This study evaluated the effect of preoperative intrathecal administration of a low dose of morphine on intraoperative fentanyl requirements in dogs undergoing spinal surgery. Eighteen dogs undergoing cervical or thoracolumbar laminectomy were randomly assigned to receive intrathecal morphine (MG group) or no treatment (CG group). The MG group had significantly lower hourly fentanyl consumption and lower predicted plasma fentanyl concentrations compared to the CG group. This suggests that a low dose of preoperative intrathecal morphine has a sparing effect on intraoperative fentanyl requirements in dogs undergoing spinal surgery. No adverse effects were observed from the intrathecal morphine administration.

1. Sparing Effect of a Low Dose of Intrathecal Morphine on Fentanyl
Requirements During Spinal Surgery: A Preliminary Clinical
Investigation in Dogs
LORENZO NOVELLO, Med Vet, Diplomate ESRA, FEDERICO CORLETTO, CertVA, Diplomate ECVAA,
ROBERTO RABOZZI, Med Vet, and SIMON R. PLATT, BVM&S, Diplomate ECVN & ACVIM (Neurology)
Objective—To evaluate the effect of preoperative intrathecal administration of a low dose of mor-
phine on intraoperative fentanyl requirements in dogs undergoing cervical and thoracolumbar
spinal surgery.
Study Design—Prospective randomized clinical study.
Animals—Dogs (n¼ 18) matched by surgical procedure administered intrathecal morphine (MG)
or no-treatment (control group, CG).
Methods—After premedication with romifidine (4 mg/kg, intravenously) and induction with prop-
ofol, anesthesia was maintained with sevoflurane in oxygen. Intrathecal morphine 0.03 (0.023–
0.034)mg/kg was administered at lumbar level 41 (25–65) minutes before surgery in MG. Ketamine
(0.5 mg/kg) was administered hourly, starting before incision. Fentanyl infusion (1.2 and 4.2 mg/kg/h
in MG and CG, respectively) was administered after a loading dose (5 and 10 mg/kg in MG and CG,
respectively), and boluses were given if an increase 420% in heart rate and arterial blood pressure
was observed. Total amount of fentanyl administered was recorded, to calculate hourly require-
ments and predict plasma concentration using a computer simulation.
Results—Hourly fentanyl consumption and predicted plasma concentrations at the time of response
to surgery were significantly lower in MG compared with CG.
Conclusions—Preoperative administration of a low dose of intrathecal morphine has a sparing
effect on intraoperative fentanyl requirements.
Clinical Relevance—Preoperative intrathecal administration of a low dose of morphine at the
lumbar level represented a safe and effective mean of providing intraoperative analgesia in dogs
undergoing cervical and thoracolumbar spinal surgery.
r Copyright 2008 by The American College of Veterinary Surgeons
INTRODUCTION
PATIENTS UNDERGOING vertebral laminectomy
may experience severe postoperative pain potentially
leading to prolonged recovery, with increased postoper-
ative morbidity and complications.1–3
Intramuscular
(IM) or intravenous (IV) administration of opioids1–4
and/or non-steroidal anti-inflammatory drugs2,3
are com-
monly used to provide analgesia in the perioperative pe-
riod. General, spinal, epidural, and local anesthesia have
been proposed as anesthetic techniques for spinal surgery,
each technique having advantages and disadvantages.5–7
Neuraxial anesthesia provides satisfactory analgesia,8
and neuraxial opioids can provide profound postopera-
Clinical cases were from the Animal Health Trust, Newmarket, UK and Dick White Referrals, Six Mile Bottom, Newmarket, UK.
Address reprint requests to Lorenzo Novello, Med Vet, Diplomate ESRA, c/o ISVRA—Italian Society of Veterinary Regional
Anaesthesia and Pain Medicine, via Meucci 13, 30016 Jesolo, Venezia, Italy. E-mail: novello@isvra.org.
Submitted March 2007; Accepted August 2007
From the Department of Anaesthetics, the Queen’s Veterinary School Hospital, Cambridge, UK; Dick White Referrals, Six Mile
Bottom, Newmarket, UK; the Department of Anaesthetics, Addenbrookes’ Hospital, Cambridge, UK; Ambulatorio Veterinario
Adriatico, Vasto, Chieti, Italy; and the Department of Small Animal Medicine & Surgery, College of Veterinary Medicine, University of
Georgia, Athens, GA.
r Copyright 2008 by The American College of Veterinary Surgeons
0161-3499/08
doi:10.1111/j.1532-950X.2007.00358.x
153
Veterinary Surgery
37:153–160, 2008

2. tive analgesia superior to patient-controlled IV analgesia
techniques9–11
with fewer side effects compared with IM
and subcutaneous administration.12–14
In dogs, preemp-
tive epidural administration of morphine reduces halo-
thane minimum alveolar concentrations (MAC)15
and
induces long-lasting analgesia superior to that provided
by repeated IM administration of oxymorphone and
ketoprofen.16
Although deliberate intrathecal adminis-
tration of a low dose of morphine has already been
reported in dogs17,18
it is not commonly used to provide
analgesia for spinal surgery in small animals, and its
contribution to intraoperative antinociception has not
been investigated.
Our purpose was to evaluate the effect of preoperative
intrathecal administration of a low dose of morphine on
intraoperative fentanyl requirements in dogs undergoing
cervical and thoracolumbar spinal surgery. We hypoth-
esized that intrathecal morphine administration may sig-
nificantly improve antinociception and analgesia for
spinal surgery. We tested this by assessing intrathecal
morphine sparing effect on fentanyl requirements during
surgery.
MATERIALS AND METHODS
After informed owner consent, 18 American Society of
Anesthesiologists physical status I and II dogs undergoing
elective cervical or thoracolumbar laminectomy were enrolled.
Dogs that had neurologic signs for 43 months, chronic pain-
ful conditions, any contraindication to meningeal puncture, or
had been administered opioid or non-steroidal/steroid anal-
gesics during the week before surgery were excluded from the
study. Dogs were matched by surgical procedure, and using
a computer-generated random list were assigned to 1 of 2
groups: treatment with intrathecal morphine (MG), or no-
treatment (control group; CG).
On the day of surgery, dogs were administered romifidine
(5 dogs in MG; 5 in CG) or no-premedication (2 MG, 2 CG)
according to their temperament and the anesthetist’s prefer-
ence. Romifidine (0.004 mg/kg diluted 1:10 in normal
saline [0.9% NaCl] solution) was administered IV over 5 min-
utes in the cephalic vein through a preplaced catheter. General
anesthesia was induced by administering propofol IV to effect
(7 dogs) or midazolam (0.1 mg/kg) IV followed by propofol
to effect (11 dogs), and maintained with sevoflurane in
oxygen administered via a small animal circle breathing
system. Delivered fresh gas flow was 150 mL/kg/min for
the first 10 minutes after induction of anesthesia, then was
decreased to 10 mL/kg/min until the end of the procedure.
Intermittent positive pressure ventilation was imposed in
all dogs using a small animal volume-controlled time-cycled
ventilator (Merlin, Vetronic Services, Newton Abbot, Devon,
UK), and ventilator settings were adjusted to maintain an end-
tidal CO2 concentration of 36–40 mmHg. Five dogs (2 MG, 3
CG) had an MRI scan immediately before surgery.
Lactated Ringer’s solution was infused throughout the
procedure at 5 mL/kg/h, and adjusted to maintain an ade-
quate cardiovascular homeostasis. Ketamine (0.5mg/kg
IV slowly) was administered 10 minutes before surgery, then
every 60 minutes. A fentanyl loading dose followed by a con-
stant rate infusion (CRI) was administered throughout the
surgical procedure to provide background analgesia, and fur-
ther boluses (2 mg/kg in MG, 3 mg/kg in CG) administered in
the event of a contemporary increase in heart rate (HR) and
systolic blood pressure, 420% compared with their value
immediately before starting surgery. Fentanyl CRI was deliv-
ered at 1.2mg/kg/h in MG, and 4.2mg/kg/h in CG based on
our previous experience. A loading dose of fentanyl (4–5 mg/kg
in MG, 5–10 mg/kg in CG) was administered IV according to
HR and blood pressure, 6–10 minutes before the first incision.
Fentanyl was discontinued at the end of surgery, and IV
morphine (0.1 mg/kg) was administered to CG dogs 10
minutes before stopping fentanyl.
Total amount of fentanyl used in each dog was recorded,
and hourly fentanyl requirements calculated and compared
using a Mann–Whitney U-test. Based on anesthetic records, a
retrospective computer simulation was performed to predict
fentanyl plasma concentrations during surgery in each dog. To
calculate predicted plasma concentration at the time of re-
sponse to surgical stimulation a notebook (Travelmate
371TMi, Acer Inc., Taiwan) and a Target-Controlled Infu-
sion (TCI) Software (CCIP, Department of Anaesthesia and
Intensive Care, The Chinese University of Hong Kong, Hong
Kong) implemented with the pharmacokinetic model of fent-
anyl in dogs19
were used.
Inspiratory and end tidal concentrations of O2 and
CO2, electrocardiogram, pulse oximetry, heart and respirato-
ry rates, and temperature were monitored continuously
(Dynascope 5300, Fukuda-Denshi, Old Woking, UK). Insp-
iratory and end tidal sevoflurane concentrations were moni-
tored in 5 dogs (2 MG, 3 CG) and sevoflurane vaporizer
settings were recorded in all dogs. Invasive arterial blood
pressure was measured continuously after placing a catheter
in the right dorsal pedal artery, using a precalibrated trans-
ducer connected to a monitor. Palpebral reflex, eyeball posi-
tion, mucous membrane color, capillary refill time, pulse
quality, and jaw tone were clinically assessed at 5-minute in-
tervals. All relevant data were manually recorded at 5-minute
intervals.
Before preparation of the surgical site and placement of an
indwelling urinary catheter, a low dose of morphine (0.03mg/
kg) was administered intrathecally to MG dogs according to a
previously reported technique.17
Briefly, using a paraspinous
approach, a spinal needle was inserted at L5-6 or L6-7 level
and advanced into the dorsal subarachnoid space. Morphine
was administered slowly after the stylet was removed and
cerebrospinal fluid appearance at the hub of the needle con-
firmed the subarachnoid location of the tip. For practical and
ethical reasons, intrathecal administration of placebo was not
performed in CG dogs.
Data were analyzed using R statistical software (R Foun-
dation for Statistical Computing, Vienna, Austria; http://
www.R-project.org). Although Kolmogoroff–Smirnov (KS)
test suggested that some variables were normally distributed,
154 INTRATHECAL MORPHINE FOR SPINAL SURGERY

3. normal probability plots were used to further assess the dis-
tribution, because of the low power of KS test in small
samples. Because normal probability plots departed from
linearity, we used Mann–Whitney test to compare variables
between groups. Results are therefore expressed as median
(range). The area under the curve for sevoflurane setting was
calculated to compare the sevoflurane exposure between
groups.
RESULTS
Only 14 dogs were ultimately included in the study.
Two dogs were excluded because of incomplete data col-
lection, and 2 dogs because matching was not possible.
Final dog grouping was cervical dorsal laminectomy
(3 dogs/group) cervical ventral slot laminectomy (2 dogs/
group), and thoracolumbar (T11-L1) laminectomy
(2 dogs/group). All dogs were anaesthetized by the same
anesthetist (L.N.), and surgery was performed by
5 different surgeons (2 ECVN Diplomates, 1 ACVIM
[Neurology] Diplomate, 2 ECVN residents).
Body weight and age for MG dogs were 7.9 kg (range,
3–65 kg) and 106 months (range, 38–144 months) and for
CG dogs, 24.5 kg (range, 10–30 kg) and 84 months
(range, 5–140 months). There were 5 males (2 neutered)
and 2 spayed females in MG, 4 males and 3 spayed fe-
males in CG. Romifidine was administered 125 minutes
(range, 80–132 minutes) and 118 minutes (range, 80–147
minutes) before beginning surgery in MG and CG, re-
spectively. Seven MG dogs were administered 0.03mg/kg
(range, 0.023–0.034mg/kg) morphine intrathecally, at L5-
6 or L6-7 level, 41 minutes (range, 25–65 minutes) before
surgery. Anesthetic and surgical times were 240 minutes
(range, 142–336 minutes) and 115 minutes (range, 70–225
minutes) in MG, and 270 minutes (range, 195–358 min-
utes) and 151 minutes (range, 95–254 minutes) in CG,
respectively.
Sevoflurane vaporizer settings ranged from 1.8% to
3.0% in both groups, and the area under the curve for
sevoflurane exposure was not significantly different
(P ¼ .22) between groups. End-tidal sevoflurane concen-
tration monitored in 2 MG dogs and 3 CG dogs ranged
from 1.8% to 2.2% in both groups. Median HR and
systolic arterial pressure immediately before the first in-
cision (baseline) were 68 beats/min (b.p.m.; range,
54–91 b.p.m.) and 102 mmHg (range, 96–110mmHg) in
MG, and 64b.p.m. (range, 49–82 b.p.m.) and 109 mmHg
(range, 88–128mmHg) in CG; however, there was no
significant difference between groups for either variable
(P ¼ .71 and .13, respectively).
Administration of further fentanyl boluses was neces-
sary on 16 occasions in MG (5 dogs) and 13 occasions in
CG (7 dogs). Hourly fentanyl requirements were signifi-
cantly lower (P ¼ .004) for MG (4.75; range, 3.94–
6.61mg/kg/h) than for CG (10.29mg/kg/h; range, 6.35–
12.02 mg/kg/h; Fig 1). Predicted median fentanyl plasma
concentration at the time of response to surgical stimu-
lation was significantly lower (Po.0001). in MG (0.68ng/
mL; range, 0.4–0.85 ng/mL) than in CG (1.22 ng/mL;
range, 0.88–1.8ng/mL; Fig 2). Two dogs in MG did not
require a fentanyl bolus during surgery, and predicted
plasma concentrations did not fall below 0.45 and
0.39ng/mL.
DISCUSSION
Our most relevant finding was that preemptive ad-
ministration of a low dose of intrathecal morphine had a
significant sparing effect on the hourly amount of fent-
anyl required to prevent an increase in HR and systolic
arterial pressure 420% in dogs undergoing spinal sur-
gery under balanced anesthesia. A clinically significant
contribution of intrathecally administered morphine to
intraoperative antinociception can therefore be speculat-
ed. No side effects and complications related to intra-
thecal morphine administration, or neurologic sequelae
of spinal puncture were noticed before discharge.
Accordingly, we suggest that preoperative intrathecal ad-
ministration of a low dose of morphine may be useful as
part of balanced anesthetic protocol in dogs undergoing
spinal surgery. We are currently investigating its contri-
bution to postoperative analgesia, and results will be
reported separately.
Fig 1. Hourly fentanyl requirements (lg/kg/ h). Median,
25th and 75th percentile, and range. Dogs in control group
required significantly more fentanyl per unit of time than dogs
administered intrathecal morphine (P ¼ .004).
155
NOVELLO ET AL

4. Intrathecally administered morphine has a smaller
distribution volume, longer permanence in the cerebro-
spinal fluid, greater spinal bioavailability, slower clear-
ance, and the greatest rostral spread compared with other
opioids, because it is more hydrophilic.20
These charac-
teristics result in a longer-lasting selective spinal analge-
sia21
and the unique ability to provide analgesia distant to
the site of administration,22
but also in a slow onset of
action.23
Accordingly, we administered intrathecal mor-
phine 25–65 minutes before surgery to allow onset of
drug effect. The dose required when the drug is admin-
istered intrathecally is lower compared with the epidural
route,24
resulting in negligible plasma concentrations
without systemic side effects, but still providing analgesia
as effective as continuous epidural patient controlled an-
esthesia in humans.10
The ability of morphine to provide distant analgesia
has been confirmed in dogs: preemptive epidural admin-
istration reduced halothane requirements in an experi-
mental model of front limb stimulation,15
provided
effective analgesia after lateral thoracotomy,25
and had
superior analgesic effects compared with IV administra-
tion.26
Because intrathecal administration in this species
is usually the consequence of a failed epidural ap-
proach,27–29
the technique has not been thoroughly in-
vestigated and the intrathecal dose is usually calculated
by reducing the intended epidural dose (0.1–0.3mg/kg)
by 50–60%.30
Preoperative intrathecal administration of
a dose as high as 0.15mg/kg has been associated with
bradycardia and hypotension responsive to atropine, and
transient myoclonus, hyper-reflexia, hyperesthesia, atax-
ia, hind limb paresis, and urinary retention.29
Although
many pharmacologic interventions were instituted, nalox-
one was not administered to reverse the clinical signs,
therefore causes other than a sole morphine overdose
cannot be definitely ruled out. The average intrathecal
dose administered in humans is 2 mg per patient, and
some studies proved that further decreasing the dose may
reduce the incidence of side effects without affecting an-
algesia.31,32
Incidentally, the administration of a low dose
of intrathecal morphine has already been successfully re-
ported in dogs.17,18
Our study seems to confirm those
findings, as side effects or complication to morphine were
not detected in any dog.
Study limitations were related to the use of clinical
cases: the study was not blinded; some dogs had MRI
before surgery; romifidine was not administered to all
dogs; intrathecal morphine dose was not identical in all
dogs; ketamine was administered during surgery; 2 differ-
ent protocols for fentanyl administration were used; and
end-tidal sevoflurane concentration was only measured in
5 dogs.
Blinding of the anesthetist would have required pre-
paring fentanyl infusions in different dilutions to achieve
the same infusion rate and the same volume for the
loading doses and the intraoperative boluses in the
2 groups. Use of an intrathecal injection of placebo was
considered to be associated with significant ethical
implications. To avoid these problems, we decided
that blinding was not critical, because the intraoperative
administration of further boluses of fentanyl was
strictly guided by a defined hemodynamic response (i.e.
increase in HR and systolic arterial pressure 420%
above baseline).
Dogs that had MRI before surgery had a longer an-
esthetic period. Because hypothermia may affect anes-
thetic depth and hemodynamic response, we monitored
body temperature and prevented hypothermia, ensuring
that all dogs had body temperature 436.51C.
Romifidine was not administered to all dogs. Romifi-
dine is an a-2 adrenergic agonist closely related to
clonidine, and is usually administered to provide seda-
tion. In dogs, most of the hemodynamic modifications
induced by romifidine 5 mg/kg IV are within the normal
range of healthy awake dogs and typically return to
baseline value within 60 minutes after administration.33
Intraoperative analgesic and anesthetic-sparing effects
of this dose of romifidine have not been reported, and
romifidine pharmacokinetic and pharmacodynamic data
are not currently available in dogs. In our study,
romifidine was not administered to 2 dogs in each group.
These dogs had MRI, therefore, part of the anesthesia
was conducted by a different anesthetist, who did not
Fig 2. Predicted plasma fentanyl concentration (ng/mL)
when hemodynamic response to surgical stimulation occurred.
Median, 25th and 75th percentile, and range. Dogs in control
group responded at significantly higher predicted plasma con-
centrations than dogs administered intrathecal morphine
(Po.0001).
156 INTRATHECAL MORPHINE FOR SPINAL SURGERY

5. consider romifidine necessary. Fentanyl requirements,
predicted fentanyl plasma concentrations at time of
hemodynamic response, and sevoflurane vaporizer set-
tings in these dogs did not differ from those of dogs ad-
ministered romifidine within the same group. Although
published data and the results of this study suggest that
the analgesic, cardiovascular, and anesthetic-sparing
effects of such a low dose administered at least 80 min-
utes before incision were probably negligible, it would
have been ideal that all dogs were administered the same
preanesthetic medication.
All dogs were administered a low dose of intraoper-
ative ketamine, as it has been suggested, although not
demonstrated, that ketamine may affect postoperative
analgesic requirement by preventing development of
central sensitization and acute tolerance to opioids.
Currently, no data support a direct anesthetic or analge-
sic effect of doses as low as 0.3–0.5 mg/kg/h, or suggest
a pharmacokinetic interaction with fentanyl, which
in our setting would be present in both groups. We can-
not exclude, although it is extremely unlikely, that an
interaction between the low dose of ketamine adminis-
tered and the intrathecal morphine may have partially
accounted for the observed sparing effect on fentanyl
requirements.
End-tidal agent monitoring would have ensured that
all dogs were exposed in principle to an equipotent
amount of inhalational agent. For technical reasons, end-
tidal agent monitoring was performed in only a few dogs
(5 of 14). In the attempt to reduce the variability of ex-
posure to sevoflurane, we used high fresh gas flow rates at
the beginning of anesthesia. This approach, in addition to
chemical–physical properties of sevoflurane, is expected
to result in a rapid decrease of sevoflurane uptake over-
time, with minimal difference between the delivered and
the end-tidal concentrations, as confirmed by data col-
lected from dogs in which end-tidal agent concentration
was measured.
We used 2 different protocols to administer fentanyl
intraoperatively. A possible approach to investigate the
sparing effect of intrathecal morphine on fentanyl re-
quirements would consist of using the same infusion rate
and the same bolus dose of fentanyl in both treatment
and control groups. The effect of morphine would then
be assessed according to the number of extra boluses re-
quired to warrant hemodynamic stability. Such an
approach, which would be apparently logical, would re-
sult in 2 possible scenarios. Fentanyl boluses would be
administered very frequently in CG, if the infusion rate
were chosen taking into account a possible effect of mor-
phine to avoid opioid overdose in MG. On the other
hand, signs of opioid overdose would be apparent in MG,
if the infusion rate were chosen assuming that intrathecal
morphine does not contribute to intraoperative anti-
nociception, and this assumption proves to be incorrect.
The same considerations apply to the boluses of fentanyl
administered to maintain hemodynamic stability.
Unpublished pilot data collected in a few dogs con-
firmed these 2 possible scenarios, and were considered
when planning this study. An inherent risk of this ap-
proach to testing our hypothesis is to bias the study
results by using fixed boluses of fentanyl, which may po-
tentially cause an over-estimation of fentanyl require-
ments,34
reflected by wide fluctuations of fentanyl plasma
concentration. Although we observed a significant differ-
ence in hourly fentanyl consumption, we deemed possible
that the study design may have in part affected the re-
sults. In attempting to test this possibility, we performed
a retrospective pharmacokinetic simulation, as already
described,34,35
using commercially available TCI software
implemented with the pharmacokinetics of fentanyl in
dogs.19
The computer simulation showed that hemody-
namic response to surgery appeared at similar predicted
plasma concentrations, which were very consistent within
the same dog (Figs 3 and 4) and the same group, but
significantly different between MG and CG (Po.0001).
Because both the measured hourly fentanyl consumption
and the predicted fentanyl plasma concentration at the
time of response to surgery are significantly different be-
tween the 2 groups, and predicted plasma concentration
at the time of response to surgery is very consistent,
any bias caused by the study design can be reasonably
ruled out.
Although predicted plasma concentrations are not a
substitute for measured plasma concentrations, it is wide-
ly accepted they are representative of the concentration
time profile, and measurement of actual plasma concen-
trations are not strictly required in clinical research.34,36,37
We relied on the data published by Sano, who charac-
terized the disposition of fentanyl in conscious dogs not
administered any other drug. The presence of anesthetic
drugs may result in pharmacokinetic interactions through
direct pharmacokinetic effects, as well as through the
physiologic effect of other agents, depending on drugs
being used. Although the use during anesthesia of a
pharmacokinetic model established in conscious patients
may underestimate the attained plasma concentrations,
this method has been used by other groups,36,37
and in
our study this is very unlikely to have altered our main
findings because the order of the interaction is likely to
have been similar in both groups.
The pharmacokinetics of fentanyl have been charac-
terized by Murphy in dogs undergoing enflurane anes-
thesia.38
Therefore we calculated the predicted fentanyl
plasma concentration using Murphy’s model to investi-
gate whether a different pharmacokinetic model could
affect our conclusions. Despite a significant difference in
clearance and steady state distribution volume, the 2
157
NOVELLO ET AL

6. models are characterized by a very similar elimination
rate constant (K10). The predicted plasma fentanyl con-
centrations calculated according to Sano’s model were
29% ( 5%) lower than those calculated for Murphy’s
model. This difference was very consistent over time. We
repeated the statistical analysis using the predicted plas-
ma concentrations calculated by Murphy’s model, and
the difference between the MG and CG was still signifi-
cant (Po.0001). Enflurane anesthesia causes a greater
reduction in cardiac output, arterial pressure, and portal
blood flow and oxygenation than sevoflurane in dogs
maintained at 1.0–1.5MAC.39,40
These differences may
explain why similar infusion schemes result in higher
predicted plasma concentrations performing the simula-
tion using Murphy’s model, as we observed.
In summary, lumbar intrathecal administration of
30mg/kg morphine 40 minutes before surgery contributes
significantly to antinociception during cervical and thor-
acolumbar spinal surgery, allowing reduction in the
amount of systemic opioids needed to control the car-
diovascular response to surgery. No side effects or com-
plications of spinal morphine were observed in the 14
dogs studied.
ACKNOWLEDGMENTS
The authors gratefully thank all staff and neurologists,
anesthesiologists, radiologists, nurses, and interns at the
Animal Health Trust and Dick White Referrals for their
cooperation and dedicated assistance during the study.
Authors also acknowledge the assistance of Y. Tam (Prince
of Wales Hospital, Hong Kong) in the implementation of
the CCIP software, and Anthony Absalom (Adden-
Fig 4. Time course of predicted plasma fentanyl concentration with relevant surgical times in a patient allocated to control group.
As in Fig 3, predicted concentrations at the time of hemodynamic response are very consistent throughout surgery, irrespective of
interval of bolus administration.
Fig 3. Time course of predicted plasma fentanyl concentration with relevant surgical times in a dog allocated to morphine group.
Predicted concentrations at the time of hemodynamic response are very consistent throughout surgery, irrespective of interval of
bolus administration.
158 INTRATHECAL MORPHINE FOR SPINAL SURGERY

7. brookes’ Hospital, Cambridge, UK) in the revision of the
paper.
REFERENCES
1. Joshi GP, McCarroll SM, O’Rourke K: Postoperative anal-
gesia after lumbar laminectomy: epidural fentanyl infusion
versus patient-controlled intravenous morphine. Anesth
Analg 80:511–514, 1995
2. Raisis AL, Brearley JC: Anaesthesia, analgesia and support-
ive care, in Platt SR, Olby NJ (eds): BSAVA Manual of
Canine and Feline Neurology (ed 3). UK, BSAVA, 2004,
pp 337–354
3. Sharp NJH, Wheeler SJ: Postoperative care, in Sharp NJH,
Wheeler SJ (eds): Small Animal Spinal Disorders. Diag-
nosis and Surgery. Philadelphia, PA, Elsevier, 2005,
pp 339–362
4. Jellish WS, Thalji Z, Stevenson K, et al: A prospective ran-
domized study comparing short- and intermediate-term
perioperative outcome variables after spinal or general an-
esthesia for lumbar disk and laminectomy surgery. Anesth
Analg 83:559–564, 1996
5. Mandel RJ, Brown MD, McCollough NC III, et al: Hypo-
tensive anesthesia and autotransfusion in spinal surgery.
Clin Orthop Relat Res 154:27–33, 1981
6. Matheson D: Epidural anaesthesia for lumbar laminectomy
and spinal fusion. Can Anaesth Soc J 7:149–157, 1960
7. Scoville WB: Epidural anesthesia and lateral position for
lumbar disc operations. Surg Neurol 7:163–164, 1977
8. Barron DW, Strong JE: Postoperative analgesia in major or-
thopaedic surgery. Epidural and intrathecal opiates. An-
aesthesia 36:937–941, 1981
9. Bourke DL, Spatz E, Motara R, et al: Epidural opioids dur-
ing laminectomy surgery for postoperative pain. J Clin
Anesth 4:277–281, 1992
10. Brown DR, Hofer RE, Patterson DE, et al: Intrathecal an-
esthesia and recovery from radical prostatectomy: a pro-
spective, randomized, controlled trial. Anesthesiology
100:926–934, 2004
11. Loper KA, Ready LB, Nessly M, et al: Epidural morphine
provides greater pain relief than patient-controlled intra-
venous morphine following cholecystectomy. Anesth Analg
69:826–828, 1989
12. Yeager MP, Glass DD, Neff RK, et al: Epidural anesthesia
and analgesia in high-risk surgical patients. Anesthesiology
66:729–736, 1987
13. Rawal N: Spinal opioids. Curr Opin Anaesthesiol 10:350–
355, 1997
14. Shulman M, Sandler AN, Bradley JW, et al: Post-thoraco-
tomy pain and pulmonary function following epidural and
systemic morphine. Anesthesiology 61:569–575, 1984
15. Valverde A, Dyson DH, McDonell WN: Epidural morphine
reduces halothane MAC in the dog. Can J Anaesth 36:629–
632, 1989
16. Troncy E, Junot S, Keroack S, et al: Results of preemptive
epidural administration of morphine with or without bup-
ivacaine in dogs and cats undergoing surgery: 265 cases
(1997–1999). J Am Vet Med Assoc 221:666–672, 2002
17. Novello L, Platt SR: Low-dose intrathecal morphine for
postoperative analgesia after cervical laminectomy. Vet
Regional Anaesth Pain Med 4:9–17, 2006 (www.isvra.org)
18. Novello L, Carobbi B, Bacon NJ, et al: Partial hemipelvec-
tomy with rapid recovery of ambulation and long-lasting
analgesia after intrathecal administration of low-dose hy-
pobaric bupivacaine and morphine. ECVS Meeting, Dub-
lin, Ireland, June 2007
19. Sano T, Nishimura R, Kanazawa H, et al: Pharmacokinetics
of fentanyl after single intravenous injection and constant
rate infusion in dogs. Vet Anaesth Analg 33:266–273, 2006
20. Ummenhofer WC, Arends RH, Shen DD, et al: Comparative
spinal distribution and clearance kinetics of intrathecally
administered morphine, fentanyl, alfentanil, and sufentanil.
Anesthesiology 92:739–753, 2000
21. Goyagi T, Nishikawa T: Oral clonidine premedication en-
hances the quality of postoperative analgesia by intrathecal
morphine. Anesth Analg 82:1192–1196, 1996
22. Gray JR, Fromme GA, Nauss LA, et al: Intrathecal mor-
phine for post-thoracotomy pain. Anesth Analg 65:873–
876, 1986
23. Liu SS, McDonald SB: Current issues in spinal anesthesia.
Anesthesiology 94:888–906, 2001
24. Chadwick HS, Ready LB: Intrathecal and epidural morphine
sulfate for post-cesarean analgesia—a clinical comparison.
Anesthesiology 68:925–929, 1988
25. Pascoe PJ, Dyson DH: Analgesia after lateral thoracotomy in
dogs. Epidural morphine vs. intercostal bupivacaine. Vet
Surg 22:141–147, 1993
26. Popilskis S, Kohn DF, Laurent L, et al: Efficacy of epidural
morphine versus intravenous morphine for post-thoraco-
tomy pain in dogs. J Vet Anaesth 20:21–25, 1993
27. Heath RB, Broadstone RV, Wright M, et al: Using bupiva-
caine hydrochloride for lumbosacral epidural analgesia.
Compend Contin Educ Pract Vet 11:50–55, 1989
28. Skarda RT: Local and regional anesthetic and analgesic
techniques: dogs, in Thurmon JC, Tranquilli WJ, Benson
GJ (eds): Lumb Jones’ veterinary anesthesia (ed 3). Bal-
timore, MD, Williams Wilkins, 1996, pp 426–447
29. Kona-Boun JJ, Pibarot P, Quesnel A: Myoclonus and urinary
retention following subarachnoid morphine injection in a
dog. Vet Anaesth Analg 30:257–264, 2003
30. Torske KE, Dyson DH: Epidural analgesia and anesthesia.
Vet Clin North Am Small Anim Pract 30:859–874, 2000
31. Yamaguchi H, Watanabe S, Motokawa K, et al: Intrathecal
morphine dose–response data for pain relief after
cholecystectomy. Anesth Analg 70:168–171, 1990
32. Milner AR, Bogod DG, Harwood RJ: Intrathecal adminis-
tration of morphine for elective caesarean section. A
comparison between 0.1 mg and 0.2 mg. Anaesthesia 51:
871–873, 1996
33. Pypendop BH, Verstegen JP: Cardiovascular effects of rom-
ifidine in dogs. Am J Vet Res 62:490–495, 2001
34. Howie MB, Cheng D, Newman MF, et al: A randomized
double-blinded multicenter comparison of remifentanil
versus fentanyl when combined with isoflurane/propofol
for early extubation in coronary artery bypass graft sur-
gery. Anesth Analg 92:1084–1093, 2001
159
NOVELLO ET AL

8. 35. Lotsch J, Dudziak R, Freynhagen R, et al: Fatal respiratory
depression after multiple intravenous morphine injections.
Clin Pharmacokinet 45:1051–1060, 2006
36. Albertin A, Casati A, Bergonzi PC, et al: Effect of two target-
controlled concentrations (1 and 3ng/mL) of remifentanil on
MACBAR of sevoflurane. Anesthesiology 100:255–259, 2004
37. Albertin A, Dedola E, Bergonzi PC, et al: The effect of add-
ing two target-controlled concentrations (1 and 3 ng mL1
)
of remifentanil on MACBAR of desflurane. Eur J An-
esthesiol 23:510–516, 2006
38. Murphy MR, Olson WA, Hug CC Jr: Pharmacokinetics of
3H-fentanyl in the dog anesthetized with enflurane. Anes-
thesiology 50:13–19, 1979
39. Mutoh T, Nishimura R, Kim HY, et al: Cardiopulmonary
effects of sevoflurane, compared with halothane, enflurane,
and isoflurane, in dogs. Am J Vet Res 58:885–890, 1997
40. Frink EJ, Morgan SE, Coetzee A, et al: The effects of sevo-
flurane, halothane, enflurane, and isoflurane on hepatic
blood flow and oxygenation in chronically instrumented
greyhound dogs. Anesthesiology 76:85–90, 1992
160 INTRATHECAL MORPHINE FOR SPINAL SURGERY