Materials and Methods

Animals—Eight clinically normal adult mixed-breed sexually intact female cats (weight range, 2.0 to 4.0 kg) were included in the study. The cats were housed according to the principles of the University Research Ethical Committee, and the study was approved by the Institutional Animal Care Committee (Protocol No. 106/2004-CEEA). Health status was assessed via serum biochemical analyses, CBC, ECG, and venous blood gas analysis. The cats were housed in groups and fed dry laboratory cat food supplemented with canned food. Food was withheld for 12 hours and water was withheld for 2 hours prior to each experimental procedure.

Phase I—Cats underwent 2 treatments (remifentanil infusions) in a randomized crossover design with an interval of at least 10 days between treatments. Cats were administered acepromazine^a (0.05 mg/kg) IM. After 35 minutes, a 22-gauge catheter was placed in a cephalic vein. After 40 minutes, anesthesia was induced with propofol^b (6.0 mg/kg) administered IV over a 30-second period. Following endotracheal intubation, the cats were positioned in right lateral recumbency over a circulating warm water blanket.^c Throughout the experiment, body temperature was maintained at 36.4° to 37.8°C with the aid of a forced air blanket.^d By use of a volume-cycled ventilator,^e intermittent positive-pressure ventilation was instituted to maintain ETCO₂ at 30 to 35 mm Hg during anesthesia. Because ETCO₂ generally underestimates PaCO₂ by approximately 5 mm Hg, it was assumed that an ETCO₂ of 30 to 35 mm Hg would result in PaCO₂ values within reference range (35 to 40 mm Hg). Oxygen flow rate was set at 200 mL/kg/ min during anesthesia. The peak inspiratory pressure was adjusted to 7 to 10 cm H₂O, and respiratory rate was adjusted to maintain ETCO₂ within the expected range. Airway gas samples were continuously collected (200 mL/min) from the proximal end of the endotracheal tube into an infrared gas analyzer^f for ETCO₂ determinations.

By use of 2 syringe infusion pumps,^g,h a fixed-rate infusion of propofol (0.3 mg/kg/min) was administered with 1 of 2 randomly assigned infusion rates of remifentanilⁱ: 0.2 μg/kg/min (treatment 1) or 0.3 μg/kg/min (treatment 2). The propofol and the assigned remifentanil infusions were administered simultaneously at a total rate infusion of 4.3 mL/kg/h during a 90-minute period. Infusions were started immediately after orotracheal intubation. No additional fluids were administered IV.

Cardiopulmonary measurements—After induction of anesthesia, adhesive electrodes were placed for a continuous lead II ECG. Heart rate and rhythm were obtained from the ECG tracings recorded by a multiparametric monitor.^j Via aseptic technique, a femoral artery was surgically exposed to facilitate catheterization with a 24-gauge catheter connected to a blood pressure transducer system^j for SAP, DAP, and MAP determinations. The zero reference point of the pressure transducer was set at the level of the heart. Blood gases were measured^k from samples collected from the femoral artery; values were corrected to esophageal temperature.^j After completion of the study, the arterial catheter was removed and the incisions in the subcutaneous tissue and skin were sutured.

All cardiopulmonary measurements were recorded at 15, 30, 45, 60, 75, and 90 minutes after induction of anesthesia. For each data sampling interval, incidences of bradycardia (defined as HR < 100 beats/min¹⁸) and hypotension (defined as MAP < 60 mm Hg) were determined for both remifentanil treatment groups.

Antinociceptive response—For each cat, the response to a noxious stimulus was assessed by one of the investigators (MAC) who was unaware of the infusion rate of remifentanil administered. The noxious stimulus consisted of electrical stimulation (50 V, 50 Hz, and 10 milliseconds) applied SC via two 25-gauge needles placed 2 cm apart on the lateral aspect of the radius-ulna area of the left forelimb.^l Electrical stimulation was maintained for 60 seconds or less if a response was detected. A response was defined as detection of gross purposeful movement of the head or extremities. The noxious stimulus was applied at 15, 30, 45, 60, 75, and 90 minutes after induction of anesthesia, after cardiopulmonary measurements were completed.

Phase II—A period of at least 10 days was allowed to elapse after completion of phase I before the same cats were used in phase II. All 8 cats underwent ovariohysterectomy performed by the same investigator (AFML). For the purposes of data collection, specific points during the procedure were defined: immediately before skin incision (T₁), at the midpoint of skin incision (T₂), immediately after celiotomy (T₃), during grasping and ligation (just before excision) of the left (T₄) and right ovary (T₅), at the time that the uterus was clamped for performing the hysterectomy (T₆), at the midpoint of closure of the abdominal wall (T₇), and at the midpoint of skin closure (T₈).

The anesthesia protocol was the same as that used during phase I, with the exceptions that a variable infusion regimen of remifentanil was used and that only a circulating warm water pad^c was used in an attempt to prevent hypothermia. For phase II, the initial infusion rate of remifentanil was 0.2 μg/kg/min, which was adjusted on the basis of cardiovascular responses to surgical stimulation. If MAP increased by > 20% from baseline value in response to surgical stimulation (considering baseline MAP as the value recorded immediately before one of the aforementioned surgical events), the surgical stimulus was immediately interrupted and the infusion rate of remifentanil was increased by 5% from the previous value. A period of at least 3 minutes at the new infusion rate was allowed before continuing surgery. The 3-minute waiting period after a change in remifentanil infusion was chosen because in humans, remifentanil CRI rapidly results in steady-state plasma concentrations (< 2 minutes after starting the CRI), even in the absence of a loading dose.¹³ The 5% increase in remifentanil infusion rate was repeated until the increase in MAP was prevented and, thereafter, the CRI was maintained at that rate. On the basis of pilot study findings, periods of maximal nociceptive stimulation (resulting in more pronounced changes in arterial pressure) appeared to be associated with T₄ and T₅. Therefore, opioid infusion was increased as previously described if an increase in MAP > 20% was evident in response to traction of the ovarian pedicles. Further increases in remifentanil infusion rate were performed until the MAP change in response to ovarian pedicle traction was blocked. If an increase in remifentanil infusion rate was necessary during T₄ or T₅, opioid infusion was decreased to the initial rate (0.2 μg/kg/min) for at least 3 minutes prior to T₆ and increased again, if necessary, following the same described protocol until the end of the procedure. The total volume of propofol and remifentanil infusions was initially 4.3 mL/kg/h. No additional fluids were administered IV during surgery. Both propofol and remifentanil infusions were discontinued at the end of surgery.

During phase II, arterial blood gases were measured 15 and 30 minutes after starting the remifentanil infusion. Cardiovascular variables (HR, SAP, DAP, and MAP) were measured at all time points from T₁ through T₈. The time periods T₄, T₅, and T₆ were considered the periods of maximal intensity of noxious stimulation. Incidences of bradycardia and hypotension were determined for each interval. Meloxicam^m (0.2 mg/kg, SC) and buprenorphineⁿ (0.02 mg/kg, SC) were administered for postoperative analgesia at the end of surgery during phase II. Duration of surgery (time from skin incision to completion of skin closure) and duration of anesthesia (time elapsed from induction of anesthesia until cessation of drug infusions) were recorded during phase II.

Anesthesia recovery characteristics—During both phases, cats' recoveries from anesthesia were monitored; time to extubation, time to head lift, time to attain sternal recumbency, and time to standing (defined as ability to ambulate without assistance) were recorded. All aforementioned intervals were considered as time elapsed from cessation of the CRI to observation of the specified event.

Statistical analysis—For data collected during phase I, statistical analyses were performed by use of ANOVA followed by a Tukey test for comparisons between treatment groups. Within group comparisons of data obtained from 30 to 90 minutes after commencing drug infusion with data obtained 15 minutes after commencing the infusion regimen were performed by use of repeated-measures ANOVA followed by a Dunnet test. Because of the categorical nature of the data, a 1-tailed 2-proportions test was used to compare antinociceptive responses to electrical stimulation between treatment groups. During phase II, an ANOVA followed by a Tukey test was used for data comparisons among all time points. Differences were considered significant at a value of P < 0.05. Data are expressed as mean ± SEM.

Results

Phases I and II were completed successfully in all 8 cats. Cardiac dysrhythmias were not detected in any cat during propofol-remifentanil anesthesia in either phase of the study.

Phase I—There were no significant differences in HR and arterial pressures between treatment groups (Table 1). When compared with values at 15 minutes after starting the remifentanil CRI, HR and arterial pressures in both groups were significantly increased during some data collection time points until the end of the observational period.

Bradycardia (defined as HR < 100 beats/min) was detected in 2 of the 8 cats during CRI of remifentanil at either dosage. In 1 cat, bradycardia (HR range, 72 to 98 beats/min) was evident at 15 through 45 minutes after starting the infusion of 0.2 μg of remifentanil/kg/min. When the same cat was anesthetized with the higher infusion rate of the opioid, bradycardia (HR range, 77 to 94 beats/min) was also present for a similar period. In the other cat, bradycardia (HR range, 68 to 98 beats/ min) was evident at 15 to 60 minutes after starting the 0.2 μg of remifentanil/kg/min. The HR in that cat was < 100 beats/min (range, 68 to 94 beats/min) at 30, 45, 75, and 90 minutes after starting the 0.3 μg/kg/min infusion of the opioid.

Hypotension (MAP < 60 mm Hg) was detected in 3 and 2 cats during CRI of remifentanil at the 0.2 and 0.3 μg/kg/min infusion rates, respectively. Mean arterial pressure remained < 60 mm Hg only during the early portion of the observational period (15 and 30 minutes). During the hypotensive period, MAP ranged from 49 to 58 mm Hg in treatment group 1 and from 49 to 54 mm Hg in treatment group 2. Comparisons among time points revealed that arterial blood gas variables and respiratory rate were not significantly changed during phase I (Table 1).

In treatment group 1 cats (those administered remifentanil at the lower infusion rate), purposeful movement in response to electrical stimulation was detected in 1 to 3 of the 8 cats at the various data collection times (Table 2). With the exception of purposeful movement in 3 of the 8 cats at 15 minutes after starting the higher opioid infusion rate, no movement in response to noxious stimulation was detected in any cat receiving treatment 2 (those administered remifentanil at the higher infusion rate) during the remaining observational period. The incidence of purposeful movement among the anesthetized cats was significantly less at 30 minutes after starting drug infusions (P = 0.02) in cats receiving the higher opioid infusion rate (no cats had purposeful movement), compared with cats receiving the lower opioid infusion rate (3 cats had purposeful movement). Although not significant (P = 0.06) because of the sample size, the incidence of movement in cats receiving the higher remifentanil infusion rate was lower at 45 minutes after starting the opioid infusions (movement was detected in 2 cats in treatment group 1 versus no movement by cats in treatment group 2).

Mean ± SEM time from cessation of the CRI to extubation was significantly longer in treatment group 2 (8.9 ± 1.3 minutes) than in treatment group 1 (4.9 ± 0.8 minutes). With treatments 1 and 2, mean time to head lift (79 ± 12 minutes and 71 ± 6 minutes, respectively), mean time to sternal recumbency (107 ± 12 minutes and 102 ± 10 minutes, respectively), and mean time until standing (125 ± 16 minutes and 140 ± 17 minutes, respectively) did not differ significantly.

Phase II—Arterial blood gas variables remained within reference ranges during anesthesia in all 8 cats (reference ranges for arterial pH, PaCO₂, PaO₂, and HCO₃⁻ concentration were 7.34 to 7.44, 31 to 43 mm Hg, 450 to 550 mm Hg, and 18.5 to 22.0 mmol/ L, respectively). The infusion rate of remifentanil was maintained at 0.2 μg/kg/min in all cats from T₁ until T₃. However, in all cats, remifentanil infusion had to be increased as a result of the cardiovascular response to ovariectomy. In 2 cats, the 5% increase in opioid infusion rate was performed 7 times until the MAP change in response to ovarian pedicle traction was blocked. When the left ovarian pedicle was grasped and ligated (T₄), remifentanil infusion rates ranged from 0.21 to 0.25 μg/kg/min. When the same procedure was performed in the right ovary (T₅), opioid infusion rates ranged from 0.21 to 0.27 μg/kg/min. The remifentanil infusion rates at T₆ (clamping of the uterus for hysterectomy) and at T₇ (midpoint of abdominal wall closure) were similar (range, 0.20 to 0.23 μg/kg/min), whereas at T₈, the remifentanil infusion rate was decreased to 0.2 μg/kg/min in all cats. There were no significant differences among remifentanil infusion rates at the various data collection points.

In 6 of the 8 cats, remifentanil infusion could be maintained at the initial rate (0.2 μg/kg/min) from hysterectomy (T₆) until the end of surgery (T₈). In 1 cat, opioid infusion was maintained at 0.23 μg/kg/min after ovariectomy (through T₆ and T₇) and was later reduced to 0.2 μg/kg/min at T₈. In the other cat, opioid infusion was maintained at 0.22 μg/kg/min during T₆ and was then reduced to the initial infusion rate for the remainder of the procedure.

At T₄ through T₈, mean HR and MAP were significantly higher than values recorded at T₁ through T₃ (Table 3). Six cats had a HR < 100 beats/min. Bradycardia was detected in 2 cats at T₁ (immediately before skin incision) through T₃ (immediately after celiotomy), whereas 4 cats had bradycardia at T₂ (during skin incision) and T₃ only. After starting the ovariectomy procedure (T₄) through T₈, none of the 8 cats had bradycardia. Mean arterial pressure remained < 60 mm Hg only during the initial phase of anesthesia (T₁) in 2 cats, and another cat was hypotensive at T₁ through T₃.

Among the 8 cats, mean body temperature was significantly lower at T₅ through T₈, compared with values at T₁ through T₃. Mean durations of anesthesia and surgery were 58 ± 22 minutes and 37 ± 23 minutes, respectively. Mean times from cessation of the CRI to achieve extubation, head lifting, and sternal recumbency were 6 ± 1 minutes, 102 ± 36 minutes, and 112 ± 34 minutes, respectively. Mean time until standing was 114 ± 34 minutes.

Discussion

The present study was undertaken to evaluate the infusion rate of remifentanil necessary to control responses to 2 types of nociceptive stimulation (electrical stimulation and surgery) in propofol-anesthetized cats. The remifentanil infusion rates used during phase I were based on the infusion rates used for clinical anesthesia in humans.¹² During phase I, cats underwent nociceptive somatic stimulation by means of electrical stimulation applied to the subcutaneous tissue of the forelimb. By use of this type of noxious stimulation, it was evident that administration of 0.2 μg of remifentanil/kg/min via CRI in propofol-anesthetized cats did not result in a satisfactory plane of anesthesia; 1 to 3 cats had gross purposeful movement in response to the stimulation at some time point during the study. By contrast, increasing the remifentanil CRI rate from 0.2 to 0.3 μg/kg/min provided a more adequate depth of anesthesia; with the exception of the early maintenance phase (15 minutes after starting drug CRIs), none of the cats receiving treatment 2 responded to electrical stimulation.

The use of more than 1 method of noxious stimulation is recommended in studies involving assessments of the depth of anesthesia.¹⁹ For this reason, a second phase of the present study was performed in cats undergoing surgical stimulation (ovariohysterectomy). To administer remifentanil during this phase, a variable infusion regimen was adopted instead of maintaining a CRI. We hypothesized that maintenance of a satisfactory level of anesthesia could be achieved by adjusting the remifentanil infusion to control hemodynamic responses to surgery. During phase II, the highest mean dose of remifentanil used to control autonomic responses was 0.23 ± 0.01 μg/kg/min at T₄ and T₅ (left and right ovarian pedicle traction and ligation, respectively). In addition, clinical signs of surgical anesthesia (eg, absence of jaw tone and adequate muscle relaxation) were evident throughout surgery. These results indicated that propofol (0.3 mg/kg/min) combined with remifentanil infusion rates ranging from 0.2 to 0.23 μg/kg/min provided a satisfactory plane of anesthesia for performing ovariohysterectomy in cats.

On the basis of our study results, it appeared that electrical stimulation (phase I) is a more potent noxious stimulus than ovariohysterectomy (phase II) because the remifentanil infusion rate that provided adequate depth of anesthesia during ovariohysterectomy was lower than the infusion rate that provided the most suitable depth of anesthesia during electrical stimulation of the forelimb. This hypothesis is corroborated by experiments²⁰ performed in dogs and rabbits during inhalant anesthesia in which skin incision was a less potent noxious stimulus than electrical or mechanical (tail clamping) stimulation.

For noxious stimulation to be considered supramaximal, the intensity of the stimulus has to be increased to a certain threshold above which there are no further changes in the incidence of the observed response (eg, gross purposeful movement).^20–23 Indeed, different types and intensities of noxious stimulation may not induce the same degree of response. The mean ± SEM minimal infusion rate of propofol necessary to abolish the response to a tetanic electrical stimulus applied to the ulnar nerve in at least 3 of 6 cats was 0.10 ± 0.02 mg/kg/min.²⁴ In that study,²⁴ the minimal infusion rate of propofol necessary to abolish the response to clamping the tail or a phalanx with a hemostat in a similar number of cats was 0.15 ± 0.03 mg/kg/min and 0.21 ± 0.02 mg/kg/min, respectively. The apparent difference in minimal infusion rates of propofol suggests that the tetanic electrical stimulus applied to the ulnar nerve was not a supramaximal noxious stimulus, compared with the phalanx or tail clamp. However, those investigators did not report the intensity of electrical current applied to the ulnar nerve. The intensity of electrical current is an important factor to be considered because smaller electrical currents may not provide supramaximal noxious stimulation.²³ The electrical stimulation used in the present study is considered to provide supramaximal noxious stimulation, and similar adjustments have been used during minimum alveolar concentration studies^20–22 in dogs.

In addition to the different intensities of nociceptive stimulation applied during the 2 phases of the present study, different end points were used to judge whether depth of anesthesia was adequate during each phase. During phase I, the remifentanil infusion rate that would abolish purposeful movement in response to supramaximal electrical stimulation was evaluated, whereas during phase II, the infusion rate required to inhibit the cardiovascular response to surgery was assessed. Abolishment of gross purposeful movement in response to a supramaximal nociceptive stimulus does not imply that the autonomic (cardiovascular) response is also abolished. Results of 1 study¹⁹ in cats have indicated that end-tidal concentrations of inhalant anesthetics necessary to inhibit autonomic responses are relatively higher than those that inhibit purposeful movement.

Although the higher remifentanil CRI provided the most suitable depth of anesthesia from 30 to 90 minutes after starting the opioid infusion, purposeful movement in response to noxious stimulation was detected in 3 of the 8 study cats during the early phase of remifentanil infusion (ie, at the 15-minute time point). One may hypothesize that within this period (0 to 15 minutes), either plasma propofol or remifentanil concentration was insufficient to provide an adequate depth of anesthesia. However, data from a study¹³ in humans indicated that remifentanil rapidly results in steady-state plasma concentrations (within a period of < 2 minutes) even in the absence of a loading dose. The rapid equilibrium in plasma concentration is also associated with rapid establishment of the effects in the CNS as a result of the short blood-brain equilibrium time.¹³ Further studies are necessary to elucidate the pharmacokinetics of remifentanil and propofol in cats.

Depending on the method of noxious stimulation, it is evident that the propofol CRI used in the present study per se did not induce an adequate level of anesthesia in cats because the concomitant administration of 0.3 μg of remifentanil/kg/min was necessary to inhibit purposeful movement in response to electrical stimulation of the forelimb in all study cats. In humans, when propofol is used alone for maintaining surgical anesthesia, relatively higher doses are required than those required in a balanced anesthesia technique incorporating drugs such as opioids.⁹ In methadone-treated dogs undergoing ovariohysterectomy, when a variable infusion regimen propofol was combined with a CRI of remifentanil (0.6 μg/kg/min), the mean infusion rate of propofol required to maintain surgical anesthesia was 0.33 mg/kg/min.¹⁵

All cats in our study received acepromazine prior to induction of anesthesia and one should consider that this phenothiazine tranquilizer could have potentiated the hypnotic effects of propofol. In cats, propofol requirements for induction of anesthesia were reduced from 8 to 6 mg/kg after tranquilizer premedication.²⁵ However, this finding contrasts those of most other studies^26–28 in which premedication with acepromazine alone did not significantly alter the propofol induction dose in cats. On the basis of those reports, it appears that acepromazine either had no effect on the anesthetic properties of propofol or that this effect was minimal.

Propofol anesthesia may reduce arterial blood pressure via a decrease in peripheral vascular resistance and via myocardial depression.^29,30 In cats, induction and maintenance of anesthesia with propofol (6.6 mg/kg loading dose, followed by 0.22 mg/kg/min) was associated with stable hemodynamics.⁷ In the present study, cats were premedicated with acepromazine and this drug could have influenced cardiovascular variables because it is known to decrease blood pressure via blockade of peripheral α-adrenergic receptors.³¹ A reduction in HR and SAP was detected in a previous study³² in which cats were premedicated with acepromazine and anesthetized with propofol.

In cats, an HR < 100 beats/min may lead to cardiovascular function impairment.¹⁸ Despite the fact that mean HR values remained higher than this limit throughout phase I of the study, 2 of the 8 cats had persistent bradycardia (HR < 100 beats/min for more than 1 data collection time point) during this phase. During phase II, mean HR values were < 100 beats/min during the early period of anesthesia (T₂ and T₃) and 6 of the 8 cats had HR < 100 beats/min at > 1 data collection time point. Vagally mediated bradycardia is a common effect of pure M opioid receptor agonists, and it is likely that the reduction in HR in the cats of the present study was attributable to the vagotonic effects of remifentanil.³³ Although bradycardia was detected in some cats, MAP values remained > 60 mm Hg during both phases of the study and the incidence of persistent hypotension (MAP < 60 mm Hg at > 1 data collection time point) was low (1/8 cats during each phase of the study).

Treatment of vagally mediated bradycardia with an anticholinergic agent is recommended, especially when low HRs are associated with hypotension. In anesthetized dogs, reversal of vagally induced bradycardia with the anticholinergic glycopyrrolate results in an improvement in arterial blood pressure.³⁴ In another clinical study,¹⁵ 2 of 15 dogs anesthetized with propofol and remifentanil received atropine sulfate to correct bradycardia (HR < 60 beats/min). In our study, concomitant bradycardia and hypotension at > 1 data collection time point was detected in only 1 cat during both treatments in phase I, and this may have justified anticholinergic treatment in that individual. However, we decided not to administer anticholinergics for treatment of bradycardia because this would represent a confounding factor on cardiovascular data.

Pure μ (OP3) opioid receptor agonists are known to cause excitement and sympathetic stimulation in cats when high doses are administered.¹⁷ Gaumann et al³⁵ reported that a single bolus of sufentanil (25 μg/kg, IV) caused a significant increase in HR and arterial blood pressure in cats anesthetized with halothane.³⁵ The effects were attributed to central sympathetic stimulation or to epinephrine release caused by the drug.³⁵ In cats undergoing orthopedic procedures, a propofol infusion rate of 0.19 mg/kg/min combined with fentanyl (20 μg/kg/h) resulted in better maintenance of arterial blood pressure than the use of inhalant anesthetics (isoflurane or sevoflurane) in combination with a similar fentanyl infusion regimen.³⁶ In the present study, the improvement of arterial blood pressure over time in the cats during phase I could be related to a decrease in the influence of the propofol bolus used for induction of anesthesia and to an increase in sympathetic tone as a result of the CNS stimulant effect induced by the opioid. Noxious stimulation probably did not have a major impact on the cardiovascular changes detected during phase I because the electrical stimulus was performed after collection of the cardiopulmonary data. During phase II, it is possible that increased sympathetic tone caused by surgical stimulation could also have contributed to the increases in arterial blood pressure and HR detected at T₄ through T₈.

The use of propofol anesthesia in animals is associated with respiratory depression leading to hypercapnia.³⁷ The ultra–short-acting opioid agent remifentanil is also a potent respiratory depressant agent.¹² Therefore, artificial ventilation of the lungs is highly recommended during propofol and remifentanil anesthesia.¹⁵ Because of the use of artificial ventilation, ETCO₂ and PaCO₂ values were unaltered during the present study. The ETCO₂ value at 15 minutes after starting the infusion of 0.3 μg of remifentanil/kg/min in phase I was significantly lower than values at 45 to 90 minutes, but this occurred when ventilator settings were still being adjusted to achieve an ETCO₂ of 30 to 35 mm Hg. Despite this difference in ETCO₂, Pa^CO₂ did not change significantly during either phase of the present study.

In the present study, acepromazine was administered to the cats prior to induction of anesthesia, and that drug is known to induce hypothermia as a result of peripheral vasodilation and interference with the thermoregulatory center of the CNS.³¹ Hypothermia may reduce the clearance of anesthetics administered IV and may result in prolonged recovery from anesthesia.^38,39 In humans anesthetized with a constant infusion of propofol, plasma propofol concentrations were 28% greater during hypothermia (body temperature, 34°C) than during normothermia (body temperature, 37°C).³⁸ During phases I and II in the present study, mean esophageal temperatures at the end of anesthesia were approximately 37.5° and 34.7°C, respectively. This difference in body temperature between phases is likely related to the methods of hypothermia prevention and treatment used; during phase I, a circulating warm water heating pad was combined with a forced warm air blanket for maintaining body temperature, whereas during phase II, only a heating pad was used.

In humans, remifentanil is rapidly inactivated via the actions of nonspecific esterases, a mechanism that is independent of hepatic biotransformation and renal excretion.^12,40,41 The relative importance of this metabolic pathway has not been clarified in cats. However, the rapid recovery from neuromuscular blockade induced by atracurium, a drug inactivated via nonspecific esterases and via Hofmann's elimination, supports the hypothesis that esterase hydrolysis may contribute to drug inactivation in feline species.⁴² Although the rapid elimination of remifentanil has yet to be proven in felids, it appears that propofol infusion was the likely cause for the prolonged recovery from anesthesia in the cats of the present study. Anesthetic recovery in cats receiving a propofol CRI for > 30 minutes may be prolonged because of the low capacity for hepatic glucuronidation in felids.^17,43,44 In a recent study,⁴⁴ mean time from cessation of propofol infusion to the observation of cats walking without ataxia was increased from 80 to 148 minutes as the duration of the infusion was increased from 30 to 150 minutes. During both phases of our study, times for the cats' recovery from propofolremifentanil anesthesia were substantially prolonged, compared with recovery times recorded after maintenance of anesthesia with an inhalant agent.⁴⁵ Although maintenance of anesthesia with propofol and remifentanil for 90 minutes resulted in mean times until standing of 125 and 140 minutes (with low and high dose of remifentanil, respectively), maintenance of anesthesia in cats for a similar period of time with an inhalant anesthetic (isoflurane) resulted in mean times until standing of 10 minutes.⁴⁵

In the present study, a combination of CRIs of propofol (0.3 mg/kg/min) and remifentanil (0.2 to 0.23 μg/ kg/min) appeared to provide adequate clinical anesthesia in cats undergoing ovariohysterectomy. However, it was evident that the electrical stimulation used was a more potent noxious stimulus than surgery because a higher infusion rate of remifentanil (0.3 mg/kg/min) was necessary to reduce the incidence of stimulus-induced movement. It is noteworthy that, in cats, recovery from propofol-remifentanil anesthesia was substantially prolonged, compared with recovery from inhalant anesthesia.