t

t
ne

ne

ne

ne
GENERAL PRINCIPLES SECTION 1 

e. 11 

e.

e.

e.
re

fre

fre

fre
f
ks

ks

ks

ks
Pharmacokinetics
oo

oo

oo

oo
eb

eb

eb

eb
m

m
t

t
ne

ne

ne

ne
combine. This underpins what is termed the target concentra-
OVERVIEW tion strategy. Individual variation in response to a given dose
e.

e.

e.

e.
of a drug is often greater than variability in the plasma
re

fre

fre

fre
We explain the importance of pharmacokinetic concentration at that dose. Plasma concentrations (Cp) are
analysis and present a simple approach to this topic. therefore useful in the early stages of drug development
sf

ks

ks

ks
We explain how drug clearance determines the (see later), and in the case of a few drugs plasma drug
k

steady-state plasma concentration during constant-rate concentrations are also used in routine clinical practice to
oo

oo

oo

oo
drug administration and how the characteristics of individualise dosage, to achieve the desired therapeutic
absorption and distribution (considered in Ch. 9) plus effect while minimising adverse effects in each individual
eb

eb

eb

eb
metabolism and excretion (considered in Ch. 10) patient – an approach known as therapeutic drug monitoring,
m

determine the time course of drug concentration in
blood plasma during and following drug administra-
tion. The effect of different dosing regimens on the
time course of drug concentration in plasma is
t
m
often abbreviated TDM. Table 11.1 shows examples of some
drugs where a therapeutic range of plasma concentrations
has been established, enabling TDM. Concentrations of
drug in other body fluids (e.g. urine,1 saliva, cerebrospinal

et
explained. Population pharmacokinetics is mentioned fluid, milk) may add useful information.
ne

ne

ne

n
briefly, and a final section considers limitations to Formal interpretation of pharmacokinetic data consists
e.

e.

e.

e.
the pharmacokinetic approach. of fitting concentration-versus-time data to a model (whether
abstract or, more usefully, physiologically based) and
fre

fre

fre

re
determining parameters that describe the observed behav-

sf
INTRODUCTION: DEFINITION AND USES iour. The parameters can then be used to adjust the dose
ks

ks

ks

k
OF PHARMACOKINETICS regimen to achieve a desired target plasma concentration,
oo

oo

oo

oo
The pharmacologically active concentration range is esti-
Pharmacokinetics is the branch of pharmacology dedicated mated from experiments on cells, tissues or laboratory
eb

eb

eb

eb
to determining the fate of chemical substances administered animals, and modified as data emerge from early human
to a living organism – ‘what the body does to the drug’. pharmacology trials, which often test single doses of the
m

In practice this involves the measurement and formal
interpretation of changes with time of drug and drug
metabolite concentrations in plasma, urine and sometimes
t
m
new drug administered to successive groups of volunteers
in progressively increasing doses - single ascending dose
(SAD) studies (Ch. 8). Some descriptive pharmacokinetic

t
other accessible regions of the body, in relation to dosing. characteristics can be estimated directly by inspecting the
ne

ne

ne

ne
It provides a framework for understanding what happens time course of drug concentration in plasma following
to a drug when given to an animal or human, where it dosing – important examples,2 illustrated more fully later,
e.

e.

e.

e.
goes in the body, and how quickly, that enables one to are the maximum plasma concentration following a given
fre

fre

fre

re
understand the effects that it produces. In contrast, phar- dose of a drug administered in a defined dosing form (Cmax)
macodynamics (‘what the drug does to the body’), describes and the time (Tmax) between drug administration and achiev- sf
ks

ks

ks

events consequent on interaction of the drug with its receptor ing Cmax. Other pharmacokinetic parameters are estimated
ok
or other primary site of action. The distinction is useful, mathematically from experimental data; examples include
oo

oo

oo

although the words cause dismay to etymological purists. volume of distribution (Vd) and clearance (CL), concepts that
eb

eb

eb

eb

‘Pharmacodynamic’ received an entry in a dictionary of have been introduced in Chapters 9 and 10, respectively,
1890 (‘relating to the powers or effects of drugs’) whereas and to which we return below. This approach applies both
m

pharmacokinetic studies only became possible in the latter
part of the 20th century with the development of sensitive,
specific and accurate physicochemical analytical techniques,
especially high-performance chromatography and mass
t
to classical low molecular-weight drugs and to macromo-
lecular biopharmaceuticals (Ch. 5), although qualitative
aspects of absorption, distribution and elimination are, of
course, very different and pharmacokinetic parameters

t
ne

ne

ne

ne

spectrometry, for measuring drug concentrations in biologi- differ markedly – for example, antibodies have evolved to
cal fluids. The time course of drug concentration following persist for long periods after exposure to antigen, and
e.

e.

e.

e.

dosing depends on the processes of absorption, distribution, therapeutic antibodies commonly have low rates of clearance
fre

fre

fre

fre

metabolism and excretion (ADME) that we have considered and long elimination half-lives in consequence.
qualitatively in Chapters 9 and 10.
ks

ks

ks

ks

In practice, pharmacokinetics usually focuses on con-

centrations of drug in blood plasma, which is easily sampled
oo

oo

oo

oo

1
via venepuncture, since plasma concentrations are assumed Clinical pharmacology became at one time so associated with the
measurement of drugs in urine that the canard had it that clinical
usually to bear a clear relation to the concentration of drug
eb

eb

eb

eb

pharmacologists were the new alchemists – they turned urine into

in extracellular fluid surrounding cells that express the airline tickets.
receptors or other targets with which drug molecules 143
m

2
Important because dose-related adverse effects often occur around Cmax.
t

t
 t

t
11

ne

ne

ne

ne
SECTION 1    General Principles

e.

e.

e.

e.
re

fre

fre

fre
• situations where this model is inadequate, and
f Table 11.1  Examples of drugs where therapeutic introduce a two-compartment model;
ks

ks

ks

ks
drug monitoring (TDM) of plasma concentrations is • situations where clearance varies with drug
used clinically
oo

oo

oo

oo
concentration (‘non-linear kinetics’);
Category Example(s) See chapter • situations (such as paediatric pharmacokinetics) where
eb

eb

eb

eb
only a few samples are available and population
Immunosuppressants Ciclosporin, 27 kinetics may be used.
m

m
tacrolimus
Cardiovascular Digoxin 22 Finally, we consider some of the limitations inherent in
the pharmacokinetic approach. More detailed accounts are
Respiratory Theophylline 17, 29 provided by Atkinson et al. (2012), Birkett (2010) and
t

t
ne

ne

ne

ne
CNS Lithium, phenytoin 48, 46 Rowland and Tozer (2010).
e.

e.

e.

e.
Antibacterials Aminoglycosides 52
re

fre

fre

fre
Anticancer drugs Methotrexate 57 DRUG ELIMINATION EXPRESSED AS
CLEARANCE
sf

ks

ks

ks
k

The overall clearance of a drug by all routes (CLtot) is the

oo

oo

oo

oo
fundamental pharmacokinetic parameter describing drug
USES OF PHARMACOKINETICS elimination. It is defined as the volume of plasma which
eb

eb

eb

eb
Knowledge of the pharmacokinetic behaviour of drugs in contains the total amount of drug that is removed from
m

animals and man is crucial in drug development, both to
make sense of preclinical toxicological and pharmacological
data3 and to decide on an appropriate dose and dosing
regimen for clinical trials (see Ch. 60). Drug regulators
t
m
the body in unit time. It is thus expressed as volume per
unit time, e.g. mL/min or L/h. Renal clearance (CLren), an
important component of CLtot, was described in Chapter
10.

et
have developed concepts such as bioavailability and bio- The overall clearance of a drug (CLtot) is the sum of
ne

ne

ne

n
equivalence (Ch. 9) to support the licensing of generic versions clearance rates for each mechanism involved in eliminating
e.

e.

e.

e.
of drugs produced when originator products lose patent the drug, usually renal clearance (CLren) and metabolic
protection. Understanding the general principles of phar- clearance (CLmet) plus any additional appreciable routes of
fre

fre

fre

re
macokinetics is also important in clinical practice, to elimination (faeces, breath, etc.). It relates the rate of elimina-

sf
understand the rationale of recommended dosing regimens, tion of a drug (in units of mass/unit time) to the plasma
ks

ks

ks

k
to interpret drug concentrations for TDM and to adjust concentration, Cp:
oo

oo

oo

oo
dose regimens rationally, and to identify and evaluate
possible drug interactions (see Chs 9 and 10). In particular, Rate of drug elimination = Cp × CLtot (11.1)
eb

eb

eb

eb
intensive-care specialists and anaesthetists dealing with a
severely ill patient often need to individualise the dose Drug clearance can be determined in an individual subject
m

regimen depending on the urgency of achieving a thera-
peutic plasma concentration, and whether the pharmacoki-
netic behaviour of the drug is likely to be affected by illness
t
m
by measuring the plasma concentration of the drug (in
units of, say, mg/L) at intervals during a constant-rate
intravenous infusion (delivering, say, X mg of drug per

t
such as renal impairment or liver disease. h), until a steady state is approximated (Fig. 11.1A). At
ne

ne

ne

ne
steady state, the rate of input to the body is equal to the
SCOPE OF THIS CHAPTER rate of elimination, so:
e.

e.

e.

e.
We describe:
fre

fre

fre

re
X = Css × CLtot (11.2)
• how total drug clearance determines steady-state sf
ks

ks

ks

plasma concentration during continuous

Rearranging this,
ok
administration;
oo

oo

oo

• how drug concentration versus time can be described X

o

by a simple model in which the body is represented as CLtot = (11.3)

eb

eb

eb

eb

a single well-stirred compartment, of volume Vd. This Css

describes the situation before steady state (or after
m

where CSS is the plasma concentration at steady state, and

drug is discontinued) in terms of elimination
CLtot is in units of volume/time (L/h in the example given).
half-life (t1/2);
For many drugs, clearance in an individual subject is
independent of dose (at least within the range of doses
t

t
ne

ne

ne

ne

used therapeutically – but see the section on saturation

kinetics later for exceptions), so knowing the clearance
e.

e.

e.

e.

3
For example, doses used in experimental animals often need to be enables one to calculate the dose rate needed to achieve a
much greater than those in humans (on a ‘per unit body weight’ basis),
fre

fre

fre

fre

desired steady-state (‘target’) plasma concentration from

because drug metabolism is commonly much more rapid in rodents
– methadone (Ch. 43) is one of many such examples. When using Eq. 11.2.
ks

ks

ks

ks

animal data to estimate a ‘human equivalent dose’ in planning the CLtot can also be estimated by measuring plasma con-
first-in-human study, doses of low molecular-weight drugs are centrations at intervals following a single intravenous bolus
oo

oo

oo

oo

normalised (so-called ‘allometric scaling’) to estimated body surface dose of, say, Q mg (Fig. 11.1B):
area rather than to body weight. Paediatricians commonly use the same
eb

eb

eb

eb

approach, estimating appropriate doses for babies and young children Q

from adult human doses in terms of dose/unit of estimated body CLtot = (11.4)
144 AUC0 −∞
m

surface area rather than dose/kg body weight.

t

t
 t

t
11

ne

ne

ne

ne
Pharmacokinetics

e.

e.

e.

e.
re

fre

fre

fre
f A
ks

ks

ks

ks
Dose, Q
oo

oo

oo

oo
CSS (oral)
CSS=X/CL
eb

eb

eb

eb
concentration

Absorption kabs
Plasma
m

m
Dose, Q Volume Vd
(intravenous)
0
t

t
Single well-stirred
ne

ne

ne

ne
Time compartment
(Infusion) X mg/h)
e.

e.

e.

e.
B kexc kmet
re

fre

fre

fre
100
sf

Excretion Metabolism
ks

ks

ks
concentration
k

Fig. 11.2  Single-compartment pharmacokinetic model.

oo

oo

oo

oo
Plasma

50 This model is applicable if the plasma concentration falls

exponentially after drug administration (as in Fig. 11.1).
eb

eb

eb

eb
m

m
0
Bolus (Q mg) Time effects in patients with disordered kidney or liver function
may differ from those observed in healthy volunteer subjects.
t

et
In early phase clinical trials these measures of drug exposure
ne

ne

ne
C

n
100 are determined at each dose level and the protocol includes
e.

e.

e.

e.
C0 ‘stopping rules’ to avoid dose increments that caused toxicity
Plasma concentration

Vd=Q/C0 during animal experiments. See Chapter 9, and Birkett,

fre

fre

fre

re
2010 for a fuller account.
(log scale)

sf
Note that these estimates of CLtot, unlike estimates based
ks

ks

ks
10

k
on the rate constant or half-life (see later), do not depend
oo

oo

oo

oo
on any particular compartmental model.
eb

eb

eb

eb
1
Bolus (Q mg) Time SINGLE-COMPARTMENT MODEL
m

m
Fig. 11.1  Plasma drug concentration–time curves.
(A) During a constant intravenous infusion at rate X mg/h,
Consider a highly-simplified model of a human being, which
indicated by the horizontal bar, the plasma concentration (C)
consists of a single well-stirred compartment, of volume
t

t
Vd (distribution volume), into which a quantity of drug Q
ne

ne

ne

increases from zero to a steady-state value (CSS); when the
infusion is stopped, C declines to zero. (B) Following an
e.
ne
is introduced rapidly by intravenous injection, and from
which it is removed either by being metabolised or by

e.

e.

e.
intravenous bolus dose (Q mg), the plasma concentration rises
abruptly and then declines towards zero. (C) Data from panel (B) being excreted (Fig. 11.2). For most drugs, Vd is an apparent
fre

fre

fre

re
plotted with plasma concentrations on a logarithmic scale. The volume rather than the volume of an anatomical compart-
straight line shows that concentration declines exponentially. ment. It links the total amount of drug in the body to its sf
ks

ks

ks

Extrapolation back to the ordinate at zero time gives an estimate concentration in plasma (see Ch. 9). The quantity of drug
ok
in the body immediately after it is administered as a single
oo

oo

oo

of C0, the concentration at zero time, and hence of Vd, the

o

volume of distribution. bolus is equal to the administered dose Q. The initial

eb

eb

eb

eb

concentration, C0, will therefore be given by:

m

Q
C0 = (11.5)
where AUC0–∞ is the area under the full curve4 relating Cp Vd
to time following a bolus dose given at time t = 0. AUC0–∞
provides an integrated measure of tissue exposure to the In practice, C0 is estimated by extrapolating the linear portion
t

t
ne

ne

ne

ne

drug in units of time multiplied by drug concentration. of a semilogarithmic plot of Cp against time back to its
Together with Cmax it informs as to drug effects, both desired intercept at time 0 (Fig. 11.1C). Cp at any time depends on
e.

e.

e.

e.

and toxic, and so is important in anticipating possible effects the rate of elimination of the drug (i.e. on its total clearance,
fre

fre

fre

fre

in humans from those observed during animal pharmacol- CLtot) as well as on the dose and Vd. Many drugs exhibit
ogy and toxicology experiments, and in anticipating how first-order kinetics, where the rate of elimination is directly
ks

ks

ks

ks

proportional to drug concentration. (An analogy is letting

your bath drain down the plug hole where the water,
oo

oo

oo

oo

analogous to drug, initially rushes out whereas the last bit

4
The area is obtained by integrating from time = 0 to time = ∞, and is
always takes an age to drain away. Contrast this with
eb

eb

eb

eb

designated AUC0–∞. The area under the curve has units of time – on the
abscissa – multiplied by concentration (mass/volume) – on the so-called zero-order kinetics where the water is pumped out
of the bath at a constant rate.) With first-order kinetics 145
m

ordinate; so CL = Q/AUC0–∞ has units of volume/time as it should.

t

t
 t

t
11

ne

ne

ne

ne
SECTION 1    General Principles

e.

e.

e.

e.
re

fre

fre

fre
of 1/time. It represents the fraction of drug in the body
f A 10 eliminated per unit of time. For example, if the rate constant
ks

ks

ks

ks
is 0.1/h this implies that one-tenth of the drug remaining
Plasma concentration (µmol/L)
oo

oo

oo

oo
(kel = 0.2/h) in the body is eliminated each hour.
8
The elimination half-life, t1/2, is the time taken for Cp to
eb

eb

eb

eb
decrease by half, and is equal to ln2/kel (=0.693/kel). The
6 plasma half-life is therefore determined by Vd as well as
m

m
by CLtot. It enables one to predict the time course of Cp after
a bolus of drug is given or after the start or end of an
4 b infusion, when Cp is rising to its steady-state level or declin-
(kel = 0.05/h)
ing to zero.
t

t
ne

ne

ne

ne
(kel = 0.05/h) When a single-compartment model is applicable, the
2 b´
drug concentration in plasma approaches the steady-state
e.

e.

e.

e.
a value approximately exponentially during a constant infu-
re

fre

fre

fre
0 sion (see Fig. 11.1A). When the infusion is discontinued,
0 3.5 13.9 25 50 the concentration falls exponentially towards zero with the
sf

ks

ks

ks
Hours same half-life: after one half-life, the concentration will have
k

fallen to half the initial concentration; after two half-lives,

oo

oo

oo

oo
t value
for a
it will have fallen to one-quarter the initial concentration;
t value after three half-lives, to one-eighth; and so on. It is intuitively
eb

eb

eb

eb
for b & b´ obvious that the longer the half-life, the longer the drug
m

m
will persist in the body after dosing is discontinued. It is
B 10 less obvious, but nonetheless true, that during chronic drug
Plasma concentration (µmol/L, log scale)

administration, the longer the half-life, the longer it will

take for the drug to accumulate to its steady-state level:
t

et
5
one half-life to reach 50% of the steady-state value, two to
ne

ne

ne

n
b reach 75%, three to reach 87.5% and so on. This is extremely
e.

e.

e.

e.
helpful to a clinician deciding how to start treatment. If the
2 b´ drug in question has a half-life of approximately 24 h, for
fre

fre

fre

re
example, it will take 3–5 days to approximate the steady-

sf
1 state concentration during a constant-rate infusion. If this
ks

ks

ks

k
is too slow in the face of the prevailing clinical situation, a
oo

oo

oo

oo
a loading dose may be used in order to achieve a therapeutic
0.5
concentration of drug in the plasma more rapidly (see later).
eb

eb

eb

eb
The size of such a dose is determined by the volume of
distribution (Eq. 11.5).
m

m
0.2

0 25 50 EFFECT OF REPEATED DOSING

Hours Drugs are usually given therapeutically as repeated doses
t

t
rather than single injections or a constant infusion. Repeated
ne

ne

ne

ne
Fig. 11.3  Predicted behaviour of single-compartment
model following intravenous drug administration at time 0.
injections (each of dose Q) give a more complicated pattern
than the smooth exponential rise during intravenous infusion,
e.

e.

e.

e.
Drugs a and b differ only in their elimination rate constant, kel.
Curve b′ shows the plasma concentration time course for a but the principle is the same (Fig. 11.4). The concentration
fre

fre

fre

re
smaller dose of b. Note that the half-life (t1/2) (indicated by will rise to a mean steady-state concentration with an
broken lines) does not depend on the dose. (A) Linear approximately exponential time course, but will oscillate
sf
ks

ks

ks

(through a range Q/Vd). The smaller and more frequent

ok
concentration scale. (B) Logarithmic concentration scale.
the doses, the more closely the situation approaches that
oo

oo

oo

of a continuous infusion, and the smaller the swings in

eb

eb

eb

eb

concentration. The exact dosage schedule, however, does

drug concentration decays exponentially (Fig. 11.3), being not affect the mean steady-state concentration, or the rate
m

described by the equation: at which it is approached. In practice, a steady state is

effectively achieved after three to five half-lives. Speedier
−CLtot attainment of the steady state can be achieved by starting
Ct = C0 exp t (11.6)
Vd with a larger dose, as mentioned earlier. Such a loading
t

t
ne

ne

ne

ne

dose is sometimes used when starting treatment with a drug

(Note that exp is another way of writing ‘e to the power with a half-life that is long in the context of the urgency
e.

e.

e.

e.

of’, so this has the same form as Ct= C0.e−kt.) of the clinical situation, as may be the case when treating
fre

fre

fre

fre

Taking logarithms to the base e (written as ln): cardiac dysrhythmias with drugs such as amiodarone
or digoxin (Ch. 22) or initiating anticoagulation with
−CLtot
ks

ks

ks

ks

ln Ct = ln C0 − t (11.7) heparin (Ch. 25).

Vd
oo

oo

oo

oo

EFFECT OF VARIATION IN RATE  

Plotting Ct on a logarithmic scale against t (on a linear OF ABSORPTION
eb

eb

eb

eb

scale) yields a straight line with slope −CLtot/Vd. The constant If a drug is absorbed slowly from the gut or from an injec-
146 CLtot/Vd is the elimination rate constant kel, which has units tion site into the plasma, it is (in terms of a compartmental
m

m
t

t
 t

t
11

ne

ne

ne

ne
Pharmacokinetics

e.

e.

e.

e.
re

fre

fre

fre
f
ks

ks

ks

ks
15 C
oo

oo

oo

oo
Plasma concentration (µmol/L)
B
eb

eb

eb

eb
10
m

m
A

Fig. 11.4  Predicted behaviour of single-

t

t
ne

ne

ne

ne
compartment model with continuous or 5
intermittent drug administration. Smooth
e.

e.

e.

e.
curve A shows the effect of continuous infusion
for 4 days; curve B the same total amount of Plasma
re

fre

fre

fre
drug given in eight equal doses; and curve C
sf

the same total amount of drug given in four 0

ks

ks

ks
k

equal doses. The drug has a half-life of 17 h

A Infusion at 200 µmol/day
oo

oo

oo

oo
and a volume of distribution of 20 L. Note that
in each case a steady state is effectively B Injection 100 µmol twice daily
eb

eb

eb

eb
reached after about 2 days (about three
half-lives), and that the mean concentration C Injection 200 µmol once daily
m

m
0 1 2 3
reached in the steady state is the same for all
three schedules. Days
t

et
ne

ne

ne

n
e.

e.

e.

e.
model) as though it were being slowly infused at a variable
fre

fre

fre

re
Pharmacokinetics rate into the bloodstream. For the purpose of kinetic model-

sf
ling, the transfer of drug from the site of administration to
ks

ks

ks
• Total clearance (CLtot) of a drug is the fundamental

k
the central compartment can be represented approximately
oo

oo

oo

oo
parameter describing its elimination: the rate of by a rate constant, kabs (see Fig. 11.2). This assumes that
elimination equals CLtot multiplied by plasma the rate of absorption is directly proportional, at any
eb

eb

eb

eb
concentration. moment, to the amount of drug still unabsorbed, which
• CLtot determines steady-state plasma concentration is at best a rough approximation to reality. The effect of
m

m
(CSS): CSS = rate of drug administration/CLtot. slow absorption on the time course of the rise and fall
• For many drugs, disappearance from the plasma of the plasma concentration is shown in Fig. 11.5. The
follows an approximately exponential time course. curves show the effect of spreading out the absorption
t

t
Such drugs can be described by a model where the of the same total amount of drug over different times.
ne

ne

ne

ne
body is treated as a single well-stirred compartment of In each case, the drug is absorbed completely, but the
volume Vd. Vd is an apparent volume linking the peak concentration appears later and is lower and less
e.

e.

e.

e.
amount of drug in the body at any time to the plasma sharp if absorption is slow. In the limiting case, a dosage
fre

fre

fre

re
concentration. form that releases drug at a constant rate as it traverses
• Elimination half-life (t1/2) is directly proportional to Vd the ileum (Ch. 9) approximates a constant-rate infusion. sf
ks

ks

ks

Once absorption is complete, the plasma concentration

ok
and inversely proportional to CLtot.
declines with the same half-time, irrespective of the rate of
oo

oo

oo

• With repeated dosage or sustained delivery of a drug,

o

the plasma concentration approaches a steady value

absorption.
eb

eb

eb

eb

within three to five plasma half-lives. ▼  For the kind of pharmacokinetic model discussed here, the area
• In urgent situations, a loading dose may be needed to under the plasma concentration–time curve (AUC) is directly pro-
m

portional to the total amount of drug introduced into the plasma

achieve therapeutic concentration rapidly.
compartment, irrespective of the rate at which it enters. Incomplete
• The loading dose (L) needed to achieve a desired initial absorption, or destruction by first-pass metabolism before the drug
plasma concentration Ctarget is determined by Vd: L = reaches the plasma compartment, reduces AUC after oral administration
t

t
Ctarget × Vd.
ne

ne

ne

ne

(see Ch. 9). Changes in the rate of absorption, however, do not affect
• A two-compartment model is often needed. In this AUC. Again, it is worth noting that provided absorption is complete,
e.

e.

e.

e.

case, the kinetics are biexponential. The two the relation between the rate of administration and the steady-state
components roughly represent the processes of plasma concentration (Eq. 11.3) is unaffected by kabs, although the
fre

fre

fre

fre

size of the oscillation of plasma concentration with each dose is reduced

transfer between plasma and tissues (α phase) and
if absorption is slowed.
ks

ks

ks

ks

elimination from the body (β phase).

• Some drugs show non-exponential ‘saturation’
oo

oo

oo

oo

kinetics, with important clinical consequences, MORE COMPLICATED KINETIC MODELS

especially a disproportionate increase in steady-state
eb

eb

eb

eb

plasma concentration when daily dose is increased. So far, we have considered a single-compartment pharma-
cokinetic model in which the rates of absorption, metabolism 147
m

m
t

t
 t

t
11

ne

ne

ne

ne
SECTION 1    General Principles

e.

e.

e.

e.
re

fre

fre

fre
f A B
ks

ks

ks

ks
Plasma aminophylline concentration (µmol/L)
oo

oo

oo

oo
10

Plasma concentration (arbitrary units)

eb

eb

eb

eb
t abs 20
m

m
0h
1h
3h
5 6h
Oral
t

t
ne

ne

ne

ne
10
e.

e.

e.

e.
Intravenous
re

fre

fre

fre
Effective conc.
sf

ks

ks

ks
0 0
k

0 8 16 24 0 2 4 6 8
oo

oo

oo

oo
Hours Hours
eb

eb

eb

eb
Fig. 11.5  The effect of slow drug absorption on plasma drug concentration. (A) Predicted behaviour of single-compartment model
with drug absorbed at different rates from the gut or an injection site. The elimination half-time is 6 h. The absorption half-times (t1/2 abs)
m

m
are marked on the diagram. (Zero indicates instantaneous absorption, corresponding to intravenous administration.) Note that the peak
plasma concentration is reduced and delayed by slow absorption, and the duration of action is somewhat increased. (B) Measurements of
plasma aminophylline concentration in humans following equal oral and intravenous doses. (Data from Swintowsky, J.V., 1956. J. Am.
Pharm. Assoc. 49, 395.)
t

et
ne

ne

ne

n
e.

e.

e.

e.
and excretion are all assumed to be directly proportional
fre

fre

fre

re
to the concentration of drug in the compartment from which

sf
transfer is occurring. This is a useful way to illustrate some
ks

ks

ks
Oral dose

k
basic principles but is clearly a physiological oversimplifica-
oo

oo

oo

oo
tion. The characteristics of different parts of the body, such
as brain, body fat and muscle, are quite different in terms Absorption kabs
eb

eb

eb

eb
of their blood supply, partition coefficient for drugs and
the permeability of their capillaries to drugs. These differ-
m

m
ences, which the single-compartment model ignores, can k12
Intravenous Central Peripheral
markedly affect the time courses of drug distribution and dose compartment k21 compartment
action, and much theoretical work has gone into the (1) (2)
t

t
mathematical analysis of more complex models (see
ne

ne

ne

ne
Atkinson et al., 2012; Rowland & Tozer, 2010). They are
kexc kmet
beyond the scope of this book, and perhaps also beyond
e.

e.

e.

e.
the limit of what is actually useful, for the experimental
fre

fre

fre

re
data on pharmacokinetic properties of drugs are seldom Excretion Metabolism
accurate or reproducible enough to enable complex models sf
ks

ks

ks

Fig. 11.6  Two-compartment pharmacokinetic model.

to be tested critically.
ok
The two-compartment model, which introduces a separate
oo

oo

oo

‘peripheral’ compartment to represent the tissues, in com-

eb

eb

eb

eb

munication with the ‘central’ plasma compartment, more

closely resembles the real situation without involving are plotted semilogarithmically (Fig. 11.7). If, as is often
m

excessive complications. the case, the transfer of drug between the central and
peripheral compartments is relatively fast compared with
TWO-COMPARTMENT MODEL the rate of elimination, then the fast phase (often called
The two-compartment model is a widely used approxima- the α phase) can be taken to represent the redistribution
t

t
ne

ne

ne

ne

tion in which the tissues are lumped together as a peripheral of the drug (i.e. drug molecules passing from plasma to
compartment. Drug molecules can enter and leave the tissues, thereby rapidly lowering the plasma concentra-
e.

e.

e.

e.

peripheral compartment only via the central compartment tion). The plasma concentration reached when the fast
fre

fre

fre

fre

(Fig. 11.6), which usually represents the plasma (or plasma phase is complete, but before appreciable elimination has
plus some extravascular space in the case of a few drugs occurred, allows a measure of the combined distribution
ks

ks

ks

ks

that distribute especially rapidly). The effect of adding a volumes of the two compartments; the half-time for the
second compartment to the model is to introduce a second slow phase (the β phase) provides an estimate of kel. If a
oo

oo

oo

oo

exponential component into the predicted time course of drug is rapidly metabolised or excreted, the α and β phases
the plasma concentration, so that it comprises a fast and are not well separated, and the calculation of separate
eb

eb

eb

eb

a slow phase. This pattern is often found experimentally, Vd and kel values for each phase is not straightforward.
148 and is most clearly revealed when the concentration data Problems also arise with drugs (e.g. very fat-soluble drugs)
m

m
t

t
 t

t
11

ne

ne

ne

ne
Pharmacokinetics

e.

e.

e.

e.
re

fre

fre

fre
f 5
ks

ks

ks

ks
Diazepam 105 µmol orally 20 Dose

Blood alcohol concentration (mmol/L)

administered
oo

oo

oo

oo
Dose
Plasma diazepam concentration (µmol/L)
eb

eb

eb

eb
(mmol/kg)
2
14.1
m

m
1.0 Slow phase 10
t? = 30 h 10.9
t

t
ne

ne

ne

ne
0.5
e.

e.

e.

e.
7.6
re

fre

fre

fre
Fast phase Absorption
t? = 1.2 h 4.3
sf

0
ks

ks

ks
0 60 90 120
k

0.2
oo

oo

oo

oo
Time after ingestion (minutes)
Fig. 11.8  Saturating kinetics of alcohol elimination in
eb

eb

eb

eb
0.1 humans. The blood alcohol concentration falls linearly rather
m

m
0 12 24 than exponentially, and the rate of fall does not vary with dose.
Hours (From Drew, G.C. et al., 1958. Br. Med. J. 2, 5103.)

Fig. 11.7  Kinetics of diazepam elimination in humans

following a single oral dose. The graph shows a
t

et
ne

ne

ne
semilogarithmic plot of plasma concentration versus time. The

n
experimental data (black symbols) follow a curve that becomes not show metabolic saturation. Another consequence
e.

e.

e.

e.
linear after about 8 h (slow phase). Plotting the deviation of the is that the relationship between dose and steady-state
plasma concentration is steep and unpredictable, and it
fre

fre

fre

re
early points (pink shaded area) from this line on the same
coordinates (red symbols) reveals the fast phase. This type of does not obey the proportionality rule implicit in Eq. 11.3

sf
for non-saturating drugs (see Fig. 49.6 for another example
ks

ks

ks
two-component decay is consistent with the two-compartment

k
model (see Fig. 11.6) and is obtained with many drugs. (Data related to ethanol). The maximum rate of metabolism sets
oo

oo

oo

oo
from Curry, S.H., 1980. Drug Disposition and Pharmacokinetics. a limit to the rate at which the drug can be administered;
Blackwell, Oxford.) if this rate is exceeded, the amount of drug in the body
eb

eb

eb

eb
will, in principle, increase indefinitely and never reach a
steady state (see Fig. 11.9). This does not actually happen,
m

m
because there is always some dependence of the rate of
for which it is unrealistic to lump all the peripheral tissues elimination on the plasma concentration (usually because
together. other, non-saturating metabolic pathways or renal excretion
t

t
contribute significantly at high concentrations). Nevertheless,
ne

ne

ne

ne
SATURATION KINETICS steady-state plasma concentrations of drugs of this kind
In the case of some drugs, including ethanol, phenytoin vary widely and unpredictably with dose. Similarly,
e.

e.

e.

e.
and salicylate, the time course of disappearance of drug variations in the rate of metabolism (e.g. through enzyme
fre

fre

fre

re
from the plasma does not follow the exponential or biex- induction) cause disproportionately large changes in the
ponential patterns shown in Figs 11.3 and 11.7 but is initially plasma concentration. These problems are well recognised sf
ks

ks

ks

linear (i.e. drug is removed at a constant rate that is for drugs such as phenytoin, an anticonvulsant for which
ok
independent of plasma concentration). This is often called plasma concentration needs to be closely controlled to
oo

oo

oo

zero-order kinetics to distinguish it from the usual first-order achieve an optimal clinical effect (see Ch. 46, Fig. 46.4).
eb

eb

eb

eb

kinetics that we have considered so far (these terms have Drugs showing saturation kinetics are less predictable in
their origin in chemical kinetic theory). Saturation kinetics clinical use than ones with first-order kinetics, so may be
m

is a better term, because it conveys the underlying mecha-
nism, namely that a carrier or enzyme saturates and so as
the concentration of drug substrate increases, the rate of
elimination approaches a constant value. Fig. 11.8 shows
t
rejected during drug development if a pharmacologically
similar candidate with first-order kinetics is available
(Ch. 60).
Clinical applications of pharmacokinetics are summarised

t
ne

ne

ne

ne

the example of ethanol. It can be seen that the rate of disap- in the clinical box.
pearance of ethanol from the plasma is constant at approxi-
e.

e.

e.

e.

mately 4 mmol/L per h, irrespective of dose or of the plasma

fre

fre

fre

fre

concentration of ethanol. The explanation for this is that

the rate of oxidation by the enzyme alcohol dehydrogenase POPULATION PHARMACOKINETICS
ks

ks

ks

ks

reaches a maximum at low ethanol concentrations, because

▼  In some situations, for example when a drug is intended for use
of limited availability of the cofactor NAD+ (see Ch. 49,
oo

oo

oo

oo

in chronically ill children, it is desirable to obtain pharmacokinetic

Fig. 49.6). data in a patient population rather than in healthy adult volunteers.
Saturation kinetics has several important consequences
eb

eb

eb

eb

Such studies in children are inevitably limited and samples for

(Fig. 11.9). One is that the duration of action is more strongly drug analysis are often obtained opportunistically during clinical
dependent on dose than is the case with drugs that do 149
m

care, with limitations as to quality of the data and on the number

t

t
 t

t
11

ne

ne

ne

ne
SECTION 1    General Principles

e.

e.

e.

e.
re

fre

fre

fre
f A Normal kinetics B Saturating kinetics
ks

ks

ks

ks
150 150
oo

oo

oo

oo
25
40
Plasma concentration (µmol/L)

Plasma concentration (µmol/L)

20
eb

eb

eb

eb
m

m
100 100
30
t

t
ne

ne

ne

ne
15
50 50
20
e.

e.

e.

e.
re

fre

fre

fre
10 10
sf

ks

ks

ks
0 0
k

0 2 4 6 8 10 0 2 4 6 8 10
oo

oo

oo

oo
Days Days
eb

eb

eb

eb
Therapeutic range 10 Dose (µmol/kg)
m

m
Fig. 11.9  Comparison of non-saturating and saturating kinetics for drugs given orally every 12 h. (A) The curves showing an
imaginary drug, similar to the antiepileptic drug phenytoin at the lowest dose, but with linear kinetics. The steady-state plasma
concentration is reached within a few days, and is directly proportional to dose. (B) Curves for saturating kinetics calculated from the
t

et
ne

ne

ne
known pharmacokinetic parameters of phenytoin (see Ch. 45). Note that no steady state is reached with higher doses of phenytoin, and

n
that a small increment in dose results after a time in a disproportionately large effect on plasma concentration. (Curves were calculated with
e.

e.

e.

e.
the Sympak pharmacokinetic modelling program written by Dr J.G. Blackman, University of Otago.)
fre

fre

fre

re
sf
ks

ks

ks

k
Uses of pharmacokinetics
oo

oo

oo

oo
eb

eb

eb

eb
• Pharmacokinetic studies performed during drug reduced in a patient with renal impairment (Ch. 52);
development underpin the standard dose regimens – the dose increment needed to achieve a target plasma
m

approved by regulatory agencies.
• Clinicians sometimes need to individualise dose regimens
to account for individual variation in a particular patient
(e.g. a neonate, a patient with impaired and changing
t
m
concentration range of a drug such as phenytoin with
saturation kinetics (Ch. 46, Fig. 46.4) is much less than
for a drug with linear kinetics.
• Knowing the approximate t1/2 of a drug can be very

t
ne

ne

ne

ne
renal function, or a patient taking drugs that interfere useful, even if a therapeutic concentration is not known:
with drug metabolism; see Ch. 10). – in correctly interpreting adverse events that occur
e.

e.

e.

e.
• Drug effect (pharmacodynamics) is often used for such some considerable time after starting regular treatment
fre

fre

fre

re
individualisation, but there are drugs (including some (e.g. benzodiazepines; see Ch. 45);
anticonvulsants, immunosuppressants and – in deciding on the need or otherwise for an initial sf
ks

ks

ks

antineoplastics) where a therapeutic range of plasma loading dose when starting treatment with drugs such
ok
oo

oo

oo

concentrations has been defined, and for which it is as digoxin and amiodarone (Ch. 22).
o

useful to adjust the dose to achieve a concentration in • The volume of distribution (Vd) of a drug determines the
eb

eb

eb

eb

this range. size of loading dose needed. If Vd is large (as for many
• Knowledge of kinetics enables rational dose adjustment. tricyclic antidepressants), haemodialysis will not be an
m

For example: effective way of increasing the rate of elimination in

– the frequency of dosing of a drug such as gentamicin treating overdose.
eliminated by renal excretion may need to be markedly
t

t
ne

ne

ne

ne
e.

e.

e.

e.

of samples collected from each patient. Population pharmacokinet-

fre

fre

fre

fre

ics addresses how best to analyse such data. Fitting data from all LIMITATIONS OF PHARMACOKINETICS
subjects as if there were no kinetic differences between individuals,
ks

ks

ks

ks

and fitting each individual’s data separately and then combining

Some limitations of the pharmacokinetic approach will be
the individual parameter estimates, each have obvious shortcom- obvious from the above account, such as the proliferation
oo

oo

oo

oo

ings. A better method is to use non-linear mixed effects modelling of parameters in even quite conceptually simple models.
(NONMEM). The statistical technicalities are considerable and beyond There are also limitations in the usefulness of monitoring
eb

eb

eb

eb

the scope of this chapter: the interested reader is referred to Sheiner drug concentrations in plasma as an approach to reducing
150 individual variability in drug response (see Ch. 12). Two
m

et al. (1997).
t

t
 t

t
11

ne

ne

ne

ne
Pharmacokinetics

e.

e.

e.

e.
re

fre

fre

fre
main assumptions underpin the expectation that by relating a nuclear receptor or when an active metabolite is involved.
f
response to a drug to its plasma concentration we can reduce Because of the blood–brain barrier, plasma concentrations
ks

ks

ks

ks
variability of response by accounting for pharmacokinetic rarely reflect local drug concentrations in the brain, so,
oo

oo

oo

variation – that is, variation in ADME. They are:
1. That plasma concentration of a drug bears a precise
eb
oo
with exception of lithium (Ch. 48) and some antiepileptic
drugs (Ch. 46), monitoring of plasma concentrations is not

eb

eb

eb
clinically useful.,
relation to the concentration of drug in the immediate
The second assumption is untrue in the case of drugs
environment of its target (receptor, enzyme, etc.).
m

m
that form a stable covalent attachment with their target,
2. That drug response depends only on the
and so produce an effect that outlives their presence in
concentration of the drug in the immediate
solution. Examples include the antiplatelet effects of aspirin
environment of its target.
and clopidogrel (Ch. 25) and the effect of some monoamine
t

t
ne

ne

ne

ne
While the first of these assumptions is very plausible for oxidase inhibitors (Ch. 48). In other cases, drugs in thera-
those few drugs that work through a target in the circulating peutic use act only after delay (e.g. antidepressants, Ch.
e.

e.

e.

e.
blood (e.g. a fibrinolytic drug working on fibrinogen) and 48), or gradually induce tolerance (e.g. opioids, Ch. 43) or
re

fre

fre

fre
reasonably plausible for a drug working on an enzyme, physiological adaptations (e.g. corticosteroids, Ch. 34) that
ion channel or G protein–coupled or kinase-linked receptor alter the relation between concentration and drug effect in
sf

ks

ks

ks
located in the cell membrane, it is less likely in the case of a time-dependent manner.
k
oo

oo

oo

oo
eb

eb

eb

eb
REFERENCES AND FURTHER READING
m

m
Atkinson, A., Huang, S.M., Lertora, J., Markey, S. (Eds.), 2012. Principles Lippincott Williams & Wilkins, Baltimore. Online simulations by H.
of Clinical Pharmacology, third ed. Academic Press, London. (Includes Derendorf and G. Hochhaus. (Excellent text; emphasises clinical
detailed section on pharmacokinetics) applications)
Birkett, D.J., 2010. Pocket Guide: Pharmacokinetics Made Easy.
t

et
McGraw–Hill Australia, Sydney. (Excellent slim volume that lives up to Population pharmacokinetics
ne

ne

ne
the promise of its title) Sheiner, L.B., Rosenberg, B., Marethe, V.V., 1997. Estimation of

n
Rowland, M., Tozer, T.N., 2010. Clinical Pharmacokinetics and population characteristics of pharmacokinetic parameters from routine
e.

e.

e.

e.
Pharmacodynamics. Concepts and Applications. Wolters Kluwer/ clinical data. J. Pharmacokinet. Biopharm. 5, 445–479.
fre

fre

fre

re
sf
ks

ks

ks

k
oo

oo

oo

oo
eb

eb

eb

eb
m

m
t

t
ne

ne

ne

ne
e.

e.

e.

e.
fre

fre

fre

re
sf
ks

ks

ks

ok
oo

oo

oo

o
eb

eb

eb

eb
m

m
t

t
ne

ne

ne

ne
e.

e.

e.

e.
fre

fre

fre

fre
ks

ks

ks

ks
oo

oo

oo

oo
eb

eb

eb

eb

151
m

m
t