.

No unusual investigations and the administration of no specificdrug was requested from the patients for the purpose of thisstudy, which conformed with local ethical considerations andthe principles outlined in the Declaration of Helsinki.

Statistical analysis. For each series of experiments, the variability of the ribavirinconcentration was estimated by determination of the relativestandard deviation (17, 18). The dose-normalized concentrationof ribavirin was expressed as the concentration in serum (inmicrograms per milliliter) adjusted to the daily ribavirin doseaccording to the patient's weight (milligrams per day per kilogramof body weight).

For the retrospective comparison of serum ribavirin concentrationsaccording to the outcome for the patients, a nonparametric Mann-Whitneytest was used in order to avoid the making of assumptions aboutthe distributions of the measured variables. P values <0.05were considered significant.

RESULTS

Top

Results

Ribavirin assay validation. Typical chromatograms of blank serum, blank serum with IS, serumsupplemented with a medium-range calibrator, and patient serumcontaining ribavirin are shown in Fig. 1. The retention timesof ribavirin and the IS were 6.7 and 11.0 min, respectively.

View larger version (20K):
[in this window]
[in a new window]
FIG. 1. Typical HPLC chromatograms of extracts of control human serum (A), control human serum spiked with 2.5 µg of IS per ml (B), control human serum spiked with 2.5 µg of IS per ml and 3.0 µg of ribavirin per ml (C), and serum from a patient receiving ribavirin (ribavirin concentration, 2.5 µg/ml) (D).

A plot of the ratio of the height of ribavirin/height of theIS versus the ribavirin concentration provides a linear relationshipdescribed by the equation y = 0.977 (± 0.161) x - 0.0173(± 0.095), with r equal to 0.988 ± 0.007. Thislinear correlation was verified over a range of concentrationsfrom 0.1 to 10.0 µg/ml.

Extraction recovery, expressed as a percentage, was definedas the ratio of the slope of the calibration curve for the extractedanalytes to the slope of the calibration curve for the nonextractedanalytes. The recovery was 87%.

Within-day and between-day precisions are given in Table 2.The lower limit of quantification and the limit of detectionwere set at 0.2 and 0.06 µg/ml, respectively, for extractedsamples.

View this table:
[in this window]
[in a new window]
TABLE 2. Within-day and between-day variabilities of ribavirin concentrations in serum samples measured by HPLC with UV detection^a

The concentrations of ribavirin in serum stored at 4°C werenot reduced over 96 h; the coefficient of variation of the slopesof the linear regression lines was 14%.

In contrast, storage of dried residues at -20°C for 72 hinduced a gradual decrease in the ribavirin concentration. Theslopes of the linear regression lines were 0.944 for fresh samplesand 0.787, 0.623, and 0.591 after storage at -20°C for 24,48, and 72 h, respectively. After storage for 72 h at -20°C,the concentration of ribavirin was reduced by 37% compared withthe concentration in fresh samples.

Three freeze-thaw cycles with spiked samples induced a trendtoward a loss of ribavirin. The concentration of ribavirin wasreduced by 33% compared with the concentration in fresh samplesafter three freeze-thaw cycles.

There was no influence of the matrix on the concentration ofribavirin (coefficient of variation, 6%). In addition, the nucleosideanalogues didanosine, stavudine, zidovudine, lamivudine, abacavir,zalcitabine, acyclovir, and ganciclovir did not interfere withthe analytical method.

Serum ribavirin concentrations during combination treatment. The mean absolute concentration of ribavirin at W12 was 2.67± 0.21 µg/ml (n = 24). The ribavirin concentrationat W24 (3.23 ± 0.28 µg/ml; n = 24) was similarto that at W12. The mean relative concentrations of ribavirinadjusted according to the dose of ribavirin received by thepatient (in milligrams per kilogram per day) were 0.20 ±0.02 µg/ml/dose (n = 24) at W12 and 0.24 ± 0.02µg/ml/dose (n = 24) at W24. The coefficients of variationfor these adjusted concentrations were 44 and 48% at W12 andW24, respectively. The ranges of ribavirin concentrations atW12 and W24 were 0.07 to 0.35 and 0.08 to 0.48 µg/ml/dose,respectively.

In order to evaluate the interest in quantification of ribavirinin serum in the management of patients treated for HCV infection,we retrospectively compared the concentrations of ribavirinaccording to the outcome status of the patients tested. In theresponder group of patients (n = 11), the relative concentrationsof ribavirin were 0.24 ± 0.03 and 0.28 ± 0.03µg/ml/dose at W12 and W24, respectively, while in patientswith treatment failures (n = 13) they were 0.16 ± 0.20and 0.21 ± 0.32 µg/ml/dose at W12 and W24, respectively.These relative concentrations of ribavirin were significantlydifferent between patients who were responders and those whowere nonresponders (P < 0.030 at W12 and P < 0.049 atW24 [Mann-Whitney test]) (Fig. 2).

View larger version (18K):
[in this window]
[in a new window]
FIG. 2. Concentration of ribavirin in sera from 11 responders and 13 nonresponders after 12 and 24 weeks of ribavirin and interferon combination therapy. Data are presented as box plots in which 50% of the values lie within the box. The horizontal lines drawn through the middle of the boxes represent the median values. The top and bottom of each box are the 10th and 90th percentiles of all values, respectively. P values were 0.03 and 0.05 for responder patients versus nonresponder patients after 3 and 6 months of treatment, respectively.

DISCUSSION

Top

Discussion

This study presents the validation of a simple bioanalyticalmethod for the quantification of ribavirin in human serum. Itsfeasibility for the monitoring of drug therapy for HCV-infectedpatients was also investigated.

The method was adapted from a previous report (7) by changingthe detection wavelength (207 nm) and modifying the SPE andpurification steps. Indeed, a methanolic solution containing3% formic acid was used to elute ribavirin in the SPE procedurein order to improve the evaporation step. Finally, the stabilityof ribavirin during sample storage and interference with othernucleoside analogues were assessed.

The method described here yielded linear results, with a lowerlimit of quantification of 0.2 µg/ml with samples fromwhich ribavirin was extracted. In addition, the limit of detection(0.06 µg/ml) was in the same range as those previouslydescribed for HPLC methods with UV detection (19, 20). The sensitivity,recovery, linearity of the calibration curve, and intra-andinterassay precisions of the present method agreed with thevalues specifically recommended for validation of bioanalyticalmethods (17, 18) and allowed the quantification of ribavirinover a wide range of clinically relevant concentrations (9).

No interference by other nucleotide analogues was found. Thisspecificity is due to the SPE procedure, which is based on theprinciple that ribavirin and other molecules with vicinal hydroxylgroups specifically bind to phenyl boronic acid at high pH.Since didanosine, stavudine, zidovudine, lamivudine, abacavir,zalcitabine, acyclovir, and ganciclovir have no vicinal hydroxylgroup in their structures, analytical interference was no longerexpected. It was important to validate this point, since patientscoinfected with HCV and HIV are more frequently treated witha combination of ribavirin and other nucleoside inhibitors.This combination treatment sometimes results in increased toxicityfor mitochondria (10) that could be managed by adaptation ofthe dose to match the clinical characteristics of these patients.

Analysis of the stability of ribavirin under various storageconditions showed that the ribavirin concentration remains stablein serum stored for 4 days at 4 or -20°C, whereas it tendsto decrease in dried residues stored at -20°C for 3 daysas well as in samples undergoing three freeze-thaw cycles. Thesedata suggest that storage of dried residues as well as freeze-thawcycles should be avoided for the preservation of ribavirin inhuman serum, whereas storage of serum samples at 4 or -20°Censures the good preservation of ribavirin for 4 days.

It has been suggested that only 27% of the pharmacokinetic variabilityin ribavirin clearance relies on the main cofactors influencingthis clearance, including body weight, gender, age, and serumcreatinine levels (8). Therefore, the bioavailability of ribavirinis probably controlled by many other, unknown parameters. Inthis context, a reliable and simple method for the quantificationof ribavirin in serum could be highly helpful for the managementof patients with HCV infection receiving combination treatment.Indeed, not only response rates but also the frequency of adverseevents are probably dependent on the interindividual variabilityof serum ribavirin concentrations. The monitoring of serum ribavirinconcentrations could improve the efficacy and the safety ofthe combination treatment. The minimum dose of ribavirin requiredfor efficacy has not been clinically established. Initial clinicaltrials have adopted a standard dose defined on the basis ofthe results of pilot studies (2, 4). The recommended dose ofribavirin was chosen to be 1,000 or 1,200 mg according to thepatient's weight (less than or more than 75 kg, respectively).A subsequent recommendation for adaptation of the dose accordingto body weight was made by the manufacturer (800, 1,000, and1,200 mg for body weights of <65, 65 to 85, and >85 kg,respectively). To our knowledge, these recommendations are basednot on direct measurements of ribavirin concentrations but onextrapolation of statistical relations among dose, body weight,and response status.

Thus, the analytical method for the quantification of ribavirinpresented in this paper could be a useful tool that can be usedto define the minimum efficient dose of ribavirin that takesinto account all the factors affecting pharmacokinetic variabilityinstead of body weight only. Subsequently, it will probablybe used to define new rules for dose adjustment for the managementof ribavirin-related adverse events.

The secondary objective of the study was to explore a possiblerelation between the bioavailability of ribavirin and the efficacyof combination therapy. Our study population was chosen in orderto represent the most likely group of patients who could benefitfrom monitoring of ribavirin concentrations: patients infectedwith HCV genotype 1, patients undergoing first treatment forcure, and patients undergoing treatment with the standard combinationof interferon and ribavirin. The response status of these patientswas defined as treatment success (patients with sustained virologicalresponses) or treatment failure (nonresponders and patientswho were responders but who then had relapses). Our data suggestthat high serum ribavirin levels could be one factor predictiveof long-term treatment success. A recent clinical trial withpegylated interferon supports our observation by showing thatthe doses of the two medications (ribavirin and interferon)received by the patient are significant predictors of a sustainedvirological response (P = 0.015 for ribavirin and P = 0.002for pegylated interferon) (11). Another study confirmed thatthe ribavirin concentration at treatment week 4 is positivelycorrelated to the initial virological response (a negative resultby an HCV-specific reverse transcription-PCR at week 24) (8).Our data suggest that this relation could be extended to thesustained virological response (a negative PCR result at week24 of posttreatment follow-up). In this context, monitoringof the ribavirin concentration at the time of treatment initiationcould help to adapt the ribavirin dose to the patient's pharmacologicalcharacteristics in order to improve efficacy in the case oflow bioavailability or to prevent severe adverse events in thecase of high bioavailability.

In conclusion, the present study describes a simple analyticalmethod for the quantification of ribavirin in human serum thatcould be a useful tool for the monitoring of ribavirin concentrationsin patients with chronic HCV infection in order to improve theefficacy and the safety of therapy with ribavirin plus interferon.

ACKNOWLEDGMENTS

This study was supported by a grant from Schering-Plough Laboratories.

FOOTNOTES

* Corresponding author. Mailing address: Laboratory of Pharmacology, Albert Michallon Hospital, BP217, 38043 Grenoble Cedex 09, France. Phone: 33 (0)4 76 76 54 92. Fax: 33 (0)4 76 75 89 38. E-mail: FStanke{at}chu-grenoble.fr.

REFERENCES

Top