Review

For reprint orders, please contact reprints@expert-reviews.com

Pharmacokinetic–
pharmacodynamic modeling
in acute and chronic pain:
an overview of the
recent literature
Expert Rev. Clin. Pharmacol. 4(6), 719–728 (2011)

Christian Martini1, In acute and chronic pain, the objective of pharmacokinetic–pharmacodynamic (PKPD) modeling
Erik Olofsen1, is the development and application of mathematical models to describe and/or predict the time
Ashraf Yassen2, course of the pharmacokinetics (PK) and pharmacodynamics (PD) of analgesic agents and link
PK to PD. Performing population PKPD modeling using nonlinear mixed effects modeling allows,
Leon Aarts1
apart from the estimation of fixed effects (the PK and PD model estimates), the quantification
and Albert Dahan†1 of random effects as within- and between-subject variability. Effect-compartment models and
1
Department of Anesthesiology, mechanism-based biophase distribution models that incorporate drug-association and
Leiden University Medical Center,
2330 RC Leiden, The Netherlands
-dissociation kinetics are applied in PKPD modeling of pain treatment. Mechanism-based models
2
Global Clinical Pharmacology and enable the quantification of the rate-limiting factors in drug effect owing to drug distribution
Exploratory Development, Astellas versus receptor kinetics (since receptor kinetics are nonlinear they are discernable from the linear
Pharma Global Development effect-compartment kinetics). It is a helpful technique in understanding the complex behavior
Europe, Leiderdorp, The Netherlands
†
Author for correspondence: of specific analgesics, such as buprenorphine, but also morphine and its active metabolite
Tel.: +31 715 262 301 morphine-6-glucuronide, especially with respect to the reversal of opioid-induced side effects,
a.dahan@lumc.nl most importantly life-threatening respiratory depression. One approach in chronic pain studies
is the application of mixture models. Mixture models do not necessarily need to take PK data
into account and allow the objective differentiation of measured responses to analgesics into
specific response subgroups, and as such, may play an important role in analyzing Phase I and
II analgesia studies. Appropriate application of PKPD modeling leads to the improvement of
current therapeutics with respect to dose design and outcome, understanding the interaction
of analgesics within complex chronic pain disease processes and may play an important role in
drug development. In the current article, novel observations using the aforementioned techniques
on opioids, NSAIDs, epidural analgesia, ketamine and GABA-ergic drugs in acute and chronic
pain are discussed.

Keywords : acute pain • analgesics • buprenorphine • chronic pain • effect compartment • epidural analgesia
• ketamine • modeling • morphine • opioids • pain • pharmacodynamics • pharmacokinetics • PKPD modeling

The objective of pharmacokinetic–pharma- (infinitely small) effect compartment to describe

codynamic (PKPD) modeling is the develop- the delay in effect (i.e., quantification of drug
ment and application of mathematical models effect in terms of drug concentration at the
to describe and/or predict the time course of target or effect site [CE ] vs measured effect).
dose-to-concentration (pharmacokinetics [PK]) In compartmental models, the concentration–
and concentration-to-effect (pharmacodynam- time profiles are described/explained by drug
ics [PD]) of pharmacological active agents in transfer between interconnected hypothetical
health and disease [1,2] . The PK model part of compartments, mimicking drug absorption,
PKPD models describes the macrodistribution distribution and elimination. Drug distribu-
of the drug (drug distribution kinetics). Effect- tion towards the target site is described by the
compartment PKPD models add a hypothetical plasma effect-site equilibration constant (ke0 ; or

www.expert-reviews.com 10.1586/ECP.11.59 © 2011 Albert Dahan ISSN 1751-2433 719

 Review Martini, Olofsen, Yassen, Aarts & Dahan

its half-life [t½ke0 ] = ln2/ke0 ), whereas the concentration–effect PKPD modeling of acute antinociception & pain
relationship is often described by a sigmoid maximum effect Morphine & its metabolites
(Emax) model of the form: Morphine, first extracted from opium in 1806, is still considered
the gold standard in the treatment of severe acute and chronic
Effect = Ac /(1 + Ac) pain, although it has no clear superiority in efficacy or tolerability
over other opioids [11] . In 2005, Lötsch listed six PKPD studies on
where A = CE /C50 (with C50 as the measure of drug potency or morphine [12] . Studies were performed either on a surrogate meas-
the effect-site or steady-state concentration causing 50% of the ure of opioid effect, pupil size, or on electrical noxious cutaneous
effect, and g as the Hill-factor). The first to propose a first-order stimulation. One of the studies discussed demonstrated that sex
rate constant between plasma and an effect-compartment was was a significant covariate on pain relief responses, with greater
the Italian, Segre, in a pivotal article published in 1968 [3] . Segre morphine potency in women (C50 = 250 nM in men vs 150 nM in
described the transfer function of norepinephrine’s effect on the women) (Table 1) but with a slower onset/offset time (blood effect-
circulatory system and estimated a delay of 15 s between changes site equilibration t½ke0 for pain relief response [pain tolerance] was
in plasma epinephrine concentration and blood pressure in a cat, 1.6 h in men vs 4.8 h in women), without any sex differences in
showing that epinephrine’s target site (i.e., the ‘biophase’) is not morphine’s PK [13] . These data were later confirmed in a systematic
in plasma. This concept was later further developed by Hull and review and meta-analysis and account for the difference in opioid
Sheiner et al., in two separate studies, both in 1979 [4,5] . Effect- consumption between men and women in postoperative patient-
compartment models are empirical models that do not describe controlled analgesia morphine studies [14] . Interestingly, while
the physiological or mechanistic pathway of the drug from its con- studies on pupil size had similar estimates for t½ke0 as observed in
centration in plasma to effect. Recently, novel mechanism-based pain studies, the potency of morphine was much greater for miosis
biophase distribution models were developed [1] . For example, (C50 ranging from 17 to 24 nM, without occurrence of any sex
mechanism-based PKPD models may incorporate concepts from differences). Collectively, these data indicate that morphine has
receptor theory describing the drug–receptor interaction (receptor an onset time that is much smaller than observed in any of the
kinetics) in terms of drug-association and -dissociation kinetics other opioids currently in clinical use [12–14] . In comparison, t½ke0
and enable the quantification of the rate-limiting factor in drug (derived from yet another surrogate end point of opioid effect:
effect owing to drug distribution versus receptor kinetics [6] . changes in EEG power spectrum) is 5–6 min for fentanyl and
Rather than analyzing individual PKPD data, population sufentanil, 1–2 min for alfentanil and remifentanil, 8–19 min
PKPD analyses (i.e., performing a simultaneous analysis on the for methadone and 17 min for piritramide [12] . This makes mor-
whole population dataset) has become the acceptable approach in phine a relatively slow and consequently difficult to control opioid
current PKPD modeling studies. A population analysis will lead analgesic. Note, however, that currently no comparable data on
to an estimate of the population PK and PD estimates, identify morphine’s effect on EEG-related parameters are available.
the sources of variability that influence PK and PD, estimate the Morphine is rapidly metabolized in the liver into the (presum-
magnitude of between-subject variability, as well as intrasubject ably) inactive morphine-3-glucuronide (M3G) and the active
variability, and allow for the estimate of random residual vari- morphine-6-glucuronide (M6G). Animal studies indicate that
ability (random effects) [7] . To separate random (i.e., variability M6G is a μ-opioid receptor agonist with greater potency than
parameters) from fixed effects (i.e., the PK and PD parameter esti- morphine and rapidly crosses the blood–brain barrier (BBB) [15] .
mates), nonlinear mixed-effects modeling is required [8] . Various By contrast, PKPD studies in humans revealed a three- to five-
statistical packages are available to perform nonlinear mixed- fold lower potency than observed in animals after intravenous
effects modeling, of which NONMEM® is considered the most M6G administration coupled with a slow transfer of the drug
state-of-the-art [7–10] . NONMEM was developed by Lewis Sheiner across the BBB (C50 = 750 nM and t½ke0 6–8 h) [12] . Using data
and Stuart Beal at the University of California in San Francisco obtained from healthy volunteers that were injected with either
(CA, USA) and was initially developed to analyze PK data but intravenous morphine or intravenous M6G, we were able to con-
is currently applied to model PK, PD and PKPD datasets [9,10] . struct a population PKPD model of morphine’s metabolism into
Particularly in studies on the pharmacology of analgesics and anes- M6G and determine M6G’s contribution to morphine analge-
thetics, drug–drug interactions, population PK, PD and PKPD sia [16] . The fraction of morphine metabolized into M6G was
analyses using NONMEM is currently the gold standard [8,10] . 6.0 ± 0.2% (median value ± standard error); M6G formation was
In this article, we will discuss population effect-compartment sex independent. Simulation studies demonstrated that repetitive
and mechanism-based PKPD and PD modeling studies pub- morphine infusions (0.1 mg/kg at 8-h intervals) resulted in stable
lished in the last 5 years on the effects of various analgesic agents M6G effect-site (‘brain’) concentrations of 10–20 nM that con-
in acute (in volunteers and postoperative patients) and chronic tributed approximately 8 (women) and 15% (men) to the analgesic
pain patients. End points covered are the relief of experimen- response. Under conditions of renal impairment, the effect-site
tal and surgery- and disease-induced pain (i.e., antinociception M6G concentration rose by a factor of ten and its estimated con-
and analgesia). In addition, we will discuss PKPD models of tribution to effect increased by a factor of two. These findings
naloxone reversal of opioid-induced respiratory depression, the derived from modeling studies and simulations are important as
most life-threatening side effect of potent opioids. they tempered the initial aspirations of M6G as the replacement

720 Expert Rev. Clin. Pharmacol. 4(6), (2011)

 Pharmacokinetic–pharmacodynamic modeling in acute & chronic pain Review

Table 1. Population pharmacodynamic model estimates for various analgesic agents and naloxone using
pain relief or respiration as the effect parameter: studies performed in healthy volunteers or patients.
Drug t½ke0 kon (ml/ koff† Kd C50 Effect parameter Population Ref.
ng/min) (min-1) (koff /kon)
Morphine (men) 1.6 h 250 nM Antinociception V [13]

Morphine 4.8 h 125 nM Antinociception V [13]

(women)
Morphine-6- 6–8 h 750 nM Antinociception V [12,16]
glucuronide
Morphine 1.2 h 0.85 0.14 160 nM Respiration V [17]

Morphine-6- 2.7 h 0.04 0.03 880 nM Respiration V [17]

glucuronide
Naloxone 5–8 min 0.5–2 nM Respiration V [17]

Morphine 1.7 h 124 nM Postoperative analgesia P [20]

Morphine-6- 3h 13 nM Postoperative analgesia P [20]

glucuronide
Buprenorphine 75 min 0.25 0.01 0.09 nM Respiration V [6]

Fentanyl 16.4 min >100 >100 1140 ng/ml Respiration V [6]

Buprenorphine 155 min 0.06 0.08 2.66 nM Antinociception V [24]

S-ketamine ‡
373–2200 ng/ml §
Antinociception V [28]

S-ketamine 11 days ¶
10.5 ng/ml Chronic CRPS pain relief P [34]

Pregabalin 11 h 1.54 µg/ml Chronic fibromyalgia pain relief P [35]

SC-75416 5500 ng/ml Postoral surgery pain relief P [29]

Rofecoxib 284 ng/ml Postoral surgery pain relief P [29]

Valdecoxib 68 ng/ml Postoral surgery pain relief P [29]

Ibuprofen 6840 ng/ml Postoral surgery pain relief P [29]

Ibuprofen 28 min 10,200 ng/ml Postoral surgery pain relief P [30]

Acetaminophen 53 min 10 mg/l Post-tonsillectomy pain relief P [32]

†
t½koff is calculated as ln2/koff.
‡
No hysteresis observed between plasma concentration and effect.
§
C50 varied depending on the nociceptive assay employed (see text).
¶
The value of 11 days represent a rate-constant form disease modulation (t½k) unrelated to the blood effect-site equilibration t½.
C50 : Steady-state concentration causing 50% of the effect; CRPS: Complex regional pain syndrome; Kd: Equilibrium dissociation constant; ke0 : Effect-site equilibration
constant; Koff: Receptor dissociation constant; Kon: Receptor-association constant; P: Patients; t1/2: Half-life; V: Volunteers.

of morphine in the treatment of severe acute pain and exemplify receptor association constant, koff or the receptor disociation con-
the strength of population PKPD modeling studies in the devel- stant of morphine and M6G, and ke0 of naloxone. Naloxone is
opment of new (analgesic) drugs [12,15,16] . The aspirations were a µ-opioid receptor antagonist used both in clinical practice to
based on early observations in animals and later in humans that reverse opioid-induced respiratory depression and in the treat-
M6G causes less respiratory depression and possibly less nausea ment of opioid addiction. Since it was first synthesized in 1960 by
and vomiting compared with the parent compound [15] . However, Jack Fishman and further developed through the early 1970s by
its low potency but especially its slow passage across the BBB and Harold Blumberg [18] , little progress has been made on the policies
accumulation in renal impairment make it an analgesic that is even required to administer appropriate doses of naloxone that cause
more difficult to control than morphine in perioperative patients rapid reversal with a limited chance of renarcotization. The study
and should be avoided in patients with reduced renal function. by Olofsen et al. [17] and especially the additional model simula-
Previously discussed studies used effect-site PKPD modeling and tions are insightful and enable the development of specific guide-
consequently remained uninformed on receptor kinetics. Using lines on the reversal of the toxic effects of morphine and its active
a mechanism-based PKPD approach, Olofsen et al. described metabolite M6G in clinical practice. The parameter estimates of
naloxone reversal of morphine- and M6G-induced respiratory the study are given in Table 1. Morphine and M6G both display
depression [17] . They estimated the following pharmacokinetic slow receptor association/dissociation kinetics with an apparent
parameters: ke0 or the biophase distribution constant, kon or the potency (Kd = koff/kon) similar to C50 values obtained in previous

www.expert-reviews.com 721
 Review Martini, Olofsen, Yassen, Aarts & Dahan

studies (see earlier). Naloxone had rapid receptor kinetics with t½ke0 resumes within minutes) [19] . A more rapid reversal of fentanyl
of approximately 5 min on average, while its apparent estimated just requires a greater dose of naloxone, as naloxone reversal is not
potency was greater for M6G reversal than for morphine reversal. dictated by receptor kinetics of the agonist (fentanyl).
The latter observation is hard to explain as naloxone is a competi- Two recent studies analyzed the effect of morphine on
tive antagonist at the opioid receptors (i.e., its effect is described by acute postoperative pain using a population PKPD analysis in
a single equilibrium dissociation constant). Possibly, the differences NONMEM. Mazoit et al. modeled morphine’s effect taking the
in potency may be related to the ability of M6G to increase the potential effects of M3G and M6G into account in patients in
affinity of naloxone for the µ-receptor [17] . Since the elimination t½ acute pain following a variety of surgical interventions with a
of naloxone is between 15 and 30 min, naloxone is a limiting factor mean age of 51 years [20] . They assumed additive agonistic effects
in the reversal of (long-acting) opioids. However, in addition, the from morphine and M6G and an antagonistic effect from M3G.
slow receptor kinetics of morphine (and M6G) limits naloxone’s Using a sigmoid Emax model with pain (as determined from a visual
ability to disperse the opioid agonist from the receptor rapidly. analogue scale [VAS] scoring system) as effect variable, they esti-
This has clinical consequences as increasing the naloxone dose mated a tenfold greater potency of M6G compared with morphine
will not increase the speed of morphine reversal (i.e., the increase (C50 = 124 [morphine] vs 13 nM [M6G]) and an inhibitory effect
in ventilation over time), although the magnitude and duration from M3G with C50 of 880 nM. The analgesic effect of M6G was
of reversal will increase (Figure 1). Taken together, a sufficient high delayed significantly relative to morphine (t½ke0 : 1.7  [morphine],
dose of naloxone is required, preferably administered as continu- 3  [M6G] and 3 h [M3G]). Age, sex and weight did not improve
ous infusion, to reverse morphine-induced respiratory depression the PD model. The M6G potency is much greater than observed
with a reduced chance of renarcotization; the speed of reversal is in earlier studies in volunteers [12,16,17] and, although one may argue
not affected by the naloxone dose. Note that these observations that in the Mazoit study measurements were made for 24–48 h
regard only opioid agonists that have slow receptor association/ (compared with 6–10 h in previous studies), the M6G potency is
dissociation kinetics (such as morphine, M6G and buprenorphine, an indirect estimate derived from a study where morphine, but not
see later) but are not valid for opioids with rapid kinetics such as M6G, was administered. Consequently, the true potency of M6G
fentanyl or any of the other fenylpiperidines (with the exception remains unknown, in comparison with previous studies where only
of remifentanil; if the remifentanil infusion is stopped, breathing M6G was administered. Of further interest is the observation of

Antagonist dose koff opioid agonist

100 100

high 5
Ventilation (% of baseline)
Ventilation (% of baseline)

80 80

60 60 1

low
40 40

20 20 0.2

0 0
0 60 120 180 240 0 60 120 180 240

Time (min) Time (min)

Figure 1. Effect of naloxone on opioid-induced respiratory depression. (A) Mechanism-based pharmacokinetic–

pharmacodynamic modeling of the ability of different doses of an opioid antagonist (naloxone, given at t = 55 min, low dose = 0.1 mg,
high dose = 0.4 mg) to reverse the opioid-induced respiratory depression from an opioid agonist with relatively low koff value (opioid
agonist injected at t = 0 min). The speed of reversal is independent of the antagonist dose and is determined by the agonist opioid
receptor koff value. (B) Effect of variations in the koff value of an opioid agonist (injected at t = 0 min over 90 s) on the effect of a fixed
dose of an opioid receptor antagonist (naloxone, given at t = 55 min over 90 s) on reversal of opioid-induced respiratory depression.
Koff values are relative with 0.2 min-1 = 0.2 × typical value, 1 = typical value, 5 = 5 × typical value. Here, the typical value = 0.03 min-1.
Koff: Receptor dissociation constant.
Data taken from [17] .

722 Expert Rev. Clin. Pharmacol. 4(6), (2011)

 Pharmacokinetic–pharmacodynamic modeling in acute & chronic pain Review

an antagonistic or hyperalgesic effect of M3G with a relatively analgesic that provides long-term analgesia without ceiling over
slow distribution into the biophase. This suggests that reduced the clinical dose-range and has a limited effect on the respira-
morphine efficacy may coincide with M3G’s hyperalgesic effect, tory systems. In case life-threatening respiratory depression does
requiring the switch to other opioid analgesics in some patients. occur, it can be antagonized but requires high doses and continu-
Abbou Hammoud et al. modeled the effect of morphine titra- ous infusions of naloxone. These data demystify some of the old
tion in the first postoperative hours on pain relief (using VAS) in beliefs around buprenorphine that can still be found in older text-
patients (mean age of 62 years) following orthopedic surgery [21] . books, including the existence of a ceiling in analgesic effect and
They used an indirect response Emax model with variations in pain inability to reverse buprenorphine-induced respiratory depression.
expressed by the equilibrium of K in (a zero-order rate constant of
the appearance of pain) and Kout (a first-order rate constant of the Ketamine
disappearance of pain) with morphine inhibiting K in. They added The third ‘senior’ analgesic that we discuss is the 50-year-old
a virtual kinetic compartment, as no PK data were available. They N-methyl-d-aspartate receptor (NMDAR) antagonist ketamine.
estimated a value for ED50 (dose of morphine causing 50% pain Initially developed as an anesthetic, multiple studies show that
relief) of 10.2 ± 0.8 mg with significant covariates including initial at low (subanesthetic) doses, ketamine is a potent analgesic in
VAS (direct postoperative VAS just before dosing), and the use the treatment of both acute and chronic pain [25] . In the last few
of additional pain medication (but not age, sex or weight). The years, a series of investigations on population PKPD modeling in
model is of interest as it predicts that greater titration doses are children and adults has been published. The studies in children
associated with a significant reduction in time to achieve analgesia were aimed at the anesthetic properties of ketamine and will not
(especially in patients with a greater initial VAS), albeit at the be discussed here [26,27] . In healthy volunteers, Sigtermans et al.
expense of increased toxicity (sedation in their study). tested the effect of increasing doses of the S(+)-enantiomer of
ketamine (S-ketamine) on antinociception using two different
Buprenorphine nociceptive assays: heat pain (a fixed-heat stimulus was applied to
There is a renewed and increasing interest in this 30-year-old semi­ the arm giving a baseline [i.e., predrug] VAS of 6 cm or greater)
synthetic opiate derived from the morphine precursor, thebaine. and tolerance to electrical pain (a 10-Hz stimulus train increasing
Currently, buprenorphine is applied in the treatment of chronic at 0.5 mA/s) [28] . Using an effect-compartment population PKPD
pain using a relatively novel patch formulation and in the treatment ana­lysis yielded a significant difference in ketamine’s potency in
of addiction (Subutex™ or in combination with naloxone Subutex the two assays (Table 1) with a sixfold greater potency for pain relief
[Reckitt Benckiser, Slough, UK]). Buprenorphine is a partial ago- of heat pain. Of further interest is the observation of hyperalgesic
nist at the µ-opioid receptor and displays typical opioid behavior responses following S-ketamine withdrawal that was modeled with
(analgesia, sedation, nausea, delayed gastric emptying and respira- a linear trend term. This phenomenon has recently been described
tory depression). Yassen et al. performed a series of experiments in in experimental and clinical studies. A delay between concentra-
men using a mechanism-based population PKPD approach [6,22–24] . tion and effect (that is hysteresis) was not observed for any of the
The most important observations from these studies are: antinociceptive end points. This indicates an almost immediate
passage of S-ketamine across the BBB and rapid receptor kinetics.
• In the clinically relevant intravenous dose range of
Similar observations were made previously on the effect of keta-
0.05–0.6 mg/kg, buprenorphine displayed a lack of ceiling
mine racemic mixture on EEG slowing in adults and on estimates
effect for antinociception (related to the absence of saturation
of arousal and recall memory during anesthesia in children [26] .
of receptor occupancy) [24] ;
• Rate-limiting factors in the onset and offset of antinociceptive COX inhibitors
effect are biophase distribution and slow receptor association– Kowalski et al. constructed PKPD models to describe the effect
dissociation (Table 1) [24] ; of the experimental COX-2 inhibitor SC-75416 (Pfizer, MI, USA)
and performed trial simulations to enhance the further develop-
• In contrast to antinociception, ceiling was observed for respira-
ment of this analgesic [29] . SC-75416 is a selective COX-2 inhibitor
tory depression at high buprenorphine concentrations (the
that has anti-inflammatory and analgesic activity. They studied
intrinsic activity of buprenorphine is 56%) [23] ;
postoral surgery patients and performed comparisons with pla-
• Reversal of buprenorphine-induced respiratory depression with cebo, ibuprofen, rofecoxib and valdecoxib. In an initial study,
naloxone is possible, however, owing to the slow receptor kinet- an oral solution of SC-75416 was followed by a capsule formula-
ics of buprenorphine, in combination with the fast elimination tion. They constructed logistic effects models in NONMEM
kinetics of naloxone, naloxone is best administered as a con- with three main components: placebo effect (modeled by an
tinuous infusion rather than a single bolus injection (optimal exponential function) and drug effect (modeled by a sigmoid
dosage is 2–4 mg naloxone per h) [6] . Emax function); random effect; and investigating pain relief. The
estimated EC50 values for SC-7541, rofecoxib, valdecoxib and
Naloxone reversal is slow, however, and cannot be accelerated ibuprofen were 5.5, 0.3, 0.07 and 6.8 µg/ml, respectively. The
by giving higher naloxone doses (see earlier). The findings from most important results of the study were: the study and trial
these modeling studies indicate that buprenorphine is an opioid simulations prompted further study developments (pursuing a

www.expert-reviews.com 723
 Review Martini, Olofsen, Yassen, Aarts & Dahan

high-dose strategy) that otherwise might not have been consid- analysis enables the development of predictive epidural anesthe-
ered; the approach resulted in time and cost savings by not having sia, in this case, analgesia models that may improve therapeutic
to repeat the postoral surgery study and formulation rework was outcome. An indirect assessment of the epidural drug concentra-
carried out to optimize drug delivery. tion was made by estimation of the time course of systemic drug
Li et al. performed a PKPD analysis of dental pain relief by ibu- absorption from the epidural space (by measurement of the drug
profen [30] . Their aim was to assess the efficacy of a novel effervescent concentration after an epidural injection). Simultaneously, they
formulation in comparison with the standard tablets of NSAIDs. assessed the level of sensory blockade over time. Tested drugs were
In a group of patients following third-molar extraction surgery, the the local anesthetics levobupivacaine and ropivacaine. The epidural
effect of the two formulations and placebo were modeled using an segments were modeled by central and peripheral absorption com-
Emax pain relief model and hazard models to analyze time to pain partment- and effect-sites. Differences were observed in the onset/
relief and remedication. The data show differences in PK parameters offset half-lives between the two drugs (15 min for levobupivacaine
between the two formulations with a greater oral absorption rate and 25 min for ropivacaine). Age was a significant modifier with
for the effervescent ibuprofen and greater effect-site concentrations respect to onset/offset half-lives (increasing age reduced t½ for lev-
between t = 0 and 2.5 h following dosing. This caused more rapid obupivacaine at all segments) and anesthetic potency (increasing
pain relief and less remedication for the effervescent formulation age increased the potency at segments Th12 and higher for rop-
compared with standard ibuprofen. The authors constructed nomo- ivicaine). The model enables individualized dosing depending on
grams to correlate time to pain relief with PK profiles. Overall, these age, sex and drug, and gives a predictive indication of the time
data indicate the usefulness of PKPD modeling in the development course of the sensory blockade (Figure 2) .
of new formulations of analgesic drugs (see later).
PKPD modeling in chronic pain
Acetaminophen Ketamine
Another example where model-based analysis was used to aid the In a reanalysis of a study on the long-term (12 week) effects
development of a new formulation is the study by Green et al. [31] . of a 100-h S-ketamine infusion on pain relief in chronic pain
They performed PKPD analysis on pain relief data following patients owing to complex regional pain syndrome type 1, Dahan
administration of four different acetaminophen (paracetamol) et al. used an inhibitory sigmoid effect-compartment model of
formulations (Tylenol®, Panadol® Rapid and two that used new the form: PAIN score = BASELINE PAIN/(1 + [CE /C50]g), and
formulation technologies from Imaginot Pty, Ltd., Brisbane, onset/offset rate constant k [34] . This very simple model satisfac-
Australia) to explore possible differences in PD outcomes. The torily described the 12-week pain relief data with a value for C50
authors used the PKPD data from Anderson et al. to do a simu- of 10 ng/ml and t½k of 10.9 days (95% CI: 5–21 days). The C50
lation on the effects of the new formulation [32] . PK data were for relief of chronic (neuropathic) pain is approximately 50-times
obtained from volunteers that received a single dose of 1 g of the lower than that observed for the relief of acute nociceptive pain (in
four formulations; PD data were obtained from simulations on an volunteers) and suggests a different mode of action of S-ketamine
earlier published PD model that quantified pain relief following in these two distinct pain states. This an example derived from
tonsillectomy. Interestingly, their data indicate that the placebo chronic pain data of the power of PKPD modeling in understand-
effect was appreciable and there was a potential difference between ing the complex behavior of analgesic agents. The parameter t½k
acetaminophen and placebo at t = 0 and 30 min postdosing only. is not synonymous to the blood effect-site equilibration parameter
With respect to the comparison between formulations, the authors t½ke0. While the later parameter described the passage of the drug
conclude that their analysis suggests a significant reduction in to the postulated receptor site, parameter k reflects the dynamics
time to onset of effect with the newer formulations. The current of disease modulation by S-ketamine. The plasma concentration
approach of learning drug behavior from already published data- of S-ketamine drops rapidly below detection values (within hours)
sets was used ‘because running a confirmatory clinical trial was following the withdrawal from ketamine, while its analgesic effect
prohibitively expensive’ [31] . persists for weeks, probably owing to the initiation of a complex
cascade of events of which NMDAR desensitization is the first
Epidural anesthesia/analgesia step of this disease-modulatory process.
Epidural anesthesia and analgesia is obtained by an injection of a Assuming that PK of S-ketamine is irrelevant in describing
local anesthetic, an opioid or their combination into the epidural its prolonged analgesic effect in chronic pain, in addition a
space. The injected drugs spread over the epidural space, diffuse (population-based) time-series analysis was performed using a
towards the intrathecal space (where the drugs diffuse into and mixture model and autoregressive (Kalman) filter constructed in
bind to neuronal tissue) and into blood vessels of the epidural space. NONMEM. The model assumes an exponential return of pain
Depending on the concentration used and type of drug injected, values towards baseline pain with autoregressive factor F, which
motor and sensory blockade develops and dissipates. Onset/off- is comparable to parameter k. Indeed, the observed value of t½F
set times and spread of the effect across the spinal segments vary corresponded well with the range observed for k: 15–41 days.
and depend on the kinetic and dynamic properties of the injected The mixture model allows the estimation of the chance of an
anesthetic/opioid. Olofsen et al. are the first to develop a popula- analgesic effect to ketamine (or placebo) by objectively subdi-
tion PKPD model of epidural anesthesia and analgesia [33] . Their viding the patient population into four subgroups: two types of

724 Expert Rev. Clin. Pharmacol. 4(6), (2011)

 Pharmacokinetic–pharmacodynamic modeling in acute & chronic pain Review

nonresponders and two types of responders (Figure 3). This type of

analysis is attractive as it rapidly allows a subdivision of analgesic
responses into various subgroups that may be defined a priori T4
without the need for PK data. Mixture models using time-series T5
T6
analysis may be especially advantageous when quantifying the T7
analgesic efficacy of drugs in terms of onset/offset time, division T8
T9
into subgroups, magnitude of variability and placebo effect, or T10
when quantifying the effect of drugs that act locally, such as T11
local applications of lidocaine or capsaicine. T12
L1
L2
Pregabalin L3
L4
Pregabalin is an analog of GABA and acts at the a2d protein
L5
associated with voltage-gated calcium channels. It has anxiolytic, S1
antiepileptic and analgesic properties. Byon et al. performed a S2
S3
pregabalin exposure-response analysis in fibromyalgia chronic S4
pain patients [35] . They modeled the probability of an observed S5
specific daily pain score (on an 11-point categorical scale) using a 0 2 4 6 8
proportional odds logistic regression model in NONMEM with Time (h)
additive components: baseline pain, placebo response (exponential
function) and drug effect (Emax and linear models were tested). 1
The logistic regression approach is used when pain or pain relief
is scored with an ordered categorical scale and was first intro-
duced by Sheiner in 1994 [36] . Byon et al. incorporated data from 0.8
four separate randomized, placebo-controlled studies including
2758 patients with daily dosing for 8–14 weeks in the dose range 0.6
of 150–600 mg two- to three-times per day [35] . The average drug
P(block)

concentration was estimated from the pregabalin dose and renal

function. The Emax drug effect model gave the best results and 0.4
they estimated an EC50 of 1.4 µg/ml (equivalent to 174 mg/ day
in a patient with creatinine clearance of 100 ml/min), the delay in 0.2
drug effect had a t½ of 11 h. Sex and age were significant covari-
ates, with increasing values for Emax with increasing age and in men
compared with women. 0
0 2 4 6 8
Expert commentary & five-year view Time (h)
In the last 5–6 years, only a limited number of studies have been
published on population PKPD modeling of analgesic agents used Figure 2. Population pharmacokinetic–pharmacodynamic
in acute and/or chronic pain. The majority of studies focused on modeling of epidural analgesia. (A) Blockade probabilities
older analgesics, including morphine, buprenorphine and keta- after an epidural injection of ropivacaine 125 mg in a 50-year-old
patient. The size of the dots is proportional to the probability of
mine, while studies on newer agents were scarce and limited to
blockade with 90% (solid line), 75% (broken line) and 50%
morphine’s metabolite M6G, the experimental NSAID SC-75416 (dotted line) isoeffect lines. Time is expressed on the x-axis and
and pregabalin. We believe that population PKPD modeling is the dermatome level on the y-axis. (B) Probability of blockade,
an important tool, not only in the description of the time course P(block), for dermatomes S5 (broken line), Th11 (solid line) and
of drug effect (which enables the appropriate use of the drug), Th7 (dotted line) versus time.
Data taken from [32] .
but equally important to understand the interaction of a specific
pharmacological agent within the often complex disease processes
and possibly even understand the complexities of the disease itself. development of new formulations or the switch towards new indi-
Furthermore, model-based analysis using a population PKPD cations of already registered analgesics. Particularly in the develop-
approach is useful in the development of new formulations of ment of new analgesic agents, population PKPD modeling enables
older drugs such as was shown for ibuprofen [29,30] and acetami- simultaneous assessment of the various properties of analgesics
nophen [31] . Simulation-based PKPD modeling may be a cheap by calculation of utility or safety functions. There are various
but valid alternative to expensive large clinical trials [31] . possible calculations of the therapeutic utility functions possible
In the upcoming 5 years, we foresee various advances in the such as the probability of analgesic effect minus the probability
use of population PKPD models in acute and chronic pain, most of side effect. In an animal study, Yassen et al. performed a logis-
importantly in the development of new analgesic agents, in the tic regression analysis to characterize the relationship between

www.expert-reviews.com 725
 Review Martini, Olofsen, Yassen, Aarts & Dahan

further a priori descriptions with patients

that experience at least 50% pain relief
10 10
Subgroup 1 Subgroup 2 throughout the treatment period as most
8 8
clinically relevant. Apart from drug effect
per se, the variability within the different
groups is an important marker of drug
Pain score

Pain score
6 6
efficacy. Moreover, patients in the differ-
4 4 ent response groups may be further exam-
ined to link specific patient characteristics
2 2 (such as parameters derived from quanti-
tative sensory testing, duration of disease,
0 0 severity of pain symptoms and variability in
0 2 4 6 8 10 12 0 2 4 6 8 10 12 pain symptoms) to drug efficacy, eventually
Time (weeks) Time (weeks) allowing prediction of drug response based
entirely on these pretreatment assessments
10 10 [39] . We predict an important place for mix-
Subgroup 3 Subgroup 4
ture modeling in NONMEM in the devel-
8 8
opment of new drugs or new indications of
already registered drugs.
Pain score

Pain score

6 6
The chronification of pain is complex
and poorly understood. Both peripheral
4 4
and central processes play important roles,
2 2
although the specifics vary per disease. To
fully understand the effect of analgesics on
0 0
these processes, disease-modulatory PKPD
0 2 4 6 8 10 12 0 2 4 6 8 10 12 models are required that describe the effects
Time (weeks) Time (weeks) of specific analgesics within the various
pain pathways [31] . These models need to
Figure 3. Mixture analyses of the effect of a 1-week ketamine treatment
be dynamic as processes vary over time. For
(shaded bar) on pain scored in patients with chronic pain. The analysis objectively example, it may well be that chronic pain
divided the total population response into four categories, with (A) no response to in its initial phase has predominant periph-
treatment, (B) a response in the treatment week only, (C) a long-term response with a eral inflammatory components, while over
slow return towards pretreatment baseline and (D) full recovery. time structural plastic changes within the
Data taken from [33] .
spinal cord or at supraspinal sites develop.
drug exposure and two effects of buprenorphine and fentanyl: Abandanes et al. use PET to predict brain target occupancy of
antinociception and respiratory depression [37] . Odds ratios for duloxetine, a serotonin-reuptake inhibitor with analgesic proper-
both end points were calculated and were, for buprenorphine, 29 ties, with various dosing regimens, using a mechanism-based PK
and 2 for antinociception and respiratory depression, respectively, brain target occupancy models [40] . They define 50% receptor
while for fentanyl these values were 3 and 2.5. The safety index, occupation (OC50) as the drug plasma concentration that achieves
calculated as the odds ratio for analgesia divided by the odds ratio 50% receptor occupation with OC50 = koff/kon. Applying these
for a side effect of 1.2 for fentanyl and 14 for buprenorphine, sug- techniques to specific pharmacological targets of chronic pain
gests a distinct margin of safety for these two opioids, with greater within the different timeframes of the lifecycle of the disease and
safety for buprenorphine compared with fentanyl. Respiratory linking receptor occupation to pain relief is an example of the
depression, as a complication of potent opioid analgesics, is seldom building of disease-modulatory PKPD models. Such models will
studied, while respiratory events are potentially life-threatening increase our insight into the disease process (such as the process of
side effects of these agents [38] . Hence, we encourage the industry chronifcation of pain) and will improve treatment development.
to determine the safety index in the development of new analgesic
agents. Particularly when developing potent analgesic agents or Financial & competing interests disclosure
new formulations (such as for fentanyl and sufentanil), PKPD Ashraf Yassen is an employee of Astellas Pharma Europe BV, Leiderdorp, The
studies on opioid side effects (including respiratory depression, Netherlands. Astellas Pharma Europe BV has active research programs in
nausea/vomiting, delayed gastric emptying and sedation) should clinical and preclinical discovery for the treatment of chronic pain. The authors
be part of the preregistration study process. have no other relevant affiliations or financial involvement with any organiza-
Another advancement would be the use of PD mixture mod- tion or entity with a financial interest in or financial conflict with the subject
els in NONMEM to predefine response groups into responders matter or materials discussed in the manuscript apart from those disclosed.
and nonresponders. Both responders and nonresponders require No writing assistance was utilized in the production of this manuscript.

726 Expert Rev. Clin. Pharmacol. 4(6), (2011)

 Pharmacokinetic–pharmacodynamic modeling in acute & chronic pain Review

Key issues
Population pharmacokinetic–pharmacodynamic modeling of analgesics enables:
• Estimation of the population pharmacokinetic and pharmacodynamic parameters, including potency and onset/offset parameters.
• Identification of the sources of variability that influence pharmacokinetics and dynamics.
• Estimation of the magnitude of between-subject variability as well as intrasubject variability.
• Individualized dosing.
• Predictive assessment of the time course of drug effect.
• Provision of aid in the development of new formulations of analgesic compounds.
Unmet needs include:
• Development of disease-modulation-oriented pharmacokinetic–pharmacodynamic modeling models in acute and chronic pain.
• Systematic application of mixture models in the development of novel analgesic agents to identify specific response subgroups,
allowing optimization of therapy and outcome.
• Extended use of utility functions in the development of novel analgesics by performing simultaneous pharmacokinetic–pharmacodynamic
modeling of effect (analgesia) and side effects (toxicity).

buprenorphine and fentanyl in humans 13 Sarton E, Olofsen E, Romberg R et al.

References and are able to separate distribution Sex differences in morphine analgesia: an
Papers of special note have been highlighted as: experimental study in healthy volunteers.
kinetics to the postulated effect-site
• of interest
(defined by first-order rate constant ke0) Anesthesiology 93, 1245–1254 (2000).
•• of considerable interest
from receptor association–dissociation •• First population PKPD analysis on the
1 Danhof M, de Jonghe J, deLange ECM,
kinetics (described by rate constants kon differential antinociceptive effect of
della Pasqua O, Ploeger BA, Voskuyl RA.
and koff ) for buprenorphine. By contrast, intravenous morphine in men and women
Mechanism-based pharmacokinetic–
fentanyl’s delay can only be described in demonstrating both differences in potency
pharmacodynamic modeling: biophase
distribution, receptor theory, and terms of ke0. and onset/offset times.
dynamical systems ana­lysis. Annu. Rev. 7 Miller R. Population pharmacokinetics. 14 Niesters M, Dahan A, Kest B et al. Do
Pharmacol. Toxicol. 47, 357–400 (2007). In: Principles of Clinical Pharmacology. sex differences exist in opioid analgesia? A
2 Minto CF, Schnider TW. Contributions of Atkinson AJ, Daniels CE, Dedrick RL, systematic review and meta-analysis of
PK/PD modeling to intravenous anesthesia. Grudzinakis CV, Markey SP (Eds). human experimental and clinical studies.
Clin. Pharmacol. Ther. 84, 37–38 (2008). Academic Press, CA, USA, 113–119 Pain 151, 61–68 (2010).
(2001). 15 Dahan A, van Dorp E, Smith TW,
3 Segre G. Kinetics of interaction between
drugs and biological systems. Farmaco. Sci. 8 Heeremans EH, Proost JH, Eleveld DJ, Yassen A. Morphine-6-glucuronide (M6G)
23(10), 907–918 (1968). Absalom AR, Struys MM. Population for postoperative pain relief. Eur. J. Pain
pharmacokinetics and pharmacodynamics 12, 403–411 (2008).
• First proposal of a first-order rate constant in anesthesia, intensive care and pain
to describe the delay between drug plasma 16 Romberg R, Olofsen E, Sarton E,
medicine. Curr. Opion. Anaesthesiol. 23, den Hartigh J, Taschner PE, Dahan A.
concentration and effect, here applied to 479–484 (2010). Pharmacokinetic–pharmacodynamic
epinephrine’s effect on blood pressure
9 Beal BL, Sheiner LB, Boeckman AJ, modeling of morphine-6-glucuronide-
in the cat. Bauer RJ. NONMEM User’s Guide. Icon induced analgesia in healthy volunteers:
4 Hull CJ. Pharmacokinetics and Development Solutions, MD, USA absence of sex differences. Anesthesiology
pharmacodynamic. Br. J. Anaesth. 51, (1989–2009). 100, 120–133 (2004).
579–594 (1979). Shafer SL, Struys MMRF. Mixed effect
10 17 Olofsen E, van Dorp E, Teppema L et al.
5 Sheiner L, Stanski LB, Vozeh S, Miller RD, modeling in analgesia trials. Anesth. Analg. Naloxone reversal of morphine- and
Ham J. Simultaneous modeling of 107(1), 9–10 (2008). morphione-6-glucuronide-induced
pharmacokinetics and pharmacodynamics: 11 Bekkering GE, Soares-Weiser K, Reid K respiratory depression in healthy volunteers:
Application to d-tubocurarine. Clin. et al. Can morphine still be considered to a mechanism-based pharmacokinetic–
Pharmacol. Ther. 25, 358–371 (1979). be the standard for treating chronic pain? pharmacodynamic modeling study.
6 Yassen A, Olofsen E, Romberg R et al. A systematic review including pair-wise and Anesthesiology 112, 1417–1427 (2010).
Mechanism-based PK–PD modeling of the network meta-analyses. Curr. Med. Res. 18 Garfield, E. Current Comments: The 1982
respiratory depressant effect of Opin. 27(7), 1477–1491 (2011). John Scott Award goes to Jack Fishman
buprenorphine and fentanyl in healthy 12 Lötsch J. Pharmacokinetic– and Harold Blumberg for synthesis and
volunteers. Clin. Pharmacol. Ther. 81, pharmacodynamic modeling of opioids. investigation of naloxone. Essays of an
50–58 (2007). J. Pain Sympt. Manag. 29(5 Suppl.), Information Scientist 6, 121–130 (1983).
•• Using a mechanism-based S90–S103 (2005). 19 Dahan A, Aarts L, Smith TW. Incidence,
pharmacokinetic–pharmacodynamic •• Excellent article on the pharmacodynamic reversal, and prevention of opioid-induced
(PKPD) model, the authors quantify the properties of currently used respiratory depression. Anesthesiology
respiratory depressant effect of opioid analgesics. 112(1), 226–238 (2010).

www.expert-reviews.com 727
 Review Martini, Olofsen, Yassen, Aarts & Dahan

20 Mazoit JX, Butscher K, Samii K. Morphine buprenorphine in healthy volunteers. 32 Anderson BJ, Holford NH, Woollard GA,
in postoperative patients: pharmacokinetics Anesthesiology 104, 1232–1242 (2006). Kanagasundaram S, Mahadevan M.
and pharmacodynamics of metabolites. 25 Noppers I, Niesters M, Aarts L, Smith T, Perioperative pharmacodynamics of
Anesth. Analg. 105(1), 70–78 (2007). Sraton E, Dahan A. Ketamine for acetaminophen analgesia in children.
• Population PKPD analysis on the treatment of chonic non-cancer pain. Anesthesiology 90(2), 411–421 (1999).
contribution of morphine-6-glucoronide Expert Opin. Pharmacother. 11(14), 33 Olofsen E, Burm AG, Simon MJ,
and morphine-3-glucoronide to 2417–2429 (2010). Veering BT, van Kleef JW, Dahan A.
postoperative morphine analgesia showing 26 Herd DW, Anderson BK, Keene NA, Population pharmacokinetic–
greater morphine-6-glucoronide than Holford NK. Investigating pharmacodynamic modeling of epidural
morphine potency and an antagonistic pharmacodynamics of ketamine in children. anesthesia. Anesthesiology 109(4), 664–674
effect of morphine-3-glucoronide. Paediatr. Anaesth. 18(1), 36–42 (2008). (2008).

21 Abou Hammoud H, Simon N, Urien S, 27 Herd DW, Anderson BK, Holford NK. 34 Dahan A, Olofsen E, Sigtermans M et al.
Riou B, Lechat P, Aubrun F. Intravenous Modeling the norketamine metabolite in Population pharmacokinetic–
morphine titration in immediate children and the implications for analgesia. pharmacodynamic modeling of ketamine-
postoperative pain management: population Paediatr. Anesth. 17(9), 831–840 (2007). induced pain relief of chronic pain. Eur.
kinetic–pharmacodynamic and logistic J. Pain 5(3), 258–267 (2011).
28 Sigtermans M, Dahan A, Mooren R et al.
regression ana­lysis. Pain 144, 139–146 S(+)-ketamine effect on experimental pain 35 Byon W, Ouellet D, Chew M et al.
(2009). and cardiac output: a population Exposure-response analyses of the effects of
pharmacokinetic–pharmacodynamic pregabalin in patients with fibromyalgia
• Population PD analysis of morphine-
modeling study in healthy volunteers. using daily pain scores and patient global
induced pain relief in postoperative
Anesthesiology 111, 892–903 (2009). impression of change. J. Clin. Pharmacol.
patients using an indirect response model.
50, 803–815 (2010).
22 Yassen A, Olofsen E, Kan J, Dahan A, 29 Kowalski KG, Olson S, Remmers AE,
Hutmacher MM. Modeling and simulation 36 Sheiner LB. A new approach to the analysis
Danhof M. Animal-to-human
to support dose selection and clinical of analgesic drug trials illustrated with
extrapolation of the pharmacokinetic and
development of SC-75416, a selective bromfenac data. Clin. Pharmacol. Ther.
pharmacodynamic properties of
COX-2 inhibitor for the treatment of acute 56(3), 309–322 (1994).
buprenorphine. Clin. Pharmacokinet.
46(5), 433–447 (2007). and chronic pain. Clin. Pharmacol. Ther. 37 Yassen A, Olofsen E, Kan J, Dahan A,
83, 857–866 (2008). Danhof M. Pharmacokinetic–
23 Yassen A, Olofsen E, Romberg R et al.
30 Li H, Mandema J, Wada R et al. Modeling pharmacodynamic modeling of the
Mechanism-based PK/PD modeling of the
the onset and offset of dental pain relief by effectiveness and safety of buprenorphine
respiratory depressant effect of
ibuprofen. J. Clin. Pharmacol. DOI: and fentanyl in rats. Pharm. Res. 25(1),
buprenorphine and fentanyl in healthy
10.1177/0091270010389470 (2011) 183–193 (2008).
volunteers. Clin. Pharmacol. Ther. 81,
50–58 (2007). (Epub ahead of print). 38 Okie S. A flood of opioids, a rising tide of
31 Green B, Chandler S, MacDonald G, deaths. N. Eng. J. Med. 363(21),
•• Mechanism-based population PKPD 1981–1983 (2010).
analysis on the respiratory effect of Elliott G, Roberts MS. Quantifying pain
relief following administration of a novel 39 Von Korff M, Dunn KM. Chronic pain
burprenorphine where both t½ke0 and
formulation of paracetamol reconsidered. Pain 138, 267–276 (2008).
receptor kinetics were successfully
(acetaminophen). J. Clin. Pharmacol. 50, 40 Abandanes S, van der Aart J, Barletta JA
estimated showing that rate-limiting
1406–1413 (2010). et al. Prediction of repeat-dose occupancy
factors in the onset and offset of effect are
dependent on biophase distribution and • PKPD study to assess the feasibility of from single-dose data: characteristics of the
new paracetamol formulations using relationship between plasma
slow receptor association–dissociation.
pharmacokinetic data linked to a pharmacokinetics and brain target
24 Yassen A, Olofsen E, Romberg R, Sarton occupancy. J. Cereb. Blood Flow Metab. 31,
simulated pain relief dataset derived from
E, Danhof M, Dahan A. Mechanism-based 944–952 (2011).
an earlier published study. This approach
pharmacokinetic-pharmacodynamic
modeling of the antinociceptive effect of is a cheap but valid alternative to a large
clinical trial.

728 Expert Rev. Clin. Pharmacol. 4(6), (2011)