Hydromorphone metabolites: Isolation and identification from pooled urine samples of a cancer patient

Article abstract of DOI:10.1080/00498250110119090

1. Hydromorphone-3-glucuronide, dihydromorphine, dihydroisomorphine, dihydromorphine-3-glucuronide and dihydroisomorphine-3-glucuronide were isolated from a cancer patient's urine and identified as metabolites of hydromorphone by comparison with synthetic standards using LC/MS/MS with gradient elution. 2. The relative urinary recovery of dihydroisomorphine-3-glucuronide was estimated to be 17-fold higher than previously reported. 3. Three new metabolites, including hydromorphone-3-sulphate, norhydromorphone and nordihydroisomorphine, were tentatively identified.

Full text of DOI:10.1080/00498250110119090

xenobiotica, 2002, vol. 32, no. 5, 427±439
Hydromorphone metabolites: isolation and
identi®cation from pooled urine samples of a
cancer patient
M. ZHENG , K. M. MCERLANE * and M. C. ONG
{}
{
{
{ Faculty of Pharmaceutical Sciences, University of British Columbia, 2146 East Mall,
Vancouver, BC V6T 1Z3, Canada
{ Chronic Pain Day Program, St Paul’s Hospital, Vancouver, BC V6Z 1Y6, Canada
Received 28 August 2001
1. Hydromorphone-3-glucuronide, dihydromorphine, dihydroisomorphine, dihydro-
morphine-3-glucuronide and dihydroisomorphine-3-glucuronide were isolated from a
cancer patient’s urine and identi®ed as metabolites of hydromorphone by comparison
with synthetic standards using LC/MS/MS with gradient elution.
2. The relative urinary recovery of dihydroisomorphine-3-glucuronide was estimated
to be 17-fold higher than previously reported.
3. Three new metabolites, including hydromorphone-3-sulphate, norhydromorphone
and nordihydroisomorphine, were tentatively identi®ed.
Introduction
Hydromorphone, or dihydromorphinone, is a semisynthetic potent opioid that
is about eight times more potent than morphine (Mahler and Forrest 1975). Since
its ®rst clinical use in 1926 (Eddy 1993), hydromorphone has been commonly used
to relieve both cancer pain (Bruera et al. 1988, Ferris et al. 1990, Moulin et al.
1991, Vanier et al. 1993, Hays et al. 1994) and postoperative pain (Dougherty et al.
1989, Brodsky et al. 1990, Wallenstein et al. 1990, Searle et al. 1994, Liu et al.
1995). Hydromorphone provides a useful alternative to morphine in situations
where adverse e€ects, such as nausea, sedation and myoclonus, remain signi®cant
problems (Librach 1991). In addition, the high water solubility of hydromorphone
hydrochloride (®ve times more soluble than morphine sulphate (Budavari 1989) is
an important advantage over morphine when concentrated preparations or small
volume delivery, such as subcutaneous infusion, are required. Subcutaneous
infusion of hydromorphone has been widely accepted for the management of
chronic cancer pain (Ferris et al. 1990, Moulin et al. 1991, Bruera et al. 1993,
Vanier et al. 1993).
Structurally, hydromorphone di€ers from morphine in two respects, namely
the saturation of the 7,8-double bond and the replacement of the 6-hydroxy group
by a 6-keto group. Although hydromorphone is commonly prescribed for pain
management, information on its metabolic fate is scant. Roerig et al. (1973)
reported the isolation of hydromorphone-3-glucuronide and dihydromorphine-
3-glucuronide from rabbit urine after hydromorphone administration. A more
* Author for correspondence; e-mail: tuit@interchange.ubc.ca.
} Present address: Metabolism and Pharmacokinetics, Pharmaceutical Research Institute, Bristol-
Myers Squibb Co., Wallingford, CT06067, USA
#
Xenobiotica ISSN 0049±8254 print/ISSN 1366±5928 online
2002 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
DOI: 10.1080/00498250110119090

428
M. Zheng et al.
extensive study by Cone et al. (1977) in several species, including humans,
indicated that hydromorphone was metabolized to conjugated hydromorphone,
dihydromorphine (6a-hydroxy hydromorphone) and dihydroisomorphine (6b-
hydroxy hydromorphone). The precise identities of these conjugates, however,
were not characterized. The overall urinary recoveries of these metabolites plus
hydromorphone in humans only accounted for 43.5% of the administered dose.
Dihydromorphine and dihydroisomorphine have been reported to exhibit 65 and
21% of the analgesic activity of hydromorphone, respectively, as determined by
mechanical clip test (Small et al. 1938).
Morphine undergoes extensive metabolism and forms morphine-3-glucuro-
nide, morphine-6-glucuronide, morphine-3-sulphate, normorphine and some
other minor metabolites (Boerner et al. 1975). Both morphine-6-glucuronide
and normorphine possess analgesic activity (Hanna et al. 1990, Gong et al. 1991,
Osborne et al. 1992, Zaw-Tun and Bruera 1992, Thompson et al. 1995). In view of
the structural similarities between hydromorphone and morphine and the unac-
counted mass balance, the existence of additional unidenti®ed metabolites of
hydromorphone was considered a possibility. Thus, the objectives of the present
study were to isolate and identify the above known metabolites as well as to search
for the presence of new metabolites in cancer patients’ urine using HPLC/mass
spectrometry/mass spectrometry (LC/MS/MS).
Materials and methods
Materials
Hydromorphone hydrochloride was a generous gift from Knoll Pharma, Inc. (Mississauga, ON,
Canada). Morphine sulphate¢5H2O was obtained from BDH, Inc. (Toronto, ON, Canada). Amberlite
XAD-2 resin was obtained from Supelco (Oakvill, ON, Canada). Palladium on activated carbon (Pd
10%) was obtained from Aldrich (Milwaukee, WI, USA). Formamidinesulphinic acid and 2,3,4-tri-O-
acetyl-1¬-bromo-1-deoxy-d-glucopyranuronyl methyl ester were obtained from Sigma Chemical Co.
(St Louis, MO, USA). Lithium hydroxide¢H2O was obtained from Anachemia Science (Montreal, PQ,
Canada). Tri¯uoroacetic acid (TFA) was obtained from Fisher (Fair Lawn, NJ, USA). D6-DMSO was
obtained from Matheson (Miamisburg, OH, USA). Other reagents were of analytical grade and the
solvents were of HPLC grade. Puri®ed water was produced using a Milli-Q water puri®cation system
(Mississauga, ON, Canada).
Isolation of hydromorphone and its metabolites from patient urine
The urine sample was obtained from a female patient with multiple myeloma and who had been
under chronic hydromorphone monotherapy for analgesia for 2 weeks (subcutaneous infusion, 4 mg
every 3 h) in St Paul’s hospital, Vancouver, British Columbia. The urine was collected over 48 h. An
aliquot of 1300 ml urine was applied to an Amberlite XAD-2 resin column …4 £ 18 cm†. The column was
washed with 180 ml water and the analytes were subsequently eluted with 240 ml methanol. The
methanol eluent was concentrated under vacuum. The residue was reconstituted in 13ml water and the
pH was adjusted to 9.5±10 with 5% ammonium hydroxide. The aqueous solution was extracted with
25 ml ethyl acetate (twice). The organic and aqueous phases were separated and both were concentrated
under vacuum. Before the LC/MS/MS analysis, the residues from the aqueous phase were dissolved in
1 ml water, whereas the residues from the organic phase were dissolved in 700 ml methanol. An aliquot of
100 ml was taken from each sample and both were further diluted 20 times. An aliquot of 20 ml of the
®nal solution was analysed by LC/MS/MS.
LC/MS/MS system
The LC/MS/MS system consisted of a HP Series II 1090 liquid chromatograph (Hewlett Packard,
Avondale, PA, USA), a Zorbax SB-phenyl column (4.6 mm £ 25cm) (Rockland Technologies, Inc.,
Newport, DE, USA) and a VG Quattro triple quadrupole mass spectrometer equipped with an
electrospray interface (Fisons, Altrincham, UK).

Metabolism of hydromorphone in man
429
The separation of hydromorphone and its metabolites was achieved by HPLC using gradient
elution. Mobile phase A consisted of 0.05% TFA in 10% methanol; mobile phase B consisted of 0.05%
TFA in 90% methanol. The initial elution condition was 10% of B maintained for 6 min and increased to
30% of B over the next 8 min and 80% of B over a further 16 min. The ¯ow rate was 0.7ml min¡1. The
LC e‚uent was split and 10% of the e‚uent was introduced into MS.
The electrospray source was operated with a capillary voltage of 3.4kV, a cone voltage of 29V and a
source temperature of 100 8C. The MS/MS analysis was performed in the multiple reaction monitoring
(MRM) mode. The ®rst MS (MS1) was static at a preselected parent ion masses. The selected parent
ions were introduced into the collision cell and the fragmentation occurred at a collision energy of 70 eV.
Argon was used as a collision gas at a pressure of 3 £ 10¡3 mbar. The second MS (MS2) was static at
pre-determined daughter ion masses produced from the fragmentation in the collision cell. A dwell time
of 0.5s was used for monitoring.
Synthesis of dihydromorphine
Dihydromorphine was synthesized by catalytic hydrogenation of morphine following the method
reported by Small and Lutz (1932). Accordingly, 40 mg 10% palladium on activated carbon was added
to 3.3 g (8 mmol morphine base) of morphine sulphate in 85 ml water. This solution was shaken with
hydrogen at 44 psi for 45 min. After ®ltration, the pH of the reaction mixture was adjusted to 9.8 by the
addition of 1.5 m Na2CO3 solution and dihydromorphine base separated as needles. The product was
recrystallized from methanol with a yield of 2.3 g (92%), m.p. 150±155 8C with decomposition at 155±
157 8C (lit. m.p. 155±157 8C with decomposition (Small and Lutz 1932). The 1H-NMR spectrum in d6-
DMSO exhibited signals at: ¯2:22 (s, 3H, NCH3; 3.80 (m, 1H, 6b-H); 4.44 (d, J ˆ 5 Hz, 1H, 5b-H);
6.40, 6.53 (d, J ˆ 8Hz, 2H, 1-H, 2-H), 8.73 (s, 1H, 3-OH).
The hydrochloride salt was obtained by dissolving the base in methanol and treating with ethanolic
HCl solution. LC/MS: ‰M ‡ 1Š‡ 288 (100%); IR (Nujol mull): 3390cm¡1 (free OH), 3140 cm¡1 (bonded
OH), 2687 cm¡1 (tertiary amine salt); UV ¶max (water): 284 nm; UV ¶max (0.1 N NaOH): 296 nm. Anal.
calcd. for C17H21NO3¢HCl: C, 63.05; H, 6.86; N, 4.33. Found: C, 62.95; H, 7.05; N, 4.31.
Synthesis of dihydroisomorphine
The synthesis of dihydroisomorphine was accomplished by adaptation of a method for stereospeci®c
synthesis of 6-b-hydroxy metabolites of naltrexone and naloxone (Chatterjie et al. 1975). Water (20 ml)
was added to dissolve 1.3g (4mmol) hydromorphone hydrochloride. To the above solution was added,
drop wise with magnetic stirring, 1.7 g (16 mmol) acid dissolved in 170 ml 0.2 N NaOH. The reaction
mixture was heated on a water bath at 80±858C with magnetic stirring for 3.5h. The mixture was
allowed to cool and the pH was adjusted to 9.8 with sodium hydroxide solution. The solution was
extracted with 150 ml ethyl acetate four times. The base was obtained by evaporation of the organic
extract to dryness with a rotary evaporator. The product was recrystallized from methanol with a yield
of 0.3g (26%), m.p. 225±227 8C with decomposition at 2278C. The 1H-NMR spectrum in d6-DMSO
exhibited signals at: ¯2:29 (s, 3H, NCH3); 3.10 (m, 1H, 6¬-H); 4.12 (d, J ˆ 6 Hz, 1H, 5b-H); 6.45, 6.54
(d, J ˆ 8 Hz, 2H, 1-H, 2-H), 8.95 (s, 1H, 3-OH).
The hydrochloride salt was obtained by dissolving the base in methanol and treating with ethanolic
HCl solution. LC/MS: ‰M ‡ 1Š‡ 288 (100%). IR (Nujol mull): 3363cm¡1 (free OH), 3127 cm¡1 (bonded
OH), 2644 cm¡1 (tertiary amine salt); UV ¶max (water): 282 nm; UV ¶max (0.1 N NaOH): 296 nm. Anal.
calcd. for C17H21NO3¢HCl: C, 63.05; H, 6.86; N, 4.33. Found: C, 63.19; H, 6.83; N, 4.31.
Synthesis of hydromorphone-3-glucuronide
Hydromorphone-3-glucuronide was prepared by adaptation of a method described by Berrang et al.
(1975) for the preparation of morphine-3-glucuronide. To a solution of hydromorphone base (2.3g,
8 mmol) and lithium hydroxide (0.3g) in methanol (13 ml) was added 2.3 g (6 mmol) of 2,3,4-tri-O-
acetyl-1¬-bromo-1-deoxy-d-glucopyranuronic methyl ester as a solid. The mixture was stirred and
allowed to stand at room temperature for 4h. A solution of lithium hydroxide (0.68g) in water (13 ml)
was then added and the mixture stirred for an additional 1 h. The pH of the mixture was reduced to
about 9.5 with acetic acid. Unreacted hydromorphone (1.2 g) was recovered by extraction with 4 vol
ethyl acetate. The remaining reaction mixture was concentrated to a viscous liquid under vacuum and
taken up in 29 ml methanol. The solution was treated with ethanolic hydrochloric acid solution to pH 7±
8 and crystallization occurred at ¡10 8C overnight. The crystals were ®ltered and recrystallized from
methanol/water (50/50v/v). The yield was 0.54 g (15%); m.p. 225±227 8C with decomposition. The 1H-
NMR spectrum in d6-DMSO exhibited signals at: d2:40 (s, 3H, ˆN‡HCH3), 4.92 (s, 1H, 5b-H), 5.23
(d, J ˆ 7 Hz, 1H, 10-H), 6.65, 6.82 (d, J ˆ 9Hz, 2H, 1-H, 2-H). LC/MS: ‰M ‡ 1Š‡ 462 (100%); IR
(Nujol mull): 3650±3110 cm¡1 (broad, bonded and free OH), 2670cm¡1 (tertiary ammonium ion),
1724cm¡1 …¸C
Ð
Ð
O†, 1590cm¡1 (7COO¡); UV ¶max (water): 278 nm; UV ¶max (0.1 N NaOH), 278 nm.
Anal. calcd. for C23H27NO9¢3H2O: C, 53.59; H, 6.26; N, 2.72. Found: C, 53.61; H, 6.31; N, 2.73.

430
M. Zheng et al.
Synthesis of dihydromorphine-3-glucuronide
Dihyromorphine-3-glucuronide was prepared following the same procedures as for the synthesis of
hydromorphone-3-glucuronide. To a solution of dihydromorphine hydrochloride (1.0 g, 3.1 mmol) and
lithium hydroxide (0.25g) in methanol (5 ml) was added 1.0 g (2.5mmol) of 2,3,4-tri-O-acetyl-1¬-
bromo-l-deoxy-d-glucopyranuronic methyl ester as a solid. After recrystallization, a yield of 0.36 g
(25%) was obtained; m.p. 230±2338C with decomposition. The 1H-NMR spectrum in d6-DMSO
exhibited signals at: d2:56 (s, 3H, ˆN‡NCH3), 3.86 (m, 1H, 6b-H), 4.53 (d, J ˆ 5 Hz, 1H, 5b-H), 5.19
(d, J ˆ 7 Hz, 1H, 10-H), 6.59, 6.70 (d, J ˆ 9Hz, 2H, 1-H, 2-H). LC/MS: ‰M ‡ 1Š‡ 464 (100%); IR
(Nujol mill): 3650±3110 cm¡1 (broad, bonded and free OH), 2660 cm¡1 (tertiary ammonium ion),
1597cm¡1 (7COO¡); UV ¶max (water): 280 nm; UV ¶max (0.1 N NaOH): 280 nm. Anal. calcd. for
C23H29NO9¢2.5H2O: C, 54.32; H, 6.74; N, 2.75. Found: C, 54.33; H, 6.58; N, 2.76.
Synthesis of dihydroisomorphine-3-glucuronide
Dihyroisomorphine-3-glucuronide was prepared following the same procedures as for the synthesis
of hydromorphone-3-glucuronide. To a solution of dihydroisomorphine base (0.4 g, 1.4 mmol) and
lithium hydroxide (0.056g) in methanol (3ml) was added 0.4 g (1 mmol) of 2,3,4-tri-O-acetyl-1¬-
bromo-l-deoxy-d-glucopyranuronic methyl ester as a solid. After recrystallization, a yield of 0.27 g
(42%) was obtained; m.p. 245±2488C with decomposition. The 1H-NMR spectrum in d6-DMSO
exhibited signals at: d2:55 (s, 3H, ˆN‡HCH3), 4.23 (d, J ˆ 7 Hz, 1H, 5b-H), 5.30 (d, J ˆ 7 Hz, 1H, 10-
H), 6.64, 6.80 (d, J ˆ 9 Hz, 2H, 1-H, 2-H). LC/MS: ‰M ‡ 1Š‡ 464 (100%); IR (Nujol mull): 3650±
3110cm¡1 (broad, bonded and free OH), 2650 cm¡1 (tertiary ammonium ion), 1585 cm¡1 (7COO¡);
UV ¶max (water): 278 nm; UV ¶max (0.1 N NaOH): 278nm. Anal. calcd. for C23H29NO9¢2H2O: C,
55.30; H, 6.66; N, 2.80. Found: C, 55.28; H, 6.53; N, 2.79.
Results
Synthesis
Dihydromorphine, dihydroisomorphine and the 3-O-glucuronides of hydro-
morphone, dihydromorphine and dihydroisomorphine were synthesized by adap-
tation of methods for the preparation of structurally similar compounds.
Dihydroisomorphine was synthesized according to the method of stereospeci®c
synthesis of the 6b-hydroxy derivatives of naltrexone and naloxone (Chatterjie et
al. 1975). The 6b-hydroxy derivatives of naltrexone and naloxone were synthe-
sized with no indication of the presence of 6a isomers. In our case, however, we
observed the presence of about 10% of the 6a isomer (dihydromorphine). Dihy-
droisomorphine was further puri®ed by recrystallization. The orientation of the
6b-hydroxy group was con®rmed by the observed up®eld shifts in NMR due to
5b-H and 6a-H, as compared with the values due to 5b-H and 6b-H on
dihydromorphine (Okuda et al. 1964).
The 3-O-glucuronide conjugates were prepared, with modi®cations, following
the method of Berrang et al. (1975) for the synthesis of morphine-3-glucuronide.
This method simpli®ed the procedure of Yoshimura et al. (1968). The crystal-
lization of these glucuronides from methanol was facilitated by adjusting the pH to
7±8 with ethanolic hydrochloric acid solution. No precipitation was observed after
1.2 mg of each of the glucuronides was added to 1.5 ml 0.1 N AgNO3 solution,
which indicated that hydrochloride salt was not formed for these glucuronides. In
contrast, 0.5 mg hydromorphone hydrochloride showed a strong positive result.
The b-con®guration of the glycosidic linkage in the glucuronides was sup-
ported by the NMR results that showed a resonance signal at d5.19±5.30 due to the
axial 10-H on the glucuronic acid moiety, whereas a signal would have appeared at
around d6 31 due to an equatorial 10-H if it were the a-isomer (Yoshimura et al.
:
1968). An absence of a bathochromic shift in UV
in the alkaline medium,
which would occur in compounds possessing a free phenolic group as seen with
¶
max

Metabolism of hydromorphone in man
431
dihydromorphine and dihydroisomorphine, further supported the formation of the
3-O-glucuronide conjugate. Furthermore, NMR results also con®rmed the for-
mation of the 3-O-glucuronide conjugate because a singlet due to the 3-hydroxy
proton, which appeared at d8 73 and 8.95 for dihydromorphine and dihydroiso-
:
morphine, respectively, was not seen with hydromorphone-3-glucuronide. The IR
results showed a strong peak at 1585±1597 cm¡1 and a broad absorption around
2650±2670 cm¡1, suggesting the existence of a carboxylate anion and a positive
charge on the piperidine nitrogen.
Identi®cation of hydromorphone and its metabolites in cancer patient urine by
LC/MS/MS
The separation of hydromorphone and the ®ve synthetic standards was
achieved by a gradient elution followed by MS/MS detection performed in the
MRM mode, which is a highly speci®c technique for rapid screening of biological
sample for known or new metabolites. Figure 1 shows the mass chromatograms of
the six standards and table 1 presents the mass/charge ratios …m z† and the
=
retention times of the parent and daughter ions. The daughter ions of the three
glucuronide conjugates were produced by the common loss of the glucuronic acid
moiety (m z 176), whereas the three free compounds produced a common
=
=
daughter ion of m z 185. The relative elution orders of dihydromorphine-3-
glucuronide to dihydroisomorphine-3-glucuronide, and dihydromorphine to di-
hydroisomorphine had been previously determined on a HPLC-UV system with
the same gradient elution program.
Figure 1. MRM traces of standards: dihydromorphine-3-glucuronide, 5.73 min; hydromorphone-3-
glucuronide, 6.87 min; dihydroisomorphine-3-glucuronide, 7.49min; dihydromorphine,
9.15min; hydromorphone, 12.88 min; dihydroisomorphine, 13.71 min.

432
M. Zheng et al.
Table 1. LC/MS/MS characteristics of hydromorphone and its metabolites.
Retention times (min)
Parent …M ‡ 1†‡
…m=z†
Daughter
…m=z†
Compound
Standard
Urine
Dihydromorphine-3-glucuronide
Hydromorphone-3-glucuronide
Dihydroisomorphine-3-glucuronide
Dihydromorphine
Hydromorphone
Dihydroisomorphine
464
462
464
288
286
288
288 …M ‡ 1†‡
296 …M ‡ 1†‡
288 …M ‡ 1†‡
185
5.73
6.87
7.49
9.15
12.88
13.71
5.62
6.66
7.39
9.15
12.78
13.71
185
185
Figures 2 and 3 show the mass chromatograms of metabolites isolated from
patient urine samples dissolved in methanol and water, respectively. The presence
of all of the ®ve metabolites plus hydromorphone were con®rmed by comparing
the retention times with those of the standards (table 1). Owing to the di€erences
in hydrophilicity of the compounds, the three glucuronide conjugates are present
at much higher levels in the aqueous sample than in the methanolic sample,
whereas the unconjugated compounds are present in much higher concentrations
in the methanolic sample than in the aqueous sample.
To obtain an estimation of the relative levels of the six compounds in the above
urine samples, a standard solution was prepared with known amounts of the six
standards added (table 2). This standard solution was analysed under the same
chromatographic conditions as used for the urine sample, but without prior
extraction. The peak areas for each compound were recorded. Only compounds
with similar physicochemical properties were compared because the glucuronide
Figure 2. MRM traces of the urinary extract dissolved in methanol: hydromorphone-3-glucuronide,
6.87 min; dihydroisomorphine-3-glucuronide, 7.49 min; dihydromorphine, 9.15 min, hydro-
morphone, 12.78 min; dihydroisomorphine, 13.71 min.

Metabolism of hydromorphone in man
433
Figure 3. MRM traces of the urinary extract dissolved in water: dihydromorphine-3-glucuronide,
5.62 min; hydromorphone-3-glucuronide, 6.66 min (B); dihydroisomorphine-3-glucuronide,
7.39 min; hydromorphone-3-sulphate, 8.53 min (C); dihydromorphine, 9.36 min; hydromor-
phone, 13.09 min; dihydroisomorphine, 13.71 min. Note that the peak at 8.53 min that appeared
in E was apparently due to the fragmentation of hydromorphone-3-sulphate in the MS1.
conjugates and the unconjugated compounds might have considerably di€erent
recoveries during the isolation with Amberlite XAD-2 column, which could lead
to an invalid comparison. As shown in table 2, when corrected for the same
concentration, the peak area ratios between dihydromorphine/hydromorphone/
dihydroisomorphine standards are about 2.3/1.0/2.4. Thus, the relative levels of
these three compounds in urine, when corrected for this response factor, are about
1/18/3, respectively. Likewise, the relative levels of the conjugated dihydromor-
phine/hydromorphone/dihydroisomorphine in urine are about 1/82/37, respect-
ively. As the reported urinary recovery of hydromorphone-3-glucuronide
Table 2. Peak areas of hydromorphone and its metabolites as determined by MRM.
Peak areas
Concentration of
standard (mg ml¡1
Compound
Dihydromorphine
Hydromorphone
)
Standard
Urine
5.0
5.0
2.5
57 794
24 751
30 136
18 305
142499
58 050
Dihydroisomorphine
(60 272){
10 908
Dihydromorphine-3-glucuronide
2.9
2 688
(18 807)
14 525
13 947
Hydromorphone-3-glucuronide
Dihydroisomorphine-3-glucuronide
5.0
3.3
169569
111656
(21 132)
{ Values in parentheses are peak areas corrected for 5mgml¡1
.

434
M. Zheng et al.
accounted for 37% of the administered dose (Cone et al. 1977), the urinary
recovery of dihydroisomorphine-3-glucuronide in the current study would ac-
count for 17% of the given dose. This is much higher than the result of Cone et al.
who reported that the urinary recovery of total dihydroisomorphine, present
almost exclusively as conjugate, only accounted for 1.0% of the given dose.
Using the same calculation, the urinary recovery of dihydromorphine-3-glucur-
onide was about 0.45%, which is in general agreement with the urinary recovery of
0.1% reported by Cone et al. when considering the possible analytical errors
involved.
Tentative identi®cation of hydromorphone-3-sulphate, norhydromorphone and
nordihydroisomorphine
Because of the structural similarities between hydromorphone and morphine,
the same LC/MS/MS conditions were used for screening additional previously
unidenti®ed metabolites of hydromorphone. The aqueous and the methanolic
samples were screened for conjugated and unconjugated metabolites, respectively.
The parent ion and daughter ion masses were determined for the proposed
structures of the new metabolites. The potential metabolites screened are sum-
marized in table 3.
The presence of previously unidenti®ed hydromorphone-3-sulphate is shown
in ®gure 3C. A signi®cant amount of hydromorphone-3-sulphate underwent
fragmentation to hydromorphone in MS1 before entering the collision cell as
shown by the presence of a peak at a retention time of 8.53 min (®gure 3E), which
coincided with the retention time of hydromorphone-3-sulphate. The presence of
other potential conjugates could not be con®rmed (®gure 4).
The presence of unconjugated nordihydroisomorphine and norhydromorphone
is shown in ®gure 5C, D, respectively. The peak at 11.77 min in ®gure 5C is
proposed to be nordihydroisomorphine with reference to the relative elution order
between dihydroisomorphine and hydromorphone (®gure 5A, B). Both nordihy-
Table 3. LC/MS/MS characteristics of the potential metabolites of hydromorphone.
Parent (M ‡ 1)+
(m/z)
Daughter
(m/z)
Retention times (min)
for urine
Compound screened
Hydromorphone-3-sulphate
366
448
286 (M ‡ )+
8.53
n.d.
Norhydromorphone-3-glucuronide
272 (M‡ 1)+
(norhydromorphone)
274 (M‡ 1)+
Nordihydromorphine- or
nordihydroisomorphine-3-
glucuronide
450
640
n.d.
n.d.
(nordihydromorphine or
nordihydroisomorphine)
Dihydromorphine- or
dihydroisomorphine-di-
glucuronide
464 (M‡ 1)+
and
288 (M‡ 1)+
Nordihydromorphine
Nordihydroisomorphine
Norhydromorphone
274
274
272
185
185
185
n.d.
11.77
11.49
n.d., not detected.

Metabolism of hydromorphone in man
435
Figure 4. Screening for additional urinary conjugates in the urinary extract dissolved in water.
Channel A was for monitoring the possible fragmentation of dihydromorphine- or dihydroiso-
morphine-di-glucuronide to dihydromorphine or dihydroisomorphine glucuronide; channel B
for dihydromorphine- or dihydroisomorphine-di-glucuronide to dihydromorphine or dihydro-
isomorphine; dihydroisomorphine-3-glucuronide, 7.49min (C); hydromorphone-3-glucuronide,
6.66min (D); channel E for nordihydromorphine glucuronide or nordihydroisomorphine
glucuronide to nordihydromorphine or nordihydroisomorphine; channel F for norhydromor-
phone glucuronide to norhydromorphone.
droisomorphine and norhydromorphone appear to be present at much lower levels
than the other three metabolities (®gure 5A, B).
Discussion
Although hydromorphone is being increasingly used for pain management, the
metabolic pro®le of this opioid analgesic is poorly understood. Hydromorphone-3-
glucuronide and dihydromorphine-3-glucuronide have been identi®ed as metabo-
lites in the rabbit (Roerig et al. 1973). A more extensive study by Cone et al. (1977)
examined the di€erences in the metabolism of hydromorphone between di€erent
species. In particular, the urinary recoveries of dihydromorphine and dihydroiso-
morphine in guinea pig, mostly present in unconjugated forms, accounted for 8.3
and 19.3% of the administered dose, respectively, which were much higher than
the level in rat, dog, rabbit and human. In human, the urinary recoveries of
dihydromorphine and dihydroisomorphine were only 0.1 and 1.0%, respectively,
with only the conjugated forms detected. The exact identities of these conjugates,
however, were not determined. In chicken, opioids with a hydromorphone
structure, such as naltrexone and naloxone, were preferentially metabolized to
6 -hydroxy metabolites (Roerig et al. 1976).
¬

436
M. Zheng et al.
Figure 5. Screening for additional unconjugated metabolites in the urinary extract dissolved in
methanol: Dihydromorphine, 9.09min; hydromorphone, 12.52 min; dihydroisomorphine,
13.55 min; nordihydroisomorphine, 11.77 min; norhydromorphone, 11.49 min.
In the present study, both free and conjugated dihydromorphine and dihy-
droisomorphine were identi®ed in the patients’ urine. More speci®cally, the
conjugated metabolites were identi®ed as dihydromorphine-3-glucuronide and
dihydroisomorphine-3-glucuronide. A surprising ®nding of the study was that the
urinary recovery of dihydroisomorphine-3-glucuronide was about 17-fold higher
than the result reported by Cone et al. (1977). A possible explanation is the
interindividual variations in the metabolism of hydromorphone. The subjects
studied by Cone et al. were healthy, adult male prisoners. In the present study, the
subject was a female patient su€ering from multiple myeloma. Large variations in
hepatic metabolism of hydromorphone have been reported in normal healthy
volunteers (Parab et al. 1988). The particular physiological conditions of our
patient could cause even larger deviation from the already highly variable
metabolism of hydromorphone among the healthy population. Since the objective
of the present study was the identi®cation of the metabolites rather than quantita-
tion, a more accurate quantitation assay and a larger patient number are required
to con®rm the above di€erences between the two studies.
Hydromorphone-3-glucuronide and unchanged hydromorphone were also
identi®ed in the patient’s urine. The overall urinary recovery of hydromorphone,
dihydromorphine and dihydroisomorphine (free plus conjugated) in humans was
about 43.5% as determined by Cone et al. (1977). According to the results of the
present study, although dihydroisomorphine was present at a much higher level,
there is still a lack of mass balance between the administered dose and the urinary
recovery. Alternate excretion routes such as biliary excretion and/or the presence
of additional metabolites are possible explanations (Cone and Darwin 1978).

Metabolism of hydromorphone in man
437
Figure 6. Chemical structures of hydromorphone and its metabolites detected in this study.
* Metabolites tentatively identi®ed but not con®rmed due to lack of standards.
Because of the structural similarities between hydromorphone and morphine, we
explored the presence of previous unidenti®ed metabolites. As expected, hydro-
morphone-3-sulphate, norhydromorphone and nordihydroisomorphine were ten-
tatively identi®ed as new metabolites of hydromorphone by LC/MS/MS. The
retention times of these new metabolites were in general agreement relative to the
other identi®ed metabolites.
In analogy to morphine-6-glucuronide, the existence of hydromorphone-6-
glucuronide has always been speculated (Lee et al. 1991, Hagen et al. 1995). The 6-
glucuronides of dihydromorphine and dihydroisomorphine have been referred as
hydromorphone-6-glucuronide based on the work of Cone et al. (1977). As we
understand, however, Cone et al. did not discuss the presence of hydromorphone-
6-glucuronide and, furthermore, they had never made the above comparison. As a
matter of fact, as already mentioned, the exact identities of the dihydromorphine
and dihydroisomorphine conjugates were not characterized by Cone et al.
It is believed that, structurally, hydromorphone-6-glucuronide is di€erent
from the 6-glucuronides of dihydromorphine and dihydroisomorphine. Hydro-
morphone-6-glucuronide is formed via the conversion of the 6-keto group on
hydromorphone to its enol form followed by further glucuronidation, whereas the
6-glucuronides of dihydromorphine and dihydroisomorphine are simply formed
through the glucuronidation of the 6-hydroxy group on dihydromorphine and
dihydroisomorphine. There has been no report on glucuronidation via a keto
group. In addition, neither hydromorphone-6-glucuronide nor the 6-glucuronides
of dihydromorphine and dihydroisomorphine were detected in the present study.

438
M. Zheng et al.
Thus, unlike morphine, the formation of hydromorphone-6-glucuronide is prob-
ably not favoured for hydromorphone.
In conclusion, we have identi®ed an unchanged hydromorphone, hydromor-
phone-3-glucuronide, and both free and conjugated dihydromorphine and dihy-
droisomorphine in the cancer patient’s urine using LC/MS/MS. More speci®cally,
the conjugated dihydromorphine and dihydroisomorphine were identi®ed as
dihydromorphine-3-glucuronide and dihydroisomorphine-3-glucuronide. The
above metabolites were con®rmed by comparison of the retention times with
those of the synthetic standards using LC/MS/MS. In addition, we have for the
®rst time tentatively identi®ed three new metabolites of hydromorphone including
hydromorphone-3-sulphate, norhydromorphone and nordihydroisomorphine
(®gure 6). Some of the disagreements between the present study and previous
work need to be investigated further.
Acknowledgements
The work was partially supported by Knoll Pharma, Inc. (Canada). The
authors thank Dr Frank Abbott, Mr Wei Tang and Mr Roland Burton for
technical assistance in the LC/MS/MS experiments.
References
Berrang, B., Twine, C. E., Hennessee, G. L. and Carroll, F. I., 1975, Synthesis of morphine-3-
glucuronide. Synthetic Communications, 5, 231±236.
Boerner, U., Abbott, S. and Roe, R. L., 1975, The metabolism of morphine and heroin in man. Drug
Metabolism Reviews, 4, 39±73.
Brodsky, J. B., Chaplan, S. R., Brose, W. G. and Mark, J. B. D., 1990, Continuous epidural
hydromorphone for postthoracotomy pain relief. Annals of Thoracic Surgery, 50, 888±893.
Bruera, E., Brenneis, C., Michaud, M., Macmillan, K., Hanson, J. and Macdonald, R. N., 1988,
Patient-controlled subcutaneous hydromorphone versus continuous subcutaneous infusion for
the treatment of cancer pain. Journal of the National Cancer Institute, 80, 1152±1154.
Bruera, E., Maceachern, T., Macmillan, K., Miller, M. J. and Hanson, J., 1993, Local tolerance to
subcutaneous infusions of high concentrations of hydromorphone: a prospective study. Journal
of Pain Symptoms and Management, 8, 201±204.
Budavari, S. (ed.), Merck Index, 1989, 11th edn (Rahway: Merck).
Chatterjie, N., Inturrisi, C. E., Dayton, H. B. and Blumberg, H., 1975, Stereospeci®c synthesis of
the 6- -hydroxy metabolites of naltrexone and naloxone. Journal of Medicinal Chemistry, 18, 490±
492.
Cone, E. J. and Darwin, W. D., 1978, Simultaneous determination of hydromorphone, hydrocodone
and their 6¬- and 6- -hydroxy metabolites in urine using selected ion recording with methane
chemical ionization. Biomedical Mass Spectrometry, 5, 291±295.
Cone, E. J., Phelps, B. A. and Gorodetzky, C. W., 1977, Urinary extraction of hydromorphone and
metabolites in humans, rats, dogs, guinea pigs, and rabbits. Journal of Pharmaceutical Sciences,
66, 1709±1713.
Dougherty, T. B., Baysinger, C. L., Henenberger, J. C. and Gooding, D. J., 1989, Epidural
hydromorphone with and without epinephrine for post-operative analgesia after cesarean
delivery. Anesthesia and Analgesia, 68, 318±322.
Eddy, N. B., 1933, Dilaudid (dihydromorphinone hydrochloride). Journal of the American Medical
Association, 100, 1032±1035.
Ferris, F. D., Kerr, I. G., De Angelis, C., Sone, M and Hume, S., 1990, Inpatient narcotic infusions
for patients with cancer pain. Journal of Palliative Care, 6, 51±59.
Gong, Q.-L., Hedner, T., Hedner, J., BjOrkman, R. and Nordberg, G., 1991, Antinociceptive and
ventilatory e€ects of the morphine metabolites: morphine-6-glucuronide and morphine-3-
glucuronide. European Journal of Pharmacology, 193, 47±56.

Metabolism of hydromorphone in man
439
Hagen, N., Thirlwell, M. P., Dhaliwal, H. S., Babul, N., Harsanyi, Z. and Darke, A. C., 1995,
Steady-state pharmacokinetics of hydromorphone and hydromorphone-3-glucuronide in cancer
patients after immediate and controlled-release hydromorphone. Journal of Clinical
Pharmacology, 35, 37±44.
Hanna, M. H., Peat, S. J., Woodham, M., Knibb, A. and Fung, C., 1990, Analgesic ecacy and CSF
pharmacokinetics of intrathecal morphine-6-glucuronide: comparison with morphone. British
Journal of Anaesthesia, 64, 547±550.
Hays, H., Hagen, N., Thirwell, M., Dhaliwal, H., Babul, N., Harsanyi, Z. and Darke, A. C., 1994,
Comparative clinical ecacy and safety of immediate release and controlled release
hydromorphone for chronic severe cancer pain. Cancer, 74, 1808±1816.
Lee, J. W., Pedersen, J. E., Moravetz, T. L., Dzerk, A. M., Mundt, A. D. and Shepard, K. V., 1991,
Sensitive and speci®c radioimmunoassays for opiates using commercially available materials. I:
Methods for the determinations of morphine and hydromorphone. Journal of Pharmaceutical
Sciences, 80, 284±288.
Librach, S. L., 1991, Adverse e€ects of opioids. In The Pain Manual: Principles and Issues in Cancer
Pain Management (Montreal: Pegasus Healthcare), pp. 58±66.
Liu, S., Carpenter, R. L., Mulroy, M. F., Weissman, R. M., McGill, T. J., Rupp, S. M. and Allen,
H. M., 1995, Intravenous versus epidural administration of hydromorphone. Anesthesiology, 82,
682±688.
Mahler, D. L. and Forrest, W. H., Jr., 1975, Relative analgesic potencies of morphine and
hydromorphone in postoperative pain. Anesthesiology, 42, 602±607.
Moulin, D. E., Kreeft, J. H., Murray-Parsons, N. and Bouquillon, A. I., 1991, Comparison of
continuous subcutaneous and intravenous hydromorphone infusions for management of cancer
pain. Lancet, 337, 465±468.
Okuda, S., Yamaguchi, S., Kawazoe, Y. and Tsuda, K., 1964, Studies on morphine alkaloids. I.
Nuclear magnetic resonance spectral studies on morphine alkaloids. Chemical and
Pharmaceutical Bulletin, 12, 104±112.
Osborne, R., Thompson, P., Joel, S., Trew, D., Patel, N. and Slevin, M., 1992, The analgesic activity
of morphine-6-glucuronide. British Journal of Clinical Pharmacology, 34, 130±138.
Parab, P. V., Ritschel, W. A., Coyle, D. E., Gregg, R. V. and Denson, D. D., 1988, Pharmacokinetics
of hydromorphone after intravenous, peroral and rectal administration to human subjects.
Biopharmaceutics and Drug Disposition, 9, 187±199.
Roerig, S., Fujimoto, J. M. and Wang, R. I. H., 1973, Isolation of hydromorphone and
dihydromorphine glucuronides from urine of the rabbit after hydromorphone administration
(37291). Proceedings of the Society for Experimental Biology and Medicine, 143, 230±233.
Roerig, S., Fujimoto, J. M., Wang, R. I. H., Pollock, S. H. and Lange, D., 1976, Preliminary
characterization of enzymes for reduction of naloxone and naltrexone in rabbit and chicken liver.
Drug Metabolism and Disposition, 4, 53±58.
Searle, N. R., Roy, M., Bergeron, G., Perrault, J., Roof, J., Heermans, C., Courtemanche, M.,
Demers, C. and Carter, R., 1994, Hydromorphone patient-controlled analgesia (PCA) after
coronary artery bypass surgery. Canadian Journal of Anaesthesia, 41, 198±205.
Small, L. F., Eddy, N. B., Mosettig, E. and Himmelsbach, C. K., 1938, Studies on drug addiction.
United States Public Health Reports, Suppl. 138, 1±35.
Small, L. F. and Lutz, R. E., 1932, Chemistry of the opium alkaloids. United States Public Health
Reports, Suppl. 103, 159.
Thompson, P. I., Joel, S. P., John, L., Wedzicha, J. A., Maclean, M. and Slevin, M. L., 1995,
Respiratory depression following morphine and morphine-6-glucuronide in normal subjects.
British Journal of Clinical Pharmacology, 40, 145±152.
Vanier, M.-C., Labrecque, G., Lepage-Savary, D., Poulin, E., Provencher, L. and Lamontagne, C.,
1993, Comparison of hydromorphone continuous subcutaneous infusion and basal rate
subcutaneous infusion plus PCA in cancer pain. Pain, 53, 27±32.
Wallenstein, S. L., Houde, R. W., Portenoy, R., Lapin, J., Rogers, A. and Foley. K. M., 1990,
Clinical analgesic assay of repeated and single doses of heroin and hydromorphone. Pain, 41, 5±
13.
Yoshimura, H., Oguri, K. and Tsukamoto, H., 1968, Metabolism of drugs. LX. The synthesis of
codeine and morphine glucuronides. Chemical and Pharmaceutical Bulletin, 16, 2114±2119.
Zaw-Tun, N. and Bruera, E., 1992, Active metabolites of morphine. Journal of Palliative Care, 8, 48±
50.