Gradient RP-HPLC method for the determination of potential impurities in dabigatran etexilate in bulk drug and capsule formulations

2.1 Chemicals and reagents

HPLC grade acetonitrile, ammonium formate and formic acid were obtained from Merck, Darmstadt, Germany. Triethylamine was purchased from J.T.baker, Mallinckrodt Inc., Phillipsburg, NJ, USA. Water used for preparing mobile phase was purified using Millipore Milli-Q plus purification system. DAB and process related substances were gift samples from a local manufacturing company in Chennai, India.

2.2 Apparatus and conditions

HPLC: The system used for method development, forced degradation and validation study was Agilent HPLC 1200 series (Agilent Technologies, Deutschland GmbH, Waldbronn, Germany) with Ezchrom elite software equipped with G1322A degasser, G1311A quaternary pump, G1315C auto injector, G16A column oven and G1315C DAD detector.

The HPLC columns used in this study were as follows:

• Inertsil ODS 3V, 250 mm × 4.6 mm, 5 μm (GL Sciences, Japan).

• Inertsil C8, 250 mm × 4.6 mm, 5 μm (GL Sciences, Japan).

• ACE C18, 250 mm × 4.6 mm, 5 μm (Advanced Chromatography, Scotland).

• Zorbax XDB-C18, 250 mm × 4.6 mm, 5 μm (Agilent technologies Inc., USA).

• Zorbax XDB-C8, 250 mm × 4.6 mm, 5 μm (Agilent technologies Inc., USA).

2.3 Chromatographic conditions

The stationary phase used was Inertsil ODS-3V, 250 mm × 4.6 mm, 5 μm particles (GL Sciences Inc., Japan). The mobile phase-A contains 20 mM of ammonium formate with 0.1% of triethylamine (pH: 5.0 adjusted with formic acid) and mobile phase-B contains acetonitrile. The flow rate was 1.0 mL min⁻¹. The developed gradient program was 0.01 min – 30% B, 5.0 min – 30% B, 15.0 min – 35% B, 30.0 min – 45% B, 45.0 min – 55% B, 54.0 min – 65.0% B, 55.0 min – 30.0% B and 60.0 min – 30.0% B. The detection was performed at 255 nm using a DAD detector. The injection volume was 10 μL. The column temperature was 30 °C.

2.4 LC-MS conditions

The LC-ESI/MS studies were carried out on Q-TOF (Agilent, USA) 4100 Quadrupole mass spectrometer, ion source voltage 5500 V, de clustering potential 40 V, entrance potential 10 V, with the nebulizer gas as nitrogen at 40 psi. The LC part consisted of Agilent 1260 series HPLC system with binary gradient pump with a degasser and an autosampler. Inertsil ODS-3V, 250 mm × 4.6 mm, 5 μm particle size was used for chromatographic separation. Mobile phase-A as 20 mM ammonium formate (pH 5.0 with formic acid) and mobile phase-B as acetonitrile was used. The flow rate was maintained at 1.0 mL min⁻¹ and the injection volume is 10 μL. The gradient program was 0.01 min – 30% B, 5.0 min – 30% B, 15.0 min – 35% B, 30.0 min – 45% B, 45.0 min – 55% B, 54.0 min – 65.0% B, 55.0 min – 30.0% B and 60.0 min – 30.0% B. The column temperature was 30 °C.

2.5 Preparative HPLC

A Shimadzu LC-8A Preparative LC equipped with SPD-10 AVP, UV–VIS detector (Shimadzu Corporation, Analytical Instruments Division, Kyoto, Japan) was used. ACE C18 (250 mm × 21.2 mm, particle size 10 μm) columns were employed for separation and isolation of impurity using ammonium formate (20 mM) as mobile phase-A and ACN as mobile phase-B with varying compositions and flow rate of 15–20 mL min⁻¹ and detection was carried out at 255 nm.

2.6 NMR spectroscopy

¹H and ¹³C NMR spectra were acquired with AVANCE (Bruker, Switzerland) 300 and 400 MHz spectrometers at room temperature, 22 °C using DMSO-d6 as solvent and tetramethylsilane (TMS) as internal standard. The data acquisition and processing was carried out by XWIN-NMR software.

2.7 FT-IR spectroscopy

The FT-IR spectra were recorded as KBr pellets using PerkinElmer Spectrum-2 FT-IR Spectrometer (PerkinElmer Life and Analytical Sciences, Shelton, CT, USA).

2.8 Differential scanning calorimeter (DSC)

The DSC Q200 (TA Instruments, Waters LLC) was used for melting point determination.

2.9 Preparation of standard solutions

Working solutions containing 1000 μg mL⁻¹ DAB and 100 μg mL⁻¹ process related substance stocks were prepared for the determination of related substances. Mixed and individual stock solutions (100 μg mL⁻¹) of the impurities (denoted as impurity-1 to impurity-15) were diluted to get a concentration about 1.5 μg mL⁻¹ in diluent, which is in the ratio of 70:30 (mobile phase-A:acetonitrile). The specification limit considered for validation studies was 0.15% for each impurity. Pradaxa (150 mg) capsule powder solution equivalent to 1000 μg mL⁻¹ of the drug was prepared by dissolving in the diluent and was filtered through a 0.45 μm nylon membrane filter. The resulting solution was analyzed by HPLC for the estimation of related substances.

3.1 Synthetic process and identification of impurities

Method development was initiated by using crude samples at various stages of synthetic process. A brief synthetic scheme along with origination of by-products during the synthesis is detailed in Fig. 1, where condensation of key starting materials compound_1 and compound_2 gives a salt form of compound_3, which is subsequently reacted to ethanolic hydrochloric acid, ethanol and base to give the salt of ethyl 3-{[(2-{[(4-carbamimidoylphenyl)amino]methyl}-1-methyl-1H-benzimidazol-5-yl)carbonyl](pyridin-2-yl)amino}propanoate (compound_4). The reaction between compound_4 and n-hexyl chloroformate, in the presence of potassium carbonate and solvent gives dabigatran etexilate, finally treating dabigatran etexilate with methane sulfonic acid in the presence of solvent to give dabigatran etexilate mesylate salt. The mesylate salt (1:1) of dabigatran etexilate is known to be therapeutically useful and is commercially marketed.

As a common practice, efforts should be made to identify and characterize all unknown impurities in the drug substance due to the ever increasing demand from regulatory agencies to manufacture high purity drug substances. Impurity profiling of drugs is the most important issue in modern pharmaceutical analysis (Rao and Nagaraju, 2003; Görög, 2000; Ahuja, 1998) for developing process technology to manufacture high purity drug substance. Hence, the method should be capable of separating all potential related impurities, including starting materials, by-products and degradants.

3.2 Optimization of chromatographic conditions

Drug impurity profiling, i.e., identification, structure elucidation and quantitative determination of impurities and degradation products in bulk drug materials and pharmaceutical formulations is one of the most important fields of activities in modern pharmaceutical analysis (Bartos and Gorog, 2008). The main objective of this chromatographic method is to get the separation of DAB from all the process related impurities, by-products and degradants. The knowledge of drug properties such as log P and pK_a was important, as most of the pH-related changes in retention occur at pH values within 1.5 units of the pK_a value and log P gives an indication about hydrophobicity or polarity nature of the molecule. The selectivity of a separation also depends upon the chemistry of the analyte i.e. the impurity-4 was depyridynyl form of impurity-3 and impurity-11 and 12 were chain isomers of hexyl group present on impurity-10 and impurity-13 and impurity-14 were differed by one methyl group. The main purpose of the method development was to optimize selectivity of these critical closely eluting pair of compounds by altering mobile and stationary phases.

The blend containing 1000 μg mL⁻¹ of DAB and 1.5 μg mL⁻¹ of each impurity was prepared and used for method optimization. To start the method optimization the effect of pH (from 4.0 to 6.0) of ammonium formate buffer was studied to understand the elution and separation of impurities. To improve peak shapes 0.1%TEA was added as an additive to the buffer. Preliminary experiments on method development were carried out on Inertsil ODS 3V column using pH 5.5–20 mM ammonium formate with 0.1% of triethylamine:ACN (60:40) (v/v) in isocratic mode. Six impurities were merged and the same kind of separation was observed by changing organic modifier as methanol. It was found that gradient elution resulted better separation than isocratic mode of elution on this column with acetonitrile as organic modifier but only the critical pair’s viz. impurity-11 and impurity-12, impurity-13 and impurity-14 separation was not observed. Attempts were made with gradient elution to separate the critical pairs changing different column chemistries but no selectivity was observed. The best results were obtained using the chromatographic conditions given in Section 2.3 and the typical chromatogram is shown in Fig. 2(a). The method development trials are shown in Table 2.

Close

Figure 2 Typical chromatograms of (a) Dabigatran etexilate spiked with 0.15% of all impurities in final chromatographic conditions (pH 5.0), (b) bulk drug stability sample containing unknown impurities, and (c) Pradaxa capsule test sample.

Export to PPT

The affinities of all impurities toward different commercially available columns were also evaluated to get the best resolution between critical pair of impurities. Inertsil ODS 3V column at pH 5.0 was selected to get the good resolution between the critical pairs, which was due to its high surface area, more hydrophobicity, and high reproducibility compared with other commercially available columns. In this optimized method all the peaks were symmetrical and show the tailing factor ⩽1.3 and the resolution ⩾2.0 from blend chromatogram of system suitability. System suitability parameters are described in Table 1.

3.3 Analysis of bulk drugs stability samples

During the analysis of active pharma ingredient (API) stability sample, three impurities were observed, of which one was imp-8 and other two were unknown. Unknown were eluted around retention time 43.3 min (unknown-1) and 24.4 min (unknown-2) (Fig. 2(b)). However, they did not match with any of the process related substances by comparing retention times. The compounds imp-8, unknown-1(imp-15) and unknown-2 were present in 0.19%, 0.26% and 0.09% respectively and remaining impurities were not detected. Attempts were made to evaluate these unknown impurities in stability sample using LC-MS/MS analysis using the conditions given in Section 2.5. The obtained fragmentation pattern for unknown-1 and unknown-2 was comparable and structures were derived with the help of known DAB and impurity-8 mass spectrums interpretation. Two unknown impurities comparison with DAB and impurity-8 detailed mass fragmentation shown in Figs. 5 and 6.

3.4 Degradation studies

Degradation study gives information about origin of degradant impurities and also useful to check the stability-indicating power of developed method. The forced degradation condition was attempted to get degradation not less than 10% (in order to get a mass balance between purity and assay). Solution state forced degradation of DAB was carried out by treating the 1.0 mg/ml solutions with 0.1 M HCl for 5 min at 75 °C, 0.25 M NaOH for 30 min at 75 °C and 3% H₂O₂ for 1 h at room temperature. Solid state forced degradation (thermal, photolytic and humidity) of the drug was conducted by exposing DAB to (1) heat at 105 °C for 120 h, (2) white fluorescent light of 1.2 × 10⁶ lux followed by UV light of 200 W h/m² and (3) 90% R.H. at 25 °C for 120 h. All these solutions are analyzed by the developed method conditions mentioned in Sections 2.3 and 2.4. One degradant which was observed in bulk drug stability sample i.e. unknown-1 was detected at 0.6% level in humidity condition and no forced degradation was observed in heat and exposure with light. Peak purity test was carried out on the stressed samples of DAB by using DAD detector. Assay studies were carried out for stress samples against qualified reference standard and the mass balance (% assay + % degradation) is calculated (Table 4). Stress chromatograms are illustrated in Fig. 3(a)–(d).

Close

Figure 3 Typical chromatograms of forced degradation of dabigatran etexilate at control condition (a) acid hydrolysis, (b) base hydrolysis, (c) peroxide stress, and (d) humidity stress.

Export to PPT

3.5 Enrichment of unknown-1 and unknown-2 impurities

Degradation in controlled conditions (i.e. ⩽10%) gives a conclusion that unknown-1 was increasing in humidity stress. Based on mass spectral data and degradation studies attempts were made to enrich the impurity using forced degradation with water. The unknown-1 impurity increased about 9% in water stress at 75 °C for 30 min and about 88% when extending the stress time to 12 h (Fig. 4(a) and (b)). Intentional addition of 0.1 N HCl to 88% degraded water stress sample and holding for 5 min give a rise of unknown-2 impurity (Fig. 4(c)). The unknown-2 impurity formation was also observed in intentional rigorous stress of DAB using 1 M HCl (Fig 4(d)). The stress solutions were injected into LC-MS/MS analysis to generate unknown impurities (unknown-1 and 2) mass spectral data and these were matching with unknown impurity mass spectrums of bulk drug stability sample.

Close

Figure 4 Typical chromatograms of (a) water stress of dabigatran etexilate, (b) rigorous water stress of dabigatran etexilate, (c) rigorous water hydrolysis followed by acid hydrolysis for 5 min of dabigatran etexilate, and (d) rigorous acid hydrolysis of dabigatran etexilate.

Export to PPT

3.6 Isolation of unknown-1 impurity using preparative LC

In order to isolate the unknown-1 impurity from water stress of DAB, the isolation conditions for preparative LC used were as mentioned in Section 2.5. Approximately 0.2 g of sample was injected on to column every time and this loading was continued until the sufficient amount of unknown-1 was isolated (∼20 injections). The fractions containing >95% of unknown-1 were collected, concentrated using rotary evaporator at 30 °C under vacuum and finally lyophilized to obtain solid product.

3.7 Degradation pathway for unknown-1 and unknown-2

During the experiment-enrichment of unknown impurities, it was observed that intentional addition of acid to water stress of dabigatran etexilate gives rise of unknown-2 impurity (Fig. 4(c)). Based on stress study and mass spectrum fragmentation data the degradation pathway for unknown-1 and unknown-2 is proposed as shown in Fig. 7.

3.8 Structural elucidation of unknown-1

In the positive mode ESI-MS analysis of the unknown-1 exhibited a protonated molecular ion peak m/z 629. The difference of mass between impurity and DAB is 1 amu, indicating deletion and addition on DAB. During degradation studies this was major impurity observed in water hydrolysis of DAB. The imine group (—C⚌NH) on DAB hydrolysis to carbonyl (—C⚌O) leads to increase of 1 amu mass for impurity. This was evidenced by major product ions observed from ESI-MS/MS of DAB m/z 526,500, 434, 350, 332 and 306 and of unknown-1(impurity-15) were at m/z 527,435, 351,333 and 308 (Fig. 5). The ¹H, ¹³C NMR and FT-IR spectral data of DAB and unknown-1 impurity are given in Table 3 and Fig. 8 respectively. ¹H NMR spectral data of unknown-1 showed chemical shift similarity with DAB in all positions except the signals corresponding to —NH₂ in DAB was absent in impurity. From ¹³C spectral data the change in chemical shift values at eighth position of carbon gives an evidence of hydrolysis of imine group (—C⚌NH) on DAB to a carbonyl (—C⚌O). From this spectral information the structure was confirmed as ethyl 3-{[(2-{[(4-{[({hexyloxy}carbonyl)amino]carbonyl}phenyl)amino]methyl}-1-methyl-1H-benzimidazol-5-yl)carbonyl](pyridin-2-yl)amino}propanoate. Determined the melting point by DSC for unknown-1 and DAB, the results are 185 °C and 178 °C respectively.

Close

Figure 5 ESI-MS/MS fragmentation patterns for dabigatran etexilate and unknown-1 impurity.

Export to PPT

Close

Figure 6 ESI-MS/MS fragmentation patterns for Impurity-8 and unknown-2 impurity.

Export to PPT

Close

Figure 7 Proposed degradation pathway for unknown-1 and unknown-2 impurity.

Export to PPT

Close

Figure 8 FT-IR spectrums of (a) dabigatran etexilate and (b) unknown-1 (Impurity-15).

Export to PPT

3.9 Structural elucidation of unknown-2

The impurity unknown-2 exhibited a protonated molecular ion peak at m/z 601 in positive ion ESI-MS analysis; the difference of mass between DAB and impurity was 27 amu and difference of mass between unknown-1 and this impurity was 28 amu, indicating absence of —CH₂CH₃ i.e. ester hydrolysis of corresponding unknown-1 (Fig. 7). Hydrolysis (ester hydrolysis) of DAB (m/z 628) mainly results in the formation of a compound impurity-8 (m/z 600). Similarly the hydrolysis of unknown-1 (m/z 629) results in a formation of compound unknown-2 having m/z 601. This was evidenced by comparing ESI-MS/MS major product ions of impurity-8 m/z 452, 434 and 332 and of unknown-2 were at m/z 453, 435 and 333 (Fig. 6). From this mass spectrum fragmentation pattern comparison, the Unknown-2 structure was predicted as 3-{[(2-{[(4-{[({hexyloxy}carbonyl)amino]carbonyl}phenyl)amino]methyl}-1-methyl-1H-benzimidazo1-5-yl) carbonyl] (pyridin-2-yl)amino} propanoic acid.

3.10 Method validation