Yuriy V. Pidpruzhnykov, a* Valerii E. Sabko,a Volodymyr V. Iurchenkoa and Igor A. Zupanets
Meldonium [3-(2,2,2-trimethylhydrazinium) propionate dihydrate, inner salt] is an effective drug to treat ischemia, stenocardia, heart attack and some other diseases. Studies of the mechanism of action of meldonium and its clinical effectiveness have been described elsewhere (Sjakstea and Kalvinsh, 2006; Dambrova and Kalvinsh, 2002). The brand named drug of meldonium, MildronateW, produced by JSC Grindeks, Latvia, has been approved and is widely used in medical practice in CIS countries. In these countries, mortality caused by heart/vascular diseases has been reported to be more than 50% according to Gorbas et al. (2010).
The aim of this work was to run a bioequivalence study of the generic meldonium drug, VasonatW (solid gelatinous capsules containing 250 mg of active compound) produced by JSC Olainfarm, Latvia, and the brand named one, MildronateW, produced by JSC Grindeks, Latvia. The pharmacokinetic studу of oral meldonium was conducted in 24 healthy volunteers. The pharmacokinetic profiles of orally administered meldonium were obtained for the first time. In previous publications (Zupanets et al., 2009), clinical aspects of the study have been described, and recommendations for the administration and monitoring of arterial blood pressure during treatment have been made. However, the development, validation and detailed description of the analytical methods used in this study have not been previously reported.
Meldonium has been classified as a hydrazinium base of low molecular weight, and its determination is rather difficult. HPLC-MS/MS is the most suitable method for this, as described in Lv et al. (2007) and Peng et al. (2010). The sample preparation procedure described in Lv et al. (2007) is difficult because of sample evaporation and reconstitution. Moreover, the use of L-carnitine as an internal standard is rather doubtful because of the high content of endogenous L-carnitine in plasma. The chromatographic conditions using a C18 column chosen by Peng et al. (2010) are questionable, since the highly polar analyte is not retained on a reversed-phase column and is eluted with the dead volume. To ensure the retention on a reversed-phase column for the control of impurities in meldonium, an ionpairing reagent was used (The State Pharmacopoeia of the Russian Federation, 2008). The ion-paired agents frequently affect the MS/MS detection sensitivity. Hydrophilic interaction chromatography (HILIC) is most effective for highly polar substances, in contrast to reversed-phase columns. Hmelnickis et al. (2008) describe the successful application of HILIC column for quantification of impurities in meldonium substance. The advantages of UPLC-MS/MS for pharmacokinetic studies as compared with traditional HPLC are described in Li et al. (2008).
This article presents the new UPLC-MS/MS method for meldonium quantification in plasma on a HILIC column. The developed method satisfies the regulatory requirements regarding the selectivity, matrix effect, accuracy, precision and other parameters (FDA/CDER, 2001; EMEA/CHMP, 2009). The advantage consists of simple sample preparation and short analysis time. The method has been successfully applied in a bioequivalence study of oral meldonium drugs.
Reagents, solvents and materials
Trimethylhydrazinium propionate (TMHP) reference standard (purity 100.1%), was provided by JSC Olainfarm, Latvia. The reference drug, MildronateW (capsules 250 mg, batch no. 031207, produced by JSC Grindeks, Latvia), and the test drug, VasonatW (capsules 250 mg, batch no. 10508 produced by JSC Olainfarm, Latvia), were provided by JSC Olainfarm, Latvia. The following substances were studied as candidates for the internal standard (IS): glycine betaine hydrochloride (BG; purity 100%, Sigma), L-carnitine (LC; purity 99%, Sigma) and thriethanolamine (TEtA; purity 98%, Merck). Chemical structures of TMHP, LC, BG and TEtA are shown in Fig. 1. Methanol and acetonitrile of HPLC grade were obtained from LAB-SCAN. Formic acid (Puriss. p.a.) was obtained from Riedel-de-Haen. Water was purified using Elix-35 (Millipore, USA), with final purification using Milli-Q Gradient (Millipore, USA). All other chemicals were of analytical grade. Drug-free human plasma from healthy volunteers was kindly provided by JSC Biopharma, Ukraine; each batch was accompanied by a certificate of quality and stored at °30 °C prior to use.
The LC-MS/MS analyses were performed using a Waters Acquity UPLCW system with a Waters Acquity UPLCW BEH HILIC column (50 x 2.1 mm, 1.7 μm; both Milford, MA, USA) in combination with a Quattro Micro API triple quadrupole mass spectrometer (Micromass, Manchester, UK) equipped with an electrospray interface. MassLynx 4.1 software was used for instrument control and data processing. Chromatography was performed at 40 C at a flow rate of 0.3 mL/min with a run time of 1.4 min. The mobile phase consisted of water, acetonitrile and 200 mM formic acid (adjusted to pH 3.0 with 12.5% ammonia hydroxide) at a ratio of 25:70:5 (v/v/v). The autosampler was kept at 15 C and 5 μL samples were injected. Data acquisition was performed in the positive-ion mode. The capillary voltage was set at 300 V, the source temperature was set at 150 °C and the desolvation temperature was set at 500 °C. The desolvation gas flow (nitrogen) was set at 700 L/h; the collision gas pressure (argon) was set at 2.8 μbar. Detection of the ions was performed in the multiple reaction monitoring mode (MRM), monitoring the ion combinations of m/z 147.15 ! 58.1, 150.2 ! 69.9, 118.09 ! 58.6 and 162.18 ! 59.9 for TMHP, TEtA, BG and LC respectively. The collision energy was set at 12, 12, 18 and 22 V and the cone voltage was set at 20, 15, 25 and 15 V for TMHP, TEtA, BG and LC, respectively. The MRM parameters were determined by means of the autotune option. Massspectra of TMHP, TEtA and their products are presented in Fig. 2.
The stock solutions of TMHP, BG and LC (200 μg/mL) were prepared in methanol–water (1:4; v/v). The stock solution of TEtA (400 μg/mL) was prepared in the mobile phase. Appropriate dilutions were made in the mobile phase for TMHP to produce working solutions of 0.1, 0.2, 0.5, 1, 2.5, 5, 10, 15, 30 and 60 μg/mL. These solutions were used to prepare a calibration curve. Other sets of working solutions of TMHP were made in the mobile phase for preparation of quality control (QC) samples. Stock solution of TEtA was used as a working IS solution. All solutions were stored at approximately 5 C for a month. Calibration samples and QC samples were prepared by spiking 100 μL of blank human plasma with the appropriate amount of the analytes (10 μL) and IS (10 μL).
IS stock solution (10 μL) equivalent to 40 μg TEtA and methanol (10 μL) were added to 100 μL of plasma sample, and vortexed. After the addition of 900 μL of methanol, the mixture was vortexed again following by centrifugation for 15 min at 20,000 rpm on a benchtop centrifuge (Sigma 3K30C). Supernatant (200 μL) was transferred to the glass autosampler vials containing 600 μL of the mobile phase. The vials were capped and vortexed, and 5 μL was injected onto the UPLC-MS/MS system.
A full validation of the developed method was carried out in accordance with the requirements of guidance (FDA/CDER, 2001). The method was validated for selectivity, lower limit of quantification (LLOQ), calibration curve, within-run and between-run precision and accuracy, recovery, procedure of the samples dilution and stability.
Three batches of blank plasma were used to determine the selectivity. TMHP was added to plasma in a quantity that corresponded to LLOQ (10 ng/mL); the IS stock solution was added to achieve the 40 μg/mL concentration in plasma. The given solutions as well as the solutions of blank plasma prepared as described above were analyzed, and the peaks areas from the corresponding chromatograms were compared.
The LLOQ was assessed using five solutions prepared from the blank plasma with the nominal concentration of 10 ng/mL TMHP in plasma. We analyzed the given solutions and the samples of blank plasma, which were prepared according to the ‘Sample preparation’ section.
The linearity was evaluated using 10 calibration samples with TMHP concentrations in plasma of 10, 20, 50, 100, 250, 500, 1000, 1500, 3000 and 6000 ng/mL. Each solution contained in plasma 40 μg/mL of IS.
The accuracy and the precision of the method were evaluated using the model biosamples, which were prepared in the group of five for each of three TMPH concentrations in blank plasma equivalent: 30, 2000 and 4500 ng/mL. The experiment was conducted on three separate occasions. Within-run and between-run precisions were calculated using the formula %RSD = (SD/M) x 100, where M is the mean of the experimentally determined concentrations and SD is the standard deviation of M. Accuracy was defined as the percent relative error (%RE) and was calculated using the formula %RE = (E - T) x 100/T, where E is the experimentally determined concentration and T is the theoretical concentration.
The extraction recovery of TMHP from plasma was determined at five concentrations (three replicates) and for the IS at the concentration used in the assay (40000 ng/mL) by comparing the areas of extracted samples with blank plasma extracts fortified with drug post extraction. The matrix effect was investigated using six batches of human plasma. For the analyte and the IS, the matrix factor (MF) was calculated for each batch of plasma by calculating the ratio of the peak area in the presence of matrix (measured by analysing blank matrix spiked with analyte at a concentration of 30 ng/mL after extraction (3 LLOQ) to the peak area in the mobile phase (EMEA/CHMP, 2009).
The validity of the 10-fold dilution procedure of model biosample with blank plasma was assessed by accuracy and precision parameters. Three different concentrations of THMP (2, 6 and 12 μg/mL for plasma equivalent) were evaluated by analyzing five aliquots of each concentration.
In accordance with the regulatory requirements, stability testing was performed (FDA/CDER, 2001). All stability samples were prepared from blank plasma with the addition of the corresponding quantity of TMHP solution in the mobile phase. The short-term temperature stability for two aliquots at nominal concentrations of 30 and 8000 ng/mL was examined following sample defrosting and 4 h incubation at room temperature. The freeze–thaw stability was examined for two aliquots at nominal concentrations of 30 and 8000 ng/mL after three freeze–thaw cycles. The long-term stability was determined by analyzing three aliquots at nominal concentrations of 30 and 1200 ng/mL and 12 μg/mL stored at -70 °C for 55 days. The stability of processed samples in the autosampler for three aliquots at nominal concentrations of 30, 1200 and 4500 ng/mL was examined after 18 h. Stock solution stability was assessed after 1 month of storage at a temperature of 2–7 °C.
The bioequivalence study for oral formulation of TMHP (250 mg capsules) was conducted in accordance with the Declaration of Helsinki, Good Clinical Practice and Good Laboratory Practice requirements. The protocol and associated informed consent forms were reviewed and approved by Central Ethical Committee of Ministry of Healthcare of Ukraine and Local Ethical Committee (license no. 852 issued 29 December 2006), and the informed consent forms were signed by the volunteers. Twenty-four healthy volunteers (aged 18–45) were enrolled in this study. The volunteers met the requirements of the inclusion/exclusion criteria. They had no history of cardiovascular, hepatic, renal, gastrointestinal, hematologic or nervous disease, or any acute or chronic diseases or drug allergy, and had stopped using any drugs 2 weeks prior to study enrollment. Physical examination and laboratory tests showed no abnormal findings. The volunteers were randomized into two groups of 12. The crossover design was applied to this study when one subgroup was administered the study drug followed by administration of the reference drug; another subgroup was administered the reference drug followed by administration of the study drug. All subjects were administered a single dose of 250 mg after overnight fasting. The subjects were required to refrain from smoking, alcohol and caffeine, and were under direct medical supervision at the study site.
Fifteen blood samples were drawn from each subject at each study time point. The blood sampling was conducted according to the following schedule: 0 h (pre-dose), 15, 30 and 45 min, and 1, 1.5, 2, 2.5, 3, 4, 5, 6, 8, 12 and 24 h post-dose. The samples were transferred into sodium heparinized tubes (6 mL volume) and centrifuged within 1 min after collection. The centrifugation was conducted in a temperature-stabilized rotary chamber at 10 C for 5 min at 3000 rpm. The plasma was divided into two aliquots (the first one was for the analysis, the second for back-up), but the volume was not less than 1.3 mL for each aliquot, and transferred into the pre-labeled cryotubes. Plasma was kept at -70 °C at the study site and at the bioanalytical laboratory as well. The time period between the blood collection and the plasma transfer into a -70 °C freezer did not exceed 15 min. In total, 720 frozen samples were submitted to the bioanalytical laboratory for assessment.
The data analysis of pharmacokinetic parameters was performed using Win-Nonlin professional software version 5.2 (Pharsight, Mountain View, CA, USA) using a noncompartmental approach to the model of assessment of NCA Model 200. Area under the plasma concentration curve (AUC0–24) values (t being the time of the last plasma concentration measured) were calculated by the linear trapezoidal rule. The AUC from 0 to infinity (AUC0–1) was calculated by the software. The first-order rate constant, Kel, was estimated by linear regression of time vs log of the concentration. Theerminal half-life (T1/2) was also calculated.
The bioequivalence was confirmed by the approach based on the statistically assessed 90% confidence intervals for the ratio of the average maximum plasma concentration (Cmax) and AUC0–t parameters for the reference and test drugs. The drugs were considered as bioequivalent if the 90% confidence interval was within 0.80–1.25 (80–125%) for the ratio of the average Cmax and AUC0–t parameters.
Selection of IS
One of the main requirements for the IS is its absence in the biomatrix on an endogenous level. The following substances were investigated as possible IS for the TMHP determination: LC, BG and TEtA. Chromatograms of blank plasma (top parts) and plasma spiked with the substances under investigation (bottom parts) are shown in Fig. 3. As is apparent from the figure, the blank plasma contains large amounts of LC and BG. The content of endogenous LC in blood plasma was assessed using a calibration curve within the LC range from 0.01 to 10 μg/mL. The acquired results showed that plasma contained LC at the level of 10 μg/mL, which agrees with Li et al. (2007). According to our estimates, the endogenous content of BG is comparable with the content of LC. Since high leveld of LC and BG in plasma were determined, these substances could not be used as IS. As TEtA was not detected in plasma, we used this compound as the IS for this study.
Samples preparation and chromatographic conditions
Acetonitrile and methanol were tested as the agents for plasma proteins precipitation and methanol showed better results. For recovery assessment, the following plasma–methanol ratios were tested: 1:5, 1:7 and 1:9. The recovery value approached 100% for the plasma–methanol ratios of 1:7 and 1:9. The ratio 1:9 was used for the further sample preparation.
The Acquity UPLCW BEH HILIC column is specially designed for polar compounds with low molecular weight, such as those containing quaternary nitrogen atoms. As expected, the allowable retention of the hydrophilic TMPH occurred on the mobile phase containing organic solvent at a rather high level (70% of acetonitrile). The analyte peak was quite symmetric, which made it possible to conduct the chromatography under the isocratic condition. The time of data registration was 1.2 min. Two minutes were required for one chromatogram, taking into account the time for sampling. Such a short period of time allowed the chromatography of each study sample to be performed twice. The TMPH concentration was calculated for the mean value of two parallel injection responses. Significant time reduction on chromatography by the UPLC-MS/MS method is an advantage of the new approach as compared with the traditional HPLC-MS/ MS. This is essential for conducting a bioequivalence study. In the given study, 720 biosamples were analyzed, making up 12 analytical runs. Ten calibration standards were prepared, six QC samples, one solution of each matrix sample processed with IS (Z-sample) and blank sample, thereby the total number of the samples for the routine analysis was 936.
Matrix effect, selectivity and recovery
The matrix effect study showed that the ion suppression of TMHP is practically absent under the selected conditions. MFTMHP and MFTEtA calculated from the six batches of matrix showed high stability. The IS normalized MF (MFN) was also calculated by dividing the MF of the analyte by the MF of the IS (FDA/CDER, 2001). The results obtained satisfy the requirements (EMEA/CHMP, 2009), as the RSD of MFN calculated from the six batches of matrix (11.0%) is ≤15%.
From each of three batches of plasma for selectivity evaluation two blank samples were prepared and also two model samples that contained 10 ng/mL of TMHP concentration and 40 μg/mL of IS. For each of the three batches of plasma, the mean ‘noise’ peak area was less than 20% of the lower limit of quantification for the analyte – 6.8, 10.1 and 4.5% respectively – which satisfies the acceptance criteria for selectivity (EMEA/ CHMP, 2009). Selectivity was verified during routine analysis.
Calibration curve and lower limit of quantification
The calibration curve was generated from 10 calibration standards prepared from blank plasma over the TMPH concentration range of 10–6000 ng/mL. The results of the blank sample and Z-sample were not taken into consideration when constructing the calibration function.
First we generated a calibration graph using our data in the linear form with a weight multiplier (as for the solution on the biomatrix, so for the mobile phase) and obtained results that did not satisfy the acceptance criteria (FDA/CDER, 2001).
At the same time, a quadratic model with the weight function 1/x (i.e. 1/concentration) more accurately described the calibration curve. The coefficients of the regression equation of calibration curve y = ax2 + bx + c had the following values:
The back-calculated concentrations of the calibration standards complied with the acceptance criteria (FDA/CDER, 2001). The quadratic form of the chosen calibration graph was continuous and reproducible for both the process of validation and routine analysis. The mean value of the coefficient of determination (R2) was 0.993. The validated value of LLOQ was 10 ng/mL. Eight independent model biosamples of the mentioned concentration and eight blank biosamples were analyzed in order to obtain this value. The mean signal-to-noise ratio was 15.4. The precision of TMHP quantification was 6.9% and the accuracy was 11.7%, which complies with the requirements (FDA/CDER, 2001).
Quintuple determinations were conducted in the model biosamples for each of three chosen concentrations and showed that accuracy and also within-run and between-run precision were better (Table 1) than prescribed by the acceptance criteria (FDA/CDER, 2001).
Since the pharmacokinetic studies of the oral drugs of TMHP were conducted for the first time and the TMHP concentration range was unknown, we performed validation of the 10-fold dilution of biosamples with blank plasma. Validation was performed according to precision and accuracy parameters. The obtained results (RSD < 4.9%, |RE| < 8.3%) for all concentration ranges confirmed the validity of the dilution procedure.
The obtained stability results (|RE| < 11.8% for all kinds of stability and concentration ranges) indicate that data comply with the acceptance criteria (FDA/CDER, 2001). It was established that the model biosamples were stable for 55 days of storage at the temperature not higher than -70 °C. Storage time of the real biosamples did not exceed 40 days.
Figure 4 shows representative chromatograms of plasma sample after administration of TMHP and spiked IS. Mean plasma concentration–time profiles for 24 volunteers after oral administration of the test and the reference drug are presented in Fig. 5. To the best of our knowledge, such profiles have been obtained for the first time.
Vasonat in capsules (250 mg) produced by JSC Olainfarm, Latvia had a similar pharmacokinetic profile to MildronatW in capsules (250 mg) produced by JSC Grindeks, Latvia. The pharmacokinetic curves of both drugs (VasonatW and MildronateW) had a statistically significant double peak: the first peak was at the time at which Cmax was reached (Tmax), about 1.5 h, and the second was at a second Tmax of about 5 h. The double peak phenomenon for TMHP was observed for the first time. This phenomenon has been described previously for several drugs, and various probable reasons of this phenomenon have been considered (Davies et al., 2010). We carried out mathematic modeling of TMHP absorption in different segments of the gastrointestinal tract in order to explain the double peak phenomenon, which is observed after oral administration of the drug, but not after its intravenous injection (Peng et al., 2010). The results of such modeling after oral administration of the drug were described by Zupanets et al. (2010), where it was shown that such TMHP behavior could be associated with two areas of its absorption in gastrointestinal tract segments: the duodenum and the empty gut.
Statistical analysis of all acquired data was performed employing the ANOVA algorithm (Win-Nonlin professional software version 5.2). The main pharmacokinetic parameters and its standard deviations are listed in Table 2. The parametric 90% confidence intervals for geometric mean ratio meldonium test/reference drug ranged from 87.7 to 99.2% (point estimate 93.0%) for Cmax and from 95.5 to 102.1% (point estimate 98.9%) for AUC0–24. These findings suggest that the drugs are bioequivalent.
A new method of quantification of meldonium in blood plasma using UPLC-MS/MS and a HILIC column was developed. The method proved to be rapid and simple. The method validation indicated the conformity of its properties with the established requirements. The method was successfully applied to the pharmacokinetic study of generic and branded oral drugs of meldonium (250 mg, in capsules), which was carried out for the first time. It was determined that the mean pharmacokinetic curve has a statistically significant double peak.
UPLC-MS/MS bioequivalence study of oral drugs of meldonium