Quantitative Assessment of Poorly Soluble Anticoagulant Rivaroxaban by Microemulsion Electrokinetic Chromatography

Quantitative Assessment of Poorly Soluble Anticoagulant Rivaroxaban by Microemulsion... Abstract Microemulsion electrokinetic chromatography (MEEKC) is an electrophoretic methodology based on the separation of compounds by a microemulsionated electrolyte. There are few options for the evaluation of the stability and content of the oral anticoagulant rivaroxaban (RIV) in pharmaceutical formulations. RIV has low water solubility and undergoes ionization only under restricted pH conditions (pH < 1 or pH > 13), thus, hindering the application of free zone capillary electrophoresis as an analytical method. Therefore, the work aimed at developing and validating a stability-indicating MEEKC method for the analysis of RIV in pharmaceutical formulations. Separation was performed in a fused-silica capillary applying a voltage of 30 kV. The microemulsion system consisted of 13 mM tetraborate, pH 9.75 + 1.2% SDS + 1.0% ethyl acetate + 2.4% butanol. The linearity range was 25–150 μg mL−1, with r = 0.9982. Drug degradations were performed in acid and basic media (HCl 1 M and NaOH 0.1 M, respectively), oxidation with 3%H2O2, 60°C temperature and exposure to UV-C radiation. No interferences with RIV or internal standard peaks were detected. Method robustness was accessed through Plackett–Burman experimental design, after evaluation of model validity. Trueness values between 100.49 and 100.68% were obtained with repeatability. The method developed was found appropriate for quality control of RIV tablets, as a consistent analytical technique that is considered less damaging to the environment due to its low consumption of organic reagents. Introduction Application of capillary electrophoresis (CE) in pharmaceutical analysis includes the assay of drugs, chiral separation, analysis of pharmaceutical excipients and determination of drug related impurities (1). CE separations are generally more efficient, can be performed on a faster time scale and require only nanolitre injection volumes (2). Reduced sample volume and low consumption of reagents and organic solvents, may reduce the cost of analysis and promote less environmental damage. Free solution CE (FSCE) methodology is based on the separation of charged and neutral compounds in aqueous media under the influence of an electric field; it is easy to implement and has wide application in the pharmaceutical analysis (3). The micellar electrokinetic chromatography (MEKC) mode applies micellar solutions of ionic surfactants for compound separation and has become one of the most popular techniques for the separation of small neutral molecules (4). Microemulsion electrokinetic chromatography (MEEKC) is a surfactant-based CE technique. Cosurfactants, usually short chain alcohols, along with surfactants are used to reduce surface tension between oil and water, producing nanometer sized oil droplets. The MEEKC system provides efficient separation for charged and neutral compounds based on drug partitioning into the droplet (5, 6). Rivaroxaban (RIV) is a novel oral anticoagulant prescribed as prophylaxis of thromboembolism and deep vein thrombosis (7). The RIV molecule (Figure 1) presents high ultraviolet absorption, which assists in the development of analytical methods with UV detectors. However, it has important limitations related to its lower water solubility (log P = 2.18) and reduced ionization in pH values between 1 and 13. These circumstances can increase the complexity of developing a reliable quantitative method for RIV analysis by CE. Chromatographic methods have been reported for analysis of the RIV pharmaceutical dosage form by high and ultra performance liquid chromatography (8, 9) yet, there are no studies approaching RIV assessment through MEEKC methodologies until now. Therefore, the objective of this survey was to develop and validate a stability indicating method able to detect degradation products (DPs) within the drug product and still quantify RIV in a pharmaceutical formulation with confidence by MEEKC technique. The full assessment of drug quality has great significance because it can assure safe and reliable treatments. Figure 1. View largeDownload slide Chemical structure of rivaroxaban. Figure 1. View largeDownload slide Chemical structure of rivaroxaban. Experimental Chemical and reagents RIV reference standard substance (99.63%) was obtained from BOC Science (New York, USA), and hydrochlorothiazide reference substance, applied as the internal standard (IS), was supplied by USP (Rockville, USA). Xarelto® tablets (Bayer Schering Pharma AG, Germany) containing 20 mg of RIV per dose were obtained from commercial sources. Excipients contained in the pharmaceutical dosage form (croscarmellose sodium, hypromellose, microcrystalline cellulose, iron oxide red, monohydrate lactose, macrogol, magnesium stearate, sodium lauryl sulfate, titanium dioxide) were all of the pharmaceutical grade and acquired from different distributors. Acetonitrile (ACN) obtained by Merck (Darmstadt, Germany) was HPLC grade. N-butanol, ethyl acetate, sodium tetraborate and sodium dodecyl sulfate (SDS) were purchased from Synth (São Paulo, Brazil). Water was purified using a Milli-Q system (Millipore, Bradford, USA). Apparatus CE experiments were performed on an Agilent 3DCE system (Agilent Technologies, Waldbronn, Germany), equipped with a photodiode array detector (DAD), an automatic sample injector, a temperature controlling system (4–60°C), and power supply able to deliver up to 30 kV. Data acquisition and analysis were done by CE ChemStation software (Agilent Technologies). Solutions Standard and sample stock solutions Standard stock solution of RIV (1.0 mg mL−1) was prepared in ACN. The solution was filtered through a 0.45 μm membrane filter, transferred to an amber volumetric flask and kept under refrigeration. Hydrochlorothiazide (IS) stock solution (1.0 mg mL−1) was also prepared in ACN following the same procedures described above. For analyses, an appropriate aliquot of RIV stock solution, spiked with IS stock solutions, was diluted in microemulsion electrolyte. To prepare RIV sample solutions (1 mg mL−1), tablets were weighed and crushed to fine powder. An appropriate amount was transferred to an amber volumetric flask and dissolved with ACN. The resultant suspension was filtered to remove insoluble excipients; a second filtration was performed through a 0.45 μm membrane filter. Placebo stock solution (prepared with excipients of RIV pharmaceutical formulation) followed the same procedure described for RIV sample solution. For analysis, an appropriate aliquot of RIV sample stock solution, spiked with IS stock solutions, was diluted in microemulsion electrolyte to the desired concentration. Preparation of microemulsion electrolyte The optimized electrolyte was prepared daily by mixing ethyl acetate and SDS at high speed (around 600 rpm) followed by butanol. A previously prepared 13 mM tetraborate solution (pH 9.75, adjusted with 1 M NaOH) was added at the end. The mixture was stirred for two more minutes to obtain a stable and optically transparent microemulsion electrolyte. Electrophoretic procedure Separations were carried out in an uncoated fused-silica capillary (Agilent Technologies, Germany) of 50 μm id and 48.5 cm total length (40 cm effective length). Before the first use, the capillary was conditioned flushing 1.0 M NaOH for 45 min, and water for 15 min. At the beginning of each working day, the capillary was rinsed with 0.1 M NaOH for 15 min, water for 10 min and then with microemulsion for 10 min. Before each analysis, the capillary was conditioned by rinsing it with 0.1 M NaOH for 3 min, water for 1.5 min and microemulsion for another 3 min. Samples were injected using the hydrodynamic procedure for 5 s at 50 mbar. The UV detector was set at 250 nm. Capillary temperature was kept constant at 35°C and the voltage applied was 30 kV. Validation of MEEKC method Validation of the proposed method was performed following the analytical parameters: selectivity, linearity, limit of detection (LOD), limit of quantification (LOQ), precision, accuracy and robustness for RIV assessment, based on ICH Guidelines (10, 11). In order to statistically assess the effect of small changes in method parameters and drug quantification, Plackett–Burman experimental design was applied to evaluate method robustness. MODDE® 11 software (Umetrics, Sweden) was used for data processing. System suitability parameters were monitored to ensure that the CE system and the method developed are capable of providing good quality reliable data based on USP 32 requirement (12). Results Optimization of electrophoretic conditions FSCE conditions were tested using different aqueous solutions with diverse pH values and concentrations. As can be seen in Figure 2A, RIV migration on FSCE electrolyte (50 mM ammonium acetate, pH 7) occurs together with the electro-osmotic flow (EOF) signal, preventing proper drug quantification. Analysis of RIV on MEKC electrolytes was unsuccessful. Although the micellar technique is frequently used for non-ionized pharmaceutical compounds, RIV did not present a coherent detection on MEKC mode. Figure 2. View largeDownload slide Electropherograms of RIV 60 μg mL−1 acquired during method development. (A) FSCE condition, electrolyte of 50 mM ammonium acetate, pH 7.0, applied voltage of 25 kV. (B) MEEKC system with 13 mM tetraborate pH 9.00 + 0.81% octane + 6.61% butanol + 3.31% SDS microemulsion electrolyte, applied voltage of 22 kV. Figure 2. View largeDownload slide Electropherograms of RIV 60 μg mL−1 acquired during method development. (A) FSCE condition, electrolyte of 50 mM ammonium acetate, pH 7.0, applied voltage of 25 kV. (B) MEEKC system with 13 mM tetraborate pH 9.00 + 0.81% octane + 6.61% butanol + 3.31% SDS microemulsion electrolyte, applied voltage of 22 kV. Another separation approach, MEEKC has been mostly applied to pharmaceutical analysis (5) especially due to its capacity to elucidate both water-soluble and insoluble drugs. The majority of published works applying MEEKC used an aqueous solution with high pH (13–16). In order to access the feasibility of RIV analysis under high pH, RIV liability was tested in basic solutions (pH from 8 to 10) protected from light, up to 6 h at 30°C. According to analyses performed on a previously validated HPLC method, RIV remained stable and without reduction in drug concentration. Therefore, alkaline aqueous solutions were tested without major concerns, always with drug sample freshly prepared on the electrolyte. Prior to initial tests, it was observed in relevant literature which microemulsion systems have a greater number of applications in pharmaceutical analysis in MEEKC mode (13, 17–19) based on which different microemulsionated electrolytes were tested. Variations were proposed in the constitution, pH and concentration of aqueous solutions, types of cosurfactants and lipophilic phase. Microemulsion constituted of 13 mM tetraborate pH 9.00 + 0.81% octane + 6.61% butanol + 3.31% SDS presented reproducible results, however, it generated an undesirably high current of 98 μA for 22 kV applied, and overly long migration time for RIV (8.9 min). Alternatively, ethyl acetate has a lower surface tension when compared to heptane and octane oils. Hence, surfactant concentration can be reduced and consequently, the voltage applied across the capillary may be increased without generation of excessively high current. A microemulsion system consisting of 13 mM tetraborate solution (pH 9.75), 1.0% ethyl acetate, 2.4% butanol and 1.2% SDS showed great results for RIV analysis. The applied voltage of 25 kV was switched to 30 kV, reducing drug migration time from 4.2 to 3.4 min, with a resultant stable current of 56 μA. The temperature was increased from 25 to 35°C, improving the peak signal. Prior to method validation procedure, several pharmaceutical compounds (glibenclamide, duloxetine, hydrochlorothiazide, ranitidine, atenolol, prasugrel chlorhydrate and clopidogrel) were investigated as IS. Hydrochlorothiazide standard substance presented the most suitable results (good UV absorption at 249 nm and resolution higher than 2 with RIV and DPs. Therefore, RIV assays were always evaluated according to the RIV/IS areas ratio. IS migration time was reproducible at 2.6 min (Figure 3A). Figure 3. View largeDownload slide Electropherograms of RIV analysis by MEEKC. (A) RIV standard substance; (B) RIV under 1 M HCl medium at 60°C; (C) RIV exposure to 0.1 M NaOH medium under 60°C; and (D) RIV solution after exposure to UVC radiation. All the electropherograms were performed at nominal conditions 13 mM tetraborate (pH 9.75), 1.0% ethyl acetate, 2.4% butanol and 1.2% SDS microemulsion; 35°C capillary temperature and 30 kV of applied voltage. Figure 3. View largeDownload slide Electropherograms of RIV analysis by MEEKC. (A) RIV standard substance; (B) RIV under 1 M HCl medium at 60°C; (C) RIV exposure to 0.1 M NaOH medium under 60°C; and (D) RIV solution after exposure to UVC radiation. All the electropherograms were performed at nominal conditions 13 mM tetraborate (pH 9.75), 1.0% ethyl acetate, 2.4% butanol and 1.2% SDS microemulsion; 35°C capillary temperature and 30 kV of applied voltage. Method validation Method selectivity was evaluated to ensure the absence of interferences in RIV quantification. Placebo sample was tested to assess whether excipients contained in the pharmaceutical product could impair RIV or IS detection. The powder mixture of excipients was produced following the usually applied concentrations (20). Interference of pharmaceutical excipients from RIV formulations was evaluated and no disruption was detected for RIV and IS peaks. Protocols of forced degradation were designed to evaluate the analytical conditions for a stability study of RIV and to ensure that DPs do not interfere in RIV or IS migration times. Drug stock solutions at 1 mg mL−1 were prepared in ACN for the stress degradation studies. Appropriate aliquots were diluted with pure water, 1 M HCl, 0.1 M NaOH, and 3% H2O2 media to a final concentration of 500 μg mL−1. These solutions were kept at 60°C up to 10 hours or until they reached 50% drug degradation. RIV concentration was analyzed for each time sample. Acid and alkaline solutions were neutralized to stop the reaction. RIV photostability was tested exposing drug solution to UVC radiation at room temperature in a mirrored chamber (1.0 × 0.12 × 0.17 m3) equipped with UVC lamps (Ecolume ZW®, 254 nm, 15 W). For CE analyses, appropriate aliquots of the above solutions were spiked with IS solution and diluted to RIV theoretical concentration of 100 μg mL−1. Peak purity of RIV and DPs detected in electropherograms was determined using the Agilent Chemstation software tools. Exposure of RIV sample solutions to UVC radiation caused drug photolysis, generating two detectable DPs at 1.9 and 2.1 min (Figure 3D). Drug concentration had a reduction of ~28% after 2 h. Hydrolysis was achieved after drug exposure to 0.1 M NaOH and 1.0 M HCl, both at 60°C. In alkaline medium, RIV reached 50% degradation after 0.5 h with the formation of one DP detected at 3.6 min (Figure 3C), whereas in acid medium after 4.0 h RIV concentration had been reduced by around 24% and one DP was detected at 2.4 min (Figure 3B). RIV exposure to thermal and oxidative stress did not result in drug degradation and DP formation, respectively. Compliance with the linear method response was evaluated by linear regression analysis calculated by the least square regression and ANOVA (analysis of variance) (α = 0.05). Standard plots were constructed with six concentration levels (25, 50, 100, 125 and 150 μg/mL) on 3 consecutive days. As shown in Table I, the method developed exhibited adequate linearity (r = 0.9982). According to the ANOVA statistical evaluation, the method presented significant linear regression (P < 0.05) and non significant deviation from linearity (P > 0.05). Limits of quantification and detection (LOQ and LOD) were obtained based on the signal-to-noise ratio (10:1 for LOQ and 3:1 for LOD) obtained for the RIV migration time range. Values obtained for RIV LOD and LOQ are also shown in Table I. Table I. Linearity Data (n = 7), Limit of Detection (LOD) and Limit of Quantification (LOQ) Values for RIV Assessment by MEEKC Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  aMean of three replicates. Table I. Linearity Data (n = 7), Limit of Detection (LOD) and Limit of Quantification (LOQ) Values for RIV Assessment by MEEKC Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  aMean of three replicates. The precision of the method for quantification of RIV was verified through repeatability (intra-day) and intermediate precision (inter-day) analyses. Results were expressed as relative standard deviation (RSD). Six independent samples were prepared on the same day by diluting 1 mL of RIV stock solution and 1 mL IS stock solution in 10 mL volumetric flask with electrolyte, with a final concentration of 100 μg mL−1 for RIV and IS. Intermediate precision was assessed by carrying out the analysis on 3 different days. Repeatability and intermediate precision RSD values presented in Table II were under 1.0% for both cases, which demonstrates adequate precision of the developed method. Method accuracy was determined through the recovery of different amounts of RIV standard solution added to placebo solution. Sample solutions for accuracy were prepared at three concentration levels (low, intermediate and high) equivalent to 50, 100 and 150 μg mL−1 of RIV, and always spiked with IS at 100 μg mL−1 (n = 3, for each added concentration). Accuracy was calculated as the percentage of drug added recovered from the placebo solution. The mean recovery data for RIV was within 99.05–100.66%, satisfying the acceptance criteria for the study (Table III). Table II. Intra-day and Inter-day Results for Assessment of Method Precision for RIV   Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66    Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66  Table II. Intra-day and Inter-day Results for Assessment of Method Precision for RIV   Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66    Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66  Table III. Accuracy of MEEKC Method for RIV Analysis Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  aPrepared in three replicates. Table III. Accuracy of MEEKC Method for RIV Analysis Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  aPrepared in three replicates. Five factors related to electrolyte constitution (tetraborate, SDS, ethyl acetate and butanol concentration and pH range) were modeled for robustness assessment by Plackett–Burman fractional design on binary levels. The experimental design was composed of eight runs with factors scrambled at two levels (−1, +1) and two central points equivalent to validated conditions. Controllable factors modified and method responses (RIV content, migration time, IS/RIV resolution) are summarized in Table IV. The F-test was applied to assess model fitness (21). F-values for RIV content, migration time, and resolution were 4.98, 96.52 and 4.76, respectively. Considering the degrees of freedom for model error and pure error, Ftab = 215.17. All responses assessed had Fobs < Ftab (α = 0.05), stating the adequacy of the Plackett–Burman model to evaluate method robustness with no lack of fit. The effect of changed factors on responses is presented in Figure 4. Whenever the standard error bar is larger than column effect and also crosses y = 0, the factor is considered not significant for the response presented. Therefore, as presented in Figure 4, none of the changes performed had a significant effect on the RIV assay for content, which demonstrated the robustness of the method. Figure 4 also shows that the factor variations did not change RIV migration time significantly. As for resolution between IS and RIV peaks, it was found that changes from −1 to +1 levels of factors pH, SDS and butanol had significant effects on the resolution, pH and butanol reducing and SDS increasing values. Table IV. Plackett–Burman Experimental Design and Response Values Assessed at Each Run for MEEKC Method Robustness Evaluation Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  a RIV content regarding amount declared by manufacturer. b Migration time. c Resolution between IS and RIV electrophoretic peaks. d Central point and nominal microemulsion: 13 mM tetraborate pH 9.75 + 1.2% SDS + 1.0% Ethyl acetate + 2.4% butanol. Table IV. Plackett–Burman Experimental Design and Response Values Assessed at Each Run for MEEKC Method Robustness Evaluation Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  a RIV content regarding amount declared by manufacturer. b Migration time. c Resolution between IS and RIV electrophoretic peaks. d Central point and nominal microemulsion: 13 mM tetraborate pH 9.75 + 1.2% SDS + 1.0% Ethyl acetate + 2.4% butanol. Figure 4. View largeDownload slide Graphical representation of the effects from changes in selected factors on RIV sample content, migration time (Tm) and resolution between IS and RIV peaks. Method robustness assessed through Plackett–Burman design. Column effect larger than error bar represents a significant response (α = 0.05). Figure 4. View largeDownload slide Graphical representation of the effects from changes in selected factors on RIV sample content, migration time (Tm) and resolution between IS and RIV peaks. Method robustness assessed through Plackett–Burman design. Column effect larger than error bar represents a significant response (α = 0.05). System suitability parameters were accessed on different days of method validation, through Agilent Chemstation software method validation. Average values of RIV migration time (3.36), theoretical plates (19,267), peak symmetry (1.37), and resolution (9.00). Discussion In the CE method development study, several electrolyte solutions were tested in order to achieve the highest sensitivity and shortest analysis time for RIV assessment. Since RIV is a weak amphoteric compound, pKa values (1.01 and 13.36), its full ionization in aqueous media with pH values usually applied in CE (from 3 to 12) is quite limited. Therefore, RIV molecule is mainly non-ionized when applying FSCE conditions and ionic separation is unfeasible. Although MEKC mode can separate adequately non ionized molecules, in the experiments performed the analyses of RIV by MEKC showed results with poor reproducibility. On the other hand, the evaluation of RIV in a microemulsion CE system presented better results (Figure 2B). MEKC micelles present a rigid structure engendered by surfactant and aqueous media, providing a more hostile environment for the analysis of molecules with hydrophobic characteristics such as RIV. Whereas, droplets have a more flexible surface, which allows an easier penetration of solutes and therefore, RIV as a molecule of low water solubility and neutral in a major range of pH (2–12), had its separation by partitioning with the droplet. The influence of different analytical parameters on resolution, electric current generated and peak symmetry were considered by method optimization. Ethyl acetate was chosen as the hydrophobic core for the MEEKC system since it has a lower surface tension than oils such as heptane and octane. Thus, the concentration of surfactant in the solution can be reduced, which enables the application of a higher voltage (up to 30 kV) across the capillary, without producing undesirable current levels (above 100 uA). The major reason for the high current generated (resulting of the high-ionic strength) is the relatively high concentrations of surfactants needed to form the microemulsion when using heptane or octane (3, 19). Application of IS is highly recommended in CE quantitative analysis to correct intrinsic errors caused by variability in pressure injection mode, voltage and EOF (22). An ideal IS presents good absorption at a selected wavelength; IS peak must have a proper resolution in relation to analyte and DPs (R ≥ 2), and should not increase analysis time. Working with an IS in CE analysis increases the precision and accuracy of quantitative measurements, therefore, we selected an IS for MEEKC analyses. As presented in the method validation results, it was possible to develop a reliable MEEKC system to quantify RIV in pharmaceutical form and also to observe the existence of DP among samples. All validation parameters are in accordance with international guidelines. Plackett–Burman experimental design was chosen for robustness assessment due to its fast application and the possibility of evaluating the effect of many factors at once. The goodness of fit test was evaluated by an F-test of sum of squares for lack of fit and pure error, assuring the fitness of the model to each response. Effects plot is a graphical representation of the magnitude and significance of evaluated factors on selected responses; non significant variations were found in RIV assay for sample content. Significant changes in IS and RIV resolution do not affect RIV quantification and method robustness since this value was never smaller than 3.0, as presented in Table IV. Those results ensure the capability of the method to remain reliable after undergoing minor variations. System suitability parameters demonstrate the qualification of the electrophoretic system for pharmaceutical analysis. Concluding remarks A stability indicating method based on MEEKC was developed for quantitative determination of RIV in a pharmaceutical formulation. Method performance in terms of accuracy and precision with analysis time of fewer than 4 min demonstrates that microemulsion systems can be successfully applied in drug analysis. According to statistical analysis, the method developed was found to be sensitive, specific, robust and linear for the intended purpose. Therefore, the MEEKC method with PDA detection can be applied as a reliable and simple technique in quality control of RIV tablets, and an analytical alternative with reduced environmental damage and low financing costs. Conflict of interest statement Authors declare the absence of conflict of interest. References 1 Altria, K., Marsh, A., Griend, C.S.; Capillary electrophoresis for the analysis of small-molecule pharmaceuticals; Electrophoresis , ( 2006); 27: 2263– 2282. Google Scholar CrossRef Search ADS PubMed  2 Holland, L.A., Chetwyn, N.P., Perkins, M.D., Lunte, S.M.; Capillary electrophoresis in pharmaceutical analysis; Pharmaceutical Research , ( 1997); 14: 372– 388. Google Scholar CrossRef Search ADS PubMed  3 Altria, K.D.; Background theory and applications of microemulsion electrokinetic chromatography; Journal of Chromatography A , ( 2000); 892: 171– 186. Google Scholar CrossRef Search ADS PubMed  4 Ryan, R., Donegan, S., Power, J., Altria, K.; Advances in the theory and application of MEEKC; Electrophoresis , ( 2010); 31: 755– 767. Google Scholar CrossRef Search ADS PubMed  5 McBride, B.F.; Preliminary assessment of the critical differences between novel oral anticoagulants currently in development; Journal of Clinical Pharmacology , ( 2005); 45: 1004– 1017. Google Scholar CrossRef Search ADS PubMed  6 Çelebier, M., Reçber, T., Koçak, E., Altinöz, S.; RP-HPLC method development and validation for estimation of rivaroxaban in pharmaceutical dosage forms; Brazilian Journal of Pharmaceutical Sciences , ( 2013); 6: 359– 366. Google Scholar CrossRef Search ADS   7 Wingert, N.R., Nunes, M.A.G., Barden, A.T., Gomes, P., Muller, E.I., Flores, E.M.M., et al.  .; Ultra-performance LC-ESI/Q-TOF MS for the rapid analysis of rivaroxaban: method validation using experimental design for robustness evaluation; Current Analytical Chemistry , ( 2015); 11: 124– 129. Google Scholar CrossRef Search ADS   8 ICH. Harmonised Tripartite Guideline: Validation of Analytical Procedure: text and methodology Q2 (R1). Geneve, Switzerland ( 2005). 9 ICH. Harmonised Tripartite Guideline: Stability Testing of New Drug Substance and Products: Methodology Q1A (R2), ( 2003). 10 Rowe, R.C., Sheskey, P.J., Quinn, M.E. (eds.); Handbook of Pharmaceutical Excipients , 6th ed. Pharmaceutical Press, Berlin, Germany, ( 2009). 11 USP. The United States Pharmacopeia 37th Rockville, MD ( 2014). 12 Altria, K.D. (ed).; Capillary electrophoresis guidebook , Vol. 52. Human Press, New Jersey, ( 1996). 13 Mahuzier, P.E., Prado, A., Clark, B.J., Kedor-Hackmann, E.R.M., Altria, K.D. An introduction to the theory and application of microemulsion electrokinetic chromatography. CE Currents, ( 2003). 14 Wen, T., Zhao, X., Luo, G., Wang, J., Wang, Y., Yao, B.; Comparison of microemulsion electrokinetic chromatography and solvent modified micellar electrokinetic chromatography on rapid separation of heroin, amphetamine and their basic impurities; Talanta , ( 2007); 71: 854– 860. Google Scholar CrossRef Search ADS PubMed  15 Lynen, F., Saavedra, L., Nickerson, B., Sandra, P.; Evaluation of a multiarray system for pharmaceutical analysis by microemulsion electrokinetic chromatography; Talanta , ( 2001); 84: 724– 729. Google Scholar CrossRef Search ADS   16 Nussbaumer, S., Fleury-Souverain, S., Schappler, J., Rudaz, S., Veuthey, J.-L., Bonnabry, P.; Quality control of pharmaceutical formulations containing cisplatin, carboplatin, and oxaliplatin by micellar and microemulsion electrokinetic chromatography (MEKC, MEEKC); Journal of Pharmaceutical and Biomedical Analysis , ( 2011); 55: 253– 258. Google Scholar CrossRef Search ADS PubMed  17 Miola, M.F., Snowden, M.J., Altria, K.D.; The use of microemulsion electrokinetic chromatography in pharmaceutical analysis; Journal of Pharmaceutical and Biomedical Analysis , ( 1998); 18: 785– 797. Google Scholar CrossRef Search ADS PubMed  18 Yu, L., Chu, K., Ye, H., Liu, X., Yu, L., Xu, X., et al.  .; Recent advances in microemulsion electrokinetic chromatography; Trends in Analytical Chemistry , ( 2012); 34: 140– 151. Google Scholar CrossRef Search ADS   19 Yang, H., Ding, Y., Cao, J., Li, P.; Twenty-one years of microemulsion electrokinetic chromatography (1991–2012): a powerful analytical tool; Electrophoresis , ( 2013); 34: 1273– 1294. Google Scholar CrossRef Search ADS PubMed  20 Altria, K.D., Elder, D.; Overview of the status and applications of capillary electrophoresis to the analysis of small molecules; Journal of Chromatography A , ( 2004); 1023: 1– 14. Google Scholar CrossRef Search ADS PubMed  21 Montgomery, D.C. (ed).; Design and analysis of experiments , 5th ed. John Willey & Sons, Inc, USA, ( 2001); pp. 177– 180. 22 Jouyban, A., Kenndler, E.; Impurity analysis of pharmaceuticals using capillary electromigration methods; Electrophoresis , ( 2008); 29: 3531– 3551. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Chromatographic Science Oxford University Press

Quantitative Assessment of Poorly Soluble Anticoagulant Rivaroxaban by Microemulsion Electrokinetic Chromatography

Loading next page...
 
/lp/ou_press/quantitative-assessment-of-poorly-soluble-anticoagulant-rivaroxaban-by-8y1vOXer0J
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0021-9665
eISSN
1945-239X
D.O.I.
10.1093/chromsci/bmy036
Publisher site
See Article on Publisher Site

Abstract

Abstract Microemulsion electrokinetic chromatography (MEEKC) is an electrophoretic methodology based on the separation of compounds by a microemulsionated electrolyte. There are few options for the evaluation of the stability and content of the oral anticoagulant rivaroxaban (RIV) in pharmaceutical formulations. RIV has low water solubility and undergoes ionization only under restricted pH conditions (pH < 1 or pH > 13), thus, hindering the application of free zone capillary electrophoresis as an analytical method. Therefore, the work aimed at developing and validating a stability-indicating MEEKC method for the analysis of RIV in pharmaceutical formulations. Separation was performed in a fused-silica capillary applying a voltage of 30 kV. The microemulsion system consisted of 13 mM tetraborate, pH 9.75 + 1.2% SDS + 1.0% ethyl acetate + 2.4% butanol. The linearity range was 25–150 μg mL−1, with r = 0.9982. Drug degradations were performed in acid and basic media (HCl 1 M and NaOH 0.1 M, respectively), oxidation with 3%H2O2, 60°C temperature and exposure to UV-C radiation. No interferences with RIV or internal standard peaks were detected. Method robustness was accessed through Plackett–Burman experimental design, after evaluation of model validity. Trueness values between 100.49 and 100.68% were obtained with repeatability. The method developed was found appropriate for quality control of RIV tablets, as a consistent analytical technique that is considered less damaging to the environment due to its low consumption of organic reagents. Introduction Application of capillary electrophoresis (CE) in pharmaceutical analysis includes the assay of drugs, chiral separation, analysis of pharmaceutical excipients and determination of drug related impurities (1). CE separations are generally more efficient, can be performed on a faster time scale and require only nanolitre injection volumes (2). Reduced sample volume and low consumption of reagents and organic solvents, may reduce the cost of analysis and promote less environmental damage. Free solution CE (FSCE) methodology is based on the separation of charged and neutral compounds in aqueous media under the influence of an electric field; it is easy to implement and has wide application in the pharmaceutical analysis (3). The micellar electrokinetic chromatography (MEKC) mode applies micellar solutions of ionic surfactants for compound separation and has become one of the most popular techniques for the separation of small neutral molecules (4). Microemulsion electrokinetic chromatography (MEEKC) is a surfactant-based CE technique. Cosurfactants, usually short chain alcohols, along with surfactants are used to reduce surface tension between oil and water, producing nanometer sized oil droplets. The MEEKC system provides efficient separation for charged and neutral compounds based on drug partitioning into the droplet (5, 6). Rivaroxaban (RIV) is a novel oral anticoagulant prescribed as prophylaxis of thromboembolism and deep vein thrombosis (7). The RIV molecule (Figure 1) presents high ultraviolet absorption, which assists in the development of analytical methods with UV detectors. However, it has important limitations related to its lower water solubility (log P = 2.18) and reduced ionization in pH values between 1 and 13. These circumstances can increase the complexity of developing a reliable quantitative method for RIV analysis by CE. Chromatographic methods have been reported for analysis of the RIV pharmaceutical dosage form by high and ultra performance liquid chromatography (8, 9) yet, there are no studies approaching RIV assessment through MEEKC methodologies until now. Therefore, the objective of this survey was to develop and validate a stability indicating method able to detect degradation products (DPs) within the drug product and still quantify RIV in a pharmaceutical formulation with confidence by MEEKC technique. The full assessment of drug quality has great significance because it can assure safe and reliable treatments. Figure 1. View largeDownload slide Chemical structure of rivaroxaban. Figure 1. View largeDownload slide Chemical structure of rivaroxaban. Experimental Chemical and reagents RIV reference standard substance (99.63%) was obtained from BOC Science (New York, USA), and hydrochlorothiazide reference substance, applied as the internal standard (IS), was supplied by USP (Rockville, USA). Xarelto® tablets (Bayer Schering Pharma AG, Germany) containing 20 mg of RIV per dose were obtained from commercial sources. Excipients contained in the pharmaceutical dosage form (croscarmellose sodium, hypromellose, microcrystalline cellulose, iron oxide red, monohydrate lactose, macrogol, magnesium stearate, sodium lauryl sulfate, titanium dioxide) were all of the pharmaceutical grade and acquired from different distributors. Acetonitrile (ACN) obtained by Merck (Darmstadt, Germany) was HPLC grade. N-butanol, ethyl acetate, sodium tetraborate and sodium dodecyl sulfate (SDS) were purchased from Synth (São Paulo, Brazil). Water was purified using a Milli-Q system (Millipore, Bradford, USA). Apparatus CE experiments were performed on an Agilent 3DCE system (Agilent Technologies, Waldbronn, Germany), equipped with a photodiode array detector (DAD), an automatic sample injector, a temperature controlling system (4–60°C), and power supply able to deliver up to 30 kV. Data acquisition and analysis were done by CE ChemStation software (Agilent Technologies). Solutions Standard and sample stock solutions Standard stock solution of RIV (1.0 mg mL−1) was prepared in ACN. The solution was filtered through a 0.45 μm membrane filter, transferred to an amber volumetric flask and kept under refrigeration. Hydrochlorothiazide (IS) stock solution (1.0 mg mL−1) was also prepared in ACN following the same procedures described above. For analyses, an appropriate aliquot of RIV stock solution, spiked with IS stock solutions, was diluted in microemulsion electrolyte. To prepare RIV sample solutions (1 mg mL−1), tablets were weighed and crushed to fine powder. An appropriate amount was transferred to an amber volumetric flask and dissolved with ACN. The resultant suspension was filtered to remove insoluble excipients; a second filtration was performed through a 0.45 μm membrane filter. Placebo stock solution (prepared with excipients of RIV pharmaceutical formulation) followed the same procedure described for RIV sample solution. For analysis, an appropriate aliquot of RIV sample stock solution, spiked with IS stock solutions, was diluted in microemulsion electrolyte to the desired concentration. Preparation of microemulsion electrolyte The optimized electrolyte was prepared daily by mixing ethyl acetate and SDS at high speed (around 600 rpm) followed by butanol. A previously prepared 13 mM tetraborate solution (pH 9.75, adjusted with 1 M NaOH) was added at the end. The mixture was stirred for two more minutes to obtain a stable and optically transparent microemulsion electrolyte. Electrophoretic procedure Separations were carried out in an uncoated fused-silica capillary (Agilent Technologies, Germany) of 50 μm id and 48.5 cm total length (40 cm effective length). Before the first use, the capillary was conditioned flushing 1.0 M NaOH for 45 min, and water for 15 min. At the beginning of each working day, the capillary was rinsed with 0.1 M NaOH for 15 min, water for 10 min and then with microemulsion for 10 min. Before each analysis, the capillary was conditioned by rinsing it with 0.1 M NaOH for 3 min, water for 1.5 min and microemulsion for another 3 min. Samples were injected using the hydrodynamic procedure for 5 s at 50 mbar. The UV detector was set at 250 nm. Capillary temperature was kept constant at 35°C and the voltage applied was 30 kV. Validation of MEEKC method Validation of the proposed method was performed following the analytical parameters: selectivity, linearity, limit of detection (LOD), limit of quantification (LOQ), precision, accuracy and robustness for RIV assessment, based on ICH Guidelines (10, 11). In order to statistically assess the effect of small changes in method parameters and drug quantification, Plackett–Burman experimental design was applied to evaluate method robustness. MODDE® 11 software (Umetrics, Sweden) was used for data processing. System suitability parameters were monitored to ensure that the CE system and the method developed are capable of providing good quality reliable data based on USP 32 requirement (12). Results Optimization of electrophoretic conditions FSCE conditions were tested using different aqueous solutions with diverse pH values and concentrations. As can be seen in Figure 2A, RIV migration on FSCE electrolyte (50 mM ammonium acetate, pH 7) occurs together with the electro-osmotic flow (EOF) signal, preventing proper drug quantification. Analysis of RIV on MEKC electrolytes was unsuccessful. Although the micellar technique is frequently used for non-ionized pharmaceutical compounds, RIV did not present a coherent detection on MEKC mode. Figure 2. View largeDownload slide Electropherograms of RIV 60 μg mL−1 acquired during method development. (A) FSCE condition, electrolyte of 50 mM ammonium acetate, pH 7.0, applied voltage of 25 kV. (B) MEEKC system with 13 mM tetraborate pH 9.00 + 0.81% octane + 6.61% butanol + 3.31% SDS microemulsion electrolyte, applied voltage of 22 kV. Figure 2. View largeDownload slide Electropherograms of RIV 60 μg mL−1 acquired during method development. (A) FSCE condition, electrolyte of 50 mM ammonium acetate, pH 7.0, applied voltage of 25 kV. (B) MEEKC system with 13 mM tetraborate pH 9.00 + 0.81% octane + 6.61% butanol + 3.31% SDS microemulsion electrolyte, applied voltage of 22 kV. Another separation approach, MEEKC has been mostly applied to pharmaceutical analysis (5) especially due to its capacity to elucidate both water-soluble and insoluble drugs. The majority of published works applying MEEKC used an aqueous solution with high pH (13–16). In order to access the feasibility of RIV analysis under high pH, RIV liability was tested in basic solutions (pH from 8 to 10) protected from light, up to 6 h at 30°C. According to analyses performed on a previously validated HPLC method, RIV remained stable and without reduction in drug concentration. Therefore, alkaline aqueous solutions were tested without major concerns, always with drug sample freshly prepared on the electrolyte. Prior to initial tests, it was observed in relevant literature which microemulsion systems have a greater number of applications in pharmaceutical analysis in MEEKC mode (13, 17–19) based on which different microemulsionated electrolytes were tested. Variations were proposed in the constitution, pH and concentration of aqueous solutions, types of cosurfactants and lipophilic phase. Microemulsion constituted of 13 mM tetraborate pH 9.00 + 0.81% octane + 6.61% butanol + 3.31% SDS presented reproducible results, however, it generated an undesirably high current of 98 μA for 22 kV applied, and overly long migration time for RIV (8.9 min). Alternatively, ethyl acetate has a lower surface tension when compared to heptane and octane oils. Hence, surfactant concentration can be reduced and consequently, the voltage applied across the capillary may be increased without generation of excessively high current. A microemulsion system consisting of 13 mM tetraborate solution (pH 9.75), 1.0% ethyl acetate, 2.4% butanol and 1.2% SDS showed great results for RIV analysis. The applied voltage of 25 kV was switched to 30 kV, reducing drug migration time from 4.2 to 3.4 min, with a resultant stable current of 56 μA. The temperature was increased from 25 to 35°C, improving the peak signal. Prior to method validation procedure, several pharmaceutical compounds (glibenclamide, duloxetine, hydrochlorothiazide, ranitidine, atenolol, prasugrel chlorhydrate and clopidogrel) were investigated as IS. Hydrochlorothiazide standard substance presented the most suitable results (good UV absorption at 249 nm and resolution higher than 2 with RIV and DPs. Therefore, RIV assays were always evaluated according to the RIV/IS areas ratio. IS migration time was reproducible at 2.6 min (Figure 3A). Figure 3. View largeDownload slide Electropherograms of RIV analysis by MEEKC. (A) RIV standard substance; (B) RIV under 1 M HCl medium at 60°C; (C) RIV exposure to 0.1 M NaOH medium under 60°C; and (D) RIV solution after exposure to UVC radiation. All the electropherograms were performed at nominal conditions 13 mM tetraborate (pH 9.75), 1.0% ethyl acetate, 2.4% butanol and 1.2% SDS microemulsion; 35°C capillary temperature and 30 kV of applied voltage. Figure 3. View largeDownload slide Electropherograms of RIV analysis by MEEKC. (A) RIV standard substance; (B) RIV under 1 M HCl medium at 60°C; (C) RIV exposure to 0.1 M NaOH medium under 60°C; and (D) RIV solution after exposure to UVC radiation. All the electropherograms were performed at nominal conditions 13 mM tetraborate (pH 9.75), 1.0% ethyl acetate, 2.4% butanol and 1.2% SDS microemulsion; 35°C capillary temperature and 30 kV of applied voltage. Method validation Method selectivity was evaluated to ensure the absence of interferences in RIV quantification. Placebo sample was tested to assess whether excipients contained in the pharmaceutical product could impair RIV or IS detection. The powder mixture of excipients was produced following the usually applied concentrations (20). Interference of pharmaceutical excipients from RIV formulations was evaluated and no disruption was detected for RIV and IS peaks. Protocols of forced degradation were designed to evaluate the analytical conditions for a stability study of RIV and to ensure that DPs do not interfere in RIV or IS migration times. Drug stock solutions at 1 mg mL−1 were prepared in ACN for the stress degradation studies. Appropriate aliquots were diluted with pure water, 1 M HCl, 0.1 M NaOH, and 3% H2O2 media to a final concentration of 500 μg mL−1. These solutions were kept at 60°C up to 10 hours or until they reached 50% drug degradation. RIV concentration was analyzed for each time sample. Acid and alkaline solutions were neutralized to stop the reaction. RIV photostability was tested exposing drug solution to UVC radiation at room temperature in a mirrored chamber (1.0 × 0.12 × 0.17 m3) equipped with UVC lamps (Ecolume ZW®, 254 nm, 15 W). For CE analyses, appropriate aliquots of the above solutions were spiked with IS solution and diluted to RIV theoretical concentration of 100 μg mL−1. Peak purity of RIV and DPs detected in electropherograms was determined using the Agilent Chemstation software tools. Exposure of RIV sample solutions to UVC radiation caused drug photolysis, generating two detectable DPs at 1.9 and 2.1 min (Figure 3D). Drug concentration had a reduction of ~28% after 2 h. Hydrolysis was achieved after drug exposure to 0.1 M NaOH and 1.0 M HCl, both at 60°C. In alkaline medium, RIV reached 50% degradation after 0.5 h with the formation of one DP detected at 3.6 min (Figure 3C), whereas in acid medium after 4.0 h RIV concentration had been reduced by around 24% and one DP was detected at 2.4 min (Figure 3B). RIV exposure to thermal and oxidative stress did not result in drug degradation and DP formation, respectively. Compliance with the linear method response was evaluated by linear regression analysis calculated by the least square regression and ANOVA (analysis of variance) (α = 0.05). Standard plots were constructed with six concentration levels (25, 50, 100, 125 and 150 μg/mL) on 3 consecutive days. As shown in Table I, the method developed exhibited adequate linearity (r = 0.9982). According to the ANOVA statistical evaluation, the method presented significant linear regression (P < 0.05) and non significant deviation from linearity (P > 0.05). Limits of quantification and detection (LOQ and LOD) were obtained based on the signal-to-noise ratio (10:1 for LOQ and 3:1 for LOD) obtained for the RIV migration time range. Values obtained for RIV LOD and LOQ are also shown in Table I. Table I. Linearity Data (n = 7), Limit of Detection (LOD) and Limit of Quantification (LOQ) Values for RIV Assessment by MEEKC Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  aMean of three replicates. Table I. Linearity Data (n = 7), Limit of Detection (LOD) and Limit of Quantification (LOQ) Values for RIV Assessment by MEEKC Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  Parameters  RIV data  Concentration rangea (μg mL−1)  25.0–150.0  Intercept  −0.0736  Slope  0.0355  Coefficient of correlation (r)  0.9982  LOD (μg mL−1)  1.68  LOQ (μg mL−1)  5.60  aMean of three replicates. The precision of the method for quantification of RIV was verified through repeatability (intra-day) and intermediate precision (inter-day) analyses. Results were expressed as relative standard deviation (RSD). Six independent samples were prepared on the same day by diluting 1 mL of RIV stock solution and 1 mL IS stock solution in 10 mL volumetric flask with electrolyte, with a final concentration of 100 μg mL−1 for RIV and IS. Intermediate precision was assessed by carrying out the analysis on 3 different days. Repeatability and intermediate precision RSD values presented in Table II were under 1.0% for both cases, which demonstrates adequate precision of the developed method. Method accuracy was determined through the recovery of different amounts of RIV standard solution added to placebo solution. Sample solutions for accuracy were prepared at three concentration levels (low, intermediate and high) equivalent to 50, 100 and 150 μg mL−1 of RIV, and always spiked with IS at 100 μg mL−1 (n = 3, for each added concentration). Accuracy was calculated as the percentage of drug added recovered from the placebo solution. The mean recovery data for RIV was within 99.05–100.66%, satisfying the acceptance criteria for the study (Table III). Table II. Intra-day and Inter-day Results for Assessment of Method Precision for RIV   Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66    Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66  Table II. Intra-day and Inter-day Results for Assessment of Method Precision for RIV   Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66    Intra-day (n = 6)  Inter-day (n = 18)  Day 1  Day 2  Day 3  RIV (assay,%)  100.49  100.68  100.54  100.57  RSD (%)  0.70  0.24  0.96  0.66  Table III. Accuracy of MEEKC Method for RIV Analysis Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  aPrepared in three replicates. Table III. Accuracy of MEEKC Method for RIV Analysis Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  Nominal concentration (μg mL−1)  Mean concentration founda (μg mL−1)  Recovered values (%)  RSD (%)  50  50.90  100.66  0.66  100  100.17  99.05  0.96  150  151.33  99.76  1.45  aPrepared in three replicates. Five factors related to electrolyte constitution (tetraborate, SDS, ethyl acetate and butanol concentration and pH range) were modeled for robustness assessment by Plackett–Burman fractional design on binary levels. The experimental design was composed of eight runs with factors scrambled at two levels (−1, +1) and two central points equivalent to validated conditions. Controllable factors modified and method responses (RIV content, migration time, IS/RIV resolution) are summarized in Table IV. The F-test was applied to assess model fitness (21). F-values for RIV content, migration time, and resolution were 4.98, 96.52 and 4.76, respectively. Considering the degrees of freedom for model error and pure error, Ftab = 215.17. All responses assessed had Fobs < Ftab (α = 0.05), stating the adequacy of the Plackett–Burman model to evaluate method robustness with no lack of fit. The effect of changed factors on responses is presented in Figure 4. Whenever the standard error bar is larger than column effect and also crosses y = 0, the factor is considered not significant for the response presented. Therefore, as presented in Figure 4, none of the changes performed had a significant effect on the RIV assay for content, which demonstrated the robustness of the method. Figure 4 also shows that the factor variations did not change RIV migration time significantly. As for resolution between IS and RIV peaks, it was found that changes from −1 to +1 levels of factors pH, SDS and butanol had significant effects on the resolution, pH and butanol reducing and SDS increasing values. Table IV. Plackett–Burman Experimental Design and Response Values Assessed at Each Run for MEEKC Method Robustness Evaluation Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  a RIV content regarding amount declared by manufacturer. b Migration time. c Resolution between IS and RIV electrophoretic peaks. d Central point and nominal microemulsion: 13 mM tetraborate pH 9.75 + 1.2% SDS + 1.0% Ethyl acetate + 2.4% butanol. Table IV. Plackett–Burman Experimental Design and Response Values Assessed at Each Run for MEEKC Method Robustness Evaluation Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  Exp. No.  Tetraborate (mM)  SDS (%)  Ethyl acetate (%)  Butanol (%)  pH  RIV content (%)a  Tmb  Resolutionc  1  15  1.0  0.8  2.4  9.50  100.62  3.13  10.43  2  15  1.0  0.8  2.6  10.00  99.85  3.04  3.75  3  15  1.0  1.2  2.6  9.50  99.18  3.10  8.70  4  11  1.4  1.2  2.6  9.50  99.93  3.01  10.07  5  15  1.4  1.2  2.6  10.00  100.4  3.55  7.70  6  11  1.0  0.8  2.4  10.00  100.9  3.04  5.90  7  11  1.4  1.2  2.4  10.00  100.22  3.01  10.00  8  11  1.4  0.8  2.4  9.50  100.7  3.40  15.00  9d  13  1.2  1.0  2.5  9.75  100.61  3.30  8.70  10d  13  1.2  1.0  2.5  9.75  100.33  3.27  9.30  a RIV content regarding amount declared by manufacturer. b Migration time. c Resolution between IS and RIV electrophoretic peaks. d Central point and nominal microemulsion: 13 mM tetraborate pH 9.75 + 1.2% SDS + 1.0% Ethyl acetate + 2.4% butanol. Figure 4. View largeDownload slide Graphical representation of the effects from changes in selected factors on RIV sample content, migration time (Tm) and resolution between IS and RIV peaks. Method robustness assessed through Plackett–Burman design. Column effect larger than error bar represents a significant response (α = 0.05). Figure 4. View largeDownload slide Graphical representation of the effects from changes in selected factors on RIV sample content, migration time (Tm) and resolution between IS and RIV peaks. Method robustness assessed through Plackett–Burman design. Column effect larger than error bar represents a significant response (α = 0.05). System suitability parameters were accessed on different days of method validation, through Agilent Chemstation software method validation. Average values of RIV migration time (3.36), theoretical plates (19,267), peak symmetry (1.37), and resolution (9.00). Discussion In the CE method development study, several electrolyte solutions were tested in order to achieve the highest sensitivity and shortest analysis time for RIV assessment. Since RIV is a weak amphoteric compound, pKa values (1.01 and 13.36), its full ionization in aqueous media with pH values usually applied in CE (from 3 to 12) is quite limited. Therefore, RIV molecule is mainly non-ionized when applying FSCE conditions and ionic separation is unfeasible. Although MEKC mode can separate adequately non ionized molecules, in the experiments performed the analyses of RIV by MEKC showed results with poor reproducibility. On the other hand, the evaluation of RIV in a microemulsion CE system presented better results (Figure 2B). MEKC micelles present a rigid structure engendered by surfactant and aqueous media, providing a more hostile environment for the analysis of molecules with hydrophobic characteristics such as RIV. Whereas, droplets have a more flexible surface, which allows an easier penetration of solutes and therefore, RIV as a molecule of low water solubility and neutral in a major range of pH (2–12), had its separation by partitioning with the droplet. The influence of different analytical parameters on resolution, electric current generated and peak symmetry were considered by method optimization. Ethyl acetate was chosen as the hydrophobic core for the MEEKC system since it has a lower surface tension than oils such as heptane and octane. Thus, the concentration of surfactant in the solution can be reduced, which enables the application of a higher voltage (up to 30 kV) across the capillary, without producing undesirable current levels (above 100 uA). The major reason for the high current generated (resulting of the high-ionic strength) is the relatively high concentrations of surfactants needed to form the microemulsion when using heptane or octane (3, 19). Application of IS is highly recommended in CE quantitative analysis to correct intrinsic errors caused by variability in pressure injection mode, voltage and EOF (22). An ideal IS presents good absorption at a selected wavelength; IS peak must have a proper resolution in relation to analyte and DPs (R ≥ 2), and should not increase analysis time. Working with an IS in CE analysis increases the precision and accuracy of quantitative measurements, therefore, we selected an IS for MEEKC analyses. As presented in the method validation results, it was possible to develop a reliable MEEKC system to quantify RIV in pharmaceutical form and also to observe the existence of DP among samples. All validation parameters are in accordance with international guidelines. Plackett–Burman experimental design was chosen for robustness assessment due to its fast application and the possibility of evaluating the effect of many factors at once. The goodness of fit test was evaluated by an F-test of sum of squares for lack of fit and pure error, assuring the fitness of the model to each response. Effects plot is a graphical representation of the magnitude and significance of evaluated factors on selected responses; non significant variations were found in RIV assay for sample content. Significant changes in IS and RIV resolution do not affect RIV quantification and method robustness since this value was never smaller than 3.0, as presented in Table IV. Those results ensure the capability of the method to remain reliable after undergoing minor variations. System suitability parameters demonstrate the qualification of the electrophoretic system for pharmaceutical analysis. Concluding remarks A stability indicating method based on MEEKC was developed for quantitative determination of RIV in a pharmaceutical formulation. Method performance in terms of accuracy and precision with analysis time of fewer than 4 min demonstrates that microemulsion systems can be successfully applied in drug analysis. According to statistical analysis, the method developed was found to be sensitive, specific, robust and linear for the intended purpose. Therefore, the MEEKC method with PDA detection can be applied as a reliable and simple technique in quality control of RIV tablets, and an analytical alternative with reduced environmental damage and low financing costs. Conflict of interest statement Authors declare the absence of conflict of interest. References 1 Altria, K., Marsh, A., Griend, C.S.; Capillary electrophoresis for the analysis of small-molecule pharmaceuticals; Electrophoresis , ( 2006); 27: 2263– 2282. Google Scholar CrossRef Search ADS PubMed  2 Holland, L.A., Chetwyn, N.P., Perkins, M.D., Lunte, S.M.; Capillary electrophoresis in pharmaceutical analysis; Pharmaceutical Research , ( 1997); 14: 372– 388. Google Scholar CrossRef Search ADS PubMed  3 Altria, K.D.; Background theory and applications of microemulsion electrokinetic chromatography; Journal of Chromatography A , ( 2000); 892: 171– 186. Google Scholar CrossRef Search ADS PubMed  4 Ryan, R., Donegan, S., Power, J., Altria, K.; Advances in the theory and application of MEEKC; Electrophoresis , ( 2010); 31: 755– 767. Google Scholar CrossRef Search ADS PubMed  5 McBride, B.F.; Preliminary assessment of the critical differences between novel oral anticoagulants currently in development; Journal of Clinical Pharmacology , ( 2005); 45: 1004– 1017. Google Scholar CrossRef Search ADS PubMed  6 Çelebier, M., Reçber, T., Koçak, E., Altinöz, S.; RP-HPLC method development and validation for estimation of rivaroxaban in pharmaceutical dosage forms; Brazilian Journal of Pharmaceutical Sciences , ( 2013); 6: 359– 366. Google Scholar CrossRef Search ADS   7 Wingert, N.R., Nunes, M.A.G., Barden, A.T., Gomes, P., Muller, E.I., Flores, E.M.M., et al.  .; Ultra-performance LC-ESI/Q-TOF MS for the rapid analysis of rivaroxaban: method validation using experimental design for robustness evaluation; Current Analytical Chemistry , ( 2015); 11: 124– 129. Google Scholar CrossRef Search ADS   8 ICH. Harmonised Tripartite Guideline: Validation of Analytical Procedure: text and methodology Q2 (R1). Geneve, Switzerland ( 2005). 9 ICH. Harmonised Tripartite Guideline: Stability Testing of New Drug Substance and Products: Methodology Q1A (R2), ( 2003). 10 Rowe, R.C., Sheskey, P.J., Quinn, M.E. (eds.); Handbook of Pharmaceutical Excipients , 6th ed. Pharmaceutical Press, Berlin, Germany, ( 2009). 11 USP. The United States Pharmacopeia 37th Rockville, MD ( 2014). 12 Altria, K.D. (ed).; Capillary electrophoresis guidebook , Vol. 52. Human Press, New Jersey, ( 1996). 13 Mahuzier, P.E., Prado, A., Clark, B.J., Kedor-Hackmann, E.R.M., Altria, K.D. An introduction to the theory and application of microemulsion electrokinetic chromatography. CE Currents, ( 2003). 14 Wen, T., Zhao, X., Luo, G., Wang, J., Wang, Y., Yao, B.; Comparison of microemulsion electrokinetic chromatography and solvent modified micellar electrokinetic chromatography on rapid separation of heroin, amphetamine and their basic impurities; Talanta , ( 2007); 71: 854– 860. Google Scholar CrossRef Search ADS PubMed  15 Lynen, F., Saavedra, L., Nickerson, B., Sandra, P.; Evaluation of a multiarray system for pharmaceutical analysis by microemulsion electrokinetic chromatography; Talanta , ( 2001); 84: 724– 729. Google Scholar CrossRef Search ADS   16 Nussbaumer, S., Fleury-Souverain, S., Schappler, J., Rudaz, S., Veuthey, J.-L., Bonnabry, P.; Quality control of pharmaceutical formulations containing cisplatin, carboplatin, and oxaliplatin by micellar and microemulsion electrokinetic chromatography (MEKC, MEEKC); Journal of Pharmaceutical and Biomedical Analysis , ( 2011); 55: 253– 258. Google Scholar CrossRef Search ADS PubMed  17 Miola, M.F., Snowden, M.J., Altria, K.D.; The use of microemulsion electrokinetic chromatography in pharmaceutical analysis; Journal of Pharmaceutical and Biomedical Analysis , ( 1998); 18: 785– 797. Google Scholar CrossRef Search ADS PubMed  18 Yu, L., Chu, K., Ye, H., Liu, X., Yu, L., Xu, X., et al.  .; Recent advances in microemulsion electrokinetic chromatography; Trends in Analytical Chemistry , ( 2012); 34: 140– 151. Google Scholar CrossRef Search ADS   19 Yang, H., Ding, Y., Cao, J., Li, P.; Twenty-one years of microemulsion electrokinetic chromatography (1991–2012): a powerful analytical tool; Electrophoresis , ( 2013); 34: 1273– 1294. Google Scholar CrossRef Search ADS PubMed  20 Altria, K.D., Elder, D.; Overview of the status and applications of capillary electrophoresis to the analysis of small molecules; Journal of Chromatography A , ( 2004); 1023: 1– 14. Google Scholar CrossRef Search ADS PubMed  21 Montgomery, D.C. (ed).; Design and analysis of experiments , 5th ed. John Willey & Sons, Inc, USA, ( 2001); pp. 177– 180. 22 Jouyban, A., Kenndler, E.; Impurity analysis of pharmaceuticals using capillary electromigration methods; Electrophoresis , ( 2008); 29: 3531– 3551. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Journal of Chromatographic ScienceOxford University Press

Published: Apr 18, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off