Jacob Bell
New Member
Excretion patterns of cannabinoid
metabolites after last use in a group of
chronic users
The urinary excretion patterns of 86 chronic cannabis users were examined after their last cannabis use
by two common screening methods, the semiquantitative EMIT-d.a.u. and the qualitative EMIT-st (Syva
Company). We demonstrated that under very strictly supervised abstinence, chronic users can have positive
results for cannabinoids in urine at 20 ng/ml or above on the EMIT-d.a.u. assay for as many as 46
consecutive days from admission, and can take as many as 77 days to drop below the cutoff calibrator for
10 consecutive days. For all subjects, the mean excretion time was 27 days. Subject excretion patterns were
clearly biphasic, with initial higher rates of excretion not sustained. During the subsequent period of
leveling off, most subjects had one or more separate sequences of cannabinoid-negative urine test results,
lasting a mean of 3 days each and followed by at least one positive result. Demographic, body type, and
drug history variables proved to be only moderate predictors of excretion patterns. Findings were discussed
in the context of potential clinical and forensic application. (CLIN PHARMACOL THE 1985;38:572-8.)
George M. Ellis, Jr., B.A., Marian A. Mann, B.A., Barbara A. Judson, M.S.,
N. Ted Schramm, M.A., and Agop Tashchian, CPT MC, USNR San Diego, Calif.
Although it is generally agreed that delta-9-tetrahydrocannabinol
(G19-THC) is the major psychoactive
component of the cannabis (marijuana) plant, there are
other cannabinoid constituents, as well as metabolites,
that have psychoactive effects. One area of scientific
interest is the lengthy retention time of the major metabolites,
which seem to persist in the system for up to
several weeks after repeated exposure. Very few studies
have been made of the actual parameters and profiles
of cannabis excretion. Lemberger et al.1'2 reported that
after intravenous injection, cannabis metabolites are excreted
in the urine and feces for more than a week in
both nonusers and chronic users. Hollister and Kanter'
reported that a metabolite was found in the urine of
nonusers for 72 hours after a single exposure. In the
most extensive study to date, Dackis et al.4 used the
EMIT-d.a.u. (Syva Company) methodology in a locked
ward setting and reported on seven heavy users who
had positive urinalysis results from 14 to 36 days after
their last known cannabis use.
We examined the urinary excretion patterns of
From the Treatment Department, Naval Drug Rehabilitation Center.
Received for publication March 14, 1985; accepted July 25, 1985.
Reprint requests to: George M. Ellis, Jr., Director of Drug Abuse
Services, Behavior Research Incorporated, 3456 Camino del Rio
North, Suite 101, San Diego, CA 92108.
chronic (regular) cannabis abusers after their last use.
This investigation will provide data on the number of
days that urine samples of chronic users produce cannabinoid-
positive results and examine the various data
elements that help establish the subjects' urine excretion
patterns.
METHODS
Subjects. Our subjects were 86 male and female enlisted
service personnel undergoing residential treatment
for drug and alcohol dependence at the Naval Drug
Rehabilitation Center in San Diego. Included in the
study were arriving personnel with cannabinoid-positive
urine samples for a minimum of the first 3 days
after admission, drawn from a sample of 274 entering
personnel that were tested over a 5-month period. Subjects
were predominantly male (97.7%) and Caucasian
(72.1%) and ranged in age from 18 to 34 years
(X = 22.8 years). They had a history of marijuana use
from 2 to 21 years (X = 8.9 years) and most (60.5%)
used the drug once a day or more often. Over half
(55.8%) were abusers of more than one drug, and almost
one third (30.2%) were considered heavy users
of one or more drugs besides marijuana.
Procedure. Subjects provided first morning voids under
direct observation and precise chain-of-custody collection
procedures. All urine samples were tested by
both the semiquantitative EMIT-d.a.u. and the qualitative
EMIT-st (Syva Company) cannabinoid assays.
Specific gravities, void order, and any prescribed medications
were noted. Urine was collected daily until
subjects had 10 consecutive cannabinoid-negative tests
as determined by EMIT-d.a.u. Within the first week
after arrival, data were collected in personal interviews
for subject age, sex, body type (lean, average, obese),
weight, and a height-weight integrated index (adapted
from U.S. Navy combined standards). Self-reported
data were also obtained on length of time since last use
of cannabis before treatment entry; route of cannabis
administration; amount administered during each use;
history of cannabis use for lifetime, 6 months, 1 month,
and 1 week before treatment entry; and history of alcohol
and other drug use.
While it is impossible to state categorically that no
marijuana could have been used by subjects while in
the Center, the closely supervised conditions made this
extremely unlikely. Subjects were never allowed to
leave the facility for any reason without staff escort
(total Center staff size was 150). Subjects were monitored
by staff members 24 hours a day, 7 days a week,
including every 15 minutes throughout the night. Subjects
were singled out for special staff attention and
urine results were carefully scrutinized each testing day
and matched with staff clinical observations to reaffirm
subject abstinence.
EMIT methodologies and equipment. According to
the product literature, EMIT homogeneous enzyme immunoassays
for cannabinoids are designed to detect a
major urinary metabolite of A9-THC, 11-nor-A9-THC-
9-carboxcylic acid, or a combination of one or more
analyte equivalents such as 11-hydroxy-A9-THC, 8,13-
hydroxy-A9-THC, or 8,i3-11-hydroxy-A9-THC. The
semiquantitative EMIT-d.a.u. assay system estimates
metabolite concentration with a low or cutoff calibration
of 20 ng/ml, a 95% detection level of 50 ng/ml, a
medium calibrator of 75 ng/ml, and a detection range
to 82.5 ng/ml (all results estimated to be -?---82.5 ng/ml
are reported as 82.5 ng/ml). The amount of cannabinoids
read from the calibration curve is approximate
and expressed as analyte equivalents. The qualitative
EMIT-st drug detection system identifies metabolite
presence with a low or cutoff calibration of 100 ng/ml
and a 95% detection level of 200 ng/ml.
Definitions. The cannabinoid excretion pattern was
broken into five basic components: time to first negative
result, duration, extended duration, number of negative
"gaps," and length of the longest negative "gap" (Fig.
1). The time to the first negative result was defined as
the number of cannabinoid-positive days after admission
until a subject first produced a cannabinoid-negative
urine sample. Duration was the number of days
after admission, negative or positive, until the subject
provided the last positive urine sample. Extended duration
was the number of days in the subject's duration
added to the number of days of claimed abstinence
before admission to treatment. Number of negative gaps
was defined as the number of times a pattern of cannabinoid-
positive urine tests turned negative then returned
positive. Length of the longest negative gap
was the greatest number of days such a negative
sequence lasted.
Data analyses. The excretion pattern for each subject
was graphed for each method. Excretion outcome data
were analyzed for the whole sample and for a by-group
categorization based on self-reported light, moderate,
or heavy cannabis use over the last month and last 6
months before admission. Wilcoxon two-sample rank
tests were done to determine differences between the
groups on urine test, demographic, and history van-ables. Pearson product-moment correlations were calculated
to examine statistical relationships between
variables. Because of the complexity of the study components,
a stepwise regression procedure was deter-mined to be best suited for an exploratory multivariate
analysis.
RESULTS
Excretion patterns. Examination of the excretion
patterns for the 86 study subjects by EMIT-d.a.u.
showed substantial variations in form, slope, and number
of days with cannabinoid-positive results. Representative
subject excretion patterns are shown in Fig.
2. Patterns showed generally declining values after admission
to treatment, with day-to-day fluctuations in
estimated concentrations evident in most cases. Nonetheless,
there were no obvious large peaks in the excretion
patterns that suggested continued use during the
Center stay. Inspection of the excretion graphs revealed
a biphasic pattern. After a fairly steep drop in excretion
levels, metabolite concentrations often fluctuated
around the low or cutoff calibration level (20 ng/ml),
causing prominent negative gaps in the graphs of both
the EMIT-d.a.u. and EMIT-st results. There seemed to
be no relationship between special events (i.e., extra
physical exercise), medications taken, or weight loss
and major shifts in excreted cannabinoid concentration.
Data were obtained primarily (91.2%) from the first
void of the morning. Of the five subjects who provided
two urine samples in the same morning, one showed
an increase in estimated concentration between successive
voids; the others showed estimated decreases of
from 5 to 42 ng/ml as determined by EMIT-d.a.u.
Time to first negative result, duration, and extended
duration. Table I shows descriptive statistics for the
excretion pattern outcome variables for all subjects.
For both testing methods, and especially for the more
sensitive EMIT-d.a.u. procedures, subjects showed
broader ranges and greater means than previously described
in anecdotal reports or in the literature. When
subjects were divided into light, moderate, and heavy
use categories (Table I), the expected directional increases
in these outcome variables were observed in
most cases. However, there was substantial overlap in
the ranges observed. Wilcoxon two-sample rank tests
between groups showed significant differences for both
methods in time to first negative result and in duration
for only the light/heavy and moderate/heavy user comparisons.
Extended duration was significantly different
only for the light/heavy user comparison determined
by EMIT-d . a . u .
Table II shows the frequency of duration of cannabinoid
excretion patterns for both testing methods.
When one considers the time it took to maintain negative
results for 10 consecutive days, almost one third
(32.6%) of the subjects (most of them heavy users)
remained positive by the EMIT-d.a.u. at 20 ng/ml for
more than a month, with four subjects remaining positive
for >2 months.
Number of negative gaps and length of longest negative
gap. For both testing methods, there was a similarly
wide range in the number of negative gaps (Table
I). Negative gaps were noted in 87.2% of our subjects
by EMIT-d.a.u. testing (75 of 86 cases) and in 47.7%
of our subjects by EMIT-st testing (41 of 86 cases).
Only a few of the demographic and history variables
correlated with the number of negative gaps.
There was a wide range in the longest negative gap
observed (Table I), with the Emit-st range of particular
note. The length of the longest negative gap extended
>3 days in 39.5% of our subjects as determined by
EMIT-d.a.u. testing (34 of 86 cases) and in 24.4% of
our subjects as determined by EMIT-st testing (21 of
86 cases). Of particular interest is that six subjects
(7.0%) had a longest negative gap as determined by
EMIT-st >10 days. It should be remembered that the
length of the longest gap in EMIT-d.a.u. testing was
limited by operational definition to only 10 days, the
criterion for termination from the study.
Wilcoxon two-sample rank tests showed little con-sistency in significant differences between groups in
either the number of negative gaps or the length of the
longest negative gap.
As expected, excretion variables correlated most
highly with the other excretion variables (with r values
ranging up to 0.89). Many of the history and demographic
variables were also significantly related to the
excretion variables, but the correlation coefficients were
considerably smaller (with r values ranging up to 0.44)
than those observed between the excretion variables
themselves.
When all variables were placed in the stepwise
regression procedure with time to first negative result,
duration, and extended duration, the excretion outcome
variables proved to the greatest predictors, accounting
for between 72.1% and 100.0% of the variability. When
the excretion outcome variables were removed, the 12
demographic and history variables selected for evaluation
were only modest contributors to a predictive
model, either singly or in complement (Table III). For
the EMIT-d. a.u., although the overall hypothetic models
were significant (P < 0.001), the percent of variability
accounted for by variables in the time to first
negative result, duration, and extended duration regressions
were relatively slight (35.7%, 42.7%, and 48.3%,
respectively). Similar results were seen in the EMITst
analyses, with the variables in the time to first negative
result, duration, and extended duration regressions
accounting for 41.4%, 31.8%, and 59.2% of the variability,
respectively, despite overall significant models
(P < 0.001). Three of the demographic and history
variables (age, weight, and having a history of cannabis
use more than once a day) seemed to be the most prominent,
but their overall predictive value could only be
judged as moderate. DISCUSSION
Despite considerable clinical and forensic interest,
there have been surprisingly few published data on the
length of cannabinoid excretion in the urine of chronic
users after last use. Important data were provided by
Dackis et al. ,4 whose seven subjects had a mean of 24
days with cannabinoid-positive results (with an upward
range of 36 days) before a first negative result was
observed. A published case study' gave data on a
chronic smoker who provided cannabinoid-positive
urine samples for 15 consecutive days after reportedly
ceasing use, and who was still positive after 20 days
of testing. (Both of these studies used EMIT-d.a.u.
testing with a 20 ng/ml cutoff.) Our results are in general
agreement with those data and substantially extend
the published findings. Urine samples from our 86
chronic users were still cannabinoid positive up to 46
days (R = 16 days) as determined by EMIT-d.a.u.
before the first negative result, and took up to 77 days
= 27 days) to drop below the 20 ng/ml cutoff for
10 consecutive days (Table I). Both the extended range
of positive days and the variations in the excretion outcome
variables for both EMIT-d.a.u. and EMIT-st were
unexpected and not previously reported.
Our study validates the reliability and usefulness of
the EMIT-d.a.u. method for determining urinary excretion
patterns after chronic cannabinoid use. Our results
also provide new information that should be taken
into account in the interpretation of urine test results
for cannabinoids, regardless of the method used. Current
applications of urinalysis, such as in surveillance
of identified users, should be reevaluated in light of our
data.
Our data generally confirm previous research that
chronic users can take, on the average, about a month
to achieve consistently negative urine samples as
determined by a sensitive urine assay such as the
EMIT-d.a.u. that uses a 20 ng/ml cutoff calibration
(Table I).
In terms of the duration of positive results, well over
half (60.6%) of our chronic users had cannabinoidpositive
results for %21 days after last use, almost one
third (32.7%) had positive results for >30 days, and
4.7% had positive results for >60 days before dropping
below the EMIT-d.a.u. 20 ng/ml cutoff for 10 consecutive
days (Table II). There was a natural break observed
in the distribution at 50 days, which based on
these data seems to be a reasonable upward limit for
the evaluation of most chronic users (Table II). The five
subjects who had positive results for >50 days had been
heavy users for at least 12 years; however, there were
other subjects with similar cannabis histories who had
cannabinoid-positive results for <50 days.
Of those chronic users who had a cannabinoid-positive
result on the EMIT-st, only a few (15.1%) had
positive results on the EMIT-st for 21 days, while a
majority (64.4%) had positive results for 10 days
before returning to a consistently negative pattern
(Table II).
Although there were statistically significant differences
in the means of many of the excretion variables
between subjects grouped by self-reported light, moderate,
and heavy use, the amount of overlap in the
ranges was noteworthy. While it is not clear how individual
data points would be affected by urine samples
collected randomly during the day, the outside ranges
could aid in the interpretation of test results. Thus if a
subject claimed to be a light user, one would expect to
see a first negative test result on the EMIT-d.a.u. within
approximately 18 days after ceasing use and the daily
test results to start being consistently negative no more
than a month after last use.
Although many were statistically significant, few of
the demographic and history variables showed substantial
correlative relationships with the excretion pattern
components. Based on the stepwise regression analysis,
demographic and history variables showed little usefulness
as predictors of excretion patterns (Table III).
There was a lack of real predictive prominence of body
type, weight, and height-weight index, despite suggestions
of previous research in animals.' Weight alone
seemed to show the most promise. Age was evident in
almost every regression; however, its overall importance
appeared relatively small. Also demonstrating little
impact were the various cannabis history variables.
Although expected, the identification of a clear biphasic
pattern in urine has not been previously reported
in the literature on cannabinoid excretion. Also of interest
is the fluctuating effect in the day-to-day EMITEMIT-
d.a.u. determinations, especially around the cutoff level
that created the strings of negative gaps. This phenomenon
is not too surprising for a long-persisting analyte,
but might tend to confuse the proper interpretation of
results for those not familiar with this phenomenon.
This can be partly attributed to the semiquantitative
nature of the assay and the differing reactivity with the
metabolites, to differing amounts of the metabolites
being released in the body, and to increased test variability
when measuring values at the extreme end of
the calibration curve. Once a subject had a urine test
result that went below the 20 ng/ml cutoff, no subsequent
test ever went above the medium calibrator (75
ng/ml) and only five subjects (6.7% of the subjects with
negative EMIT-d.a.u. gaps) had positive urine samples
>50 ng/ml.
Estimates of time since last use derived from a single
urine test result or even a spaced series of results would
seem to have limited forensic value. Our study reaffirms
that a positive urine test result from a chronic user does
not necessarily indicate recent or continued use. Because
of large individual variability, a positive test result
may reflect use within the past few hours, days,
weeks, or even months, depending on the cutoff calibration
level selected. Because of the special sensitivity
demonstrated by the EMIT assay, which cross-reacts
with a broad metabolite mix rather than a single prominent
metabolite such as the 9-carboxy-THC (COOHTHC),
it is clear from previous comparative research
that the duration of positive results would also be affected
when an alternate method is used.7'8 Additionally,
some urine specimens have been found to have
lower concentrations of free COOH-THC than expected
based on unpredictable rates of metabolization to its
glucuronide ester.'" Even EMIT results must be interpreted
with some small amount of caution because
assay response may vary slightly with the relative ratio
of free COOH-THC and this unstable glucuronide conjugate.*
Finally, the impact of the differences of drug
potency (plant genetics), the potential interactive effect
of previous alcohol and other drug abuse, and the differences
in bioavailability of 6,9-THC in the route and
pattern of administration due to the experience of the
user would seem important, but such studies were beyond
the scope of our investigation.
*Willette RE: Personal communication, 1985.
Donald Avoy, M.D., and David Himmelberger, M.S., of
Syva Company, Palo Alto, Calif., provided operational sup-port and technical assistance in the project. Avram Goldstein,
M.D., Addiction Research Foundation, Palo Alto, Calif.;
Robert Willette, Ph.D., Research Design, Annapolis, Md.;
and John Irving, CDR MSC, USN, Naval Medical Command,
Washington, D. C., provided helpful guidance in the draft
review. Lyn Chilcote of the Naval Drug Rehabilitation Center
provided technical review of project materials. Hospital
Corpsmen Phillip Mucha, Rande Stanvitch, and F. Lincoln
Merritt of the Naval Drug Rehabilitation Center conducted
most of the urinalyses.
References
Lemberger L, Silberstein SD, Axelrod J, Kopin IJ. Marijuana:
Studies on the disposition and metabolism of
delta-9-tetrahydrocannabinol in man. Science
1970;170:1320-2.
Lemberger L, Tamarkin NR, Axelrod J, Kopin IJ. Delta-
9-tetrahydrocannabinol: Metabolism and disposition in
long-term marihuana smokers. Science 1971;173:72-4.
Hollister LE, Kanter SL. Laboratory verification of
"heavy" and "light" users of cannabis. Drug Alcohol
Depend 1980;5:151-2.
Dackis CA, Pottash ALC, Annitto W, Gold MS. Persistence
of urinary marijuana levels after supervised abstinence.
Am J Psychiatry 1982;139:1196-8.
Clinical summary addendum, EMIT-d.a.u. and EMITSt
urine cannabinoid assays. Palo Alto, Calif: Syva Company,
1982:40-4.
Garrett ER, Hunt CA. Pharmacokinetics of A9-tetrahydrocannabinol
in dogs. J Pharm Sci 1977;66:395-407.
Jones AB, ElSholy HW, Arafat ES, ElSohly MA. Analysis
of the major metabolite of A9-tetrahydrocannabinol
in urine. IV. A comparison of five methods. J Anal Toxicol
1984;8:249-51.
Irving J, Leeb B, Foltz RL, Cook CE, Bursey JT, Willette
RE. Evaluation of immunoassays for cannabinoids in
urine. J Anal Toxicol 1984;8:192-6.
Peat MA, Finkle BS, Deyman ME. Laboratory evaluation
of immunoassay kits for the detection of cannabinoids
in biological fluids. In: Hawks RL, ed. The analysis
of cannabinoids in biological fluids. Rockville, Md: National
Institute on Drug Abuse 1982:85-98; NIDA Research
Monograph 42.
Foltz RL. Analysis of cannabinoids in physiological specimens
by gas chromatography/mass spectrometry. In:
Baselt RC, ed. Advances in analytical toxicology, vol I.
Foster City, Calif: Biomedical Publications, 1984:125-
57.
Williams PL, Moffat AC. Identification in human urine
of A9-tetrahydrocannabino1-11-oic acid glucuronide: A
tetrahydrocannabinol metabolite. I Pharm Pharmacol
1980;32:445-8.
Source: Excretion patterns of cannabinoid
metabolites after last use in a group of chronic users
metabolites after last use in a group of
chronic users
The urinary excretion patterns of 86 chronic cannabis users were examined after their last cannabis use
by two common screening methods, the semiquantitative EMIT-d.a.u. and the qualitative EMIT-st (Syva
Company). We demonstrated that under very strictly supervised abstinence, chronic users can have positive
results for cannabinoids in urine at 20 ng/ml or above on the EMIT-d.a.u. assay for as many as 46
consecutive days from admission, and can take as many as 77 days to drop below the cutoff calibrator for
10 consecutive days. For all subjects, the mean excretion time was 27 days. Subject excretion patterns were
clearly biphasic, with initial higher rates of excretion not sustained. During the subsequent period of
leveling off, most subjects had one or more separate sequences of cannabinoid-negative urine test results,
lasting a mean of 3 days each and followed by at least one positive result. Demographic, body type, and
drug history variables proved to be only moderate predictors of excretion patterns. Findings were discussed
in the context of potential clinical and forensic application. (CLIN PHARMACOL THE 1985;38:572-8.)
George M. Ellis, Jr., B.A., Marian A. Mann, B.A., Barbara A. Judson, M.S.,
N. Ted Schramm, M.A., and Agop Tashchian, CPT MC, USNR San Diego, Calif.
Although it is generally agreed that delta-9-tetrahydrocannabinol
(G19-THC) is the major psychoactive
component of the cannabis (marijuana) plant, there are
other cannabinoid constituents, as well as metabolites,
that have psychoactive effects. One area of scientific
interest is the lengthy retention time of the major metabolites,
which seem to persist in the system for up to
several weeks after repeated exposure. Very few studies
have been made of the actual parameters and profiles
of cannabis excretion. Lemberger et al.1'2 reported that
after intravenous injection, cannabis metabolites are excreted
in the urine and feces for more than a week in
both nonusers and chronic users. Hollister and Kanter'
reported that a metabolite was found in the urine of
nonusers for 72 hours after a single exposure. In the
most extensive study to date, Dackis et al.4 used the
EMIT-d.a.u. (Syva Company) methodology in a locked
ward setting and reported on seven heavy users who
had positive urinalysis results from 14 to 36 days after
their last known cannabis use.
We examined the urinary excretion patterns of
From the Treatment Department, Naval Drug Rehabilitation Center.
Received for publication March 14, 1985; accepted July 25, 1985.
Reprint requests to: George M. Ellis, Jr., Director of Drug Abuse
Services, Behavior Research Incorporated, 3456 Camino del Rio
North, Suite 101, San Diego, CA 92108.
chronic (regular) cannabis abusers after their last use.
This investigation will provide data on the number of
days that urine samples of chronic users produce cannabinoid-
positive results and examine the various data
elements that help establish the subjects' urine excretion
patterns.
METHODS
Subjects. Our subjects were 86 male and female enlisted
service personnel undergoing residential treatment
for drug and alcohol dependence at the Naval Drug
Rehabilitation Center in San Diego. Included in the
study were arriving personnel with cannabinoid-positive
urine samples for a minimum of the first 3 days
after admission, drawn from a sample of 274 entering
personnel that were tested over a 5-month period. Subjects
were predominantly male (97.7%) and Caucasian
(72.1%) and ranged in age from 18 to 34 years
(X = 22.8 years). They had a history of marijuana use
from 2 to 21 years (X = 8.9 years) and most (60.5%)
used the drug once a day or more often. Over half
(55.8%) were abusers of more than one drug, and almost
one third (30.2%) were considered heavy users
of one or more drugs besides marijuana.
Procedure. Subjects provided first morning voids under
direct observation and precise chain-of-custody collection
procedures. All urine samples were tested by
both the semiquantitative EMIT-d.a.u. and the qualitative
EMIT-st (Syva Company) cannabinoid assays.
Specific gravities, void order, and any prescribed medications
were noted. Urine was collected daily until
subjects had 10 consecutive cannabinoid-negative tests
as determined by EMIT-d.a.u. Within the first week
after arrival, data were collected in personal interviews
for subject age, sex, body type (lean, average, obese),
weight, and a height-weight integrated index (adapted
from U.S. Navy combined standards). Self-reported
data were also obtained on length of time since last use
of cannabis before treatment entry; route of cannabis
administration; amount administered during each use;
history of cannabis use for lifetime, 6 months, 1 month,
and 1 week before treatment entry; and history of alcohol
and other drug use.
While it is impossible to state categorically that no
marijuana could have been used by subjects while in
the Center, the closely supervised conditions made this
extremely unlikely. Subjects were never allowed to
leave the facility for any reason without staff escort
(total Center staff size was 150). Subjects were monitored
by staff members 24 hours a day, 7 days a week,
including every 15 minutes throughout the night. Subjects
were singled out for special staff attention and
urine results were carefully scrutinized each testing day
and matched with staff clinical observations to reaffirm
subject abstinence.
EMIT methodologies and equipment. According to
the product literature, EMIT homogeneous enzyme immunoassays
for cannabinoids are designed to detect a
major urinary metabolite of A9-THC, 11-nor-A9-THC-
9-carboxcylic acid, or a combination of one or more
analyte equivalents such as 11-hydroxy-A9-THC, 8,13-
hydroxy-A9-THC, or 8,i3-11-hydroxy-A9-THC. The
semiquantitative EMIT-d.a.u. assay system estimates
metabolite concentration with a low or cutoff calibration
of 20 ng/ml, a 95% detection level of 50 ng/ml, a
medium calibrator of 75 ng/ml, and a detection range
to 82.5 ng/ml (all results estimated to be -?---82.5 ng/ml
are reported as 82.5 ng/ml). The amount of cannabinoids
read from the calibration curve is approximate
and expressed as analyte equivalents. The qualitative
EMIT-st drug detection system identifies metabolite
presence with a low or cutoff calibration of 100 ng/ml
and a 95% detection level of 200 ng/ml.
Definitions. The cannabinoid excretion pattern was
broken into five basic components: time to first negative
result, duration, extended duration, number of negative
"gaps," and length of the longest negative "gap" (Fig.
1). The time to the first negative result was defined as
the number of cannabinoid-positive days after admission
until a subject first produced a cannabinoid-negative
urine sample. Duration was the number of days
after admission, negative or positive, until the subject
provided the last positive urine sample. Extended duration
was the number of days in the subject's duration
added to the number of days of claimed abstinence
before admission to treatment. Number of negative gaps
was defined as the number of times a pattern of cannabinoid-
positive urine tests turned negative then returned
positive. Length of the longest negative gap
was the greatest number of days such a negative
sequence lasted.
Data analyses. The excretion pattern for each subject
was graphed for each method. Excretion outcome data
were analyzed for the whole sample and for a by-group
categorization based on self-reported light, moderate,
or heavy cannabis use over the last month and last 6
months before admission. Wilcoxon two-sample rank
tests were done to determine differences between the
groups on urine test, demographic, and history van-ables. Pearson product-moment correlations were calculated
to examine statistical relationships between
variables. Because of the complexity of the study components,
a stepwise regression procedure was deter-mined to be best suited for an exploratory multivariate
analysis.
RESULTS
Excretion patterns. Examination of the excretion
patterns for the 86 study subjects by EMIT-d.a.u.
showed substantial variations in form, slope, and number
of days with cannabinoid-positive results. Representative
subject excretion patterns are shown in Fig.
2. Patterns showed generally declining values after admission
to treatment, with day-to-day fluctuations in
estimated concentrations evident in most cases. Nonetheless,
there were no obvious large peaks in the excretion
patterns that suggested continued use during the
Center stay. Inspection of the excretion graphs revealed
a biphasic pattern. After a fairly steep drop in excretion
levels, metabolite concentrations often fluctuated
around the low or cutoff calibration level (20 ng/ml),
causing prominent negative gaps in the graphs of both
the EMIT-d.a.u. and EMIT-st results. There seemed to
be no relationship between special events (i.e., extra
physical exercise), medications taken, or weight loss
and major shifts in excreted cannabinoid concentration.
Data were obtained primarily (91.2%) from the first
void of the morning. Of the five subjects who provided
two urine samples in the same morning, one showed
an increase in estimated concentration between successive
voids; the others showed estimated decreases of
from 5 to 42 ng/ml as determined by EMIT-d.a.u.
Time to first negative result, duration, and extended
duration. Table I shows descriptive statistics for the
excretion pattern outcome variables for all subjects.
For both testing methods, and especially for the more
sensitive EMIT-d.a.u. procedures, subjects showed
broader ranges and greater means than previously described
in anecdotal reports or in the literature. When
subjects were divided into light, moderate, and heavy
use categories (Table I), the expected directional increases
in these outcome variables were observed in
most cases. However, there was substantial overlap in
the ranges observed. Wilcoxon two-sample rank tests
between groups showed significant differences for both
methods in time to first negative result and in duration
for only the light/heavy and moderate/heavy user comparisons.
Extended duration was significantly different
only for the light/heavy user comparison determined
by EMIT-d . a . u .
Table II shows the frequency of duration of cannabinoid
excretion patterns for both testing methods.
When one considers the time it took to maintain negative
results for 10 consecutive days, almost one third
(32.6%) of the subjects (most of them heavy users)
remained positive by the EMIT-d.a.u. at 20 ng/ml for
more than a month, with four subjects remaining positive
for >2 months.
Number of negative gaps and length of longest negative
gap. For both testing methods, there was a similarly
wide range in the number of negative gaps (Table
I). Negative gaps were noted in 87.2% of our subjects
by EMIT-d.a.u. testing (75 of 86 cases) and in 47.7%
of our subjects by EMIT-st testing (41 of 86 cases).
Only a few of the demographic and history variables
correlated with the number of negative gaps.
There was a wide range in the longest negative gap
observed (Table I), with the Emit-st range of particular
note. The length of the longest negative gap extended
>3 days in 39.5% of our subjects as determined by
EMIT-d.a.u. testing (34 of 86 cases) and in 24.4% of
our subjects as determined by EMIT-st testing (21 of
86 cases). Of particular interest is that six subjects
(7.0%) had a longest negative gap as determined by
EMIT-st >10 days. It should be remembered that the
length of the longest gap in EMIT-d.a.u. testing was
limited by operational definition to only 10 days, the
criterion for termination from the study.
Wilcoxon two-sample rank tests showed little con-sistency in significant differences between groups in
either the number of negative gaps or the length of the
longest negative gap.
As expected, excretion variables correlated most
highly with the other excretion variables (with r values
ranging up to 0.89). Many of the history and demographic
variables were also significantly related to the
excretion variables, but the correlation coefficients were
considerably smaller (with r values ranging up to 0.44)
than those observed between the excretion variables
themselves.
When all variables were placed in the stepwise
regression procedure with time to first negative result,
duration, and extended duration, the excretion outcome
variables proved to the greatest predictors, accounting
for between 72.1% and 100.0% of the variability. When
the excretion outcome variables were removed, the 12
demographic and history variables selected for evaluation
were only modest contributors to a predictive
model, either singly or in complement (Table III). For
the EMIT-d. a.u., although the overall hypothetic models
were significant (P < 0.001), the percent of variability
accounted for by variables in the time to first
negative result, duration, and extended duration regressions
were relatively slight (35.7%, 42.7%, and 48.3%,
respectively). Similar results were seen in the EMITst
analyses, with the variables in the time to first negative
result, duration, and extended duration regressions
accounting for 41.4%, 31.8%, and 59.2% of the variability,
respectively, despite overall significant models
(P < 0.001). Three of the demographic and history
variables (age, weight, and having a history of cannabis
use more than once a day) seemed to be the most prominent,
but their overall predictive value could only be
judged as moderate. DISCUSSION
Despite considerable clinical and forensic interest,
there have been surprisingly few published data on the
length of cannabinoid excretion in the urine of chronic
users after last use. Important data were provided by
Dackis et al. ,4 whose seven subjects had a mean of 24
days with cannabinoid-positive results (with an upward
range of 36 days) before a first negative result was
observed. A published case study' gave data on a
chronic smoker who provided cannabinoid-positive
urine samples for 15 consecutive days after reportedly
ceasing use, and who was still positive after 20 days
of testing. (Both of these studies used EMIT-d.a.u.
testing with a 20 ng/ml cutoff.) Our results are in general
agreement with those data and substantially extend
the published findings. Urine samples from our 86
chronic users were still cannabinoid positive up to 46
days (R = 16 days) as determined by EMIT-d.a.u.
before the first negative result, and took up to 77 days
= 27 days) to drop below the 20 ng/ml cutoff for
10 consecutive days (Table I). Both the extended range
of positive days and the variations in the excretion outcome
variables for both EMIT-d.a.u. and EMIT-st were
unexpected and not previously reported.
Our study validates the reliability and usefulness of
the EMIT-d.a.u. method for determining urinary excretion
patterns after chronic cannabinoid use. Our results
also provide new information that should be taken
into account in the interpretation of urine test results
for cannabinoids, regardless of the method used. Current
applications of urinalysis, such as in surveillance
of identified users, should be reevaluated in light of our
data.
Our data generally confirm previous research that
chronic users can take, on the average, about a month
to achieve consistently negative urine samples as
determined by a sensitive urine assay such as the
EMIT-d.a.u. that uses a 20 ng/ml cutoff calibration
(Table I).
In terms of the duration of positive results, well over
half (60.6%) of our chronic users had cannabinoidpositive
results for %21 days after last use, almost one
third (32.7%) had positive results for >30 days, and
4.7% had positive results for >60 days before dropping
below the EMIT-d.a.u. 20 ng/ml cutoff for 10 consecutive
days (Table II). There was a natural break observed
in the distribution at 50 days, which based on
these data seems to be a reasonable upward limit for
the evaluation of most chronic users (Table II). The five
subjects who had positive results for >50 days had been
heavy users for at least 12 years; however, there were
other subjects with similar cannabis histories who had
cannabinoid-positive results for <50 days.
Of those chronic users who had a cannabinoid-positive
result on the EMIT-st, only a few (15.1%) had
positive results on the EMIT-st for 21 days, while a
majority (64.4%) had positive results for 10 days
before returning to a consistently negative pattern
(Table II).
Although there were statistically significant differences
in the means of many of the excretion variables
between subjects grouped by self-reported light, moderate,
and heavy use, the amount of overlap in the
ranges was noteworthy. While it is not clear how individual
data points would be affected by urine samples
collected randomly during the day, the outside ranges
could aid in the interpretation of test results. Thus if a
subject claimed to be a light user, one would expect to
see a first negative test result on the EMIT-d.a.u. within
approximately 18 days after ceasing use and the daily
test results to start being consistently negative no more
than a month after last use.
Although many were statistically significant, few of
the demographic and history variables showed substantial
correlative relationships with the excretion pattern
components. Based on the stepwise regression analysis,
demographic and history variables showed little usefulness
as predictors of excretion patterns (Table III).
There was a lack of real predictive prominence of body
type, weight, and height-weight index, despite suggestions
of previous research in animals.' Weight alone
seemed to show the most promise. Age was evident in
almost every regression; however, its overall importance
appeared relatively small. Also demonstrating little
impact were the various cannabis history variables.
Although expected, the identification of a clear biphasic
pattern in urine has not been previously reported
in the literature on cannabinoid excretion. Also of interest
is the fluctuating effect in the day-to-day EMITEMIT-
d.a.u. determinations, especially around the cutoff level
that created the strings of negative gaps. This phenomenon
is not too surprising for a long-persisting analyte,
but might tend to confuse the proper interpretation of
results for those not familiar with this phenomenon.
This can be partly attributed to the semiquantitative
nature of the assay and the differing reactivity with the
metabolites, to differing amounts of the metabolites
being released in the body, and to increased test variability
when measuring values at the extreme end of
the calibration curve. Once a subject had a urine test
result that went below the 20 ng/ml cutoff, no subsequent
test ever went above the medium calibrator (75
ng/ml) and only five subjects (6.7% of the subjects with
negative EMIT-d.a.u. gaps) had positive urine samples
>50 ng/ml.
Estimates of time since last use derived from a single
urine test result or even a spaced series of results would
seem to have limited forensic value. Our study reaffirms
that a positive urine test result from a chronic user does
not necessarily indicate recent or continued use. Because
of large individual variability, a positive test result
may reflect use within the past few hours, days,
weeks, or even months, depending on the cutoff calibration
level selected. Because of the special sensitivity
demonstrated by the EMIT assay, which cross-reacts
with a broad metabolite mix rather than a single prominent
metabolite such as the 9-carboxy-THC (COOHTHC),
it is clear from previous comparative research
that the duration of positive results would also be affected
when an alternate method is used.7'8 Additionally,
some urine specimens have been found to have
lower concentrations of free COOH-THC than expected
based on unpredictable rates of metabolization to its
glucuronide ester.'" Even EMIT results must be interpreted
with some small amount of caution because
assay response may vary slightly with the relative ratio
of free COOH-THC and this unstable glucuronide conjugate.*
Finally, the impact of the differences of drug
potency (plant genetics), the potential interactive effect
of previous alcohol and other drug abuse, and the differences
in bioavailability of 6,9-THC in the route and
pattern of administration due to the experience of the
user would seem important, but such studies were beyond
the scope of our investigation.
*Willette RE: Personal communication, 1985.
Donald Avoy, M.D., and David Himmelberger, M.S., of
Syva Company, Palo Alto, Calif., provided operational sup-port and technical assistance in the project. Avram Goldstein,
M.D., Addiction Research Foundation, Palo Alto, Calif.;
Robert Willette, Ph.D., Research Design, Annapolis, Md.;
and John Irving, CDR MSC, USN, Naval Medical Command,
Washington, D. C., provided helpful guidance in the draft
review. Lyn Chilcote of the Naval Drug Rehabilitation Center
provided technical review of project materials. Hospital
Corpsmen Phillip Mucha, Rande Stanvitch, and F. Lincoln
Merritt of the Naval Drug Rehabilitation Center conducted
most of the urinalyses.
References
Lemberger L, Silberstein SD, Axelrod J, Kopin IJ. Marijuana:
Studies on the disposition and metabolism of
delta-9-tetrahydrocannabinol in man. Science
1970;170:1320-2.
Lemberger L, Tamarkin NR, Axelrod J, Kopin IJ. Delta-
9-tetrahydrocannabinol: Metabolism and disposition in
long-term marihuana smokers. Science 1971;173:72-4.
Hollister LE, Kanter SL. Laboratory verification of
"heavy" and "light" users of cannabis. Drug Alcohol
Depend 1980;5:151-2.
Dackis CA, Pottash ALC, Annitto W, Gold MS. Persistence
of urinary marijuana levels after supervised abstinence.
Am J Psychiatry 1982;139:1196-8.
Clinical summary addendum, EMIT-d.a.u. and EMITSt
urine cannabinoid assays. Palo Alto, Calif: Syva Company,
1982:40-4.
Garrett ER, Hunt CA. Pharmacokinetics of A9-tetrahydrocannabinol
in dogs. J Pharm Sci 1977;66:395-407.
Jones AB, ElSholy HW, Arafat ES, ElSohly MA. Analysis
of the major metabolite of A9-tetrahydrocannabinol
in urine. IV. A comparison of five methods. J Anal Toxicol
1984;8:249-51.
Irving J, Leeb B, Foltz RL, Cook CE, Bursey JT, Willette
RE. Evaluation of immunoassays for cannabinoids in
urine. J Anal Toxicol 1984;8:192-6.
Peat MA, Finkle BS, Deyman ME. Laboratory evaluation
of immunoassay kits for the detection of cannabinoids
in biological fluids. In: Hawks RL, ed. The analysis
of cannabinoids in biological fluids. Rockville, Md: National
Institute on Drug Abuse 1982:85-98; NIDA Research
Monograph 42.
Foltz RL. Analysis of cannabinoids in physiological specimens
by gas chromatography/mass spectrometry. In:
Baselt RC, ed. Advances in analytical toxicology, vol I.
Foster City, Calif: Biomedical Publications, 1984:125-
57.
Williams PL, Moffat AC. Identification in human urine
of A9-tetrahydrocannabino1-11-oic acid glucuronide: A
tetrahydrocannabinol metabolite. I Pharm Pharmacol
1980;32:445-8.
Source: Excretion patterns of cannabinoid
metabolites after last use in a group of chronic users
