Comparison of Heart Rate and Blood Pressure with Toe Pinch and Bispectral Index for Monitoring the Depth of Anesthesia in Piglets

Abstract

Determining depth of anesthesia (DOA) is a clinical challenge in veterinary medicine, yet it is critical for the appropriate oversight of animals involved in potentially painful experimental procedures. Here, we investigated various parameters used to monitor conscious awareness during surgical procedures and refined the application of noxious stimuli to anesthetized animals. Specifically we used a common stimulus, a compressive toe pinch (TP), to determine physiologic changes that accompanied a positive or negative motion response in isoflurane-anesthetized piglets. A positive response was defined as any reflexive withdrawal, whereas a negative response was defined as the absence of motion after stimulation. We also assessed the utility of the bispectral index (BIS) for its ability to predict a motion response to TP. The average of BIS values over 1 min (BIS_mean) was recorded before and after TP. In piglets with a positive response to TP, heart rate (HR), but not blood pressure (BP), increased significantly, but receiver operating characteristic (ROC) analysis revealed that HR was not a sensitive, specific predictor of TP motion response. Both before and after TP, BIS_mean was a strong predictor of a positive motion response. We conclude that HR and noninvasive BP changes are not clinically reliable indicators of anesthetic depth when assessed immediately after a peripherally applied compressive force as an indicator stimulus; however, BIS_mean and response TP are acceptable for assessing DOA in piglets maintained under isoflurane anesthesia.

Ensuring an appropriate depth of anesthesia (DOA) is a clinical and humane imperative for animals undergoing invasive procedures, yet there is a paucity of reliable standardized assessments in broad use in veterinary medicine. Current practice dictates that DOA is determined by evaluating changes in hemodynamic and respiratory variables and, in the absence of neuromuscular blockade, by observable withdrawal movement in response to a noxious stimulus. In research animals, the noxious stimulus is typically a nonstandardized compressive force applied to a toe, and absence of a pedal withdrawal reflex is the metric used to identify when animals are at an appropriate surgical plane of anesthesia prior to incision or invasive procedures.

Alternative methods to assess DOA may aid in the refinement of humane animal use in research. Although motor response to stimulation has been the most commonly used parameter to clinically assess depth of sedation and anesthesia in animals, this response may be independent of input from the forebrain and is mediated at the level of the spinal cord.^1,51-53 Alternatively, EEG measures are increasingly being used to measure the depth of sedation, anesthesia, and awareness.⁴³ Because hypnosis and amnesia are difficult to assess in some patients, namely in nonverbal patients such as animals and young human pediatric patients, the possibility of using refined EEG measures to determine the awareness or recall ability of patients is very promising.

The ancillary EEG measure that has been researched most extensively is the bispectral index monitor (BIS monitor, Covidien, Mansfield, MA). The bispectral index (BIS) is a parameter derived from the EEG; it is calculated by comparing the patient's EEG with a database of EEG recordings from sedated and conscious human volunteers that have been matched to observed DOA and drug doses. The proprietary algorithm used to determine BIS takes into account 3 main factors:^37,61,62 1) bicoherence or phase coupling, the amount of waveforms that are in phase together, 2) the power spectrum, the ratio of low- (δ) to high- (β) frequency waves; and 3) the percentage isoelectricity (burst suppression ratio) in the recording.^37,62 These metrics are based on the observation that the conscious mind has an EEG recording that is very random, high-frequency, and low-amplitude, reflecting desynchronized neuronal firing, whereas the unconscious mind has an EEG recording that is low-frequency, high-amplitude, and synchronized. Tailoring anesthetic administration to changes in BIS monitoring has been associated with a decrease in the incidence of intraoperative awareness as measured by postoperative recollection of surgical events.⁴⁴ In addition, the titration of anesthetic regimens to intraoperative BIS monitoring in humans has been associated with decreased time to first spontaneous breathing, eye opening, extubation, time to response to verbal command, and time spent in the postanesthesia care unit.^7,42 Improvements in these short-term outcomes are essential in the research environment, where reliable assessments in high-throughput systems are the desired goal yet difficult to achieve by using large-animal models.

BIS monitoring has not been restricted to human patients. In fact, it has been used in dogs, cats, horses, calves, goats, pigs, rabbits, chickens, alpacas, elk, dolphins, and a wild raptor to measure DOA, cerebral dysfunction, and sleep cycles, with varying success depending on species, anesthetic, and stimulus.^{2,4,6,16-20,23,29,37-41,54,56,62} Due to their frequency of use as surgical models in biomedical research, laboratory swine may benefit substantially from the investigation of BIS monitoring. Previous studies evaluating the use of BIS in pigs found varying utility of this modality to measure depth of anesthesia in this species. Overall, BIS decreases with the administration of anesthetics in swine.^{18,23,38,40,41,55,56} Although some investigators found low sensitivity of BIS in predicting response to movement⁴ and poor correlation to a visual analog scale of DOA,²³ others found that BIS correlated well to clinical measures of DOA⁴⁰ and that its utility depended on the dose of anesthetic administered.¹⁸ Changes in heart rate (HR) and blood pressure (BP) in swine did not affect BIS values during abdominal surgery.³⁸ The use of different anesthetic protocols in these studies makes comparisons across them difficult. The age, size, and breed of pigs differed between the studies as well, and the effect of these variables on BIS in pigs is unknown. For these reasons, BIS should be evaluated in pigs of different breeds and ages with different anesthetics.

In the present study, our objectives were to examine physiologic responses to a commonly applied compressive toe pinch (TP) without standardized equipment and to relate the motor response to TP and BIS as assessments of hypnosis in piglets. Our hypothesis was that a positive response to TP would correspond to changes in HR, BP, and level of hypnosis as measured by BIS. Outcomes from this study were intended to examine the veracity of concerns regarding the subjectivity of TP as a measure of anesthetic depth. To our knowledge, this study is the first to specifically relate BIS to motion response to noxious stimuli during the use of inhaled volatile anesthetics in swine.

Materials and Methods

Animals.

Female, purpose-bred Yorkshire-cross pigs (Sus scrofa domestica; age, 4 wk; weight: average, 8.9 kg; range, 7.3 to 10.8 kg) were obtained from a commercial dealer (Archer Farms, Darlington, MD). The source herd is free of pseudorabies, brucellosis, porcine respiratory and reproductive syndrome, and swine influenza virus. The animals were obtained for an approved protocol involving induced-injury models for which they would be anesthetized. Briefly, the surgical procedure included a crescent-shaped skin incision over the right orbit and a 1.5- to 2.0-cm diameter craniotomy to access the sensorimotor cortex of the piglets. A mechanical indenter then was used to produce a shallow depression in the cortical surface. All 33 pigs recovered from anesthesia once, and 26 of these animals were used for terminal experiments (tissue collection) 6 d later, where euthanasia occurred prior to recovery. The data presented here were obtained at time points that did not coincide with surgical manipulation of these animals but were prior to the manipulation (for assessment of response to noxious stimuli) or after the end of the manipulation and prior to extubation (for assessment of anesthetic recovery). The pigs were housed in indoor runs in an AAALAC-accredited facility and fed a standard laboratory diet (LabDiet 5081, St Louis, MO) with water provided via automatic watering system free choice. All procedures performed were approved by the University of Pennsylvania IACUC and in accordance with the Guide for the Care and Use of Laboratory Animals.³¹ Given the opportunistic nature of our experimental design, where recruited subjects were simultaneously enrolled in other studies with the requisite anesthesia induction and maintenance protocol, group sizes were not uniform.

Anesthesia and monitoring.

All pigs (n = 33) were premedicated with ketamine (20 mg/kg) and xylazine (2 mg/kg) IM prior to mask induction with 3% to 3.5% isoflurane. Endotracheal intubation was followed by a single dose of buprenorphine (0.02 mg/kg IM) in all animals. All pigs were mechanically ventilated during the experimental procedures (respiratory rate, 30 to 35 breaths/min; tidal volume, 8 to 10 mL/kg; peak airway pressure, 20 cm H₂O), beginning at the time of intubation. Anesthesia was maintained by using isoflurane with a target concentration during the procedure of 1.25%, as measured by end-tidal concentration (Et_ISO). In our experience, this target Et_ISO, in concert with these preanesthetic sedative and analgesic medications, provides an adequate level of anesthesia during the procedures, as measured by the absence of a pedal withdrawal reflex and no change in HR or BP during surgical incision, and maintains hemodynamic stability. Published minimal alveolar concentrations are usually higher when isoflurane is used as a sole agent in pigs,^40,64 but have been reported at a similar concentration as we provided here.¹⁸ HR, pulse oximetry (maintained greater than 97%), respiratory rate (30 to 35 breaths/min), end-tidal carbon dioxide (maintained at 35 to 45 mm Hg), Et_ISO, and systolic, diastolic, and mean arterial BP were measured by using a commercially available veterinary anesthetic monitor (Cardell MAX12 DUO HD, Midmark, Versailles, OH). Rectal temperature was maintained between 36 to 38 °C with the use of a circulating warm-water blanket (Gaymar, Orchard Park, NY). HR was recorded from the pulse wave of a peripherally placed pulse oximeter on the hindfoot opposite the blood pressure cuff. According to the manufacturer, this modality has an accuracy of ± 2 bpm. Noninvasive blood pressure was measured by using oscillometry immediately proximal to the tarsus, through a cuff whose width was approximately 2/3 the circumference of the pig's leg in the area where the cuff was placed. These monitoring modalities were chosen because they are noninvasive and easily adopted by scientific investigators.

Noxious stimulus.

A toe pinch (TP) was applied at 15-min intervals. All TP were performed in the absence of concurrent surgical manipulation. Motor response to TP, in the form of reflex withdrawal, was evaluated and designated as the standard-of-practice method for determining DOA. This included any such motor response that occurred within 30 s after application of the TP. To mimic the TP most commonly practiced by investigators at our institution, a uniform TP technique was performed by a single operator as follows: Carmalt forceps were used to apply a force on either the medial or lateral dewclaw (digit 2 or 4) at the level of the coronary band. The same pair of forceps was used throughout this study. The contact surface of the hemostat was midway along the curve of the instrument, not at the distal tip nor adjacent to the hinge. The pressure applied when closing the forceps on the digit was as consistent as possible and sufficient to blanch the subungual tissue, but the clasp of the hemostat was never locked. Force was held for 1 to 2 s and then released. Although there were occasionally residual depressions on the digit, force was not reapplied to any digit when a break in the integrity of the skin occurred (1 animal). Each pig received an average of 2.9 (range, 1 to 5) TP events per anesthetic event. A positive response was defined as any motion of the pig in response to the applied stimulus, including reflex withdrawal or purposeful movement of the neck, head, or any limb. No spontaneous or purposeful motion was noted, and all responses were therefore withdrawal of the limb in response to TP. A negative response was defined as the absence of such movement in response to the stimulus.

All pigs (n = 33) were evaluated for hemodynamic and motion response to TP. Prestimulation parameters were recorded at 30 s before the application of each stimulus. The poststimulation values for HR recorded were the peak (greatest) change from the prestimulation values over a 30-s period after application of the stimulus. Poststimulation BP was measured after initiating a new cycle on the oscillometric device at 30 s from application of the stimulus. Prior to initiation of the experimental injury procedure, tissue collection, or prior to extubation, isoflurane was titrated in 0.2% increments to lower concentrations to obtain at least one positive TP event from each animal. These positive events were included in the analysis. Data for HR, BP, BIS, and motion were recorded only after at least 15 min of equilibration at the decreased anesthetic dose.

BIS monitoring.

BIS values were recorded by using the BIS Monitor (A2000 version 3.3, Aspect Medical, Newton, MA) in a subset of all pigs (n = 22). Data were collected on a laptop computer through a serial port in 5-s increments. Factory default settings were used (smoothing rate, 15 s). Data with a signal quality index of less than 50 as reported by the monitor were discarded. Impedance was checked continuously by the monitor and kept at 10 kΩ or lower. After clipping of hair and degreasing with isopropyl alcohol solution, BIS pediatric sensor electrodes were placed in a frontotemporal configuration, as previously described in dogs.⁹ A diagram of the lead placement is shown in Figure 1. With this configuration, the BIS sensor was not moved to accommodate the surgical procedures, but recording did not continue during surgical manipulation. Recordings began at least 30 min after premedication, to minimize the effect of ketamine on the readings. This 30-min interval was based on a small (n = 3 piglets) pilot study to determine the duration of paradoxical increases in BIS caused by ketamine. BIS_mean was calculated as the average BIS value for a 1-min interval. BIS_mean was used to minimize effect of second-to-second variation in this measurement. BIS_mean was recorded before and after each noxious stimulus for all animals in this subgroup (n = 22 piglets) and at the time of extubation for some animals (n = 12 piglets).

Figure 1.

BIS pediatric lead placement on the piglet head. The numbered circles indicate the individual electrodes and their associated number as marked by the manufacturer. C, coronal plane; S, midsagittal plane.

Anesthetic recovery.

Readiness for extubation was compared with Et_ISO and BIS_mean in a subset of pigs that underwent survival procedures. Extubation criteria for the survival procedure included: spontaneous ventilation, appropriate oxygen saturation (>97% by pulse oximetry), normocarbia as measured by end-tidal carbon dioxide (35 to 45 mm Hg), the presence of a positive TP response, and sustained head lift. Time to extubation from cessation of isoflurane administration was recorded in minutes. Et_ISO and BIS_mean were recorded (n = 9 for Et_ISO and n = 12 for BIS_mean) at the time of extubation. Nine of the pigs had both Et_ISO and BIS_mean values, but the exhaled anesthetic concentration was not available for the remaining 3 pigs, in which only BISmean was recorded. On the basis of the experimental protocols, the total anesthetic time averaged 85 min (range, 59 to 115 min) for pigs. Timelines of the procedures are provided in Figure 2.

Figure 2.

Timelines for survival and terminal piglet experiments. The length along the horizontal axis is not to scale with the duration of each portion of the experiment.

Statistical analysis.

Statistical analysis was performed and graphs were generated by using SigmaPlot (version 12, Systat Software, San Jose, CA) or JMP 11 (SAS Institute, Cary, NC). Prestimulation and poststimulation variables were compared by using paired t tests when data were normally distributed or Wilcoxon signed-rank tests when the data were not normally distributed according to the Shapiro–Wilk test. Percentage changes in HR and BP values (positive and negative) were calculated as: (poststimulation value – prestimulation value) / prestimulation value × 100%. Mean values for groups were compared by using t tests, when normally distributed data were compared between 2 groups, or alternatively by using Mann–Whitney rank-sum tests. One-way ANOVA was used when comparing more than 2 groups. Linear regression analysis was used to investigate the effect of Et_ISO on BIS_mean at each time point and the effect of anesthesia depth, as measured by BIS_mean at the time of isoflurane cessation, on the time to extubation. Receiver operating characteristic (ROC) analysis was used to determine the ability of changes in HR, BP, and BIS_mean to predict a positive or negative motion response to TP, and AUC was calculated as a measure of fit, where AUC = 0.5 signifies no relationship and AUC = 1.0 is perfectly predictive.¹⁴ Logistic regression was used to test the significance of these relationships. Relationships were determined to be statistically significant at a P value of less than 0.05. All data are presented as mean ± 1 SD unless otherwise noted.

Results

Response to TP.

Prestimulation and poststimulation values were evaluated for positive and negative motor responses. Data were categorized according to the concentration of Et_ISO at the time of TP: low, 0.2% to 0.5% (n = 33); intermediate, 0.6% to 0.9% (n = 25); and high, greater than or equal to 1.0% (n = 25). However, one-way ANOVA revealed that the percentage change in HR and BP after positive and negative TP events did not vary with isoflurane concentration (that is, the increase in HR for pigs with positive events was not dependent on their isoflurane subgroup); therefore results were pooled for all further analyses. A statistically significant difference between prestimulation anesthetized baseline and postTP values was found for both HR (P = 0.042) and BP (P = 0.032) in the positive response to TP events (n = 58 TP events in 33 pigs) but not for the negative response to TP events (n = 87 TP events in 33 pigs; P = 0.027 and P = 0.27, respectively). Comparing percentage changes in HR and BP between positive and negative events revealed a significant difference in percentage change in HR (positive: 1.0% ± 0.4%; negative, 0.4% ± 0.3%; P = 0.027) but not in percentage change in BP (positive, 2.0% ± 6.5%, negative, 1.2% ± 7.0%; P = 0.27; Figure 3). However, the area under the ROC curve for HR (0.618, data not shown) demonstrated that although significant, the percentage change in HR was not strongly predictive of a motion response to toe pinch.

Figure 3.

Changes in physiologic parameters after toe pinch (TP), as a percent change from prestimulation value ± SEM, for both positive and negative events in isoflurane anesthetized animals (33 total animals). Asterisks signify statistically significant difference between groups (P < 0.05).

BIS_mean, and TP.

To evaluate whether BIS is a predictor for positive or negative TP responses, we compared BIS_mean before and after TP in isoflurane-anesthetized pigs. For negative responders to TP, BIS_mean before stimulation (56.9 ± 7.9) did not differ from that after TP (56.7 ± 7.9; P = 0.34). However, BIS_mean was significantly increased after stimulation among positive responders to TP (before, 67.7 ± 9.6; after, 70.3 ± 9.4; P = 0.04; Figure 4). BIS_mean values before and after stimulation were both strongly predictive of motion in response to TP during isoflurane anesthesia of pigs, according to ROC AUC (Figure 5). The ROC analysis revealed that at a poststimulation BIS_mean of 64.9, there was 96% sensitivity and 91% specificity for detecting a positive TP response in isoflurane-anesthetized pigs; in comparison, the prestimulation BIS_mean was as sensitive but less specific. At the optimal cut point of a BIS_mean of 59.9, there was 95% sensitivity and 73% specificity. No pig had a positive response to TP below a prestimulation BIS_mean value of 51.1. Poststimulation BIS_mean was as low as 45.3 in positive responders.

Figure 4.

BISmean values are plotted before and after positive and negative TP events. The boxes extend the interquartile range with the mean indicated by the central line. The whiskers extend from the 5th to 95th percentile; dots indicate outliers; *, significant (P < 0.05) difference between groups. The sample size (n) in this figure represents the number of TP events from 33 piglets.

Figure 5.

BISmean (A) before and after stimulation and (B) sensitivity and specificity to predict TP response in isoflurane-anesthetized piglets, plotted as a ROC curve; the AUC value shows excellent sensitivity and specificity. The optimal cut point for maximizing sensitivity and specificity is the point where the yellow tangential line contacts the ROC curve.

Anesthetic recovery.

BIS_mean and Et_ISO showed a significant negative linear relationship (Figure 6), such that higher Et_ISO was associated with lower BIS_mean values. We compared BIS_mean after stimulation and Et_ISO between positive responders to TP, negative responders to TP, and pigs clinically ready for extubation, as a measure of anesthetic recovery (Figure 7). BIS_mean poststimulation differed significantly between positive (73.9 ± 5.9) and negative (54.8 ± 7.5) responders to TP (P < 0.001) and between positive responders and those deemed clinically ready for extubation (88.2 ± 5.3; P < 0.001), indicating that BIS is predictive of gradations in anesthetic depth at low levels of anesthesia. Mean Et_ISO at the time of TP differed between positive (0.50% ± 0.16%) and negative (0.94% ± 0.28%) responders (P < 0.001), but no difference was detected when we compared average concentrations between positive TP responders and pigs ready for extubation (0.30% ± 0.25%; P > 0.05). No significant correlation was seen between the BIS_mean at time of turning off the isoflurane and the time to extubation (P = 0.803; data not shown).

Figure 6.

BIS_mean and Et_ISO before TP stimulation are strongly correlated, as determined by linear regression (BIS_mean before stimulation = 79.976 – (23.442 × Et_ISO); P < 0.001). These data were obtained from 27 anesthetic events in 15 pigs.

Figure 7.

(A) Et_ISO and (B) BIS_mean are compared at time points when piglets had a positive or negative TP responseor at extubation (Extube). The boxes extend the interquartile range with the mean indicated by the central line. The whiskers extend from the 5th to 95th percentile. Dots indicate outliers. *, Values differ significantly (P < 0.05; one-way ANOVA) between groups.

Discussion

The objectives of this study were to examine physiologic responses to a commonly used TP stimulus and determine the relationship between motor response to TP and bispectral index as an assessment of hypnosis in piglets enrolled in biomedical research protocols. We defined TP as a compressive stimulus to a dewclaw at the level of the coronary band. A positive response to this TP included any reflexive motion of the pig, and a negative response was defined as the absence of any motor response. The TP was used as the ‘gold standard’ for monitoring of anesthetic depth for the experiment. We evaluated the ability of BIS to predict movement in response to the TP stimulus.

Physiologic parameters and motion response to noxious stimuli form much of the basis of anesthetic effectiveness and monitoring in veterinary medicine. Changes in cardiovascular parameters have previously been evaluated in pigs in response to doses of different anesthetics as well as in response to noxious stimuli.^18,24,38 Our findings suggest that, compared with withdrawal, neither changes in HR nor noninvasively measured BP in response to brief TP are sensitive indicators of anesthetic depth. Even when a statistically significant change in HR was detected, this represented a difference of only 1 to 2 bpm. Not only is this difference clinically insignificant when observing a HR of 100 to 130 bpm, but the limit of detection of changes in pulse rate using this commercial anesthetic monitor is ±2 bpm, as indicated in the manufacturer's instruction manual. A true difference may not have been detected, and this finding may represent measurement error. This finding is in contrast to a previous study, which found that invasively measured mean arterial BP was the most sensitive of the parameters evaluated in detecting nociception in pigs under isoflurane anesthesia.²⁴ The discrepancies between the previous and our current results may be explained by the difference in direct compared with indirect measurement of HR and BP, the different duration and intensities of noxious stimulation applied, or our more balanced anesthetic regimen that includes analgesic administration and adjunct hypnotics. Specifically, the previous study used isoflurane as a single anesthetic agent without additional analgesics.²⁴ In the current study, pigs anesthetized with isoflurane were first premedicated with ketamine and xylazine, which have analgesic properties, and all animals received buprenorphine at the time of endotracheal intubation. The nociceptive blockade provided by these agents may have inhibited reflex autonomic responses to stimuli. Each of these drugs has been shown to decrease the minimal alveolar concentration of isoflurane.^{12,30,50,63,65} In addition, xylazine has well-documented depressive effects on the cardiovascular system, causing decreased cardiac output and hypotension after an initial phase of hypertension, by increasing vagal tone and decreasing sympathetic tone.⁶⁶ Ketamine is an N-methyl-d-aspartate receptor antagonist that limits central sensitization of pain to stimuli. Ketamine frequently is used in combination with xylazine for its sympathomimetic effects, which should dampen the pronounced depression of the cardiovascular system. Buprenorphine has its antinociceptive effect through actions on the opioid receptors of the central and peripheral nervous system.⁶⁶ The authors of the previous study did not separate animals that responded with motion to the stimuli from those that did not respond, and changes in the parameters reported were evaluated solely based on anesthetic level, defined by decreasing end-tidal isoflurane concentration from an initial finding of 20% burst suppression ratio on EEG. Importantly, although another study revealed differences in HR, mean arterial BP, and diastolic BP between pigs of different treatment groups according to anesthetic regimen and dosing, no differences were detected within an anesthetic regimen between those animals that received a noxious stimulus and those that did not, even when motion was detected.¹⁸ In contrast, our study evaluated positive and negative responders to a noxious stimulus under isoflurane anesthesia and found the percentage change in HR, but not noninvasive BP, was significant, albeit at a clinically irrelevant level. Although invasive BP measures would be more precise and sensitive than are measures used here, we chose noninvasive methods for this study because investigators at our institution who are performing experiments on anesthetized swine most often use these monitoring modalities.

In the present study, we focused on stimuli that are easily administered and practical for routine use. This feature contrasts with studies that determine minimal alveolar concentration with prolonged and intense tail or digit clamps as stimuli and use only purposeful, but not reflex, motion to determine the anesthetic requirement in the experimental setting.⁴⁹ The types of stimulation used in such studies could not be performed multiple times during a clinical or experimental procedure, nor would it be reasonable to pause the surgical intervention for a lengthy assessment, thus extending the total anesthesia time. Therefore, to reflect real-world monitoring situations by research or veterinary staff, we chose a brief stimulus and assessed all motion, including reflex motion, in response to the stimulus. This strategy is only possible because we used no neuromuscular blocking agents in our anesthetic regimen. Compression of the medial or lateral claw with forceps was the chosen method of stimulation, in light of previous evaluations of the sensitivity of this anatomic location in producing motor response inversely proportional to the administration of anesthetic.^13,22

BIS previously has been evaluated for its ability to determine anesthetic depth in pigs.^{4,18,23,38,40,41,55,56} Similar to our findings regarding a strong correlation between BIS_mean and Et_ISO, these reports seem to agree that, in general, BIS decreases with increasing anesthetic dose in pigs. There is disparity in the literature regarding whether BIS correlates to clinical scales of sedation depth^40,41 or not.²³ BIS may be useful in differentiating ‘light’ from ‘deep’ anesthesia but is not useful in differentiating moderate from deep sedation.¹⁸ Our focus was detection of motion in response to noxious stimuli, because this method is common practice prior to experimental manipulation that would otherwise cause pain in the laboratory setting. Our finding that BIS was a very good predictor of motion response in anesthetized pigs was in contrast to a previous report,⁴ in which BIS was a poor predictor of motion in response to surgical stimulation in pigs. In that study, the anesthetic combination used was ketamine and azaperone,⁴ and BIS has been shown to be unreliable when ketamine is used,^{15,26,28,60,68} except at a very low dose.⁴⁵ We therefore chose not to include any data obtained within 30 min of ketamine dosing in the analyses. Ketamine is a dissociative agent and increases cortical activity⁵⁷ causing paradoxical increase in BIS despite deeper clinical levels of anesthesia.¹¹ For this reason, BIS monitoring may be inappropriate for animals that are primarily anesthetized with ketamine as the main component of their drug therapy. In addition, this ketamine-associated effect on BIS may explain why a previous study found an average BIS value of greater than 85 in animals that did not respond to noxious stimulation.⁴ Several reports in human patients describe a benefit in using BIS to predict motion in response to stimuli in the context of various anesthetics.^33,59,69 However, it is important to note that the utility of BIS for predicting motion in response to stimuli in sedated or anesthetized animals may be agent-specific, because anesthetic drugs have different effects on the EEG.⁴⁶ In a small pilot experiment (n = 10 pigs, data not shown) not presented here, we did not find BIS to be useful in predicting motion response during a total intravenous anesthesia regimen of fentanyl–midazolam–dexmedetomidine. We concluded that BIS was useful when determining motion response in piglets anesthetized with isoflurane after premedication with ketamine, xylazine, and buprenorphine, and we do not advise the extrapolation of the precise BIS values found here to other anesthetic agents without thorough evaluation. Buprenorphine, used as an analgesic in this study, has been shown to alter EEG parameters that may be reflected as changes in BIS. Intramuscular or epidural administration of buprenorphine in sheep before orthopedic surgery significantly lowered the median power frequency and θ:β ratios on EEG.⁴⁷ Tiletamine–zolazepam (Telazol, Zoetis, Florham Park, NJ) is commonly used as a premedication in swine. Due to their similar pharmacologic actions,³⁵ tiletamine and ketamine may have similar effects on BIS, that is, paradoxical increases in BIS despite deepening levels of sedation. Benzodiazepines, on the other hand, have been shown to induce a decrease in BIS, in line with clinical correlates of sedation.⁵ In a study evaluating EEG parameters in rabbits anesthetized with tiletamine–zolazepam with or without xylazine, the animals that received tiletamine–zolazepam alone showed significant changes in spectral edge frequency and β:δ ratios only in the absence of noxious stimulation;⁶⁷ all combinations using xylazine significantly decreased these parameters. Therefore, the value of BIS for anesthetic monitoring in different species anesthetized with premedication or anesthetics different that those we evaluated needs to be specifically evaluated.

In addition, we noted a significant difference in prestimulation compared with poststimulation values for BIS in positive, but not negative, responders to TP (Figure 3). We speculate that this difference is due either to continued emergence over time from anesthesia, the effect of stimulation on degree of hypnosis, or subtle changes in muscle activity, but our experiments were not powered sufficiently to determine which of these scenarios is most likely. Increases in BIS have been noted in human patients in response to noxious stimuli with no change in anesthetic delivery. One study compared BIS in the context of different intensities of noxious stimuli and saw a significant increase in BIS after the application of thermal stimulation.¹⁰ No utility was found for BIS in differentiating moderate from severe stimuli, according to patients’ reports before surgery; a motion response was not noted because neuromuscular blocking agents were used as part of the anesthetic regimen.¹⁰ Furthermore, noxious stimuli have been associated with changes in other EEG parameters in animals, usually consistent with increased arousal.⁴⁶ Interestingly, a paradoxical synchronization in the EEG can be seen after noxious stimuli.⁴⁶ This effect may explain the increased variability of BIS after TP in our study (Figure 4). In a previous study, EEG parameters such as burst suppression ratio, total power, spectral edge frequency, and median frequency, all of which may play a role in the BIS value, did not significantly change in response to noxious stimuli applied to pigs for 30 s on various body parts at different anesthetic doses.²⁴

On the basis of the predictive value of BIS to identify positive responders to TP in piglets with normal neuromuscular function, the pedal withdrawal reflex may be a good correlate of cortical activity reflective of consciousness in isoflurane-anesthetized pigs despite the fact that this index is not a direct measure of consciousness. One group found that EMG evaluation of reflex withdrawal in response to electrical stimulation more closely approximated the minimal alveolar concentration in isoflurane-anesthetized pigs than previously was reported for humans, although reflexes still remain intact until anesthetic levels greater than the minimal alveolar concentration are administered.⁶⁴ This finding, combined with our data, suggests that reflexes mediated at the level of the spinal cord may be used to approximate the depression of higher neural centers to the level of a surgical plane of isoflurane anesthesia in pigs, in that an absence of a reflex is likely to ensure absence of consciousness when paralytic agents are not used. The exact relationship between the minimal alveolar concentration of isoflurane and the abolition of reflexes, however, is likely to be species-dependent,²¹ and all of the animals used in the current study were piglets; therefore these results should be extrapolated cautiously to other species. It is also important to note that isoflurane may act on subcortical, brainstem, and spinal levels to affect patient motion. Our experiments were not designed to be able to determine the location of the disruption of the signaling pathway but simply to investigate the association of BIS with the presence or absence of motion response to TP. In light of our findings on the lack of obvious short-term changes in HR and noninvasive BP with concurrent positive TP, monitoring only these parameters in paralyzed experimental subjects may lead to a delay in addressing light anesthesia during procedures. Neuromuscular blockade is associated with an increased incidence of patient awareness in some reports in humans.⁵⁸ BIS_mean may therefore aid in the refinement of animal use to ensure the appropriate anesthetic depth in the presence of neuromuscular blockade, in conjunction with currently used methods of assessing increases in hemodynamic parameters. Given that HR and BP were not sensitive in predicting motion in response to TP and that BIS_mean was predictive, BIS should be considered as an adjunct to anesthetic monitoring in cases where experimental animals are given paralytic agents during potentially painful procedures. This practice will help identify animals that have insufficient anesthetic delivery before hemodynamic changes, thus improving animal wellbeing.

The purpose of general anesthesia is to ensure amnesia, analgesia, and hypnosis. Because the ethical consequences of subject awareness during an experimental procedure outweigh the cost of mild decreases in hemodynamic values, the tendency in laboratory animal medicine is to overanesthetize animals. BIS, however, may aid in anesthetic refinement by adding a ‘too deep’ measure, evident by a very low BIS value (for example, less than 40) or a high burst suppression ratio presented on the monitoring interface. Determining excessive depth is impossible when solely monitoring motor response to noxious stimuli, and BIS monitoring can be used in the presence of neuromuscular blockade when motor response cannot. The value in identifying when animals are anesthetized excessively includes stability in cardiovascular and respiratory parameters, as well as savings in the amount of anesthetic agent delivered and decreased time to recovery, all of which improve efficiency and personnel safety. However, DOA, as measured by BIS_mean, may not predict time to recovery after anesthesia. A previous report found that BIS values and recovery times differed between 2 groups of pigs that received different anesthetics after the discontinuation of anesthesia, and—counterintuitively—the group with the higher BIS took longer to recover.³⁸ In that previous report, the authors did not attempt to correlate recovery time within a group of pigs given a single anesthetic regimen. In the current case, we evaluated a group of pigs that received the same anesthetic regimen and found that even within a single anesthetic group, BIS_mean at the time of cessation of isoflurane administration did not correlate with time to extubation.

In these studies, BIS was evaluated as a marker of hypnosis, but true confirmation of awareness is impossible in pigs due to communication barriers. We assumed that BIS is reflective of awareness because of the study of BIS and other EEG parameters in humans^36,70 and the similarities in anatomy, physiology, and clinical response to anesthetic drugs between humans and pigs.³⁶ Although the concern might arise that BIS would be altered by the injury model produced, a study by our group found no significant effect of focal cortical injury on BIS in piglets.³²

Because the TP method relies on the application of force by an operator, that force is likely to be inconsistent. In rodents, modified hemostats^3,8 and von Frey filaments^27,34 have been used to standardize the stimulus in studies evaluating anesthetic regimens. In large animal models, more intense noxious stimulation usually is required and is commonly produced by compressing hemostatic forceps on the digit or interdigital web. Criticisms of this method include a lack of consistency in the magnitude and the location of the stimulus being applied (both between users and between events for a single user), potential tissue damage secondary to the crush injury, and inability to determine excessive anesthetic depth. Specialized devices have been developed to standardize distally applied mechanical force to animals but were not used in the present study because they are not commonly used by investigators at our institution.^25,48 Instead, we used the tissue reaction to compression (blanching) to help standardize the stimulus. Here, the goal was to evaluate a stimulus, already in practice, that is easily adopted by research and veterinary staff in the experimental setting. Although unlikely, the manual application of pressure with hemostatic forceps might have been sufficiently variable to mask or highlight differences between groups. Because a single observer scored the response to TP, observer bias in these data cannot be eliminated. There were, however, no instances when the response to TP was ambiguous, and a clear withdrawal reflex was noted in every case. Most piglets (26 of 33) were examined twice during the study, once during a survival procedure, and once for a terminal procedure. Because these procedures were separated by 6 d, they were treated as separate events. We did not examine the effect of repeated anesthesia, and this perspective should be studied in the future. Furthermore, assessing motion response to TP is not a direct measure of awareness. If this method is to be used routinely, it must be systematically evaluated for its effectiveness as a metric for lack of awareness, or hypnosis, in anesthetized animals. Another potential limitation of our study was the inability to collect data on changes in respiratory rate, tidal volume, and end-tidal carbon dioxide due to the mechanical ventilation of the piglets. In practice, these respiratory values can be obtained readily and potentially evaluated in response to TP for the assessment of DOA in spontaneously breathing pigs.

In summary, changes in HR and BP, as measured in the current study, were not a good measure of DOA when evaluated immediately after application of a TP stimulus. BIS_mean might be a valuable adjunct in determining depth of anesthesia in piglets anesthetized with isoflurane at the tested doses and in the context of premedication with ketamine and xylazine and buprenorphine analgesia. In pigs with a positive response to TP, HR, but not BP, was increased significantly in isoflurane-anesthetized pigs. However, BIS_mean was a much stronger predictor of response in these animals. In addition, BIS_mean was superior to Et_ISO in differentiating piglets with a positive TP response from those ready for endotracheal extubation. Therefore, we support the use of BIS to complement the response to TP when evaluating the depth of anesthesia in isoflurane-anesthetized piglets.