Anal. Chem. 2006, 78, 7891-7895
Correspondence
Major Degradation Product Identified in Several Pharmaceutical Formulations against the Common Cold J. Wong, L. Wiseman, S. Al-Mamoon, T. Cooper, L.-K. Zhang, and T.-M. Chan*
Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, New Jersey 07033, Schering-Plough HealthCare Products, 3030 Jackson Avenue, Memphis, Tennessee 38151
Different pharmaceutical preparations against the common cold contain acetaminophen, phenylephrine hydrochloride, and chlorpheniramine maleate. A degradation product had been discovered in these preparations after short- and long-term stability studies. This degradation product was isolated and found to be an adduct of phenylephrine and maleic acid. An account of the isolation and characterization of this compound was published. Our interest in this area led us to synthesize the compound, and we found that the synthesized compound does not have the same spectroscopic properties described in the original paper. Our subsequent work identified the structure of the degradation product as a “Michael addition” product of phenylephrine and maleic acid. Phenylephrine is a decongestant1 used in commercial drug products as a replacement for pseudoephedrine and is being evaluated by various drug manufacturers or has already been marketed. Its most common form is the hydrochloride salt. The combination of phenylephrine with other active ingredients, namely, chlorpheniramine maleate and acetaminophen, has been reported to produce a small amount of a degradant with a molecular mass of 283. This degradant has been reported2 as being the reaction product of phenylephrine and maleic acid and was assigned a structure of 3-hydroxy-N-[2-hydroxy-2-(3hydroxyphenyl)ethyl]-N-methylsuccinamic acid (compound II, Figure 1). During pharmaceutical development3 in which phenylephrine HCl and dexbrompheniramine maleate were combined one at a time with various common tablet excipients, we observed a degradant with the same molecular mass of 283, similar HPLC profile (RRT), and UV spectrum. Compound II (3-hydroxy-N-[2hydroxy-2-(3-hydroxyphenyl)ethyl]-N-methylsuccinamic acid) was synthesized and found to have NMR spectra that are different * To whom correspondence should be addressed. E-mail: tze-ming.chan@ spcorp.com. (1) 21 CFR Part 341. (2) Marin, A.; Espada, A.; Vidal, P.; Barbas, C. Anal. Chem. 2005, 77, 471477. (3) International Conference on Harmonization (ICH) Q8: Guidance for Industry. Pharmaceutical Development. 10.1021/ac0611263 CCC: $33.50 Published on Web 09/27/2006
© 2006 American Chemical Society
from those reported in the literature.2 Another possible compound that has the same molecular formula was then synthesized, and its NMR spectra matched those reported in the literature. We report that the correct structure of the degradant should be 2-[2hydroxy-2-(3-hydroxyphenyl)ethyl]methylaminosuccinic acid (compound III, Figure 1). EXPERIMENTAL SECTION Materials. Phenylephrine HCl was from Boehringer Ingelheim (Ingelheim, Germany). Excipient blends were from ScheringPlough Healthcare Products (Memphis, TN). Acetonitrile and water were Optima grade from Fisher Scientific (Pittsburgh, PA). Sodium perchlorate, (99%) was from Acros Organics/Fisher Scientific (Pittsburgh, PA). Perchloric acid (70% ACS grade) was from Acros Organics/Fisher Scientific (Pittsburgh, PA). Synthesis. Compounds II and III were synthesized in Schering-Plough Research Institute. The compounds were synthesized from I in reactions shown in Figure 1. Details of the syntheses will be published in the future. High-Performance Liquid Chromatography (HPLC). The HPLC system consisted of an Agilent Technologies (Palo Alto, CA) 1100 quaternary pump with a photodiode array detector set at 215, 265, and 275 nm. HPLC conditions were very similar to those described in the paper by Marin and co-workers2 (method 1). The analytical column was a Discovery HS PEG column 4.6 × 150 mm (5 µm) (Supelco, Bellefonte, PA) with the column temperature set to 35 °C. The flow rate was 1.0 mL/min and injection volume 10 µL. The mobile phase was sodium phosphate buffer 20 mM at pH 7.0/actonitrile 90:10 (v/v). For better resolution, another HPLC method (method 2) was used. The analytical column was Zorbax SB-CN 4.6 × 250 mm (5 µm) (Agilent) with the column temperature set to 25 °C. The flow rate was 1.0 mL/min and injection volume 10 µL. The mobile phase consisted of (A) 0.1% (v/v) perchloric acid with 100 mM sodium perchlorate and (B) acetonitrile. Mobile phases were filtered through 0.45-µm membrane and degassed before use. A 4-min initial isocratic (94:6 A/B) run was followed by a 16-min gradient with increasing percent B to elute the peaks (10:90 A/B), followed by regeneration of the starting conditions. Analytical Chemistry, Vol. 78, No. 22, November 15, 2006 7891
Figure 1.
Figure 2. (1) Standard solution, (2) solution of stressed active and excipient, (3) solution of stressed active and excipient with III added, and (4) solution of stressed active and excipient with II added. 7892
Analytical Chemistry, Vol. 78, No. 22, November 15, 2006
Figure 3.
Figure 4.
Excipient Compatibility/Stability Study. Two sets of active/ excipient mixes were made. One set consisted of heating at 60 °C for 1 month. The second set consisted of the identical components as the first except that ∼10% (w/w) water was added followed by heating to 60 °C for 1 month. Both sets of mixes contained the actives phenylephrine HCl and dexbrompheniramine maleate along with a single tablet excipient. Both sets of active/excipient mixes (with and without water) produced multiple degradants, and samples were analyzed by HPLC. HPLC Sample Preparation. The concentration of phenylephrine HCl in standard and sample solutions was ∼0.5 mg/mL in diluent. The dilute used was water/acetonitrile 80:20 (v/v). Excipient mixes were added quantitatively into volumetric flasks with the aid of some acetonitrile followed by dilution with diluent to a phenylephrine concentration of 0.5 mg/mL. Filtration, centrifugation, or both were performed as necessary to obtain a clear sample solution for analysis. RESULTS Both sets of active/excipient mixes (with and without water) produced multiple degradants. The addition of water to the active/
excipient mix generally resulted in higher amounts of degradation products. Among them was one with molecular mass of 283 (referred as the unknown in the rest of this paper). The amount of this particular degradation product produced over all active/ excipient sets made was 1.5-11.7%, by area, relative to phenylephrine HCl standard response. Analysis. Using the HPLC conditions in the paper of Marin and co-workers (method 1), the unknown degradation product has a RRT of 0.75 (phenylephrine, RRT ) 1.0). Compound III has the same RRT, while II has a RRT of 0.79. For better resolution, a different column (method 2) was used. Figure 2 presents an overlay of four chromatograms as analyzed by method 2. HPLC analysis of the stressed active/excipient mix produced an unknown peak at RRT 0.90 (phenylephrine RRT, 1.0). Spiking the solution with a solution of III resulted in an increased response of the peak (chromatogram 3, Figure 2). For comparison, a solution of stressed active/excipient was spiked with a solution of II (chromatogram 4, Figure 2). Spiking the stressed active/excipient solution with III matches the retention time of the unknown peak and the response is enhanced. Spiking Analytical Chemistry, Vol. 78, No. 22, November 15, 2006
7893
Table 1. Assignments of the Proton and Carbon NMR Resonances for Compounds II and IIIa compound II 13C
1 2 3 4 1′ 2′ 3′ 4′ 5′ 6′ 7′ 8′ 1′′
δ (ppm)
172.55, 172.3, 172.2 39.2, 39.25, 39.3 64.8, 65.0, 65.2, 65.4 172.15, 172.1, 171.9 56.2, 56.4, 56.5 70.5, 70.7, 70.75, 70.9 144.75, 145.0, 145.3, 145.4 112.7, 112.8, 112.9 157.2, 157.25, 157.3, 157.4 114.0, 114.1, 114.2 129.0, 129.1, 129.2 116.5, 116.55, 116.6 34.1, 34.3, 36.4, 36.5
compound III 1H
13C
δ (ppm)
2.33 2.54, 2.33 2.60, 2.39, 2.67 4.60, 4.69, 4.75 3.16 3.49, 3.32, 3.22 3.70, 3.43 4.62, 4.64, 4.71, 4.73 6.76, 6.79, 6.81 6.62, 6.64 7.09, 7.11, 7.12 6.73, 6.77, 6.78 2.78, 2.85, 2.96, 3.00
δ (ppm)
171.3/174,4, 174.7 32.5, 33.1 66.0, 66.8 171.3/174.4, 174.7 62.1, 62.6 69.0, 69.6 143.8, 143.9 113.9, 114.0 159.1 116.4 130.9 118.1, 118.2 39.1, 39.5
1H
δ (ppm)
2.90 3.09, 2.92 3.10 4.27, 4.31 3.21 3.42, 3.24 3.46 5.04 6.88 6.72 7.11 6.88 2.95, 2.99
a NMR spectra were obtained on a Varian Inova 500-MHz spectrometer at 25 °C. Spectra of II were obtained in DMSO-d , and those of III were 6 obtained in methanol-d4.
Figure 5. Proton NMR spectrum of III, obtained on a Varian Inova 500-MHz spectrometer. Spectrum was obtained in D2O. Since the resonances were sensitive to the pH of the solution, a small amount of NaOD was added to the D2O solution until the spectrum was nearly identical to that reported by Marin and co-workers.
the solution with II does not match the retention time of the unknown peak. The UV spectra of the unknown peak and of III are given in Figure 3. An overlay of the spectra is presented in Figure 4. Characterization of Compounds II and III. The compounds were characterized by mass spectrometry and NMR spectroscopy. In their mass spectrum, both compounds give a M + H+ ion of 284, and the high-resolution mass spectrometry confirms a molecular formula of C13H17NO6 for both compounds. Structures of both compounds were determined by 2D NMR. Since the structures contain an additional chiral center, both synthetic samples are a mixture of two diastereomers, and two sets of peaks were observed in NMR. All proton and carbon resonances in the spectra were assigned, and the connectivities between proton and carbon nuclei were confirmed by 2D NMR 7894 Analytical Chemistry, Vol. 78, No. 22, November 15, 2006
techniques such as HSQC, HSQCTOCSY, and HMBC. The proton and carbon assignments are summarized in Table 1. The NMR spectra of II are more complicated; each diastereomer in the sample also exist as two rotomers due to the restricted rotation around the NC4 amide bond. Therefore, there are four sets of peaks in the NMR spectra due to the four rotomers. In the HMBC spectrum, correlation was observed between the N-methyl proton (H1′′) and the methylene carbon C1′, and between H1′′ and C4 (172 ppm). No HMBC correlation was observed between H1′′ and C3 as expected since they are separated by four bonds. In the HMBC spectrum of III, correlation was observed between the N-methyl proton (H1′′) and the methylene carbon C1′, and between H1′′ and C3, which is the same as observed by Marin and co-workers.2 No HMBC correlation was observed between H1′′ and the carbonyl carbons (160-170 ppm) as expected. NMR spectra (proton, carbon, HSQC, and HMBC spectra) of III obtained in D2O were nearly identical to those reported by Marin and co-workers.2 The proton NMR spectrum shown in Figure 5 was nearly identical to the one reported by Marin and co-workers.2 CONCLUSIONS Two new compounds were synthesized and fully characterized by MS and NMR. These compounds are potential candidates for a degradation product in pharmaceutical formulation against the common cold containing phenylephrine and maleates (chlorpheniramine maleate or dexbrompheniramine maleate). In the literature,2 this degradation product was assigned the structure of 3-hydroxy-N-[2-hydroxy-2-(3-hydroxyphenyl)ethyl]-N-methylsuccinamic acid. The syntheses of these two compounds allow us to conclude that this degradation product was misidentified, and its structure should be 2-[2-hydroxy-2-(3-hydroxyphenyl)ethyl]methylaminosuccinic acid. It is a “Michael addition” product between phenylephrine and maleic acid, which should be readily formed under fairly mild conditions such as the stressed conditions used for the excipient compatibility and stability studies.
ACKNOWLEDGMENT We thank Dr. B. Pramanik, Dr. P. McNamara, and Dr. M. Senior for their helpful discussions, Dr. Mohammed Kabir for providing formulation samples, and Ms H. Yun for library research.
Received for review June 21, 2006. Accepted September 11, 2006. AC0611263
Analytical Chemistry, Vol. 78, No. 22, November 15, 2006
7895
