Download Intrathecal Ketorolac in Dogs and Rats

TOXICOLOGICAL SCIENCES 80, 322–334 (2004)
DOI: 10.1093/toxsci/kfh168
Advance Access publication May 12, 2004
Intrathecal Ketorolac in Dogs and Rats
Tony L. Yaksh,*,1 Kjersti A. Horais,* Nicolle Tozier,* Michael Rathbun,* Phil Richter,* Steve Rossi,* Marjorie Grafe,†
Chuanyao Tong,‡ Carol Meschter,§ J. Mark Cline,‡ and James Eisenach‡
*University of California, San Diego, Department of Anesthesiology, Research Laboratory-0818, 9500 Gilman Drive, La Jolla, California 92093-0818; †Oregon
Health and Science University, Department of Pathology, Portland, Oregon 97239; ‡Wake Forest University School of Medicine, Department of Anesthesiology,
Winston-Salem, North Carolina 27157; and §Comparative Biosciences, Inc., 2672 Bayshore Parkway, Suite 515, Mountain View, California 94043
Received February 13, 2004; accepted May 2, 2004
This study was conducted to assess spinal safety of the cyclooxygenase inhibitor ketorolac in dogs and rats. Beagle dogs were
prepared with lumbar intrathecal catheters and received continuous
spinal infusions of 5 mg/ml ketorolac (N 5 6), 0.5 mg/ml ketorolac
(N 5 8), or saline vehicle (N 5 6) at 50 ml/h (1.2 ml/day) for 28 days.
No systematic drug or dose-related changes were observed in motor
function, heart rate, or blood pressure. Histological examination
revealed a mild pericatheter reaction in all groups with no drug
or dose related effect upon spinal pathology at the lumbar site of
highest drug concentration. Cisternal CSF protein was elevated for
all treatment groups at necropsy, and cisternal glucose was within
normal range for all treatment groups, though three dogs displayed
decreases in cisternal glucose. Significant reductions in hematocrit
were noted, and increased incidence of gastric bleeding at necropsy
was observed in animals receiving ketorolac. Intrathecal ketorolac
kinetics revealed a biphasic clearance: t1/2s 5 10.3 and 53 min,
respectively. After initiation of infusion (0.5 mg and 5 mg/ml/
50 ml/h), lumbar CSF concentrations of ketorolac were 3.8 and
52.7 mg/ml, respectively. Bolus and continuous infusion of intrathecal ketorolac resulted in significant reduction of lumbar CSF
PGE2 concentrations. In rats, with intrathecal catheters, four
daily bolus deliveries of saline or ketorolac (5 mg/ml/10 ml) had
no effect upon spinal histology or upon spinal cord blood flow.
These data indicate that intrathecal ketorolac in two species at
the dose/concentrations employed does not induce evident spinal
pathology but diminishes spinal prostaglandin release.
Key Words: nonsteroidal anti-inflammatory drug (NSAID);
spinal; cyclo-oxygenase; COX; prostaglandin.
Following tissue injury and inflammation a behaviorally
defined hyperalgesia can be identified in animal models and in
humans (Yaksh et al., 1999a). Current thinking suggests that this
behavioral syndrome results from the chronic activation of primary afferent input and/or the presence of circulating cytokines
that initiate a spinal cascade that serves to sensitize spinal nociceptive processing. An important component of this cascade
appears to be the synthesis and extracellular movement of
1
To whom correspondence should be addressed. Fax: (336) 716-3220.
E-mail: [email protected].
Toxicological Sciences vol. 80 no. 2
#
prostaglandins through an activation of spinal cyclooxygenase
(Svensson and Yaksh, 2002, Yaksh et al., 2001). Thus, previous
studies have shown that a variety of COX inhibitors, such as
acetylsalicylate, ibuprofen, and ketorolac given intrathecally in
rat models will diminish hyperalgesic states at intrathecal doses
thatare100to300timessmallerthanthoserequiredaftersystemic
delivery and that have no evident effects after systemic delivery
(MalmbergandYaksh,1992a,b;Yaksh,1982;Yakshetal.,2001).
Ketorolac is a water-soluble nonsteroidal anti-inflammatory
drug (NSAID) with potent inhibitor effects upon both cyclooxygenase isozymes. Administered systemically in humans,
ketorolac has potent antihyperalgesic actions in postoperative
pain (Gillies et al., 1987). Its limiting effects are gastrointestinal
ulceration, effects upon renal blood flow, and depressed platelet
function. These effects are believed to be common to the blockade of peripheral systemic cyclooxygenase isozyme activity
(Eisenberg et al., 1994). These potent effects of cyclo-oxygenase
(COX) inhibitors on hyperalgesic states in humans suggest that,
as in preclinical models, the spinal delivery of COX inhibitors
may also display significant efficacy with diminished systemic
side effects.
To consider the possible reaction of intrathecal ketorolac in
humans, it is necessary to establish the pharmacokinetics and
the effects upon spinal cord after intrathecal delivery in
well-defined preclinical models that are believed to robustly
define the safety (or lack thereof) of the test articles in question.
These studies should be undertaken at the maximum usable
formulation of the GMP (Good Manufacturing Practice)
product. The present study was conducted to obtain such data.
The dog and rat are species routinely used to satisfy regulatory
requirements for drug safety evaluation and are chosen because
there is considerable knowledge of these species’ responses to a
wide variety of intrathecal drugs (Yaksh and Malkmus, 1999;
Yaksh et al., 1999b).
MATERIALS AND METHODS
All animal work was carried out according to protocols approved by
the Institutional Animal Care and Use Committee of the University of California,
Society of Toxicology 2004; all rights reserved.
323
INTRATHECAL KETOROLAC IN DOGS
San Diego and was in compliance with the USDA Animal Welfare Act (USDA
Title 9 Code of Federal Regulations Part 3, Federal Register 15 February 1991).
within 30 s were included in the study. After surgery and testing with
lidocaine, the animals recovered for at least five days. Animals with any
evident neurologic deficit were excluded from the study.
Animals
Test Article
Dogs. Beagle dogs (Covance, Inc., Princeton, NJ and Marshall Farms,
Inc., North Rose, NY), 10 to 24 months in age and weighing 6 to 12 kg, were
individually housed in runs with wood shavings and given ad libitum access to
food and water with a daily 12-h light/dark cycle. Kennels were maintained
within the range of 62–82 F. Animals were adapted for a minimum of 10 days
prior to surgery. A nylon vest (Alice King Chatham Medical Arts, Hawthorne,
CA) was placed on each dog approximately 48 h prior to intrathecal catheter
placement surgery for acclimation.
Rats. Male Sprague-Dawley rats (Harlan-Industries, Indianapolis, IN),
250–370 g, were individually housed and given ad libitum access to food
and water in a regulated temperature (66–76 F) and illumination (12 h light/
12 h dark) environment. Body weight was assessed before implantation of the
intrathecal catheter and daily during study drug treatment.
Surgical Preparation
Dog. To permit chronic delivery of drug into the lumbar intrathecal
space, a catheter was passed from the cisterna magna to the lumbar intrathecal
space. This placement was accomplished approximately 72 h prior to intrathecal dosing. The antibiotic sulfamethoxazole-trimethoprim (240 mg tablet,
15–25 mg/kg, po, twice daily) was given 48 h prior to surgery and for
48 h after surgery. Dogs received atropine (0.04 mg/kg, im) prior to sedation
with xylazine (1.5 mg/kg, im). Anesthesia was induced with isoflurane, 1–2%
in N2O/oxygen (60/40%) with a mask followed by endotracheal intubation
and spontaneous ventilation. Animals were continuously monitored for
oxygen saturation, inspired and end tidal values of isoflurane, CO2, N2O
and oxygen and heart and respiratory rates.
Surgical areas were shaved and prepared with chlorhexadine scrub and
solution. Using sterile technique, the cisterna magna was exposed and a
small incision (1–2 mm) was made in the dura. APE-10 l catheter (fabricated
with polyethylene tubing and sterilized by electron beam irradiation) was
inserted and passed caudally approximately 40 cm to a level corresponding
to the L2–3 segment. Dexamethasone sodium phosphate (0.25 mg/kg, im) was
administered just after catheter placement. The external catheter was then tunneled subcutaneously to exit at the left scapular region. Upon closure of the
incision, isoflurane was discontinued and butorphanol tartrate (Torbugesic; 0.04
mg/kg, im) was administered to relieve postoperative pain as necessary.
For pharmacokinetic studies, dogs were prepared as described above, but a
second catheter (polyethylene, PE50) was also placed in addition to the
infusion catheter for purposes of lumbar cerebrospinal fluid (CSF) sampling.
The sampling catheter was passed such that the tip was approximately 2 cm
cephalad to the infusion site. The catheter was externalized similarly on the
contralateral side.
Following recovery, the nylon vest was again placed on the animal (animals having been previously acclimated to the vest prior to surgery), and an
infusion pump (PANOMAT C-10; Disetronic Medical Systems, Saint Paul,
MN) secured in the vest pocket where it was connected to the externalized
end of the PE-10 intrathecal catheter. An infusion of 0.9% (w/v) Sodium
Chloride for Injection, USP at 50 ml/h was initiated to ensure catheter patency
prior to infusion of Test Article.
Rats. For neurotoxicologic assessment, rats were anesthetized with 2–
4% halothane in oxygen/air. Polyethylene catheters were inserted through a
small incision in the atlanto-occipital membrane and passed 8 cm caudally
to the level of the lumbar enlargement, as previously described (Yaksh and
Rudy, 1976). To confirm correct placement of the catheters, 10 ml of 2%
lidocaine was injected, followed by a 10 ml 0.9% saline flush the day after
surgery. All animals developing a bilateral motor block of the hind limbs
Ketorolac (5 mg/ml; Acular PFTM ; Allergan, Irvine, CA) was supplied as
a sterile additive-free saline solution. Sterile saline (0.9% (w/v) Sodium
Chloride for Injection, USP, Abbott Laboratories, North Chicago, IL; 10 ml/vial)
was used as the control solution and as flush for the catheter. From each cartridge
change of the test article treatment group, drug samples were randomly
collected during the filling of the pump cartridges and analyzed for test article
concentration.
Dog Procedures
These studies had three phases: dose ranging, safety, and pharmacokinetics.
Table 1 indicates the dog treatment assignments.
Dose-ranging phase. Two dogs were assigned to receive the maximum
concentrations (5 mg/ml) at the maximum intrathecal infusion rate routinely
employed for spinal delivery in this model (100 ml/h). In the event of deleterious
side effects, the animal was terminally anesthetized and underwent necropsy and
histopathology.
Safety phase. Dogs were randomly assigned to receive the lumbar intrathecal infusion of the maximum available concentration of ketorolac (5 mg/ml of
TABLE 1
Summary of Dog Group Assignment and Treatment
Dog ID
Gender
Weight
(kg)
Treatment
group
3376087
3381293
Male
Male
7.2
9.4
5 mg/ml Ketorolac
test article at 100 ml/h
(2.4 ml/day)
385 8979
389 3243
389 4631
OFR-1
YZR-1
OER-1
389 2328
388 4376
Male
Male
Male
Male
Male
Male
Male
Male
13.0
11.8
12.8
9.0
9.2
9.6
12.4
12.2
0.5 mg/ml Ketorolac
test article at 50 ml/h
(1.2 ml/day)
CPH AVL
CPH AYU
CPH AUI
CPH AZX
CPH ATY
CPH AER
Male
Male
Male
Female
Female
Female
11.0
10.4
10.6
9.0
7.6
6.0
5 mg/ml Ketorolac
test article at 50 ml/h
(1.2 ml/day)
CPH AWK
CPH AVH
CPH AXA
CPH ASL
CPH ATW
CPH APB
Male
Male
Male
Female
Female
Female
7.8
11.4
10.2
7.8
9.8
7.6
Saline control article
at 50 ml/h (1.2 ml/day)
CRDAIC
CRDATI
3359565
3377130
Male
Male
Male
Male
11.8
12.3
8.6
8.6
Pharmacokinetics-prostanoids
sampling
324
YAKSH ET AL.
ketorolac), 0.5 mg/ml ketorolac, or an equal volume of saline delivered at 50 ml/h
for a minimum of 28 days, unless euthanasia was deemed necessary.
Pharmacokinetics phase. Serum and lumbar intrathecal CSF samples (0.5
and 0.3 ml, respectively) were taken before and at intervals after the intrathecal
injection of ketorolac (5 mg/ml) as a 1 ml bolus and assayed for ketorolac. After a
3–5 day interval, intrathecal infusion of ketorolac (5 mg/ml) at a rate of 50 ml/h
was initiated. Before and at 6–12 h after the initiation of infusion, serum, and
lumbar CSF samples were taken and assayed for ketorolac. In addition, CSF
samples were split and assayed for PGE2.
Dog Clinical Observations
A physical examination of each dog was conducted prior to its inclusion into
the study. Animals were observed for health status and clinical signs at least twice
daily. Catheters and pumps were checked at least once daily.
Neurological examination. For safety phase animals, a neurological examination was undertaken prior to the initiation of drug delivery and within three
days prior to sacrifice (e.g., on or about study Day 25–28). Examinations were
performed without knowledge of treatment group.
Periodic observation. After surgery, each dog’s rectal temperature was
recorded daily. Specific behavioral indices chosen to assess each animal’s
state of arousal, muscle tone, and motor coordination were undertaken twice
daily throughout the in-life phase of the study, as described in detail elsewhere
(Yaksh et al., 1997). Body weights, heart rate, and blood pressure measurements
were recorded on Days 5, 3, prior to initiation of test or control article, at
regular 4-day intervals, and on the day of necropsy.
Measurement of cardiovascular parameters. Measurements were made
using a tail cuff manometer (Dinamap 8100, Criticon). Respiration rates were
measured by observation of chest expansion and contraction.
Nociception response. The thermally evoked skin twitch response was
measured using a probe with approximately 1 cm surface area maintained at
approximately 62.5 6 0.5 C with a feedback controller. The probe was applied
to two shaved areas of the back (right and left lateral) at the thoracolumbar level of
the spine. Typically, this results in a brisk contraction of the local musculature
within 1–3 s of probe placement. Upon the appearance of this response, the probe
was removed and the latency to response recorded. Failure to respond within 6 s
was cause to remove the probe and assign that value (6 s) as the latency.
For analysis purposes, the nociceptive response was presented as the mean of
the two latencies.
Dog Laboratory Measurements
Blood analyses. Blood was collected from each dog by cephalic venipuncture, after an overnight fast of approximately 16 h (no food was allowed but free
access to water was provided) for a complete blood count and blood chemistry
panel.
Complete blood counts included the following analyses: leukocyte, erythrocyte, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular
hemoglobin, mean corpuscular hemoglobin concentration, platelet count, differential leukocyte count (absolute and relative), reticulocyte count, and blood cell
morphology evaluation.
Blood chemistry panels included the following analyses: sodium, cholesterol,
potassium, aspartate aminotransferase, chloride, alanine aminotransferase, total
protein, alkaline phosphatase, albumin, triglycerides, calcium, globulin (calculated), phosphorous, albumin/globulin ratio, creatinine, total and direct bilirubin,
glucose, and urea nitrogen. After initiation of test or control article, weekly
hematocrit values were determined on site over the duration of the safety
phase of the study.
Drug samples. To determine drug concentration in the blood, duplicate
cephalic blood samples (41 ml) were collected on wet ice and prepared as
serum, then transferred to Eppendorf tubes and stored at 20 6 10 C until
analysis. To determine test article concentration in the cerebrospinal fluid
(CSF), samples of CSF (approximately 0.3 ml) were collected on ice then
transferred to Eppendorf tubes and stored at 20 6 10 C until analysis.
Ketorolac analysis. Serum and CSF samples were quantitatively extracted
by C-18 reverse phase cartridge chromatography and eluted with acetonitrile.
Concentrated eluates were injected onto an HPLC equipped with a Phenomenex
‘‘Prodigy’’ C-18 reverse phase column (250 mm 3 4.6 mm). Peaks were detected
withanAgilent Model1100UV detectorset ata wavelengthof313.Allunknowns,
standards, and controls contained an equal quantity of internal standard, 200 ng of
indoprofen. With this assay, a single peak for ketorolac is found at the retention
time of10.68min,the internalstandardindoprofenelutesat12.09min.Plasma and
CSF extracts showed no interfering chromatography throughout the integration
time period. The absolute sensitivity of the ketorolac assay was 5 ng/ml, and the
coefficient of variation was 510% within the concentration range 5–500 ng/ml.
Dog Necropsy
Allanimalswere anesthetizedwith anivdoseofsodiumpentobarbital(35 mg/kg
or dose to effect). Blood and cisternal CSF (by percutaneous puncture with a 22 ga
spinal needle) were then sampled. The aorta was then catheterized though a thoracotomy and the blood cleared by perfusion with approximately four liters of 0.9%
(w/v)salinefollowedbyapproximatelyfourlitersof10%neutralbufferedformalin.
Spinal dissection. After perfusion, the dura of the spinal cord was exposed
by laminectomy. Methylene blue dye was injected through the intrathecal catheter to determine the catheter patency, the location of the tip, and dye spread
around the tip. The condition of the spinal cord was noted and a photographic
(digital) record of the spinal cord/catheter relationship was made. The cord was
cut (taking care to keep the dura intact) into four sections: cervical (section A),
thoracic (section B), high lumbar (catheter tip region, section C), and lower
lumbar (just caudal to catheter tip section D), for histopathological analysis.
In addition to the spinal cord, portions of gastrointestinal tract, kidney, and
bladder were also retained
Tissue preparation. Formalin-fixed tissues were trimmed into cassettes,
processed and embedded in paraffin, sectioned at approximately 3–5 mm and
stained with hematoxylin and eosin. In addition, special stains were performed on
the spinal cord in the block at the catheter tip region (block C). These sections
were immunohistochemically stained for glial fibrillary acidic protein (GFAP)
(DAKO, Carpenteria, CA; Kit 20334, Lot # CH096302), Gram stain
(McDonald’s Gram Stain, American Master, Lodi, CA; Kit # KTMGS) for
bacteria and Grolcott’s Methenamine Silver (GMS), (Grolcott’s Methenamine
Silver Stain Kit # KTMGSMIC, American Master, Lodi, CA) for fungi. Spinal
cord sections were evaluated by light microscopy independently by two
pathologists (C.M. and M.G.) without knowledge as to treatment. A text description was written for each section. The severity of the inflammation was graded on
a scale of 0–3 for increasing severity, and the animals were then forced ranked in
order of increasing severity. To more precisely rank the pathology scores, all
slides were re-reviewed, still without knowledge of treatment groups, and within
each score category an additional increment of 60.25 could be given if
the reaction was slightly more or less severe than that of animals with whole
number scores.
Rat Procedures
There were two phases to the rodent studies: intrathecal toxicity and spinal
cord blood flow.
Intrathecal toxicity. To assess the response of the intrathecal space to
repeated bolus delivery, rats were randomized to receive 50 mg ketorolac or
an equal volume (10 ml) 0.9% saline daily by intrathecal bolus for four days
(n 5 10 per group). Arousal, motor coordination, paw withdrawal latency to
radiant heat, and general behavior were assessed daily after each injection, as
previously described (Yaksh et al., 1995). On day 6, animals were deeply
INTRATHECAL KETOROLAC IN DOGS
anesthetized with pentobarbital, perfused transcardially with saline followed by
paraformaldehyde, the spinal cords removed, fixed, then sectioned and stained
with hematoxylin-eosin. Sections of spinal cord near the catheter tip in the lumbar
region, and distant from the tip in the thoracic region and cervical region were
examined by a pathologist unaware of treatment group.
Spinal cord blood flow. Rats were anesthetized with sevoflurane. Femoral
venous and arterial catheters were inserted; then a tracheotomy was performed
for controlled ventilation. A small laminectomy was performed at the atlantooccipital membrane, the dura excised, and a tissue well created. A PE 50 catheter
was tunneled under the skin of the back, exiting above the spinal cord surface, and
the well perfused with saline solution through this catheter. Laser-doppler flow
was performed by placement of a flowmeter probe directly over the spinal cord
surface. Following baseline measurements, the spinal cord was continuously
perfused with saline, followed by ketorolac, 2.5 and then 5 mg/ml. Blood pressure
and heart rate were monitored throughout the experiment.
Statistical Analysis
Data of the respective observations (e.g., mmHg, beat/min, mg/ml) are
expressed as means and SEs, or as medians and quartiles. Continuous normal
data were compared using repeated measures and one-way ANOVA with post
hoc Bonferroni tests for multigroup comparisons. Distribution-free testing of
data, such as ranked spinal cord pathology, was performed using the Jonckheere
test for comparison of three or more groups with post hoc Mann-Whitney U
tests for multigroup comparisons. All statistical comparisons were made at the
p 5 0.05 level of significance. Pharmacokinetic data were analyzed using PK
Solutions 2.0 (#1999, Summitt Research Services, Ashland, OH) statistical
software for noncompartmental pharmacokinetic data analysis. Analysis
included half-lives, descriptive curve parameters, curve area, statistical moment,
and volume of distribution calculations.
RESULTS
Dog Investigations
All animals survived the surgical preparation without evidence of significant neurological deficit due to the implantation
of the catheters.
Dose-Ranging Phase
There were no significant changes in motor function, cardiovascular status, or body temperature in either dog receiving 5 mg/
ml, 100 ml/h. However, due to prominent tarry stools, a significant lowering of hemtocrit (pre vs. post: 47% to 27% and 41% to
25%) and deteriorating physical condition, ketorolac infusion (5
mg/ml, 100 ml/h) of dog 3376087 was terminated on day 13, and
this animal was euthanized on day 18. Dog 3381293 displayed
blood in stools and some deterioration of physical status, but was
infused until euthanasia on infusion day 34. It was considered that
these outcomes reflected the systemic effects arising from the
daily dose of ketorolac of 2 mg and 1.3 mg/kg/day, respectively,
based on body weight at the time of termination of infusion.
Serum ketorolac concentrations at sacrifice were 7.83 and
3.36 mg/ml, respectively. Inspection of the spinal cord revealed
no pathology (see Table 4). Based upon these results, the ketorolac concentration of 5 mg/ml was retained, but the infusion rate
for the safety phase was reduced to 50 ml/h.
325
Safety Phase
All vehicle dogs, six of eight 0.5 mg/ml test article dogs and
five of the six 5 mg/ml test article dogs completed the 28-day
infusion sequence. The two 0.5 mg/ml dogs (3892328 male,
3884376 male) were euthanized at day 13 and 23, respectively,
and the single 5 mg/ml test article dog (CPH AER, female) was
sacrificed at day 10 of infusion.
Behavioral and Physiological Observations
Neurological examination. Baseline behavioral scores for
arousal, muscle, and motor coordination were scored as ‘0’ or ‘1’
at the start of infusion. In vehicle animals, all neurological signs
were judged minor and limited to mild increases in hind limb
motor tone and minimal gait abnormalities. Behavioral manifestations, i.e., arousal, muscle tone, and motor coordination,
were typically scored as ‘0’ or ‘1’ throughout the interval of
infusion. In the 0.5 mg/ml treatment group, six of the eight
displayed minimal neurologic changes comparable to the vehicle group, with behavioral scores typically ‘0’ or ‘1’. However,
two animals, 389 2328 and 388 4376, were euthanized at day 13
and 23, respectively, due to progressive hypertonicity of the neck
and hind limbs, and pain behavior. These conditions developed
between days 9 and 13 for dog 389 2328 and days 14 to 20 in dog
388 4376. In the 5 mg/ml treatment group, as in the vehicletreated group, all abnormalities identified were minimal and
typically reflected mild hind limb and minimal gait impairment,
with behavioral scores for arousal, muscle tone, and motor
coordination typically scored as ‘0’ or ‘1’ throughout the interval
of infusion. A single 5 mg/kg dog (CPH AER, female) was
sacrificed at day 10 of infusion due to the acute development
of a lateral recumbent posture and whole body stiffness.
Physiological measures. Pre-infusion values for temperature (101.8 6 0.6 C), respiration (30 6 4 b/m), heart rate (125 6
21 b/m), and blood pressure (diastolic: 79 6 16; mean: 99 6 17;
systolic: 136 6 17 mmHg) were comparable for each group.
Over the interval of the infusion, no treatment group displayed
any clinically significant trend in these indices in any animal,
even in those undergoing early sacrifice.
Body weight. Mean body weight upon initiation of infusion
was 9.7 6 0.5 kg and did not differ between treatment groups.
Neither the test nor control article treatments had a significant
effect upon mean body weight. CPH AER displaying significant
falls in hematocrit and euthanized at 10 days showed a 13%
decrease in body weight over the interval of test article administration (see below).
Nociceptive response. Baseline thermal escape latencies
were 3.6 6 0.4 s. There were no significant changes in this
thermal escape latency over time in any treatment group.
Clinical Chemistry
Blood chemistry values were not different among the
three groups at surgery, and the medians for selected
326
YAKSH ET AL.
TABLE 2
Group Summary of Blood/CSF Clinical Chemistry Taken at Surgery and at Sacrifice and Individual Data for Animals Showing
Pathology or Early Sacrifice
Blood
WBC
(#/ml)
Neutrophil
(#/ml)
CSF
RBC
(#/ml)
Hematocrit
(%)
Pre-treatmenta
All animals
10 (9–11)
6 (5–7)
7 (6–8) 47 (45–50)
Post-treatment groupsa
Vehicle (n 5 6)
9 (8–100)
6 (5–7)
8 (7–8) 49 (47–49)
0.5 mg/ml (n 5 5)
8 (5–10)
5 (3–8)
6 (6–7) 43 (39–45)
5.0 mg/ml (n 5 6)
14 (10–16) 10 (6–12)
5 (4–7) 36 (25–43)
Post-treatment individual dogs showing notable pathology or early sacrifice
3892328 (13 d) (0.5 mg/ml)b 12
11
7
50
388 4376 (23d) (0.5 mg/ml)
7
7
7
49
385 8979 (28d) (5 mg/ml)
18
17
6
40
CPHAER (10d) (5 mg/ml)
24
22
4
29
337 6087 (18d) (5 mg/ml)d
15
11
4
25
12
9
4
27
338 1293 (28 d) (5 mg/ml)d
Protein
(mg/dl)
Glucose
(mg/dl)
11 (10–12)
69 (66–72)
41 (25–151)
178 (36–1050)
63 (5–186)
66 (63–71)
65 (56– 67)
59 (40–63)
115 (10–1015)
38 (19–292)
560 (184, 1741)
3800
300
1200
66
44
18
c
1320
c
c
c
592
36
40
63
WBC
(#/ml)
Observations
4 (1–36)
c
800
40
CB. WBH
SF, WGH
CB, SF
CB, GI, SF, WBH
GI
GI
Note. Rank-order comparison across treatment groups (Jonckheere) showed no statistically significant dose-dependent trends for blood or CSF clinical chemistry.
p values ranged from p 4 0.10 to p 4 0.50. Abbreviations: CB: cisternal blood; SF: serous fluid along external catheter track; GI: gastrointestinal bleeding and
ulceration; WBH: whole body hypertonicity.
a
Median with 25%–75% quartiles.
b
Sacrifice day.
c
Insufficient sample.
d
Infusion rate at 100 ml/h as compared to 50 ml/h for all other animals.
values are presented in Table 2. Examination and statistical
comparison of these parameters at the time of sacrifice reveals
no systematic dose-dependent effects. As noted, several animals
underwent early sacrifice or showed distinct reactions, and theirdata are presented separately to permit independent assessment.
Hematocrit/RBC. As indicated in Table 2, there were no
systematic changes between surgery and sacrifice in animals
receiving vehicle. Dogs receiving 0.5 mg/ml ketorolac typically
displayed a slight ( < 15–20%) reduction in hematocrit and red
blood cell count. With the 5 mg/ml infusion (50 ml) treatment,
one of six dogs displayed a prominent drop in hematocrit
and RBC count. Red cell count paralleled the changes noted
in the respective hematocrit. As noted above, both animals
receiving the higher infusion dose showed prominent drops in
hemtocrit values.
WBC/neutrophils. Animals in vehicle and ketorolac treatments typically displayed little or no change in WBC and neutrophil counts (Table 2). Treatment animals undergoing early
sacrifice or showing prominent dysfunction typically showed an
increase in both WBC and neutrophil count.
CSF analysis. Summaries of cisternal CSF clinical analysis results are presented in Table 2. Mean CSF protein and
glucose were not different among treatment groups at the time of
surgery.
Vehicle and both test article groups displayed increases in
protein at sacrifice. Cisternal glucose levels were within
expected ranges before and at the time of necropsy, except
for animal 385 8979 in the 0.5 mg/ml ketorolac group. White
blood cell counts showed an increase over the 28-day period
that was similar in both vehicle and test article groups.
Differential cell counts showed a similar profile for both
vehicle and test article groups. Statistical comparison of
these parameters at the time of sacrifice reveals no dosedependent differences among treatment groups ( p 4 0.10).
In the animals undergoing early sacrifice, cisternal CSF protein concentrations were notably elevated (see Table 2). CSF
glucose levels were comparable at sacrifice in all animals
with the exception of low glucose levels in 3858979, 388437,
and 3376087, the latter two of which underwent early
sacrifice.
Necropsy Observations
Catheter. Examination of the external catheter tracts typically revealed a well-healed sheath encasing the catheter as it
exited from the cisternal membrane. The presence of serous fluid
was noted along the sc catheter track in two 0.5 mg/kg dogs
(3884376; 3858979) and one 5.0 mg/kg dog (CPH AER) (see
Table 2). Intrathecal injection of 0.2 ml of methylene blue dye
uniformly revealed coloration at the tip indicating patency and
continuity.
Spinal cord. No visually evident abnormalities were typically present in the spinal canal at the catheter sites at necropsy.
327
INTRATHECAL KETOROLAC IN DOGS
However, in two 0.5 mg/kg dogs (3892328; 3858979) and one
5.0 mg/kg dog (CPH-AER), unclotted blood was noted in the
cisterna and upper cervical spinal cord.
Other tissues examined. Administration of 5 mg/ml ketorolac (50 ml/h) was associated with gastrointestinal ulceration in
two of six animals (CPH AYU and CPH AZX) and, as noted
above, in both animals receiving 5 mg/ml at 100 ml/h. Small
intestinal mucosal infiltration with lymphoid and monocytic
cells, and edema was present in dog CPH AER. Mild pulmonary
congestion and thrombosis were noted in Dog CPH AER. Tissue
from other organs was not examined by the histopathologist,
however, no other gross abnormalities were noted at necropsy. With the low (0.5 mg/ ml) infusion concentration no evidence of gastric bleeding or other untoward systemic effects
were noted.
Spinal Histopathology
Table 3 presents the summary of the incidence of pathology
observations and scores for the four treatment groups. Table 4
presents the individual pathology score presented by animal and
spinal level. In general, observations may be considered in terms
of local catheter compression of the spinal cord, the presence of
dural and arachnoid infiltrates, and root and spinal infiltrates at
the three levels of the spinal cord that were systematically examined. Several points summarize the overall assessment.
Spinal cord compression. A modest spinal cord compression adjacent to the catheter was noted with an equal incidence in
all four treatment groups, along the axis of the catheter but not
below the catheter tip (Table 3; see Fig. 1).
Local inflammatory reaction. The incidence of inflammatory infiltrates in dura, nerve root, and spinal cord was noted with
approximately equal incidence observed at all levels in all four
treatment groups (Table 3). At the catheter tip site (section C),
where drug concentrations are highest, the local reaction was
judged to be minimal to mild (score of 1.25 or less) in four of six
control animals, five of eight 0.5 mg/ml treated animals, and five
of six animals in the 5 mg/ml treated group. In these animals, a
small fibrous track inside the dural sac represented the catheter
track (see Figure 1A). Minimal to mild chronic inflammation
was associated with the tracks, consisting of small numbers of
lymphocytes, plasma cell, and macrophages around the catheter
site and in the adjacent dura. In the remaining two animals in the
control group (CPHATW, CPH AVH), three animals in the
0.5 mg/ml treated group (3892328, 3884376, 3858979) and
one animal in the 5 mg/ml (CPH AER) treated group, moderate
to severe inflammation (score of 1.5 or greater) was present in
block C (catheter tip). One dog had white matter vacuolization
TABLE 3
Summary of Histological Observations by Group
Histopathology observations (incidence by group)
Spinal
compression by
catheter
Dura infiltrate
(fb/lym/pmn)
Arachnoid infiltrate
(lym/pmn)
Nerve root
infiltrate
(lym/pmn)
Spinal cord
infiltrate
(lym/pmn)
Spinal level
Cervical
Saline 50 ml/h
0.5 mg/ml 50 ml/h
5 mg/ml-50 ml/h
5 mg/ml-100 ml/h
2/6
2/7
3/6
1/2
5/6
4/7
5/6
2/2
3/6
6/7
6/6
1/2
1/6
2/7
3/6
2/2
0/6
2/7
2/6
0/2
0.75
1.0
1.5
(1.5)
Thoracic
Saline 50 ml/h
0.5 mg/ml 50 ml/h
5 mg/ml-50 ml/h
5 mg/ml-100 ml/h
1/6
1/7
2/6
2/2
6/6
7/7
6/6
2/2
3/6
7/7
6/6
2/2
1/6
1/7
1/6
2/6
2/7
2/6
0.87
1.0
1.5
(1.9)
Lumbar (tip)
Saline 50 ml/h
0.5 mg/ml 50 ml/h
5 mg/ml-50 ml/h
5 mg/ml-100 ml/h
1/6
4/7
1/6
0/2
5/6
7/7
6/6
2/2
2/6
7/7
6/6
1/2
1/6
2/7
3/6
1/2
1/6
1/7
2/6
1/2
1.0
1.0
1.0
1.5 (?)
Lumbar (distal)
Saline 50 ml/h
0.5 mg/ml 50 ml/h
5 mg/ml-50 ml/h
5 mg/ml-100 ml/h
0/6
0/7
0/6
0/2
5/6
1/7
5/6
2/2
5/6
7/7
5/6
2/2
4/6
0/7
2/6
1/2
2/6
1/7
2/6
0/2
1.0
0.75
1.0
1.0(?)
Treatment
Pathology
score
median
Note. Chi square analysis across treatments (vehicle, 0.5 mg/ml, and 5.0 mg/ml) for each class of histological observation vs. treatment at each spinal level revealed
no statistical significance at a critical value corresponding to p 5 0.10. Abbreviations: fb: fibroblasts; lym: lymphocytes; pmn: polymorhonuclear leucocytes.
328
YAKSH ET AL.
FIG. 1. Representative histology for dogs receiving 50 ml/h saline (A and D) or 5 mg/ml ketorolac (B and C) for 28 days. Abbreviations: C: catheter, SAS:
subarachnoid space. Arrows indicate the dura. Top left: Dog CPH ATW Saline, spinal cord section D (lumbar). Top right: CPHAVL Ketorolac, spinal cord
section D (lumbar). Bottom left: Dog CPH AYU Ketorolac, spinal cord section D (lumbar). Bottom right: Dog CPH AXA Saline, spinal cord section D (lumbar).
TABLE 4
Summary of Spinal Histopathology
Spinal pathology score
Group
Dog
Sex
Control
Control
Control
Control
Control
Control
0.5 mg/ml
0.5 mg/ml
0.5 mg/ml
0.5 mg/ml
0.5 mg/ml
0.5 mg/ml
0.5 mg/ml
0.5 mg/ml
5 mg/ml
5 mg/ml
5 mg/ml
5 mg/ml
5 mg/ml
5 mg/ml
CPH APB
CPH ASL
CPH ATW
CPH AVH
CPH AWK
CPH AXA
OFR-1
OER-1
YZR-1
389 2343
389 4631
385 8979
389 2328a
388 4376a
CPH AERa
CPH ATY
CPH AUI
CPH AVL
CPH AYU
CPH AZX
F
F
F
M
M
M
M
M
M
M
M
M
M
M
F
F
M
M
M
F
Cervical
Thoracic
Lumbar (Tip)
Lumbar (Distal)
0.75
0.75
1.25
1.75
0.75
0.75
0.75
1
1
1
1.25
3
2
2
3
2
1
2
0.75
1
0.75
0.75
1
2
1
0.75
1
1
1
1.25
1
3
1.25
2
3
1.25
3
1.75
1
1
0.75
0.75
2.25
2.25
1
1
0.75
1
1.25
1
1
3
2
2
3
1
1
1
1
1
1
0.75
3
2.25
0.75
1
1
1
0.75
0.75
0.75
3
1
0.75
3
1
2.25
1
1
0
Note. Inflammationseveritykey:0 5 noabnormalfindings;1 5 mild;2 5 moderate;3 5 severe;incrementsof60.25inscoreindicatedslightlylessormoresevere
than whole number score. Control 5 saline at a volume of 1.2 ml/day, 0.5 and 5 mg/ml ketorolac intrathecally at a volume of 1.2 ml/day. Rank order comparison across
treatmentgroupsateachspinallevelwascarriedoutusingtheJonckheererankedtestfororderedmeansateachspinallevel.Comparisonofgrouppathologyscoresacross
treatment at lumbar distal, lumbar tip and cervical were not statistically significant (p 50.10), while comparison across dose at the thoracic level was (p 5 0.05).
a
Animals euthanized early (day 13, 23, and 10 of infusion respectively).
INTRATHECAL KETOROLAC IN DOGS
329
(without inflammation); this was focal and not associated with
the catheter site. Consistent with this summary, the median
pathology scores for sections C and D were not different across
treatment group (Jonckheere test p 4 0.10). Consideration of the
individual animals presented in Table 3 emphasizes that the most
severe histopathology reactions were observed in animals with
low CSF glucose at sacrifice (3884376; 3858979, 3376087).
At sites cephalad to the catheter tip (section A: cervical and B:
thoracic) a continuation of the reaction observed distally
was typically noted (see Figs. 1B and 1D). The sections with
inflammation showed variable proportions of lymphocytes,
neutrophils, plasma cells, and macrophages. In some animals,
the inflammation involved spinal nerve roots and epi/perineurium. A variable degree of fibrosis was noted at the catheter
site and in adjacent dura, sometimes with neovascularization.
Two animals (389 2328 and CPH ATW) had focal lymphocytic
infiltration of the white matter adjacent to the catheter site and
one (CPH AVL) had a focal lymphocytic infiltrate along the
Virchow-Robin spaces. Two animals (385 8979, 0.5 mg/ml
ketorolac and CPH AER, 5 mg/ml ketorolac) had extensive
infiltration of the spinal cord parenchyma by predominantly
neutrophils, with focal necrosis and abscess formation. Statistical comparison across groups revealed a statistical significance
by dose for section B (thoracic: p50.05) but not section
A (cervical: p40.10).
In all treated and control groups, spinal GFAP staining was
within normal limits, even at sites adjacent to the catheter site
indicating alack of reactive gliosis.
Gram and GMS stains undertaken in block C (catheter tip)
revealed no signs of bacteria or of fungal elements.
Toxicokinetics of Safety Animals
Infusate. Analysis of injectate (ketorolac 5 mg/ml) indicated that representative aliquots contained 5375 6 58 mg/ml
of ketorolac (N 5 5).
Serum. No animal showed ketorolac levels in their serum
prior to initiation of infusion. The vehicle treatment group did
not show detectable drug levels in serum after 28 days of infusion. In animals receiving ketorolac infusion, the mean serum
ketorolac concentrations on day 4, the first time point after infusion initiation, were 0.18 6 0.05 mg/ml and 1.14 6 0.25 mg/ml
for 0.5 mg/ml and 5 mg/ml ketorolac infusions, respectively.
Serum concentrations measured at the time of sacrifice did not
differ for these respective treatments. In the one animal (CPH
AER, 6.0 kg) in the 5 mg/ml test article treatment group that was
euthanized on day 10 of infusion because of gastric ulceration,
maximum serum ketorolac levels were 2.08 mg/ml. As expected,
small animals (such as CPH AER) displayed the highest serum
ketorolac concentrations.
Cisternal CSF. No animals showed any drug levels in
pretreatment cisternal CSF. The vehicle group did not show
drug in cisternal CSF after 28 days of infusion. After 28 days
FIG. 2. Ketorolac (mg/ml) in lumbar CSF and serum of three dogs after
bolus lumbar intrathecal injection (T 5 0) of ketorolac (5 mg/1 ml). (Kinetic
data for these experiments are presented in Table 5.)
of continuous infusion, the test article treatment groups had mean
cisternal CSF drug levels of 0.06 6 0.03 mg/ml and 0.16 6 0.09
mg/ml (0.5 mg/ml and 5 mg/ml groups, respectively) in the
cisternal CSF. The mean cisternal CSF to serum ratio measured
at the time of necropsy was 0.32 6 0.15 and 0.10 6 0.07 for the
0.5 mg/ml and 5 mg/ml treatments, respectively.
Intrathecal Pharmacokinetics
All pK animals survived the experiments without disability.
One animal CRD ATI failed to sample after the initial studies and
was removed from further analysis. At necropsy, the intrathecal
infusion catheters were typically located at L1/2 and the sampling
catheter was approximately 2 cm proximal to that of the infusion
catheter. Examination of the spinal cords revealed no evidence of
gross pathology or any sign of blood in the intrathecal space.
Bolus intrathecal injections and ketorolac concentrations.
The bolus intrathecal delivery of ketorolac (5 mg/ml/1 ml) in
three dogs resulted in an initial peak concentrations in lumbar
330
YAKSH ET AL.
TABLE 5
Summary of Calculated Lumbar CSF PK Parameters for Bolus
Dose (5 mg/ml/1 ml) of Lumbar Intrathecal Ketorolac in Dogs
Parameters
Dog
CRDAIC
Dog
3359565
Dog
3377130
Grouped
dataa
D half life (min)
E half life (min)
C initial (mg/ml)
MRT-area (min)
AUC obs (mg 3 min/ml)
Vd obs (ml)
17.9
81.2
1381
46.2
39620
14.8
7.6
71
6304
23.4
99278
5.2
5.5
38
4443
11.7
43402
6.3
10.3
52.7
2663
34.5
54119
7
Note. D, distribution; E, elimination; C, concentration; MRT, mean residence
time; AUC area under the curve; obs, observed; Vd, volume of distribution.
a
Regression analysis run on grouped data.
FIG. 4. PGE2 (% of control concentration) and Ketorolac (mg/ml)
(Mean 6 SD; N 5 3) after bolus intrathecal injection (T 5 0) of ketorolac
(5 mg/1 ml) in lumbar CSF sampled at T 5 15, 30, and 240 min **p 50.05,
paired Student’s t-test).
significant reduction (48%) in the concentrations of lumbar
PGE2 at 30 min with significant recovery being observed
by 4 h (see Fig. 4). After the initiation of intrathecal ketorolac
infusion (0.5 mg/ml) in three dogs, lumbar PGE2 concentrations
were significantly depressed from baseline 1271 6 18 to
87 6 22 at 24 h.
FIG. 3. Left: Lumbar CSF Ketorolac concentrations (mean and individual
data) at T 5 15 min (Pre) and 24–48 h after the initiation of intrathecal
infusion of ketorolac (0.5 and 5 mg/ml/100 ml/h) in three dogs. Right: Ratio
of lumbar CSF ketorolac measured versus that measured during infusion with
0.5 mg/ml at 100 ml/h.
CSF and plasma, and this declined in a biphasic fashion (Fig. 2).
Calculations of lumbar CSF kinetic parameters, assuming a onecompartment model, are presented in Table 5. At the time of
peak serum concentration, the CSF/serum ratios were approximately 30:1.
Continuous infusion and ketorolac concentrations. To
determine steady state levels of ketorolac in CSF, lumbar
CSF samples were harvested before and at 24–48 h after
initiation of ketorolac infusion in three animals. As indicated
in Figure 3, ketorolac concentrations at 0.5 and 5.0 mg/ml at 100
ml/h were on the order of 3.8 and 53 mg/ml, respectively.
Effects on Lumbar CSF PGE2 Levels
Mean lumbar CSF PGE2 concentrations measured in six dogs
prior to the delivery of ketorolac were 1224 6 26 pg/ml. After the
bolus delivery of ketorolac (5 mg/ml; 1 ml), there was a
Rat Investigations
Safety Phase
All animals completed the series of injections, and there were
no alterations in behavior. Ketorolac-injected animals did not
differ from saline-injected animals in assessment of motor tone,
coordination, and arousal (data not shown). Similarly, weight
did not differ before intrathecal drug treatment (288 6 19 g in
saline; 304 6 21 g in ketorolac) at the end of drug treatment
(289 6 22 g in saline; 305 6 18 g in ketorolac), or on any day
during treatment (data not shown). Latency to paw withdrawal
did not differ prior to or following intrathecal drug delivery and
baseline withdrawal latencies did not change over the time
course of drug treatment (data not shown).
Necropsy. At post mortem, spinal cords, meninges, and
nerve roots appeared grossly normal in all animals. Perfusion
was inadequate in one animal (in the saline group).
Histopathology. Histological examination showed a rim of
fibrin, fibroblasts, and inflammatory cells around the catheter
site. Histological changes were observed frequently in both
groups, but with no difference in incidence between groups.
At the lumbar level, the number of animals with no significant
INTRATHECAL KETOROLAC IN DOGS
331
preclinical evaluation of ketorolac safety by this route must be
ascertained. Ketorolac (Acular PF) is a sterile preservative-free
formulation provided in concentrations of 5 mg/ml. We examined continuous 28-day intrathecal infusion of the highest available concentration of drug at a dose that could be tolerated by the
dog and 1/10 this highest tolerated concentration. In the rat,
the standard bolus injection volume is 10 ml and, accordingly,
the highest concentration was delivered in that volume, and the
effects of the injections on spinal blood flow were examined.
Intrathecal Safety Assessment
FIG. 5. Effect of intrathecal space perfusion of saline (solid squares),
ketorolac, 2.5 mg/ml (open circles), or ketorolac, 5 mg/ml (solid triangles) on
change in spinal cord blood flow velocity (SCBFV) as measured by laserDoppler flowmetry, in anesthetized rats (N 5 5 per group). No significant
change over time was observed with any treatment.
lesions, and minimal multifocal white matter vacuolation was
three, five, and one in saline-treated animals and one, seven, and
two in ketorolac-treated animals. The white matter vacuolation
was swelling of myelin sheaths, some of which contained myelin
fragments and cellular debris. These changes, when present,
were in all funiculi and at all dermatomal levels, not consistent
with a focal lesion of the cord.
Spinal cord blood flow. Laser-Doppler flow remained stable
following establishment of the experimental model, and was not
altered over 30 min of saline perfusion. Ketorolac, either at the
2.5 or 5.0 mg/ml concentration, had no effect on SCBF, as
measured by relative change in blood flow velocity (Fig. 5).
DISCUSSION
Spinal Cyclo-oxgenase
Cyclo-oxygenase (COX) inhibitors suppress hyperalgesia
that arises secondary to tissue injury. In contrast to the broadly
held belief that NSAID have a peripheral action, these agents
exert a potent antihyperalgesic action following spinal delivery
(Malmberg and Yaksh, 1992a,b; Yaksh, 1982; Yaksh et al.,
2001). This spinal effect reflects three observations: 1) cyclooxygenases are constitutively expressedin spinal parenchyma; 2)
afferent input evokes spinal prostaglandins release; 3) prostanoids exert a facilitatory effect on spinal nociceptive neurons
(Svensson and Yaksh, 2002; Yaksh et al., 2001). While it is clear
that systemically delivered NSAIDs can exert central effects, the
relative central bioavailability of agents such as ketorolac is poor
(Rice et al., 1993). Given the peripheral side effect (e.g., antiplatelet action), the central effects of a systemic agent may be
limited by the peripheral dose effects. This raises the possibility
of considering spinal delivery. Cancer patients received pain
relief from intrathecal lysine acetylsaliscylate, suggesting a
spinal action in humans (Devoghel, 1983, 1993; Pellerin et al.,
1987). Nevertheless, prior to human intrathecal delivery,
Spinal drug safety has two components: untoward physiological effects mediated by spinal and nonspinal systems, and local
tissue toxicity. In the first case, the effects are proportional to the
total delivered dose. Such outcomes reflect a direct effect on
spinal or extra-spinal mechanisms and often define the maximum dose of drug to which the dog can be exposed. The limiting
response for opiates is defined by respiratory depression
(Sabbe et al., 1994); for clonidine it is hypotension and sedation.
(Yaksh et al., 1994) Conversely, tissue toxicity is dependent
upon the concentration to which the local tissue proximal to
the delivery site is exposed. Accordingly, the most robust test
of safety will be consideration of the tissue reaction to the highest
concentration that can be delivered. In this case, the available
GMP, preservative-free formulation of ketorolac is 5 mg/ml.
Dose/Concentration Selection
In the dose-ranging series, we observed that 5 mg/ml, 100 ml/h
(total dose of 12 mg/day, approximately 1–1.5 mg/kg) resulted in
a reduced hematocrit and gastrointestinal bleeding. Accordingly, the infusion rate in was lowered to 50 ml/h to reduce
the total systemic dose by half. This permitted us to retain the
highest local spinal exposure and reduce, but not abolish,
systemic toxicity.
Intrathecal Ketorolac Effects
Under the present treatment paradigm, all animals including
the saline-control dogs displayed a mild local meningeal inflammatory reaction and elevated cisternal CSF protein at the time
of sacrifice. Inflammation at the catheter tip site was noted with
approximately equal incidence in control, 0.5 mg/ml, and
5 mg/ml treated group. At the site of highest drug concentration
(the catheter tip), there were only modest effects in all groups and
none were statistically drug/dose related. Interestingly, group
scores revealed an elevation of pathology score only for thoracic
sections. No difference was noted in cervical sections. This lack
of difference in pathology scores for lumbar sections proximal to
the catheter tip where drug level is highest makes it difficult to
interpret the significance of the thoracic response. Finally,
although the higher dose of ketorolac (5.0 mg/ml/100 ml/h)
was not systematically examined, the incidence and magnitude
332
YAKSH ET AL.
of the spinal pathology was not different from that observed at
the lower doses. These incidental observations over a 10-fold
range of concentration and a 20-fold range of dosing confirm
the lack of an effect of ketorolac at these concentrations on
spinal tissue. These results are in accord with those of Gallivan
et al. (2000) who studied repeated epidural bolus delivery of
ketorolac in dogs and observed no untoward effects on
spinal structure.
While there were no group differences, three animals had
severe acute inflammation in the spinal cord parenchyma, two
(3858979 at all spinal levels and 3884376 at thoracic level) in the
0.5 mg/ml group and one (CPH AER at all levels) in the 5 mg/ml
group. These observations reflect several variables. First, these
animals displayed systemic neutropenia and reduced cisternal
CSF glucose (when determined), a profile suggestive of a bacterial infection. The severity of the effect upon CSF could not be
determined at necropsy from the third animal with severe spinal
cord inflammation. This animal was the smallest animal and had
the most severe gastrointestinal and adverse systemic responses.
Gram stains undertaken at the catheter tip were negative in all
animals. This, however, does not exclude that a peri-catheter
infection at other loci contributed to this picture. This suggestion
is supported by the observation of serous fluid along the external
subcutaneous catheter tract at necropsy in these animals.
Neuraxial bleeding. Blood was noted in the cisterna and
upper cervical cord at sacrifice in two 0.5 mg/kg and one 5.0
mg/kg dog. No bleeding was observed in other animals or vehicle-treated dogs. As there was a percutaneous cisternal tap
approximately 30 min prior to death, it is possible that the observation of local cisternal blood reflects an event secondary to a
traumatic puncture. The presence of persistent blood ketorolac
might facilitate such bleeding. The fact that bleeding was not
observed in any animal intraparenchymally, or at the lumbar
level where drug concentration was highest supports a local
event at the cisterna. The importance of persistent anticoagulation in this outcome is suggested by the absence of cisternal
bleeding in vehicle animals and in pharmacokinetic dogs receiving intrathecal ketorolac by bolus and by infusion for two-day
intervals.
Rat Studies
The absence of specific intrathecal drug related toxicity in the
dog was mirrored by the studies carried out in the rat. Here
moderate catheter-related pathological changes were noted.
These effects were not altered by ketorolac. Further, consistent
with previous work suggesting that prostaglandins have minimal
effects on central perfusion (Siesjo and Nilsson, 1982), local
ketorolac had no effect upon spinal blood flow.
1999), which are driven by local drug concentration and duration
of exposure. (Yaksh et al., 1999b) The most robust assessment of
toxicity is provided by continuous delivery of the drug into the
local intrathecal space at the highest available concentration.
(Yaksh et al., 1999b) Accordingly, we employed the highest
concentration and dose consistent with survival and applied
this by continuous infusions for an interval of 28 days. Given
an absence of tissue effect, an important question is how sensitive is the test model in displaying toxicity. To date, a variety of
agents have examined in this dog model, including baclofen
(Sabbe et al., 1993); clonidine (Yaksh et al., 1994); adenosine
(Chiari et al., 1999); neostigmine (Yaksh et al., 1995); and brainderived nerve growth factor (Yaksh et al., 1997). Morphine
given for 28 days as a bolus was also without effect (Sabbe
et al., 1994). In contrast, the opioid peptide DPDPE (Horais
et al., 2003) and high concentrations of morphine (Yaksh et al.,
2003) are associated with toxicity after 28-day continuous infusion. The canine model is thus both robust and sensitive.
Intrathecal Ketorolac Kinetics
Intrathecal ketorolac showed a biphasic clearance after bolus
delivery. This clearance is consistent with a small moderately
lipid soluble molecule that is cleared from the CSF compartment
first by an initial distribution followed by clearance. Inulin, a
small molecule shows a similar biphasic clearance with similar
T1/2 values (Yaksh et al., 1997). Importantly, continuous infusion of 0.5 and 5 mg/ml gave rise to lumbar CSF concentration of
approximately 4 and 53 mg/ml respectively, suggesting that the
CSF kinetics of intrathecally infused ketorolac are linear.
Intrathecal Ketorolac Efficacy
In the dog, the available pain models involve thermal escape
(e.g., skin twitch). COX inhibitors have little effect upon such
acute nociceptive processing (Svensson and Yaksh, 2002).
Accordingly, based on rodent data (Svensson and Yaksh,
2002), we anticipated that IT ketorolac would have no effect
upon this end point. However, measurement of prostaglandins
levels in the lumbar cerebrospinal fluid revealed depression of
the basal levels after bolus and infusion. Suppression was greatest with continuous infusion. Delivery of 1.2 mg over 24 h (e.g.,
0.5 mg/ml) resulted in a greater reduction in CSF PGE2
levels than when 5 mg was delivered as a bolus. As suppression
of prostanoid synthesis reflects the essential mechanism of
ketorolac, we conclude that achieving CSF concentrations of
4 mg/ml of ketorolac in the CSF leads to a prominent biological
action.
Intrathecal Ketorolac Therapeutic Ratio
Model Reliability
Spinal agents may lead to local reactions including aseptic
inflammation, neuronal death, and demyelination (Weller,
In the rat, it has been shown that intrathecal; ketorolac (10 mg/
10 ml) bolus can produce a potent analgesia (Malmberg and
Yaksh, 1992a.b., 1993). In the present study, we undertook
333
INTRATHECAL KETOROLAC IN DOGS
repeated bolus delivery of 50 mg/10 ml) of ketorolac. This is the
maximum dose that can be delivered given the standard injection
volume of 10 ml and available ketorolac concentration. At this
maximum concentration/dose, no untoward spinal toxicology
was noted. Accordingly, we suggest that the therapeutic index
(toxic dose/effective dose) for analgesia in the rodent exceeds
five. In so far as concerns the dog, the suppression of PGE2
concentration is a surrogate marker. We assert that a maximal
suppression of CSF-PGE2 concentration as compared to control
was observed with continuous infusion of 0.5 mg/ml yielding a
CSF concentration of 4 mg/ml. As spinal tissue toxicity was not
observed at a mean steady CSF concentration of ketorolac of
approximately 50 mg/ml (produced by 5 mg/ml), we suggest that
the therapeutic ratio as measured by ketorolac concentration in
the CSF or infused dose readily exceeds a factor of 10.
Relationship of Preclinical Safety Data to Humans
The ‘robustness’ of the assertion of intrathecal safety is
defined by the degree to which the preclinical exposure exceeds
that to which the human will be exposed. Accordingly, to the
degree that local concentration defines toxicity, we conclude that
28 days of exposure to a local CSF concentration of approximately 50 mg/ml (as produced by the continuous infusion of
5 mg/ml) has minimal drug related toxicity. As for bolus
delivery, we note that after the bolus delivery of 5 mg in 1 ml,
the initial CSF concentration was approximately 3000 mg/ml or a
factor of 60 (i.e., 3000/60) greater. In previous work, we noted
that the local acute dilution volume of the lumbar space of the
dog is conservatively 1/3 of the human, i.e., for any given dose
the human will dilute the bolus injection by a factor three times
greater than the dog (Sabbe et al., 1993). Therefore, assuming
linear kinetics we would argue that the initial bolus dose of
ketorolac in humans would be 5 mg/ml 3 1/60 3 3 5 250
mg/ml. These are admittedly crude calculations, but provide a
conservative first approximation of the dose and concentration
for bolus delivery that should give a local concentration in the
human that was tolerated for 28 days in the preclinical model.
Based on these studies, a Phase 1 bolus dose ranging trial
was undertaken with intrathecal ketorolac in humans with no
significant events (Eisenach et al., 2002).
With regard to continuous infusion, it is clear that in the dog,
the limiting side effects at the 5 mg/ml concentration was total
dose, At doses greater than 6 mg/day (corresponding to approximately 0.6 mg/kg over 24 h in a 10 kg dog or chronic serum
concentrations of 1.1 mg/ml) systemic gastrointestinal effects
and reductions in hematocrit were evident. In humans, a single
bolus analgesic dose of ketorolac (approximately 0.4 mg/kg
yields peak plasma levels of 2–3 mg/ml; Brocks and Jamali,
1992) can indeed yield significant and long lasting (24 h)
changes in coagulation parameter (Niemi et al., 2000). Doubtless chronic exposure with these serum concentrations would be
poorly tolerated and such considerations must temper any
chronic ketorolac application in humans.
ACKNOWLEDGEMENT
This work was sponsored by NIH GM48085.
REFERENCES
Brocks, D. R., and Jamali, F. (1992). Clinical pharmacokinetics of ketorolac
tromethamine. Clin. Pharmacokinet. 23(6), 415–427.
Chiari, A., Yaksh, T. L., Myers, R. R., Provencher, J., Moore, L., Lee, C. S., and
Eisenach, J. C. (1999). Preclinical toxicity screening of intrathecal adenosine
in rats and dogs. Anesthesiology 91, 824–832.
Devoghel, J. C. (1983). Small intrathecal doses of lysine-acetylsalicylate relieve
intractable pain in man. J. Int. Med. Res. 11, 90–91.
Devoghel, J. C. (1993). Intrathecal injection of lysine-acetylsalicylate in man
with intractable cancer pain. In Progress in Pharmacology and Clinical
Pharmacology (I. Jurna and T. L. Yaksh, Eds.), Vol. 10, pp. 111–118. Gustav
Fischers Verlag.
Eisenberg, E., Berkey, C. S., Carr, D. B., Mosteller, F., and Chalmers, T. C.
(1994). Efficacy and safety of nonsteroidal antiinflammatory drugs for cancer
pain: A meta-analysis. J. Clin. Oncol. 12, 2756–2765.
Eisenach, J. C., Curry, R., Hood, D. D., and Yaksh, T. L. (2002). Phase I safety
assessment of intrathecal ketorolac. Pain 99(3), 599–604.
Gallivan, S. T., Johnston, S. A., Broadstone, R. V., Jortner, B. S., and Reimer, M.
(2000). The clinical, cerebrospinal fluid, and histopathologic effects of
epidural ketorolac in dogs. Vet. Surg. 29, 436–441.
Gillies, G. W., Kenny, G. N., Bullingham, R. E., and McArdle, C. S. (1987). The
morphine sparing effect of ketorolac tromethamine. A study of a new,
parenteral non-steroidal anti-inflammatory agent after abdominal surgery.
Anaesthesia 42, 727–731.
Horais, K., Hruby, V., Rossi, S., Cizkova, D., Meschter, C., Dorr, R., and
Yaksh, T. L. (2003). Effects of chronic intrathecal infusion of a partial
differential opioid agonist in dogs. Toxicol. Sci. 71, 263–275.
Malmberg, A. B., and Yaksh, T. L. (1993). Pharmacology of the spinal action of
ketorolac, morphine, ST-91, U50488H, and L-PIA on the formalin test and
an isobolographic analysis of the NSAID interaction. Anesthesiology 79,
270–281.
Malmberg, A. B., and Yaksh, T. L. (1992a). Antinociceptive actions of
spinal nonsteroidal anti-inflammatory agents on the formalin test in the rat.
J. Pharmacol. Exp. Ther. 263, 136–146.
Malmberg, A. B., and Yaksh, T. L. (1992b). Hyperalgesia mediated by spinal
glutamate or substance P receptor blocked by spinal cyclooxygenase inhibition. Science 257, 1276–1279.
Niemi, T. T., Backman, J. T., Syrjala, M. T., Viinikka, L. U., and Rosenberg, P. H.
(2000). Platelet dysfunction after intravenous ketorolac or propacetamol.
Acta Anaesthesiol. Scand. 44(1), 69–74.
Pellerin, M., Hardy, F., Abergel, A., Boule, D., Palacci, J. H.,
Babinet, P., Wingtin, L. N., Glowinski, J., Amiot, J. F., Mechali, D., et al.
(1987). [Chronic refractory pain in cancer patients. Value of the spinal
injection of lysine acetylsalicylate. 60 cases]. Presse Med. 16,
1465–1468.
Rice, A. S., Lloyd, J., Bullingham, R. E., and O’Sullivan, G. (1993). Ketorolac
penetration into the cerebrospinal fluid of humans. J. Clin. Anesth. 5(6),
459–462.
Sabbe, M. B., Grafe, M. R., Mjanger, E., Tiseo, P. J., Hill, H. F., and
Yaksh, T. L. (1994). Spinal delivery of sufentanil, alfentanil, and morphine
in dogs. Physiologic and toxicologic investigations. Anesthesiology 81,
899–920.
Sabbe, M. B., Grafe, M. R., Pfeifer, B. L., Mirzai, T. H., and Yaksh, T. L. (1993).
Toxicology of baclofen continuously infused into the spinal intrathecal space
of the dog. Neurotoxicology 14, 397–410.
334
YAKSH ET AL.
Siesjo, B. K., and Nilsson, B. (1982). Prostaglandins and the cerebral circulation.
Adv. Prostaglandin. Thromboxane. Leukot. Res. 10, 367–380.
Svensson, C. I., and Yaksh, T. L. (2002). The spinal phospholipase-cyclooxygenase-prostanoid cascade in nociceptive processing. In Annual Reviews in
Pharmacology and Toxicology (A. K. Cho, T. Blaschke, P. A. Insel, and
H. H. Loh, Eds.), Vol. 42, pp. 553–583. Annual Reviews, Palo Alto, CA.
Yaksh, T. L., Hua, X. Y., Kalcheva, I., Nozaki-Taguchi, N., and Marsala, M.
(1999a). The spinal biology in humans and animals of pain states generated by
persistent small afferent input. Proc. Natl. Acad. Sci. USA. 96, 7680–7686.
Yaksh, T. L., and Malkmus, S. A. (1999). Animal models of intrathecal
and epidural drug delivery. In Spinal Drug Delivery (T. L. Yaksh, Ed.),
pp. 317–344. Elsevier Science B.V., Amsterdam.
Weller, R. O. (1999). Reaction of intrathecal and epidural spaces to infection
and inflammation. In Spinal Drug Delivery (T. L. Yaksh, Ed.), pp. 297–315.
Elsevier Science B.V., Amsterdam.
Yaksh, T. L., Rathbun, M., Jage, J., Mirzai, T., Grafe, M., and Hiles, R. A. (1994).
Pharmacology and toxicology of chronically infused epidural clonidine HCl in
dogs. Fundam. Appl. Toxicol. 23, 319–335.
Yaksh, T. L. (1982). Central and peripheral mechanisms for the antianalgesic
action of acetylsalicylic acid. In Acetylsalicylic Acid: New Uses for an Old
Drug (J. M. Barnet, J. Hirsh, and J. F. Mustard, Eds.), pp. 137–152. Raven
Press, New York.
Yaksh, T. L., Rathbun, M. L., Dragani, J. C., Malkmus, S., Bourdeau, A. R.,
Richter, P., Powell, H., Myers, R. R., and Lebel, C. P. (1997). Kinetic
and safety studies on intrathecally infused recombinant-methionyl
human brain-derived neurotrophic factor in dogs. Fundam. Appl. Toxicol.
38, 89–100.
Yaksh, T. L., Dirig, D. M., Conway, C. M., Svensson, C., Luo, D., and
Isakson, P. C. (2001). The acute antihyperalgesic action of NSAIDs and
release of spinal PGE2 is mediated by the inhibition of constitutive spinal
COX-2 but not COX-1. J. Neurosci. 21, 5847–5853.
Yaksh, T. L., Grafe, M. R., Malkmus, S., Rathbun, M. L., and Eisenach, J. C.
(1995). Studies on the safety of chronically administered intrathecal neostigmine methylsulfate in rats and dogs. Anesthesiology 82, 412–427.
Yaksh, T. L., Horais, K. A., Tozier, N. A., Allen, J. W., Rathbun, M., Rossi, S. S.,
Sommer, C., Meschter, C., Richter, P. J., and Hildebrand, K. R. (2003). Chronically infused intrathecal morphine in dogs. Anesthesiology 99, 174–187.
Yaksh, T. L., Rathbun, M. L., and Provencher, J. C. (1999b). Preclinical safety
evaluation for spinal drugs. In Spinal Drug Delivery (T. L. Yaksh, Ed.),
pp. 417–437. Elsevier Science B.V., Amsterdam.
Yaksh, T. L., and Rudy, T. A. (1976). Chronic catheterization of the spinal
ub-arachnoid space. Physiol. Behav. 17, 1031–1036.
Yaksh, T. L., and Svensson, C. (2001). Role of spinal cyclo-oxygenases in
nociceptive processing. In Therapeutic Roles of Selective COX-2 Inhibitors
(J. R. Vane and R. M. Botting, Eds.), pp. 168–190. William Harvey Press,
London.