Effects of isoflurane on somatosensory-evoked potentials in calves: A pilot study

Abstract

Somatosensory evoked potentials (SSEP) are used to monitor sensory function and are often recorded under general anesthesia. The objective of the study was to evaluate the effects of isoflurane on SSEPs in calves as it has not been reported. Eight calves (mean age: 40 days), were included in the study. Calves were anesthetized with a randomized sequence of four different isoflurane partial pressures. Blood gas analysis was performed before each measurement. SSEP were induced by repeated stimulation of the common dorsal digital nerve III. SSEPs were recorded from the lumbo-sacral junction (s-SSEP) and the head (c-SSEP). Latency and inter-amplitude of each peak were measured. For s-SSEP: One negative (N_sp1) and two positive (P_sp1 and P_sp2) peaks were identified in all tracings except for two calves. There was a significant effect of isoflurane on the latency of P_sp2 (P = 0.01). Inter-amplitude decreased significantly with PaO₂, PaCO₂ and temperature (P < 0.05). P_sp2 latency decreased with PaO₂ (P = 0.01). For c-SSEP: two positive (P_c1 and P_c2) and two negative (N_c1 and N_c2) peaks were identified. There were identifiable peaks for the analysis of P_c1 latencies only. There was a significant positive linear relation between end-tidal isoflurane partial pressure (ET_iso) and P_c1 latency (P = 0.04). None of the co-variables had a significant effect on the latency of P_c1 (P > 0.1). Isoflurane has a major impact on the recording of c-SSEP. Recording should be done at the lowest ET_iso as possible, and anesthesia parameters should be kept constant.

Introduction

Somatosensory-evoked potentials (SSEPs) are surface electrical potentials recorded from the nervous system after repeated stimulation of a peripheral sensory nerve. Somatosensory-evoked potentials can be recorded from any point between the stimulation site and the cerebral cortices, but are usually recorded on the head or along the spinal cord. They are used to monitor the sensory function of a patient during a procedure that may cause sensory loss (1) or to assess the efficacy of regional anesthesia, such as after epidural injection of local anesthetics (2,3). When the action potential resulting from the peripheral stimulation is stopped or slowed, whether by a physical insult or by a chemical agent, the amplitudes and latencies of SSEPs may be reduced or prolonged, respectively.

In veterinary medicine, SSEP monitoring is often done using general anesthesia, as the patient must stay still during the stimulation to avoid movement artefact. Volatile anesthetics, such as isoflurane, are commonly used in veterinary medicine to maintain anesthesia. The effects of isoflurane on SSEPs has been reported in humans (4) and in rats (5). In those species, isoflurane reduces the amplitude and increases the latency of SSEPs at low partial pressure (1,4,5). Volatile anesthetic agents have an inhibitory effect on transmission of action potentials with the depressing effect being greater on synaptic transmission than on axonal conduction (2,3,6). Sub-cortical SSEPs are therefore more resistant to increasing partial pressures of isoflurane than those of cortical origin (4,7). This study is part of a broader study in which SSEP will be used to monitor nerve blockade under general anesthesia, as reported in ponies (8). To the authors’ knowledge, there are no published reports on the effects of isoflurane on SSEPs in calves.

The objective of this study was to evaluate the effects of isoflurane on SSEPs in calves. Our hypothesis was that isoflurane would increase latencies and decrease amplitudes of SSEPs in calves.

Materials and methods

Holstein calves enrolled in this study were healthy, based on clinical and neurological examinations. This research project was approved by the Ethics Committee for the Use of Animals at the Université de Montréal.

Animal management

Housing consisted of 2 identical stalls in the contagious section of the hospital, with 5 calves housed per stall. They were fed with milk powder [0.15 L/kg body weight (BW)] and had free access to water. The calves were examined by a board-certified large animal medicine specialist and had ultrasound as part of another study on the day after their arrival. They were anesthetized for this study on the morning of the following day. The calves were fed the evening before anesthesia and milk was not withdrawn before anesthesia.

Anesthesia was induced in each calf with isoflurane using a face-mask. All calves were then intubated and maintained with varying partial pressures of isoflurane in oxygen via a rebreathing circuit. The calves were positioned in left lateral recumbency. Two catheters were inserted: a 14-gauge (Angiocath 14G, 3.25 inch; BD, Franklin Lakes, New Jersey, USA) in the right jugular vein for venous fluid access and a 20-gauge (insyte/w 20G, 1.16 inch; BD) in the right auricular artery for recording arterial pressure and analyzing arterial blood gas. The arterial catheter was connected to a pressure transducer (Edward Life Sciences, Irvine, California, USA) via non-compliant tubing (Edward Life Sciences). Monitoring was carried out using a multi-parameter anesthesia monitor (LifeWindow LW6000; Digicare Biomedical Technology, Boynton Beach, Florida, USA), which included pulse oximetry, measurement of end-tidal carbon dioxide (ET_CO2) and isoflurane (ET_iso), direct arterial blood pressure, core temperature (naso-esophageal probe) and peripheral skin temperature (probe placed on the medial aspect of the metatarsus), and the electrocardiogram (ECG). A certified technician calibrated all devices before the experimental period, as specified by the manufacturer. The capnograph and gas analyzer were calibrated using a standardized gas containing 10% CO₂, 50% nitrous oxide, 1.5% isoflurane, and 38.5% azote (Agent End-Tidal Calibration; Smiths Medical, Dublin, Ohio, USA).

Invasive blood pressure was calibrated using a sphygmomanometer. The ECG and the pulse oxymeter were calibrated using a simulator (Fluke, Montreal, Quebec). The ECG was recorded using base-apex leads and monitoring was done in lead II. Core temperature was maintained using forced-air warming blankets (SurgiVet Equator Convective Warming System; Smiths Medical). Mean arterial pressure was maintained above 65 mmHg during anesthesia using a constant rate infusion of dobutamine (Sandoz, Boucherville, Quebec) and/or phenylephrine (Neo-Synephrine; Hospira, Lake Forest, Illinois, USA) as needed. Solutions were diluted to 0.1 mg/mL in 0.9% saline. Blood gas analysis (ABL 80 Flex CO-OX; Radiometer Canada, London, Ontario) included pH, arterial partial pressure of carbon dioxide (P_aCO₂) and of oxygen (P_aO₂) and was done before each SSEP measurement. Each calf was mechanically ventilated to maintain both the P_aCO₂ and P_aO₂ within reference ranges (35 to 50 mmHg for P_aCO₂ and above 85 mmHg for P_aO₂).

The experimental protocol is summarized in Figure 1. In order to avoid any confounding effect from the duration of anesthesia, 8 different sequences of 4 alveolar partial pressures were randomly generated for the experiment. Briefly, all combinations of 4 alveolar partial pressures were generated and randomly assigned to each calf using the “random” function of Excel (Microsoft, Redmond, Washington, USA). To allow rapid changes in ET_iso, high flow rate of oxygen (4 L/min) was used when changing ET_iso. As soon as the targeted ET_iso was reached, oxygen flow rate was reduced to decrease unnecessary pollution and oxygen and isoflurane utilisation; vaporizer setting was adjusted to keep the ET_iso at the desired level. Measurements of SSEP were done 15 min after the targeted ET_iso was obtained. The sequences were allocated to each calf randomly.

Figure 1.

Experimentation protocol. SSEP — somatosensory-evoked potential. ET_iso — end-tidal isoflurane partial pressure. ET_iso 1 to 4 refers to the chronological sequence of the different ET_iso but not the real ET_iso tracings, which were randomized. During equilibrium, ET_iso was set at 1.3%.

For the first 2 calves, the ET_iso was 1.0%, 1.3%, 1.6%, and 1.9%. As no SSEPs were observed for 1 calf at any ET_iso, it was decided to measure SSEP at lower ET_iso. For the other 6 calves, the ET_iso tracings were 0.7%, 1.0%, 1.3%, and 1.6%. Calves were euthanized at the end of the study protocol for another unrelated study by administering 110 mg/kg BW pentobarbital (Euthanyl Forte; Bimeda-MTC Animal Health, Cambridge, Ontario) intravenously.

SSEP recording

Somatosensory-evoked potentials (SSEPs) were evoked by repeated stimulation of a peripheral nerve, the common dorsal digital nerve III. Stimulation was applied to the dorso-lateral aspect of the right metatarsus, approximately 5 cm above the metatarsophalangeal joint, overlying the common dorsal digital nerve III. The nerve was stimulated percutaneously, using 2 subcutaneous needles (Neuroline Monopolar 25 × 0.36 mm; Ambu, Ballerup, Denmark) separated by 2 cm with the anodal electrode located distally. The stimulations had a square waveform, 0.1 ms duration, applied at 2.7 stimulations per second, as described by Strain et al (9). Intensity was fixed at 6 mA based on a pilot study in which a satisfactory spinal signal was obtained without causing purposeful movement at low partial pressure of anesthesia.

Somatosensory-evoked potentials were recorded using a computer- based electrodiagnostic system (Sierra II Wedge; Cadwell Laboratories, Kennewick, Washington, USA). This is a modification of the technique described in previous studies using adult cows (5,9). The digitization rate was 76.8 kHz. Based on a pilot study, artefact rejection function was not used as the absence of SSEP induced by isoflurane causes all signals to be rejected. The SSEPs recorded at the lumbosacral junction were called spinal SSEP (s-SSEP) and those recorded at the cerebral cortices were called cortical SSEP (c-SSEP).

The same electrodes were used for all recordings and grounding (Neuroline Subdermal 12 × 0.4 mm; Ambu). For the s-SSEP recording, electrodes were placed as follows: the negative recording electrode at the lumbo-sacral junction in the ligament between L6 and S1 and the positive electrode at the right ilium. Electrodes for the c-SSEP recording were placed as described by Strain et al (9): the negative electrode at the vertex on the upper surface of the head and the positive electrode at the midpoint of the inter-orbital line. A grounding electrode was placed on the left ilium. Recording band pass was 1 to 1000 Hz. An average of 1000 stimulations was collected for each recording.

The duration of the stimulation was about 6 min and 10 s. Before all recordings, the impedance was measured at each electrode and placement of the electrodes was adjusted accordingly to reduce the impedance. Impedance was less than 5 kilo-ohms. The person (GT) measuring the SSEP was blinded to the level of anesthesia used during the measurement. He could not see the ET_iso as the monitor did not face him and a blanket covered the calf once the electrodes were in place so he could not see the level of anesthesia.

Consecutive positive and negative peaks were identified for spinal and cortical SSEP. The peaks were identified independently in a randomized manner by 3 of the authors (GT, PB, and JP), all of whom were blinded to the ET_iso. Randomization was done by allocating a random number to each tracing using a random number table and then sorting them by ascending order. Blinding was done by removing any identification from the tracings.

Measurements consisted of latencies and peak-to-peak amplitudes. Latency was defined as the time from stimulus to peak for each deflection. Peak-to-peak amplitude was defined as the difference in amplitude between each successive peak (Figure 2). Latencies were corrected for the length of the animal, as described in a previous study (9). The distances between the stimulation site and the recording sites were measured using the following anatomical landmarks: positive needle, tarsus, femoro-patellar ligament, ilium, and the recording electrode at the lumbosacral junction and at the occipital crest. Distances were measured 3 times by 1 person (GT) and the mean of the 3 measurements was used as the measured distance. The following equation was then used:

Figure 2.

Illustrated tracings showing latencies and peak-to-peak amplitudes of spinal and cortical recordings. A — Cortical somatosensoryevoked potentials (c-SSEPs). P_c1, N_c1, P_c2, and N_c2 identify the different peaks. 1 to 4: c-SSEP latency of P_c1, N_c1, P_c2, and N_c2 respectively. 5 to 7: P_c1 to N_c1, N_c1 to P_c2, P_c2 to N_c2 peak-to-peak amplitudes respectively. B — Spinal somatosensory-evoked potentials (s-SSEPs). P_sp1, N_sp1, and P_sp2 identify the different peaks. 1 to 3: s-SSEP latency of P_sp1, N_sp1, and P_sp2 respectively. 4 to 5: P_sp1 to N_sp1 and N_sp1 to P_sp2 peak-to-peak amplitudes.

$$\text{Corrected\hspace{0.17em}latency}=\text{measured\hspace{0.17em}latency}\times \text{mean\hspace{0.17em}distance}/\text{measured\hspace{0.17em}distance}$$

where: the mean distance is the mean of the distances of all calves. This allows the latencies of individuals of different sizes to be compared (9). The SSEPs were recorded so that negative deflections were displayed upward.

Data analysis

Reproducibility of the results of the 3 authors was analyzed by the intra-class correlation coefficient obtained from a random-effects model relating the variation among observers to the variation within observer. A linear model for repeated measures was used to analyze the effect of isoflurane on latencies and amplitudes of s-SSEP and c-SSEP. Goodness of the fit of the model was not determined as it is not straightforward to determine for a linear mixed model like the one we used (repeated measurements at the level of the individual). The fit statistics associated with the model are only useful when comparing different models (a question of relative fit instead of absolute fit). This was not the goal of our statistical procedure since we wanted to evaluate the full model with all cofactors. Co-variables were P_aCO₂, P_aO₂, core and surface temperatures, and mean arterial pressure. All variables were analyzed as quantitative variables. Statistical significance was attributed when P < 0.05. Statistical analysis was carried out using the statistical software SAS, Version 9.3 (SAS Institute, Cary, North Carolina, USA).

Results

Eight Holstein calves were used in this study. Gender was not recorded as the model used was for repeated measure at the individual level and the effect of gender on SSEP would therefore not alter our conclusions. Mean age of the calves was 40 d (27 to 48 d). Mean estimated weight was 96 kg (70 to 110 kg). Anesthesia was uneventful except for 1 calf that was hypoxemic (P_aO₂ between 53 and 70 mmHg) during the procedure. Pneumonia was suspected based on blood gas results and the finding of purulent material in the tracheal tube at extubation. Cortical SSEPs could not be measured on this calf even at the lowest isoflurane partial pressure, but s-SSEPs were measured. This calf was excluded from the analysis. All calves received dobutamine (1 to 2 μg/kg BW per minute) and phenylephrine (0.5 to 1 μg/kg BW per minute). Representative SSEP recordings, both spinal and cortical, are shown in Figure 3. Results of blood gases and physiological data at each ET_iso are summarized in Table I.

Figure 3.

Typical tracings from different individuals observed during the study. A to D — cortical somatosensory-evoked potentials (c-SSEPs). E to F — Spinal somatosensory-evoked potentials (s-SSEPs). A — Typical c-SSEP with 4 well defined peaks: P_c1, N_c1, P_c2, and N_c2. B — c-SSEP in which the different peaks were difficult to identify. C — c-SSEP in which only P_c1 could be identified. D — c-SSEP in which no peaks could be identified. E — A typical s-SSEP with its well defined 3 peaks: P_sp1, N_sp1, and P_sp2. F — s-SSEP in which movement artefacts were observed after the s-SSEP was recorded.

Table I.

Results of blood gas analyses and values of mean arterial pressure at different end-tidal isoflurane partial pressure. Values are presented as median (min; max)

ETiso	Tcore	T surface	MAP	PaCO2	PaO2
0.7	38.0 [37.1; 38.4]	37.9 [32.0; 39.5]	104.0 [87.0; 136.0]	34.0 [33.0; 37.0]	478.0 [97.0; 526.0]
1.0	37.9 [36.7; 38.4]	36.0 [34.2; 38.3]	87.0 [73.0; 139.0]	37.0 [29.0; 49.4]	418.0 [90.0; 553.0]
1.3	37.9 [36.6; 38.5]	37.0 [33.2; 38.4]	88.0 [81.0; 105.0]	31.0 [27.0; 53.1]	456.0 [63.0; 510.0]
1.6	37.9 [36.9; 38.4]	36.1 [33.0; 37.6]	86.0 [70.0; 114.0]	34.0 [27.0; 54.0]	411.0 [137.0; 513.0]
1.9	38.0 [37.6; 38.4]	36.4 [35.6; 37.1]	91.5 [78.0; 105.0]	51.5 [51.3; 51.6]	331.5 [212.0; 451.0]

ET_iso — end-tidal isoflurane partial pressure in percentage; MAP — mean arterial pressure in millimeters of mercury (mmHg);

P_aCO₂ — arterial partial pressure of carbon dioxide in mmHg; P_aO₂ — arterial partial pressure of oxygen in mmHg.

S-SSEP

Reproducibility coefficient was above 0.8 in all but 1 reading of s-SSEP. As the results of all authors were similar, the results shown and analyzed are those of the first author. Three peaks, 2 positives (P_sp1 and P_sp2) and 1 negative (N_sp1), were identified in all tracings except for those from 2 calves. No s-SSEP could be recorded at any ET_iso (1.0%, 1.3%, 1.6%, and 1.9%) for 1 of these calves. The recording electrodes fell out of the other calf during the recording as the calf moved at ET_iso = 0.7%. The test could not be redone due to time constraints. The recordings for this calf were obtained for all other ET_iso (1.0% to 1.6%). Movement artefact after stimulations was observed in most calves at a light plane of anesthesia. The number of tracings with movement-induced peaks is summarized in Table II. Typical tracings are shown in Figure 3. Values of latencies and peak-to-peak amplitudes are summarized in Table III.

Table II.

Number of s-SSEP tracings showing movement artifact at different end-tidal isoflurane partial pressures (ET_iso)

	End-tidal partial pressure of isoflurane (%)
	0.7	1.0	1.3	1.6	1.9
Numbers of movement artifacts	4/5	3/7	1/7	1/7	0/2

Table III.

Amplitudes and latencies of cortical (c-SSEPs) and spinal somatosensory-evoked potentials (s-SSEPs) at different end-tidal isoflurane partial pressures (ET_iso). Values are presented as median (min; max)

	End-tidal concentration of isoflurane
	0.7 (n = 6)	1.0 (n = 8)	1.3 (n = 8)	1.6 (n = 8)	1.9 (n = 2)
Pc1	39.85 [37.90; 41.80] (n = 4)	43.95 [40.20; 64.80] (n = 6)	50.40 [41.80; 86.30] (n = 5)	50.40 [44.10; 59.40] (n = 3)	50.00 (n = 1)
Nc1	50.75 [46.90; 52.70] (n = 4)	62.90 (n = 1)	72.25 [70.30; 74.20] (n = 2)	70.50 [66.80; 74.20] (n = 2)	68.80 (n = 1)
Pc2	63.10 [59.00; 66.40] (n = 4)	82.80 (n = 1)	87.10 [85.50; 88.70] (n = 2)	83.00 [75.00; 91.00] (n = 2)	77.30 (n = 1)
Nc2	87.30 [79.70; 119.50] (n = 4)	122.70 (n = 1)	116.60 [106.60; 126.60] (n = 2)	125.00 [121.50; 128.50] (n = 2)	158.60 (n = 1)
Pc1–Nc1	0.68 [0.35; 1.27] (n = 4)	1.09 (n = 1)	0.73 [0.30; 1.15] (n = 2)	1.01 [0.44; 1.57] (n = 2)	0.90 (n = 1)
Nc1–Pc2	0.49 [0.10; 1.97] (n = 4)	0.62 (n = 1)	0.20 [0.06; 0.34] (n = 2)	0.33 [0.20; 0.45] (n = 2)	0.08 (n = 1)
Pc2–Nc2	1.89 [1.64; 3.73] (n = 4)	1.95 (n = 1)	1.08 [0.71; 1.44] (n = 2)	1.89 [1.46; 2.32] (n = 2)	1.61 (n = 1)
Psp1	13.29 [10.17; 15.18] (n = 4)	12.24 [10.17; 14.34] (n = 6)	12.73 [10.54; 14.34] (n = 6)	13.04 [10.54; 14.34] (n = 6)	12.72 (n = 1)
Nsp1	16.54 [13.81; 18.42] (n = 4)	16.75 [13.81; 18.84] (n = 6)	16.59 [13.81; 19.67] (n = 6)	17.05 [13.81; 19.67] (n = 6)	16.35 (n = 1)
Psp2	22.69 [20.72; 24.49] (n = 4)	27.48 [22.21; 31.75] (n = 6)	32.05 [22.21; 34.75] (n = 6)	31.37 [26.22; 35.07] (n = 6)	27.58 (n = 1)
Psp1–Nsp1	1.81 [1.33; 2.83] (n = 4)	1.87 [0.84; 2.48] (n = 6)	1.62 [0.88; 2.69] (n = 6)	1.87 [0.97; 2.43] (n = 6)	1.33 (n = 1)
Nsp1–Psp2	2.72 [2.09; 3.01] (n = 4)	2.53 [1.08; 3.47] (n = 6)	2.21 [1.22; 3.79] (n = 6)	2.63 [1.15; 3.52] (n = 6)	1.83 (n = 1)

Isoflurane had a significant effect on latency of P_sp2 [P = 0.01, slope (β) = 6.57 (SEM = 2.15)] but not on latency of P_sp1 and N_sp1 [P = 0.45, β = 0.25 (0.33) and P = 0.20, β = 0.34 (0.25), respectively]. There was no significant effect of isoflurane on any of the peak-to-peak amplitudes [P_sp1 to N_sp1: P = 0.66, β = 0.09 (0.20); N_sp1 to P_sp2: P = 0.51, β = 0.20 (0.30)].

Temperature had no effect on P_sp1 latency [P = 0.10, β = −0.20 (0.11)], N_sp1 latency [P = 0.38, β = −0.08 (0.09)] or P_sp2 latency [P = 0.50, β = 0.26 (0.38)]. However, P_sp1 to N_sp1 peak-to-peak amplitude and N_sp1 to P_sp2 peak-to-peak amplitude decreased with temperature [P = 0.02, β = −0.09 (0.03) and P = 0.01, β = −0.17 (0.05) respectively].

Mean arterial pressure (MAP) had no effect on P_sp1 to N_sp1 peak-to-peak amplitude [P = 0.32, β = −0.005 (0.0045)], N_sp1 to P_sp2 peak-to-peak amplitude [P = 0.51, β = −0.005 (0.007)], P_sp1 latency [P = 0.23, β = 0.012 (0.01)], N_sp1 latency [P = 0.79, β = −0.002 (0.008)], or P_sp2 latency [P = 0.47, β = −0.04 (0.049)].

Arterial partial pressure of carbon dioxide (P_aCO₂) had no effect on P_sp1 latency [P = 0.69, β = 0.02 (0.06)], N_sp1 latency [P = 0.21, β = −0.07 (0.05)], or P_sp2 latency [P = 0.98, β = 0.004 (0.14)]. However, P_sp1 to N_sp1 peak-to-peak amplitude and N_sp1 to P_sp2 peak-to-peak amplitude decreased with P_aCO₂ [P = 0.003, β = −0.06 (0.01) and P = 0.003, β = −0.08 (0.02) respectively].

Arterial partial pressure of oxygen (P_aO₂) had no effect on P_sp1 latency [P = 0.29, β = −0.002 (0.002)] or N_sp1 latency [P = 0.48, β = −0.001 (0.001)]. However, P_sp1 to N_sp1 peak-to-peak amplitude and N_sp1 to P_sp2 peak-to-peak amplitude decreased significantly with P_aO₂ [P = < 0.0001, β = −0.0046 (0.0006)] and P = < 0.0001, β = −0.006 (0.001) respectively]. Additionally, P_sp2 latency increased with P_aO₂ [P = 0.01, β = 0.019 (0.007)].

C-SSEP

Reproducibility coefficient was below 0.5 in 9 out of 16 readings for latency and below 0.7 in 5 out of 16 readings for amplitudes of c-SSEPs. Researchers met to review the readings and reach a consensus for each reading. For each peak, a consensus was considered reached when all 3 researchers agreed that either the peak was identifiable beyond any doubt or the peak could not be reasonably identified. Researchers were blinded to the ET_iso. Only the group consensus results were analyzed and are presented for c-SSEP.

Four peaks, 2 positives (P_c1 and P_c2) and 2 negatives (N_c1 and N_c2), were identified at ET_iso of 0.7% in 4 calves. Peaks were not observed for each calf at every ET_iso, as shown in Figure 3. Latencies and peak-to-peak amplitudes at each ET_iso are summarized in Table III.

There were enough identifiable peaks to analyze P_c1 latencies only. There was a significant positive linear relation between ET_iso and P_c1 latency [P = 0.04, β = 7.62 (3.24)]. None of the co-variables had a significant effect on the latency of P_c1: temperature [P = 0.63, β = −0.51 (1.02)], MAP [P = 0.78, β = 0.02 (0.08)], P_aCO₂ [P = 0.13, β = −0.69 (0.42)], and P_aO₂ [P = 0.67, β = 0.006 (0.01)].

Discussion

The objective of the study was to evaluate the effects of isoflurane on cortical- and spinal-SSEPs in calves. Increasing expired isoflurane partial pressures has a major effect on cortical activity and effects only P_sp2 for s-SSEP. To the best of our knowledge, the effects of isoflurane, or any other volatile anesthetics, on SSEPs have not previously been reported in cows. The effects of isoflurane on SSEPs have been studied in humans (10,11) and in rats (5,12), however, with similar results.

The c-SSEP tracings obtained in this study were different from those obtained in conscious adult cows by Strain et al (9). They found that there were 2 sequential biphasic waveforms, although the first deflection was negative rather than positive as with our results (9). Recording electrodes in our study and in this previous study (9) were positioned similarly. There are 4 main differences between this previous study (9) and ours. First, calves were used in the present study and not adult cows. In humans, the characteristics of SSEPs in children gradually changed until adolescence when they developed an adult pattern (13). It is therefore possible that SSEPs of calves look different than SSEPs of adult cows. Second, tracings were obtained under anesthesia, which in itself could modify SSEPs. Reversal of the first deflection has not been reported as an effect of isoflurane or any other volatile anesthetics. As anesthesia is one of the variables between the study of Strain et al (9) and this study, isoflurane anesthesia could be the cause of the apparent reversal of the peaks. Third, we did not use the same stimulus intensities as used in the previous study (9) for the peripheral nerve stimulation. Strain et al (9) reported using an average stimulus of 24 mA (12 to 31 mA) to evoke the SSEPs from the hind limb, using the same nerve as in our study. The optimal stimulation for each animal was defined as the amplitude just below the one triggering movement. An intensity of 20 mA is recommended for human SSEPs, although some authors use current levels as high as 100 mA. A higher intensity of stimulation is painful, however, if the patient is not anesthetized (1). We used an intensity of 6 mA, based on a pilot study in 2 calves in which this intensity was found adequate to trigger both c-SSEP and s-SSEP signals without causing movement at the lowest partial pressures of anesthesia. It is possible that different stimulus intensities stimulate different fibers as they might have different thresholds (14), resulting in different c-SSEP. Fourth, we used a different recording band pass than Strain et al (9). According to Møller (1), changing the upper limit of the band pass filter can cause distortion of the waveform and may cause the peak to appear inverted. We chose a band pass of 1 to 1000 Hz based on the pilot study in which this band pass was found adequate to record SSEP.

Amplitudes of c-SSEP have been reported to decrease, while c-SSEP latencies increase, after isoflurane is administered in humans (4). In our study, the identification of c-SSEP was impeded by isoflurane, as indicated by the low coefficient of reproducibility. Only at lower partial pressures of isoflurane was the quality of the signal adequate and peaks easily identifiable. In our study, P_c1 appears to be less suppressed by isoflurane than the other peaks, as P_c1 was identified more often than the other peaks at ET_iso > 1.0%. In a study in humans (15), the late cortical waveform components were more sensitive to volatile anesthetics, with a marked attenuation at partial pressures above 0.5 times the minimum alveolar concentration (MAC). Based on the published MAC of 1.14 in adult cows (16), we used from 0.6 to 1.7 times the MAC value. In the study of Cantalapiedra et al (16), however, the MAC was determined in adult cows. In humans, infants have higher MAC value than adults (17). Therefore, it is possible that calves also have higher MAC value than adult cows. We chose not to determine MAC value in this study due to time constraints and lack of clinical relevance. Additionally, variation of MAC value among calves of different ages was addressed by using a mixed linear regression model for repeated measurement at the level of the individual. In humans, earlier cortical waveforms, thought to be of subcortical origin, are less sensitive to volatile anesthetics (15). Based on the latency of the peaks observed in our study, and their response to isoflurane, we hypothesized that P_c1 has a subcortical origin, whereas the other peaks have a cortical origin. Morphologic changes have also been described in humans where during deep isoflurane anesthesia, the c-SSEP transforms into a simple monophasic wave (15,18). In our study, at ET_iso of 0.7%, 4 peaks could be observed in most calves, and at ET_iso above 1.0%, only 1 peak was observed in most calves. Isoflurane supressed peaks of suspected cortical origin (N_c1, P_c2, N_c2) at ET_iso above 1.0%, whereas peaks of suspected sub-cortical origin were recordable events at higher ET_iso.

The description of s-SSEPs has not been reported in cattle. In rats (5,19), isoflurane increased the latency and decreased the amplitude of the P_sp2 peak, but had no effect on either P_sp1 or N_sp1, as was the case in our study. In ponies, s-SSEPs were recorded under isoflurane anesthesia with a morphology that was similar to that obtained in our study (8).

As s-SSEPs have not been previously reported in calves, the origin of each peak is not known. The spinal recordings were done at the lumbosacral junction. The morphology of the s-SSEP recordings in humans (14) and in rats (5,19) was similar to what was observed in this study. In those species, it is hypothesized that the first positive peak (P_sp1) is the afferent volley approaching the epidural electrode. The afferent volley is then transmitted to the second-order neurons or inter-neurons to create the first negative peak (N_sp1). Finally, the afferent volley is relayed to the inhibitory interneurons, which transmit these volleys to the afferent terminals to produce the slow positive wave (P_sp2) (14). The generation of the second positive wave, however, seems to involve supra-spinal mechanisms, which could be modified by isoflurane. This mechanism could explain why the latency of P_sp2 was increased by the administration of isoflurane in our study.

In calves that moved the stimulated limb after each stimulus, additional electrical peaks could be recorded at the lumbosacral junction after the normal triphasic tracings of the s-SSEP, but before the cortical activity. The electrical activity was likely the result of motor neuron activity. As the motor activity preceded the recorded cortical activity, we hypothesized that the movement was a local reflex, although an electromyographic artefact cannot be ruled out. The motor activity was not observed at higher ET_iso. As isoflurane and other volatile anesthetic agents have a more profound effect on the ventral horn than the dorsal horn (12), we hypothesized that, while the efferent response of this reflex arc was inhibited at the higher partial pressure of expired isoflurane, the s-SSEP, as measured at the lumbosacral junction, remained.

In our study, reduction of P_aO₂ decreased the s-SSEP peak-to-peak amplitude significantly and increased P_sp2 latency, but changes of P_aO₂ did not significantly affect the other peak-to-peak amplitudes or the latencies. Furthermore, 1 of the calves was hypoxemic and no c-SSEP could be measured on this calf. Since s-SSEP was recorded and the impedance of cortical electrodes was within recommended limits of the device, the lack of c-SSEP was not due to a failure to stimulate. Human c-SSEPs were not affected by breathing gas mixtures of 17.2%, 14.2%, and 11.1% oxygen (20). While no blood gas analyses were done in this earlier study, the authors stated that, based on the alveolar gas equation, this should correspond to a P_aO₂ of 80, 60, and 40 mmHg, respectively. In dogs, in more severe hypoxic condition (gas mixture of 4.5% of oxygen, P_aO₂ = 14 ± 1 mmHg), cortical peak-to-peak amplitude decreased to 17% of control and latencies increased to 111% and 107% of control (21). In humans, similar results were observed in severe hypoxic condition (gas mixture of 6.5% of oxygen and below) (20), although no blood gas measurements were done during the study to match the alteration of amplitudes and latencies with P_aO₂. Cortical SSEPs also seem to be more sensitive to hypoxia than spinal and subcortical SSEPs, presumably because the latter are more tolerant to hypoxia than those from the cerebral cortex (23). In our study, both general and neurological examinations were normal before anesthesia for the hypoxemic calf. Based on the absence of neurological signs, the presence of s-SSEP, and the hypoxemia, we hypothesized that failure to obtain c-SSEP was due to hypoxemia. In our study, the other calves had a P_aO₂ above 63 mmHg, which was presumably not low enough to cause any modification of SSEP, based on studies in humans (20,24) and in dogs (25). Further study is needed on the effect on P_aO₂ on the peak-to-peak amplitude and latency of s-SSEP in calves.

Peak-to-peak amplitude of s-SSEPs was also reduced with decreased temperature and P_aCO₂ in our study, but neither of those parameters had any effects on latencies. Mean arterial pressure (MAP) had no effect on s-SSEPs. In humans (20,24) and in dogs (25), P_aCO₂ from 20 to 50 mmHg had minimal effects on latencies and amplitude. Below 20 mmHg, a small reduction (2% to 4%) of latency and an increase of amplitude of spinal SSEP recorded from the scalp of conscious patients was observed (20). In contrast, peak-to-peak amplitude of s-SSEP was reduced with decreased P_aCO₂ in our study. Further study is needed on the effect of P_aCO₂ on the peak-to-peak amplitude and latency of s-SSEPs in calves.

The relationship between temperature and SSEP is complex and varies between cooling and warming (24). Amplitude can either decrease or increase depending on the origin of the signal (26). It has been reported that MAP decreases s-SSEP and c-SSEP amplitude, but only when below 50 mmHg (23,24). In our study, MAP was kept above 65 mmHg for the duration of the monitoring and accordingly, did not significantly affect SSEP. Therefore, even if the effects of physiological parameters on latency and peak-to-peak amplitude were mild in this study, they should be kept within physiological range and any changes in P_aO₂, P_aCO₂, and temperature should be avoided when measuring SSEPs, especially if it is necessary to compare 2 readings.

In many calves, peaks were difficult to identify for c-SSEP especially at higher ET_iso. Reproducibility among observers was low for the identification of peaks in c-SSEP tracings, which led to a review of the tracings to reach a consensus to identify each of the peaks. During this review, at ET_iso > 1.0%, less than 50% of peaks other than P_c1 could be identified with a high degree of certainty. At higher ET_iso, latter peaks began to disappear and their amplitude was similar to background noise, thus rendering their identification difficult. Only peaks for which a consensus was reached were included in the analysis. Artefact rejection function was not used in this study based on a pilot study, as the absence of SSEP induced by isoflurane causes the signal to be rejected. Although no duplicates of the measurements were taken due to time constraints placed upon this study, using duplicates might have helped to identify smaller peaks. Despite the absence of duplicates and artefact rejection function, the presence of consensus to identify peaks reduced the impact of background noise on the c-SSEP identification.

Spinal somatosensory-evoked potentials (s-SSEPs) were always recorded except in 1 calf, for which no cortical SSEPs were recorded either. Since impedance was within recommended limits of the device and no cortical or s-SSEPs were recorded, we hypothesized that this failure was due to improper stimulation of the nerve. We tried to adjust the stimulation for this calf at the beginning of the experiment, but due to time constraints we decided to proceed with what we thought was the proper place for the electrode. The presence of s-SSEP was useful in order to differentiate an absence of c-SSEP due to a failure to stimulate the peripheral nerve from the effect of isoflurane.

At higher concentration of isoflurane, MAP was kept above 65 mmHg with low doses of dobutamine or phenylephrine. As these drugs can cross the blood brain barrier, their effect on SSEP cannot be excluded. No study reports their effect on SSEP, however, and vasopressors and inotropes are used in human medicine when assessing SSEP. We chose to treat MAP to keep it above 65 mmHg as it has been reported that hypotension affects SSEP (23,24).

Based on the results of our study, isoflurane cannot be considered an ideal agent if monitoring c-SSEP is required. Sevoflurane and desflurane have been used to measure cortical and sub-cortical SSEP in humans at 1 MAC (4,27). Their effects are similar to isoflurane, but may permit their use at higher multiples of MAC. Injectable agents could also be used to reduce inhalant partial pressure as is done in human medicine (4). No report has been published on the effects of injectable drugs on SSEP in cattle. The effects of injectable agents on SSEP in humans have been reviewed elsewhere (4). Among injectable agents, dexmedetomidine, opioids, and benzodiazepines could be used in bovines to reduce the isoflurane partial pressure in order to obtain better SSEP tracings.

In conclusion, isoflurane was shown to have a major impact on the recording of c-SSEP in response to a standardized stimulus. It increased the latency of P_c1 and prevented the reliable identification of peaks at all tracings of ET_iso. Isoflurane can be used to record c-SSEP, but should be kept to the lowest partial pressure possible. It also increased the latency of P_sp2 in s-SSEP, but did not affect the other latencies or the peak-to-peak amplitudes. Spinal SSEPs were useful to identify successful stimulation of peripheral nerves. Recording of SSEPs in calves should be done at the lowest ET_iso possible and this should be kept constant throughout the monitoring period in order to interpret any changes in the tracings. Furthermore, when measuring SSEP, physiological parameters should be kept within physiological range and any changes in P_aCO₂, P_aO₂, and temperature should be avoided.