Acta Tropica 107 (2008) 145–149 Contents lists available at ScienceDirect Acta Tropica journal homepage: www.elsevier.com/locate/actatropica The influence of food on the pharmacokinetics of piperaquine in healthy Vietnamese volunteers Trinh Ngoc Hai a,1 , Sofia Friberg Hietala b,1 , Nguyen Van Huong a , Michael Ashton b,∗ a Pharmaceutical Unit, Department of Malaria Treatment and Research, National Institute of Malariology, Parasitology and Entomology (NIMPE), BC10200 Tu liem, Hanoi, Viet Nam b Unit for Pharmacokinetics and Drug Metabolism, Institute of Neuroscience and Physiology, Sahlgrenska Academy at Gă oteborg University, Box 431, 405 30 Gă oteborg, Sweden a r t i c l e i n f o Article history: Received November 2007 Received in revised form 16 May 2008 Accepted 19 May 2008 Available online 24 May 2008 Keywords: Malaria Piperaquine Food–drug interactions Pharmacokinetics a b s t r a c t The combination piperaquine and dihydroartemisinin is emerging as first line treatment of uncomplicated falciparum malaria in Southeast Asia The aim of this study was to determine the influence of a standard Vietnamese meal on the single-dose pharmacokinetics of piperaquine when administered in combination with dihydroartemisinin, and to gain extended data on the terminal half-life of piperaquine in healthy Vietnamese volunteers Subjects were randomly assigned to take a single oral dose of piperaquine phosphate (640 mg) + dihydroartemisinin (80 mg) together with a standardized Vietnamese meal (n = 16) or to remain fasting for h following drug intake (n = 16) Frequent blood sampling was conducted during 36 h, followed by weekly samples for weeks The pharmacokinetic parameters of piperaquine were determined by noncompartmental analysis The median (80% central range) AUC0–last was 11.5 (6.9–17.3) h mg/L in fed and 13.9 (2.8–19.3) h mg/L in fasting subjects, indicating a considerable variability in exposure in both groups The estimated overall oral clearance was 0.27 (0.12–1.49) L/(h kg), the volume of distribution during the terminal elimination phase was 230 (102–419) L/kg and estimated terminal half-life was 18 (5–93) days This study did not demonstrate a significant impact of a standardized Vietnamese meal on the oral absorption of piperaquine © 2008 Elsevier B.V All rights reserved Introduction The combination piperaquine (PQ) phosphate and dihydroartemisinin (DHA), is first line treatment of uncomplicated falciparum malaria in Vietnam The current dose regimen in Vietnam is two tablets, each containing 320 mg PQ phosphate and 40 mg DHA, administered at 0, 6, 24 and 36 or 48 h The efficacy and safety of this PQ + DHA combination treatment has been established in a number of studies (Denis et al., 2002; Karunajeewa et al., 2004; Tran et al., 2004) The same total dose administered in a once daily regimen for days is also highly effective (Ashley et al., 2005; Tangpukdee et al., 2005; Mayxay et al., 2006; Smithuis et al., 2006; Hasugian et al., 2007; Ratcliff et al., 2007) Despite increasing clinical use, the pharmacokinetics of PQ have only recently been described, in healthy subjects as well as in adult and pediatric patients (Hung et al., 2004; Roshammar et al., ∗ Corresponding author Tel.: +46 31 786 34 12; fax: +46 31 786 32 84 E-mail address: Michael.Ashton@pharm.gu.se (M Ashton) These authors contributed equally to the work 0001-706X/$ – see front matter © 2008 Elsevier B.V All rights reserved doi:10.1016/j.actatropica.2008.05.013 2006; Ahmed et al., 2008; Karunajeewa et al., 2008; Tarning et al., 2008) The absorption of PQ has been shown to be discontinuous with plasma profiles exhibiting multiple peaks (Sim et al., 2005; Roshammar et al., 2006; Ahmed et al., 2008) PQ disposition is characterized by a multiphasic profile with an exceptionally slow terminal elimination during which the half-life may exceed month in the adult (Tarning et al., 2005) Average oral clearance (CL/F) values between 0.9 and 1.4 L/(h kg) and volume of distribution values between 103 and 716 L/kg have been reported (Hung et al., 2004; Roshammar et al., 2006; Ahmed et al., 2008; Tarning et al., 2008) PQ is a highly lipophilic base, with a water/octanol partition (log P) as high as 6.11 (Warhurst et al., 2007) Combining PQ with a high-fat meal resulted in a twofold increase of the AUC as well as a marked increase in maximum concentrations (Cmax ) (Sim et al., 2005) The fatty food intake did not appear to alter the time to reach Cmax , or influence the frequency of multiple peaks in the plasma concentration of PQ The clinical implication of an increased AUC due to food intake is unclear It has been proposed that an increased exposure due to better absorption with food may cause a greater risk for side effects rather than increase efficacy (Sim et al., 2005) 146 T.N Hai et al / Acta Tropica 107 (2008) 145–149 The aim of this study was to determine the influence of a standard Vietnamese meal on the pharmacokinetics of PQ when administered in combination with DHA, and to further explore the terminal elimination half-life of PQ in healthy Vietnamese volunteers variation (CV) in quality control samples was 13%, 14% and 14% and the intra-day accuracy, expressed as mean percentage of nominal value, was 110% (n = 4), 107% (n = 4) and 92% (n = 5) at 20, 198 and 1235 g/L, respectively The lower limit of quantification (LLOQ) was set at g/L with an intra-day CV of 14%, and an accuracy of 96% (n = 5) All concentrations are expressed in terms of PQ base Method 2.1 Subjects and study design 2.3 Pharmacokinetic and statistical analysis Thirty-two healthy Vietnamese adult subjects were included in the study The number of subjects was calculated to have an 80% power to detect a difference in mean AUCs equal to, or greater than, one standard deviation Subjects were included if they were healthy according to a physical examination including a resting electrocardiogram, blood pressure and heart rate measurement, standard biochemistry (total bilirubin, creatinine, AST, ALT, erythrocytes, hemoglobin and leucocytes) and medical history Subjects were excluded if they had a history of antimalarial intake within months prior to inclusion, had known allergies to the study drugs or excipients, were pregnant or planning to become pregnant within months Following an overnight fast subjects received two tablets of CV.Artecan (Pharmaceutical Company 26, Ho Chi Minh City, Vietnam), each containing 40 mg DHA and 320 mg PQ phosphate (equivalent of 171.5 mg of PQ base) as a single dose under supervision of study personnel Subjects were randomly assigned to take the study drugs together with a standardized Vietnamese meal (n = 16) or to remain fasting for another h following drug intake (n = 16) The meal consisted of one fried egg and a meat soup (pork, beef, rice, vegetables and beans) and contained approximately 482 kcal (17 g fat, 27 g protein and 53 g carbohydrates) This meal represents approximately one-fourth of the average daily caloric intake in the adult Vietnamese population (Thang and Popkin, 2004) Blood samples were obtained through an indwelling venous catheter at −5 (pre-dose), 0.5, 1, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 24, 28, 32 and 36 h and by venepuncture in the mornings on days 7, 14, 21, 28, 35, 42 and 49 Blood samples were drawn into heparinized Vacutainer® tubes which were inverted ten times by hand and centrifuged for at 3000 × g Plasma aliquots were transferred to plastic cryotubes (Nunc, Hereford, UK) and frozen at −70 C until transported in dry ice for analysis ă at Goteborg University The safety of the treatment was assessed with a questionnaire completed on days 1, 2, 7, 14, 21, 35 and 49 The study was conducted in accordance with the principles laid down in the Helsinki declaration and the standards established for Good Clinical Practice (GCP) at the Clinical Unit of the National Institute of Malariology, Parasitology and Entomology (NIMPE) in Hanoi, Vietnam Study approval was obtained from the Ethics and Human Research Committee at NIMPE and from the Ministry of Health, Vietnam The pharmacokinetic parameters of PQ were determined by noncompartmental analysis in WinNonlin version (Pharsight Corporation, California, USA) The area under the concentration–time curve (AUC) was calculated using linear interpolation between increasing concentrations and logarithmic interpolation between declining concentrations The AUC0–last was defined as the area under the concentration time curve from the time of dose until the last measurable concentration above LLOQ The AUClast–∞ was extrapolated from the predicted concentration at the time of the last concentration above the LLOQ (AUClast–∞ = Cpred /z ) The terminal elimination constant, z , was determined from the slope of on average five sampling points (range 3–7) The mean percentage of the AUC0–∞ extrapolated beyond the last sample above LLOQ was 37% and ranged from 7% to 90% The poor description of the AUClast–∞ was caused by the very slow terminal elimination in some subjects The median and the 80% central range of the pharmacokinetic parameters were calculated in Microsoft Excel (Microsoft Corporation, Washington, USA) The estimated pharmacokinetic parameters for the fed and the fasting state were compared using the Mann–Whitney two sample rank-sum test in SPSS 12.0.1 for Windows (SPSS Inc., Illinois, USA) 2.2 Chemical assay PQ plasma concentrations were determined using a previously described method (Lindegardh et al., 2005) In brief: PQ was separated from plasma components using solid-phase extraction (MPC solid-phase extraction 96-well plates, Millipore AB, Solna, Sweden) followed by concentration determination using HPLC (Chromolith ă Performance column, Chromtech AB, Hagersten, Sweden) with UV detection at 347 nm Detector response (peak heights) and nominal plasma concentrations were log-transformed to generate linear calibration curves (range 5–5000 g/L) The inter-day coefficient of Results The demographic profiles did not differ between the two study groups Median (range) age was 26 (19–45) and 31 (19–56) years and the body mass index was 21 (18–26) and 22 (17–25) kg/m2 in fed and fasting subjects, respectively The male:female ratio was 13:3 in fed and 14:2 in fasting subjects The mean (range) bodyweight normalized dose was 6.00 (4.90–7.46) and 6.02 (4.64–7.15) mg PQ base/kg in fed and fasting subjects, respectively All subjects complied with the study schedule and there were no adverse effects reported in either study group during clinic stay or during the 49 days of follow-up The median (80% central range) AUC0–last was 11.5 (6.9–17.3) h mg/L in fed and 13.9 (2.8–19.3) h mg/L in fasting subjects There was a considerable interindividual variability in exposure as shown by the approximately 20-fold range (max–min) in AUC0–last in both groups There was less variability in AUC0–24 , i.e early exposure, which had a range of and h mg/L in fed and fasting subjects, respectively Median (80% central range) AUC0–24 was 2.2 (1.4–3.6) h mg/L in fed and 1.7 (0.7–3.6) h mg/L in fasting subjects No statistically significant difference in exposure between fed and fasting subjects was observed (Table 1) As indicated in Fig 1, piperaquine elimination was marked by a very slow terminal phase in many subjects The estimated overall median (80% central range) CL/F was 0.27 (0.12–1.49) L/(h kg), corresponding values for Vz /F and t1/2z were 230 (102–419) L/kg and 18 (5–93) days Multiple peaks were a prominent feature of the PQ concentration–time profiles in both fasting and fed subjects, occurring in 26 out of the 32 individuals 147 T.N Hai et al / Acta Tropica 107 (2008) 145–149 Table Piperaquine pharmacokinetic parameters for fed and fasting healthy subjects after an oral single dose of piperaquine phosphate and dihydroartemisinin determined by noncompartmental analysis Parameter Fed (n = 16) Median Cmax (g/L) Tmax (h) AUC0–24 (h mg/L) AUC0–last (h mg/L) AUC0–∞ (h mg/L) CL/F (L/(h kg)) Vz /F (L/kg) t1/2z (day) 212 2.2 11.5 20.9 0.3 193 18 Fasting (n = 16) 80% central range 130–368 3–9 1.4–3.6 6.9–17.3 8.8–40.5 0.13–0.69 146–390 6–72 Discussion This study did not show a significant impact of a standardized Vietnamese meal on the pharmacokinetics of PQ The interindividual variability in exposure was considerable as shown by the 20-fold range in AUC0–last in both fed and fasting subjects The multiple peaks described in previous studies (Sim et al., 2005; Roshammar et al., 2006; Ahmed et al., 2008) occurred in both groups Given the long elimination half-life, a parallel group design was chosen as recommended by the FDA (CDER, 2003) The number of subjects required was calculated based on the assumption that a clinically relevant difference in AUCs would be greater than one standard deviation Our study showed a major interindividual difference in AUC0–last with CVs of 40% and 53% for fed and fasting subjects, respectively Thus the between-group difference would have had to be around 50% to be identified as statistically significant in this study Sim et al (2005) found a twofold increase in PQ exposure following intake with a high-fat meal They proposed that the risk of side effects may increase if PQ is administered with food Sim Median 130 1.7 13.9 23.1 0.27 262 20 Mann–Whitney 80% central range 50–407 1–9 0.7–3.6 2.8–19.3 3.9–49.5 0.13–1.61 93–483 5–101 P 0.22 0.54 0.13 0.36 0.78 0.87 0.29 0.98 et al (2005) used a high-fat (37 g) test meal as recommended by the FDA, designed to achieve the maximum effect on GI physiology and systemic drug availability In contrast, our study meal, as the normal Vietnamese diet, contained considerably less fat Our results indicate that there is no increased risk of adverse effects due to concomitant low-fat food intake during treatment with PQ There are three published studies describing the pharmacokinetics of PQ following a single oral dose in healthy subjects (Sim et al., 2005; Liu et al., 2007; Ahmed et al., 2008) The dose-normalized total exposure in fasting subjects (AUC0–∞ ) was approximately three times greater in our study compared to studies by Sim et al (2005) and Ahmed et al (2008) (based on geometric mean values) but only half of that described by Liu et al (2007) (arithmetic mean) In keeping with the differences in AUC0–∞ the estimated CL/F varies between studies, with ours being approximately one-third of the CL/F reported by Sim et al (2005) and Ahmed et al (2008) Liu et al (2007) report a mean CL/F of 0.02–0.03 L/(h kg) following a single dose of PQ However, using their reported mean AUC0–∞ a much Fig Piperaquine plasma concentrations over time following a single-oral dose of piperaquine phosphate and dihydroartemisinin in fed (panel A, n = 16) and fasting healthy individuals (panel B, n = 16) 148 T.N Hai et al / Acta Tropica 107 (2008) 145–149 higher CL/F of approximately 0.1–0.2 L/(h kg) can be calculated (CL/F = dose/AUC0–∞ ), suggesting some caution when evaluating these results Different drug products were used in these studies and it is possible that differences in drug formulation significantly influenced the bioavailability of PQ A relatively small change in the fraction of dose absorbed, from, e.g 10% to 20%, would result in a doubling of the AUC0–∞ and a halving of the CL/F Further, the ethnicities and gender distribution of the study populations differed The studies by Sim et al (2005) and Ahmed et al (2008) were both conducted in Caucasian volunteers Liu et al (2007) studied Chinese volunteers while the current study was conducted in a Vietnamese population Liu et al (2007) suggest a gender difference in pharmacokinetics, with a significantly higher AUC0–∞ in female subjects which appeared to hold true also for dose per body weight normalized AUCs However, the relevance of the reported normalizations for all other pharmacokinetic parameters, except Cmax , is unclear Sim et al (2005) had an equal but small number of male and female subjects (4/4), also in the current study, with only 5/32 subjects being female, the number was too small too justify a separate analysis Ahmed et al (2008) did not include any female subjects Values for CL/F (Table 1) indicate PQ to be a drug of low organ extraction (maximal extraction = (CL/F)/organ blood flow ≤ 0.3) In the rat, renal clearance of PQ was found to be negligible with a fraction excreted as unchanged drug less than 1% after intravenous administration (Tarning et al., 2007) Assuming that non-renal elimination predominates also in human beings the hepatic extraction can therefore be assumed to be about or less than 30% depending on the value of absolute bioavailability (F) The low organ extraction suggests the oral bioavailability of PQ to depend on solubility in the gastrointestinal tract rather than any major first-pass metabolism Together with a remarkably large volume of distribution, the low extraction results in an elimination half-life of substantial length Earlier studies have reported a long and variable elimination of PQ with mean half-lives of 11–33 days (Hung et al., 2004; Sim et al., 2005; Tarning et al., 2005; Roshammar et al., 2006; Ahmed et al., 2008) The present study confirms the presence of a prolonged elimination phase The estimated median half-life was 18 days Obviously, individual half-lives longer than the duration of sampling were estimated with poor accuracy Nevertheless, these PQ concentrations sustained on a 20–50 g/L level may contribute to a post-treatment prophylactic effect (Tarning et al., 2008) lasting several weeks Although there have been attempts to model the multiple peak kinetics of PQ (Roshammar et al., 2006), their presence lend non-compartmental analysis suitable for rich data Multiple peaks occurring within hours of drug intake could be the result of erratic dissolution and/or absorption regulated by gastric emptying since two tablets were administered It could also result from precipitation and slow re-dissolution as the lipophilic base passes from the acidic environment of the stomach to the more alkaline small intestine However, as indicated in Fig 1, PQ concentrations continued to oscillate during the whole study period The multiple peaks have been suggested to result from enterohepatic recirculation of PQ (Sim et al., 2005) However, results in the rat suggest that biliary excretion of PQ is quantitatively too low (