An Analysis of the Duration of Fentanyl and Its Metabolites

in Urine and Saliva

Jeffrey H. Silverstein, MD*, Michael F. Rieders, PhDZ, Matthew McMullin, M S ~ ,Steven
Schulman, MD*, and Kenneth Zahl, MD*
*Departmentof Anesthesiology, The Mount Sinai Medical Center, New York, New York and $National Medical Services,
Inc., Willow Grove, Pennsylvania

This study was undertaken to determine if metabolites was undetectable. Norfentanyl was present in larger
of fentanyl might be useful in the detection and mon- quantities than fentanyl immediately postoperatively
itoring of substance abuse. The presence of fentanyl and was detected in all patients at 48 h and in 4 of 7
and two of its metabolites in the urine and saliva of patients at 96 h. Despropionylfentanyl was not de-
seven female patients receiving small doses (110 ? 56 tected in any of the urine specimens tested. Neither
pg) of fentanyl was studied up to 96 h from the time of fentanyl nor its metabolites could be detected consis-
administration. Fentanyl and its two metabolites (nor- tently at any time in saliva. Saliva testing does not ap-
fentanyl and despropionylfentanyl) were extracted pear to be a viable alternative to urine testing based on
from samples and analyzed by gas chromatography/ this study. Urinary norfentanyl might be considered as
mass spectrometry. Unchanged fentanyl was detect- the substance of choicewhen testing for fentanyl abuse.
able in urine in all patients immediately postopera-
(Anesth Analg 1993;76:618-21)
tively and in 3 of 7 patients at 24 h. By 72 h, fentanyl

U rine testing is presently fundamental to the chemically dependent individuals and to determine if
detection and monitoring of chemically de- saliva would provide a viable alternative for the de-
pendent individuals (1). Fentanyl is the drug tection of fentanyl and its metabolites.
of choice for addicts in the anesthesia workplace (2).
Due to its short elimination half-life (t"P = 219 min) Methods
and approximately 90% metabolism (31, fentanyl is
This study was approved by the Institutional Review
difficult to detect in urine. Murphy et al. suggested
Board of The Mount Sinai Medical Center and in-
that norfentanyl would persist in the urine (4), and
formed consent was obtained from all participants.
Hammargren and Henderson, using a difficult extrac-
ASA I and I1 patients presenting for ovum retrieval
tion and detection method, were able to demonstrate
were included in this study. Seven female patients who
the presence of both fentanyl and norfentanyl in urine received less than 400 pg of fentanyl for operative an-
(5). The time course of the presence of these com- algesia were eligible for participation. The anesthetic
pounds in urine or saliva and the feasibility of using regimen for ovum retrieval consisted of intravenous
these methods as a monitor for substance abuse have sedative/analgesic technique with incremental bolus
not been evaluated. We tested urine and saliva for the doses of fentanyl, midazolam, thiamylal, propofol,
presence of fentanyl and two of its metabolites (nor- lidocaine, and /or droperidol. Medications used in con-
fentanyl and despropionylfentanyl) (Figure 1) from junction with fentanyl were not analyzed in the study,
individuals receiving less than 400 pg of fentanyl as a but were determined not to interfere with the analytical
component of intravenous analgesia for outpatient method (data not shown). Participants were asked to
surgery. We sought to determine the feasibility of provide urine (10-mLbottles) and saliva (1teaspoon or
using this method as a surveillance test or screen for approximately 4 mL) specimens preoperatively, imme-
~ ~~~~ ~~
diately postoperatively, and subsequently at 12,24,48,
Presented in part at the Annual Meeting of the American Society 72, and 96 h. Urine and saliva specimens were stored
of Anesthesiologists,San Francisco, California,October 1991. at -4°C (for sanitary reasons) and shipped frozen
Accepted for publication September 29, 1992. (-20°C) to one of the authors for analysis.
Address correspondenceand reprint requests to JeffreyH. Silver-
stein, MD, Department of Anesthesiology, Box 1010, Mount Sinai Fentanyl and its two metabolites were extracted
Medicaf Center, I Gustave L. Levy PIace, New York, NY 10029-6574. from patient samples using pH-controlled acid-base,

01993 by the International Anesthesia Research Society

618 he& Analg 1993;7661€&21 0003-2999/93/$5.00
 ANESTH ANALG SILVERSTEIN ET AL. 619
iw3;76:61azi DURATION OF FENTANYL. METABOLITES IN URINE

Fentanyl and Metabolites Table 1. Method Validation

Fentanyl Norfentanyl
0
Limit of detection (ng/mL) 0.1 0.1
Reporting limit (ng/mL) 0.2 0.2
Run-to-run precision at 10 5.9 6.3
ng/mL (% coefficient of variance)
Accuracy (at 10 ng/mL) +8.3% *9.7%
Fentanyl Analvte stabilitv at -10°C >6 mo >6 mo

Scientific, Folsom, CA). The operating conditions were

a helium carrier gas at a head pressure of 10.0 psi, col-
umn oven temperature programmed from 150°C to
Norfentanyl
Despropionylfentanyl 300°C at 35"C/min, final time 1 min, initial and inlet
purge time were both 2 min, injector and detector tem-
Figure 1. Fentanyl is metabolized in the body to norfentanyl and
despropionylfentanyl. Two other metabolites, hydroxyfentanyl and peratures were 270°C and 300"C, respectively. The vol-
norhydroxyfentanyl, which are found in lower quantities in human ume injected on the column was 2 pL. Gas
serum, were not evaluated in this study. chromatographic/mass spectral (GC/MS) analyses
were performed on a Hewlett-Packard Model 5988A
solvent extraction and back extraction (6-8). The ex- system using the same chromatographic parameters
tracted internal standard utilized was 8-meth- described above. The electron impact mass spectral
oxyloxapine (8-MLOX) (Aldrich Chemical Company, data were acquired in the selected ion monitoring
Milwaukee, WI). Aliquots (2 mL) of standards, con- mode (SIM) for the following ions (rn/z>:fentanyl, 146,
trols, blanks, and patient samples were placed into 189, 245; norfentanyl, 132, 158, 231; despropionylfent-
tubes. One milliliter of 10% sodium hydroxide and 100 anyl, 146, 189, 259; 8-MLOX, 247, 287 (5).
pL of 8-MLOX (10 ng/100 pL) were added to each tube. The manufacturer's certified primary standards uti-
After equilibration for 5 min, 4 mL of water/ lized were: fentanyl citrate (Janssen Pharmaceutica,
isopropanol (3:2, v/v), plus 6 mL of petroleum ether Piscataway, NJ), despropionylfentanyl, and norfenta-
were added. The tubes were capped and rotomixed for nyl (Goldmark Biologicals, Phillipsburg, NJ). Fentanyl,
10 min at 20 rpm, centrifuged for 5 min at 1400 g, and norfentanyl, despropionylfentanyl, and 8-MLOX dem-
allowed to stand for 10 min. The upper layer was trans- onstrated baseline separation in this chromatographic
ferred carefully to a tube containing 2 mL of 0.05 N HC1. system with typical retention times of 6.85, 6.22, 7.01,
The tubes were vortexed for 1 rnin, then centrifuged for and 7.37 min, respectively.
5 min at 1400 g. The lower aqueous layer was trans- All specimens were analyzed by GC/NP. Those
ferred to a tube containing 0.2 mL of 5%NaOH, and the specimens that were initially positive for fentanyl
tubes were briefly vortexed. Final extraction was and/or its metabolite(s1were analyzed by GC/MS for
achieved by adding 0.3 mL of dichloromethane, vor- confirmation of identity and quantitation. A represen-
texing for 1min, and then centrifuging for 2 min at 1400 tative sampling of negative patient specimens were run
g. The bottom organic layer, containing fentanyl and its as blank controls along with the positive patient sam-
metabolites, was transferred to a new tube for analysis. ples for both GC/NPand GC/MS. The validation char-
Norfentanyl and despropionylfentanyl were deriva- acteristics for the assays of fentanyl and norfentanyl are
tized by adding 5 pL of a mixture of to1uene:butyric presented in Table 1.
anhydride:4-dimethylaminopyridine(catalyst) (90:9:1)
to each tube. The tubes were heated at 70°C for 22 min Results
under a stream of nitrogen to prevent oxidation. The
residue was reconstituted with 30 pL of toluene and A representative GC/NP chromatogram is presented in
transferred to an auto sampler vial and sealed with a Figure 2. Demographicdata and results for each subject
crimp top. Gas chromatographic (GC) analyses were are presented in Table 2. The mean fentanyl dose ad-
performed with a Hewlett-Packard Model 5890A GC ministered was 110 ? 56 pg. The specimens were col-
equipped with a nitrogen phosphorus detector (GC/ lected at the times requested 2 2 h as reported by the
NP), 7673A liquid auto sampler,and a 3396A integrator. participants. The reporting limit for this assay is 0.2
The injections were made in the splitless mode into a ng/mL for both fentanyl and norfentanyl. Samples re-
4-mm glass injection port liner, followed by an analyt- ported as negative represent levels <0.2 ng/mL. Fent-
ical 15-m X 0.32-mm internal diameter fused silica cap- anyl was detected in all immediate postoperative urine
illary column with a bonded 0.15-pm DB-17 film U&W specimens, in 3 of 7 specimens at 12 and 24 h, and
620 SLVERSTEIN ET AL. -TI4 ANALG
DURATION O F FENTANYL METABOLITES IN URINE 1993;7661&21

Table 2. Demographic Data and Results

Urine fentanyl (ng/mL) Urine norfentanyl (ng/mL)
Other
Age (yr)" Fentanyl ( p g ) b medications' Postop 12 h 24 h 48 h 72 h 96 h Postop 12 h 24 h 48 h 72 h 96 h
35 200 M 4 mg 0.8 0.2 0 0.9 0 0 5.8 13.8 5.3 4.7 1.3 0.3
43 170 M 4 mg 2.0 0 0.2 0 0 0 5.4 3.7 1.9 2.6 0.2 0
D1.25
32 100 M 4 mg 0.9 0 0 0 0 0 11.5 10.2 3.1 1.7 0.7 0.7
L 25 mg
T 250 mg
38 100 M 2 mg 0.5 0 0 0 0 0 10.9 7.6 0 0.5 0 0
P 90 mg
L 50 mg
40 100 M 2 mg 4.7 0.4 0.4 0 0 0 10.0 23.0 3.3 0.8 0.6 0.5
D 0.25 mg
T 250 mg
38 50 M 4 mg 1.9 0.3 0.6 0 0 0 8.5 20.0 12.0 0.4 0.7 0
P 100 mg
D 1.25 mg
42 50 M 4 mg 0.8 0 0 0 0 0 10.0 1.1 0 0.6 0.4 0.6
P 150 mg
D 1.25 mg
, Average age = 38.3 i 3.9 yr.
Average fentanyl dose = 110 i 56 pg.
M = midazolam; P = prupofol; T = thiamylal; D = droperidol; L = Lidocaine.

in one specimen at 48 h. Fentanyl could not be detected Discussion

at 72 or 96 h. By contrast, norfentanyl was detected in
greater quantities at all times and in 4 of 7 samples at In this study, seven female patients undergoing ovum
96 h (Figure 3). Despropionylfentanylwas not detected retrieval received analgesia with 1200 pg of fentanyl.
in any of the urine specimens. Initial screening of the Urine and saliva specimens were analyzed for the pres-
preoperative specimen from one patient was positive ence of fentanyl and two of its metabolites, norfentanyl
by GC/NP for norfentanyl (0.7 ng/mL), but GC/MS and despropionylfentanyl. Our results are consistent
analysis did not confirm the initial positive GC/NP with the known disposition of fentanyl in humans
screen. which involves excretion of approximately 90% of an
The requested quantities of saliva, 5 mL, proved dif- administered dose as metabolites (5).At 72 h, fentanyl
ficult to obtain and quantities of 1-2 mL were typically could not be detected in any urine sample, but nor-
produced. Neither fentanyl nor despropionylfentanyl fentanyl was detected in 86% of samples. Des-
were detected in any saliva specimens. Norfentanyl propionylfentanyl was not detected at any time in any
was detected in two saliva specimens from one patient of the urine specimens in this study.
who received 100 pg of fentanyl, but could not be de- Hammargren and Henderson described a different
tected in any of the other 43 specimens. extraction and detection procedure for fentanyl and
norfentanyl(5). They evaluated an unspecified number
of individuals, but did not determine the time course
II of the presence of these normetabolites in urine.

f l McClain and Hug (3) and Murphy et al. (4)suggested

that norfentanyl would persist in urine as confirmed by

Q - I I/
a
our results.
The doses of fentanyl used in this study (50-200
pg) are small compared to the quantities that may be
used by a fentanyl addict. Our interest is the moni-
toring of the recovering addict as well as the detec-
tion of the early addict. The doses studied are appro-
priate to these situations. Fentanyl metabolism and
Figure 2. Gas chromatographic/nitrogen phosphorus chromato- excretion into urine were highly variable among indi-
graph showing separation of fentanyl and three metabolites and
8-methoxyloxapine (8-MLOX), the internal standard. The numbers viduals. There was no discernible relationship be-
represent retention times on the chromatographiccolumn. tween the administered fentanyl dose and the levels
 SKVERSTEIN ET AL. 621
DURATION OF FENTANYL METABOLITES I
N URINE

Duration of Norfentonyl a n d Fentanyl in Urine careful confirmation procedures in urine drug testing.
means i S D
20 -
. This patient denied use of fentanyl or other analgesics,
and GC/MS analysis made clear that the substance was
18
0 Norfentonyl not norfentanyl.She was the same patient for which the
16 Fentoriyi saliva norfentanyl samples also were positive. We have
no explanation for the differences in this patient's re-
c.1 4
E
\
sults. Although the GC/NP screen will identify a sub-
y 12 stance that has similar chromatographic characteristics

I
v 1
c
.? 10
as, in this example, norfentanyl, the molecular struc-
c
0 ture was distinctly different than norfentanyl, as re-
solved by mass spectrometry.
8 6
Saliva proved difficult to collect according to the pa-
tients studied. Four samples contained volumes insuf-
4 ficient to analyze. Norfentanyl was detected in two sa-
2
liva specimens from one patient at 12 h (0.6 ng/mL)
and at 72 h (1.9 ng/mL). Fentanyl and despropio-
0 ~

POST-OP 1 2 HOURS 24 HOURS 48 HOURS 72 HOURS 9 6 HOURS nylfentanyl were not detected in any saliva sample.
Collection Time
Thus, it would appear that saliva sampling is not a
Figure 3. Utilizing an assay with a reporting limit of 0.2 ng/mL, good alternative to urine sampling.
norfentanyl was detected in 4 of 7 samples at 96 h. Fentanyl levels Norfentanyl is detectable in the urine for a consid-
approached the reporting limit by 12 h. erably longer period of time than fentanyl. Drug testing
programs either for monitoring recovering addicts or
of fentanyl and norfentanyl detected in urine. Some screening health care professionals with access to fent-
of the variability may be attributable to differences in anyl might consider testing for norfentanyl rather than
urine specific gravity or creatinine clearance. The ef- fentanyl. The ability to detect norfentanyl from 48 to 96
fect, if any, of the other coadministered drugs on fent- h postadministration will enhance the quality and util-
anyl metabolism and excretion is not known. Al- ity of monitoring programs. Further, larger scale stud-
though the present study included only female ies should be undertaken to better understand the im-
subjects, there are no data to suggest that fentanyl plications of urine testing for fentanyl and its
metabolism is significantly different in males. metabolites. Radioimmunoassays under development
Other medications coadministered for sedation (Ta- will require similar confirmation.
ble 2) were not monitored or analyzed in this study. It
was determined, however, that neither these sub- References
stances nor other common substances of abuse inter- 1. Canavan DI. Screening: urine drug tests. Md Med J 1987;36:
fere with the quantitative identification of fentanyl or 229-33.
2. Ikeda R, Pelton C. Diversion programs for impaired physicians.
its metabolites (data not shown). Because individuals West J Med 1990;152:617-21.
who abuse drugs frequently abuse more than one sub- 3. McClain DA, Hug CC Jr. Intravenous fentanyl kinetics. Clin
stance, the detection of fentanyl or its metabolites Pharmacol Ther 1980;28:106-14.
4. Murphy MR, Hug CC Jr,McClain DA. Dose-independent phar-
among the mix of other drugs and metabolites repre- macokinetics of fentanyl. Anesthesiology 1983;59:53740.
sents realistic conditions. 5. Hammargren WR, Henderson GL. Analyzing normetabolites of
Situations leading to urine fentanyl or fentanyl me- the fentanyls by gas chromatography/electron capture detec-
tabolite concentrations of less than 0.2 ng/mL will re- tion. J Anal Toxicol 1988;12:183-91.
6. Van Rooy HH, Vermeulen MI', Bovill JG.The assay of fentanyl
sult in a false negative finding. False positive results are and its metabolites in plasma of patients using gas chromatog-
unlikely because the method requires that the GC re- raphy with alkali flame ionisation detection and gas
tention time index and the MS ion ratios be within re- chromatography-mass spectrometry. J Chromatogr 1981;223:
85-93.
stricted limits. False positives can result, however, from 7. Goromaru T, Matsuura H, Yoshimura N, et al. Identification and
sample mix up, unexpected carryover ("ghosting") quantitative determination of fentanyl metabolites in patients
from a previous high concentration sample, and other by gas chromatography-mass spectrometry. Anesthesiology
erroneous analytical and handling techniques. The 1984;61:73-7.
8. Gillespie TJ,Gandolfi AJ,Maiorino RM, Vaughan RW. Gas chro-
presence of norfentanyl on the initial screening (by matographic determination of fentanyl and its analogues in hu-
GC/NP) for one patient highlights the necessity of man plasma. J Anal Toxicol 1981;5133-7.