See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/40039812

Determination of naltrexone and 6β-naltrexol in human blood: Comparison

of high-performance liquid chromatography with spectrophotometric and
tandem-mass-spectrometric detection

Article  in  Analytical and Bioanalytical Chemistry · November 2009

DOI: 10.1007/s00216-009-3301-z · Source: PubMed

CITATIONS READS

10 538

8 authors, including:

Sonja Brünen Felix Korf

Johannes Gutenberg-Universität Mainz Institute of group analysis
9 PUBLICATIONS   102 CITATIONS    4 PUBLICATIONS   13 CITATIONS   

SEE PROFILE SEE PROFILE

Falk Kiefer Christoph Hiemke

Central Institute of Mental Health Johannes Gutenberg-Universität Mainz
457 PUBLICATIONS   8,900 CITATIONS    891 PUBLICATIONS   15,836 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Sybil-AA: Systems Biology of Alcohol Addiction View project

alcohol addiction View project

All content following this page was uploaded by Christoph Hiemke on 23 May 2015.

The user has requested enhancement of the downloaded file.

Anal Bioanal Chem
DOI 10.1007/s00216-009-3301-z

ORIGINAL PAPER

Determination of naltrexone and 6β-naltrexol in human

blood: comparison of high-performance liquid
chromatography with spectrophotometric
and tandem-mass-spectrometric detection
Sonja Brünen & Ralf Krüger & Susann Finger &
Felix Korf & Falk Kiefer & Klaus Wiedemann &
Karl J. Lackner & Christoph Hiemke

Received: 20 August 2009 / Revised: 3 November 2009 / Accepted: 5 November 2009

# Springer-Verlag 2009

Abstract We present data for a comparison of a liquid-
chromatographic method coupled with tandem mass
spectrometry (LC-MS/MS) and a high-performance
liquid-chromatographic method with column switching
and UV spectrophotometric detection. The two methods
were developed for determination of naltrexone and 6β-
naltrexol in blood serum or plasma aiming to be used for
therapeutic drug monitoring to guide the treatment of
patients with naltrexone. For the high-performance liquid
chromatography (HPLC)/UV detection, online sample
cleanup was conducted on Perfect Bond C18 material
with 2% (vol/vol) acetonitrile in deionized water. Drugs
were separated on a C18 column using 11.5% (vol/vol)
acetonitrile and 0.4% (vol/vol) N,N,N,N-tetramethylethy-
lenediamine within 20 min. LC-MS/MS used naltrexone-d3
S. Brünen : S. Finger : F. Korf : C. Hiemke (*)
Department of Psychiatry and Psychotherapy,
University of Mainz,
Untere Zahlbacherstraße 8,
55131 Mainz, Germany
e-mail: hiemke@mail.uni-mainz.de
R. Krüger : K. J. Lackner
Institute of Clinical Chemistry and Laboratory Medicine,
University of Mainz,
55131 Mainz, Germany
F. Kiefer
Central Institute for Mental Health, University of Heidelberg,
68159 Mannheim, Germany
K. Wiedemann
Department of Psychiatry and Psychotherapy,
University of Hamburg,
20146 Hamburg, Germany
and 6β-naltrexol-d4 as internal standards. After protein
precipitation, the chromatographic separation was performed
on a C18 column by applying a methanol gradient (5–100%,
vol/vol) with 0.1% formic acid over 9.5 min. The HPLC/UV
method was found to be linear for concentrations ranging
from 2 to 100 ng/ml, with a regression correlation coefficient
of r2 >0.998 for naltrexone and 6β-naltrexol. The limit of
quantification was 2 ng/ml for naltrexone and 6β-naltrexol.
For the LC-MS/MS method the calibration curves were
linear (r²>0.999) from 0.5 to 200 ng/ml for both substances,
and the limit of quantification was 0.5 ng/ml. The concen-
trations measured by the two methods correlated significant-
ly for both substances (r²>0.967; p<0.001). Both methods
could be used for therapeutic drug monitoring. The HPLC/
UV method was advantageous regarding automatization and
costs, whereas LC-MS/MS was superior with regard to
sensitivity.
Keywords Naltrexone . 6ß-Naltrexol .
Therapeutic drug monitoring . High-performance liquid
chromatography/UV detection . Liquid chromatography/
tandem mass spectrometry
Introduction
Naltrexone is a potent µ-opiate receptor antagonist that is
used in the treatment of opiate and alcohol dependence. It is
rapidly metabolized by the cytosolic enzyme family
dihydrodiol dehydrogenases to its active main metabolite
6β-naltrexol [1] (Fig. 1). The naltrexone metabolite 6α-
naltrexol is not a product of naltrexone in humans [2] and
 S. Brünen et al.

Fig. 1 Metabolic pathway of

naltrexone
HO HO HO
DD (1-4)

O O + O
OH H
OH H OH H
N N N
H H
O HO HO

main metabolite (active)

naltrexone 6β-naltrexol 6α-naltrexol

OH OH
O O

O O
OH H OH H
N N
H
O HO
2-hydroxy-3-methoxynaltrexone 2-hydroxy-3-methoxy-6β-naltrexol

its pharmacological effects are unclear. 6β-Naltrexol has
weaker pharmacological potency than naltrexone [3, 4].
Because of the longer half-life of 12 h compared with 4 h
for naltrexone, however, 6β-naltrexol concentrations in
blood are markedly higher than those of the mother
compound [5]. Therefore the metabolite contributes signif-
icantly to the clinical observable effects of naltrexone.
Different analytical procedures have been described for
naltrexone and in a few cases also for 6β-naltrexol.
Reported methods used thin-layer chromatography [6],
high-performance liquid chromatography (HPLC) with
electrochemical detection [7–11], gas chromatography with
flame ionization detection [12, 13], and gas chromatogra-
phy coupled with mass spectrometry (MS) [14]. The latest
developments use liquid chromatography (LC) with MS
detection [15] and LC-MS/MS [16–18], primarily for
determination of naltrexone in animal plasma. Existing
methods use time-consuming off-line sample cleanup [7–
11, 16, 18], including liquid-phase or solid-phase extraction
steps, which is disadvantageous for a routine therapeutic
drug monitoring (TDM) survey aiming to report results
within a single day after blood withdrawal [19]. Our goal
was to develop a HPLC/UV method which needs minimum
sample preparation and has sufficiently low limit of quanti-
fication, covering the currently discussed therapeutic range for
TDM. On the other hand, LC-MS/MS has better sensitivity,
and we decided to develop a second, MS-based method suited
for lower plasma levels. This is also necessary for further
studies concerning the therapeutic range, which is still under
discussion. A recently reported LC-MS/MS method [17] used
online sample cleanup and column switching which allows
rapid and automated analysis. The analytical system,
however, was a very sensitive LC-MS/MS system of the
newest generation which is not yet available in every
laboratory. Therefore, we aimed to develop a robust method
for use with a standard MS instrument without the need for
extensive sample treatment or enrichment.
To evaluate if a method achieves the criteria for use in a
TDM laboratory, it is necessary to bring concentrations into
clinical context. Vereby et al. [12] have shown that 100 mg
orally administered naltrexone did not completely antago-
nize the effect of 25 mg intravenously administered heroin
when the plasma level of naltrexone was below 2 ng/ml.
Two other groups showed an antagonizing effect even when
naltrexone plasma levels were quite low (0.3 ng/ml) [20,
21]. Naltrexone, however, is rapidly metabolized to its active
main metabolite 6β-naltrexol [1]. In a pharmacokinetic study
following single-dose administration, the mean ratio of the
concentration of the parent drug to the concentration of the
metabolite was 0.3:10 after 8 h [22]. This indicates that
6β-naltrexol is the predominant pharmacological agent under
treatment with naltrexone. Regarding pharmacokinetic studies
and doses that are used for the treatment of patients with
alcohol dependence, it seems likely that steady-state plasma
concentrations ranging between 3 and 40 ng/ml for naltrexone
and between 35 and 70 ng/ml for 6β-naltrexol may be
expected [23]. A correlation between the sum of the plasma
levels and a therapeutic effect was not examined up to now.
Here, we developed two different methods for determina-
tion of naltrexone and 6β-naltrexol in serum or plasma aiming
to be used for therapeutic drug monitoring to guide the
treatment of patients with naltrexone: a LC-MS/MS method
and a HPLC method with column switching and UV
spectrophotometric detection. The two methods were validat-
ed, applied to the analysis of patient samples, and compared
with regard to their applicability in clinical routine.
Determination of naltrexone and 6β-naltrexol in blood: comparison of HPLC-UV with LC-MS/MS
Patients and methods
Patients and plasma samples
We determined plasma levels of 186 samples from patients
receiving naltrexone. The patients had been included in two
clinical trials. The study was conducted according to the
European good clinical practice guidelines [24] and the
World Medical Association Declaration of Helsinki study
[25] and was approved by two ethics committees. All
participants provided written informed consent. Blood was
drawn under steady-state conditions on the morning before
the first dose. Plasma was prepared by centrifugation of
blood samples at 4,000g for 10 min and stored at -80°C
until assayed.
Chemicals and reagents
Naltrexone, 6β-naltrexol, and Optigrade acetonitrile were
obtained from LGC Promochem (Wesel, Germany),
naltrexone-d3 was obtained from Biomol (Hamburg,
Germany), and 6β-naltrexol-d4 was obtained from Toronto
Research Chemicals (North York, Canada). Acetic acid was
purchased from Merck (Darmstadt, Germany) and N,N,N,
N-tetramethylethylenediamine was obtained from Sigma
(Taufkirchen, Germany). Formic acid puriss. p.a. for MS
was obtained from Fluka Chemie (Buchs, Switzerland),
water of HPLC gradient grade was purchased from J.T.
Baker (Griesheim, Germany), and methanol of HPLC
grade was obtained from Fisher Scientific (Schwerte,
Germany).
Calibrators and quality control samples
Naltrexone and 6β-naltrexol calibrators and control samples
as well as the internal standards naltrexone-d3 and 6β-
naltrexol-d4 were prepared from stock solutions (1 mg/ml in
methanol). Quality control samples for measuring precision,
accuracy, and recovery and standards for measuring linearity
and for preparation of calibrators were first prepared by
dilution of stock solutions in blank plasma (human plasma
samples obtained from healthy volunteers). Quality control
samples and standards were prepared by using separate
prepared stock solutions. Further quality controls were then
prepared by pooling twice ten plasma samples from patients
under naltrexone therapy with different comedications.
Samples for recovery determination were prepared by
dilution of stock solutions in methanol for the LC-MS/MS
method and in methanol/water (60:40 vol/vol) for the HPLC/
UV method. Overall, samples with 11 different concen-
trations were prepared for both methods (0.1–200 ng/ml).
The internal standard solutions for the LC-MS/MS method
were prepared by dilution of the stock solution in methanol.
The protein precipitation/internal standard solution for the
LC-MS/MS method (8:2 vol/vol methanol/0.2 M ZnSO4,
33 ng/ml naltrexone-d3 and 6β-naltrexol-d4) was prepared
freshly before each precipitation.
Samples could be stored in the dark at -20°C for several
months without measurable decomposition.
Sample extraction
For HPLC/UV analysis, samples were thawed and centri-
fuged at 13,000 g for 5 min and supernatants were
transferred into autosampler tubes.
For LC-MS/MS analysis, protein precipitation was the
only cleanup step before chromatographic separation. Three
hundred microliters of protein precipitation/internal stan-
dard solution was added to 200µl plasma. The mixture was
vortexed for 30 s and centrifuged at 13,000g for 5 min. The
supernatant was transferred into autosampler tubes.
HPLC/UV conditions
For simultaneous determination of naltrexone and its
active main metabolite 6β-naltrexol, HPLC with column
switching and UV spectroscopy was used. HPLC/UV
analysis was performed using an Agilent 1100 system
obtained from Bio-Rad (Munich, Germany) consisting of
two HPLC pumps, an autosampler, a thermostatted
column set at 40°C with an electric six-port switching
valve coupled to the autosampler, and a variable-
wavelength UV detector which was set at 225 nm. The
HP ChemStation chromatography data system was used
for data acquisition and integration. Two hundred micro-
liters were injected and an online sample cleanup was
carried out, which was conducted on Perfect Bond C18
material of 20-µm particle size (MZ-Analysentechnik,
Mainz, Germany), using 2% (vol/vol) acetonitrile in
deionized water as the eluent at a flow rate of 1.0 ml/min.
However, online extraction enabled injection of higher
samples volumes, which in turn leads to lower detection
limits. In our method, 100µl were injected twice before
the analyte was flushed onto the analytical column.
Proteins and other interfering matrix constituents of the
serum were separated and washed to waste. After 7 min,
the electric six-port valve switched and the second pump
transported the sample by back-flush mode into the
analytical column. Naltrexone and 6β-naltrexol were
separated on ODS Hypersil C18 material (3µm; column
size 150 mm×3.0 mm inner diameter) at a flow rate of
0.7 ml/min using 11.5% (vol/vol) acetonitrile and 0.4%
(vol/vol) N,N,N,N-tetramethylethylenediamine in deion-
ized water adjusted to pH 6.0 with acetic acid. Naltrexone
and 6β-naltrexol could be analyzed within 20 min by
evaluation of the peak height.

LC-MS/MS conditions
LC-MS/MS analysis was performed using a triple-
quadrupole mass spectrometer (TSQ Quantum Discovery
Max, Thermo Scientific, Fremont, CA, USA) equipped with
a gradient HPLC pump and autosampler (Surveyor, Thermo
Scientific, Fremont, CA, USA). Electrospray ionization in
positive ion mode and multireaction monitoring (MRM)
were applied. Protonated species [M+H]+ were monitored
using the following MRM transitions: for naltrexone m/z=
342 → 324, for 6β-naltrexol m/z = 344 → 326, and for the
internal standards naltrexone-d3 and 6β-naltrexol-d4 m/z =
345 → 327 and m/z = 348 → 330, respectively. For all
transitions, 21-V collision energy, 150-V tube lens offset,
and 0.1-min scan time were used. The electrospray ioniza-
tion source parameters were set as follows: spray voltage
4.0 kV, sheath gas 35 AU, auxiliary gas 5 AU, capillary
temperature 240°C, source collision-induced dissociation
0 V, capillary offset 35 V. The divert valve was set to waste
until 3.5 min to prevent salts and impurities entering the ion
source.
Chromatographic gradient separation was performed on an
MZ-Aqua Perfect C18 column (5µm; 2.1 mm×150 mm)
(MZ-Analysentechnik, Mainz, Germany) using a column
temperature of 30°C and a flow rate of 0.5 ml/min. Eluent A
consisted of water containing 0.1% (vol/vol) formic acid and
eluent B was methanol with 0.1% (vol/vol) formic acid.
Starting conditions were 95% eluent A and 5% eluent B. A
5.5-min gradient from 5% eluent B to 70% eluent B was
applied, followed by a 2-min wash step at 100% eluent B
from 5.5 to 7.5 min, and 2 min equilibration time at 5%
eluent B from 7.5 to 9.5 min. The injection volume was 10µl.
Validation
For assay validation the validation software program
Valistat which was developed by the Gesellschaft für
Forensische und Toxikologische Chemie was used. The
examination was carried out according to ISO 5725 and
DIN 32645. Selectivity was ascertained by determination of
six different blank samples without internal standards and
also two blank samples with internal standard with the LC-
MS/MS method. Linearity was verified with six calibration
samples (2, 5, 10, 20, 50, 100 ng/ml) for the HPLC/UV
method and eight calibration samples (0.5, 1, 2, 5, 10, 20,
100, 200 ng/ml) for the LC-MS/MS method. The calibra-
tion curve was measured over 6 days. The Grubbs test, the
F test, and the Mandel test were implemented in the
software program, and the calibration curves had to have a
correlation coefficient r²>0.99. Precision and accuracy
were determined using the quality control samples contain-
ing 2, 5, 20, and 100 ng/ml for the HPLC/UV method and
0.5, 5, 50, and 100 ng/ml for the LC-MS/MS method.
S. Brünen et al.
Precision was calculated as within-run, between-run, and
between-day imprecision. Precision was also determined by
using pooled plasma samples from patients under naltrex-
one therapy and was calculated as within-run imprecision.
Accuracy was calculated against the weighed-in quantity of
the control samples.
The limits of quantification were accepted when the
precision was better than 15% with a signal-to-noise ratio
greater than 6. The limit of detection was defined as a
signal-to-nose ratio greater than 3.
Recovery was determined by analyzing samples with
two different concentrations of the drugs (5 and 100 ng/ml).
For the LC-MS/MS method we measured drug concen-
trations in plasma samples after protein precipitation and in
control samples consisting of analytical eluent supple-
mented with drugs and protein precipitation solution. To
determine the recovery rates for the HPLC/UV method,
solvent samples were measured without and with inclusion
of cleanup columns. Interferences were checked with 33
psychoactive drugs, primarily antipsychotics and antide-
pressants. They were prepared in drug-free plasma and
tested for possible interferences.
Statistical analyses
The statistical analyses were performed by the statistical
software program SPSS version 17.0. For the method
comparison, a bivariate correlation by the Pearson test and
the t test was carried out. Statistical significance was
predefined as p<0.05.
Results
HPLC/UV method
The HPLC/UV method with column switching enabled
automatic analysis of naltrexone and 6β-naltrexol within
20 min (Fig. 2). Linearity was tested in a range between 2
and 100 ng/ml for both substances. The calibration curves
were linear according to linear regression analysis, which
revealed correlation coefficients of r²>0.998. The Grubbs
test showed no straggler or outlier and the F test
demonstrated homogeny for all six runs. Linearity was
proven by the Mandel test. Accuracy and precision results of
the method, analyzed by measuring three concentrations of
quality control samples and a pooled plasma sample, are given
in Table 1. For both naltrexone and 6β-naltrexol, the limit of
quantification was 2 ng/ml and the limit of detection was
1.2 ng/ml. Recoveries were determined with and without a
cleanup column with plasma and solvent samples and were
found to be higher than 85% for the low-concentration and
the high-concentration ranges. Concerning the tested inter-
Determination of naltrexone and 6β-naltrexol in blood: comparison of HPLC-UV with LC-MS/MS

mAU Table 2 Retention times of other possible coadministered psychotro-

15 pic drugs measured by the HPLC method

11.681 6β-naltrexol

12.942 naltrexone
Substance Retention time (min)
10
Naltrexone 12.9
5 6β-Naltrexol 11.7
Antipsychotics
0 Clozapine ND
Haloperidol 16.8
N-Desmethylperazine ND
-5
Olanzapine ND
Perazine ND
-10
Quetiapine ND
Sulpiride 16.5
-15
Zotepine ND
9 10 11 12 13 14 min
Antidepressants
Fig. 2 High-performance liquid chromatography UV (HPLC/UV) Amitriptyline ND
chromatogram of naltrexone and 6β-naltrexol (50 ng/ml of each Citalopram ND
compound)
Doxepin 18.4
Fluoxetine ND
ferences, none of these drugs had a retention time interfering Imipramine ND
with the elution of naltrexone or 6β-naltrexol determined Maprotiline ND
with the HPLC/UV method (Table 2). Mirtazapine ND
N-Desmethylcitalopram ND
LC-MS/MS method N-Desmethylclomipramine ND
N-Desmethyldoxepin ND
One important step during LC-MS/MS method develop- N-Desmethylimipramine ND
ment was to find an appropriate gradient, since it was N-Desmethylmaprotiline ND
necessary to separate naltrexone from 6β-naltrexol owing N-Desmethylsertraline ND
to overlapping isotope patterns: the first isotope peak O-Desmethylvenlafaxine 18.5
(basically 13C) of 6β-naltrexol has the same nominal mass Paroxetine ND
as the monoisotopic peak of the internal standard Sertraline ND
naltrexone-d3 (m/z 345.1). Differentiation would require Tianeptine 16.3
very high mass-spectrometric resolution, because the m/z Trimipramine ND
values differ by only 0.17 mDa (345.18899 for
Venlafaxine ND
C1913C1H26O4N1 versus 345.18882 for C20H21D3O4N1).
Other substances
Since the most abundant fragmentation reaction is the loss
Donepezil 16.3
of water (-18 Da), collision-induced dissociation does not
Oxazepam 16.3
result in specific fragment ions with acceptable intensity,
and consequently these two substances can not be discrim- ND not detectable

Table 1 Coefficients of variation of imprecision (%) and accuracy (%) for determination of a weighed-in quantity naltrexone/6β-naltrexol using
the high-performance liquid chromatography (HPLC)/UV method

Naltrexone 6β-Naltrexol

Low 1 Low 2 Middle High Pool Low 1 Low 2 Middle High Pool

Concentration (ng/ml) 2 5 20 100 2.1 2 5 20 100 38.7

Within-run imprecision (%) 13.3 11.2 14.5 2.52
Between-day imprecision (%) 5.5 5.3 5.2 3.0 6.9 7.0
Between-run imprecision (%) 3.0 4.2 5.0 2.6 6.9 7.0
Accuracy (%) 98.9 94.7 99.4 100.0 99.6 98.9
 S. Brünen et al.

RT: 0.0 - 7.0

100
RT: 4.5 The within-run precision, intraday prescision, and between-
TIC F: + c SRM
day precision were between 2.1 and 3.3%, and accuracies
50 RT: 4.1 were found to be between 96.7 and 101.3% (Table 3). For
both substances, the limit of quantification was 0.5 ng/ml
0 RT: 4.1
100 NL: 2.18E3 and the limit of detection was 0.2 ng/ml. Figure 4 shows a
naltrexone chromatogram of a sample with a concentration of 0.5 ng/
Relative Abundance

342.10@-21.00
50 [324.10-324.11] ml. Comparison of methanolic standards versus spiked
0 plasma samples (including precipitation step) showed high
100 RT: 4.5
NL: 1.72E5 recoveries of 95% for the low-concentration range (5 ng/ml)
6β-naltrexol
6β-naltrexol 344.10@-21.00 and 99% for the high-concentration range (100 ng/ml).
50 [326.10-326.11]
Naltrexone, naltrexone-d3, 6β-naltrexol, and 6β-naltrexol-d4
0 peaks were free of interferences from endogenous com-
100 RT: 4.1
NL: 1.03E5
pounds in the plasma. Of the 33 drugs tested for interfer-
naltrexone-d3 RT: 4.5 345.10@-21.00 ences, citalopram, desmethylperazine, clozapine, and
50
[327.10-327.11]
desmethylclozapine had similar masses. None of these
0
100 RT: 4.5 substances were detectable under the LC-MS/MS conditions
NL: 1.09E5
described.
6β-naltrexol-d4
6β-naltrexol-d4 348.10@-21.00
50
[330.10-330.11]
Comparison of LC-MS/MS and HPLC/UV methods
0
0 1 2 3 4 5 6 7
Time [min] For comparison of the HPLC/UV method and the LC-MS/
MS method, 186 plasma samples from patients under
Fig. 3 Representative chromatogram, which demonstrates the sepa-
ration of naltrexone and 6β-naltrexol, and the interference in the naltrexone treatment were analyzed by both methods. The
naltrexone-d3 multireaction monitoring trace due to the first isotope LC-MS/MS method found measurable concentrations of
peak of 6β-naltrexol, which has the same nominal mass as the naltrexone and 6β-naltrexol in 95 and 135 samples,
monoisotopic peak of naltrexone-d3. The concentrations in this patient respectively. In six of the samples that tested positive with
sample were calculated as 1.9 ng/ml for naltrexone and as 65.4 ng/ml
for 6β-naltrexol the LC-MS/MS method, the HPLC/UV method failed to
detect measurable concentrations of either naltrexone or
6β-naltrexol.
inated by MRM. With the gradient described above, Both methods found similar concentrations. The mean±the
naltrexone and its metabolite were separated by 0.4 min at standard deviation concentration of naltrexone was 13±20 ng/
retention times of 4.1 and 4.5 min for naltrexone and 6β- ml when using the HPLC/UV method and 14±20 ng/ml when
naltrexol, respectively. No shift of retention times was using the LC-MS/MS method. For 6β-naltrexol, the mean
observed for the deuterated compounds. An example of the concentrations were 58±33 and 57±32 ng/ml, respectively.
separation is shown in Fig. 3. The total run time was As shown in Figs. 5 and 6, the concentrations measured by
9.5 min, including the equilibration time. Standard curves the two methods correlated significantly (p <0.001) over the
for naltrexone and 6β-naltrexol in drug-free plasma were whole range of calibration for both naltrexone (r2 =0.996)
linear in a range of 0.5–200 ng/ml. Linearity was shown by and 6β-naltrexol (r2 =0.967). This was also found for
the Mandel test and linear regression analysis (r²>0.999). samples with either low (naltrexone less than 12 ng/ml;

Table 3 Coefficients of variation of imprecision (%) and accuracy (%) for determination of a weighed-in quantity naltrexone/6β-naltrexol using
the liquid chromatography (LC)–tandem mass spectrometry (MS/MS) method

Naltrexone 6β-Naltrexol

Low 1 Low 2 Middle High Pool Low 1 Low 2 Middle High Pool

Concentration (ng/ml) 0.5 5 50 100 0.56 0.5 5 50 100 33.75

Within-run imprecision (%) 12.6 8.18 11.2 2.13
Between-day imprecision (%) 3.3 2.5 2.1 2.7 2.8 2.4
Between-run imprecision (%) 2.5 2.3 2.1 2.4 2.8 1.0
Accuracy (%) 98.4 101.3 97.9 96.7 98.9 98.2
Determination of naltrexone and 6β-naltrexol in blood: comparison of HPLC-UV with LC-MS/MS

RT: 0.0-7.0
RT:4.5
100 RT:4.1
TIC F: MS 28
50

0
100 NL: 8.02E2
RT: 4.1 342.10@ -21.00
50
naltrexone [324.10 -324.11]
Relative Abundance

0
100 NL: 4.87E3
RT: 4.5 344.10@ -21.00
50 6β-naltrexol [326.10 -326.11]

0
100 RT: 4.1 NL: 1.01E5
345.10@ -21.00
50 naltrexone-d3 [327.10 -327.11]

0
100 RT: 4.5 NL: 1.37E5 Fig. 6 Determination of 6β-naltrexol: correlation of the concentra-
6β-naltrexol -d4 348.10@ -21.00 tions measured by the HPLC/UV method and the LC-MS/MS method
50 [330.10 -330.11]

0
0 1 2 3
Time [min]
Fig. 4 Chromatogram of a standard sample containing 0.5 ng/ml
naltrexone and 0.5 ng/ml of its metabolite 6β-naltrexol
6β-naltrexol less than 50 ng/ml) or high concentrations of
the drugs (naltrexone more than 12 ng/ml; 6β-naltrexol more
than 50 ng/ml). For naltrexone, the correlation coefficients
(r2) were 0.996 and 0.996, respectively, and for 6β-naltrexol
they were 0.963 and 0.905, respectively, and always reached
high levels of significance (p<0.001).
4 5 6 7 Discussion
We have presented two new methods for determination of
naltrexone and its main active metabolite 6β-naltrexol. The
HPLC method with column switching and UV detection
accommodated a minimum of sample preparation. The
calibration curves were linear over a range of 2–100 ng/ml
and the limits of quantification were 2 ng/ml for naltrexone
and 6β-naltrexol. For the LC-MS/MS-method, the calibra-
tion curves were linear from 0.5 to 200 ng/ml and the lower
limit of quantification was 0.5 ng/ml for both substances.
For the LC-MS/MS method, a single-step sample prepara-
tion was used. Since the method was developed for

Table 4 Method characteristics including costs for the HPLC/UV and

LC-MS/MS methods for determination of naltrexone and 6β-naltrexol

HPLC/UV LC-MS/MS

Cleanup procedure Online solid-phase Off-line protein

extraction precipitation
Chromatographic run 20 9.5
time (min)
Consumption of organic 100 ml (thereof 215 ml
solvent/100 samples 80 ml at backflow)
Sample capacity/column 100 >600
Linearity (ng/ml) 2–100 0.5–200
LLOQ (ng/ml) 2 0.1
Imprecision (%) <11.5 <8.5
Inaccuracy (%) <5.5 <3
Costs according to [27] 17.30 51.90
Fig. 5 Determination of naltrexone: correlation of the concentrations (€/sample)
measured by the HPLC/UV method and the liquid chromatography/
tandem mass spectrometry (LC-MS/MS) method LLOQ lower limit of quantification

application in a routine TDM laboratory, we decided to use
deuterated internal standards. This required separation of
naltrexone-d3 and 6β-naltrexol owing to overlapping
isotope patterns (see earlier) and led to higher limits of
quantification, longer run times, and lower sensitivity. In
spite of this limitation, LC-MS/MS was advantageous.
Using deuterated substances led to markedly improved
method stability and accuracy. Analogue substances with
similar structure, such as naloxone, did not give satisfactory
results and were therefore not found to be suitable as
internal standards for LC-MS/MS of naltrexone and its
metabolite. Unexpectedly, the LC-MS/MS method achieved
better precision than the HPLC/UV method (Table 4).
Regarding actual knowledge on the use of naltrexone, it
seems likely that a limit of quantification of 2 ng/ml is
sufficient when analyzing plasma or serum in patients.
Lower concentrations can be attained by high-sensitivity
MS detectors [13]. They may be required for animal studies
or for the analysis of other body fluids such as saliva and
cerebrospinal fluid.
In comparison with existing HPLC methods [9, 10, 15],
our new HPLC/UV method was advantageous at least with
regard to two important aspects: (1) our solid-phase
extraction for sample cleanup and subsequent column
switching enabled automated analysis of serum or plasma,
whereas other methods reported so far include off-line
sample cleanup, e.g., by liquid–liquid extraction; (2) all
existing methods use electrochemical detection, which
indeed allows very low detection limits, but is susceptible
to contamination and requires long equilibration times. We
used UV detection, which is perhaps less sensitive with
regard to the limit of quantification but more robust than
electrochemical detection. UV detection, however, was
sufficiently sensitive for quantification of steady-state
concentrations of naltrexone and its metabolite in blood
plasma or serum of patients treated with therapeutic doses
of naltrexone.
Our LC-MS/MS method was also advantageous in com-
parison with existing LC-MS methods [15–18]. The method
of Iyer et al. [16] does not include determination of the active
metabolite. Other methods are disadvantageous, since they
require a precipitation–extraction step [15] or liquid–liquid
extraction [16, 18] before LC-MS or LC-MS/MS analysis.
Clavijo et al. [17] used online sample cleanup with column
switching and attained extremely low limits of quantification
using a very sensitive MS instrument of the latest generation.
However, determination of picogram quantities of naltrexone
and naltrexol may be required for pharmacokinetic studies
and research purposes, but it is not necessary for routine
TDM. We have shown that simultaneous quantification of
naltrexone and its metabolite using a standard triple-
quadrupole mass spectrometer without online extraction gives
a sufficiently low limit of quantification for TDM.
S. Brünen et al.
The general advantages of the LC-MS/MS method over the
HPLC/UV method are higher sensitivity and selectivity, better
precision, shorter chromatography time, and higher through-
put. The disadvantages are the inability to detect comedication
(owing to the mass filter, in contrast with the UV trace), the
need for expensive deuterated internal standards to achieve
sufficient accuracy and stability, and the generally more
expensive acquisition and servicing compared with the
requirements for use of the HPLC/UV method [26]. In our
comparison, both methods were found to be suitable for use
in a TDM laboratory to determine naltrexone plus 6β-
naltrexol. LC-MS/MS is advantageous when drug concen-
trations are low (Table 4). Its main shortcoming is the high
acquisition and maintenance cost. LC-MS/MS is not as yet
available in all clinical laboratories. Therefore, it was
important to show that HPLC with column switching and
UV detection also has the potential to be used for TDM of
naltrexone in routine clinical practice.
References
1. Ohara H, Miyabe Y, Deyashiki Y, Matsuura K, Hara A (1995)
Biochem Pharmacol 50:221–227
2. Malspeis L, Bathala MS, Ludden TM, Bhat HB, Frank SG,
Sokoloski TD, Morrison BE, Reuning RH (1975) Chem Pathol
Pharmacol 12:43–65
3. Cone EJ, Gorodetzky CW, Yeh SY (1974) Drug Metab Dispos
2:506–512
4. Blumberg H, Ikeda C (1976) Fed Proc 35:469
5. Dunbar JL, Turncliff RZ, Dong Q, Silverman BL, Ehrich EW,
Lasseter KC (2006) Alcohol Clin Exp Res 30:480–490
6. Wall ME, Brine DR, Perez-Reyes M (1981) Drug Metab Dispos
9:369–375
7. Meyer MC, Straughn AB, Lo MW, Schary WL, Whitney CC
(1984) J Clin Psychiatry 45:15–19
8. Derendorf H, El-Din A, El-Koussi A, Garrett ER (1984) J Pharm
Sci 73:621–624
9. O‘Connor EF, Cheng SWT, North WG (1989) J Chromatogr
Biomed Appl 491:240–247
10. Zuccaro P, Altieri I, Betto P, Pacifici R, Ricciarello G, Pini LA,
Sternieri E, Pichini S (1991) J Chromatogr 567:485–490
11. Davidson AF, Emm TA, Pieniaszek HJ Jr (1996) J Pharm Biomed
Anal 14:1717–1725
12. Vereby K, Kogan MJ, DePace A, Mulle SJ (1976) J Chromatogr
118:331
13. Reunig RH, Ashcraft SB, Morrison BE (1981) NIDA Res Monogr
28:25
14. Monti KM, Foltz RL, Chinn DM (1991) J Anal Toxicol 15:136–140
15. Valiveti S, Nalluri BN, Hammell DC, Paudel KS, Stinchcomb AL
(2004) J Chromatogr B 810:259–267
16. Iyer SS, Kellogg GE, Karnes HAT (2007) J Chromatogr Sci
45:694–700
17. Clavijo C, Bendrick-Peart J, Zhang YL, Johnson G, Gasparic A,
Christians U (2008) J Chromatogr B Anal Technol Biomed Life
874:33–41
18. Slawson MH, Chen M, Moody D, Comer SD, Nuwayser ES, Fang
WB, Foltz RL (2007) J Anal Toxicol 31:453–461
19. Baumann P, Hiemke C, Ulrich S, Eckermann G, Gaertner I,
Gerlach M, Kuss HJ, Laux G, Müller-Oerlinghausen B, Rao
Determination of naltrexone and 6β-naltrexol in blood: comparison of HPLC-UV with LC-MS/MS

ML, Riederer P, Zernig G (2004) Pharmacopsychiatry 37:243– 24. European Medicines Agency ICH Topic E 6 (R1): guideline for
265 good clinical practice; step 5 [Online]. 07.2002 [Last access 17-
20. Navaratnam V, Jamaludin A, Raman N, Mohamed M, Mansor SM 11-09]. URL: http://www.emea.europa.eu/pdfs/human/ich/
(1994) Drug Alcohol Depend 34:231–236 013595en.pdf
21. Schuh KJ, Walsh SL, Stitzer ML (1999) Psychopharmacology 25. World Medical Association Declaration of Helsinki (1997) JAMA
145:162–174 277:925–926
22. Dunbar JL, Turncliff RZ, Dong Q, Silverman BL, Ehrich EW, 26. Kirchherr H, Kühn-Velten WN (2006) J Chromatogr B Anal
Lasseter KC (2006) Alcohol Clin Exp Res 30:480–490 Technol Biomed Life Sci 843:100–113
23. Ferrari A, Bertolotti M, Dell’Utri A, Sterinieri E (1998) Drug 27. Kassenärztliche Bundesvereinigung (2009) Einheitlicher Bewer-
Alcohol Depend 52:211–220 tungsmaßstab. http://www.kbv.de/ebm2009/ebmgesamt.htm

View publication stats