Journal of Analytical Toxicology, Vol. 23, July/August 1999 Direct SemiquantitativeScreeningof Drugs of Abuse in Serum and Whole Blood by Means of CEDIA | DAU Urine lmmunoassays Stefanie Iwersen-Bergmann* and Achim Schmoldt Departmentof LegalMedicine, University of Hamburg, Butenfeld 34, D-22529 Hamburg, Germany Abstract The purpose of this study was to test the direct applicability of CEDIA DAU urine immunoassaysto serum or whole blood. The performance of the urine assays for sensitive screening of amphetamines (AMP), benzoylecgonine (BZE),benzodiazepines (BENZ), methadone (MET),opiates (OPI), and tetrahydrocannabinol carboxylic acid (THCCOOH) was evaluated on the BM/Hitacbi 911 analyzer with unpretreated serum and whole blood. The limit of detection was 0 ng/mL for all tests. Cutoff values were set from 10 to 40 ng/mL for the different assays. The assays were found to be linear between the following concentrations: AMP 0-2500 ng/mL, BZE 0-1200 ng/mL, BENZ 0-1600 ng/mL, MET 0-600 ng/mL, OPI 0-720 ng/mL, and THCCOOH 24-60 ng/mL. Precision results (within run) for different concentrations were as follows: AMP 3.1-5.7%, BZE 2.4-6.6%, BENZ 4.3-8.0%, MET 2.0-5.5%, OPI 2.8-7.6%, and THCCOOH 1.4-2.4%. Between-run results were as follows: AMP 8.7-15.5%, BZE 6.4-7.5%, BENZ8.2-15.8%, MET 2.7-5.1%, OPI 4.3-11.2%, and THCCOOH 2.6-7.4%. Sensitivity, specificity, and comparison of CEDIA semiquantitation with GC-MS quantitative results were performed on 500 original serum and whole blood samples. The data provided sufficient documentation to use the CEDIA urine-screening technique without any adaptation as a sensitive serum/whole blood screening for BZE, BENZ, MET, OPI, and THCCOOH. Se(um screeningfor amphetamines is not sensitive enough in the unchanged urine mode. It will require some adaptation to a serum mode (probably a higher sample volume [BM/Hitacbi 911] combined with protein precipitation of the sample). Introduction The widespread use of immunoassay technique for the screening of drugs of abuse in urine is based on its fast perfor- mance and good sensitivity. However, in many cases, immuno- logical analysis of blood and serum is required, for example, when no urine sample is available. Depending on the technology of the manufacturer (FPIA,EIA, KIMS)and the measured wavelength of the analyzer, protein content or turbidity of serum or hemoglobin of whole blood samples may result in high background absorbance levels (1) or error messages (2). Therefore, numerous *Author to whom correspondence shouldbe addressed. efforts have been made to adapt the urine immunoassays to serum or whole blood. Indirect application for EMIT or TDx/ADx was achieved by protein precipitation with methanol (3,4), ace- tone (5), trichloroacetic acid (6), or zinc sulfate (2). Other possi- bilities are the extraction of the blood sample (7,8) or simple dilution with saline (9). Disadvantages of these methods are that they are time consuming and probably result in a loss of sensi- tivity. This is of special interest when the result of the screening is required urgently, especially for clinical cases. Another problem specific to Germany is that suspected impaired car drivers or other road users can refuse to supply a urine sample but not a blood specimen. As a consequence, the number of cases in which Table I. Spiked Low, Medium, and High Serum Concentrations Serum concentrations (ng/mL) Assay Low Medium High AMP(methamphetamine) 500 1000 2000 BZE (benzoylecgonine) 100 300 600 BENZ (nitrazepam) 100 200 400 MET (d,l-methadone) 150 300 600 OPI(morphine) 100 300 600 THCCOOH (THCCOOH) 15 30 60 Display CEDIA (ng/mL) 1.000 Serum -* Urine ~AP Serum 20O 0 0 200 400 600 800 Concentration (ng/mL) Figure 1. Benzoylecgonine-spiked blank serum measured with urine, serum calibration, and spikedblankserum afterproteinprecipitation. Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 247 Downloaded from https://academic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
10
Embed
Direct Semiquantitative Screening of Drugs of Abuse in Serum and
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Journal of Analytical Toxicology, Vol. 23, July/August 1999
Direct Semiquantitative Screening of Drugs of Abuse in Serum and Whole Blood by Means of CEDIA | DAU Urine lmmunoassays Stefanie Iwersen-Bergmann* and Achim Schmoldt
Department of Legal Medicine, University of Hamburg, Butenfeld 34, D-22529 Hamburg, Germany
Abstract
The purpose of this study was to test the direct applicability of CEDIA DAU urine immunoassays to serum or whole blood. The performance of the urine assays for sensitive screening of amphetamines (AMP), benzoylecgonine (BZE), benzodiazepines (BENZ), methadone (MET), opiates (OPI), and tetrahydrocannabinol carboxylic acid (THCCOOH) was evaluated on the BM/Hitacbi 911 analyzer with unpretreated serum and whole blood. The limit of detection was 0 ng/mL for all tests. Cutoff values were set from 10 to 40 ng/mL for the different assays. The assays were found to be linear between the following concentrations: AMP 0-2500 ng/mL, BZE 0-1200 ng/mL, BENZ 0-1600 ng/mL, MET 0-600 ng/mL, OPI 0-720 ng/mL, and THCCOOH 24-60 ng/mL. Precision results (within run) for different concentrations were as follows: AMP 3.1-5.7%, BZE 2.4-6.6%, BENZ 4.3-8.0%, MET 2.0-5.5%, OPI 2.8-7.6%, and THCCOOH 1.4-2.4%. Between-run results were as follows: AMP 8.7-15.5%, BZE 6.4-7.5%, BENZ 8.2-15.8%, MET 2.7-5.1%, OPI 4.3-11.2%, and THCCOOH 2.6-7.4%. Sensitivity, specificity, and comparison of CEDIA semiquantitation with GC-MS quantitative results were performed on 500 original serum and whole blood samples. The data provided sufficient documentation to use the CEDIA urine-screening technique without any adaptation as a sensitive serum/whole blood screening for BZE, BENZ, MET, OPI, and THCCOOH. Se(um screening for amphetamines is not sensitive enough in the unchanged urine mode. It will require some adaptation to a serum mode (probably a higher sample volume [BM/Hitacbi 911] combined with protein precipitation of the sample).
Introduction
The widespread use of immunoassay technique for the screening of drugs of abuse in urine is based on its fast perfor- mance and good sensitivity. However, in many cases, immuno- logical analysis of blood and serum is required, for example, when no urine sample is available. Depending on the technology of the manufacturer (FPIA, EIA, KIMS) and the measured wavelength of the analyzer, protein content or turbidity of serum or hemoglobin of whole blood samples may result in high background absorbance levels (1) or error messages (2). Therefore, numerous
*Author to whom correspondence should be addressed.
efforts have been made to adapt the urine immunoassays to serum or whole blood. Indirect application for EMIT or TDx/ADx was achieved by protein precipitation with methanol (3,4), ace- tone (5), trichloroacetic acid (6), or zinc sulfate (2). Other possi- bilities are the extraction of the blood sample (7,8) or simple dilution with saline (9). Disadvantages of these methods are that they are time consuming and probably result in a loss of sensi- tivity. This is of special interest when the result of the screening is required urgently, especially for clinical cases. Another problem specific to Germany is that suspected impaired car drivers or other road users can refuse to supply a urine sample but not a blood specimen. As a consequence, the number of cases in which
Table I. Spiked Low, Medium, and High Serum Concentrations
Serum concentrations (ng/mL) Assay Low Medium High
Figure 1. Benzoylecgonine-spiked blank serum measured with urine, serum calibration, and spiked blank serum after protein precipitation.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 2 4 7
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Journal of Analytical Toxicology, Vol. 23, July/August 1999
1.000 DjsPlaY--CEDIA (ng/mL) Serum * Urine o AP Serum
800 j . d
600 . . ~ f . - - J -
200 ~
0 200 400 600 800 Concentration (ng/rnL)
Figure 2. Morphine-spiked blank serum measured with urine, serum calibra- tion, and spiked blank serum after protein precipitation.
Display CEDIA (ng/mL) 200 ~ Urine . Serum e hemoI.Serum I
150
100
50 ~ :
0 ~ ..... 50 100 200
Concentratio~ (ng/mL)
Figure 3. Benzoylecgonine-spiked blank serum, blank hemolytic serum, and blank urine measured with urine calibration.
only a blood sample (and no urine sample) is available for analysis is steadily increasing. The blood-sample Vacutainer tubes used by the police contain sodium or potassium fluoride. Because of these salts, blood samples for the analysis for drugs of abuse are more or less hemolyzed. Concerning blood samples, another important aspect is that in many cases only very limited sample material is available. Especially for forensic (impaired drivers, criminals) and clinical cases, it is necessary to detect very low concentrations of drugs of abuse. A procedure that can show sensitively which test parameters require a confirmation method saves time and assists in getting analytic data of high quality from a small quantity of sample, sometimes only 1 mL of blood.
As a first step, we tested whether the high protein content of serum samples interferes with the cloned enzyme donor immunoassay (CEDIA) and whether it is necessary to perform a special calibration with serum calibrators that are in lower con- centration ranges than those supplied by the manufacturer. To do this, we selected the opiate (OPI) and benzoylecgonine (BZE) assays as test parameters. Serum and urine blanks were spiked with morphine or BZE and measured with the urine standard cal- ibration mode (as recommended by the manufacturer) and with a self-spiked serum calibration with lower concentrations. Subsequently, hemolytic whole blood and acetone-precipitated serum samples were measured. This pretest showed the suc- cessful applicability of CEDIA DAU for the direct measurement of opiates and BZE in serum and whole blood. Therefore, a system- atic evaluation for all test parameters mentioned was begun.
Materials and Methods
250
200
Display CEDIA (ng/mL)
Urine * Serum o hemol. Serum
150
100
50
o
i
i
. J
50 1 O0 150 200
Standards Stock solutions for amphetamine (AMP), methamphetamine
(METH), methylenedioxymethamphetamine (MDMA), methyl- enedioxyamphetamine (MDA), methylenedioxyethylamphetamine (MDEA), benzodioxolbutaneamine (MBDB), nitrazepam, diaze- pare, oxazepam, nordazepam, flunitrazepam, temazepam, cloba- zam, clonazepam, chlordiazepoxide, tetrazepam, bromazepam, midazolam, BZE, cocaine, methadone (MET), morphine, codeine, dihydrocodeine, tetrahydrocannabinol (THC), and tetrahydro- cannabinol carboxylic acid (THCCOOH) and their deuterated internal standards were purchased from either Promochem GmbH
(Wesel, Germany) or Sigma Chemie (Deisenhofen, Germany). Stock solutions were diluted with ethanol to produce standard working solutions. The standard working solutions were used to
prepare serum and urine standards for the analysis.
Reagents All solvents were of reagent-grade quality.
CEDIA immunoassay reagents were supplied by Boehringer Mannheim (Mannheim, Germany). Reagents for the ADx were supplied by Abbott GmbH (Wiesbaden, Germany). Reagents and test tubes for RIA were supplied by Biermann (Bad Nauheim, Germany). Measurements were done on a y-counter (Berthold, Isernhagen, Germany). All procedures for the preparation of the
248
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Journal of Analytical Toxicology, Vol. 23, July/August 1999
immunoassay reagents were the same as those recommended by the manufacturer for the analysis of urine. The CEDIA test has two multiurine controls (for methamphetamine, BZE, morphine, nitrazepam, and methadone) and two controls (25 ng and 50 ng/mL) for THCCOOH which were used to check the accuracy
Table III. Within-Run and Between-Day Precision for CEDIA
and stability of the calibration. The 25-ng/mL THCCOOH control was routinely diluted to 20 ng/mL. The urine calibration curves proved to be stable for at least two weeks.
Within run (%CV) Between day (%CV) Assay Low Medium High Low Medium High
with morphine (10, 50, 100, 200, 300, 500, 800, 1000, and 2000 ng/mL) or BZE (10, 50, 100, 200, 300, 500, 600, and 800 ng/mL) and measured with the urine standard calibration mode (as recom- mended by the manufacturer: BZE 300, 2000, and 5000 ng/mL; OP1300, 800, and 2000 ng/mL) and subsequently with a self-spiked serum calibration with lower concentrations (BZE 200, 500, 1000 ng/mL; OPI 150, 300, 500 ng/mL). Drug-free serum and urine specimens were obtained from laboratory personnel. After these measurements an acetone precipitation was performed for the spiked serum samples as follows: aliquots of 500 tJL were added dropwise to 1 mL acetone, then vortex mixed for I rain and centrifuged at 4000 rpm for 5 rain. The clear supernatant was evapo- rated under nitrogen and reconstituted in 500 IJL of ADx buffer solution. The reconstituted sample was again measured by BM/Hitachi 911 analyzer and compared with direct serum measurements. Hemolyzed blood samples were spiked and run like serum samples to see if red color interfered with the test. The instrument settings were the same as the settings for the urine specimens rec- ommended by the manufacturer. The only excep- tion was the THCCOOH assay. Routinely, a cutoff of 20 ng/mL instead of 25 ng/mL is used.
Limit of detection, cutoff, and precision The limit of detection was calculated by adding
three standard deviations to the mean of 21 repli- cates of a blank serum. Because of the necessary low limit of detection, cutoffs were set as follows: AMP, 20 ng/mL; BZE, 10 ng/mL; BENZ, 10 ng/mL; MET, 40 ng/mL; OPI, 10 ng/mL; THCCOOH, 10 ng/mL. For within-run precision, three controls of a low, medium, and high serum concentration were assayed in 21 replicates. Between-day preci- sion was determined by assaying each of the low, medium, and high control in three replicates per day for 21 consecutive days. The low, medium, and high concentrations are shown in Table I. Linearity was analyzed by 11 dilution steps of a high concen- tration, starting 10% over the assay range.
Immunoassays (CEDIA, FPIA, RIA) Five hundred serum or whole blood hemolyzed
samples from clinical poisonings or impaired
249
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Journal of Analytical Toxicology, Vol. 23, July/August 1999
drivers were centrifuged and run on the BM/Hitachi 911 analyzer with CEDIA DAU reagent under the same conditions as urine samples. Following this initial testing CEDIA-negative samples (under the cutoff) were precipitated with acetone as described with the only difference being that 20 IJL of isopropanolic HCI was added before evaporation. The reconstituted samples were tested using the ADx analyzer and complete ADx reagent. Pretests with spiked acetone precipitated samples showed the ADx to have a nearly equal or partially better sensitivity (amphetamines) than CEDIA with the exception of the benzodiazepine test. Therefore, we used it as a method for cross checking CEDIA-negative sam- ples. Additionally, CEDIA-negative samples were retested by RIA for morphine and by HPLC for benzodiazepines. Positive samples by either screening assay with a displayed drug concentration
higher than the CEDIA cutoffs were selected for confirmation by gas chromatography-mass spectrometry (GC-MS). Radio- immunoassay (RIA) was performed according to the manufac- turers instruction. A five-point standard curve from 2.5 to 250 ng/mL was generated. The cutoff positive calibrator was 5 ng/mL. The correlation coefficient for each standard curve was 0.998 or greater.
GC-MS Quantitative GC-MS was performed on 91, 102, and 21 sam-
ples positive for BZE, opiates, and methadone, respectively, as described elsewhere (10) and on 16 samples positive for amphetamines as described elsewhere (11). Quantitative GC-MS was performed on 71 samples positive for THCCOOH as follows:
Figure 5. Results of the linearity test for AMP (A), BZE (B), BENZ (C), MET (D), OPI (E), and THCCOOH (F).
250
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Journa l o f A n a l y t i c a l Tox i co logy , Vo l . 23, Ju l y /Augus t 1999
10 ng and 20 ng of the deuterated standards (THC-d3, THC- COOH-d3) were added to 1 mL of centrifuged serum and trans- ferred to an RP18 column (Baker) conditioned with 1 mL methanol and 1 mL distilled water. The column was then washed consecutively with 2 mL distilled water, 2 mL acetic acid (0.25 mol/L), and 2 mL distilled water; dried by passing air through it for 10 min; and subsquently centrifuged for 10 min (4000 rpm). Adsorbed drugs of abuse were eluted with 3 x 0.5 mL acetone. The eluent was evaporated to dryness under a stream of nitrogen at 50~ The residue was reconstituted with 150 IJL tetrabutyl- ammonia hydroxide/dimethyl sulfoxide (2:98, v/v). After 2-rain incubation at room temperature, 50 IJL iodomethane was added and the solution vortex mixed. The reaction was stopped after an additional 5 rain with 350 IJL HCI (0.1 tool/L). Extraction of the methylated products followed twice with 2 x I mL isooctane for 1 rain. After centrifugation the organic phases were decanted into a clean tube, dried under nitrogen, and reconstituted in 50 IJL isooctane.
A 1-1JL quantity of the sample was injected into the same GC-MS system as described for opiates. The temperature was pro- grammed from 100 (3-min hold) to 180~ at 30~ and to 290~ at 10~ (8-rain hold). The MS detected the following ions of the methylated products in selected ion monitoring mode: THC-d3 316/331,THC 313/328, THCCOOH-d3 316/360/375, and THCCOOH 313/357/372.
Standard curves were generated using a six-point standard curve from 10 to 500 ng/mL for amphetamine, BZE, cocaine, codeine, dihydrocodeine, MBDB, MDA, MDEA, MDMA, metham- phetamine, morphine, and methadone. Six-point curves from 1 to 50 ng/mL were generated for THCCOOH and THC. The detection limits were 0.5 ng/mL for THCCOOH and THC; 5 ng/mL for ben- zoylecgonine, codeine, dihydrocodeine, and morphine; and 10
Table V. Results of 16 Blood Samples Analyzed by CEDIA and GC-MS for Amphetamines
GC-MS
ng/mL for amphetamine, cocaine, MBDB, MDA, MDEA, MDMA, methamphetamine, and methadone. The correlation coefficient for each standard curve generated was 0.998 or greater.
High-performance liquid chromatography (HPLC) Quantitative HPLC was performed on 47 samples positive for
benzodiazepines and for cross checking of the 453 benzodi- azepine-negative samples. Five-hundred nanograms of the internal standard papaverine was added to 1 mL of centrifuged serum. The pH was adjusted to pH 8-9, 1 mL K2HPO 4 buffer (pH 8-9) was added, and the solution was vortex mixed for 5 min. After centrifugation, the supernatant was transferred to an RP18 (Baker) column conditioned with 2 mL methanol and 2 mL dis- tilled water. The column was then washed consecutively with 2 x I mL distilled water and dried by passing air through for 10 min. Adsorbed drugs of abuse were eluted with 3 x 0.5 mL methanol. The eluent was evaporated to dryness under a stream of nitrogen at 50~ The residue was reconstituted in 50 pL methanol. HPLC analysis: a 20-1JL quantity of the sample was injected into the HPLC system. Instrumentation: LDC/Milton Roy Spectro- MonitorTMD detector, pump: TSP Constametric, wavelength 220 nm. The system was equipped with a RP18 LiChrospher| column, solvent system: (4.80 g H3PO 4 + 6.66 g K2HPO 4 in 1 L distilled water, adjusted to pH 2.3) + 458.8 g acetonitrile. Flow rate was 1 mL/min.
Standard curves were generated using a six-point standard curve from 40 to 500 ng/mL for nitrazepam, diazepam, oxazepam, nordazepam, temazepam, bromazepam, midazolam, and clo- bazam. Six-point curves from 10 to 50 ng/mL were generated for clonazepam, flunitrazepam, and tetrazepam. The detection limits were 30 ng/mL for nitrazepam, oxazepam, nordazepam, temazepam, bromazepam, and midazolam; 50 ng/mL for
diazepam and clobazam; and 10 ng/mL for clon- azepam, flunitrazepam, and tetrazepam. The cor- relation coefficient for each standard curve generated was 0.998 or greater.
Sample CEDIA Amph. Methamph. MDMA MDEA MDA MBDB no. ( n g / m L ) ( n g / m L ) ( n g / m L ) ( n g / m L ) ( n g / m L ) ( n g / m L ) ( n g / m L )
1 21 134 - 33 4 7 50 -
2 0 - - 275 - -
3 0 - - 8 2 0 - -
4 9 8 215 - - - 100 -
5 0 - - - 4 0 - -
6 223 3 1 7 - 6 6 4 - - -
7 0 - - - 2 5 0 2 0 -
8 0 - - 242 - - -
9 0 - 180 - 30 - -
10 0 70 20 20 4 0 - -
11 0 - - 164 - - 100
12 0 . . . . . 170
13 0 20 - 80 - - 8 0
14 0 - - - 127 - 148
15 0 - - - 315 - -
16 0 - - - 212 - 180
* Abbreviat ions: A m p h = Amphetamine , Me thamph = Methamphetamine, M D M A = Methy lened ioxymeth - amphetamine, MDEA = Methy lenedioxyethamphetamine, M D A = Methy lened ioxyamphetamine , MBDB = Benzodioxolbuta neamine.
Results and Discussion
Pretest Figures 1 and 2 show the results of spiked blank
serum measured with urine or serum calibration and after protein precipitation. Acetone-precipi- tated samples gave the same results as unpre- treated serum for benzoylecgonine and 10-30% lower results for morphine. The use of the serum calibration curve or the standard urine calibration curve gave approximately the same results. For that reason and ease of use, the standard urine calibration curve was selected for further determi- nations. Figures 3 and 4 show that red color due to hemoglobin does not interfere with the test. However, it has to be noted that high viscosity of blood samples (e.g., postmortem blood) may cause problems for the pipette station of the
251
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Ioumal of Analytical Toxicology, Vol. 23, July/August 1999
BM/Hitachi 911 analyzer. A precipitation step is thus recom- mended for very viscous samples to be on the safe side. Spiked blank urine samples delivered the same results as spiked serum samples. Thus, serum proteins do not interfere with the test, and direct measurement without any preliminary treatment of serum samples is therefore possible, in principle. After these successful pretests, the main study was begun.
Table Vl. Results of 91 Blood Samples Analyzed by CEDIA and GC-MS for Benzoylecgonine and Cocaine
GC-MS Case CEDIA BZE Cocaine Case CEDIA BZE Cocaine no, (ng /mt ) (ng/mt) (ng/mL) no. (ng/mt) (ng/mL) (ng/mL)
Limit of detection and precision In all assays tested the limit of detection was 0 ng/mL. This
result was corroborated by all the following measurements. Whereas urine blank measurements (i.e., dependent on turbidity, bacterial contamination, etc.) often give threshold displays of 10-30 ng/mL and even as high as 100 ng/mL, serum measure- ments show distinctively lower thresholds, most often a zero dis-
play. Of course, this is dependent on the quality of the serum sample too. Table I lists the spiked low, medium, and high serum concentrations, and Table II shows the mean concentrations with
GC-MS CEDIA. Table III lists the average within-run and between-day precision for CEDIA. For the low- THCCOOH concentration of 15 ng/mL, no coeffi- cient of variation (CV%) is given. This is because the display for about half of the samples showed 0 ng/mL as a result. However, in real samples, a combination of different cannabinoids that cross- reacts with the antibody is present. As can be seen in Table IV, CEDIA gave a signal above the cutoff for very low concentrations (cases 1, 2, 11, 13, 19, 33, and 36). Therefore, the test is sufficiently sen- sitive for the direct detection of low cannabinoid concentrations in blood.
Linearity Figures 5A-F show the results of the linearity
test for the different groups. The assays were found to be linear between the following concen- trations: AMP 0-2500 ng/mL, BZE 0-1200 ng/mL, BENZ 0-1600 ng/mL, MET 0-600 ng/mL, OPI 0-720 ng/mL, and THCCOOH 24-60 ng/mL. The pretests have already shown that linearity begins at 10 ng/mL for opiates and benzoylecgonine, respectively. Because of the very low limits of detection, very low cutoffs from 10 to 40 ng/mL were selected as follows: AMP 20 ng/mL, BENZ 10 ng/mL, MET 40 ng/mL, OPI 10 ng/mL, and THCCOOH 10ng/mL.
Method comparison Comparison of GC-MS quantitative and CEDIA
semiquantitative results are shown in Tables IV-IX and Figures 6A-D.
Amphetamines Results for 16 blood samples analyzed for
amphetamines by CEDIA and GC-MS are shown in Table V. Referring to the cutoffof20 ng/mL, 13 samples were false negatives. These results show that the amphetamine urine test is not suitable for serum testing without adaptation. One reason is the lower cross-reactivity of commonly used amphetamine derivatives such as MDMA (69%), MDEA (21%), and MBDB (60%) compared with methamphetamine. Another reason is that serum proteins may cause interferences in this test, which is not the case with other assays. This
252
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Journal of Analytical Toxicology, Vol. 23, July/August 1999
became evident when the sample volume was increased on the BM/Hitachi 911 analyzer from 3 IJL (as recommended for the urine tests) to 18 IJL to increase sensitivity for amphetamine detection. With the so-adapted calibration, spiked urine concen- trations of 150 ng methamphetamine/mL were shown as
Table VII. Results of 47 Blood Samples Analyzed by CEDIA and HPLC for Benzodiazepines (ng/mL)
146 ng/mL, whereas the same concentration in spiked serum was shown as only 20 ng/mL. Lower concentrations than 150 ng methampheta- mine/mL could not be detected in serum at all, whereas detected urine concentrations went down to 10 ng/mL (results not shown). It will therefore be necessary to adapt the urine test for a sensitive detection of amphetamines in serum. A combina- tion of enhancement of sample volume on the BM/Hitachi 911 analyzer and preliminary treat- ment of the sample (protein precipitation) is very promising. Amphetamine precipitation with organic solvents is problematic because of the loss of the very volatile amphetamines during the dry- down step of the organic supernatant. Addition of methanolic HC1 prevents this amphetamine loss, but it also causes a loss of about 30% of cannabi- noids (12). However cannabinoids can be directly measured (as will be shown later in this paper). A good alternative for organic solvents is the pre- cipitation by zinc sulfate. This method was shown to work for sensitive determination (cutoff 25 ng/mL) of methamphetamine with TDx (2), but we did not test it here.
BZE In contrast to the amphetamines, BZE could be
detected at very low concentrations. All immuno- logical positive samples (at a cutoff of 10 ng/mL) could be confirmed by GC-MS. Two false nega- tives were detected by GC-MS with results of 155 ng BZE and 1100 ng BZE/mL, respectively. These two samples were extremely hemolytic and vis- cous. As we showed in the pretest, red color does not interfere with the assay. It was the viscosity of the samples that made it difficult for the pipette to get enough sample volume. Probably, in each case, no sample got into the assay at all. Therefore, blood samples have to be controlled for their vis- cosity before running the assays. Highly viscous samples should be centrifuged or treated with ace- tone or both. The linearity test (Figure 5B) showed the CEDIA BZE assay to be linear from 0 ng/mL to 1200 ng/mL. All assays showed that at low concentrations the linearity is best (with exception for THCCOOH). The good correlation of GC-MS and CEDIA results (Table VI, Figure 6A) show that a sensitive semiquantitative BZE detec- tion is possible in serum and whole blood without any adaptation.
Benzodiazepines Direct and sensitive benzodiazepine measure-
ment is also possible in serum. No false positives and no false negatives could be detected. This is probably because the CEDIA benzodiazepine assay does not use a monoclonal antibody like the other CEDIA tests (13), but polyclonal. Because of the
254
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Journal of Analytical Toxicology, Vol. 23, July/August 1999
very good cross-reactivities, benzodiazepines with very low thera- peutic blood levels can nevertheless be detected in serum at normal therapeutic range (flunitrazepam [109%], clonazepam [71%], alprazolam [220%]). The good cross-reactivity for metabo-
Table IX. Results of 102 Blood Samples Analyzed by CEDIA and GC-MS for Opiates (Continued)
lites like 7-amino-clonazepam (96%) and 7-amino-flunitrazepam (81%) further enhances sensitivity. Another point is that the CEDIA urine Benzodiazepine Hs assay offers the opportunity of automatic glucuronidase cleavage by the analyzer. Although glu-
curonidated products do not play such a big role in serum as in urine, this step might bring a slight but important increase in sensitivity for direct serum measurements. It was possible to confirm all positive (cutoff 10 ng/mL) immunological results with another method (Table VII). All CEDIA-negative samples were tested by another immunoassay (FPIA) and by HPLC. Although the determination of drugs of abuse works well with
- FPIA after acetone precipitation (14,15), exact cut- 311 offs for all benzodiazepines are not available.
- Huang (8) showed that the FPIA urine assay was - not able to detect 70 ng of several benzodiazepines - after serum extraction with butyl chloride. Table X - shows that the CEDIA test is able to detect 10
ng/mL for some benzodiazepines. Therefore, it 32o appeared necessary to check the immuno-nega- 803 tives by the more sensitive HPLC method, but no
_ false negatives were discovered. The use of poly- clonal antibodies is very good for the specific and sensitive detection of benzodiazepines, and this is the main point when doing an immunological screening. Because of the use of polyclonal anti- bodies the correlation of HPLC to CEDIA semi- quantitative results must be very noisy when more than one benzodiazepine and metabolites are pre- sent in the sample (Tables VII and X).
Methadone All samples positive for methadone could be
confirmed by GC-MS. Correlation of CEDIA semi- quantitation with GC-MS was low (Table VIII and Figure 6C). We did not distinguish between d- and /-methadone in the study. The addition of a fifth, very low calibrator to the standard curve may improve the correlation (communication of Boehringer Mannheim). Patients who receive methadone for substitution show usually blood levels of 50-400 ng/mL. A cutoff of 40 ng/mL will therefore be low enough to detect this.
Opiates Opiate detection with CEDIA was very specific
and sensitive in serum and blood. All positives were confirmed by GC-MS. Three false negatives were found at very low (see Cases ]9, 25, and 82, Table IX) concentrations. Two of these "false neg- atives" contained no morphine at all but did con- tain morphine glucuronide (cross-reactivity of the opiate test for morphine-3-glucuronide is 81%). As already shown in the pretest (Figure 2), CEDIA has good linearity at low concentrations. Semiquantitative CEDIA results were usually higher than GC-MS results even when the
255
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022
Journal o f Analytical Toxicology, Vol. 23, July/August 1999
Table X. CEDIA Display with Different Concentrations of Benzodiazepines in Spiked Serum Samples
* Limit high = requires dilution, out of calibration range.
different cross-reactivities of codeine, morphine, morphine glu- curonide, and dihydrocodeine were taken into consideration (Table IX, Figure 6B).
THCCOOH For THCCOOH, all CEDIA results higher than 10 ng/mL could
be confirmed by GC-MS. Linearity could be detected from 24 ng/mL up to 60 ng/mL, but GC-MS results did not correlate with CEDIA semiquantitation (Table IV, Figure 6D). Similar results for THCCOOH CEDIA assay were found by Cagle et al. (12) after acetone pretreatment of blood. With a cutoff of 25 ng/mL, they found FPIA semiquantitation to show a much better correla- tion with results of GC-MS than CEDIA. Sensitivity differences of both tests were not tested. Other authors chose a 25-ng cutoff for FPIA after acetone precipitation to get presumptive positives (< 2 ng THC/mL) by GC-MS (16). Although there is almost no corre- lation between CEDIA semiquantitation and GC-MS results, all CEDIA-positives could be confirmed by GC-MS. The low correla- tion is due to the CEDIA assay's cross-reactivity to other cannabi- noids and metabolites. Therefore, very sensitive detection of cannabis misuse in unpretreated serum/blood is possible for con- centrations far below 2 ng THC/rnL.
Conclusions
The data provided sufficient documentation to use the CEDIA screening technique without any adaptation as a very reliable method to detect analytically low but toxic concentrations of BZE, BENZ, MET, OPI, and THCCOOH in serum or whole blood without any pretreatment of the samples and without any changes at the BM/Hitachi 911 analyzer for routine urine mea- surements. Serum screening for amphetamines is not sensitive enough in the unchanged urine mode. It will require some adap- tation to a serum mode (probably a higher sample volume [BM/Hitachi 911] combined with protein precipitation of the sample). This procedure saves time and assists in economic use of very small sample specimens.
Acknowledgments
The authors wish to thank Boehringer Mann- heim for support and their kind assistance.
References
1. E.L. Slightom and H.H. McCurdy. Enzyme immunoassay: novel approaches to tissue and fluid analysis. In Advances in Analytical Toxicology, Vol. 1. R.C. Baselt, Ed. Biomedical Publications, Foster City, CA, 1984, pp 9-40.
2. T.F. Simonick and V.W. Watts. Preliminary evalua- tion of the Abbott TDx for screening of d-metham-
pbetamine in whole blood specimens. J. Anal. ToxicoL 16" 115-118 (1992).
3. W.M. Asselin, J.M. Leslie, and B. McKinley. Direct detection of drugs in whole hemolyzed blood using the EMIT dau urine assays. J. Anal ToxicoL 12:207-215 (1988).
4. H.W. Peel and B.J. Perrigo. Detection of cannabinoids in blood using EMIT. J. Anal. Toxicol. 5:165-167 (1981).
5. L.J. Lewellen and H. McCurdy. A novel procedure for the analysis of drugs in whole blood by homogenous enzyme immunoassay. J. Anal. ToxicoL 12:260-267 (1988).
6. C.E. McCord and J.R. McCutcheon. Preliminary evaluation of the Abbott TDx for benzoylecgonine and opiate screening in whole blood. J. Anal. Toxicol. 12:295-297 (1988).
7. E.L. Slightom. The analysis of drugs in blood, bile and tissue with an indirect homogenous enzyme immunoassay. J. Forensic Sci. 23: 292-303 (1978).
8. W. Huang and D.E. Moody, Immunoassay detection of benzodi- azepines and benzodiazepine metabolites in blood. J. Anal. ToxicoL 19:333-342 (1995).
9. C.W. Lee and A.M. Lee. Evaluation of the Abbott TDx Analyzer. J. Anal Toxicol. 13:50-56 (1989).
10. S. lwersen, A. Schmoldt, F. Schulz, and K. PCischel. Evidence of ges- tational heroin exposure by comparative analysis of fetal and maternal body fluids, tissues, and hair in a heroin-related death. J. Anal. Toxicol. 22:296-298 (1998).
11. S. Iwersen and A. Schmoldt. One fatal and one nonfatal intoxication with tranylcypromine. Absence of amphetamines as metabolites. J. Anal. Toxicol. 20:301-304 (1996).
12. J.C. Cagle, H.H. McCurdy, Y. Mary Pan, K.J. Ayton, W.H. Wall, and E.T. Solomons. Evaluation of the CEDIA TM DAU assays and the AxSym TM system for the analysis of cannabinoids in whole blood. J. AnaL Toxicol. 21: 213-217 (1997).
13. A.H.t3. Wu, E. Forte, G. Casella, K. Sun, G. Hemphill, R. Foery, and H. Schanzenbach. CEDIA for Screening drugs of abuse in urine and the effect of adulterants. J. Forensic Sci. 40' 614-618 (1995).
14. M. Bogusz, R. Aderjahn, G. Schmitt, E. Nadler, and B. Neureither. The determination of drugs of abuse in whole blood by means of FPIA and Emit-dau immunoassays--a comparative study. Forensic Sci. Int. 48:27-37 (1990).
15. R.D. Maier, M. Erkens, H. Hoenen, and M. Bogusz. The screening for common drugs of abuse in whole blood by means of EMIT-ETS and FPIA-ADx urine immunoassays. Int. J. Leg. Med. 105:115-119 (1992).
16. C.R. Goodall and B.J. Basteyns. A reliable method for the detection, confirmation, and quantitation of cannbinoids in blood. J. Anal. Toxicol. 19:419-426 (1995).
Manuscript received June 22, 1998; revision received October 23, 1998.
256
Dow
nloaded from https://academ
ic.oup.com/jat/article/23/4/247/852853 by guest on 14 January 2022