Materials and Methods

Animals—The study was approved by an institutional animal care and use committee (protocol 189/2007). Thirty client-owned dogs scheduled for surgery because of a single tibial or femoral fracture or unilateral hip joint luxation were included in the study after owner's consent was obtained. All surgeries were performed within 3 to 5 days after trauma. Exclusion criteria included an American Society of Anesthesiologists status of ≥ 3 on a 5-point scale, abnormal laboratory data, open fractures, signs of skin infection over the lumbosacral area, pregnancy, evidence of neurologic or neuromuscular disease, and aggressive behavior.

Anesthetic procedure and instrumentation—All dogs received meloxicam (0.1 mg/kg, PO, q 24 h) and tramadol (4 mg/kg, PO, q 8 h) from the occurrence of the trauma until the day before surgery. The last preoperative doses of meloxicam and tramadol were administered 24 hours and 12 hours prior to surgery, respectively. Dogs were premedicated with meperidine (4 mg/kg, IM). A 20- or 22-gauge catheter was introduced aseptically into a cephalic vein for drug and fluid administration, and 30 minutes after premedication administration, anesthesia was induced with propofol (4 to 6 mg/kg), given IV until endotracheal intubation could be accomplished. Anesthesia was maintained with isoflurane in a mixture of air and oxygen to yield an inspired oxygen fraction of approximately 0.8, which was administered through a rebreathing circuit. Flow rates were set at 100 mL/kg/min during the first 15 minutes and reduced to 30 to 50 mL/kg/min for the remaining anesthetic period. Lactated Ringer's solution was administered IV at a rate of 5 mL/kg/h throughout anesthesia. Intraoperatively, systolic, mean, and diastolic arterial blood pressures were monitored noninvasively by use of an oscillometric device. End-tidal carbon dioxide and isoflurane concentrations were measured by use of an infrared analyzer via a sampling line that was connected to the distal end of the orotracheal tube. Adhesive surface electrodes were attached to the skin in accordance with a lead II ECG and used to monitor heart rate.

The concentration of isoflurane was maintained initially at 1.4% and was adjusted during surgery on the basis of clinical signs, including absence of palpebral reflex, absence of jaw tone, absence of spontaneous breathing efforts during mechanical ventilation, and mean arterial blood pressure between 60 and 100 mm Hg. Dogs were mechanically ventilated to maintain the end-tidal carbon dioxide concentration between 35 and 45 mm Hg. Body temperature was monitored by use of an esophageal temperature probe and was maintained near physiologic values for dogs by means of a forced warm air blanket.

During surgery, bradycardia (heart rate < 60 beats/min) associated with persistent hypotension (mean arterial blood pressure < 60 mm Hg) for > 10 minutes was treated with atropine (0.04 mg/kg, IV). If hypotension was observed despite a heart rate ≥ 60 beats/min, dogs received an IV bolus of lactated Ringer's solution (10 mL/kg over 10 minutes); dopamine infusion was initiated (5 to 10 μg/kg/min, IV) if hypotension persisted after administration of 2 subsequent fluid boluses. A fentanyl bolus (2.5 μg/kg, IV) was administered if mean arterial blood pressure increased by > 20 mm Hg above the values measured before the start of surgery.

Epidural catheter placement—Following instrumentation, each dog was positioned in sternal recumbency, the skin over the lumbosacral area was surgically prepared, and an 18- or 20-gauge Tuohy needle^a was aseptically introduced into the lumbosacral epidural space with the bevel oriented cranially. Approximately 5 cm of an 18- or 20-gauge epidural catheter^a was inserted into the epidural canal as determined from markings on the catheter. Correct needle placement into the epidural space was confirmed by the hanging-drop technique or by lack of resistance to injection of 1 mL of saline solution and by the lack of CSF or blood obtained during aspiration. In the event of inadvertent subarachnoid puncture (verified by obtaining CSF during aspiration), the dog was excluded from the study.

Dogs received 2% lidocaine^b without epinephrine (5 mg/kg) through the epidural catheter, and surgery was initiated 30 minutes later. All surgeries were performed by 1 of 3 veterinarians. Duration of surgery was considered as the time from epidural administration of lidocaine until skin closure. A single dose of ceftriaxone (30 mg/kg, IV) was administered before the start of surgery for antimicrobial prophylaxis.

Experimental design—At the end of the surgical procedure, the dogs were randomly assigned to receive 1 of 3 treatments administered through the epidural catheter as follows: morphine sulfate^c (0.1 mg/kg), neostigmine methylsulfate^d (5 μg/kg), and morphine sulfate in combination with neostigmine methylsulfate at the same doses. The final volume of each epidural treatment was corrected with saline solution to achieve a standardized volume of 0.4 mL/kg and administered over 1 minute. Randomization was performed by use of closed envelopes after random allotment in a 1:1:1 ratio to assign 10 dogs to each of the 3 treatment groups.

After administration of the experimental treatment, isoflurane-induced anesthesia was maintained for another 45 minutes. This amount of time reportedly is adequate for onset of analgesia after epidural administration of morphine in dogs,¹ although the time to onset of analgesia after epidural administration of neostigmine has not been reported. Thereafter, isoflurane administration was discontinued and the dogs were allowed to recover from anesthesia.

Postoperative evaluation—Postoperative sedation was assessed by use of a VAS. The VAS consisted of a 10-cm line representing no sedation at the left end and the most sedation possible at the right end. An observer was responsible for placing a mark on the line that corresponded to the degree of sedation for the animal. The distance between the left end of the scale and the mark was considered the VAS score.

Signs of pain were assessed by use of a VAS and a CMPS.¹⁰ The CMPS was a behavior-based scale that included the following categories: demeanor, posture, comfort, vocalization, attention to surgical wound, mobility, and response to wound touch (Appendix). Possible pain scores evaluated by use of the CMPS ranged from 0 to 10. Pain scoring was standardized. First, the assessor observed the dog from outside the kennel to evaluate its posture, attention to surgical wound, and spontaneous vocalization. The assessor then entered the kennel, called the dog by its name, and encouraged it to stand and walk. Finally, gentle pressure was applied to the surgical site. On the basis of noninteractive and interactive behaviors of the dog, the assessor scored signs of pain first by use of the VAS and thereafter by use of the CMPS on the basis of each of the 7 behavior categories. All assessments were made by a single observer who was experienced in evaluating signs of pain and sedation in dogs and was familiar with the 2 scoring systems used in this study. This observer was unaware of the epidural treatment administered to each dog.

Baseline pain and sedation scores were obtained before premedication administration. Postoperatively, signs of sedation and pain were assessed 2, 3, 4, 6, 8, 10, 12, 16, and 24 hours after administration of the epidural treatment. Additional assessments of signs of pain were performed any time the blinded observer was concerned about patient discomfort. The first measurement (time point 2 hours) was performed approximately 75 minutes after discontinuation of anesthesia because of the 45 minutes allowed from administration of the epidural treatment until discontinuation of isoflurane-induced anesthesia. In dogs with a VAS or CMPS score ≥ 4 of 10, rescue analgesia was provided and consisted of morphine administered IV (0.2 mg/kg) and epidurally (0.1 mg/kg) and meloxicam (0.2 mg/kg, IM). Dogs continued to be scored after administration of rescue analgesia, but data from these dogs were excluded from further analysis. Times until the first administration of rescue analgesia after surgery, first food consumption, first spontaneous urination, and first defecation were recorded. After evaluations at the last time point (24 hours), the epidural catheter was removed and dogs were discharged from the hospital.

Statistical analysis—Data distribution was analyzed by means of the Shapiro-Wilk normality test. Mean body weight and age of dogs in each group were compared by use of an ANOVA and a Tukey test. Differences among groups in times until first food consumption, spontaneous urination, and defecation were compared by means of the Kruskal-Wallis test followed by a Dunn test for multiple comparisons. Differences among groups and over time in pain and sedation scores were analyzed by use of a repeated-measures design with treatment and time as main effects and the interaction treatment versus time in a generalized linear model with γ distribution. Multiple comparisons were obtained by decomposing the interaction in nested effects (ie, fixing time and comparing treatments and fixing treatments and comparing time by means of the Wald test).

The χ² test and survival analysis (log rank Mantel-Cox test) were used to compare the number of dogs that required rescue analgesia and the time of administration of the first rescue, respectively. For all analyses, values of P < 0.05 were considered significant.

Results

Demographic data for 30 dogs included in the study were summarized (Tables 1 and 2). No significant differences were found among the groups for body weight, age, and surgery time.

Fentanyl (2.5 μg/kg, IV) was administered during surgery to 3, 4, and 2 of 10 dogs each in the morphine, neostigmine, and morphine-neostigmine groups, respectively. The number of fentanyl bolus doses administered to each dog ranged from 2 to 4. Atropine was administered once to 1 dog in the morphine-neostigmine group. Seven of 10 dogs in each of the morphine and morphine-neostigmine groups and 6 of 10 dogs in the neostigmine group received a single bolus of lactated Ringer's solution IV (10 mL/kg over 10 minutes) because of intraoperative hypotension, and no dog in any group required dopamine infusion.

During 24 hours, rescue analgesia was provided for 4, 7, and 2 of 10 dogs each in the morphine, neostigmine, and morphine-neostigmine groups, respectively (Table 3). The number of dogs that received rescue analgesia was significantly different among groups at 2, 3, 4, and 6 hours after surgery. Survival analysis revealed that dogs in the morphine and morphine-neostigmine groups had a significantly lower probability of receiving rescue analgesia within 24 hours than did dogs in the neostigmine group (Figure 1).

Kaplan-Meier survival analysis of 30 dogs (n = 10/treatment) that underwent orthopedic surgery of a pelvic limb and received 1 of 3 epidurally administered treatments at the end of surgery as follows: morphine (0.1 mg/kg; dashed line); neostigmine (5 μg/kg; dotted line); and morphine in combination with neostigmine at the same doses (solid line). Curves of the morphine and morphine-neostigmine groups were significantly (P < 0.05) different from the curve of the neostigmine group.

Citation: American Journal of Veterinary Research 75, 11; 10.2460/ajvr.75.11.956

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Kaplan-Meier survival analysis of 30 dogs (n = 10/treatment) that underwent orthopedic surgery of a pelvic limb and received 1 of 3 epidurally administered treatments at the end of surgery as follows: morphine (0.1 mg/kg; dashed line); neostigmine (5 μg/kg; dotted line); and morphine in combination with neostigmine at the same doses (solid line). Curves of the morphine and morphine-neostigmine groups were significantly (P < 0.05) different from the curve of the neostigmine group.

Citation: American Journal of Veterinary Research 75, 11; 10.2460/ajvr.75.11.956

• Download Figure
• Download figure as PowerPoint slide

Kaplan-Meier survival analysis of 30 dogs (n = 10/treatment) that underwent orthopedic surgery of a pelvic limb and received 1 of 3 epidurally administered treatments at the end of surgery as follows: morphine (0.1 mg/kg; dashed line); neostigmine (5 μg/kg; dotted line); and morphine in combination with neostigmine at the same doses (solid line). Curves of the morphine and morphine-neostigmine groups were significantly (P < 0.05) different from the curve of the neostigmine group.

Citation: American Journal of Veterinary Research 75, 11; 10.2460/ajvr.75.11.956

• Download Figure
• Download figure as PowerPoint slide

Because data from dogs given postoperative rescue analgesia were excluded thereafter, only 6, 3, and 8 of 10 dogs each in the morphine, neostigmine, and morphine-neostigmine groups, respectively, had postoperative evaluation data at 24 hours. No significant differences were found among groups in pain scores determined by use of the CMPS. The VAS pain score was significantly higher in the neostigmine group than in the morphine group at 2 hours and higher than in the morphine-neostigmine group at 2 and 24 hours. The VAS pain score was higher in the morphine group than in the neostigmine group at 8, 10, and 12 hours and higher in the morphine group than in the morphine-neostigmine group at 2 and 4 hours. Pain scores determined by use of the VAS and the CMPS decreased from baseline values (ie, measurements obtained before premedication administration) from 2 to 24 hours in the morphine and morphine-neostigmine groups. Compared with baseline values, VAS pain scores were lower in the neostigmine group from 2 to 24 hours and CMPS scores were lower from 3 to 24 hours (Table 4).

Postoperative sedation scores were significantly higher than baseline values from 2 to 24 hours in the morphine group and from 2 to 12 hours in the neostigmine group. Sedation scores in the morphine-neostigmine group were significantly higher than the baseline value at 2, 3, and 4 hours and were significantly lower than that of the neostigmine group at 4, 6, and 10 hours (Table 4).

During the 24-hour observation period, all dogs in each group urinated and ingested food. One dog in the neostigmine group vomited after administration of rescue analgesia at 2 hours. One dog in the neostigmine group and 2 dogs in the morphine-neostigmine group did not defecate within 24 hours after surgery. No significant difference was found among groups in the times until first urination and first food consumption, but the time until first defecation was significantly less in the neostigmine group, compared with the time for the morphine-neostigmine group. Median time until first urination was 4 hours (IQR, 4 to 12 hours), 3 hours (IQR, 1 to 9 hours), and 8 hours (IQR, 4 to 16 hours) for the morphine, neostigmine, and morphine-neostigmine groups, respectively. Median time until first defecation was 6 hours (IQR, 2 to 24 hours), 1 hour (IQR, 1 to 3 hours), and 24 hours (IQR, 4 to 24 hour) for the morphine, neostigmine, and morphine-neostigmine groups, respectively. Median time until first food consumption was 7 hours (IQR, 4 to 24 hours), 4 hours (IQR, 4 to 7 hours), and 4 hours (IQR, 4 to 5 hours) for the morphine, neostigmine, and morphine-neostigmine groups, respectively.

Discussion

In a previous study,¹¹ epidural administration of neostigmine (10 μg/kg) provided effective postoperative analgesia in most dogs (9/10 dogs) following elective ovariohysterectomy. In the present study, epidural administration of neostigmine (5 μg/kg) did not provide effective postoperative analgesia following orthopedic surgery such that 7 of 10 dogs required rescue analgesia within 24 hours after surgery. Conversely, most dogs that received epidural administration of morphine (0.1 mg/kg) alone (6/10 dogs) or in combination with neostigmine (8/10 dogs) did not require rescue analgesia within 24 hours after surgery.

Clinical assessment of postoperative signs of pain in animals is difficult because absence of verbal response makes pain scoring a subjective interpretation of the animal's behavior. Because of the subjective nature of pain scoring, potential confounders should be eliminated. In the present study, the intent was to standardize the study population as much as possible such that demographic data were homogeneous among treatment groups. The extent and intensity of surgical trauma is another factor that might influence postoperative pain scores.¹² The fact that more than 1 veterinarian performed the surgical procedures might be a source of bias in the present study because inexperienced surgeons are likely to cause more tissue trauma. However, all 3 surgeons were experienced professionals; had one of the surgeons caused more tissue trauma during surgery than the others, this fact would have influenced all groups equally because the number of surgical procedures performed by each veterinarian was evenly distributed among the treatment groups. In addition, the duration of surgeries did not differ among groups. Finally, considering that the existing scoring systems used to evaluate postoperative signs of pain in dogs are subjective, both the VAS and CMPS were used in the present study. The use of 2 scoring systems reduced the chance that any dog that was in pain did not receive rescue analgesia.

Prior to orthopedic surgery in the dogs of the present study, lidocaine was administered epidurally because the effectiveness of intraoperative analgesia provided by epidural administration of neostigmine alone had not been determined in dogs and might have resulted in insufficient analgesia in the neostigmine group. It has been reported that epidural administration of lidocaine provides effective analgesia for orthopedic surgeries on the pelvic limb in dogs for up to 2 hours.¹³ Recovery from anesthesia for dogs in the study reported here was approximately 3 hours after epidural administration of lidocaine. Although residual analgesia in the early postoperative period cannot be discounted, lidocaine would have influenced all groups equally because duration of surgery did not differ among groups.

Robinson et al¹⁴ reported that fentanyl plasma concentrations of 0.95 ng/mL were associated with analgesia in dogs. In another study¹⁵ on the pharmacokinetics of fentanyl, analgesic plasma concentrations (approx 1 ng/mL) were detected for 45 minutes after administration of a bolus (10 μg/kg, IV) of fentanyl. In the present study, a fentanyl bolus (2.5 μg/kg) was administered when mean arterial blood pressure increased in response to surgery. The maximum cumulative dose of the opioid received by each dog was 10 μg/kg, administered over approximately 2 hours of surgery. Considering that, and also because 45 minutes were allowed between the end of surgery until anesthesia recovery, a residual analgesic effect of fentanyl during the early postoperative period seemed unlikely. In addition, no relationship could be found between the administration of fentanyl and the need for rescue analgesia.

In humans, epidural administration of neostigmine with coadministration of local anesthetics prolonged the time to first postoperative rescue analgesia, compared with that for lidocaine¹⁶ or bupivacaine alone.^8,17 It is unclear whether the analgesic effect of epidurally administered neostigmine is dose related. In 2 studies^16,18 in adults and children who underwent knee surgery or genitourinary tract surgery, no apparent dose-effect relation was found for neostigmine doses ranging from 2 to 4 μg/kg. Conversely, in another study,¹⁷ an analgesic effect in patients undergoing hysterectomy was observed at 10 μg/kg but not at 5 μg/kg. Because the dose of neostigmine for epidural administration has not been established for dogs, the neostigmine dose used in the present study (5 μg/kg) was determined on the basis of previous studies^8,16 conducted on the use of this anticholinesterase agent in humans. In these studies,^8,16 effective postoperative pain control after orthopedic surgery was achieved with doses ranging from 1 to 4 μg/kg.

The mechanism of action of intrathecally injected cholinomimetic drugs such as neostigmine has been attributed to an increase in spinal cord extracellular acetylcholine concentration resulting in increased muscarinic receptor activity.⁵ It has also been reported that the increase in the concentration of nitric oxide in the spinal cord may play a role in the analgesic effect of intrathecally administered neostigmine.¹⁹

In the present study, epidural administration of neostigmine alone failed to provide effective postoperative pain control in most (7/10) dogs following orthopedic surgery. More importantly, for 6 of 10 dogs, rescue analgesia was provided within 3 hours after surgery. These results are in contrast with results for the study by Marucio et al¹¹ wherein only 1 of 10 dogs given neostigmine (10 μg/kg) epidurally received rescue analgesia within 24 hours after ovariohysterectomy. The lack of analgesic effectiveness of epidural administration of neostigmine in the present study may have resulted from the lower dose used in this study (5 μg/kg), compared with the dose used in the study by Marucio et al¹¹ (10 μg/kg). Although doses as low as 1 μg/kg have been associated with analgesia in humans,¹⁶ it is possible that higher doses of neostigmine are needed in dogs. Another reason for the lack of analgesic effectiveness after epidural administration of neostigmine in the present study may have been related to the intensity of pain. Moderate to intense pain is expected after femoral head and neck excision or femoral or tibial osteosynthesis, whereas postoperative pain following ovariohysterectomy has been considered as mild to moderate.¹² Therefore, epidural administration of neostigmine alone may be effective for postoperative pain control in dogs with mild to moderate but not severe pain.

It has been reported that the duration of analgesia following epidural administration of morphine is long, ranging from 12 to 24 hours in dogs.¹ Rescue analgesia was not required within 24 hours after surgery in any of 10 dogs that received morphine (0.1 mg/kg) epidurally before elective ovariohysterectomy.¹¹ However, in dogs that underwent orthopedic surgeries of a pelvic limb, the results of postoperative pain scores and rescue analgesia requirements within 24 hours did not differ between dogs that received epidural administration of morphine (0.2 mg/kg) or saline solution.⁴ In the present study, duration of analgesia after epidural administration of morphine ranged from 3 to 24 hours following orthopedic surgery in dogs. Therefore, postoperative pain should be evaluated closely and additional analgesics should be administered as necessary.

In humans undergoing orthopedic knee surgery, epidural administration of morphine and neostigmine, combined with bupivacaine, prolonged the time to first administration of a rescue analgesic (approx 11 hours), compared with that for bupivacaine alone (approx 4 hours).⁸ In the same study,⁸ morphine-bupivacaine and neostigmine-bupivacaine failed to significantly prolong the time to the first rescue analgesia event (approx 7 and 6 hours, respectively). In another study²⁰ in humans who underwent abdominal surgery, patients who received epidural administration of neostigmine with bupivacaine prior to surgery had lower pain scores and required significantly less epidural administration of morphine for 48 hours after surgery, compared with patients who received only bupivacaine before surgery. Analysis of the results of these previous studies^8,20 suggest an improvement in the analgesic efficacy for epidural administration of morphine when the opioid was combined with neostigmine. In the present study, VAS pain scores were lower in the morphine-neostigmine group than in the morphine group throughout the 24-hour observation period, and a significant difference between groups was observed at 2 and 4 hours. However, although fewer dogs required rescue analgesia in the morphine-neostigmine group than in the morphine group (2 vs 4 dogs), it cannot be stated from these data that there was a significant difference. One point to take into consideration is that in the studies by Kirdemir et al²⁰ and Omais et al,⁸ neostigmine was administered epidurally prior to surgery, whereas in the present study, it was administered after surgery. In another study⁹ in humans, epidural infusion of neostigmine started before thoracotomy reduced postoperative pain scores and the need for postoperative analgesics, whereas epidural infusion of neostigmine initiated after surgery did not. Therefore, further studies are necessary to evaluate the analgesic efficacy of preemptive epidural coadministration of neostigmine with morphine in dogs.

It has been reported that epidural administration of morphine may result in adverse effects such as nausea, vomiting, urinary retention, pruritus, and respiratory depression in animals.¹ Of the adverse effects, respiratory depression and urinary retention are of the most concern. In the present study, urinary retention was not observed in any dog in the morphine and morphine-neostigmine groups, and although blood gas concentrations were not measured, no clinical signs of respiratory depression such as apnea or cyanosis were observed in any dog during the 24-hour observation period.

The main adverse effects reported in humans who received epidural administration of neostigmine, alone or in combination with a local anesthetic, were nausea and vomiting, with a great variation in the incidence (ranging from 8% to 60%).^8,9,16,17,21 According to these previous studies,^8,9,16,17,21 no apparent dose-effect relation existed between the dose of neostigmine for epidural administration and the incidence of these adverse effects. In addition, the incidence of nausea and vomiting was not significantly higher in patients given neostigmine epidurally, compared with the incidence in those that did not receive the anticholinesterase agent. In dogs that underwent ovariohysterectomy, postoperative vomiting was observed in 2 of 10 dogs given neostigmine (10 μg/kg) epidurally, which was not significantly different from the proportion of dogs given morphine (0.1 mg/kg; 1/10), the combination of neostigmine with morphine (0/10), or saline solution (control treatment; 0/10).¹¹ In the present study, vomiting was observed in 1 dog in the neostigmine group and 0 dogs in the morphine-neostigmine group. Results of the present study are in agreement with previous data, which indicated a low incidence of adverse effects following epidural administration of neostigmine and that neostigmine coadministered epidurally with morphine does not increase the incidence of adverse effects induced by the opioid.

One limitation of the present study was that several dogs were excluded from pain scoring after they received rescue analgesia. As a result, VAS and CMPS pain scores from only 3 of 10 dogs in the neostigmine group were included for statistical analysis after 6 hours. Because of the small number of subjects, the power of statistical tests was reduced, which increased the chance of a type II error (false negative). An alternative approach would have been to not exclude data from dogs receiving rescue analgesia, but this might also have minimized differences among treatment groups because pain scores in dogs receiving rescue analgesia would have been lowered. In addition to comparison of postoperative pain scores in the present study, the number of dogs given rescue analgesia and the time until first rescue event (survival analysis) were compared. Although comparisons in pain scores among groups was limited in the study reported here, postoperative analgesia in the neostigmine group was considered inferior to that in the morphine and morphine-neostigmine groups on the basis of the number of dogs receiving rescue analgesia and results of rescue analysis.

In the present study, epidural administration of morphine alone or in combination with neostigmine provided effective postoperative analgesia in most dogs after orthopedic surgery on a pelvic limb, whereas epidural administration of neostigmine alone was not effective for this purpose. Analysis of the results of this study suggested a potential role for neostigmine as an adjuvant for epidural analgesia in dogs undergoing orthopedic surgeries on the pelvic limbs. Further studies with larger groups of subjects are needed to determine the possible advantages of combining neostigmine with morphine for epidural analgesia in dogs.