Pharmacokinetics Enantiomers and Synovial S. R. Cox,
PhD,
of the R(-) and S(+) of Ibuprofen in the Serum Fluid of Arthritis Patients
E. P. Gall,
MD,
K. K. Forbes
BA,
M. Gresham,
BS, and
G. Goris,
MD
Eight patients with arthritis and knee effusions received 13 doses of a single 800-mg ibuprofen tablet every 8 hours. Serum and synovial fluid samples were obtained after the first and last doses and assayed for the R(-) and S(+) enantiomers of ibuprofen by a stereospecific assay. Since only S(+)-ibuprofen inhibits cyclo-oxygenase, a description of the time course of this isomer in synovial fluid is needed for the development of suitable pharmacodynamic models. The isomers were significantly different with respect to peak concentrations and areas under the concentration-time curves (AUC) in synovial fluid levels. No significant accumulation of either isomer was observed in serum or synovial fluid levels between the first and the last doses. The steady-state concentration of both isomers fluctuated less in synovial fluid than in plasma, and the synovial fluid concentrations of the S(+) isomer were about twice that of the R(-) isomer. The mean synovial albumin concentration was about 60% of the serum albumin concentration, and the steady-state isomer AUC values in synovial fluid were significantly correlated with the corresponding serum values after the differences between the two fluids with respect to albumin concentration were corrected. The authors conclude that binding of the isomers to albumin and the serum-synovial fluid albumin ratio controls the steady-state distribution of the ibuprofen isomers into synovial fluid. The ramifications of these findings in the development of satisfactory concentration-response relationships are discussed.
N
onsteroidal anti-inflammatory drugs (NSAIDs) are the basic therapy available for inflammatory and degenerative arthritis. Much has been learned in recent years about their mechanisms of action in relieving pain and inflammation in joints. Among other mechanisms, the inhibition of prosta-
cumulation and peak levels of these drugs in synovial fluid may lag behind and be lower than those levels in plasma.2 For drugs such as ibuprofen, only the S(+) enantiomeric form inhibits cyclo-oxygenase,3 and it is important to understand the kinetics of this bioactive form of ibuprofen in synovial fluid.
glandin
synthesis,
For
vation, enzymes
and the appear
Although
the
the
prevention
drugs
are
stream, much of their self, and a characterization
NSAIDs portance. sorption,
vary
in synovial Synovial distribution,
with
of neutrophil
stabilization of to be important
protein
fluid
levels
absorbed
action
fluid
cellular in these
may
occurs of the
into
the
be of therapeutic
inflammation.2
The
abmay
ac-
From the Upjohn Company (Drs. Cox, and Goris, Mr. Forbes), Kalamazoo, Michigan, the Department of Internal Medicine (Dr. Gall), The University of Arizona, Tucson, Arizona, and the ABC Laboratories (Mr. Gresham), Columbia, Missouri. Address for reprints: S. R. Cox, Clinical Pharmacokinetics Research Unit, The Upjohn Company, 301 Henrietta Street, Kalamazoo, Ml 49007.
88
#{149} J ClIn Pharmacol
1991;31:88-94
reason, in the
we studied synovial
fluid
the
enantiomers
of patients
of ibu-
with
arthritis.
METHODS itof
im-
reflect and
this
profen
blood
in the joint time-course
levels of NSAIDs and elimination,
and
acti-
lysosomal actions.1
Patients.
A total
of eight
patients,
five
with
active
inflammatory (rheumatoid arthritis, ankylosing spondylitis, and psoriatic arthritis), and three with degenerative arthritis were enrolled in the study after they gave written, informed consent. The characteristics of these patients are summarized in Table I. All patients had knee effusions that warranted aspiration and several patients had additional stable nonrheumatic diseases. A history and a physical examination were obtained, and a complete blood count, urinalysis, and serum and urine tests of
hepatic
and
renal
function
were
performed.
Patients
IBUPROFEN
IN ARTHRITIS
TABLE Patient
Subject
Age (yr)
1
55
Female
80
2
61
Female
71
I
Demographics Other (Nonarthrltic) Serious, Controlled Disease
Weight (kg)
Sex
PATIENTS
Concurrent
Zinc sulfate furosemide, azathioprine
-
Diabetes,
Medlcations*
Glyburide,
congestive
heart failure
atenolol,
metoclopramide, amitriptyline Auranofin, carisoprodol, diazepam
Psoriasis
3
56
Male
64
4
39
Male
84
-
-
5
79
Male
78
Hypertension, angina,
Cimetidine, ferrous sulfate, hydrochlorothiazide, isorbide dinitrate, theophylline, metoproterenol, potassium
bronchitis
chloride 6
Male
43
67
Hypertension,
cardiac
Promethazine,
arrhythmias
*
7
37
Male
81
Manic-depressive disorder
8
28
Male
76
Psoriasis
Excluding
therapy with gold salts or drugs were enrolled in the been administered at a conmonths. Patients receiving
with long half-lives, from the study,
i.e., but
piroxicam, were patients receiving
NSAIDs with shorter half-lives were enrolled in the study after they completed a 72-hour NSAID-free washout period. During the washout period, the patients were permitted to receive acetaminophen, co-
deine, or propoxyphene for pain. The patients were permitted to continue receiving drug treatment as prescribed by their attending physicians for nonarthritic conditions, but the doses of any concomitant medications were required to be stable during the
study
period.
Synovial
fluid
was
obtained
from
four
tients who met the previously described criteria. These patients had not taken for at least 72 hours before joint aspiration,
samples
were
fen in synovial hour washout
MISCELLANEOUS
Lithium, amitriptyline, oxycodone, sucralfate -
codei ne, and propoxyphene.
acetaminophen,
receiving concomitant other disease-modifying study if these drugs had stant dose for at least 3
NSAIDs excluded
chlorpromazine,
prazosin, methadone, prednisone
used fluid period,
to validate (vide the
the
assay
infra). After remaining
control
pa-
enrollment any NSAIDs and these
for ibupro.
at least a 72eight patients
received
13 doses of a single 800-mg ibuprofen tablet (Motrin, Upjohn, Kalamazoo, MI) every 8 hours. The patients were instructed to take each dose with 6 fluid ounces of water and to record the time of each dose in a diary. Blood (10 mL) was drawn into a vacutainer containing no anticoagulant before the first dose and at 0.5, 1, 1.5, 2, 3, 4, 6, and 8 hours after the first dose. The second dose was administered after the 8-hour sample was drawn. Blood samples were obtained at the same times and also at 10 and 12 hours on day 5 after the last dose. Serum was gathered as soon as possible and then frozen until assayed for ibuprofen.
Synovial one knee skin and with 1% fluid ples and
fluid samples were collected from only of each patient. To obtain these samples, underlying tissues were first infiltrated lidocaine, and then I to 2 mL of synovial
was withdrawn using aseptic were drawn at 2 and 8 hours at 2, 4, 6, and 8 hours after
synovial fluid ately after they for protein
and
techniques. Samafter the first dose the last dose. The
samples were centrifuged immediwere drawn and frozen until assayed ibuprofen.
For each
patient,
one
sy-
89
COX
novial
fluid
assayed
sample
for
DACOS nostics,
from
both
total
discrete Hialeah,
each
of the
protein
and
chemistry FL).
analyzer
study
days
was
albumin
by
a
(Coulter
Diag-
Ibuprofen Assays. The serum and synovial fluid samples were assayed for each ibuprofen enantiomer by the capillary gas chromatographic procedure previously described.4 This assay, originally developed fluid with
for the
serum, samples
was from
validated the first
for four
synovial patients.
No interferences were observed from synovial fluid, but only about 80% of either isomer was recovered relative to the plasma values. As a consequence of this low recovery, calibration standards for the synovial
fluid
assay
were
prepared
from
blank
[1]
nKDP
In Equation in the fluid.
1, Pt is the total albumin concentration If drug is extensively bound to albumin (>95%) over the concentration range of interest, then the total drug concentration (C0) will be approximately equal to Cb. If the extent of drug ionization bound),
is equal steady-state
in
and synovial fluid drug concentration concentration sequently CtOtSYn/CIOtS
both drug
fluids, then concentrations
free (unin serum
will be equal. The ratio of total in synovial fluid (C0’y’’) to the
of total drug in described by Equation =
the
serum 2:
(C05)
is con-
[2]
(PSYfl/PS)
If the binding parameters are independent of albumin concentration and identical in both serum and synovial fluid, then the total drug concentration in synovial fluid is described by Equation 3: C tot “#{176}
90
#{149} J Clin Pharmacol
-
C
tot ser (.
1991;31:88-94
t syn I /0
3 should
multiple
classes
be valid of
[3]
even
if binding
independent
steady-state concentrations areas under the concentration 3 is equivalently expressed =
occurs
sites.
to
Since
the
are proportional to the time curves, Equation by Equation 4:
AUC01Ser
[4]
(PSYfl/PSC)
In Equation 4, AUCIO is the steady-state area under the concentration-time curve over a dosage interval. If these assumptions are valid for the ibuprofen isomers, a regression of AUC018” versus the product AUC0 (P1sYfl/P5er) should have a slope of unity and an intercept of zero. Equation 5 follows directly from
Equation
4:
AUCtots/AUCtotrSYfl
Theory. For simplicity, drug binding to albumin is assumed to occur at one class of independent sites with association constant K and with a total number of sites n. Drug is assumed to bind only to albumin in both serum and synovial fluid, although the binding parameters do not have to be identical in both fluids. If the free drug concentration (C1) is sufficiently low, i.e., KC1 < .1, then the bound drug concentration (Cb) may be approximately given by Equation 1: =
Equation
synovial
fluid rather than from serum. The precision of the assay for the ibuprofen enantiomers in synovial fluid, as indicated by relative standard deviations, was less than 12% for enantiomer concentrations of I to 45 1zg/mL. Accuracy of the assay, as determined from the mean recovery of quality control standards, ranged from 93% to 100% over this same concentration range.
Gb
ET AL
In Equation
uration 5, the
AUCtotsser/AUCtotrser
=
the subscripts
5,
r or s denote
of the ibuprofen isomer. ratio of the steady-state
Based AUG
isomers in synovial fluid should the ratio of the isomers in serum. Data
Analysis.
vial fluid dose were trapezoidal tions
areas
under
the
than
.5 1zg/mL
fluid
config-
be estimable
the
serum
concentration-time curves calculated through 8 hours rule. In these calculations, synovial
the
on Equation values of the
from
and
after
syno-
the
last
(AUCO8)
by
concentra-
were assumed to be zero, concentrations at the beginning of the last dosing interval were assumed to be equal to the concentrations at 8 hours, the end of the steady-state dosing interval. The terminal disposition rate constant (Xj for each isomer was calculated from a log-linear regression of the serum concentration-time data apparently within the terminal disposition phase after the last dose. The maximum concentration (Cmax) after the last dose, the time of its occurrence (Tmax), and the ratio of maximum to
and
less
The
[5]
minimum concentrations dosing interval (Cmax/Cmjn) enantiomers rameters lated for
in serum
and
Tmax,
and
during were
the steady-state tabulated for both
in synovial fluid. The paAUG0_8 were also calcuboth enantiomers in serum after the first dose. Paired t tests were used to test the significance of the differences: 1) between isomers within a matrix (either serum or synovial fluid) with respect to Cmax,
the steady-state tween matrices parameters for
pharmacokinetic with respect each isomer,
day 5 serum isomer, and
between
pharmacokinetic 4)
parameters, 2) beto the pharmacokinetic 3) between day I and parameters
day
1 and
day
for
each
5 synovial
fluid concentrations of each isomer at both the 2and the 8-hour time points. Bonferroni corrections were used in these tests to maintain an overall value of a at .05. Regression analysis was used to evaluate correlations both among the isomers with respect to
IBUPROFEN
IN ARTHRITIS
PATIENTS
presented in Figure 1. Pharmacokinetic parameters for both isomers are listed in Tables II and III. The values for synovial fluid AUG08 were calculated by assuming that the observed concentrations at the end of the steady-state dosing interval were equal to the concentrations at the beginning of the interval.
40 E
This was done to limit the samples that were obtained
Ibuprofen both serum
j10
2 0
1
2
lime Identity
r-plssma
-
4
3 After
Oru
5
5
Administration
r-synoviat
Figure 1. Average concentrations plasma and synovial fluid. Error the mean.
-
- -
7
s-plumS
-
fluid
to only a minor extent in fluid over the time course
as evidenced by the lack of significant between days I and 5 regarding serum Cmax serum AUCO8, and synovial fluid concentrations at 2 and 8 hours. The mean serum AUCO8 values for both isomers on day 5 tended to be about 17% higher than those from day 1, and this small
s-syiioviat
-
of synovial each patient.
of the study, differences
0
(hours) -
accumulated and synovial
number from
difference cumulation
of the ibuprofen enantiomers in bars indicate ±1 standard error of
was consistent with that was predicted
These results support conditions prevailed
the steady-state from the day
acdata.
5
a contention that steady-state on day 5 of the study.
pharmacokinetic parameters within a fluid and also among fluids with respect to isomer pharmacokinetic parameters. Ratios of the steady-state concentrations of the isomers in synovial fluid and serum
The concentration-time differed significantly fluid. The Gmax values
were
in serum were about 40% of the values in synovial fluid. Also, the fluctuations between the maximum and minimum ibuprofen concentrations were much less in synovial fluid than in serum. As shown in Figure 1, serum and synovial fluid concentrations of each isomer were different at 2 hours but similar at times 4 hours. The average (±SD) steady-state sy-
compared
by a paired
in serum
t-test.
RESULTS Average
concentration
ibuprofen
isomers
versus in
serum
time and
curves
for both
synovial
fluid
are
TABLE Summary
of Plasma
than
profiles of both isomers between serum and synovial were about three times higher
in synovial
Subject
(hr)
C (Mg/mi)
AUC
Pharmacokinetic
(mL/mln/kg)
0.29
5.9
0.75
1.5
27
103
2.7
1.2
2.0
17
70
0.27
1.2
0.5
33
78
0.42
(hr’)
1.5
27
112
2
2.5
13
78
0.40
3
0.5
33
87
0.38
21
4
0.5
23
73
0.42
16
5
1.0
29
107
0.32
6
1.5
36
113
0.30
7
1.0
30
92
0.33
8
2.0
21
82
0.33
8.8
Mean
1.3 ± 0.7
26
93 ± 16
0.35 ± 0.05
9.2 ± 5.7
*
SD
± 7
Minim urn ratio s ince the C,,,, was below
MISCELLANEOUS
the limit
values
Tmax
R(-).lbuprofen
1
±
the
Parameters
C,/ C,,.,,,
X,
(Mg hr/mL)
and
II
S(+)-lbuprofen T
fluid,
CI,,.
C,..,
T
(hr)
(Mg/mi)
AUC0
X
(Mg hr/mi)
(hr1) 0.32
C,,,,,,/ C,.,,,
CI., (mi/mitt/kg)
17
0.81
5.9
1.3
33
1.4
1.1
0.5
21
50
0.57
>42*
1.6
0.81
1.0
28
84
0.39
12
1.0
10
0.87
1.0
16
36
0.16
15
2.8
12
0.89
1.0
22
41
0.37
42
2.0
1.1
1.0
25
75
0.62
41
1.2
1.1 ± 0.5
24 ± 6
67
0.39 ± 0.15
8.3
of qua ntitation,
1.0 ± 0.2 0.5 Mg/rnL.
The overall
±
23
me an value was calculated
26 ± 15
±
1.5 0.6
with a ratio of 42 for this subject.
91
COX
ET AL
TABLE Summary
of Synovial
III
Fluid Pharmacokinetic
Parameters
S(+)-lbuprofen
Subject
T,,..,,
1 2 3
4
4 5
6
7 8 ± SD
Mean *
Minimum
ratio since
for
R(-)-lbuprofen AUC
20
C,,,.,,/Cmt,,
1,..,,,,
106
2.4
4
2.1
4
5.6
28
3.0 2.6
2 2
16 5.7
84 25
4
4.8
30
4 4 2
8.9 15 11 11 ± 5
54 78 67
3.1 5.7 2.1 1.7 2.5 2.0
2 2 2 4 4 2
11 3.7 3.6 2.7 5.1 6.2
58 12 19 15 29 37
3.3 >74* 3.0 2.7 3.2 2.7
2.7 ± 1.3
3.0 ± 1.1
6.4 ± 3.8
32 ± 18
the C,,
value was below
61 ± 28
the limit of quantitation,
0.5 Mg/mL.
ratio, was .53 its optical
The overall
with >
were
observed
of either
ing albumin-normalized slopes not significantly 2, 3). Other correlations
among
isomer
and
the the
serum
AUCO8
different
from
were
also
synovial
correspond-
values, unity
observed
with (Figures
between
mean
the synovial centrations = .62) and
Serum and synovial fluid levels differed significantly with respect to albumin concentration, with mean (±SD) values of 3.9 ± 0.5 and 2.3 ± 0.6 g/dL, respectively. As predicted by the model, significant
13
Cma,,/Cmtn
44
antipode.
AUCO#{216}values
AUC
8.1
to serum ibuprofen concentration as the ratio of AUCQ8 values, the R-isomer and .66 ± .27 for
correlations
Cm,,,,
4
3.25 ± 1.0
novial fluid calculated ± .14
C,,,,,,
61
value was calculated
3.5 ± 1.6
with a ratio of 7.4 for this subject.
and albumin-normalized serum conof S(+)-ibuprofen at 8 hours (P < .03, r2 between the isomers in synovial fluid to both Gmax and AUCO8 (P < .006, r2
respect
.74).
The
values
for
X5 were
calculated
for
drug
in
serum on day 5 alone, because only then were a sufficient number of time points for reliable lations. No difference was observed between isomers with respect to X, and the harmonic
there calcuthe mean
half-lives
and
hours, respect
of R-
and
respectively. to the other
S-ibuprofen The serum
were
isomers did pharmacokinetic
not
1.8
2
differ with parame-
70 E
#{163}
0
SO
0 0
2
50
0
I
p a 0
I
0 0
20
30
40
50
00
Serum S(+).ibeprots
71
00
50
100
AUC,.,, (mci hr/el)
2. Plot of S(+)-ibuprofen steady-state AUCOS values: Synovial AUCO8 versus albumin-normalized serum AUC08. The line is from a statistically significant linear regression (P < .004). Regression results: r2 = .75, slope (mean ± SE) = 1.1 ± .2, intercept not significantly different from zero. Figure
92
#{149} J Clln Pharmacol
1991;31:88-94
10
20
30
40
Serum R(-)-ihepmten
50
60 AUC,
70
00
90
(mcg Ir/mi)
Figure 3. Plot of R(-)-ibuprofen steady-state AUC08 values: Synovial AUCOS versus albumin-normalized serum AUCOS. The line is from a statistically significant linear regression (P < .00 1). Regression results: r2 = .85, slope (mean ± SE) = 0.77 ± .12, intercept not significantly different from zero.
IBUPROFEN
IN ARTHRITIS
ters except that the serum AUG08 values for 5-ibuprofen tended to be larger than those of R-ibuprofen (.05