European Journal of Pharmaceutical Sciences 68 (2015) 7886

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences

journal homepage: www.elsevier.com/locate/ejps

Population pharmacokinetics of morphine and morphine-6-glucuronide

following rectal administration A dose escalation study
Anne Brokjr a,1, Mads Kreilgaard b,2, Anne Estrup Olesen a,b,, Ulrika S.H. Simonsson c,3,
Lona Louring Christrup b,2, Albert Dahan d,4, Asbjrn Mohr Drewes a,e,1
a

Mech-Sense, Department of Gastroenterology & Hepatology, Aalborg University Hospital, Aalborg, Denmark
Department of Drug Design and Pharmacology, University of Copenhagen, Copenhagen, Denmark
Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden
d
Department of Anesthesiology, Leiden University Medical Center, Leiden, The Netherlands
e
Department of Clinical Medicine, Aalborg University Hospital, Aalborg, Denmark
b
c

a r t i c l e

i n f o

Article history:
Received 7 August 2014
Received in revised form 8 October 2014
Accepted 1 December 2014
Available online 5 December 2014
Keywords:
Morphine
Morphine-6-glucuronide
Dose escalation
Rectal administration
Population pharmacokinetics

a b s t r a c t
Introduction: To safely and effectively administer morphine as liquid formulation via the rectal route, a
thorough understanding of the pharmacokinetics is warranted. The aims were: (1) to develop a
population pharmacokinetic model of liquid rectal morphine and morphine-6-glucoronide (M6G), (2)
to simulate clinically relevant rectal doses of morphine and (3) to assess the tolerability and safety.
Material and methods: This open label, dose escalation, four-sequence study was conducted in 10 healthy
males. Three escalating doses of morphine hydrochloride (10 mg, 15 mg and 20 mg) were administered
20 cm from the anal verge. A 2 mg morphine hydrochloride dose was administered intravenously as reference. Blood samples were drawn at baseline and at nine time points post dosing. Serum was obtained
by centrifugation and assayed for contents of morphine and M6G with a validated high performance
liquid chromatographic method. Modelling was performed using NONMEM 7.2 and the rst order
conditional estimation method with interaction.
Results: A two compartment distribution model with one absorption transit compartment for rectal
administration and systemic clearance from the central compartment best described data. Systemic PK
parameters were allometric scaled with body weight. The mean morphine absorption transit time was
0.6 h, clearance 78 L/h [relative standard error (RSE) 12%] and absolute bioavailability 24% (RSE 11%).
To obtain clinically relevant serum concentrations, simulations revealed that a single morphine hydrochloride dose of 35 mg will provide sufcient peak serum concentration levels and a 46 mg dose four
times daily is suggested to maintain clinically relevant steady-state concentrations. Body weight was
suggested to be an important covariate for morphine exposure. No severe side effects were observed.
Conclusion: A population pharmacokinetic model of liquid rectal morphine and M6G was developed. The
model can be used to simulate rectal doses to maintain analgesic activity in the clinic. The studied doses
were safe and well tolerated.
2014 Elsevier B.V. All rights reserved.

Corresponding author at: Mech-Sense, Medicinsk Gastroenterologisk Afdeling,

Medicinerhuset, Mlleparkvej 4, 4. sal, 9000 Aalborg, Denmark. Tel.: +45 97663523;
fax: +45 9766 3577.
E-mail addresses: anne.brokjaer@rn.dk (A. Brokjr), mads.kreilgaard@sund.ku.
dk (M. Kreilgaard), aneso@rn.dk (A.E. Olesen), Ulrika.Simonsson@farmbio.uu.se (U.
S.H. Simonsson), llc@sund.ku.dk (L.L. Christrup), A.Dahan@lumc.nl (A. Dahan),
amd@rn.dk (A.M. Drewes).
1
Address: Mech-Sense, Medicinsk Gastroenterologisk Afdeling, Medicinerhuset,
Mlleparkvej 4, 4. sal, 9000 Aalborg, Denmark.
2
Address: Institut for Lgemiddeldesign og Farmakologi, Farmakoterapi, Jagtvej
160, Bygning 21, 2100 Kbenhavn , Denmark.
3
Address: Institutionen fr farmaceutisk biovetenskap, Uppsala biomedicinska
centrum BMC, Husarg. 3, Box 591751 24 Uppsala, Sweden.
4
Address: Afdeling Anesthesiologie, Leids Universitair Medisch Centrum, postbus
9600, 2300 RC Leiden, The Netherlands.
http://dx.doi.org/10.1016/j.ejps.2014.12.004
0928-0987/ 2014 Elsevier B.V. All rights reserved.

1. Introduction
Opioids are considered mainstay in management of moderate to
severe acute and chronic pain. They primarily produce analgesic
effect via activation of opioid receptors in the central nervous system. Additionally, peripheral opioid receptors may be responsible
for analgesia especially during inammation (Sehgal et al., 2011;
Stein and Lang, 2009). Among opioids, morphine is the golden
standard and oral administration is the cornerstone in pain management. When oral administration of morphine is not applicable,
the rectal route represents a non-invasive and cheap alternative.
Rectally administered non-standard liquid preparations, where
morphine is rapidly available for absorption, have to some extent

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

been under investigation in different patient groups (De Conno

et al., 1995; Pannuti et al., 1982; Westerling et al., 1982). However,
ndings in patients can be confounded by local and systemic pathological changes and no studies have to the authors knowledge
been performed in healthy participants. Furthermore, a thorough
understanding of the pharmacokinetics (PK) of morphine after rectal administration is lacking. Additionally, the PK of morphine-6glucoronide (M6G), a glucuronide metabolite of morphine, has
not been evaluated after rectal administration of liquid morphine.
To be able to effectively administer morphine as a liquid formulation via the rectal route it is fundamental to establish an analgesic
active dose but also to evaluate safety aspects. It is equally
important to attain a thorough pharmacokinetic understanding of
morphine and M6G. Therefore the aims of this study were: (1) to
develop a population PK model of liquid rectal morphine and
M6G; (2) to simulate clinically relevant rectal doses of morphine
and (3) to assess the tolerability and safety of three clinically relevant doses.

2. Material and methods

2.1. Participants
The study was conducted at the Department of Gastroenterology & Hepatology, Aalborg University Hospital. The participants
were all healthy males of Northern European descent aged 18
65 years. The protocol and informed consent were approved by
The North Denmark Region Committee on Health Research Ethics,
Denmark (N-20110077), the Danish Health and Medicines Authority (EudraCT identier: 201100516920) and registered in clinicaltrialsregister.eu. Participants were informed verbally and in
writing before deciding to participate. Subsequently, an informed
consent form was obtained from all participants. Medical history
and physical examination (including measurement of blood pressure and saturation) veried that each participant was healthy.
Individuals with a history of abuse of alcohol, opioids and other
drugs or with a history of drug and alcohol abuse in the near family
were not considered eligible for inclusion. Similarly, individuals
with previous severe illnesses or psychiatric diseases were
excluded. Participants were not allowed to use over the counter
analgesics 24 h prior to the experiment and were barred from
using strong analgesics altogether. The study was conducted in
accordance with the Declaration of Helsinki, local regulations and
International Conference on Harmonization (ICH) Good Clinical
Practice (GCP) guidelines. The GCP unit at Aarhus University Hospital monitored the study.

2.2. Medications and administration

The study was a single centre, open label, four-sequence design.
On the rst study day, 1 mL of morphine hydrochloride 2 mg/mL
for injection (Sygehus Apotekerne i Danmark, Denmark) was
administered as a reference formulation, as a total dose of 2 mg
morphine hydrochloride (1.5 mg free base). It was administered
intravenous via a peripheral venous catheter. Three separate study
days were used for escalating doses of morphine hydrochloride
10 mg, 15 mg and 20 mg (corresponding to 7.6, 11.4 and 15.2 mg
free base) which were administered rectally. Morphine hydrochloride 10 mg/mL for injection (SAD, Denmark) was used to formulate
the rectal dosages. It was diluted with isotonic saline (0.9% w/v) (B.
Braun, Denmark) to a concentration of 10, 15 and 20 mg/5 mL,
respectively. Medications were administered rectally, 20 cm from
the anal verge, by means of a custom designed rectal probe. It
had four channels (lumens) for medicine administration with an

79

outlet 1 cm below the tip of the probe. The probe enabled medication to be sprayed evenly onto the rectosigmoid wall (see Fig. 1).
The dosages were administered with the participant in the leftlateral position with knees and hips exed. Subsequently, the
probe was perfused with 1 mL of isotonic saline to ensure a complete drug administration. The probe was then slowly removed
and subsequently, liquid leakage was checked by visual inspection.
Following this procedure, the participant was allowed to sit
upright. The participant fasted for at least 10 h before the initiation
of each experiment. Prior to the medication administration the
participant had his bowels cleansed with an enema (Toilax 2 mg/mL
(Bisacodyl) Orion Corporation, Espoo, Finland). A washout period
of minimum one week was left between the four different
treatment schedules.
2.3. Blood sampling
Blood samples for the determination of the serum concentrations of morphine and M6G were drawn at baseline and nine times
after morphine administration via a peripheral venous catheter.
The samples were allowed to clot for a minimum of 15 min and
then centrifuged for 15 min at 3500 RPM. Serum was then pipetted
into vials in 2 mL aliquots, and subsequently stored at 80 C. A

Medicine administration site

Rectum

Anus

Rectal probe

Fig. 1. The custom designed rectal probe inserted 20 cm into the rectum. The probe
had four channels for medicine administration with an outlet 1 cm below the tip of
the probe. The probe enabled medication to be sprayed evenly onto the rectosigmoid wall.

80

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

sampling matrix design was used to allocate the participants to

individual blood sampling times. This resulted in the following
sampling intervals: 15 min, 610 min, 1115 min, 1620 min,
2230 min, 3660 min, 6690 min, 96120 min and nally 130
180 min. The participant had a blood sample drawn within each
of the nine intervals. The exact blood sampling times were noted
and consequently used in the PK calculations.
2.4. Analysis of morphine and M6G
The serum samples were assayed for contents of morphine and
M6G using a validated high performance liquid chromatographic
(HPLC) method. Protein precipitation was performed with acetonitrile. Acetonitrile contained an adequate amount of the internal
standard 2H3-morphine and 2H3-M6G (Cerilliant, Round Rock,
Texas, USA) and 100 lL 1 mM zinc sulphate. Samples were vortexed
for 2 min and then centrifuged for 5 min at 13,000 RPM. Following
this, the supernatant was transferred into a glass tube and dried
under a gentle stream of N2 at 50 C. The residues were reconstituted in 0.1% (v/v) formic acid in water. Subsequently, 20 lL of
the sample was injected into a 3 lm, 120 , 50  2.1 mm YMC-pack
ODS-AQ column (YMC Inacom, Overberg, The Netherlands) with an
ODS precolumn (Phenomenex, Utrecht, The Netherlands) at 30 C.
The mobile phase consisted of 0.1% formic acid in water with 3%
acetonitrile (Lichosolv, Merck B.V., Amsterdam, The Netherlands)
as modier. The eluent was monitored by a Quattro microAPI tandem mass spectrometer (Waters, Etten-Leur, The Netherlands).
The multiple reaction mode was used to determine the peak areas
of reaction ions from morphine, M6G and the internal standards.
The Masslynx 4.1 software (Waters, Etten-Leur, The Netherlands)
was used to integrate the data. Morphine and M6G levels were calculated using the internal standard method with a weighing factor
of 1/X. The lower limit of quantication (LLOQ) was 1 ng/mL for
both morphine and M6G, according to the lowest calibration standard included. The concentration of morphine and M6G was calculated as morphine free base. Internal quality controls of low
(morphine: 5.0 ng/mL, M6G:/2.5 ng/mL), medium (morphine:
50 ng/mL, M6G: 25 ng/mL) and high (morphine: 250 ng/mL,
M6G: 150 ng/mL) concentration were applied in every run. The
day-to-day variation expressed as the coefcient of variation for
morphine was 10.0% or less and 6.8% or less for M6G. Analysis
was performed by Department of Anesthesiology, Leiden University Medical Center, Leiden, The Netherlands.
2.5. Pharmacokinetic modelling
All intravenous and rectal serum concentrationtime proles of
morphine and M6G were modelled using NONMEM 7.2 (ICON
Development Solutions, Hanover, Maryland, USA). First order conditional estimation (FOCE) method with interaction was used.
Serum samples with concentrations down to 0.1 ng/mL (<5% of
total samples) were used for modeling purposes. Morphine doses
expressed as free base were used as dose input. Intravenous and
rectal PK data were modelled simultaneously. However, model
development was performed sequentially for morphine and M6G,
respectively, xing morphine parameters during the M6G model
development (but maintaining the morphine data in the data le).
In the nal model, all parameters for both morphine and M6G were
estimated simultaneously.
Log-normal distribution of the parameters around the typical
value was assumed:

Pi P  expgi
where Pi is the value of the parameter in the individual i, P is the
typical value of the parameter in the population, and gi is the normally distributed inter-individual variability (IIV) with mean 0 and

variance x2. Inter-occasion variability (IOV) was assessed for all

parameters (Karlsson and Sheiner, 1993). Additive, proportional,
and combined additive and proportional residual error models were
investigated.
One, two and three compartment models with clearance from
the central compartment were evaluated for morphine disposition.
Morphine absorption was modelled with a rst-order absorption
rate. Absorption delay was evaluated by estimation of a lag-time
or transit compartments modelled, either hard-coded (Erlangtype) (Rousseau et al., 2004) or by stirling approximation (Savic
et al., 2007).
Body weight was evaluated as a continuous covariate on the
systemic morphine PK parameters using a power function, either
empirically estimating the exponent of the relationship, or xing
the exponents according to allometric scaling (0.75 for clearance
and 1.0 for volume parameters) (Holford, 1996).
The formation of M6G, was modelled as a xed fraction
(fM6G = 0.14) of morphine clearance according to that published
by Ltsch et al. (Lotsch et al., 2002). For the rectal administration route, additional formation of M6G resulting from gut or
hepatic rst-pass metabolism was modelled as morphine clearance/bioavailability (CL/F) or CL/F/fadd, where fadd represents an
additional fraction involved in the formation rate of M6G. Transformation of M6G from the central morphine compartment (V2)
to the central M6G compartment (VM6G) was modelled either
direct or using Erlang-type transit compartments with rstorder rate constants in between (ktr,M6G). One or two compartment distribution was assessed for M6G with clearance from
the central M6G compartment (CLM6G). The nal model is illustrated in Fig. 2.
Perl-speaks-NONMEM (PsN) (Lindbom et al., 2004) was used for
model execution and simulations with Piranha graphical interphase (Keizer et al., 2013). Model discrimination and selection of
best t to the PK data was based on goodness-of-t plots, prediction corrected visual predictive checks (pcVPC) using Xpose 4.0
(Jonsson and Karlsson, 1999), parameters precision and reduction
in objective function values (OFV). In contrast to a regular VPC, a
pcVPC normalises the observed and simulated dependent variable
based on the typical population prediction for the median independent variable in the bin (Bergstrand et al., 2011). This enables validation of the predictive power of the model across multiple
dosages, which is particularly useful diagnosis of random effects
for studies with few subjects. Inclusion of a parameter was regarding statistically signicant (P < 0.05) for a dOFV of 3.84 for nested
structures, while P < 0.01 was used as backwards deletion criteria
for nal acceptance in the covariate analysis.
Simulation of therapeutically relevant rectal doses of
morphine was performed using the nal PK model with the
same study population as the current study (n = 10). For the
simulations and pcVPC, 2000 samples of the nal model were
performed.

2.6. Safety assessment

To assess safety the following side effects were continuously
rated: nausea, dizziness, itching, sweating, lower gastrointestinal complaints and sedation. The participants were asked
to rate the side effects at baseline and 10, 20, 30, 45, 60, 90 and
120 min after administration. The participants were asked to
report the severity of the side effects (1 = no side effects, 2 = modest side effects, 3 = moderate side effects and 4 = intolerable side
effects). In addition to this assessment, the participant was contacted approximately 24 h after drug administration in order to
obtain information regarding their wellbeing and any side effects
experienced.

81

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

V3

Rectal dose

Peripheral mor

IV dose

Q3

V M6G

CL M6G

Central M6G

A1

Ktr, F

A4

Transit mor

ka

V2

Central mor
CL

Ktr, M6G

CL/(F*fadd)*fM6G

A6

Transit M6G

Fig. 2. Final PK model diagram for simultaneous description of serum concentrationtime proles of morphine (mor) and M6G following IV or rectal administration of
morphine. A is drug amount in the compartment and numbers denote the compartments (1: dosing; 2 and 3: the central and peripheral compartments of morphine; 4:
morphine transit compartment from compartment 1 to 2; 5: central compartment of M6G; 6: transit compartment for M6G between compartment 2 and 5). Q3 is intercompartmental clearance. V is volume of distribution. CLM6G is M6G clearance. ktr, ktr,M6G and ka are rst-order rate constants between compartments, F rectal bioavailability
and fadd is additional fraction of morphine converted to M6G after rectal administration. fM6G is fraction of total systemic morphine clearance converted to M6G (xed to 0.14).

3. Results
A total of 10 healthy male participants were included in this
study. Their mean age was 25.6 years (range 2329 years), mean
height 180.1 cm (range 170196 cm) and mean weight 77.1 kg
(range 63100 kg). No leakage was observed following morphine
administration, thus, the full dose was administered in all
participants.
3.1. Pharmacokinetic model
The observed serum concentrationtime proles of morphine
and M6G following intravenous or rectal administration of morphine are presented in Fig. 3. Based on the dened model development and evaluation criteria, the morphine serum concentration
time proles were best described by a two-compartment distribution model with one absorption transit compartment for rectal
administration and systemic clearance from the central compartment (Fig. 2).
Mean absorption transit time for rectally administered morphine was calculated to 0.6 h for a typical person (i.e. 70 kg) in
the population. Formation of the metabolite M6G was modelled
as a xed fraction of morphine clearance (0.14 as reported in a previous paper (Lotsch et al., 2002)) from the central morphine compartment via a transit compartment into the central M6G
compartment. For simultaneous modelling of M6G formation after
intravenous and rectal administration (where additional M6G is
assumed to be formed by rst-pass metabolism), the M6G formation rate was best described as CL/(F fadd). The morphine and
M6G parameters for the nal PK model are summarized in Table 1.
Parameter precision <42% was observed for the nal model with
acceptable shrinkage for all random effects (<17%, except IOV for
V2 which was 42%). The nal PK model included allometric scaling
of weight on morphine clearance and volume parameters. Signicant improvement of the model was found by inclusion of IOV
for morphine absorption rate constant (ka), V2 and F in addition
to ktr,M6G for M6G. The residual unexplained variability was
described by a proportional and additive term for morphine and
a proportional term for M6G. The predictive performance of the
nal model is presented in a pcVPC (Fig. 4). The pcVPC conrmed
that the model adequately described the morphine and M6G data.

deviation) in patients treated for post-operative pain (Dahlstrom

et al., 1982). The nal PK model was used to simulate rectal morphine doses that provide equivalent analgesic exposure after single
and multiple doses, assuming equivalent serum/plasma concentration of morphine. From a population with similar demographics as
the participants in the current study, predicted morphine serum
concentrationtime proles from 2000 simulations suggest that,
a single dose of 35 mg morphine hydrochloride (equivalent to
27 mg free base of morphine) will provide median peak serum levels of around 21 ng/mL (95% prediction interval: 552 ng/mL)
(Fig. 5a). To sustain analgesic activity of morphine a maintenance
dose of 46 mg morphine hydrochloride (equivalent to 34 mg free
base of morphine) four times daily was suggested to provide a
median steady-state morphine concentration of 21 ng/mL for this
population (Fig. 5b).
Finally, the impact of body weight was simulated to demonstrate the impact on the morphine PK prole using a maintenance
dose of 46 mg morphine hydrochloride four times daily (Fig. 6).
This dose was predicted to provide a mean steady-state morphine
serum concentration of 26 and 18 ng/mL in a 63 kg and 100 kg person, respectively.
3.3. Safety assessment
All participants completed the study without complications. All
doses were well tolerated. The frequency of side effects reported
during each of the three sequences is summed up in Table 2.
Most participants rated the side effects as being modest. However, after 10 mg morphine administration; one participant
reported nausea and gastrointestinal complaints (stomach rumble)
as being moderate. After administration of 20 mg morphine, signs
of moderate dizziness were observed in one participant. None of
the participants rated the side effects as being intolerable and no
local intolerance or other gastrointestinal complications were
observed. When asked about their side effects 24 h after drug
administration, one participant, who had not otherwise expressed
any side effects during the study session, reported to have experienced sedation after the 10 mg morphine administration while
another, reported nausea after food intake following the 20 mg
morphine administration.
4. Discussion

3.2. Pharmacokinetic simulations

Therapeutically relevant steady-state plasma levels of
morphine have been reported to 21 12 ng/mL (mean standard

The population PK of morphine and M6G was modelled after

rectal and intravenous administration of morphine hydrochloride.
A two compartment distribution model with one absorption transit

82

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

(a)

10

(b)

100

Drug (ng/mL)

Drug (ng/mL)

100

10

0.1

0.1
0

(d)

(c)

100

Drug (ng/mL)

100

Drug (ng/mL)

Time (h)

Time (h)

10

0.1

10

0.1

Time (h)

Time (h)

Fig. 3. Serum concentrationtime proles of morphine (blue) and M6G (red) following administration of (a) 2 mg intravenous; (b) 10 mg rectal; (c) 15 mg rectal or (d) 20 mg
rectal morphine HCl (n = 10/dose).

Table 1
Population PK parameters of morphine and M6G from nal model, following intravenous (2 mg) or rectal (10, 15 and 20 mg) administration of morphine.
Parameter

Estimate (RSE%)

IIV (RSE%)

CL (WT/70)0.75 (L/h)
V2 (WT/70) (L)
V3 (WT/70) (L)
Q3 (WT/70)0.75 (L/h)
F (%)
ktr (h1)
ka (h1)
r1, prop error (%)
r2, add error (ng/mL)
r3, M6G prop error (%)
CLM6G (L/h)
VM6G (L)
ktr,M6G (h1)
fadd
fM6G

78 (12)
6.4 (31)
143 (14)
196 (18)
24 (11)
3.2 (30)
2.1 (9.5)
22 (12)
0.29 (41)
22 (9.7)
8.9 (7.3)
5.9 (22)
6.5 (27)
0.85 (9.2)
0.14 (FIX)

26 (23)

IOV (RSE%)
135 (16)

30 (27)
127 (34)

28 (22)
51 (20)

16 (29)
70 (26)
25 (29)

Description
Systemic clearance with allometric weight scaling
Central compartment with allometric weight scaling
Peripheral compartment with allometric weight scaling
Inter-compartmental clearance with allometric weight scaling
Rectal bioavailability
Transit compartment rate constant for rectal morphine absorption
Absorption rate constant
Proportional residual unexplained error
Additive residual unexplained error
Proportional residual unexplained error
Clearance of M6G from central M6G compartment
Volume of central M6G compartment
Transit compartment rate constant for M6G
Additional fraction converted to M6G from rectal administration
Fraction of total morphine clearance converted to M6G

IIV = inter-individual variability expressed as coefcient of variation, IOV = inter-occasion variability expressed as coefcient of variation, RSE = relative standard error
reported on the approximate standard deviation scale.

compartment for rectal administration and systemic clearance

from the central compartment best described the data. Simulation
based on the current model suggested that a single rectal morphine hydrochloride dose of 35 mg was sufcient to obtain clinically relevant peak serum concentration levels in a typical person
in the population. To maintain analgesic activity, a dose of 46 mg
four times daily was proposed to sustain clinically relevant
steady-state serum concentrations. The three studied morphine
doses in the trial (10, 15 and 20 mg) all proved safe and tolerable.

4.1. Pharmacokinetic model and simulations

The structural model and typical parameter values for morphines systemic PK are comparable with those previously published using two-compartment models (Ravn et al., 2014; Mazoit
et al., 2007). Three-compartment models have also been presented
for studies with longer sampling duration (Lotsch et al., 2002;
Dahan et al., 2004; Meineke et al., 2002). Body weight was incorporated in the nal model as allometric scaling of morphines sys-

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

83

Fig. 4. Prediction-corrected visual predictive check (pcVPC) of nal morphine and M6G PK model following administration of morphine. (a) Intravenous morphine,
arithmetic; (b) Intravenous morphine, semi-logarithmic; (c) rectal morphine; (d) M6G after rectal morphine administration; (e) M6G after intravenous morphine
administration. Blue dotsobservations; red line median of the prediction-corrected observations; Red dotted lines 2.5th and 97.5th percentile of the observed data; blue
shaded areas- 95% condence intervals for the 2.5th percentile and 97.5th percentiles of simulated data; pink shaded area 95% condence interval for the median of the
simulated data.

temic PK parameters. This nding indicates that clearance rates are

perfusion-limited, which is in good agreement with the high
extraction ratio evident by a systemic clearance value close to
liver blood ow (90 L/h for a 70 kg person) (Davies and Morris,

1993). Body weight as a covariate of morphine PK was also

reported by Ravn et al. (2014) and Dahan et al. (2004), although
the latter assumed a direct proportional correlation for their
model.

84

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

80

(a)
Morphine (ng/mL)

Morphine (ng/mL)

60

40

20

(b)

60

40

20

0
0

10

20

30

Time (h)

Time (h)

Fig. 5. Simulation of morphine serum concentrationtime proles using the nal PK model (2000 simulations of 10 participants with demographics from the current study),
following rectal administration of (a) single dose 36 mg morphine HCl targeting a median maximum serum concentration of 21 ng/mL or (b) 46 mg morphine HCl  4 with
6 h intervals targeting a median steady-state serum concentration of 21 ng/mL. Solid lines are prediction median and dotted lines are 2.5 (lower) and 97.5 (upper) percentile
for the simulated data.

50

Morphine (ng/mL)

40

30

20

10

0
0

10

15

20

25

Time (h)
Fig. 6. Simulation of typical morphine serum concentrationtime proles following
rectal administration of 46 mg morphine HCl times four with 6 h intervals for a
100 kg person (red) or a 63 kg person (blue) using the nal PK model.

Table 2
Side effects observed during each of the three sequences (10, 15 and 20 mg morphine
hydrochloride, respectively).
Type of side effect

10 mg
morphine

15 mg
morphine

20 mg
morphine

Nausea
Dizziness
Itching
Sweating
Lower gastrointestinal complaints
Sedation

2
4
0
1
1
0

2
3
0
1
0
1

1
3
0
1
0
0

Side effects in total

In this simultaneous modelling of morphine and M6G serum

concentrations after intravenous and rectal administration, the
typical value of bioavailability was estimated to 24% for rectally
administered morphine, which was found to be dose-independent.
No previous report of absolute bioavailability based on a population model for neither oral nor rectal administration of liquid morphine exits in the literature. However, using a two-stage individual
PK model, an absolute bioavailability of morphine has been
reported to 20 8.3% after oral administration of a 12 mg morphine tablet (Osborne et al., 1990). Thus, extent of morphine
absorption appears to be comparable between oral and rectal
administration in these two studies. In the present study, a custom

designed rectal probe enabled administration of morphine 20 cm

from the anal verge. Other studies have investigated the PK properties of liquid morphine after administration via the rectal route.
Westerling (1985) and Lundeberg et al. (2006) both evaluated the
PK of morphine after administration in the lower part of the rectum in children. Similarly, other studies (De Conno et al., 1995;
Pannuti et al., 1982; Westerling et al., 1982) have been performed
in order to assess the PK of morphine in an adult patient population. Although the site of administration is not accurately dened
in these studies, the description suggests that morphine has been
administered in the lower rectum. The lower part of the rectum
is drained by the middle and inferior veins, which drains into the
systemic system via the internal iliac vein. Consequently, partial
avoidance of hepatic rst-pass metabolism can be achieved. In contrast, when morphine is administered in the upper part of the rectum, the superior rectal vein drains via the portal system and
morphine will undergo rst-pass metabolism before entering the
systemic circulation (van Hoogdalem et al., 1991; de Boer and
Breimer, 1997). The low bioavailability seen in this study suggests
that absorption from the upper part of the rectum is subject to
extensive rst-pass metabolism. This is also reected in the substantially higher M6G formation rate, relative to intravenous
administration, which was inverse proportional to rectal bioavailability in this model. In the nal model, a signicant improvement
in the model performance by incorporating an additional fraction
term (fadd) to the formation rate after rectal administration was
seen, suggesting that M6G formation rate was approximately 15%
higher than CL/F alone. Population models enable assessment of
variability components for the PK of a drug. For morphine, the
absorption process appears to contribute the most to variability
between participants and study occasion (IOV for ka and IIV for
ktr). Also a large IOV was found for the central volume of distribution, V2. The variability in absorption suggested from the results of
our study is in good agreement with the results from the study of
Westerling et al. (1982). It has been suggested that poor mucosal
contact contributes to the large variation in both plasma concentration and bioavailability and that enhancement of contact
between drug and mucosa will lead to a signicant improvement
in the inter-individual variation (Westerling and Andersson, 1984).
Therapeutically relevant steady-state plasma levels of morphine
have been reported to 21 12 ng/mL in patients with post-operative pain (Dahlstrom et al., 1982). The PK model developed in the
current study is a useful tool to guide optimal clinical dosing after
rectal administration, which is not well described in current literature. To obtain clinically relevant serum concentrations with the
rectal route, simulations suggested that a single dose of 35 mg morphine hydrochloride will provide sufcient peak serum levels for a
typical person from the population for acute analgesia. To maintain

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

steady-state serum levels around 21 ng/mL for chronic pain treatment, simulations from the PK model suggest 46 mg morphine
hydrochloride four times daily may be sufcient for a typical person.
The PK model suggested body weight to be an important covariant
for morphine exposure. Simulating a multiple-dose PK prole for a
typical 63 kg person, suggested that steady-state serum levels will
be 41% higher than that of a typical 100 kg person. The model has
not been validated outside this body weight range. However, it
could be speculated that individuals weighing more than 100 kg
may have even lower morphine clearance, leading to sub-therapeutic serum concentrations, and vice versa for individuals weighing
less than 63 kg. This could warrant guidance for dose adjustments
for persons outside the normal body weight range according to
the allometric equation for total morphine clearance. To our knowledge, the current study is the rst to evaluate serum exposure of the
pharmacological active M6G metabolite after rectal administration
of liquid morphine. In vitro, M6G is reported to have 6-fold lower
binding afnity for the l-opioid receptors compared to morphine
(Kilpatrick and Smith, 2005). Using pupil diameter as a marker of
central opioid receptor activation, M6G is suggested to be almost
30-fold less potent than morphine relative to plasma exposure
(Skarke et al., 2003). Based on the current model, typical peak serum
values of M6G after rectal administration of morphine was estimated to be around 4-fold higher than morphine. Thus, it is unlikely
that M6G will contribute signicantly to morphine analgesia after a
single rectal dose.
4.2. Methodological considerations
To ensure a complete drug administration in the present study,
the rectal probe was perfused with 1 mL isotonic saline and subsequently, no leakage was observed. This implies that the full dose
was sufciently administered. Body position and movement as
well as individual rectal secretion are factors that may contribute
to variability in absorption, as seen in the present study. The participants were asked to calmly sit upright immediately after drug
administration. The movement and the inuence of gravity may
have contributed to spreading of morphine in the rectum, although
it seems unlikely with a volume of only 6 mL. Generally, spreading
of medication in the rectum is likely to be reduced by using a more
viscous solution and/or to minimize the volume. To further minimize the variability in future studies or in the clinic, it is essential
that the patient remain calm after administration. The population
of this study was healthy males, all with normal bowel movements
and before initiation of each study sequence the participant was
given a bowel cleansing enema. Thus, variation in absorption
caused by dysfunctional motility and faeces was considered
negligible. Given that the present study was performed in healthy
participants under standardized conditions and with limited
movement, it could be speculated that the observed variability in
absorption may be higher in the clinical setting. The probe used
for administration was custom-designed and thus, not commercially available. A similar device (e.g. a simple applicator) may in
theory enable patient self-administration. However, self-administration seems to be impractical due to the site of administration
in the upper rectum. In addition, as anatomical differences in hemorrhoidal venous drainage of the rectum may substantially inuence the systemic drug level achieved it is essential that the
administration device is placed correctly in the rectum. It is therefore recommended that medicine administration is performed by
professionals.
4.3. Safety assessment
This study found that liquid morphine administered rectally is
safe and well tolerable in healthy male participants in the dosage

85

regime presented in this study. Interestingly, the side effects did

not appear to be dose dependent. In two participants (one after
10 mg- and one after 20 mg morphine administration) modest side
effects were reported after the study sequence had been terminated. However, the study was not placebo-controlled and it is
complex to speculate whether these delayed side effects are causally related or related to the experimental procedure itself. In contrast to our ndings after single morphine administration, Panutti
et al. found local intolerance to liquid rectal morphine (diluted in
distilled water) in 6% of their study population (very advanced cancer patients) after 10 days of consecutive treatment (Pannuti et al.,
1982). It is unknown whether long term administration of morphine may lead to local intolerance in health or in disease.

5. Conclusions
A population PK model describing the PK prole of morphine
and M6G after rectal administration of liquid morphine was developed. Model simulations suggested a single rectal morphine dose
of 35 mg morphine hydrochloride to be sufcient to obtain clinically relevant peak serum levels in a typical person, while a dose
of 46 mg morphine hydrochloride rectally four times daily was
suggested to sustain clinically relevant steady-state serum levels
for chronic pain treatment. Body weight was shown to be signicantly related to population variability in morphine PK, and may
need to be accounted for by rectal administration to individuals
outside the normal weight range. The three studied doses of rectally administered morphine were all safe and well tolerated in
healthy male participants.

Acknowledgements
Support for this study was provided by The Danish Council for
Strategic Research and Det Obelske Familiefond. The authors
would like to thank Matias Nilsson, Mech-Sense, Department of
Gastroenterology & Hepatology, Aalborg University Hospital, Aalborg, Denmark for graphical illustration.

References
Bergstrand, M., Hooker, A.C., Wallin, J.E., Karlsson, M.O., 2011. Prediction-corrected
visual predictive checks for diagnosing nonlinear mixed-effects models. AAPS J.
13 (2), 143151.
Dahan, A., Romberg, R., Teppema, L., Sarton, E., Bijl, H., Olofsen, E., 2004.
Simultaneous measurement and integrated analysis of analgesia and
respiration after an intravenous morphine infusion. Anesthesiology 101 (5),
12011209.
Dahlstrom, B., Tamsen, A., Paalzow, L., Hartvig, P., 1982. Patient-controlled analgesic
therapy, Part IV: pharmacokinetics and analgesic plasma concentrations of
morphine. Clin. Pharmacokinet. 7 (3), 266279.
Davies, B., Morris, T., 1993. Physiological parameters in laboratory animals and
humans. Pharm. Res. 10 (7), 10931095.
de Boer, A.G., Breimer, D.D., 1997. Hepatic rst-pass effect and controlled drug
delivery following rectal administration. Adv. Drug Deliv. Rev. 28 (2), 229237.
De Conno, F., Ripamonti, C., Saita, L., MacEachern, T., Hanson, J., Bruera, E., 1995.
Role of rectal route in treating cancer pain: a randomized crossover clinical trial
of oral versus rectal morphine administration in opioid-naive cancer patients
with pain. J. Clin. Oncol.: Off. J. Am. Soc. Clin. Oncol. 13 (4), 10041008.
Holford, N.H., 1996. A size standard for pharmacokinetics. Clin. Pharmacokinet. 30
(5), 329332.
Jonsson, E.N., Karlsson, M.O., 1999. Xposean S-PLUS based population
pharmacokinetic/pharmacodynamic model building aid for NONMEM.
Comput. Methods Programs Biomed. 58 (1), 5164.
Karlsson, M.O., Sheiner, L.B., 1993. The importance of modeling interoccasion
variability in population pharmacokinetic analyses. J. Pharmacokinet.
Biopharm. 21 (6), 735750.
Keizer, R.J., Karlsson, M.O., Hooker, A., 2013. Modeling and simulation workbench
for NONMEM: tutorial on Pirana, PsN, and Xpose. CPT: Pharmacomet. Syst.
Pharmacol. 2, e50.
Kilpatrick, G.J., Smith, T.W., 2005. Morphine-6-glucuronide: actions and
mechanisms. Med. Res. Rev. 25 (5), 521544.

86

A. Brokjr et al. / European Journal of Pharmaceutical Sciences 68 (2015) 7886

Lindbom, L., Ribbing, J., Jonsson, E.N., 2004. Perl-speaks-NONMEM (PsN)a Perl
module for NONMEM related programming. Comput. Methods Programs
Biomed. 75 (2), 8594.
Lotsch, J., Skarke, C., Schmidt, H., Liefhold, J., Geisslinger, G., 2002. Pharmacokinetic
modeling to predict morphine and morphine-6-glucuronide plasma
concentrations in healthy young volunteers. Clin. Pharmacol. Ther. 72 (2),
151162.
Lundeberg, S., Hatava, P., Lagerkranser, M., Olsson, G.L., 2006. Perception of pain
following rectal administration of morphine in children: a comparison of a gel
and a solution. Paediatr. Anaesth. 16 (2), 164169.
Mazoit, J.X., Butscher, K., Samii, K., 2007. Morphine in postoperative patients:
pharmacokinetics and pharmacodynamics of metabolites. Anesth. Analg. 105
(1), 7078.
Meineke, I., Freudenthaler, S., Hofmann, U., Schaeffeler, E., Mikus, G., Schwab, M.,
Prange, H.W., Gleiter, C.H., Brockmoller, J., 2002. Pharmacokinetic modelling of
morphine, morphine-3-glucuronide and morphine-6-glucuronide in plasma
and cerebrospinal uid of neurosurgical patients after short-term infusion of
morphine. Br. J. Clin. Pharmacol. 54 (6), 592603.
Osborne, R., Joel, S., Trew, D., Slevin, M., 1990. Morphine and metabolite behavior
after different routes of morphine administration: demonstration of the
importance of the active metabolite morphine-6-glucuronide. Clin.
Pharmacol. Ther. 47 (1), 1219.
Pannuti, F., Rossi, A.P., Iafelice, G., Marraro, D., Camera, P., Cricca, A., Strocchi, E.,
Burroni, P., Lapucci, L., Fruet, F., 1982. Control of chronic pain in very advanced
cancer patients with morphine hydrochloride administered by oral, rectal and
sublingual route. Clinical report and preliminary results on morphine
pharmacokinetics. Pharmacol. Res. Commun. 14 (4), 369380.
Ravn, P., Foster, D.J., Kreilgaard, M., Christrup, L., Werner, M.U., Secher, E.L., Skram,
U., Upton, R., 2014. Pharmacokineticpharmacodynamic modelling of the

analgesic and antihyperalgesic effects of morphine after intravenous infusion

in human volunteers. Basic Clin. Pharmacol. Toxicol.
Rousseau, A., Leger, F., Le Meur, Y., Saint-Marcoux, F., Paintaud, G., Buchler, M.,
Marquet, P., 2004. Population pharmacokinetic modeling of oral cyclosporin
using NONMEM: comparison of absorption pharmacokinetic models and design
of a Bayesian estimator. Ther. Drug Monit. 26 (1), 2330.
Savic, R.M., Jonker, D.M., Kerbusch, T., Karlsson, M.O., 2007. Implementation of a
transit compartment model for describing drug absorption in pharmacokinetic
studies. J. Pharmacokinet. Pharmacodyn. 34 (5), 711726.
Sehgal, N., Smith, H.S., Manchikanti, L., 2011. Peripherally acting opioids and clinical
implications for pain control. Pain Phys. 14 (3), 249258.
Skarke, C., Darimont, J., Schmidt, H., Geisslinger, G., Lotsch, J., 2003. Analgesic effects
of morphine and morphine-6-glucuronide in a transcutaneous electrical pain
model in healthy volunteers. Clin. Pharmacol. Ther. 73 (1), 107121.
Stein, C., Lang, L.J., 2009. Peripheral mechanisms of opioid analgesia. Curr. Opin.
Pharmacol. 9 (1), 38.
van Hoogdalem, E., de Boer, A.G., Breimer, D.D., 1991. Pharmacokinetics of rectal
drug administration, Part I. General considerations and clinical applications of
centrally acting drugs. Clin. Pharmacokinet. 21 (1), 1126.
Westerling, D., 1985. Rectally administered morphine: plasma concentrations in
children premedicated with morphine in hydrogel and in solution. Acta
Anaesthesiol. Scand. 29 (7), 653656.
Westerling, D., Andersson, K.E., 1984. Rectal administration of morphine hydrogel:
absorption and bioavailability in women. Acta Anaesthesiol. Scand. 28 (5), 540
543.
Westerling, D., Lindahl, S., Andersson, K.E., Andersson, A., 1982. Absorption and
bioavailability of rectally administered morphine in women. Eur. J. Clin.
Pharmacol. 23 (1), 5964.