Materials and Methods

Animals—Forty healthy client-owned adult bitches were selected for use in the study. Mean ± SD body weight was 13.7 ± 5.9 kg, and mean age was 16 ± 8 months. Health status was assessed on the basis that results of physical examination, CBC, and serum biochemical analysis (concentrations of BUN and creatinine and activities of alkaline phosphatase and alanine transaminase) were within reference ranges. Only dogs with calm and docile behavior that did not have evidence of discomfort or aversive reactions to palpation of the abdomen while interacting with the investigator during preoperative assessment were included in the study. The study was performed in accordance with the guidelines of the Brazilian College of Animal Experimentation and was approved by an institutional animal care committee (protocol No. 514 CEEA). Owner consent was obtained for each dog included in the study.

Anesthetic and presurgical procedures—Food and water were withheld from dogs for 12 and 6 hours before anesthesia, respectively. Acepromazine maleate^a (0.05 mg/kg, IM) was administered as a preanesthetic medication. Twenty minutes later, a 20-gauge catheter^b was inserted in a cephalic vein. Anesthesia was induced by administration of propofol^c (5 mg/kg, IV) 30 minutes after administration of acepromazine. The catheter was also used for fluid administration during anesthesia (lactated Ringer's solution at a rate of 10 mL/kg/h). After endotracheal intubation, anesthesia was maintained with isoflurane^d in oxygen (50 to 100 mL/kg/min) by use of a circle breathing system.^e An adapter was inserted between the orotracheal tube and circle circuit to allow aspiration of gases for continuous measurement of ETISO and PETCO₂ by use of a multivariable monitor.^f Isoflurane vaporizer settings were adjusted to maintain surgical depth of anesthesia on the basis of clinical signs, whereas positive-pressure ventilation was instituted to maintain PETCO₂ between 35 and 45 mm Hg throughout anesthesia.

Electrodes were affixed to the skin in accordance with a lead II ECG derivation to monitor heart rate, whereas a dorsal pedal artery or femoral artery was catheterized with a 20-gauge catheter^b connected to a fluid-filled pressure transducer system^g for recording MAP throughout anesthesia.^f Calibration (zero reference) for the pressure transducer was set at the level of the manubrium. Rectal temperature, recorded by the multivariable monitor^f via a rectal temperature probe,^h was maintained at > 37°C by means of a forced-air warming blanketⁱ and an electric heating pad.

Epidural treatments—After induction of anesthesia, dogs were positioned in sternal recumbency and the skin over the lumbosacral area was surgically prepared for aseptic insertion of a spinal needle into the lumbosacral epidural space. Correct needle placement was confirmed by the lack of CSF or blood at the needle hub and by a lack of resistance to the injection of 2 to 3 mL of air with a low-resistance syringe.

The study was randomized, placebo-controlled, and double-blinded. Dogs were assigned to 4 groups (10 dogs/group). The investigator who administered the epidural drugs and assessed postoperative pain (RLM) was unaware of the epidural treatment administered to each dog. Dogs in the control group received an epidural injection of saline (0.9% NaCl) solution (0.4 mL/kg). Preservative-free neostigmine sulfate^j (10 μg/kg) was administered to a second group, preservative-free morphine sulfate^k (0.1 mg/kg) was administered to a third group, and a combination of neostigmine and morphine (10 μg/kg and 0.1 mg/kg, respectively) was administered to the fourth group. In all treatment groups, final volume administered into the epidural space (0.4 mL/kg) was achieved by the addition of physiologic saline solution.

Doses of morphine and neostigmine selected for the study were determined on the basis of other reports.^1,5 The dose of morphine (0.1 mg/kg) has been associated with negligible undesirable effects in dogs and reportedly results in postoperative analgesia that may last up to 12 hours.¹ In human patients that underwent hysterectomy, coadministration of neostigmine (5 μg/kg) and bupivacaine as an epidural injection did not improve postoperative analgesia, compared with results for epidural administration of bupivacaine alone.⁵ However, coadministration of bupivacaine and a larger dose of neostigmine (10 μg/kg) prolonged the early postoperative analgesic effects of the bupivacaine.⁵

After administration of epidural treatments, dogs were positioned in dorsal recumbency for 30 minutes before surgery commenced. Heart rate, MAP, ETISO, PETCO₂, and rectal temperature were recorded at predefined intervals during anesthesia. Duration of surgery, length of abdominal incision, and variables recorded after discontinuation of the inhalant anesthetic (interval until first head lift and interval until standing) were recorded.

Surgical procedures—Bitches underwent routine ovariohysterectomy. An experienced surgeon (BWM) performed all surgeries, and a minimally invasive surgical technique was used. Duration of the surgical procedure was consistent among dogs (approx 15 minutes). Traction and exposure of the ovaries was accomplished with a metallic hook, which enabled the surgeon to use a relatively short abdominal wall incision (3 cm), although the incision was lengthened slightly when necessary.

Assessment of postoperative analgesia—Postoperative analgesia was assessed subjectively by 1 investigator (RLM) at 1, 2, 3, 4, 6, 8, 12, and 24 hours after the end of anesthesia. The investigator was unaware of the epidural treatment administered to each dog. Two methods were used for assessment. A 10-cm horizontal line without graduation was used to determine the VAS score. The left end of the line indicated a total lack of pain, and the right end of the line indicated the strongest pain possible. The evaluator observed and interacted with each dog (eg, coaxing the dog and palpating the surgical incision) and then quantified the degree of pain by indicating a point on the VAS line. The distance between the point indicated by the evaluator and the left end of the horizontal line was measured to record the VAS score (in centimeters). The investigator also assessed postoperative analgesia in each dog by use of a descriptive scale modified from an NDS reported elsewhere¹⁴ (Appendix). The NDS contained 11 variables, and the sum of all variables could yield NDS values ranging from 0 to 24, where 0 represented a total lack of behavior indicative of pain and 24 represented behavior indicative of the most intense pain possible.

The interval until return of appetite was also evaluated by offering a meal (the same food that the dog was accustomed to eating at home). The interval elapsed from discontinuation of anesthesia until the moment a dog had evidence of an appetite was calculated as the interval until return of appetite.

Supplemental analgesia (ie, rescue analgesia) was performed with morphine (0.5 mg/kg, IM) when the VAS score was ≥ 4 cm or the NDS score was ≥ 8. Possible adverse reactions, such as emesis and other effects, were also recorded during the postoperative period.

Statistical analysis—Data were analyzed by use of a commercial software program.^l Quantitative variables were compared by use of an ANOVA followed by a Tukey method for multiple comparisons. Qualitative variables (eg, number of dogs requiring rescue analgesia in each treatment group) were compared by use of a χ² or Fisher exact test. A Friedman test was used to compare scores within each group, whereas a Kruskal-Wallis test followed by a Dunn test was used to compare scores among treatment groups. Unless otherwise stated, physiologic variables were reported as mean ± SEM, whereas scores were reported as median and range of values. Significance was set at values of P < 0.05.

Results

Demographic data and variables recorded during anesthesia and recovery from anesthesia—No significant differences were detected among treatment groups with regard to age, body weight, and variables recorded during anesthesia (duration of surgery, length of surgical incision, heart rate, MAP, PETCO₂, ETISO, and rectal temperature). Anesthesia recovery characteristics (interval until first head lift and interval until standing) were also similar among treatment groups (Table 1). Two dogs receiving neostigmine and 1 dog receiving morphine had an episode of emesis after anesthesia was discontinued.

Assessment of postoperative pain—Significantly more control dogs (n = 8) received supplemental opioid analgesia during the first 4 hours after anesthesia because of a VAS score ≥ 4 or NDS score ≥ 8 than did neostigmine-treated dogs (1). Of the 8 control dogs that received supplemental morphine, 7 received analgesic supplementation only 1 time (from 1 until 4 hours after surgery), whereas 1 dog received analgesic supplementation 2 times (1 and 2 hours after surgery). At 2 hours after surgery, supplemental analgesia was administered to 1 dog treated epidurally with neostigmine, whereas no supplemental analgesia was required throughout the observation period in dogs treated with epidural administration of morphine or the combination of morphine and neostigmine.

The highest VAS scores were recorded 2 hours after surgery in the control group (median, 2.9; range, 0 to 4.1). Compared with values for the control group, the VAS scores were significantly lower for dogs receiving morphine-neostigmine (from 1 to 4 hours) or morphine alone (from 2 to 4 hours). Epidural administration of morphine and morphine-neostigmine resulted in lower VAS scores at 2 hours, compared with scores for epidural administration of neostigmine (Table 2).

The highest median value for NDS was recorded 2 hours after surgery in the control group (median, 7; range, 4 to 10). In comparison to values for the control dogs, the NDS scores were significantly lower for dogs receiving morphine-neostigmine (from 2 to 6 hours and at 12 hours) or morphine (all time points). The NDS scores were significantly lower in the morphine treatment group, compared with values for the neostigmine treatment group, at 3, 12, and 24 hours (Table 2).

Return to appetite was delayed in control dogs (mean ± SD, 17.3 ± 8.9 hours), compared with the values for dogs receiving morphine (3.1 ± 0.3 hours), morphine-neostigmine (3.2 ± 0.4 hours), or neostigmine (8.5 ± 8.3 hours).

Discussion

In the study reported here, epidural administration of neostigmine resulted in a moderate degree of analgesia, as indicated by the significant reduction in the use of supplemental analgesia during the early postoperative period in dogs that underwent ovariohysterectomy. However, its analgesic effectiveness was inferior to that for epidural administration of morphine or morphine combined with neostigmine. In humans that underwent knee surgery after receiving intrathecal administration of bupivacaine, postoperative analgesia induced by epidural administration of morphine (total dose, 0.6 mg) was significantly more prolonged when neostigmine (total dose, 60 μg) was added to the opioid, as indicated by a significant increase in the interval until administration of the first rescue analgesic (11 hours after surgery).⁶ On the basis of that study, we expected to achieve a more prolonged and effective analgesic effect by adding neostigmine to morphine for epidural administration. However, on the basis of the results of the study reported here, it was evident that the postoperative analgesic effects after epidural administration of the combination of morphine and neostigmine were similar to the postoperative analgesic effects detected after epidural administration of morphine alone. It is possible that the lack of differences between morphine and morphine-neostigmine was related to the type of postoperative pain in our study. Ovariohysterectomy in dogs is considered to result in mild to moderate pain in most instances.¹⁵ The degree of soft tissue trauma (and hence postoperative pain) resulting from the surgical technique used in our study was relatively small. In this experiment, it would have been difficult to detect an additional analgesic effect resulting from combining any other drug with morphine because epidural administration of the opioid induced excellent postoperative analgesic effects. Therefore, it may be hypothesized that the magnitude of postoperative pain resulting from this type of surgery was not sufficient to allow detection of a possible benefit that may have been achieved by adding neostigmine to morphine.

Given the subjective nature of the clinical assessment of postoperative pain and the considerable variability that may exist among animals in terms of response to pain,¹⁵ we sought to standardize the study population on the basis of physical and laboratory assessments and also on the basis of behavior and interaction with the individual responsible for evaluating postoperative pain. Variability among dogs of the study was likely reduced because demographic data were homogenous among treatment groups. Another possible confounder that may cause bias in experiments that compare the efficacy of drugs for alleviating postoperative pain is the type of surgery and amount of tissue trauma.¹⁵ Because variations in the amount of surgical trauma may result in differences in the intensity of nociception,¹⁵ the same experienced surgeon performed all surgeries by use of the same surgical technique to provide similar amounts of tissue trauma in all treatment groups.

Although quantitative variables, such as an increase in heart rate and blood pressure, have been used for the assessment of pain in animals, these variables are influenced by a wide array of nonpainful conditions that may limit their diagnostic value.¹⁶ Behavioral changes, although effective for assessing pain in many instances, may also vary considerably among animals and may be altered in hospitalized patients.^15,16 To our knowledge, there is no single method that has been proven to be optimally effective for the clinical assessment of postoperative pain. Therefore, a combination of a VAS and NDS was used in an attempt to provide the best assessment of postoperative pain. On the basis of the results of the study reported here, the NDS was more effective than the VAS for use in detecting the need for supplemental analgesia. For the 9 supplemental analgesic interventions in the control group (7 dogs received 1 supplemental intervention each, and 1 dog received 2 supplemental interventions), supplemental analgesic was administered on the basis that only the NDS score was above the threshold (≥ 8) for 6 assessments. In the control group, supplemental morphine was administered when only the VAS score was above the threshold (ie, ≥ 4) for 1 assessment and when both the VAS and NDS scores were above the respective thresholds for another 2 assessments.

The use of supplemental analgesics for treating animals with unacceptable signs of pain is a confounding factor and may mask the evaluation of the effectiveness of an original analgesic treatment. Therefore, the VAS and NDS scores reported here should be analyzed with the caveat that both scoring systems were influenced by the use of supplemental analgesia (0.5 mg/kg of morphine, IM) during the early postoperative period in 8 control dogs and 1 dog that received neostigmine. Given the relatively large number of control dogs that received supplemental analgesia, it is likely that had none of the control dogs received supplemental analgesia, pain scores would have been higher in this group and the differences in severity of pain between the control and other treatment groups would have been more evident. However, for humane reasons, pain studies that do not incorporate a plan for rescue analgesia are not acceptable.

Although the use of supplemental opioid analgesia in 8 of 10 control dogs made it more difficult to evaluate the analgesic efficacy of the epidural treatments, a difference between the control and other treatment groups was still detected by use of the VAS and NDS. On the basis of the results for both scoring systems, epidural administration of morphine and morphine-neostigmine provided the best analgesic effect that was evident mainly during the early postoperative period (up to 6 to 8 hours after termination of surgery and anesthesia). An alternative to dealing with the masking effect of rescue analgesia on the evaluation of efficacy of an analgesic treatment is to compare the number of animals that receive rescue analgesics or the number of rescue analgesic interventions required in each treatment group.^15,17 Although the analgesic effects of epidural administration of neostigmine were inferior to those after epidural administration of morphine or morphine-neostigmine, the use of neostigmine alone still resulted in an analgesic effect because fewer dogs in that treatment group received supplemental opioid analgesia in comparison with the number of control dogs requiring supplemental analgesia.

Assessment of the interval until the return of appetite during the postoperative period may provide an inverse correlation with the degree of patient comfort, and an earlier return of appetite may be expected in painfree animals. This variable was significantly delayed in control dogs, compared with results for the other treatment groups, and this result could be interpreted as a clinical sign of patient discomfort or pain. However, the sedative effects caused by the use of supplemental morphine during the early postoperative period in 8 of 10 control dogs may have represented a confounding factor for this variable.

In another study,⁴ investigators reported that the use of 1 to 4 μg of neostigmine/kg administered via the epidural route in combination with lidocaine resulted in dose-dependent analgesia in human patients who underwent minor orthopedic procedures. Another study⁵ involving the addition of 5 and 10 μg of neostigmine/kg to bupivacaine for epidural administration prior to hysterectomy in humans revealed that only the dose of 10 μg/kg resulted in significant postoperative analgesia, compared with results for the use of bupivacaine alone. However, the apparent discrepancy in the dose of neostigmine that provided effective analgesia in that study⁵ may have been related to the greater analgesic effectiveness of epidural administration of neostigmine for somatic pain, rather than for visceral pain.^4,18 Evidence that supports this viewpoint was obtained in a study¹⁸ in which intrathecal administration of neostigmine was more effective for alleviation of somatic pain (vaginoplasty) than for alleviation of visceral pain (tubal ligation) in women. The dose of neostigmine (10 μg/kg) in the study reported here was determined on the basis of the dose recommended for hysterectomy in humans.⁵

Antinociceptive effects of neostigmine may be more intense in female than in male subjects. Evidence for a sex-related difference in the antinociceptive efficacy of this drug was obtained in a study⁹ in mice in which neostigmine was 5 times as potent for the relief of pain in female mice as it was in male mice. Spinal analgesia induced by neostigmine is related to an increase in acetylcholine concentrations within the spinal cord, which stimulates muscarinic and nicotinic cholinergic receptors.^7–9 A greater concentration of nicotinic receptors has been identified in the spinal cord of female mice, and this difference in receptor population may explain the greater analgesic efficacy in females.⁹

In humans, subarachnoid (intrathecal) administration of neostigmine is associated with a higher incidence of vomiting and nausea during the postoperative period, compared with results after epidural administration of neostigmine.^{4–6,10–13} Emesis is believed to be secondary to rostral diffusion of drugs via the CSF after intrathecal administration, and epidural administration of neostigmine reduces the incidence of this effect in humans.^{4–6,10–13} Although vomiting was recorded during the postoperative period, only 2 dogs from the neostigmine group and 1 dog from the morphine group vomited, and vomiting did not result in further complications.

Epidural administration of neostigmine provided a moderate postoperative analgesic effect in most female dogs that underwent ovariohysterectomy in the study reported here. Analgesia induced by epidural administration of neostigmine was less, compared with the analgesia induced by epidural administration of morphine or morphine combined with neostigmine. In dogs with postoperative pain as a result of ovariohysterectomy, the addition of neostigmine to morphine did not potentiate the analgesic effects of the opioid. Because neostigmine may have a role as an adjuvant for pain control when combined with other analgesics for epidural administration, additional studies are needed to assess whether neostigmine improves the postoperative analgesia provided by drugs administered via the epidural route for dogs with postoperative pain resulting from other types of surgical procedures.