Quantitation and Enantiomeric Determination of Propoxyphene
Using Capillary Zone Electrophoresis

Clay P. Phelan
U.S. Department of Justice
Drug Enforcement Administration
Southwest Laboratory
2815 Scott Street
Vista, CA 92081
[Email Address: Clay.P.Phelan -at- usdoj.gov ]

ABSTRACT: Validated Methods for the quantitation of d-propoxyphene HCl and d-propoxyphene napsylate were developed using capillary zone electrophoresis, using an uncoated capillary, a lithium phosphate buffer, and using thiamine HCl as the internal standard. The addition of a small amount of acetonitrile to the injection solvent facilitated the solubilization of d-propoxyphene napsylate. The analytes’ responses were reproducible, provided accurate recovery values, and were linear within the experimental concentration range. A chiral analysis was also conducted, using the same capillary but with 2-hydroxypropyl-β-cyclodextrin added to the run buffer. The methods were specifically developed for the analysis of pharmaceutical tablets containing d-propoxyphene HCl or d-propoxyphene napsylate, which typically are adulterated only with caffeine, aspirin, and/or acetaminophen; however, the method is applicable to analysis of a wide variety of other drugs.

KEYWORDS: Capillary Zone Electrophoresis, CZE, d-Propoxyphene HCl, d-Propoxyphene Napsylate, Chiral Analysis, Forensic Chemistry

Introduction

d-Propoxyphene is a mild narcotic analgesic found in various pharmaceutical preparations, usually as the hydrochloride or napsylate salt. These preparations typically also contain large amounts of acetaminophen, aspirin, or caffeine. This drug is prescribed for pain relief; however, it is also abused for its euphoric side effects [1], and it is therefore commonly diverted into the illicit drug trade. Currently, d-propoxyphene is a Schedule IV controlled substance in the United States; however l-propoxyphene is not controlled. This requires enantiomeric determination for all samples containing propoxyphene.

Propoxyphene is thermally labile, and will break down on a gas chromatograph. Therefore, most of the literature procedures for its analysis are based on liquid chromatographic techniques, more recently including capillary electrophoretic methods [1,2]. The analysis of controlled substances with CE, especially using specialized capillary coatings and/or run buffers, have been shown to produce highly accurate and reproducible results [3-5]. However, some of these techniques are relatively costly and complicated. The described methodologies are simple, inexpensive, and can also be utilized for a wide variety of other drugs, including the phenethylamines [6].

The first CE method for quantitation and enantiomeric determination of propoxyphene was reported in 1994 [7]. However, the methodology required the use of fairly long capillaries, resulting in long analysis times. In addition, because d-propoxyphene napsylate has a solubility limit of approximately 1.0 mg/mL in 0.01 N HCl, the quantitation samples were prepared at concentrations less than 0.6 mg/mL, and had to be sonicated for several hours prior to analysis.

Solubility problems in CE can be addressed by the use of an organic modifier in the injection solvent and/or run buffer. For d-propoxyphene napsylate, acetonitrile was determined to be an appropriate modifier. Thiamine HCl was selected as the method internal standard, as it is commercially available, inexpensive, and not found in typical pharmaceutical preparations or in most illicit drug samples.

Experimental

Preparation of Internal Standard Stock Solution
Thiamine HCl (Sigma, St. Louis, MO) was dissolved in 0.01 N HCl, for a concentration of 1 mg/mL. The deionized water used to produce the 0.01 N HCl was obtained from a Milli Q^® Gradient 10A purification system (Millipore, Bedford, MA).

Preparation of the Achiral Buffer
A 100 mM solution of phosphoric acid was prepared using deionized water and 85%+ reagent grade phosphoric acid (J.T. Baker, Phillipsburg, NJ). The solution was then titrated to a pH of 2.30 ±0.02 with solid lithium hydroxide (Sigma, St. Louis, MO). (Precise pH control is very important in CE, as it affects both migration times and selectivity.) The buffer was filtered prior to use through a 0.45 μm filter, using an Agilent (Wilmington, DE) Solvent Filter/Degasser. Because this buffer contains no preservatives, it was stored at 7 ºC, and was replaced every 6 to 8 weeks.

Preparation of the Chiral Buffer
2-Hydroxypropyl-β-cyclodextrin (Sigma, St. Louis, MO) was added to the achiral buffer such that its concentration was 20 mM. The buffer was filtered prior to use through a 0.45 μm filter.

Preparation of Capillaries
The capillary was prepared in-house, using a 50 μm ± 3 μm ID with a 363 μm ± 10 μm OD flexible polyamide-coated fused silica capillary tubing (Polymicro Technologies, Phoenix, AZ). The capillary was manually cut to a nominal length of 34 cm ± 0.5 cm using a CE column cutter equipped with a diamond blade. Both ends of the capillary were inspected under a microscope to ensure that the glass edge was straight and perpendicular to the length of the capillary tubing, and also was free of debris and defects. The detector window was produced by removing the polyamide coating using a standard window maker equipped with a 7 mm heating module (MICROSOLV^® , Long Branch, NJ). The coating on each end of the capillary was removed using a 2 mm heating module. The new capillary was initially conditioned at 40 ºC by flushing it with 1.0 N NaOH (5 minutes), 0.1 N NaOH (10 minutes), deionized water (5 minutes), and 100 mM lithium phosphate buffer (10 minutes). Subsequently, capillaries were conditioned once every 24 hours at 15 ºC by flushing them with 1.0 N NaOH (1 minutes), 0.1 N NaOH (2 minutes), deionized water (1 minutes), and 100 mM lithium phosphate buffer (2 minutes).

Sample and Standard Preparation for d-Propoxyphene Hydrochloride
d-Propoxyphene HCl standard (Sigma, St. Louis, MO) or a d-propoxyphene HCl-containing sample was accurately weighed and placed in a volumetric flask with an appropriate aliquot of thiamine HCl stock solution (1:5), and the solution was diluted to final volume with 0.01 N HCl. The concentration of the standard or sample in each solution varied between 0.2 - 0.5 mg/mL. Each solution was filtered prior to injection through a syringe equipped with a 0.45 μm filter (Acrodisc^®). For the linearity studies, eight solutions of the d-propoxyphene HCl were prepared at concentrations ranging from 0.05 to 1.3 mg/mL, with the internal standard concentration constant at 0.2 mg/mL. The standard and sample solutions were diluted with 0.01 N HCl to approximately 0.05 to 0.10 mg/mL for the chiral analyses. Enantiomers were determined using the chiral buffer.

Sample and Standard Preparation for d-Propoxyphene Napsylate
d-Propoxyphene napsylate standard or a d-propoxyphene napsylate-containing sample was accurately weighed and placed in a 100 mL volumetric flask with 4 mL acetonitrile and sonicated for approximately 5.0 minutes. An appropriate aliquot of thiamine HCl stock solution (1:5) was added, and the solution was diluted to final volume with 0.01 N HCl. The concentration of the standard or sample in each solution varied between 0.2 - 0.5 mg/mL (the amount of acetonitrile in the final solutions was approximately 4%). Each solution was filtered prior to injection through a syringe equipped with a 0.45 μm filter (Acrodisc^®). For the linearity studies, eight solutions of the d-propoxyphene napsylate were prepared at concentrations from 0.05 to 1.2 mg/mL, containing varying amounts of acetonitrile (0.2% to 4.8%). Additional solutions containing d-propoxyphene napsylate were also prepared from 0.569 to 0.737 mg/mL, containing acetonitrile ranging from 6% to 14%. The appropriate aliquot of thiamine HCl stock solution (1:5) was added, and the solution were diluted to final volume with 0.01 N HCl. The standard and the sample solutions were diluted to approximately 0.05 to 0.10 mg/mL with 0.01 N HCl for the chiral analyses. Enantiomers were determined using the chiral (20 mM 2-hydroxypropyl-β-cyclodextrin) buffer.

Capillary Electrophoresis
Experiments were performed using a Hewlett Packard 3DCE capillary electrophoresis system (Agilent Technologies, Wilmington, DE), equipped with a diode array detector set at a wavelength of 207 nm with a bandwidth of 7 nm. For quantitations, the capillary was flushed with buffer for 2.5 minutes between injections for the HCl, and for 1.0 minutes with 0.1 N NaOH and 2.0 minutes with the achiral buffer for the napsylate. The buffer was replenished after six injections to prevent depletion of electrolytes and charge. The capillary temperature was maintained at 15 ºC. The hydrodynamic injection time was 2.5 seconds at 50 mBar. The applied voltage was 14.5 kV, which was determined empirically to maintain current below 60 μA, thereby limiting Joule heating while also optimizing analysis time.

For the enantiomer determination, the capillary was flushed with buffer for 2.5 minutes between injections for both the HCl and the napsylate salts. The standard and the sample solutions were diluted to approximately 0.2 mg/mL with 0.01 N HCl. The enantiomers were determined using the chiral buffer, using the same applied voltage (14.5 kV). The d,l-propoxyphene standard and sample injection was set at 50 mBar of pressure for 1.5 seconds, followed by a second co-injection of 0.01 N HCl at 20 mBar for 1.0 second. The samples and standards of d-propoxyphene and l-propoxyphene were injected at 50 mBar of pressure for 1.5 seconds, followed by a second co-injection of the d,l-propoxyphene standard at 20 mBar for 1.0 second. The buffer was not replenished after six injections, but rather was utilized until depleted.

Results and Discussion

The objective of the study was to accurately and rapidly quantitate d-propoxyphene HCl and d-propoxyphene napsylate by CZE without interferences from adulterants or diluents. CZE permits direct analysis without requiring extractions. The use of a simple aqueous buffer for the HCl salt reduces analysis cost and allows for simple disposal; the use of acetonitrile as an organic modifier for the napsylate salt is nearly as convenient. Because neutral compounds migrate at the rate comparable to the electroosmotic flow (EOF), while negatively charged (acidic) compounds migrate at a rate slower than the EOF, these species are not detected using the presented method. Therefore, pharmaceutical tablets that contain adulterants such as caffeine, aspirin, or acetaminophen do not interfere. For example, the analysis of a 50 mg d-propoxyphene napsylate tablet containing 325 mg of acetaminophen displays no peak(s) for the acetaminophen (Figure 1).

These methods were specifically developed for pharmaceutical tablets containing d-propoxyphene HCl or d-propoxyphene napsylate, which typically are adulterated only with caffeine, aspirin, and/or acetaminophen. As noted above, these adulterants do not interfere with achiral quantitation; therefore, a selectivity study was not conducted (or required for this application). In the unlikely event that counterfeit pharmaceuticals containing additional adulterants were encountered, the alternative adulterants would have to be identified by spectroscopic means, after which the sample would have to be evaluated to ensure that the selectivity requirements were met, before proceeding with CZE analysis.

The linearity study demonstrated that the calculated errors were less than five percent, and the correlation coefficients were greater than 0.998, within the specified linear range (see Table I). The linearity studies were conducted both with the method using a 1.0 minute flush with the 0.1 N NaOH followed by a 2.0 minute flush of the achiral buffer, and with the method using a 2.5 minute flush with the achiral buffer. It was determined that the use of a 0.1 N NaOH flush resulted in a higher correlation coefficient.

The precision was determined by injecting two concentrations of analyte at the lower and upper ends of the established linear range. The %RSDs for the two concentrations did not exceed 3% for the HCl or the napsylate. Furthermore, an additional five solutions ranging from 0.569 to 0.737 mg/mL of d-propoxyphene napsylate containing varying amounts of acetonitrile (6, 8, 10, 12, or 14%) gave equivalent %RSD values. The precision was determined both with the 1.0 minute flush with the 0.1 N NaOH followed by a 2.0 minute flush of the achiral buffer, and with the 2.5 minute flush with the achiral buffer. Again, it was determined that the use of a 0.1 N NaOH flush resulted in a lower, more consistent %RSDs (Tables II and III).

The accuracy (recovery) was determined by preparing three different concentrations of the analyte with each of the following adulterants: Acetaminophen, aspirin, and caffeine. The concentrations of the analytes represented the lower, middle, and upper linear ranges (i.e., 10%, 50%, and 80%), and contained the appropriate amount of the internal standard. The samples were prepared by sonicating for 5 minutes, and were also compared to non-sonicated samples. The CZE results were compared to the actual values, and did not exceed a 5.0% difference (Table IV). However, samples that were sonicated gave lower errors.

Conclusions

CZE is an effective technique for the quantitation of pharmaceutical preparations containing d-propoxyphene HCl or d-propoxyphene napsylate. Quantitative results were shown to be accurate, reproducible, and precise, and allowed analyses to be accomplished in less than 6 minutes. The use of thiamine HCl as the internal standard was convenient, did not interfere with any known controlled adulterant, and is commercially available at low cost. Chiral separations are conveniently accomplished on the same system with the use of 2-hydroxypropyl-β-cyclodextrin in the run buffer. The described system offers an approach for routine analysis that is simple, robust, practical, and inexpensive. The methodology has been applied to a broader range of illicit drugs, including synthetic opiates and phenethylamines, under the same/similar operating conditions with equal success, and has been used to analyze a large number seized exhibits over the past two years.

Acknowledgments

The author expresses appreciation to the following staff employed at the DEA Southwest Laboratory: Senior Forensic Chemist Harry F. Skinner, and Forensic Chemists Alan Randa, Dean A. Kirby, Nathan Salazar, and Nicole C. Payne-King, for their technical contributions. The author would also like to thank Librarian Rose Mary Russo of the DEA Library (Arlington, Virginia) for her assistance in acquiring references.

References

1. Magoon T, Ota K, Jakubowski J, Nerozzi M, Werner TC. The use of neutral cyclodextrins as additives in capillary electrophoresis for the separation and identification of propoxyphene enantiomers. Analytical and Bioanalytical Chemistry 2002;373:628.

2. Weinberger R. Practical Capillary Electrophoresis, 2nd Edition, Academic Press (2000).

3. Lurie IS, Bethea MJ, McKibben TD, Hays PA, Pellegrini P, Sahai R, Garcia AD, Weinberger R. Use of dynamically coated capillaries for the routine analysis of methamphetamine, amphetamine, MDA, MDMA, MDEA, and cocaine using capillary electrophoresis. Journal of Forensic Sciences 2001;46(5):1025.

4. Lurie IS, Hays PA, Parker K. Capillary electrophoresis analysis of a wide variety of seized drugs using the same capillary with dynamic coatings. Electrophoresis 2004;25(10-11):1580.

5. Cheng W-C, Lee W-M, Chan MF, Tsui P, Dao K-L. Enantiomeric separation of methamphetamine and related analogs by capillary zone electrophoresis: Intelligence study in routine methamphetamine seizures. Journal of Forensic Sciences 2002;47(6):1248.

6. Phelan CP, Kirby DA. Manuscript in progress.

7. Lurie IS, Klein RFX, Dal Cason T, LeBelle M, Brenneisen R, Weinberger R. Chiral resolution of cationic drugs of forensic interest by capillary electrophoresis with mixtures of neutral and anionic cyclodextrins. Analytical Chemistry 1994;66:4019.

* * * * *

d-Propoxyphene	Linear Range (mg/mL)	Range of % Error	Correlation Coefficient	Y Intercept	Slope
Hydrochloride	0.0513 to 1.23	0.581 to 3.98	0.99991	0.042	19.30
Napsylate 1 (with 0.1N NaOH flush)	0.0509 to 1.22	0.11 to 2.72	0.99995	0.032	12.7813
Napsylate 1,2 (without 0.1N NaOH flush)	0.101-1.22	0.239 to 2.05	0.99981	-0.0023	12.5327
1 Standards sonicated for 5 minutes in acetonitrile before the addition of 0.01 N HCl and Thiamine HCl internal standard. Percent acetonitrile varied as follows: 0.2, 0.4, 0.8, 1.6, 2.4, 3.2, 4.0, and 4.8, respectively. 2 The percent error was 7.6 at a concentration of 0.0509 mg/mL.

* * * * *

Table II: Data for Repeatability Study for d-Propoxyphene Napsylate.

Propoxyphene Napsylate Repeatability (Sonicated)
Percent Acetonitrile Concentration of napsylate (mg/mL)	0.8% 0.203	3.2% 0.815	6% 0.737	8% 0.569	10% 0.729	12% 0.714	14% 0.602
%RSD (with 0.1N NaOH flush)	1.46	1.67	1.46	0.801	0.833	2.19	1.74
%RSD (without 0.1N NaOH flush)	1.46	1.13	2.49	2.22	0.756	1.01	0.755

* * * * *

Table III: Data for Repeatability Study for d-Prropoxyphene HCl.

d-Propoxyphene HCl
Concentration (mg/mL)	0.205	0.822
%RSD	0.372	0.987

* * * * *

Table IV: Data for Recovery Study for d-Propoxyphene Napsylate.

d-Propoxyphene Napsylate Recovery at 4% Acetonitrile (Sonicated)
Adulterant	Actual percentage d-propoxyphene napsylate (% PXP) and Calculated percent error (% error)
Acetaminophen	%PXP= 11.28 % error= 0.576	%PXP= 45.17 % error= 1.55	%PXP= 78.18 % error= 0.147
Aspirin	%PXP= 8.81 % error= 1.59	%PXP= 49.21 % error= 1.61	%PXP= 78.7 % error= 1.00
Caffeine	%PXP= 8.59 % error= 1.97	%PXP= 48.32 % error= 0.166	%PXP= 80.42 % error= 2.77

* * * * *

Table V: Data for Recovery Study for d-Propoxyphene Napsylate.

d-Propoxyphene Napsylate Recovery at 4% Acetonitrile (Non-Sonicated)
Adulterant	Actual percentage d-propoxyphene napsylate (%PXP) and Calculated percent error (% error)
Aspirin	* %PXP= N/D * % error= N/D	%PXP= 45.82 % error= 3.09	%PXP= 85.02 % error= 4.16
Caffeine	* %PXP= N/D * % error= N/D	%PXP= 52.17 % error= 4.69	%PXP= 85.90 % error= 4.91
* N/D = not determined

* * * * *

Table VI: Data for Recovery Study for d-Propoxyphene HCl.

d-Propoxyphene HCl Recovery
Adulterant	Actual Percentage d-Propoxyphene HCl (%PXP) and Calculated Percent Error (% Error)
Acetaminophen	%PXP= 11.27 % error= 0.621	%PXP= 45.81 % error= 0.185	%PXP= 79.37 % error= 0.460
Aspirin	%PXP= 10.85 % error= 0.645	%PXP= 50.71 % error= 0.424	%PXP= 77.15 % error= 0.972
Caffeine	%PXP= 9.23 % error= 1.89	%PXP= 51.18 % error= 0.840	%PXP= 83.11 % error= 0.0120

* * * * *

Figure 1: Electropherogram of Thiamine and Propoxyphene.

* * * * *

Figure 2 : The Electropherograms for the Chiral Analysis of Propoxyphene Napsylate:
         I. The racemic separation of the d,l-propoxyphene napsylate standard.
         II. The chiral analysis of a sample.
         III. The sample co-injected with the d,l-propoxyphene napsylate standard.
         IV. The standard l-propoxyphene nasylate with the d,l-propoxyphene napsylate standard.
         V. The standard d-propoxyphene nasylate with the d,l-propoxyphene napsylate standard.