”mill I (ll m lllllllllllllllll 4

3 1293 00999 7549

 

 

 

 

This is to certify that the
thesis entitled
A DUAL COLUMN SCREENING METHOD

FOR SOME DRUGS OF ABUSE

presented by

Craig William Killingbeck

has been accepted towards fulfillment
of the requirements for

M - g . degree in Caldafl hiya/twat}

% ole/(AULS

MJ 6 C
Major professokr]

0-7639 MSU is an Affirmative Action/Equal Opportunity Institution

 

 

 

Date KAN 19f (9%0
<l

 

RETURNING MATERIALS:
IV1531_] Place in 500E drop to

remove this checkout from
LIBRARlES your record. FINES will
be charged if 500E is
returned after the date
stamped below.

 

 

 

:DE 0:93:93

, 4‘
14159 fififij
MAGIC 2

RARiOR’aiw A

 

 

A DUAL COLUMN SCREENING METHOD

FOR SOME DRUGS OF ABUSE

BY

Craig William Killingbeck

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Medical Technology

1986

‘léleQS‘U

ABSTRACT
A DUAL COLUMN SCREENING METHOD FOR SOME DRUGS OF ABUSE
BY

Craig William Killingbeck

A dual column gas chromatography method to screen
for drugs of abuse is described. A Shimadzu chromatograph
with a SPB-l wide-bore capillary column as a primary or
screening column and a DB-l capillary column as the second
or confirmatory column both connected to a flame ionization
detector is used. A library of retention indices was
developed for each column on an IBM-PC XT. The results
showed excellent precision, linearity, and recovery using
the ultra rapid extraction method described. A comparison
study on urines assayed by Syva's Urine Drugs Of Abuse kit
was carried out which showed the gas chromatographic
method to have great potential for drug screening. The
other two benefits of this method are simultaneous
confirmation of results and quantitation of drugs present.
The method may prove to be rapid, inexpensive, and suitable

for operation in a clinical laboratory.

ii

TABLE OF CONTENTS

INTRODUCTION.. ............................... Page
Materials and Methods ........................ Page
Results ..... ............ ................... ..Page
Discussion ................................... Page
Summary and Conclusions ...................... Page
BIBLIOGRAPHY... ...... . ....................... Page

iii

1
9
13
28
34

35

Table

Table

Table

Table

LIST OF TABLES

Drug Standards Used In Study .......... Page
Libraries of Retention Indices........Page

Recovery Studies on SPB-l Column
After Extraction.............. ........ Page

Comparison Study of 27 Urine
Specimens ............. . ............... Page

iv

14

23

24

Figure

Figure

Figure

Figure
Figure

Figure

LIST OF FIGURES

Chromatogram Of A Single Standard (Above)
And Mixed Standards (Below)... ....... .....Page

Chromatogram Of Morphine With
Hydrocarbons (Above) And Morphine (Below).Page

Chromatogram Of A Drug-Free Urine After

ExtractionOOOOOOOOOOOO00.00.00.000... ..... Page
Chromatograms Of Recovery Studies. ....... .Page
Linearity Study of Hexobarbital ........... Page

Chromatogram Of Urine Positive For
Tetrahydrocannabinol...... ................ Page

l6

17

20
22

25

27

GC

FID

EMIT

HPLC

MS

RT

LIST OF ABBREVIATIONS

gas chromatograph
flame ionization detector

enzyme multiplied immunoassay technique
trademark of Syva Co., Palo Alto, Ca.

high pressure liquid chromatography
mass spectrometer

retention time

vi

INTRODUCTI ON

Interest in screening urine specimens for drugs of

I

abuse is still expanding. Not only are screening
programs currently being used to monitor the success of
drug rehabilitation programs, but also private industry
is starting a large number of drug screening programs for
the purpose of enhancing productivity by screening out
potential problem employees. Many clinical laboratories
are considering the addition of a drugs-of—abuse screen.
A drugs-of—abuse screen could include the following:
amphetamines, barbiturates, cocaine, pethidine, diazepam,
hydromorphine, LSD, marijuana, mescaline, methadone,
morphine, oxycodone, phencyclidine, and propoxyphene.3
Methods currently in use to screen for drugs—of—abuse
include colorimetric, immunoassay, radio-immunoassay, and
hemagglutination inhibition. Enzyme immunoassays are now
available for the nine major families of drugs of abuse.
Among the chromatographic methods available are thin-layer,
high-pressure, and gas-liquid. Further in gas-liquid
chromatography (GC) there are different detector systems

available including nitrogen-phosphorus, flame ionization,

electron capture, and mass spectrometry.

2

Colorimetric methods are available for some drugs of
abuse including barbiturates, morphine and methaqualone.
These methods have advantages of low cost, small amount of
equipment, and no need of highly-trained personnel. The
disadvantages of these methods are a distinct lack of
specificity and the ability to assay only one family of
drugs at a time. The procedures are also labor intensive,
which does not fit well into the routine of the clinical
laboratory.

Thin-layer chromatography, one of the first developed
broad spectrum screens for drugs, is currently the most
widely used method.5 More than one specimen can be
screened simultaneously. It is an inexpensive method to
screen for the presence of drugs prior to undertaking the
more elaborate methods still to be described.6 The
negative aspects of thin-layer chromatography are its labor
intensiveness and the many subjective aspects of
interpretation based on color development. This is a
method from which results must always be confirmed by
another independent procedure.7’8

Enzyme immunoassay and radioimmunoassay techniques
have been available to the clinical laboratory for many
years. These techniques, particularly Enzyme Multiplied
Immunoassay Technique (EMIT), have been developed for the
major families of the drugs of abuse which are as follows:

opiates, barbiturates, amphetamine, cocaine and metabolites,

3
benzodiazapines, propoxyphene, methadone, phencyclidine,
cannabinoids, and ethyl alcohol.5 The precision of these
techniques is adequate for routine clinical use.
Inexpensive automated hardware is readily available and
easily operable, which makes this procedure affordable
on a cost-per-test basis in terms of both labor and
reagents. Since one EMIT assay is much like another and
all are easily automated, training is not complex. The
drawbacks to this system are tied to the antibodies
developed for the assay. Laboratory scientists are
dependent on the manufacturers of kits to develop the
antibodies. Not all antibodies are equally specific.
False positives, such as dextromethorphan reacting with
the opiate assay, are an ever-present problem. Since more
than one drug is commonly present, many different kits must
be used. Also, only those classes of drugs that the
antibody is directed at may be detected. Of the thousands
of drugs available antibodies have been made to relatively
few.

High-pressure-liquid chromatography (HPLC) is not the
chromatographic method of choice for drug studies.l Its
chief use in studying drugs is for use on those compounds
with low volatility or compounds that are heat labile.6
HPLC has not been used as extensively as GC in monitoring
drugs due to detector systems that are not as sensitive

or versatile as those used in gas chromatography.

4

A choice that currently makes a lot of sense is gas
chromatography. Its potential as a screening and
confirmatory tool is almost limitless. It is a very
specific methodology, and the sensitivity is acceptable
for measuring clinically significant levels of drugs.10

Many choices exist for making use of GC. The first
choice which must be considered is the specimen itself.
While in forensic work or therapeutic drug monitoring
the specimen of choice may be tissue or blood, urine is
the commonly used specimen in drugs of abuse.3 Since
absorption into tissues or active clearance by the kidneys
takes place, urine is the best place to find evidence of
drugs of abuse in a screening environment. Although
quantitation in the urine is not perfect, there are some
relationships that can be established by measuring drugs
in urine.

Many methods of extracting drugs are currently used.
Most involve several steps including initial extraction
in solvent, centrifugation, evaporation to dryness, and
solution in solvent before analysis. Simple extraction
techniques involving only two steps have been successfully
employed.12 Solid phase extraction using disposable
supplies is rapid and efficient.

Derivatization is sometimes necessary to make a parent

compound more volatile, separable, stable, and detectable.l3

Volumes have been written about techniques and compounds

5
such as BSTFA, a silylating agent, which is currently used.
However, derivatization has the disadvantage of adding
timely steps to the drug screening procedure and can be
avoided. Many methods of drug screening on the GC that do
not involve derivatization are currently available and will
be discussed.

There is no single correct method for choosing a column
for GC, but there are many considerations.l4 Capillary
columns have the advantages of high reproducibility and
decreased analysis time.15 The literature is replete with
studies of retention times by both standardized and
nonstandardized methodologies for nonderivatized drug

analysis.6'15'l6'l7

Making use of literature and choosing
from columns with a variety of polarities such as OV-l,
SE-30, or carbowax, one can empirically rather than
theoretically choose a column and analyze the results of
retention time studies for peak separation, length of
analysis time, and reproducibility.

A very attractive option to consider when choosing a
capillary GC column is the wide-bore capillary columns.
A wide-bore column is now available and licensed by the
FDA for diagnostic use with drugs. Research has documented
that the column is both inert and efficient enough to rival
resolution provided by the narrow-bore columns.l8 A major

advantage of this column is that a splitter is not needed.

The splitter is run wide open, making the column much

easier to use in the clinical laboratory setting by
personnel who are not specially trained in gas
chromatography.

For the experienced chromatographer continuing
advances in capillary column technology will offer even
more choices. Columns have recently been prepared which
have been coated with monomers specifically designed to
separate drugs of abuse such as barbiturates,
benzodiazepines, and LSD. 9 This work on columns plus work
presently being done with detector technology will continue
to enhance the possibilities of using GC in the clinical
laboratory setting for drug screening programs.

An advantage of GC is that it is widely used as a
confirmatory method for drugs-of-abuse screening. Many
studies report success using a dual column GC method as
both a screening and confirmatory method.20'21 Ideally,
the second or confirmatory column should have a different

1 2 . . .
' 2 Instrumentation 15 available so

phase and polarity.
that a single sample injection can be entered into both
columns, or the columns can be run on different GC
instruments using one for screening and the other for
confirmation. Each column will have its own detector and
integrator.

Although drug screening by definition is principally

qualitative, quantitative data is obtainable. The routine

availability of quantitative data from GC is due

7
principally to the relatively recent advances in capillary
column technology.1

The only technique that can improve the sensitivity
and selectivity of GC is coupling the GC to a mass
spectrometer (MS).l7 This method yields more information
at the nanogram level than any other method23 plus it has
additional benefits of the mass spectrograph, which gives
the molecular weight of the parent compound from the
molecular ion and the structure of the parent compound
from the fragment ions. Several important factors will
slow the usefulness of the GC-MS in the clinical
laboratory. It is an expensive instrument to both buy

and maintain.5'23

Throughput of the instrument is quite
limited. It is a difficult instrument to operate, therefore,
calling for highly trained personnel. As a highly
sophisticated confirmatory method in an environment in
which 24 hour per day high speed throughput is not
essential, GC-MS may be the instrument of choice.
The detector of choice is the flame ionization
detector (FID) when using the GC for drug screening. This
is due to high sensitivity, broad linear range, simple
and reliable construction, and a broad general utility for
organic compounds.l'24
It goes almost without saying that whenever large

amounts of complex data are generated, such as with GC, a

computer does the same things as a human interpretor with

8
the advantages of electronic data smoothing, rapidity, and
data storage.

With the current interest in drugs-of—abuse screening
there is an opportunity to develop a methodology to meet
the needs of a hospital clinical laboratory. The method
must be rapid and inexpensive when purchasing equipment,
supplies, and labor. The method must be able to provide
confirmation that is equivalent or better to the current
standards of practice in drug screening. A dual column GC
with a wide-bore capillary column for screening following
a rapid extraction procedure will be developed to meet
these criteria. It will include a second capillary column
for confirmation, a quantitation procedure, and data
reduction on an IBM PC-XT with existing software. The
purpose of this study will be to assess the potential of

such a system for drugs-of—abuse screening.

Materials and Methods

Standards for this work were purchased from Supelco.
The drugs were diluted by the manufacturer to 1 mg/ml
in methanol. Table l is a list of drugs which were used
as standards in this study. Standards were prepared for
use by diluting from 1 mg/ml to 100 ng/ul using glass
distilled methanol. The standards then became the working

standards used throughout the project.

Table 1 Drug Standards Used In Study

Drug Standard Working Standard
l-Amphetamine 1 mg/ml lOO ng/ul
Methamphetamine 1 mg/ml lOO ng/ul
Barbital 1 mg/ml lOO ng/ul
Amobarbital 1 mg/ml lOO ng/ul
Pentobarbital 1 mg/ml lOO ng/ul
Secobarbital 1 mg/ml lOO ng/ul
Hexobarbital 1 mg/ml 100 ng/ul
Mephobarbital 1 mg/ml lOO ng/ul
Phenobarbital 1 mg/ml lOO ng/ul
Methaqualone 1 mg/ml 100 ng/ul
Methadone-HCI 1 mg/ml 100 ng/ul
Codeine Sulfate 1 mg/ml 100 ng/ul
Tetrahydrocannabinol 1 mg/ml 100 ng/ul
d'Tetrahydrocannabinol 1 mg/ml lOO ng/ul
Cannabinol 1 mg/ml 100 ng/ul

The GC used for this project was a Shimadzu GC-9A.
The instrument had a dual injection port, and the dual
detectors were of the FID type. The single integrator on
the instrument was a Chromatopac C—R2AX also by Shimadzu.
It was entirely programmable, but it had no data storage

capabilities.

10

Two columns were used on the GC. The primary or
screening column was a Supelco, wide-bore SPB-l capillary
column which had a .75 mm ID and had a bonded SE—30 phase.
The SE-30 phase of this column is the phase of choice for
drug screening.18 The splitter was run wide open. The
carrier gas used was helium. The linear velocity of this
system was 41.7 cm/sec. and the number of theoretical
plates 18,050.

The second or confirmatory column was a DB—l from
Hewlett-Packard. This column was a thin-film DB—l bonded
phase .1 mm thick that is a narrow bore 60 m long with an
internal diameter of 25 mm. The carrier gas used was
helium and the linear velocity was 26.2 cm/sec. with a
split ratio of 1:47 and a split flow of 35.9 ml/min.

The oven temperature throughout the study was programmed
to begin at 115°C and increase by 6°C per minute up to
285°C. The injector temperature was set at 300°C and the
detector temperature was 300°C. The run was programmed
to end at 40 minutes.

The integrator was also programmed to run for 40
minutes. It was programmed for the following: width 5,
attenuation 5, drift 1, slope 2000, minimum area 1000,
chart speed 15 mm/min., and slope weight 100.

The extraction method used was designed to be quick
and to make use of a minimum number of steps yet insure

70% recovery of the drugs in question.12 Two milliliters

ll

of urine were first made basic with .2 ml of 5N NaOH in a
very pointed 10 ml conical centrifuge tube, and 50
microliters of chloroform were added for the extraction.
The tube was mixed by vortex for 30 seconds and centrifuged
for 5 minutes. This procedure was repeated with the urine
made acidic with 6N HCl. Injection was made directly from
the solvent layer in the point of the centrifuge tube.

Injections were made by using hexanes as the solvent
front, injecting a mixture of hydrocarbons C10 through
C26 as markers of retention time and hexobarbital as an
internal standard for quantitation. When drug standards
were injected, .4 microliter aliquots were next drawn into
the 10 microliter syringe. When urine specimens were used,
a l microliter aliquot was taken from the acidic and basic
extractions and injected with the hexanes, hydrocarbons,
and hexobarbital.

The Shimadzu integrator was not directly interfaced
to a data reduction device. The retention times and peak
area which had been calculated by "area normalization-
method 41" of the C-R2AX integrator were then entered into
an off-line computer. The computer was an IBM PC XT which
was a stand-alone system running programs designed and
written by Sweeley et al. at Michigan State University
for the metabolic profiling of urine. The only modification

to the programs was to allow data entry by CRT rather than

by interface. The computer first stores the data from

12

each run in files for future data reduction. In metabolic
profiling standards for retention time can be found in the
urine, but in drug screening they are co-injected with

the sample. The peaks found in the sample are then
identified as belonging to the library previously built

or as a significant peak not identified. The major
component of identification is retention time, but in
addition a confidence factor is printed in percentage
which takes into account area calculations, linkage to
other peaks which are the co—injected standards in this

case, and concentration.

   

Results

The first step undertaken in this project was to
build a library of retention indices for each column used
in the study. The drugs previously listed and hydrocarbons
C10 through C26 are the individual members of this library.
Table 2 is the printed output of a library built on the
IBM PC which contains retention indices of drugs and
standards when using the Supelco SPB-l column and the
Hewlett-Packard DB-l column.

To obtain the values which were used to build the
library each standard was run alone through the entire
temperature program of the CC on the SPB-l column. Figure 1
illustrates such a run with the methadone standard. It was
run alone to guarantee that there was no chance of error
in identifying the drug in question by its retention time
alone and then run with mixtures of standards. The goal
was to run the drugs until three values were obtained
within .005 minutes of each other. This level of precision
was not obtainable. However, all three values obtained for
each drug were within .01 minutes. Subsequently, all three
values were averaged and entered into the library. This
procedure was repeated for the second column.

When building the library it became apparent that
morphine sulfate and hydrocarbon 24 both came off the
column at virtually the same time. Figure 2 shows two

back-to—back runs on the same day with the same conditions

13

14

Table 2

1. C-12

2. l-Amphetamine

3. Methamphetamine

4. C-14

5. C—16

6. Barbital

7. C-18

8. Amobarbital

9. Pentobarbital
10. Secobarbital
11. C—20
12. Hexobarbital
13. Mephobarbital
14. Phenobarbital
15. C—22
16. Methaqualone
l7. Methadone-HCl
18. Codeine Sulfate
19. Cannabidiol
20. Morphine Sulfate
21. Tetrahydrocannabinol
22. d'Tetrahydrocannabinol
23. Cannabinol
24. C—26

Libraries of Retention Indices

SPB-l

321.70)
333.30)
391.20)
420.50)
714.20)
( 820.20)
(1063.00)
(1224.30)
(1262.20)
(1336.70)
(1410.20)
(1466.20)
(1530.20)
(1617.00)
(1734.80)
(1928.70)
(1951.00)
(2309.80)
(2327.50)
(2311.00)
(2379.30)
(2406.30)
(2492.30)
(2569.00)

AAAAA

DB-l

557.70)
573.50)
629.70)
657.60)
( 951.50)
(1070.19)
(1319.60)
(1500.10)
(1541.20)
(1639.30)
(1695.10)
(1762.00)
(1831.40)
(1930.10)
(2041.40)
(2265.89)
(2289.10)
(2634.10)
( - )
(2673.00)
(2754.00)
(2783.60)
(2878.20)
(2948.70)

“AAA

15

Figure 1 Chromatogram Of A Single Standard (Above) And
Mixed Standards (Below)

 

*‘2

7

 

 

 

7/
asuoasaa HOLDEIEUHL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l___..— __._.__.- -

. :

.. ___,
L .. m.m_N_c”T/IL
+N 7 +N +H +H + AeflzvaHe

541315 1 j

 

 

 

 

 

—1

 

6 _
1 _ “G
.J._H
. .1. “I
I 21
I mmuo .HH I mi
I emuo .oH I m
mmno .m I “a
omlu .w "I3
H Hepflbumnoaa .s M “a
male .m ”mm
Hmpflbumm .m Mus
mane .4 Mus
eauo .m WI
ocflemwozmemnpoz .m _w

mcoemnpmz .H

 

Figure 1

 

 

IBOHmmV ocHQQHOE pc< Iw>OQ¢I muonumoouphm Suflz ocflflmuoz mo EMHOODMEOHSU N mudmflm

opmmasm ocflamuoz .m I

/
HSNOdSHH HOIDHIHG

II I

II I
l -- fij II (III/I7,

CHEVOEHB

 

 

 

 

 

 

 

 

 

 

7
.I.._ I _
I I II7, m
I ,, I
I _ E
f ,3
I _ , r
I\ III, m
\4 _ . _.
.: m w I I __H
II I I E
II m I I d
I I m
I I
m I II . I.
mmuo .m mfluo .4 I II
opmmasm ocflamuozlemlu .w mHIU .m m a
NNIU .m vHIU .N
, ONIU .m NHIU .H

18

which illustrate this problem. In order to make the entire
library useful, hydrocarbon C24 must be dropped from the
retention time standards. Cannabidiol was dropped from
the study when it was mishandled and all standard was lost.

To test whether or not the chosen extraction method
would be acceptable, a urine which was free of drugs was
extracted and injected on the column. It had been decided
that if too many extraneous peaks were present, the job of
entering the data into the computer by hand would not be
practical. The urine used was the author's and was free
of drugs except possibly caffiene. Figure 3 is one of
several extraction runs made using this urine specimen. In
a 25 minute run, the significant time period for all the
drugs in the study, only five extraneous peaks were
encountered.

Acceptable recovery of drugs was determined to be
70%. Recovery studies were carried out on all drugs in
the study on the Supelco SPB-l column. All drug working
standards were run in mixtures on the GC. The peak areas
of each drug standard were recorded. Next the author's
urine was spiked with an equal amount of the same drug
mixtures. The extraction procedure was carried out on
the urine specimens, and the peak areas obtained were
divided by the peak areas obtained from the previous
runs of unextracted standards. Figure 4 is a Chromatogram

of drug standards run on the GC and a Chromatogram of

 

19

Figure 3 Chromatogram Of A Drug-Free Urine After Extraction

 

20

 

 

 

em om ma NH m
_ I _ I lit _
3?: were
m

xmmm emfleflucmeuco .
xmmm emflmflucmeflco .
xmmm emflmflucmefleo .
xmom cosmeucmeflco .
xmmm emflmuucmeflca .

HNMV‘LO

 

 

 

 

HSNOdSHH HOLOHLHG

 

Figure 3

 

21

Figure 4 Chromatograms Of Recovery Studies

1. C-12

2. C-l4

3. C-16

4. C-18

5. C-20

6. C-22

7. C-24

8. Tetrahydrocannabinol
9. d'Tetrahydrocannabinol
10. Cannabinol
11. C-26

22

 

 

 

wII_I.—e — —fl-‘f “Si:7:5_I§iII:I::
[>__é.__=:_ __._. —_ :I _ _;:II .51 D —# I II——I-:-I’: I I1)
‘f::)

 

 

Ln
III _: I~ _ 7:”? —— Q. IIII IIIeIIII III II“;
Q'ZJ
<14 ”fl __H ___ _ _ w sf

 

 

 

 

 

 

 

_——‘ :.' _ —-_—I—_‘ ;’ __ __ II-
cw- III ...:—_. -——“——_..___ -.—__-‘—_._.
. ; ‘
_____—__.:-_____._ _ .. ._ ____ II
&-——hfl_ _:____ _. _‘_,_ _ :—
_ _ _,_—— ___ _ -

LIIIIIIIIIIIIIIIIIIWI_”new L4j4;;j___Humfl_I_HWI_WIfl_
HSNOdSHH HOIDHIHG ] HSNOdSHH HOLDEIHG 1

Figure 4 (cont'd)

drug-free urine spiked with an equal amount of drug

standard. Both Chromatograms have peak areas included.

It was evident at this time that the barbiturates

and morphine sulfate were not being recovered adequately

from the urine specimen.

Ramsey et al.12 was modified.

At this point the method of

The mixing and

centrifugation steps were lengthened from 5 seconds and

2 minutes to 30 seconds and 5 minutes,

improved recovery of all drugs.

respectively. This

Next an additional

extraction with 6N HCl substituting for 5N NaOH was carried

out, and the barbiturates were recovered, although

phenobarbital recovery was only 55% and was not considered

acceptable. Table 3 demonstrates the recovery of all drugs

in the library except morphine.

Table 3 Recovery Studies on SPB-l Column After Extraction

l-Amphetamine
Methamphetamine
Barbital

Amobarbital
Pentobarbital
Secobarbital
Hexobarbital
Mephobarbital
Phenobarbital
Methaqualone
Methadone-HCI
Codeine Sulfate
Morphine Sulfate
Tetrahydrocannabinol
d'Tetrahydrocannabinol
Cannabinol

125%
99.4%
108%
80.6%
81.7%
120%
118%
119.7%
55%
86%
79%
82%

No Recovery
99.7%
72%
105.8%

24

To assess the methodology for potential quantitative
work it was necessary to study linearity and precision.
Since hexobarbital had been chosen as the internal standard
for quantitation, it was used as a standard in this study.
Figure 5 is a graphic representation of a linearity study
carried out on the range of concentration from 10 ng/ul to
320 ng/ul. This range is a good representation of the
clinically significant range of drugs of abuse in urine.
Phenobarbital was also studied between 10 ng/ul and 80 ng/ul
and found to have acceptable linearity.

Finally, twenty seven urine specimens which had been
screened for drugs of abuse at St. Lawrence Hospital in
Lansing, Michigan were analyzed for drugs of abuse. Table 4
contains the results of these analyses. The screening
method used at St. Lawrence Hospital was the Syva EMIT
Urine Drug Assays which were performed on a Cobas Fara from

Hoffman La-Roche Diagnostics.

Table 4 Comparison Study of 27 Urine Specimens

Drug Cannabinoids Barbiturates Negative
Method Emit GC Emit GC Emit GC
No. of Specimens 14 10 l 1 12 12
% Concurrence 71% 100% 100%

The results in Table 4 show that all twelve urines
screened as negative by the EMIT Urine Drug Assays were
also screened as negative by the GC method. A single urine
specimen contained barbiturate according to the EMIT, and

the GC identified phenobarbital in this specimen. The EMIT

25

ng/ul

360 "

320 "

280 —‘

24o —-«

200 —“

160 —‘ 1.

120 —‘

80“ .

40-1

 

 

I l l l T’ T_ 'T
(J 40 80 120 160 200 240 280

PEAK AREA X 103

Figure 5 Linearity Study of Hexobarbital

 

 

26

system identified fourteen urines as being positive for
cannabinoids. The GC was able to pick up ten of those

same urines with a positive screen for tetrahydrocannabinol.
Figure 6 is a Chromatogram of a urine specimen for drug
screening which was positive for tetrahydrocannabinol. As
in many specimens positive for tetrahydrocannabinol, there
are many small peaks detected in the same area that are

unidentified.

27

HSNOdSHH HOIOHIHG

 

OH

Hocflnmccmooupmzmupoe Mom o>Hpflmom ocHHD m0 EmHWODmEOHQO o ousmflm

 

mmlu
Hocflnmccmooupmamupme
vNIO

NNIO

Hopflnumnoxmm

. omlv

wHIU

wHIU

wHIO

NHIU

 

 

 

 

.\

I] I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Discussion

Overall, the GC method appears to be adaptable to the
hospital clinical laboratory as a screening method for
selected drugs of abuse. By using a wide—bore capillary
column with the split run wide open the technical knowledge
necessary to operate the GC equipment was minimal. The
techniques needed to operate the equipment properly are
the same ones used to operate many other pieces of equipment
in the automated clinical chemistry lab of an average
hospital. The building of the libraries of data was
straight forward with no real problem. By running each
standard separately and then in groups a very practical type
of experience is gained, and one can get a feel for the GC
and its precision. Although this step is time consuming,

30 minutes per standard for 2 libraries, and expensive,
$30.00 per standard, it is the best way to develop confidence
in the ability of the system to identify drugs of abuse by
retention time alone.

The extraction method12 chosen was simple and rapid
and would fit nicely into the routine of a stat clinical
laboratory. Many other extraction methods or additional
extraction steps could be used to improve recovery of
drugs of abuse, but since the primary goal of the study was
not to compare recovery methods, a simple goal of 70%
recovery was used. The earliest extraction work in the

study was an indicator that the method should be slightly

28

 

29

modified to increase the percentage of recovery. The
modification suggested by Ramsey et al.12 as potentially
necessary was to increase the mixing time on the vortex

once solvent was added from 30 seconds to 1 minute and to
increase the centrifugation step from 2 minutes to 5 minutes
to improve the separation of the solvent front from the
urine specimen. A layer of precipitate always formed at

the junction, but the needle of the injection syringe was
able to pass through the urine layer and precipitate with

no contamination and withdraw the 1 ul specimen directly.
The failure of this method to extract morphine will be
addressed. An acid extraction also was added which improved
recovery of the barbiturates.

The excellent precision of retention times of standards
during the building of the library guarantees that acceptable
reproducibility is achievable for this method in identifying
drugs. A second column of different polarity will be an
adequate confirmatory method.

The recovery and linearity studies done show that
quantitation with this method is possible. A recovery study
was done on every standard, and except for morphine and
phenobarbital all results fell within the guidelines chosen
for this study. Linearity studies indicate that the method
for at least 2 standards is adequate in the clinically
relevant ranges. Retrospectively, a linearity study for
every drug standard should be done to leave no unanswered

questions.

 

30

As mentioned quantitation of drugs of abuse using the
GC is a relatively new development.16 Care must be taken
in the selection of GC instruments if quantitation is a
consideration since not all equipment may be of necessary
quality. Of course maintenance and quality control
procedures will also be necessary.

As shown in the results section, the method picked up
most of the drugs identified by EMIT. As a predictor of
negative drug screens, it matched in all cases the EMIT
system. There were few positive specimens to match most
of the screen drugs on GC. However, phenobarbital was
picked up. Cannabinoids were picked up in 10 of 14 positive
specimens or 71%. In some cases where cannabinoids were
not identified by GC there were peaks present which were
similar in retention time and size to peaks which had been
identified as tetrahydrocannabinol. This study employed
only 4 of the major cannabinoids. The Syva EMIT Urine Drug
Assays detect at least 39 of the greater than 400 THC
metabolites that exist and are already prebuilt into some
GC-MS libraries according to the director of the local
toxicology laboratory.27 It is very likely that
unidentified peaks on the GC would turn out to be
cannabinoids. It is also possible that many cannabinoid
metabolites in small quantities, an area of less than 1000
on the GC, could be reacting in combination with the

antibody in the Syva EMIT Urine Drug Assay.

 

31

Problems were identified with the method and several
changes are recommended. First of all, it would be
necessary to look carefully at the equipment used. This
work was done on a single GC with one integrator. Either
a second such GC should be set up for confirmation or,
better, a dual column system should be used. For maximum
efficiency a GC with an injector port that would split the
injection evenly to 2 columns each with their own detectors
and integrators is preferable. Such a system is available.28

As mentioned previously in this paper, the second
column used for confirmation should be of a different
polarity than the first. The polarity of the DB-l column
is different than that of SE-30, but it is not sufficiently
different to cause the drugs to elute off the column in
different order. A solution would be to use a second wide—
bore capillary column coated with SPB-35.

Data reduction was tedious, and for the purposes of
this work was done principally off-line by the operator.
The IBM PC programs of Sweeley et al. worked, but it took
too much time to enter the retention times of all standards
and significant unknown peaks to really save any analysis
time. The obvious solution would be to interface the
integrator to the IBM PC so that data acquisition would be
automatic. The interface is a standard feature of the
programs, and now that the method has been shown workable

this modification should be done. Also, the software should

 

32

be modified to reflect predictive values of the second
column confirming the findings of the first.

The extraction procedure needs modification to allow
increased recovery of opiates, in particular, morphine.
Morphine must be extracted first in strong acid to cleave
off glucuronide groups which make these conjugates very
highly water soluble.4 Then the final extraction takes
place best at pH 9.0. More control of the pH appears to
be the answer to obtaining greater recovery of phenobarbital.

Since 400 cannabinoids exist, the problem of identifying
these peaks cannot be taken for granted. At the very least
this GC method should cover the 39 cannabinoids commonly
found in urine by the Syva EMIT Urine Drug Screening kit.
Further studies of potential cannabinoid metabolites will
have to be undertaken and may be in process now due to the
intensity of interest in this area presently.

Another problem encountered was a shift of retention
times during the study. During the time that urines were
being assayed and compared to EMIT results, retention
times on the SPB-l column shifted longer by as much as
.7 minutes. Since this condition must be due to flow rate,
the apparatus was taken apart and cleaned including washing
the column. The air pressure was also adjusted, but the
retention times never returned to their original values.

At this time it was necessary to rerun the entire library

and load it into the IBM PC. The shift in retention times

 

 

33
made it very difficult to use the computer for peak
identification. The shift could also be partly due to
thermal shock of the column due to excess heating the
column to 300° at greater than 25°C per minute and cooling
the column by opening the oven door.29

Both this equipment and other GC equipment in the
laboratory have been run extensively for extended periods
of time with no shifts in retention time being encountered.
This problem should not be considered as a detriment to the
methodology. A good program of preventive maintenance and
tight quality control coupled with a moderate level of
experience running the equipment would entirely eliminate
such shifts.

This work demonstrates the potential for this drug
screening methodology in the clinical laboratory. With
the modifications suggested the problems would be resolved.
The method for screening drugs of abuse herein described
should be considered as a starting point for the novice

at drug screening with the GC.

 

 

 

 

 

Summary and Conclusions

There is a need to develop methods for clinical
laboratories to meet the demand for drugs-of—abuse screening
programs. The method must be inexpensive, rapid, accurate,
and operable by clinical laboratory personnel. Key features
of such a method would be confirmation of results and
quantitation. Of the many methods currently available for
consideration gas chromatography is one of the best choices.

This purpose of this study was to develop such a method
on a gas chromatography system. Drugs of abuse can be
identified and these results confirmed by the dual column
gas chromatography method described. The study includes a
description of excellent precision and linearity. A study
in tandem with an EMIT (trademark of Syva Co.) Urine Drug
Screening kit proved the method to be accurate. Data
reduction on an IBM PC using programs developed by Sweeley
et al. at Michigan State University for metabolic profiling
is described. Also an ultra rapid extraction method which
could easily fit into the routine of any clinical laboratory
is used. This study shows that such a method is feasible,
and it provides a starting point for the laboratory
scientist who needs to develop a practical system for

drugs-of—abuse screening.

34

 

 

 

 

BIBLIOGRAPHY

 

10

11

B I BLI OGRAP HY

Jarvie DR, Simpson D, Gas Chromatographic Screening
For Drugs And Metabolites In Plasma And Urine. Ann
Clin. Biochem. 1984; 21: 92—101

Goldbaum LR, Santinga P, Dominguez AM, A Procedure
For The Rapid Analysis Of Large Numbers Of Urine
Samples For Drugs. Clinical Toxicology 1972; 5: 369-
379

Sohn D, Byers J III, Cost Effective Drug Screening
In The Laboratory. Clinical Toxicology 1981; 18: 459—
469

Sunshine I, Methodology For Analytical Toxicology.
CRC Press, Inc. 1985; Volume I: 1—465

Bittikofer JA, Benton B, Rapid Drug Screening In The
Hospital Clinical Laboratory. Marion Labs Monograph
1982; 3-26

Sunshine I, Toxicology. CRC Press, Inc. 1978;
Volume I: 3—387

Sunshine I, Methodology For Analytical Toxicology.
CRC Press, Inc. 1985; Volume II: 1—238

Mule SJ, Routine Identification Of Drugs Of Abuse In

Human Urine I. Application Of Flourometry, Thin Layer
And Gas Liquid Chromatography. J. Chromatogr. 1971;

55: 255—266

Dawkins BG, McLafferty FW, The Mass Spectrometer As A
Detector For High Pressure Liquid Chromatography. GC
and HPLC Determination Of Therapeutic Agents. Marcel
Dekker, Inc. 1978; Volume 9, Part 1: 259-275

McChesney JD, Drugs Of Abuse And Overdose. GLC and
HPLC Determination Of Therapeutic Agents. Marcel
Dekker, Inc., Tsuji K 1978; Volume 9, Number 2: 417—473

Willette RE, Interpreting Cannibinoid Assay Results.
Syva Monitor 1986; Volume 4, Number 1: 1-7

35

 

 

12

13

14

15

16

17

18

19

20

21

22

36

Ramsey J, Campbell DB, An Ultra Rapid Method For The
Extraction Of Drugs From Biological Fluids. J.
Chromatogr. 1971; 63: 303-308

Perry JA, Feit CA, Derivitization Techniques in Gas-
Liquid Chromatography. GLC and HPLC Determination Of
Therapeutic Agents. Marcel Dekker, Inc. 1978; Volume 9,
Part 1: 137-208

Cohen JL, Column Selection In GLC. GLC and HPLC
Determination Of Therapeutic Agents. Marcel Dekker,
Inc., Tsuji K, Morozowich W 1978; Volume 9, Part 1:
47-82

Lora-Tamayo C, Rams MA, Chacon JMR, Gas Chromatographic
Data For 187 Nitrogen Or Phosphorus-Containing Drugs
And Metabolites Of Toxicological Interest Analyzed On
Methyl Silicone Capillary Columns. J. Chromatogr. 1986;
374: 73-85

Plotczyk LL, Application Of Fused-Silica Capillary
Gas Chromatography To The Analysis Of Underivatized
Drugs. J. Chromatogr. 1982; 240: 349-360

Plotczyk LL, Larson P, Advances In Fused Silica Column
Technology For The Analysis Of Underivatized Drugs.
J. Chromatogr. 1983; 257: 211-226

Monograph, Drug Screening Analysis On Wide Bore Capillary
Columns. The Supelco Reporter 1984; Volume III,
Number 5: 8-9

Chow WML, Caddy B, Preparation Of Novel, Curable
Capillary Gas Chromatographic Systems And Their
Application To The Analysis Of Underivatized Barbiturates
And Other Controlled Drugs Of Forensic Interests. J.
Chromatogr. 1986; 354: 219-229

Perrigo BJ, Peel HW, Ballantyne DJ, Use Of Dual Column
Fused Silica Capillary GC In Combination With Detector
Response Factors For Analytical Toxicology. J.
Chromatogr. 1985; 341: 81-88

Levy JM, Dual Capillary Column Analysis Of Complex
Oxygenated Mixtures. Journal Of Chromatographic
Sciences 1984; Volume 22: 149-152

Aggarwal V, Santinga P, GC For Drugs Of Abuse.
Methodology For Analytical Toxicology. CRC Press,
Sunshine I 1985; Volume I: 448-450

 

23

24

25

26

27

28

29

37

Andresen BD, Wise BL, Ng KJ, Wise GA, Mass Spectrometry
In Clinical Chemistry. Am. J. Med. Technol. 1985;
Volume 2, Number 4: 245—250

Olson CL, Theory And Instrumentation. GLC and HPLC
Determination Of Therapeutic Agents. Marcel Dekker,
Inc., Tsuji K, Morozowich W 1978; Volume 9, Part 1: 1-46

Holland JF, Martin FE, Sweeley CC, Computer Interfacing
And Data Processing. GLC and HPLC Determination Of
Therapeutic Agents. Marcel Dekker, Inc., Tsuji K,
Morozowich W 1978; Volume 9, Part 1: 385—415

Holland JF, Leary JJ, Sweeley CC, Instrumentation
Strategies For Metabolic Profiling. An Article Prepared
For Publication At Michigan State University, Department
Of Biochemistry 1986

Schutte, Geritt, Toxicology Laboratory Center, Lansing,
Mi.; Personnal Communication

Monograph, Drug Screening Analysis Using New Sample
Injection Tee And Two Wide Bore Capillary Columns.
The Supelco Reporter 1986; Volume V, Number 2, May: 4—5

Monograph, Wide Bore Capillary Column Installation
Instructions. Supelco, Number 9684: 1-7