Safety of oral robenacoxib in the cat

Abstract

King, J. N., Hotz, R., Reagan, E. L., Roth, D. R., Seewald, W., Lees, P. Safety of oral robenacoxib in the cat. J. vet. Pharmacol. Therap. 35 , 290–300.
The safety of robenacoxib, a nonsteroidal anti‐inflammatory drug with high selectivity for inhibition of the cyclooxygenase (COX)‐2 isoform of COX, was investigated in the cat in two randomized, blinded, placebo‐controlled, parallel‐group studies. Robenacoxib was administered orally to healthy young domestic short‐hair cats at dosages of 0 (placebo), 5 and 10 mg/kg once daily for 28 days (study 1) and at 0 (placebo), 2, 6 and 10 mg/kg twice daily for 42 days (study 2). The recommended minimum dosage for robenacoxib tablets in cats is 1 mg/kg once daily (range 1–2.4 mg/kg). Relative to placebo treatment, no toxicologically significant effects of robenacoxib were recorded in either study, based on general observations of health, haematological and clinical chemistry variables and urinalyses in life, and by post mortem organ weight, gross pathology and histopathology assessments. Pharmacokinetic–pharmacodynamic simulations indicated that all dosages of robenacoxib were associated with marked inhibition of COX‐2 at peak effect (median I max 97.8–99.4% inhibition) with lesser inhibition of COX‐1 (median I max 26.8–58.3% inhibition). Inhibition of the COXs was short lasting, with >10% median inhibition persisting for 4.0 h for COX‐2 and 1.5 h for COX‐1. These levels of inhibition of COX‐1 and COX‐2 twice daily with robenacoxib were not associated with any detectable toxicity, suggesting that, as previously described in dogs, the high safety index of robenacoxib in cats may be related to a combination of its high COX‐2 selectivity and short residence time in the central compartment.

Full text

Safety of oral robenacoxib in the cat
J. N. KING*
R. HOTZ

E. L. REAGAN
à
D. R. ROTH
§
W. SEEWALD* &
P. LEES
–
*Novartis Animal Health Inc., Clinical
Development, Basel, Switzerland;

Novartis
Centre de Recherches, St Aubin,
Switzerland;
à
Liberty Research Inc.,
Waverly, NY, USA;
§
Novartis Pharma AG,
Preclinical Safety, Basel, Switzerland;
–
Royal Veterinary College, Hawkshead
Campus, Hatfield, Hertfordshire, UK
King, J. N., Hotz, R., Reagan, E. L., Roth, D. R., Seewald, W., Lees, P. Safety of
oral robenacoxib in the cat. J. vet. Pharmacol. Therap. doi: 10.1111/j.1365-
2885.2011.01320.x.
The safety of robenacoxib, a nonsteroidal anti-inflammatory drug with high
selectivity for inhibition of the cyclooxygenase (COX)-2 isoform of COX, was
investigated in the cat in two randomized, blinded, placebo-controlled, parallel-
group studies. Robenacoxib was administered orally to healthy young domestic
short-hair cats at dosages of 0 (placebo), 5 and 10 mg ⁄kg once daily for
28 days (study 1) and at 0 (placebo), 2, 6 and 10 mg ⁄kg twice daily for
42 days (study 2). The recommended minimum dosage for robenacoxib tablets
in cats is 1 mg ⁄kg once daily (range 1–2.4 mg ⁄kg). Relative to placebo
treatment, no toxicologically significant effects of robenacoxib were recorded in
either study, based on general observations of health, haematological and
clinical chemistry variables and urinalyses in life, and by post mortem organ
weight, gross pathology and histopathology assessments. Pharmacokinetic–
pharmacodynamic simulations indicated that all dosages of robenacoxib were
associated with marked inhibition of COX-2 at peak effect (median I
max
97.8–
99.4% inhibition) with lesser inhibition of COX-1 (median I
max
26.8–58.3%
inhibition). Inhibition of the COXs was short lasting, with >10% median
inhibition persisting for 4.0 h for COX-2 and 1.5 h for COX-1. These levels of
inhibition of COX-1 and COX-2 twice daily with robenacoxib were not
associated with any detectable toxicity, suggesting that, as previously described
in dogs, the high safety index of robenacoxib in cats may be related to a
combination of its high COX-2 selectivity and short residence time in the central
compartment.
(Paper received 1 April 2011; accepted for publication 8 June 2011)
J. N. King, Novartis Animal Health Inc., Clinical Development, CH-4058 Basel,
Switzerland. E-mail: jonathan.king@novartis.com
INTRODUCTION
Although the need to provide adequate analgesia in the cat has
been increasingly recognized (Robertson, 2005, 2008; Lascelles
et al., 2007; Livingston, 2010), there are relatively few non-
steroidal anti-inflammatory drugs (NSAIDs) licensed for feline
use, and their usage is generally limited to administration over
short periods of time. For example, carprofen is authorized only
for single use, ketoprofen for a maximum of 5 days and
tolfenamic acid for a maximum of 3 days. Meloxicam is licensed
as a single dose of 0.3 mg ⁄kg subcutaneously in the cat and, in
the EU, at a much lower dosage of 0.05 mg ⁄kg orally for long-
term therapy of unrestricted duration. Although meloxicam is
licensed and recommended for long-term use in the cat (Gunew
et al., 2007), the manufacturer’s literature indicates that the
safety margin in this species is narrow (European Public
Assessment Report Scientific Discussion for Metacam, 2010,
http://www.ema.europa.eu) and, because of concerns on toxic-
ity, a cautious approach was proposed by Papich (2008) and
Robertson (2008).
The small number of NSAIDs licensed for use in the cat, their
limited indications and restrictions on duration of dosing may
reflect the paucity of pharmacological and toxicological data in
the cat together with the low safety margins for many NSAIDs in
this species. The poor safety margins of some NSAIDs in cats may
be a result of their slow clearance from the body (Lascelles et al.,
2007).
In recent years, NSAIDs of the coxib class, which are
preferential or selective in their inhibitory action for the
cyclooxygenase (COX)-2 isoform, have been developed for
canine use (McCann et al., 2005; King et al., 2009; Roberts
et al., 2009; Cox et al., 2010). The rationale underlying
development of the coxibs is that they should have similar
efficacy but improved safety profiles compared with the older
nonselective NSAIDs. Five drugs of the coxib class, cimicoxib,
deracoxib, firocoxib, mavacoxib and robenacoxib are now
J. vet. Pharmacol. Therap. doi: 10.1111/j.1365-2885.2011.01320.x
2011 Blackwell Publishing Ltd 1

licensed for use in the dog, but only one, robenacoxib, is
authorized for administration to cats.
In the cat, dog and rat, robenacoxib has been shown to be a
highly selective inhibitor of COX-2 (Giraudel et al., 2009a; King
et al., 2009, 2010; Schmid et al., 2010) and possesses the
analgesic, anti-inflammatory and antipyretic actions character-
istic of NSAIDs (Giraudel et al., 2009b; King et al., 2009). The
pharmacokinetic (PK) profile of robenacoxib has been described
in the cat and dog following oral and subcutaneous administra-
tion (Giraudel et al., 2009b; Jung et al., 2009; Pelligand et al.,
2009; J.N. King, M. Jung, M.P. Maurer, V.B. Schmid, W. Seewald
& P. Lees, In preparation). In both species, absorption is rapid
after oral and subcutaneous dosing, whilst clearance is also rapid
and the elimination half-life from blood is short. However, in the
cat, dog and rat, robenacoxib has been shown to persist for
longer in inflammatory exudate than in blood (King et al., 2009;
Pelligand et al., 2009; Silber et al., 2010).
The aim of this study was to complement the reported feline
studies on robenacoxib PK and pharmacodynamics (PD), admin-
istered at recommended dosages, by investigating the safety
profile of robenacoxib in the cat. The principal objective was to
investigate the safety of robenacoxib, in a tablet formulation,
administered at daily dosages of 0, 5 and 10 mg ⁄kg (28-day
study) and 4, 12 and 20 mg ⁄kg (42-day study). As the clinically
recommended dosage of robenacoxib is 1–2.4 mg ⁄kg, the
20 mg ⁄kg dosage constitutes an 8.3- to 20-fold overdose. In both
studies, safety was assessed by: (i) observations and examinations
to establish general health status; (ii) haematology and clinical
chemistry profiles; (iii) urinalyses; and (iv) post mortem investiga-
tions, including gross and histopathology. A second objective was
to use Pharmacokinetic-pharmacodynamic (PK-PD) modelling to
correlate safety data with the magnitude and time course of
inhibition of COX-1 and COX-2 in the central compartment.
MATERIALS AND METHODS
Two safety studies were conducted. In study 1, cats received 0, 5
and 10 mg ⁄kg robenacoxib once daily for 28 days at Liberty
Research Inc. (Waverly, NY, USA). In study 2, cats received 0, 2,
6 and 10 mg ⁄kg robenacoxib twice daily for 42 days at Novartis
Centre de Recherches (St Aubin, Fribourg, Switzerland). Both
studies were run in compliance with GLP and site procedures,
FDA Guidelines on the conduct of target animal safety studies
(CVM Guidelines 33 and 104), and after approval of the protocol
by institutional Animal Use and Welfare Committees.
Animals
Healthy young domestic short-hair cats were used. All cats were
in good health at the start of the study as assessed by physical
examinations plus haematological, clinical chemistry and uri-
nalyses. The acclimatization phase for the cats was approxi-
mately 2 weeks in study 1, and 3 weeks in study 2.
In study 1, nine females and nine males were used, aged 20–
21 weeks with average body weights of 2.4 kg (females) and
2.8 kg (males) at study commencement. In study 2, 16 females
and 16 males were used, aged 7.5–8.5 months with average
body weights of 2.8 kg (females) and 3.7 kg (males) at study
commencement.
Cats were housed in environmentally controlled rooms with
regulation of temperature and relative humidity. The light:dark
cycle was 12 h:12 h in study 1 and 14 h:10 h in study 2. In
study 1, cats were housed in individual cages. In study 2, cats
were group housed during the day but were housed individually
at night and during sampling periods. The cats were fed certified
Feline Diet 5003 (PMI Nutrition International, St. Louis, MO,
USA) ad libitum, renewed once daily, in study 1, and Selina

3-mix (A ⁄S Arovit Petfeed, Denmark-6600 Vejen, Denmark)
provided twice daily in study 2. In both studies, fresh drinking
water was available ad libitum. Feed consumption was measured
once daily in study 1 and twice daily in study 2. Body weight
was recorded weekly in both studies.
Baseline values (including body weight and clinical observa-
tions) before the start of dosing with the test articles were
recorded for 14 days in study 1, and 12–13 days in study 2.
Study designs
In both studies, prospective, randomized, blinded, parallel-group
designs were used. Cats were assigned to treatment groups with
three groups each of six cats (three female, three male) in study
1, and four groups each of eight cats (four female, four male) in
study 2. Randomizations were conducted separately for females
and males with additional stratification by weight within sex to
ensure that cats spanning the range of body weights of both
genders were represented in each group.
In both studies, there was a separation of responsibilities, such
that personnel conducting the clinical observations, physical
examinations, clinical pathology and gross and microscopic
pathology were blinded to the treatment each cat received. As
there were 32 cats in study 2, not all procedures could be
managed within a single day. Therefore, some procedures were
spread over 2–3 days (e.g., 3 days for the post mortems).
Test articles
In study 1, cats received nonflavoured lactose-based tablets
containing 10 mg robenacoxib, or matched placebo tablets
identical in composition to the robenacoxib tablets but excluding
the active ingredient (Novartis Animal Health, Basel, Switzer-
land). Cats were dosed once daily with 5 mg ⁄kg robenacoxib
(Group B), 10 mg ⁄kg robenacoxib (Group C) or placebo (Group A),
administered in an amount equivalent to the 10 mg ⁄kg
robenacoxib group. Combinations of whole or half tablets were
placed on the back of the tongue, and the mouth held closed
until the tablets were swallowed. All cats were treated for 28
consecutive days. As food was available ad libitum, the relation
between feed intake and test article administered was not
controlled.
In study 2, cats received flavoured tablets containing 6 mg of
robenacoxib (Onsior

; Novartis Animal Health). Cats were dosed
2J. N. King et al.
2011 Blackwell Publishing Ltd

twice daily with robenacoxib at dosages of 2, 6 and 10 mg ⁄kg
(in Groups B, C and D, respectively) or placebo (Group A). For
dosing, whole or parts of tablets were placed into hard gelatine
capsules (Dermarcam S.A., Chambesy, Switzerland). The placebo
comprised an empty gelatine capsule. Each capsule was placed
on the back of the tongue and the cats administered a small
volume of tap water to facilitate swallowing. Treatments were
given twice daily for 42–44 consecutive days. The cats were fed
twice daily 1 h after the administration of the test articles (8.00–
10.00 and 18.00–20.00).
Safety variables recorded and samples collected in life
Cats were observed twice daily for general health and behaviour.
In addition, each cat had a detailed physical examination by a
veterinarian on two occasions in study 1 (in the baseline period
and on day 21 of dosing) and on three occasions in study 2 (at
the start of acclimatization, in the baseline period and in week 4
of treatment).
Venous blood samples were taken (from a jugular vein in
study 1 and from a cephalic or brachial vein in study 2) for blood
haematology and coagulation and serum clinical chemistry
analyses. In study 1, blood samples were taken prior to drug
dosing and again at the end of the dosage period, with no record
of relation to feeding. In study 2, samples were collected from
fasted animals on days )13 ⁄)12, )7⁄)6, 14 ⁄15 and 35 ⁄36.
Anticoagulants were EDTA for haematology and sodium citrate
for coagulation variables. No anticoagulant was used for serum
chemistry variables. Variables measured included serum alanine
aminotransferase (ALT), amylase, aspartate aminotransferase
(AST), alkaline phosphatase, creatinine kinase, gamma-glu-
tamyltransferase (GGT), bilirubin, creatinine, urea (blood urea
nitrogen), glucose, calcium, chloride, inorganic phosphorus,
potassium, sodium, total protein, albumin and globulin. Hae-
matology variables included erythrocyte count, haemoglobin
concentration, haematocrit, mean cell volume, platelet
count, total leucocyte count, differential counts for neutrophils,
eosinophils, basophils, monocytes and lymphocytes, and
activated partial thromboplastin time.
Urine samples were collected on the same days as the blood
samples in study 1, and on days )13 ⁄)12 ⁄)11, )6⁄)5,
13 ⁄14 ⁄15 and immediately prior to necropsy on days
42 ⁄43 ⁄44 in study 2. Urine was collected via free catch into
litter trays except after euthanasia, when it was collected directly
with a syringe and needle from the bladder. Variables measured
included urine pH and urine-specific gravity.
Blood robenacoxib concentrations
In study 2, blood samples for measurement of robenacoxib
concentration were taken from a cephalic vein into EDTA on
days )12, )7 and 35. The 35 day samples were collected
immediately before the a.m. dosing, i.e., approximately 12 h
after the previous dose. Blood samples were stored below )16 C
before determination of blood robenacoxib concentrations using
a liquid chromatography–mass spectrometry method as
described and validated previously in cats (Jung et al., 2009).
The limit of quantification in cat blood was 3 ng ⁄mL.
Variables recorded post mortem
In study 1, after 28 days dosing, the animals were fasted
overnight, anaesthetized with ketamine (Fort Dodge Animal
Health, Overland Park, KS, USA) and then euthanized with
pentobarbital (Fort Dodge Animal Health), both drugs adminis-
tered intravenously. In study 2, after 42–44 days of treatment,
cats were euthanized by intravenous injection of pentobarbital
(Esconarkon

; G. Streuli and Co., Uznach, Switzerland). The cats
were then exsanguinated and necropsies performed, together
with organ weight measurements and macroscopic and histo-
pathology on a wide range of tissues. Selected organs and tissues
collected for weighing and histopathological investigation are
indicated in the Results section (vide infra, not all data shown).
All tissues were preserved in 10% formalin (study 1) and neutral
phosphate-buffered formalin (study 2) except for testes, epidid-
ymides and eyes with optic nerves, which were fixed in
Davidson’s solution.
Statistical analyses of safety studies
Data are reported as mean and SD. Data for most variables were
numerical and were analysed statistically by analysis of variance
(
ANOVA
) using SAS

procedure PROC MIXED (SAS

Institute Inc,
Cary, NC, USA).
ANOVA
was used for endpoints measured post-
treatment only once, and repeated measures analysis of variance
(
RMANOVA
) for endpoints measured multiple times post-treatment.
Analysis of covariance (
RMANCOVA
) was used for variables with a
baseline value and multiple measures post-treatment. Data were
log (study 2) or rank (study 1) transformed if required to give the
best estimation of a normal distribution. The following param-
eters were included in the original
ANOVA
models: treatment, sex,
treatment ·sex interaction, block and baseline (if applicable).
For
RMANOVA
and
RMANCOVA
, time and treatment ·time inter-
action were included also. Nonsignificant terms were removed
sequentially from the model. In the event of overall significance,
groups were compared with the placebo group in post hoc
analyses using linear contrasts with correction for multiple tests,
either for pooled sexes (if the treatment ·sex interaction was not
significant) or separately for each sex (if the interaction was
significant).
All calculations were carried out using the software SAS

,
Versions 8.2. (study 1) and 9.1.3. (study 2) (SAS

Institute Inc).
Two-tailed Pvalues less than 0.05 were considered significant.
PK–PD simulations
Pharmacokinetic–pharmacodynamic simulations were under-
taken using: (i) blood concentration–time data generated in a
previous investigation in 12 cats (six male, six female), in which
dosages of 1–2 mg ⁄kg robenacoxib were administered orally to
each cat on a single occasion in a cross-over study to investigate
the effect of feeding on PK variables (study 3, J.N. King, M. Jung,
Robenacoxib safety in cats 3
2011 Blackwell Publishing Ltd

M.P. Maurer, V.B. Schmid, W. Seewald & P. Lees, In prepara-
tion); and (ii) PD data generated in another previous investi-
gation that established IC
50
values for inhibition of COX-1 and
COX-2 isoforms in the cat (study 4, Giraudel et al., 2009a). Total
blood concentrations of robenacoxib were used in the analysis
because, although binding to plasma proteins is extensive
(>99%), this binding is linear (Jung et al., 2009). As oral
bioavailability of robenacoxib is higher when cats are fasted,
simulations were calculated for fasted animals (J.N. King,
M. Jung, M.P. Maurer, V.B. Schmid, W. Seewald & P. Lees, In
preparation). A total of 12 robenacoxib concentration–time
profiles were available from fasted cats. To standardize profiles,
blood concentrations were adjusted to a nominal dose of
1mg⁄kg, assuming linear PKs. A one-compartment PK model
was fitted to each of the 12 concentration–time profiles as
follows with parameters: FD ⁄V = absorbed dose ⁄volume of
distribution, K
10
= rate constant for elimination and K
a
= rate
constant for absorption, with K
a
‡K
10
.
Blood concentration (C) is given by:
C¼FD
VKa
KaK10
expðK10 tÞexpðKatÞ½;
if Ka>K10
C¼FD
VK10 texpðK10 tÞ;
if Ka¼K10
Fitting parameters involved optimizing a nonlinear function,
which was conducted iteratively. Only good model fits (n= 10)
were used subsequently. The data did not always conform, so it
was assumed pragmatically that K
a
=K
10
, which, for most of
the profiles, was in fact the case. Consequently, only two of the
three parameters, FD ⁄V and K
10,
remain. To obtain the
approximate distributions of these two parameters from the 10
pairs of values obtained, the data were re-parameterized. Natural
parameters were considered to be volume of distribution (V) and
AUC ¼FD
VK10, or logarithms thereof. Univariate inspection of
these parameters showed that log V and log AUC followed
approximately a Gaussian distribution, although both seemed to
be correlated. Therefore, a bivariate Gaussian distribution (with
nonzero correlation) was fitted to the 10 pairs of values (log
AUC, log V). For dosages other than 1 mg ⁄kg, AUC was
multiplied by the dosage.
For PD data, the standard sigmoidal (Hill) model
% inhibition ¼Imax Cn
ICn
50 þCn
was used to establish predicted profiles of inhibition of throm-
boxane (Tx)B
2
(COX-1) and prostaglandin (PG)E
2
(COX-2) with
dosages of 2, 6 and 10 mg ⁄kg of robenacoxib. I
max
represents
the maximal inhibition, C the (total blood) concentration of
robenacoxib, IC
50
the concentration of robenacoxib providing
50% of I
max
, and n the slope parameter. Geometric mean (geo-
metric SD) values for n and IC
50
(ng ⁄mL) were, respectively,
0.79 (1.48) and 7298 (2.17) for inhibition of TxB
2
, and 0.89
(1.56) and 21.0 (2.49) for inhibition of PGE
2
(study 4, Giraudel
et al., 2009a).
For simulations in 10 000 animals, values for log V
D,
log AUC,
n (TxB
2
), log ED
50
(TxB
2
), n (PGE
2
) and log ED
50
(PGE
2
) were
drawn randomly from their underlying distributions (taking into
account the correlations). Blood concentrations of robenacoxib
and corresponding inhibitions of TxB
2
and PGE
2
were calculated
using the underlying models. Medians and 90% tolerance
intervals were calculated and plotted against time. All calcula-
tions were carried out using the software SAS

, Version 9.1.3.
(SAS

Institute Inc).
RESULTS
Study 1
Nominal dosages were 0, 5 and 10 mg ⁄kg robenacoxib once
daily, respectively, in groups A, B and C. Actual daily dosages
received by Groups B and C over the 28 day dosing periods were
respectively: 5.43 ± 0.09 mg ⁄kg (males) and 5.53 ± 0.12
mg ⁄kg (females) in Group B; and 10.35 ± 0.25 mg⁄kg (males)
and 10.44 ± 0.28 mg ⁄kg (females) in Group C.
No cat died or became moribund during the study. Based on
clinical observations of general health [appetite, behaviour, body
temperature and integument, and cardiovascular (heart rate),
gastrointestinal, muscular, nervous, respiratory (respiratory
rate) and ocular systems], there were no statistically significant
or toxicologically relevant changes either from baseline mea-
surements or for the comparisons of robenacoxib with placebo-
treated cats (data not shown). The commonest sign was soft
stools, which occurred in a maximum of two of six cats per
group. Body weight increased in all groups each week through-
out the study in both females and males, with the following
exceptions: slight decrease of 0.02 kg between days 21 and 27
(Group C females); slight decrease of 0.01 kg between days 21
and 27 (Group B males); and no change in weight between days
21 and 27 (Group C males). For pooled sexes data, there were no
significant differences between the placebo- and robenacoxib-
treated groups for body weight. However, mean body weight
gain was significantly lower for Group C in the first week of
dosing compared with the gain achieved in the preceding week.
As the rate of gain in weight in this group was similar to
historical data in the colony, it was concluded to be unrelated to
treatment.
Mean daily food consumption was not significantly different in
each week of the study between treatment groups with one
exception. During week 2, food consumption was statistically
lower compared with other weeks for Group C cats. As this
occurred in a single week only, it was not considered to be
toxicologically relevant.
Summary data for selected haematology variables are pre-
sented in Table 1 (not all data shown). There were no significant
differences in predosing (day 0) values for any variable between
the placebo group and either the 5 or 10 mg ⁄kg robenac-
oxib groups. By day 28, there were significant increases in
4J. N. King et al.
2011 Blackwell Publishing Ltd

erythrocyte count, leucocyte count, monocyte count and
haematocrit, and a significant reduction in haemoglobin con-
centration in the 5 mg ⁄kg robenacoxib group (Group B)
compared with placebo (Group A). From the small magnitude
of the changes, the differences were considered to be not relevant
toxicologically. Moreover, comparison between cats receiving
10 mg ⁄kg robenacoxib (Group C) and placebo revealed only one
statistically significant difference, a numerically slight increase in
erythrocyte count.
Summary data for selected clinical chemistry variables are
presented in Table 2 (not all data shown). There were no
significant differences in predosing (day 0) values for any variable
between the placebo group and either the 5 or 10 mg ⁄kg
robenacoxib groups. By day 28, there were statistically significant
increases in total protein, albumin and globulin concentrations in
cats receiving 5 mg ⁄kg robenacoxib (Group B) compared with
placebo (Group A). In each case, the changes were numerically
small, and similar changes compared with placebo were not
obtained in cats administered 10 mg ⁄kg robenacoxib (Group C).
There were significant, but numerically small, reductions in
inorganic phosphorus concentration by day 28 in both robenac-
oxib groups relative to placebo. None of the clinical chemistry
differences were considered to be toxicologically relevant.
Urine was examined visually for appearance and microsco-
pically for formed elements and analysed for specific gravity.
There were no significant differences between placebo and
robenacoxib groups on either day 0 or day 28.
Gross pathology examinations revealed occasional findings in
all groups, but these did not follow any pattern that would
indicate a relationship to treatment or differences between
groups. These comprised small testes and thymus, enlarged
thyroid, cervical lymph nodes and ovary, small nodule on
pancreas, small discoloration on spleen and enlarged and ⁄or
mottled lymph nodes.
Initial and final body weights and post mortem organ weights
are presented in Table 3. There were no significant differences
between groups. Organ weights were also expressed (i) as a
percentage of body weights and (ii) as a percentage of brain
weights. Apart from a significant increase in pituitary ⁄body
weight ratio and pituitary ⁄brain ratio in Group B, there were no
Table 1. Summary of haematology data for
cats in study 1; mean (SD), n=6
Variable (units) Day
Group A
(placebo)
Group B
(robenacoxib
5mg⁄kg)
Group C
(robenacoxib
10 mg ⁄kg)
Erythrocyte count (10
12
⁄L) 0 8.1 (1.2) 7.7 (0.9) 8.2 (0.8)
28 8.4 (1.1) 8.8 (1.2)** 9.0 (0.8)*
Haemoglobin concentration (m
M
) 0 11.2 (1.3) 10.1 (1.3) 10.9 (0.9)
28 11.5 (1.1) 11.3 (1.5)* 11.7 (0.6)
Haematocrit (L ⁄L) 0 0.34 (0.04) 0.31 (0.04) 0.33 (0.03)
28 0.34 (0.04) 0.34 (0.05)* 0.35 (0.03)
Platelet count (10
9
⁄L) 0 221.3 (127.8) 233.3 (120.0) 315.0 (133.7)
28 207.3 (126.7) 247.3 (171.6) 274.5 (125.9)
Total leucocyte count (10
9
⁄L) 0 12.3 (5.2) 14.7 (4.9) 14.3 (4.9)
28 7.1 (2.0) 11.3 (3.7)* 9.7 (2.0)
Activated partial thromboplastin
time (sec)
0 49.7 (11.8) 56.4 (13.3) 53.7 (7.2)
28 60.1 (22.9) 59.9 (7.2) 52.5 (8.4)
Statistical difference between placebo and robenacoxib-treated cats: *P< 0.05; **P< 0.01.
Table 2. Summary of serum clinical chemis-
try data for cats in study 1; mean (SD), n=6
Variable (units) Day
Group A
(placebo)
Group B
(robenacoxib
5mg⁄kg)
Group C
(robenacoxib
10 mg ⁄kg)
Alanine aminotransferase (IU ⁄L) 0 49.3 (6.8) 59.8 (10.5) 55.3 (6.6)
28 62.7 (7.9) 74.2 (13.7) 56.7 (12.2)
Aspartate aminotransferase (IU ⁄L) 0 26.3 (3.3) 23.2 (2.4) 25.2 (2.0)
28 22.5 (2.9) 23.7 (1.9) 20.2 (4.9)
Creatine kinase (IU ⁄L) 0 629.5 (165.8) 401.5 (45.9) 435.3 (186.6)
28 402.3 (134.4) 593.0 (264.3) 464.7 (75.1)
Alkaline phosphatase (IU ⁄L) 0 131.0 (57.5) 142.7 (37.0) 116.8 (38.0)
28 117.5 (32.5) 124.3 (38.3) 92.0 (32.2)
Blood urea nitrogen (m
M
) 0 5.1 (1.2) 4.6 (0.8) 5.1 (0.8)
28 4.3 (0.9) 4.2 (0.8) 4.0 (0.8)
Creatinine (l
M
) 0 63.4 (17.2) 56.0 (7.2) 64.9 (10.7)
28 84.1 (14.5) 76.7 (16.5) 88.5 (12.5)
Total protein (g ⁄L) 0 63.7 (4.3) 62.7 (2.5) 64.0 (3.2)
28 61.8 (2.6) 64.5 (3.9)* 62.7 (2.3)
Statistical difference between placebo and robenacoxib-treated cats: *P< 0.01.
Robenacoxib safety in cats 5
2011 Blackwell Publishing Ltd

significant differences between placebo and robenacoxib groups
(data not shown).
Histopathology indicated a range of microscopic changes in
cats of all groups, with no indication of relationship to or
differences between treatments.
Study 2
Nominal dosages were 0, 2, 6 and 10 mg ⁄kg robenacoxib twice
daily, respectively, in groups A, B, C and D. Actual daily dosages
received by Groups B, C and D were respectively:
4.7 ± 0.53 mg ⁄kg (males) and 4.7 ± 0.54 mg ⁄kg (females) in
Group B; 13.3 ± 0.66 mg ⁄kg (males) and 13.2 ± 0.71 mg ⁄kg
(females) in Group C, and 21.2 ± 0.77 mg ⁄kg (males) and
21.9 ± 1.02 mg ⁄kg (females) in Group D.
No cat died or became moribund during the study. The
physical examinations [comprising assessment of general health,
integument and mucous membranes, auscultation of heart and
lungs, examination of head (including mouth, eyes and ears)
and central nervous system (including reaction to stimuli and
measurement of body temperature)] revealed no toxicologically
relevant effects in any robenacoxib-treated group, relative to the
placebo group. Vomiting was the most commonly reported
adverse reaction, but the incidence was similar in all groups.
Likewise, soft faeces were reported occasionally with no differ-
ences between groups. Although a heart arrhythmia was
recorded in a single cat, on day 37, it was transient and not
reported on other days.
Body weights and food consumption were recorded on three
occasions in the baseline period and on six occasions (once
weekly) during treatment. Body weights increased slightly or
remained constant throughout the study, with no differences
between groups and likewise no differences in food intake (data
not shown).
For most haematology variables, there were no statistically
significant differences from placebo-treated cats for all roben-
acoxib-treated groups (selected results shown in Table 4, not all
data shown). However, relative to the placebo group, mean
cellular haemoglobin concentration was decreased in Groups C
and D at day 14 but not at day 35. The changes from baseline at
day 14 were slight, comprising +0.22%, )2.98% and )4.02%
for Groups A, C and D, respectively.
Table 3. Summary of body (kg) and organ (g) weights in cats in study 1;
mean (SD), n= 6 unless stated
Organ
Group A
(placebo)
Group B
(robenacoxib
5mg⁄kg)
Group C
(robenacoxib
10 mg ⁄kg)
Initial bodyweight 2.6 (0.3) 2.6 (0.3) 2.6 (0.3)
Final bodyweight 3.0 (0.4) 3.0 (0.4) 2.9 (0.4)
Heart 13.0 (2.4) 12.9 (2.9) 12.5 (1.4)
Kidneys 20.8 (3.8) 22.1 (5.2) 19.4 (4.4)
Liver 85.4 (15.5) 90.8 (19.9) 83.2 (16.8)
Thymus 7.0 (1.0) 5.5 (2.4) 5.7 (1.5)
Spleen 11.5 (2.7) 9.9 (2.9) 10.9 (5.2)
Ovaries (n= 3) 0.46 (0.1) 0.46 (0.1) 0.41 (0.1)
Uterus with cervix (n= 3) 3.9 (0.5) 2.9 (2.0) 2.7 (0.9)
Testes (n= 3) 1.4 (0.4) 1.8 (1.1) 1.2 (0.5)
Thyroid 0.39 (0.08) 0.74 (0.6) 0.61 (0.56)
Adrenals 0.41 (0.05) 0.42 (0.1) 0.40 (0.08)
Pituitary 0.035 (0.01) 0.050 (0.02) 0.038 (0.016)
Brain 26.0 (3.8) 26.2 (1.6) 27.2 (2.8)
Differences between placebo and robenacoxib-treated cats were not
statistically significant.
Table 4. Summary of blood haematology data for cats in study 2; mean (SD), n= 8 unless stated
Variable (units) Day
Group A
(placebo)
Group B
(robenacoxib 2 mg ⁄kg)
Group C
(robenacoxib 6 mg ⁄kg)
Group D
(robenacoxib 10 mg ⁄kg)
Erythrocyte count (RBC) (10
12
⁄L) )12 10.2 (1.2) 9.7 (1.0) 10.1 (0.7) 9.5 (2.2)
)7 9.1 (1.5) 8.7 (0.8) 9.2 (1.0) 8.4 (1.9)
14 8.5 (0.9) (n= 7) 8.2 (1.1) 9.1 (0.8) 8.2 (2.1)
35 8.3 (1.1) 8.2 (1.4) 8.5 (0.6) 7.5 (1.8)
Haemoglobin concentration (m
M
))12 8.1 (0.6) 7.0 (2.1) 7.9 (0.4) 7.4 (1.4)
)7 7.3 (0.9) 6.8 (0.6) 7.2 (0.8) 6.5 (1.1)
14 7.1 (0.6) (n= 7) 6.6 (0.7) 7.2 (0.5) 6.5 (1.3)
35 7.0 (0.6) 6.7 (0.9) 6.8 (0.3) 6.2 (1.2)
Haematocrit (L ⁄L) )12 0.43 (0.03) 0.40 (0.03) 0.42 (0.01) 0.40 (0.06)
)7 0.38 (0.04) 0.35 (0.02) 0.37 (0.04) 0.35 (0.06)
14 0.37 (0.04) (n= 7) 0.35 (0.04) 0.39 (0.03) 0.36 (0.06)
35 0.36 (0.03) 0.35 (0.05) 0.36 (0.01) 0.33 (0.05)
Platelet count (10
9
⁄L) )12 344.4 (87.9) 305.1 (64.8) 359.1 (64.2) 376.0 (151.1)
)7 328.5 (120.0) 325.9 (71.7) 342.9 (118.8) 382.8 (120.3)
14 387.6 (79.7) (n= 7) 353.9 (61.7) 419.0 (77.7) 439.6 (125.0)
35 371.5 (86.7) 379.0 (73.5) 442.5 (71.0) 414.8 (178.3)
Total leucocyte count (10
9
⁄L) )12 14.7 (4.7) 18.5 (9.5) 14.6 (4.5) 17.2 (3.1)
)7 16.4 (3.9) 16.5 (6.2) 13.7 (4.2) 17.0 (3.4)
14 18.3 (4.2) (n= 7) 18.0 (7.9) 17.3 (4.8) 18.8 (5.2)
35 13.8 (3.2) 16.5 (4.0) 13.6 (3.9) 18.7 (6.5)
Differences between placebo and robenacoxib-treated cats were not statistically significant.
6J. N. King et al.
2011 Blackwell Publishing Ltd

For most clinical chemistry variables, there were no significant
differences for all robenacoxib-treated groups relative to placebo.
Selected results are shown in Table 5 (not all data shown).
Although mean AST activity was reduced in Group D cats
relative to placebo (Group A), activity of this enzyme was in fact
reduced in both Groups (A and D) at day 35, relative to baseline,
by 21% in Group A and 29% in Group D. ALT activity was
reduced in Groups B and D, compared with Group A. This
difference was because of no change in baseline values for Group
A cats (+1.8%) and small decreases in Group B ()16%) and D
()26%) animals. In Group B and D cats, GGT activity was
increased compared with values in Group A, but the differences
were due in part to a decrease from baseline values in Group A
at day 14 ()28%), with a small increase in Group B (+28%) at
day 14. There was no increase from baseline in any group
at day 35.
Serum creatinine and urea concentrations are indirect
indicators of integrity of renal function. Compared with Group
A, neither variable was significantly changed at days 14 and 35
in the three robenacoxib-treated groups, with one exception;
creatinine was increased in Group B cats. However, differences
from baseline were slight, increases of 2%, 10%, 3% and 3%
occurring in Groups A, B, C and D, respectively, at day 35.
Relative to Group A, calcium concentration increased in
Group C cats. However, the difference was attributable to greater
decreases in calcium from baseline values in Group A ()4.5%
and )2.2% at days 14 and 35) than in Group D ()1.4% and
+0.4% at days 14 and 35). Although sodium concentration was
increased in Groups C and D relative to Group A, the differences
were apparent only at day 14, when increases from baseline
were +0.05% (Group A), +1.16% (Group C) and +1.95%
(Group D).
Urinalysis revealed a decrease in urine-specific gravity over
time in all groups, but there were no differences between groups.
There were similarly no group differences in other urine
parameters including erythrocytes, leucocytes, epithelial cells,
casts, organisms or crystals.
At necropsy, few macroscopic findings were reported and
none, with the possible exception of the thymus gland, could be
considered treatment related. Small thymuses were noted in
individual cats of all groups, and this was reflected in decreased
thymus weights. These were significantly reduced in all three
groups of robenacoxib-treated cats compared with those in the
placebo group (Table 6). However, when two outlier values were
excluded from Group A, thymus weight differences in the
robenacoxib groups were not significantly different from the
placebo group. There were no other significant differences in
organ weights. Microscopic examination of the following tissues
revealed no treatment-related effects: adrenals, aorta, brain,
caecum, colon, duodenum, epididymides, eyes with optic nerves,
Table 5. Summary of serum clinical chemistry data for cats in study 2; mean (SD), n= 8 unless stated
Variable (units) Day
Group A
(placebo)
Group B
(robenacoxib 2 mg ⁄kg)
Group C
(robenacoxib 6 mg ⁄kg)
Group D
(robenacoxib 10 mg ⁄kg)
Alanine aminotransferase (IU ⁄L) )12 99.4 (37.2) 72.5 (29.0) 86.9 (52.5) 64.8 (18.0)
)7 215.3 (337.3) 68.5 (30.2) 66.9 (28.4) 97.1 (109.6)
14 153.6 (137.9) (n= 7) 56.5 (19.2)* 63.8 (30.0) 45.3 (15.2)*
35 77.5 (50.0) 82.8 (60.0) 47.9 (13.0) 42.5 (16.5)
Aspartate aminotransferase (IU ⁄L) )12 24.5 (7.3) 19.5 (5.0) 19.0 (6.5) 17.9 (6.6)
)7 38.8 (44.7) 20.4 (7.2) 18.8 (8.5) 27.4 (33.1)
14 29.1 (16.0) (n= 7) 22.8 (5.1) 20.3 (6.8) 17.0 (2.8)*
35 20.4 (6.6) 21.8 (10.3) 12.6 (2.6) 12.8 (3.2)*
Alkaline phosphatase (IU ⁄L) )12 195.0 (62.2) 153.6 (37.2) 216.9 (95.5) 159.5 (59.3)
)7 205.0 (61.0) 160.3 (47.1) 222.8 (98.5) 162.6 (51.3)
14 180.7 (62.2) (n= 7) 167.8 (70.0) 187.9 (63.7) 139.4 (38.6)
35 153.8 (64.4) 132.8 (37.1) 164.6 (61.3) 112.0 (32.2)
Gamma-glumatyltransferase (IU ⁄L) )12 11.1 (1.9) 9.4 (2.5) 9.5 (2.2) 8.9 (3.1)
)7 6.3 (1.4) 6.3 (1.3) 5.6 (1.5) 6.6 (1.3)
14 6.1 (2.0) (n= 7) 9.8 (2.0)** 6.3 (1.4) 8.6 (4.2)*
35 6.6 (1.8) 7.4 (2.4)** 6.3 (1.7) 7.3 (1.8)*
Creatinine (l
M
))12 132.3 (26.1) 130.5 (12.7) 134.5 (13.7) 133.6 (21.4)
)7 113.8 (18.7) 122.4 (16.1) 118.8 (8.4) 123.4 (18.2)
14 113.9 (17.l) (n= 7) 129.4 (13.3)* 119.3 (15.1) 119.4 (16.9)
35 124.8 (23.5) 137.6 (12.6)* 130.0 (11.3) 132.3 (19.6)
Blood urea nitrogen (m
M
))12 6.5 (1.8) 6.8 (1.4) 6.4 (1.2) 6.7 (1.3)
)7 6.3 (1.4) 7.1 (1.3) 7.2 (0.9) 6.7 (1.3)
14 8.4 (1.0) (n= 7) 8.6 (1.5) 8.5 (1.3) 9.1 (1.2)
35 9.0 (1.1) 8.9 (0.8) 10.1 (0.8) 9.6 (1.1)
Total protein (g ⁄L) )12 64.4 (1.6) 65.4 (4.2) 65.3 (2.6) 64.8 (5.1)
)7 61.5 (2.4) 60.7 (2.5) 61.4 (2.4) 61.3 (4.2)
14 60.9 (1.8) (n= 7) 61.4 (2.6) 61.0 (2.9) 60.9 (4.8)
35 58.3 (3.0) 62.1 (2.0) 61.2 (2.5) 59.5 (5.3)
Statistical differences between placebo and robenacoxib-treated cats: *P< 0.05; **P< 0.01.
Robenacoxib safety in cats 7
2011 Blackwell Publishing Ltd

femur (distal with articular surface and bone marrow), gall
bladder, heart, ileum, jejunum, kidneys, lacrimal glands, larynx,
liver, lung, lymph nodes (cervical, mesenteric and mediastinal),
mammary area, oesophagus, ovaries, oviducts, pancreas, peri-
pheral nerves, pituitary, prostate, rectum, salivary glands,
skeletal muscle, skin, spinal cord, spleen, sternum (with bone
marrow), stomach, synovial membrane (knee joint), testes,
thymus, thyroid (with parathyroid), tongue, trachea, ureters,
urinary bladder, uterus (with cervix) and vagina.
Blood robenacoxib concentrations
Blood samples collected on day 35 approximately 12 h after
dosing in study 2 were analysed for blood robenacoxib. No drug
was detected in any of the eight placebo-treated cats. Low
concentrations of robenacoxib, ‡3ng⁄mL, were detected in two
of eight cats receiving 2 mg ⁄kg (3 and 3.1 ng ⁄mL), four of eight
cats with 6 mg ⁄kg (3, 3.2, 4.6 and 5.2 ng ⁄mL) and eight of
eight cats with 10 mg ⁄kg (3.9, 4.6, 4.9, 5, 7.1, 9.6, 10.6 and
11.6 ng ⁄mL), indicating rapid clearance of the drug from blood.
PK–PD simulations
Median and 90% tolerance intervals for simulations in 100 cats
for inhibition of COX-1 (TxB
2
used as surrogate) and COX-2
(PGE
2
as surrogate) are illustrated in Figs 1 & 2. For safety
assessments, the median and upper limits of the 90% tolerance
interval are considered. At 2 mg ⁄kg, the median maximum
inhibition of COX-1 was 30.6% and inhibition exceeded 10% for
0.8 h (Fig. 1). In contrast, at the same dosage, the median
maximum inhibition of COX-2 was 97.8% and inhibition
exceeded 50% for 2.3 h (Fig. 2). More prolonged inhibition of
COX-1 and COX-2 occurred at higher dosages but, even at the
10 mg ⁄kg dosage, median COX-1 inhibition did not exceed 50%
for more than 0.5 h and the median COX-2 inhibition exceeded
50% for only 2.9 h. For the upper limit of the 90% tolerance
interval at 10 mg ⁄kg, the peak inhibition of COX-1 was 93.3%,
50% inhibition persisted for 1.5 h, and 10% inhibition persisted
for 3 h. For the upper limit of the 90% tolerance interval at
10 mg ⁄kg, the peak inhibition for COX-2 was 100%, 50%
inhibition persisted for 4.4 h, and 10% inhibition persisted for
6.1 h.
DISCUSSION
Principal findings and context to literature
The primary objective of this study was to evaluate the safety of
robenacoxib tablets in cats with repeated oral administration. It
was concluded that administration of robenacoxib was well
Table 6. Summary of body (kg) and organ (g) weights in cats in study 2; mean (SD), n= 8 unless stated
Organ Group A (placebo)
Group B
(robenacoxib 2 mg ⁄kg)
Group C
(robenacoxib 6 mg ⁄kg)
Group D
(robenacoxib 10 mg ⁄kg)
Final bodyweight 3.4 (1.0) 3.4 (0.8) 3.6 (0.6) 3.2 (0.7)
Heart 14.0 (4.5) 12.8 (3.7) 17.2 (8.6) 14.4 (4.4)
Kidneys 21.2 (8.5) 22.0 (8.7) 21.5 (5.8) 20.7 (10.3)
Liver 99.5 (32.3) 96.6 (30.3) 97.2 (18.7) 96.3 (29.8)
Thymus 2.7 (1.9) 1.5 (0.5)* 1.3 (0.5)* 1.3 (0.5)*
Thymus

1.8 (0.9) (n= 6) 1.5 (0.5) 1.3 (0.5) 1.3 (0.5)
Spleen 21.1 (8.9) 19.9 (12.4) 21.1 (8.2) 18.1 (10.3)
Testes (n= 4) 2.5 (0.5) 2.7 (0.3) 2.7 (0.3) 3.1 (1.2)
Ovaries (n= 4) 0.36 (0.1) 0.36 (0.05) 0.40 (0.05) 0.35 (0.03)
Uterus (n= 4) 3.2 (2.1) 2.8 (0.7) 3.1 (0.4) 2.7 (1.0)
Thyroid 0.26 (0.1) 0.35 (0.3) 0.25 (0.06) 0.22 (0.09) (n=7)
Adrenals 0.42 (0.1) 0.42 (0.1) 0.35 (0.07) 0.41 (0.2)
Pituitary 0.040 (0.001) 0.032 (0.01) 0.035 (0.01) 0.030 (0.01)
Brain 28.1 (2.2) 26.5 (2.7) 27.0 (2.2) 27.4 (1.7)
Statistical differences between placebo and robenacoxib-treated cats: *P< 0.01.

Excluding animal nos. 3 and 4 in Group A only as outliers.
Inhibition TxB2 (%)
0
10
20
30
40
50
60
70
80
90
100
Time (h)
0 4 8 12162024
Dosage 2 mg/kg 6 mg/kg 10 mg/kg
Fig. 1. Simulated inhibition of serum TxB
2
, as an index of cyclooxy-
genase-1 inhibition, after oral administration to fasted cats of robenac-
oxib at three dosages twice daily (study 2). Data are the median (full line)
and 90% tolerance intervals (dotted lines).
8J. N. King et al.
2011 Blackwell Publishing Ltd

tolerated at 5–10 mg ⁄kg once daily for up to 1 month (study 1)
and at 2–10 mg ⁄kg twice daily for up to 6 weeks (study 2), with
no biologically relevant treatment-related toxicity. This conclu-
sion was based on general observations of health, food and water
consumption, body weight changes, haematological, clinical
chemistry and urine analyses, and by post mortem organ weight,
gross pathology and histopathology assessments. Although
statistically significant differences from placebo were observed
for some haematological and clinical chemistry variables, it was
judged that no biologically relevant effects of robenacoxib were
present for the following reasons: (i) some changes were too
small numerically to signify toxicity; (ii) some significant
differences from placebo were because of a change with time in
placebo values, such that changes from baseline (predosing)
values occurred with placebo treatment but not in robenacoxib-
treated cats; (iii) some plasma enzyme activity changes in
robenacoxib-treated cats were because of decreases relative to
placebo treatment, when it is increased activity which signifies
potential toxicity; (iv) some differences from placebo treatment
were not dose-related but occurred, for example, in low and ⁄or
medium robenacoxib groups but not in cats receiving the high
dosage of robenacoxib; and (v) some differences from placebo
occurred in only one of the two studies. For example, some
haematological and clinical chemistry variable differences
between placebo and robenacoxib-treated groups differed
between study 1 and study 2.
The principal targets for toxicity of nonselective NSAIDs are
damage to the gastrointestinal tract, kidney and liver, and
inhibition of blood clotting (Warner et al., 1999; Flower, 2003).
No evidence of toxicity of robenacoxib to any of these systems
was detected in the cats in either study 1 or 2. Regarding the
gastrointestinal tract, no changes in serum total protein or
albumin concentrations, which would indicate protein-losing
enteropathy, were detected, and no gross or histopathological
signs of damage or ulceration were observed. In addition, no
gross or histopathological changes in the kidney or liver, and no
changes in biochemical indicators of deterioration of kidney
function (serum creatinine or urea concentration) or liver
damage (serum transferase activities) were detected. Finally, at
no dosage did robenacoxib affect the activated partial thrombo-
plastin time. Similar results were reported recently with
robenacoxib in dogs (King et al., 2011).
It is recognised that the methods used for testing robenacoxib
safety were not in all cases the most sensitive. For example, we
did not assess, during the in-life phase, the appearance of the
stomach via gastroscopy or check for the presence of occult
faecal blood (only gross appearance of the faeces was checked).
In addition, reliance was placed on plasma creatinine concen-
tration as an approximate and indirect measure of glomerular
filtration rate. Nevertheless, we conclude that the methods were
sufficiently sensitive and robust to conclude that robenacoxib
produced no biologically relevant toxicity, even at the highest
dosage of 20 mg ⁄kg daily for 42 days.
Robenacoxib tablets are registered in the European Union for
acute musculoskeletal disorders in cats at a dosage of
1–2.4 mg ⁄kg for up to 6 days. Therefore, the upper dosages
tested in this study represent 4–10 (28 days, study 1) and 8.3–
20 (42 days, study 2) multiples of the clinically recommended
daily dosage. The results of this study indicate that the safety
index of robenacoxib is high in cats, and to our knowledge is
higher than published for any other NSAID in this species
(Lascelles et al., 1995, 2007; Runk et al., 1999; Papich, 2008).
Cats were dosed with robenacoxib tablets approximately 1 h
before feeding in study 2. In study 1, cats were fed ad libitum, and
therefore it is unlikely that a large amount of food was present in
the stomach at the time of each dosing. The oral bioavailability of
robenacoxib from tablets is reduced in cats when given with the
entire daily ration, but not when co-administered with one third
of the daily ration (J.N. King, M. Jung, M.P. Maurer, V.B. Schmid,
W. Seewald & P. Lees, In preparation). In addition, the T
max
of
robenacoxib in blood is 0.5 h after oral administration in fasted
cats (J.N. King, M. Jung, M.P. Maurer, V.B. Schmid, W. Seewald &
P. Lees, In preparation). Therefore, it is concluded that oral
bioavailability was probably optimal in both studies, and
consistent with the label recommendations in the European
Union, i.e., to administer Onsior

(Novartis Animal Health)
tablets to cats either without food or with a small quantity of food.
Different tablet formulations were used in studies 1 and 2,
namely nonflavoured lactose tablets in study 1 and the flavoured
tablets presently marketed (Onsior

; Novartis Animal Health) in
study 2. However, this should have had no impact on the
conclusions of the studies as the two formulations have been
shown to be bioequivalent for both C
max
and AUC (Novartis
Animal Health data on file).
PK–PD simulations
It has been concluded in humans that inhibition of COX-1 makes
the major and possibly sole contribution to the gastrointestinal
toxicity of nonselective NSAIDs (Warner et al., 1999), although
Inhibition PGE2 (%)
0
10
20
30
40
50
60
70
80
90
100
Time (h)
04812162024
Dosage 2 mg/kg 6 mg/kg 10 mg/kg
Fig. 2. Simulated inhibition of plasma prostaglandin E
2
, as an index of
cyclooxygenase-2 inhibition, after oral administration to fasted cats of
robenacoxib at three dosages twice daily (study 2). Data are the median
(full line) and 90% tolerance intervals (dotted lines).
Robenacoxib safety in cats 9
2011 Blackwell Publishing Ltd

Wallace et al. (2000) concluded in rats that NSAID-induced
gastrointestinal damage required simultaneous inhibition of both
COX-1 and COX-2. The exact extent of inhibition of COX-1
needed to induce damage to the gastrointestinal tract is not
known (Warner et al., 1999), although it is logical that damage
will be a function of the duration as well as magnitude of
inhibition of COX-1. To explore the extent and duration of COX
isoforms inhibition by robenacoxib in cats, simulations of the
predicted inhibition of COX-1 and COX-2 in the central
compartment with the dosages of robenacoxib tested in this
study were created. All of these dosages had produced no
detectable toxicity to the gastrointestinal tract. Similar simula-
tions were reported recently for robenacoxib in dogs (King et al.,
2011). Even though robenacoxib is a highly selective inhibitor of
COX-2 in cats, with a potency ratio for 50% inhibition of COX-
1:COX-2 exceeding 500:1 (Giraudel et al., 2009a), transient
inhibition of COX-1 was predicted at early time points after
dosing with the high dosages of robenacoxib used in this study
(Fig. 1). In all cats, there was greater inhibition of COX-2
simultaneously with the slight inhibition of COX-1 (Fig. 2). As
no biologically relevant toxicity was detected with any dosage of
robenacoxib, it is concluded that 50% inhibition of COX-1 for
1.5 h, or 10% for 3 h (the upper limits of the 90% tolerance
interval at the highest robenacoxib dosage tested of 10 mg ⁄kg)
with maximal inhibition and somewhat longer inhibition of
COX-2 with robenacoxib twice daily did not affect the cats
adversely.
CONCLUSIONS
Robenacoxib tablets had an excellent safety profile in young
healthy domestic short-haired cats, when administered at daily
dosages up to 10 mg ⁄kg for 28 days and up to 20 mg ⁄kg for
42 days. The absence of toxicity occurred even though the
greatest PD effect of robenacoxib was predicted to be 50%
inhibition of COX-1 for 1.5 h, or 10% inhibition for 3 h, together
with marked inhibition of COX-2. The results of this feline study
support the conclusion, as previously reported for rats and dogs
(King et al., 2009, 2011), that the excellent safety profile of
robenacoxib may be because of a combination of PD (high
selectivity for COX-2) and PK (rapid clearance from the central
compartment with longer residence times at sites of inflamma-
tion) properties. It should be noted that the safety studies
reported in this paper were conducted in healthy young domestic
short-haired cats. The same conclusions might not apply to other
breeds, or to cats with pre-existing damage to the gastrointes-
tinal tract, kidney or liver. Results of field studies in cats with
naturally occurring diseases are required to complete the safety
profiling of robenacoxib in cats.
REFERENCES
Cox, S.R., Lesman, S.P., Boucher, J.F., Krautmann, M.J., Hummel, B.D.,
Savides, M., Marsh, S., Fielder, A. & Stegemann, M.R. (2010) The
pharmacokinetics of mavacoxib, a long-acting COX-2 inhibitor, in
young adult laboratory dogs. Journal of Veterinary Pharmacology and
Therapeutics,33, 461–470.
Flower, R.J. (2003) The development of COX2 inhibitors. Nature Reviews
Drug Discovery,2, 179–191.
Giraudel, J.M., Toutain, P.-L., King, J.N. & Lees, P. (2009a) Differential
inhibition of cyclooxygenase isoenzymes in the cat by the NSAID robe-
nacoxib. Journal of Veterinary Pharmacology and Therapeutics,32, 31–40.
Giraudel, J.M., King, J.N., Jeunesse, E.C., Lees, P. & Toutain, P.-L. (2009b)
Use of a pharmacokinetic ⁄pharmacodynamic approach in the cat to
determine a dosage regimen for the COX-2 selective drug robenacoxib.
Journal of Veterinary Pharmacology and Therapeutics,32, 18–30.
Gunew, M.N., Menrath, V.H. & Marshall, R.D. (2007) Long-term safety,
efficacy and palatability of oral meloxicam at 0.01–0.03 mg ⁄kg for
treatment of osteoarthritic pain in cats. Journal of Feline Medicine and
Surgery,10, 235–241.
Jung, M., Lees, P., Seewald, W. & King, J.N. (2009) Analytical determi-
nation and pharmacokinetics of robenacoxib in the dog. Journal of
Veterinary Pharmacology and Therapeutics,32, 41–48.
King, J.N., Dawson, J., Esser, R.E., Fujimoto, R., Kimble, E.F., Maniara,
W., Marshall, P.J., O’Byrne, L., Quadros, E., Toutain, P.-L. & Lees, P.
(2009) Preclinical pharmacology of robenacoxib: a novel selective
inhibitor of cyclooxygenase-2. Journal of Veterinary Pharmacology and
Therapeutics,32, 1–17.
King, J.N., Rudaz, C., Borer, L., Jung, M., Seewald, W. & Lees, P. (2010)
In vitro and ex vivo inhibition of canine cyclooxygenase isoforms by
robenacoxib: a comparative study. Research in Veterinary Science,88,
497–506.
King, J.N., Arnaud, J.P., Goldenthal, E.I., Gruet, P., Jung, M., Seewald, W.
& Lees, P. (2011) Robenacoxib in the dog: target animal species safety
in relation to extent and duration of inhibition of COX-1 and COX-2.
Journal of Veterinary Pharmacology and Therapeutics,34, 298–311.
Lascelles, B.D., Cripps, P., Mirchandani, S. & Waterman, A.E. (1995)
Carprofen as an analgesic for postoperative pain in cats: dose titration
and assessment of efficacy in comparison to pethidine hydrochloride.
Journal of Small Animal Practice,36, 535–541.
Lascelles, B.D., Court, M.H., Hardie, E.M. & Robertson, S.A. (2007)
Nonsteroidal anti-inflammatory drugs in cats: a review. Veterinary
Anaesthesia and Analgesia,34, 228–250.
Livingston, A. (2010) Pain and analgesia in domestic animals. In
Handbook of Experimental Pharmacology: Comparative and Veterinary
Pharmacology. Eds Cunningham, F.M., Elliott, J. & Lees, P., pp. 159–
190. Springer Verlag, London.
McCann, M.E., Rickes, E.L., Hora, D.F., Cunningham, P.K., Zhang, D.,
Brideau, C., Black, W.C. & Hickey, G.J. (2005) In vitro effects and in
vivo efficacy of a novel cyclooxygenase-2 inhibitor in cats with
lipopolysaccharide-induced pyrexia. American Journal of Veterinary
Research,66, 1278–1284.
Papich, M.G. (2008) An update on nonsteroidal anti-inflammatory drugs
(NSAIDs) in small animals. Veterinary Clinics of North America Small
Animal Practice,38, 1243–1266.
Pelligand, L., King, J.N., Toutain, P.L. & Lees, P. (2009) In vivo COX-2
selectivity of robenacoxib in a feline tissue cage model of inflammation.
Journal of Veterinary Pharmacology and Therapeutics,32 (Suppl. 1),
103–104.
Roberts, E.S., Van Lare, K.A., Marable, B.R. & Salminen, W.F. (2009)
Safety and tolerability of 3-week and 6-month dosing of Deramaxx
(deracoxib) chewable tablets in dogs. Journal of Veterinary Pharmaco-
logy and Therapeutics,32, 329–337.
Robertson, S.A. (2005) Managing pain in feline patients. Veterinary
Clinics Small Animal Practice,35, 129–146.
Robertson, S.A. (2008) Managing pain in feline patients. Veterinary
Clinics Small Animals,38, 1267–1290.
10 J. N. King et al.
2011 Blackwell Publishing Ltd

Runk, A., Kyles, A.E. & Downs, M.O. (1999) Duodenal perforation in a
cat following the administration of nonsteroidal anti-inflammatory
medication. Journal of the American Animal Hospital Association,35,
52–55.
Schmid, V.B., Seewald, W., Lees, P. & King, J.N. (2010) In vitro and
ex vivo inhibition of COX isoforms by robenacoxib in the cat: a
comparative study. Journal of Veterinary Pharmacology and Therapeutics.,
33, 444–452.
Silber, H.E., Burgener, C., Letellier, I.M., Peyrou, M., Jung, M., King, J.N.,
Gruet, P. & Giraudel, J.M. (2010) Population pharmacokinetic analysis
of blood and joint synovial fluid concentrations of robenacoxib from
healthy dogs and dogs with osteoarthritis. Pharmaceutical Research,27,
2633–2645.
Wallace, J.L., McKnight, W., Reuter, B.K. & Vergnolle, N. (2000) NSAID-
induced gastric damage in rats: requirement for inhibition of both
cyclooxygenase 1 and 2. Gastroenterology,119, 706–714.
Warner, T.D., Giuliano, F., Vojnovi, I., Bukasa, A., Mitchell, J.A. &
Vane, J.R. (1999) Nonsteroid drug selectivities for cyclo-oxygenase-
1 rather than cyclo-oxygenase-2 are associated with human
gastrointestinal toxicity: a full in vitro analysis. Proceedings of the
National Academy of Science of the United States of America,96,
7563–7568.
Robenacoxib safety in cats 11
2011 Blackwell Publishing Ltd