Pharmacokinetics and Comparative Bioavailability of Artesunate and Mefl oquine Administered Separately or as a Fixed Combination Product to Healthy Volunteers and Patients with Uncomplicated Plasmodium falciparum Malaria

1UNICEF/UNDP/World Bank/WHO Special Programme for Research & Training in Tropical Diseases (TDR), Geneva, Switzerland 2Centre for Drug Research, Universiti Sains Malaysia, Penang, Malaysia 3Clinical Epidemiology and Public Health Unit, Centre for Health Studies, CRP-Santé, Luxembourg 4School of Pharmacy & Medical Sciences, University of South Australia, Adelaide, Australia 5Hospital for Tropical Disease, Bangkok, Thailand 6Drugs for Neglected Diseases initiative (DNDi), Geneva, Switzerland 7Division of International and Humanitarian Medicine, Geneva, Switzerland §Deceased


Introduction
Malaria is a leading cause of mortality and morbidity in developing areas of the world. Every year, malaria is responsible for an estimated 250 million infections and results in approximately 1 million deaths, predominantly in children under 5 years of age (World Health Organisation, 2008).
Due to the increasing resistance to traditional anti-malarials in falciparum malaria endemic areas, the current World Health Organisation (WHO) recommendation for the treatment of uncomplicated falciparum malaria is with artemisinin-based combination therapy (ACT) (World Health Organization, 2006). These treatments generally combine rapid acting artemisinin or one of its derivatives (i.e. artesunate, artemether, dihydroartemisinin) with another more slowly eliminated anti-malarial (e.g. amodiaquine, mefloquine, piperaquine) (World Health Organization, 2006). The principle of artemisinin-based combination therapies is that the two schizontocidal drugs have independent modes of action, thereby increasing the therapeutic efficacy and delaying the development of resistance to the individual components of the combination treatment (World Health Organization, 2006). Artesunate and mefloquine combination treatments have been in use for several years, particularly on the Thai-Myanmar border. In this area of multidrug resistance, artesunate and mefloquine combination therapy has achieved consistently high efficacy rates and reduced malaria morbidity (Nosten et al., 1991).
Artesunate is rapidly absorbed and hydrolysed to its pharmacologically active metabolite, dihydroartemisinin (Navaratnam et al., 2000). Conversion of artesunate to dihydroartemisinin occurs chemically within the gastrointestinal tract and is further catalysed by tissue and blood esterases (Navaratnam et al., 2000). Whilst both artesunate and dihydroartemisinin share the critical peroxide bridge necessary for anti-malarial effect, due to artesunate's rapid conversion, it is thought that the anti-malarial efficacy of artesunate is primarily dependent on adequate systemic exposure to dihydroartemisinin. The in vitro-determined minimum inhibitory concentration (MIC) against P. falciparum of 0.1 -2.0 g/L for both compounds (de Vries and Dien, 1996).
Mefloquine is structurally related to quinine and is active against the erythrocytic stages of P. falciparum (Palmer et al., 1993). Based on in vivo and in vitro data of multi-drug resistant P. falciparum, pharmacokinetic-pharmacodynamic modelling suggests that the MIC of mefloquine is 500 g/L and the minimum parasiticidal concentration (MPC; i.e the minimum concentration to provide maximum parasite killing) is 600 g/L .
Artesunate and mefloquine combination therapy is generally available as separate 50 mg artesunate and 250 mg mefloquine tablets, individually packaged or co-blistered. Several treatment regimens have been examined with the optimal regimen, which has been adopted as standard, being 4 mg/kg artesunate administered daily for 3 days and a total mefloquine dose of 25 mg/kg World Health Organization, 2006). Mefloquine doses of 15 mg/kg and 25 mg/kg have been evaluated, and whilst the higher dose has been shown to be associated with poor absorption and drug-induced vomiting especially in young children and older adults, the lower dose has been associated with inferior efficacy and is therefore not recommended (Luxemburger et al., 1998;World Health Organization, 2006). Common practice, in order to improve both tolerability and absorption, is to administered 25 mg/ kg mefloquine as a split dose with an overall treatment regimen of: 4 mg/kg artesunate alone without mefloquine on Day 1; 4 mg/kg artesunate + 15 mg/kg mefloquine on Day 2, and; 4 mg/kg artesunate + 10 mg/kg mefloquine on Day 3 World Health Organization, 2006). Pharmacokinetic-pharmacodynamic modelling has demonstrated that with this split dose treatment regimen, whole blood mefloquine concentrations would remain above the MPC of 600 g/L for approximately 20 days as opposed to 6 days for a single dose of 15 mg/kg mefloquine . This prediction is supported by the findings of other studies Ramharter et al., 2007).
Although this treatment regimen for artesunate plus mefloquine has been proven to be efficacious, the regimen may be difficult to comply with outside of a research setting. Consequently, with the aim of simplifying the standard treatment regimen and improving patient adherence, an artesunate -mefloquine fixed dose oral co-formulation has been developed. The formulation consists of artesunate and mefloquine in a dose ratio of 1:2, with the intention of administration of daily doses of 4 mg/kg artesunate and 8 mg/ kg mefloquine for 3 days, corresponding to total doses of 12 mg/ kg and 24 mg/kg respectively, equivalent to the current total dose recommendations. This new fixed dose regimen has been shown to be as effective and tolerated as the standard regimen in a large field trial in Thailand (Ashley et al., 2006). The fixed dose co-formulation has been developed through a collaborative project coordinated by the Drugs for Neglected Diseases initiative (DNDi) and FarManguinhos, a publicly-owned technical/scientific unit of the Foundation Oswald Cruz (Fiocruz). The co-formulated product has been approved by the Brazilian regulatory agency and is used by the Brazilian malaria control programme.
In order to assess the pharmacokinetics and comparative bioavailabilities of artesunate and mefloquine administered as separate products and the developed fixed dose co-formulated product, two studies were conducted: a randomised, crossover design study in healthy volunteers and a randomised, parallel group study in patients with uncomplicated falciparum malaria.

Methods
The clinical studies were conducted at the Hospital for Tropical Disease, Mahidol University (Bangkok, Thailand) and were reviewed and approved by the institutional Human Research Ethics Committee. Participants were fully informed of the study procedures and provided written informed consent prior to study initiation. The studies were conducted in accordance with the study protocols, the Declaration of Helsinki and the principles of Good Clinical Practice.

Healthy volunteer study
The pharmacokinetics of artesunate and mefloquine administered as the Fixed and Non-Fixed formulations were assessed in 24 healthy, adult male and non-pregnant female volunteers. The study was conducted according to the randomised, cross-over design with a 90 day washout period between administrations of the study treatments. This clinical trial was registered as http://www.controlled-trials.com/ mrct/trial/673005/DNDi. Volunteers were screened and were required to meet the following eligibility criteria prior to treatment administration: age 18-50 years; negative pregnancy test for female volunteers; non-smoker (<10 cigarettes/day); no significant abnormal findings in the medical history or examination (including ECG, haematology, liver and renal function tests); no history of alcohol or drug abuse; negative malaria thick smear; no history of travel to or residency in a malarial area in the preceding 8 weeks; no history of anti-malarial medication ingestion within the preceding three months; no other medications (including over-the-counter products) within the preceding week; no known allergy to the study treatment; no concurrent clinical trial participation; adequate venous access; provision of written informed consent.
Participants were administered a single oral dose of artesunate and mefloquine as the

Uncomplicated falciparum malaria patient study
The pharmacokinetics of artesunate and mefloquine administered as the Fixed and Non-Fixed formulations were assessed in 50 male and non-pregnant female patients with slide-proven uncomplicated P. falciparum malaria. The study was conducted according to a randomised, parallel design with patients randomly allocated to receive multiple dosing with the Fixed or Non-Fixed formulation, according to a 1:1 ratio. This clinical trial was registered as http:// www.controlled-trials.com/mrct/trial/590967/DNDi.
Patients were screened and were required to meet the following eligibility criteria prior to treatment administration: age 18-65 years; not pregnant or lactating for female patients; minimum weight of 40 kg; microscopically confirmed mono-infection of P. falciparum; history or presence of fever; P. falciparum asexual stage parasitaemia <4% red blood cells; absence of clinical features of severe malaria and other significant illnesses or signs; no history of convulsions and/or psychiatric illnesses; no history of splenectomy; normal baseline ECG reading; no history of mefloquine ingestion within the preceding two months; no known contraindications or hypersensitivity to the study treatment; provision of written informed consent.
Participants allocated to the Non-Fixed formulation received daily administration of 4 mg/kg artesunate for three days as well as 15 mg/ kg mefloquine on the second day of artesunate dosing and 10 mg/kg mefloquine on the third day of artesunate dosing. Blood samples were collected for analysis of plasma artesunate and dihydroartemisinin concentrations prior to dosing (0 hr) and 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 8, 12 and 24 hours after treatment administration on the first day of dosing (Day 1). Blood samples were collected for analysis of plasma mefloquine concentrations prior to dosing on Day 2 (24 hr) and 48, 72, 120, 168, 336, 504 and 672 hours after treatment administration on Day 1. The blood samples scheduled for collection at the same time as dosing on Day 2 (24 hr) and Day 3 (48 hr) were collected prior to treatment administration.
The analysis of artesunate and dihydroartemisinin was conducted using BAS 200A HPLC-EC systems operating at reductive mode. Chromatographic separation of artesunate, dihydroartemisinin and the internal standard (artemisinin, QHS) was obtained at room temperature (25°C) on a Hypersil C4 column, 250 x 4.6 mm ID, and 5 μm particle size (Thermo Hypersil-Keystone, Bellefonte, USA). The mobile phase consisted of acetonitrile and 0.05 M acetic acid (42:58 % v/v) adjusted to pH 5.2 with 0.1 M NaOH. The flow rate was 1.5 mL/min. Detection was performed in the reductive mode at -1000 mV in an oxygen free environment. The HPLC-EC (reductive) system produced a linear response for the concentration of artesunate and dihydroartemisinin in the range of 10 ng/0.5 mL -800 ng/0.5 mL. The extraction recoveries of artesunate, dihydroartemisinin and QHS were above 86% with a coefficient of variation 13%. The lower limit of quantification for artesunate was 10 ng/0.5 mL with an accuracy of 109.9% and a coefficient of variation of 7.2%. The lower limit of quantification for dihydroartemisinin was 10 ng/0.5 mL with an accuracy of 99.1% and a coefficient of variation of 5.9%. The within-day and day-to-day precision for both artesunate and dihydroartemisinin were less than 8%.
The analysis of mefloquine was conducted using Waters 2487 (Dual Absorbance Detector) HPLC-UV systems coupled to Waters 515 HPLC pump. The analysis of mefloquine as internal standard was carried out at room temperature (25°C) using a Inertsil C8-3 column, 150 x 4.6 mm ID, 5 m particle size (GL Sciences Inc, Tokyo, Japan). The mobile phase consisted of acetonitrile, methanol and 0.05 M KH 2 PO 4 (80:500:420 % v/v) adjusted to pH 3.9 with orthophosphoric acid. The flow rate was 1.00 mL/min. Detection was performed in the UV mode at 285 nm. The extraction recoveries of mefloquine were all above 80% with a coefficient of variation less than 8%. The HPLC-UV system produced a linear response for the concentration of mefloquine in the range of 10 ng/0.5 mL -800 ng/0.5 mL. The lower limit of quantification of mefloquine was 25 ng/0.5 mL with an accuracy of 98.9% and a coefficient of variation of 8%. The method was found to be accurate with an average deviation of <5.2% from the true value. The within-day and day-to-day precision for mefloquine were less than 4%.

Pharmacokinetic and statistical analysis
Plasma analyte concentrations were utilised for calculation of pharmacokinetic parameters using a standard model-independent approach. Area under the plasma concentration time curve from time zero to the last quantifiable concentration (AUC last ) was calculated using the linear-trapezoidal method. Area under the plasma concentration time profile from zero extrapolated to infinite time (AUC inf ) was calculated as AUC last + last quantifiable concentration / terminal rate constant. The maximum observed concentration (C max ) and the time of the maximum observed concentration (T max ) were taken directly from the data without interpolation.
A linear mixed-effects analysis of variance (ANOVA) model was used to analyse dose-normalised (where appropriate), Ln-transformed AUC last , AUC inf and C max parameters. The model incorporated factors for treatment, period, sequence and subject nested in sequence for the healthy volunteer cross-over study and treatment and subject for the uncomplicated falciparum malaria patient parallel study. The residual error (error mean square) was used to construct the 90% confidence intervals for the ratio of the Non-Fixed and Fixed formulation means. In constructing these 90% confidence intervals, the Non-Fixed formulation was used as the reference treatment.
A non-parametric Wilcoxon Signed Ranks test was used to assess treatment differences for paired T max data obtained in the healthy volunteer study. A Mann-Whitney U test was used to assess treatment differences for unpaired T max data obtained in the uncomplicated falciparum malaria patient study.
Bioequivalence was concluded if the 90% confidence intervals were within the regulatory limits of 80 -125%. Significance was set at an -level of 0.05. WinNonlin Professional, Version 5.2 (Pharsight Corporation, Mountain View, CA, USA) was used for the pharmacokinetic and parametric statistical analyses. SPSS for Windows, Version 16.0 (SPSS Inc, Chicago, IL, USA) was used for non-parametric statistical analysis.

Healthy volunteer study
Twenty-four healthy volunteers completed the study. Participant demographic information is summarised in Table 1.
Plasma concentration-time profiles for artesunate, dihydroartemisinin and mefloquine are displayed in Figure 1, Figure 2 and Figure  3 respectively. Pharmacokinetic parameters are summarised in Table  2 and the ratio of Fixed/Non-Fixed data and associated 90% confidence intervals for the pharmacokinetic parameters are summarised in Figure 4.
The results indicate substantially lower artesunate and dihydroartemisinin concentrations after administration of the Fixed formulation compared to the Non-Fixed formulation (Figure 1 and Figure 2). Bioequivalence analysis of artesunate pharmacokinetic data indicated that the two formulations were not bioequivalent with significantly lower AUC last (p=0.0479) and C max (p=0.0006) values for the Fixed formulation and the 90% confidence intervals extending below the 80 -125% limits (Figure 4). Due to limited data AUC inf could not be calculated for artesunate. Similarly, the two formulations were not bioequivalent with respect to dihydroartemisinin pharmacokinetics with significantly lower C max values (p=0.0455) and the 90% confidence intervals for AUC last , AUC inf and C max outside of the 80 -125% limits ( Figure 4).
Statistical analysis of dose-normalised mefloquine data indicated that there were no significant differences between the Fixed and Non-Fixed formulations for AUC last , AUC inf and C max . In addition, the 90% confidence intervals for dose-normalised data were within the 80 -125% limits (Figure 4), indicating that the formulations were bioequivalent with respect to both the rate and extent of mefloquine absorption. A highly significant period effect was found for AUC last , AUC inf and C max (p<0.0001) with area-under-the-curve values approximately 1.5-fold higher and C max values approximately 1.8-fold higher in Period 2, irrespective of treatment sequence.

Uncomplicated falciparum malaria patient study
Forty-four patients with confirmed P. falciparum malaria completed the study, of whom 21 were administered the Fixed formulation and 23 were administered the Non-Fixed formulation. Participant demographic information is summarised in Table 1.      7 respectively. Pharmacokinetic parameters are summarised in Table  3 and the ratio of Fixed/Non-Fixed data and associated 90% confidence intervals for the pharmacokinetic parameters are summarised in Figure 8.

Non-Fixed Formulation
The results reveal significantly lower dose-normalised maximum artesunate and dihydroartemisinin concentrations (p=0.0192 and p=0.0113 respectively) after administration of a single oral dose of      (Figure 8), indicating that the formulations were bioequivalent with respect to both the rate and extent of absorption.

Discussion
The fixed-dose co-formulation was bioequivalent to the separately administered products for mefloquine in healthy volunteers and uncomplicated falciparum malaria patients but not for artesunate or dihydroartemisinin pharmacokinetics in either study population. Administration of the fixed formulation resulted in 15 -25% lower AUC last and AUC inf values and 25 -40% lower C max values for dihydroartemisinin than those observed after administration of the separate products. This reduction in the exposure to dihydroartemisinin with the fixed dose co-formulated product is most likely reflective of a less complete dissolution of the Fixed formulation for the clinical batch used (data not shown). Whilst decreased exposure may also be indicative of a reduction in the conversion of artesunate to dihydroartemisinin, the ratio of artesunate to dihydroartemisinin concentrations is the same for the two study treatments, in both the healthy volunteer and patient populations, and thus it can be concluded that the conversion of artesunate to dihydroartemisinin has not changed. Alternatively, lower dihydroartemisinin concentrations may also be indicative of a higher rate of elimination of dihydroartemisinin, perhaps due to the presence of co-administered mefloquine; however, artesunate was administered both with and without mefloquine in the patient study and the half-lives of dihydroartemisinin in both cases were similar (data not presented) thereby suggesting that the elimination of dihydroartemisinin remains unchanged.
The clinical relevance of these differences in the exposure to dihydroartemisinin is uncertain. For artesunate and mefloquine combination therapy, the efficacy is dependent upon the initial reduction in parasite biomass by artesunate and dihydroartemisinin and the killing of residual parasites by mefloquine. Due to their very short half-lives, the maximum plasma concentration of artesunate the Fixed formulation compared to the Non-Fixed formulation ( Figure  5 and Figure 6). There were no significant differences between the treatments with respect to area-under-the-curve values for artesunate and dihydroartemisinin. Bioequivalence analysis of dose-normalised artesunate and dihydroartemisinin pharmacokinetic data indicated that the two formulations were not bioequivalent with the 90% confidence intervals below the 80 -125% limits in all cases (Figure 8).
Statistical analysis of dose-normalised mefloquine data indicated that there were no significant differences between the Fixed and Non-Fixed formulations for AUC last , AUC inf and C max . In addition, the 90% confidence intervals for dose-normalised data were within the   and dihydroartemisinin is probably more important than the extent of absorption; for the longer acting mefloquine, the time that mefloquine concentrations remain above the MPC and MIC, hence extent of absorption, is important.
The plasma levels of artesunate and dihydroartemisinin observed in both healthy volunteers and malaria patients in this study are manyfold higher than that required for schizontocidal activity. Mean C max values for dihydroartemisinin observed in patients were 1233 g/L and 2043 g/L for the Fixed and Non-Fixed formulations, respectively. This compared to dihydroartemisinin the in vitro MIC values against P. falciparum of 0.1 -2.0 g/L (de Vries and Dien, 1996). Artemisinin resistance is now reported on the Thai-Cambodian border and the IC 50 s of the strains collected from this area for dihydroartemisinin are 2.3 g/L (IQR 1.1 -3.2 g/L) (Dondorp et al., 2009). In that respect, a limitation of this study is that it was designed to determine the bioavailability and bioequivalence of the new fixed formulation, and did not allow for the measurement of the plasma levels of artesunate and dihydroartemisinin after the second and third dose of artesunate in patients.
Mean mefloquine concentrations in malaria patients remained above the MIC of 500 g/L throughout the assessment period, with mean plasma concentrations at the last study time-point (672 hours = 28 days) approximately 560 g/L for both study treatments. This Interestingly, a highly significant period effect was observed for mefloquine in the healthy volunteer study with a 1.5-fold increase in AUC last and AUC inf values and a 1.8-fold increase in C max values after treatment administration in the second study period. Apparent total systemic clearance and volume of distribution were approximately 40 50% lower in Period 2; half-life was unchanged (data not presented). Whilst, due to the longer half-life of mefloquine, a carry-over effect is likely, this is estimated to account for only a 1% increase in values and does not explain the substantial increase in the rate and extent of absorption observed in the second treatment period. The underlying cause of this is unknown and warrants further investigation.
Both the loose and the fixed combination proved effective in curing malaria this small (results reported elsewhere: Krudsood et al., submitted) and in a larger clinical trial (Ashley et al., 2006). Other studies are underway (Myanmar NCT00902811, India ISRCTN70618692).
While there are obvious differences in the disposition of both antimalarial drugs between healthy volunteers and patients, the experimental design did not permit a direct comparison of noncompartmental parameters between healthy volunteers and patients. We are currently using a population pharmacokinetic approach to interrogate these differences.

Conclusion
This study has demonstration that the developed fixed dose co-formulated artesunate and mefloquine product is equivalent to the current standard treatment regimen in terms of mefloquine pharmacokinetics in both healthy volunteers and malaria patients but not in terms of artesunate and dihydroartemisinin concentrations. While the plasma concentrations of artesunate and dihydroartemisinin were adequate to achieve effective antimalarial activity, the results of this study underscore the importance of conducting bio-equivalence evaluation of new antimalarial products.