Journal of Analytical Toxicology, Vol. 16, September/October 1992

A Solid Phase Extraction Technique for the Isolation and Identification of Opiates in Urine Wei Huang, 1,2 W i l m o A n d o l l o , 1 and W i l l i a m Lee H e a r n 1

IDade County Medical Examiner Department, Number One on Bob Hope Road, Miami, Florida 33136-1133 2School of Forensic Sciences, West China University of Medical Sciences, Chengdu, Sichuan, The Peoples' Republic of China

I Abstract J A quantitative method was developed for the simultaneous analysis of morphine, codeine, hydromorphone, hydrocodone, and oxycodone in urine by gas chromatography/mass spectrometry. Samples were hydrolyzed with ~-glucuronidase and then extracted by solid phase extraction on Bond Elute CertifyTM cartridges at pH 6.8. Nalorphine was used as the Internal standard. The opiates were analyzed by full-scan electron Impact GC/MS after derlvatization with acetic anhydride--pyrldlne. The standard curves for all five drugs were linear between 50 and 1000 ng/mL, with correlation coefficients exceeding 0.99. Coefficients of variation were less than 7%. The method was applied to the analysis of postmortem urines positive by EMIT opiate assay, and the effect of the hydrolysis procedure on recovery of each drug was measured. The results Indicate that the hydrolysis procedure is effective in Increasing the recovery of all five drugs from urine. The described method enables the laboratory to identify the five opiates most commonly encountered in forensic and clinical laboratories. Its sensitivity for all five drugs Is well below GC/MS cutoffs for codeine and morphine employed In NIDA laboratories, and It provides for conclusive full-scan drug Identification.

Introduction

Most of the published methods for GC/MS confirmation of opiate-positive urine samples are designed for detection of morphine and codeine only (1-12), and several procedures provide for analysis of 6-monoacetylmorphine to distinguish heroin users from consumers of poppy seed products (12-15). Such limited methods are satisfactory for workplace drag testing, but a more comprehensive opiate confirmation procedure is desirable for postmortem toxicology and human performance drag analysis. Opiates other than heroin are commonly abused and may cause impairment, even when taken therapeutically. Oxycodone and hydrocodone are commonly prescribed as analgesics, and the latter is also employed as an antitussive. Hydromorphone is less commonly prescribed, but, being more potent than heroin, is popular with narcotic abusers. Saady and co-workers included hy-

dromorphone along with morphine and codeine in their GC/MS opiate method, but the sensitivity for hydromorphone was limited to 0.08 mg/L, even with the use of selected ion monitoring (SIM) analysis (16). Cone and Darwin described a method for the analysis of hydromorphone and hydrocodone and their hydroxylated metabolites in urine but did not include the other opiates (17). Bowie and Kirkpatrick (18) reported a method for identification of morphine, codeine, 6-monoacetylmorphine, hydromorphone, hydrocodone, oxycodone, and oxymorphone in urine by liquidliquid extraction and GC/MS operated in the SIM mode. Compound identification is more convincing when the full spectra of unknown and standard samples can be compared. Whenever possible, full spectral comparison is used in our laboratory for drag identification. However, in our experience with opiate analysis, the practical limitations of extract cleanliness and poor recovery associated with liquid-liquid extraction often preclude obtaining clean background-subtracted spectra of anatytes extracted from biological samples when their concentration is less than a few hundred nanograms per milliliter. The use of SIM is a trade-off of specificity for sensitivity. In procedures for qualitative analysis of drags in biological fluids, solid phase extraction (SPE) is gaining favor over liquidliquid extraction because of its efficiency, expediency, and cleanliness (1,2,14,19-22). Hayes et al. (1) and Dixit and Dixit (2) used SPE for the analysis of codeine and morphine but not for other opiates. Nakamura and Stall used solid phase extraction in their GC/MS method for analysis of dihydrocodeine in urine (21). Logan and co-workers used solid phase extraction and high performance liquid chromatography to separate 100 basic drags, including codeine, morphine, dihydrocodeine, hydrocodone, hydromorphone, oxycodone, and oxymorphone from urine but did not apply GC/MS to the analysis of the extracts (22). Their reported extraction recoveries for the opiates exceeded 90%. This paper describes a sensitive method for the isolation and identification of morphine, codeine, hydromorphone, hydrocodone, and oxycodone in urine using enzymatic hydrolysis, Bond Elute Certify TM solid phase extraction, and full scan electron impact GC/MS analysis. The efficiencies of hydrolysis of morphine-3-glucuronide and of extraction of all five opiates are determined. The effect of pH, matrix, and solvents on the recovery of opiates from the Bond Elute Certify TM column is discussed. The method was tested by analyzing opiate-positive forensic samples by immunoassay screening techniques.

Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission,

307

Journal of Analytical

Experimental Reagents and chemicals Methanol, ethyl acetate, dichloromethane, and isopropanol were high purity solvent grade (Burdick and Jackson). Glacial acetic acid, phosphoric acid, potassium phosphates, and potassium hydroxide were analytical reagent grade (Mallinckrodt). Acetic anhydride (analytical reagent) and pyridine (distilled in glass) for use as derivatizing reagents were also obtained from Mallinckrodt. Both were freshly purified by distillation. EMIT d.a.u. TM opiate assay reagents were obtained from Syva. 13-Glucuronidase (from E. Coli) type IX-A #G-7396, from Sigma Chemical, was employed in enzymatic hydrolysis. It was prepared by reconstituting with phosphate buffer (0.075M, pH 6.8) to give a final concentration of 2500 units/mL. Aliquots of 1.0 mL were dispensed in individual vials and stored at -20~ until needed. Elution solvent was prepared by combining 80 mL of methylene chloride with 20 mL of isopropanol, removing 2 mL of this solution, adding 2 mL of fresh ammonium hydroxide, and mixing well. The standards of nalorphine, oxycodone, hydrocodone, morphine, codeine, hydromorphone, and morphine-3-13-d-glucuronide were obtained from Alltech Associates. Nalorphine was used as an internal standard. Bond Elut Certify TM extraction columns and a Vac Elut TM vacuum manifold were provided by Analytichem International, Inc. Screw-cap tubes with round bottoms (16 x 100 mm) were silanized by vapor phase before use. Procedure Postmortem urine samples were screened for opiates by EMIT d.a.u. TM immunoassay on an Autocarousel equipped with a Lab Processor 6500 (Syva). The screening cutoff was 300 ng/mL, as recommended by the manufacturer. Enzymatic hydrolysis was performed by adding 50 gL of working stock solution of nalorphine and 1 mL working ~-glucuronidase solution to 1 mL of urine, mixing, and incubating at 37~ for 60 rain in a heating block. The samples were cooled and centrifuged before applying to columns. The final pH was 6.8. To prepare the solid phase extraction column, 2 mL of methanol and 2 mL phosphate buffer (pH 6.8) were passed through the column sequentially without allowing the sorbent to dry. Besides removing the contaminants present in the cartridge, conditioning sets the sorbent in a state that is compatible with the loading solvent and analytes of interest. Therefore, a buffer at the same pH as the sample was used as the conditioning solvent for sample loading. Urine reference standards or supernatants of the hydrolyzed samples were then added to the column and allowed to pass through by gravity. The column was washed with 1 mL 0.1M acetate buffer (pH 4.0), 2 mL deionized water, and 2 mL methanol and then allowed to dry for 5 min at full vacuum. Finally, the drugs were eluted with 3 mL of elution solvent. The eluate was collected in screw-cap tubes, evaporated to dryness, and reconstituted with 750 pL acetic anhydride and 120 ~L pyridine. The mixture was capped and heated to 60~ for 30 min in a heating block, evaporated to dryness, and reconstituted with 50 laL of dry ethyl acetate, A 1-~L aliquot was injected into the GC/MS. Instrumentation The GC/MS employed was a Finnigan 4500 system. The gas chromatograph was equipped with a 15-meter DB-5 fused silica capillary column with a 0.25-mm inside diameter and 0.25-I.trn 308

T o x i c o l o g y , Vol. 16, S e p t e m b e r / O c t o b e r

1992

film thickness from J&W Scientific. The chromatographic operating conditions were initialized at 70~ with a 1-min hold. The temperature was then programmed from 70 to 280~ at 15~ The final temperature was held for 2 rain. Injection port and transfer line temperatures were 250 and 290~ respectively. The mass spectrometer was operated in electron impact mode with 70 eV of electron energy, and a scan range of rrdz 50 to 400. All data were acquired in full scan mode. Calculations For each compound, the base peak was selected to measure the area ratio of compound to internal standard. The selected base ions were 299 for hydrocodone and 353, 327, 285, 357, and 341 for the acetates of nalorphine, morphine, hydromorphone, oxycodone, and codeine, respectively. The calculation of recovery was based on the mean of three determinations. It was performed by comparing the results of analysis of blank urines spiked with 500 ng of each opiate prior to extraction to those from the analysis of extracts of blank urine that were spiked just before derivatization. The internal standard was added to both sets of samples in the initial step of the procedure. The limit of detection was defined as the lowest concentration that would satisfy a purity of 800 or better for the full spectrum of each opiate studied when compared to the library in the data system. 1

R e s u l t s and D i s c u s s i o n Efficiency of enzymatic hydrolysis The efficiency of enzymatic hydrolysis was measured by comparing peak areas of extracts obtained from hydrolysis of morphine glucuronide to those of free morphine standard solutions analyzed under the same conditions. Assuming the free morphine standard represents 100% hydrolysis, the yield from bound morphine was consistently determined to be 80%. Analysis of forensic urine specimens (see below) demonstrated that the hydrolysis technique liberates all five opiates from their conjugated metabolites, but the unavailability of those conjugated metabolites from commercial sources precluded determinations of efficiency of their hydrolysis. Effect of pH and sample matrix on recovery The effects of pH on solid phase extraction recovery at pH 3, 4, 5, 6, 7, 8, 9, and 10 were investigated. Phosphoric acid and the phosphates of potassium were used to prepare the different buffers needed to adjust sample pH and conditioning solvent. The results for aqueous standard solutions and urine standard solutions are given in Tables I and II, respectively. It was observed that recovery of most compounds (hydrocodone, morphine, oxycodone, and codeine) from urine was independent of pH. Although hydromorphone showed higher extraction recoveries at pH 6 or less, the enzyme hydrolysis was more effective at pH 6.8, so the higher pH was used. The data in Tables I and II show that The purity search finds the pattern in the library that is most likeIy 1o be identical to the unknown. It employs an alogrilhm that makes use of the relalive intensity inlormation (i.e. ion ralio) for each of up to 50 mass peaks in the spectrum. In a sense, lhe data system is making a 50-ion ratio comparison, although impotlant mass/intensity weighting factors are included in the algo=Sthm. The purity search algorithm easily permits differentiation of analyte spectra from one another. A purity of 1000 means that the unknown spectrum is identical to the library spectrum. A value of 0 means that the spectra have no features in common. By convention and in-house expenence, the level of similanty represented by a punt,/value of 800 combined with the proper analyte retention time provides absolute assurance of a correct identification.

Journal of Analytical Toxicology, Vol. 16, S e p t e m b e d O c t o b e r 1992

in most cases the recovery from urine was slightly lower than from aqueous solution.

Evaluation of the method Linear responses were obtained for all opiates studied in the concentration range of 50 to 1000 ng/mL with a correlation coefficient of 0.99 or better (Figure 1). Coefficients of variation of 7% or better were determined at a concentration of 500 ng/mL. Limits of detection, as defined above, were found to be 50 ng/mL

Table I. pH Effect on Recovery in Aqueous Recovery(%) Compound pH3 pH4 pH5 pH6 pH7 Hydrocodone 89 Codeine 100 Oxycodone 91 Hydromorphone 90 Morphine 101

pH8

pH9 pill0

98 107 96

96 104 98

93 102 98

95 100 93

94 101 93

98 104 97

93 101 94

96 107

95 105

93 105

89 104

89 102

94 105

90 103

Table II. pH Effect on Recovery in Urine Recovery(%) Compound pH3 pH4 pH5 pH6 Hydrocodone Codeine Oxycodone Hydromorphone Morphine

0

Results of analysis of forensic case samples Forensic samples that were found to be opiate positive by immunoassay techniques were referred for identification by the above procedure. Each sample was extracted with and without enzymatic hydrolysis to compare the yield obtained. The results tabulated in Table IV show increases in concentration of up to 1357% for morphine, 236% for codeine, 193% for hydromorphone, 42% for hydrocodone, and 23% for oxycodone. Although hydrocodone appears to lack a hydroxyl group for conjugation, the increased recovery following hydrolysis indicates that it is partly excreted as a glucuronide. Perhaps the enol tautomer of its keto function participates in conjugate formation.

Table III. Analytical Method Linearity, CV, and Limit of Detection

pH7

pH8

pH9 pill0

82 88 92

79 84 88

76 84 92

77 85 88

82 89 85

81 88 83

85 93 86

88 98 85

97 92

93 88

96 89

91 98

66 83

57 84

46 90

32 94

Correlation Coefficient Limit of coefficient ofvariati0n(%) detecti0n(ng/mL)

Compound Hydromorphone Hydrocodone Morphine Codeine Oxycodone

0.9980 0.9951 0.9998 0.9999 0.9996

6.91 5.19 3.72 1.85 2.51

50 50 10 10 50

Table IV. Effect of Enzymatic Hydrolysis on Recovery of Opiates From Forensic Urine Samples

4.0 3.5

or better and corresponded to signal-to-noise ratios greater than 10 to 1 for the base peak ions. Table III lists specific information on correlation, CV, and limits of detection for each opiate studied. The overall efficiency of the procedure was determined to be 75% recovery of conjugated morphine from its glucuronide and up to 90% for the free drug.

9

MORPHINE

9 o 9

CODEINE OXYCODONE HYDROMORPHONE

HM HC MO CO OC Sample Hydrolysis (mg/I.) (mg/l_) (mg/L) (mg/I.) (mg/L)

30

1

Y N

4.94 0.54

2.5

2

Y N

19.82 1.36

2.01

3

Y N

1.40 0.24

1.5

4

Y

0.45 0.55

N 1.0

5

3.5

6

3.0= 0

Y N Y

N 100

200

300

400

SO0

600

700'

800

900

1000

7

Y

8

N Y N

CONCENTRATION (ng/ml)

Figure 1. Standardcurves of GC/MSanalysisof five opiatesin urine. The vertical axis representsthe ratio of the area of the base peakof the drug spectrumto that of the internal standard.The ions usedare describedin the text. The horizontal axis represents concentration of the drugs in nanograms per milliliter. Regression coefficients for all five drugs exceeded0.99.

1.61 0.55

15.64 13.58

4.06 3.30

6.84 4.80 0.64 0.068

0.84 0.25

2.36 0.67 6.82 4.01

0.23 0.12

HM = Hydromorphone HC = Hydrocodone MO = Morphine CO = Codeine OC = Oxycodone

309

Journal of AnalyticalToxicology,VOL 16, September/October1992

Conclusion The aim of this study was to examine the applicability of solid phase extraction as a clean, expedient, and efficient extraction technique for opiates. In preliminary experiments we evaluated a variety of derivatizing reagents, including trifluoroacetic anhydride (TFAA), pentafluoropropionic anhydride (PFPA), heptafluorobutyric anhydride (HFBA), N,O-bis-(trimethylsilyl) trifluoroacetamide + I% trimethylchlorosilane (BSTFA + 1% TMCS), and acetic anhydride-pyridine. Acetic anhydride derivatization was chosen for the current method because it yielded a single derivative for each opiate except hydrocodone, which is not acetylated. The fluoroacyl anhydrides gave variable amounts of mono- and di-acyl derivatives of hydromorphone, oxycodone, and nalorphine. The silyl derivatives were found to be less stable than the acetyl, particularly in a humid environment, and oxycodone and hydromorphone did not derivatize effectively with BSTFA-1% TMCS. One of these other derivatives can be employed if detection of 6-monoacetylmorphine is required, but at a cost of quantitative accuracy and precision. The results of our evaluation indicate that the procedure described above is able to detect and identify opiates and their metabolites within acceptable limits of detection, linearity, and precision. The extraction gives good recovery for all five opiates (80% and better with the exception of hydromorphone, which had approximately 70% recovery). The extracts yielded low GC/MS background, which enhanced the mass spectral characteristics and permitted full scan identification at levels ranging from l0 to 50 ng/mL, depending upon the drug. Also, instrument maintenance downtime is decreased because the injection port, column, and detector are exposed to less contamination. The extraction process is simpler and faster than cumbersome liquid liquid processes and lends itself to easy automation.

References 1. L.W. Hayes, W.G. Krasselt, and P.A. Mueggler. Concentrations of morphine and codeine in serum and urine after ingestion of poppy seed. Clin. Chem. 33(6): 806-808 (1987). 2. V. Dixit and V.M. Dixit. A solid phase extraction technique for preparation of drugs of abuse samples. Am. Clin. Lab. 46-51 (1990). 3. A. Dasgupta and V. Haver. GC/MS confirmation of urinary opiates by use of trimethylsilyl derivatives. J. Clin. Chem. 34(9): 1925-30 (1988). 4. R.E. Struempler. Excretion of codeine and morphine following ingestion of poppy seeds. J. Anal ToxicoL 11(3): 97-99 (1987). 5. H.N. EISohly, D.E Stanford, A.B. Jones, M.A. EISohly, H. Snyder, and C. Pedersen. Gas chromatographic/mass spectrometric analysis of morphine and codeine in human urine of poppy seed eaters. J. Forens. Sci. 33(2): 347-56 (1988). 6. R.L. Foltz, A.E Fentiman, Jr., and R.B. Foltz. GC/MSAssays for Abused Drugs in Body Fluids. National Institute on Drug Abuse Research Monograph 32, Department of Health and Human Services Publication No. (ADM) 80-1014, U.S. Government Printing Office, Washington, D.C., 1980, pp. 110-26.

310

7. W.O.R. Ebbighausen, J.H. Mowat, P. Vestergaard, and N.S. Kline. Stable isotope method for the assay of codeine and morphine by gas chromatography-mass spectrometry. A feasibility study. Adv. Biochem. Psychopharm. Raven Press, New York, Vol. 7, pp. 135-46 (1973). 8. B.D. Paul, L.D. Mell, Jr., J.M. Mitchell, J. Irving, and A.J. Novak. Simultaneous identification and quantitation of codeine and morphine in urine by capillary gas chromatography and mass spectroscopy. J. Anal. Toxicol. 9:222-26 (1985). 9. N.B. Wu Chen, M.I. Schaffer, R.L. Lin, and R.J. Stein. Simultaneous quantitation of morphine and codeine in biological samples by electron impact mass fragmentography. J. Anal ToxicoL 6: 231-34 (1982). 10. S.J. Mule and GA, Casella. Confirmation of marijuana, cocaine, morphine, codeine, amphetamine, methamphetamine, and phencyclidine by GC/MS in urine following immunoassay screening. J. Anal ToxicoL 12:102-107 (1988). 11. E.J. Cone, W.D. Darwin, and W.F. Buchwald. Assay for codeine, morphine and ten potential urinary metabolites by gas chromatography-mass fragmentography. J. Chromatogr. 275: 307-318 (1983). 12. P.A. Clarke and R.L. Foltz. Quantitative analysis of morphine in urine by gas chromatography-chemical ionization-mass spectrometry, with [N-C2H3]-morphine as an internal standard. Clin. Chem. 20(4): 465-69 (1974), 13. S.J. Muld and G.A. Casella, Rendering the "poppy seed defense" defenseless: Identification of 6-MAM in urine by GC/MS. Clin. Chem. 34(7): 1427-30 (1988), 14. J. Fehn and G. Megges. Detection of O6-monoacetylmorphine in urine samples by GC/MS as evidence for heroin use. J. Anal Toxicol. 9:134-38 (1985). 15. A. Solans, R. De La Torre, and J. Segura. Determination of morphine and codeine in urine by gas chromatography-mass spectrometry. J. Pharm. Biomed. Anal 8(8-12): 905-909 (1990). 16. J.J. Saady, N. Narasimhachari, and R.V. Blanke. Rapid, simultaneous quantification of morphine, codeine, and hydromorphone by GC/MS. J. Anal. Toxicol. 6:235-37 (1982). 17. E.J. Cone and W.D. Darwin. Simultaneous determination of hydromorphone, hydrocodone and their 6-alpha- and 6-beta-hydroxy metabolites in urine using selected ion recording with methane chemical ionization. Biomed. Mass Spec. 5:291-95 (1978). 18. L.J. Bowie and K.B. Kirkpatrick. Simultaneous determination of monoacetylmorphine, morphine, codeine, and other opiates by GC/MS. J. Anal ToxicoL 13:326-29 (1989). 19. J.T. Stewart, T.S. Reeves, and I.L. Honingberg. A comparison of SPE techniques for the assay of drugs in aqueous solutions and human plasma samples. Anal. Lett. 17(B16): 1811-26 (1984). 20. I.D. Wilson. A rapid method for the isolation and identification of drug metabolites from human urine using SPE and proton NMR spectroscopy. J. Pharm. Biomed. Anal. 4(5): 663 (1986). 21. G.R. Nakamura, W.J. Stall, and R.D. Meeks. Analysis of dihydrocodeine in urine using Sep-PakT M C-18 cartridges for sample cleanup. J. Forens. Sci. 32(2): 535-38 (1987). 22. B.K. Logan, D.R. Stafford, I.R Tebbett, and C.M Moore. Rapid screening for 100 basic drugs and metabolites in urine using cation exchange solid-phase extraction and high performance liquid chromatography with diode array detection. J. Anal. ToxicoL 14(3): 154-59 (1990). Manuscript received November 13, 1991; revision received May 29, 1992,