Nitrous Oxide Depresses Spinal F Waves in Rats

With approval of the University of California, San Francisco Committee on Animal Research, we performed two studies of young adult (approximately 3-month-old) Sprague-Dawley rats. The goal of the first study was to test the interaction of nitrous oxide with isoflurane in clinically relevant concentrations. The second study sought to test the effect of nitrous oxide when administered at concentrations high enough to provide complete surgical anesthesia (i.e., greater or equal to 1.0 MAC). All animals were allowed food and water ad libitum until the day of study.

Animal Preparation. Anesthesia was induced by inhalation of isoflurane in air, and tracheas were intubated with a 2" (5.1-cm) 16 G intravenous catheter. Anesthesia was maintained with isoflurane in oxygen and [nearly equal] 2% carbon dioxide. Mechanical ventilation (Model 804, EDCO Scientific, Chapel Hill, NC) with peak inspiratory airway pressures less or equal to 20 mmHg was adjusted to achieve normocapnia. Airway gas concentrations were monitored continuously by an infrared analyzer (CapnoMac Ultima, Datex Instrumentarium, Helsinki, Finland) using an expiratory limb dead space sampling point previously described. [1]Normothermia was maintained thermostatically through use of a heat lamp and water blanket.

Anesthesia. The end-tidal isoflurane concentration was maintained at 1.0% for 30 m before initial F-wave recording (see below). Next, 70% N²O was added to the 1.0% isoflurane-oxygen mixture, and 20 m allowed for equilibration before F-wave recording. F-wave recordings were made at 15 and 30 m after discontinuing nitrous oxide administration. Finally, 30% N²O was added to the isoflurane-oxygen mixture and allowed to equilibrate for 20 m, with a final F-wave recording made. After each F-wave measurement, the rat's motor response to 60 s of tail clamp was noted. After completing the protocol, the rats were killed with an isoflurane overdose, followed by bilateral thoracotomy.

Electromyographic (F-Wave) Recording. Late evoked electromyographic waves known as F waves may be used as an indirect indicator of spinal motor neuron excitability. [8,9,14-20]The right tibial nerve was stimulated at the popliteal fossa via subdermal electrodes driven by an isolated Grass S-88 constant voltage stimulator as described previously. [8,9]The stimulus width was fixed at 500 miscro s; intensity was varied through use of a prospectively randomized sequence of 10, 20, 30, 40, and 50 V. Evoked electromyographic activity was recorded from subdermal electrodes placed in the intrinsic muscles of the right foot, and referenced to a ground needle electrode in the right heel. The electromyograph was amplified, band-pass filtered (10 Hz-10 kHz), and digitized with 12-bit resolution at a rate of 10 kHz. The digitized electromyographic waveform data were displayed, analyzed, and stored on a Macintosh IIfx (Apple Computer, Cupertino, CA). Stimulator trigger timing, data acquisition, and analysis were automated and controlled by a LabView 3.0 program (National Instruments, Austin, TX) written for this study by one of the authors (IJR).

The M wave was quantified as the peak-to-peak voltage difference in the recorded electromyographic waveform between 0.9 and 3.0 msec after the stimulus. The F wave was quantified as the peak-to-peak voltage difference in the recorded electromyographic waveform between 5.5 and 9.0 msec after the stimulus. All recordings were manually overread to exclude artifact. The wave amplitudes were measured in 10 trials at each voltage, and the 20 M- and F-wave amplitude values that resulted from supramaximal stimulation (40 and 50 V) were averaged for further analysis. Waveforms that resulted from submaximal stimulation (10-30 V) were used to confirm the absence of H-reflex activity. [9,19].

Mean F-wave over M-wave amplitude ratios were compared through use of analysis of variance with repeated measures (StatView 4.1, Abacus Concepts, Berkeley, CA), with Newman-Keuls post-hoc testing.**** This F/M ratio was used to compensate for differences in the geometry of the recording electrodes (with respect to innervated muscle) between rats. A value of P less or equal to 0.05 was considered significant.

Anesthesia was induced by inhalation and maintained with isoflurane while transcutaneous stimulating and recording electrodes were placed as in study 1 and secured to the hind limb with tape and a splint. Six additional transcutaneous electrodes (three stimulation-ground pairs) were secured at 1-cm intervals in the proximal tail. [21]The animal was placed in an experimental hyperbaric chamber described previously [22]into which hyperbaric mixtures of nitrous oxide and oxygen (29, 46, 65, and 79% nitrous oxide, with a balance of oxygen maintaining a constant chamber pressure of 3.4 atm; Table 1) were delivered in random sequence. While still anesthetized with isoflurane, an initial low-voltage stimulation trial verified the absence of H-reflex activity. [9,19]Once pressurized, the chamber fresh gas flow rate was set to approximately 10 gas volume exchanges per hour. Before recording at the first mixture, 30 m were allowed for isoflurane washout and nitrous oxide equilibration, and 20 m for equilibration (with spontaneous ventilation) at each subsequent nitrous oxide concentration. Recurrent, evoked spinal motor neuron activity was recorded at each concentration as described for study 1, except that only 40- and 50-V stimuli were applied. After recording electromyographic data at the four hyperbaric mixtures, chamber pressure was reduced slowly to 2.0 atm. Recordings were repeated after equilibration at nitrous oxide partial pressures of 1.0 and 1.3 atm. The purpose of these lower pressure measurements was to evaluate the effect of oxygen as an anesthetic at high partial pressures. [23].

After F-wave recording at each concentration, a noxious electrical stimulus (15 V at 50 Hz) was delivered for as long as 1 m to the tail via the most distal of the three pairs of tail-stimulating electrodes. [21,22]This stimulus and electrode configuration is equivalent to tail clamp for testing anesthetic potency. [21]The animal was observed for purposeful movement of the extremities or head for 1 min after the start of stimulation. If no movement was observed, the second pair was used; if still no movement, the most proximal pair was used. This reduced the possibility of false negative responses of rat movement in the event that tail nerves were desensitized or damaged during prior stimulation.

Thirteen rats (284+/-54 g+/-SD; 9 M, 4 F) received nitrous oxide in the presence of isoflurane. Stimulation greater or equal to 40 V induced supramaximal M and F waves in all rats, and only these data are presented. Late waves due to monosynaptic reflex activity (H waves) occurred at similar latency as F waves, but achieved maximal amplitude at stimulation voltages below those that produced maximal M waves (i.e., 10 or 20 V). [19]At initial electrode placement, H waves were sought, but rarely seen. When they were detected, electrode locations were adjusted so that only F waves were elicited during the experimental protocol.

Dose Response. During steady-state isoflurane administration, nitrous oxide produced a dose-dependent decrease in F-wave amplitude (Figure 1) and no change in M-wave amplitudes. The resulting dose-dependent depression in F/M ratio is illustrated in Figure 2. The slope of the dose-response curve (42% decrease in F/M ratio between 0.8 and 1.2 MAC) is similar to that (53% decrease during the same dose interval) previously reported for isoflurane. [9].

Relation of F-Wave Depression to Tail-Clamp Response. Based on the ratio of responders to nonresponders at each mixture of nitrous oxide and isoflurane, logistic regression (JMP 1.0, SAS Institute Inc., Cary, NC) indicated that 44% N²O (29-62%: 95% confidence limits) added to 1.0% isoflurane provided the equivalent of 1.0 MAC anesthesia. At the deepest level of anesthesia that permitted movement (0% N²O in 4 rats, 30% N²O in 7 rats, and 70% N²O in 2 rats), rats had a mean F/M of 20.6+/-17.5%; at the lightest level of anesthesia blocking movement, rats had a mean F/M of 13.7+/-13.9% (Figure 3). The F/M ratio difference between responders and nonresponders was significant (P = 0.01 by paired, one-tailed t test). The mean change in F/M ratio from the "move" concentration to the "no-move" concentration was computed for each animal. The result for the 11 rats was a mean decrease of 54+/-41%. This figure included two animals in which the F/M ratio increased (mean increase, 14+/-2%) when nitrous oxide concentration was increased from the "move" to the "no-move" level: in one animal, the F/M ratio increased consequent to a greater decrease in M- than F-wave amplitude, whereas in the other, F-wave amplitude increased with increasing nitrous oxide concentration.

Eleven Sprague-Dawley rats (6M, 5F; 293+/-27 g) were studied during exposure to hyperbaric nitrous oxide. Stable electromyographic recordings (M-wave amplitude variability < 10%) could not be obtained in three (2M, 1F) animals once within the pressure chamber. Data from these rats were excluded from further electromyographic analysis. When compared with other anesthetic regimes of similar potency (nearly equal 0.8 MAC), F waves (but not M waves) recorded during nitrous oxide administered at 1.6 atm were smaller (F/M ratio = 4.6+/-2.5% SD; (Figure 2and Figure 4) than those obtained in study 1 (18.1+/-14.9% during 30% N²O/isoflurane), or those obtained previously with isoflurane [9](22+/-22%). However, the slopes of the dose-response curves of each of the three anesthetic regimens are similar when the curves produced are each normalized (to unitary value) at their respective baseline (lowest) doses (0.5-0.7 MAC).

Spontaneous tremor and head bobbing in 10/11 animals at 1-1.5 atm N²O made the determination of nitrous oxide potency more difficult than usual. Logistic regression indicated that 2.0 (1.71-2.28%: 95% confidence limits) atm N²O provided the equivalent of 1.0 MAC anesthesia.

At the minimal hyperbaric nitrous oxide pressure blocking purposeful movement in response to tail stimulation, there was a mean F/M of 3.7+/-2.9%, whereas the F/M recorded at the highest nitrous oxide pressure that permitted movement was 4.4+/-2.7% (P < 0.04 by one-tailed, paired t test).

F/M ratios at the two different hyperbaric oxygen pressures were compared with an unpaired, two-tailed t test. Motor neuron excitability did not differ when nitrous oxide partial pressure was held constant at 1.0 atm and oxygen partial pressure was set to either 1.0 atm or 2.4 arm (P > 0.38), suggesting no confounding effect from anesthetic properties of oxygen.

Before this study, there existed some uncertainty regarding MAC of nitrous oxide in rats. Earlier reports determined nitrous oxide MAC to be 1.51-1.55 atm [24]; a more recent study reports MAC to be 2.21 atm. [22]In the current study, we used the methods of the later study and arrived at a value of 2.0 atm.

Nitrous oxide appears to share with isoflurane a capacity to depress evoked F-wave amplitude. The addition of nitrous oxide to stable isoflurane anesthesia produces depression of F-wave amplitude equivalent to that produced by equipotent isoflurane-oxygen anesthesia. [9]That is, conventional partial pressures of nitrous oxide in combination with isoflurane depress the excitability of spinal motor neurons.

Nitrous oxide used as a sole agent appears to be a more potent inhibitor of recurrent motor neuron responses than equivalent MAC multiples of isoflurane alone or in combination with nitrous oxide. We found fourfold greater depression in F-wave amplitude in the presence of hyperbaric nitrous oxide compared with the anesthetics of which isoflurane was a part. Yet the tremor and head bobbing observed in these animals suggested the presence of significant motor tone, conflicting with the hypothesis that F-wave amplitude is directly related to motor neuron excitability. This dichotomy was noted in spastic patients, [25]and the suggested explanation is noted below.

M-wave amplitude was similar for all anesthetic mixtures and pressures, suggesting neither the neuromuscular junction nor the muscle per se led to changes in the F-wave.

Several factors may explain the disparate potencies of normobaric and hyperbaric nitrous oxide to diminish F-wave amplitude, and the lack of relation of F-wave amplitude to behavioral assessment of motor tone during hyperbaric exposure. We believe we excluded simple mechanical or artifactual etiologies due to differences in muscle stretch and geometry that resulted from splinting and fixation within the hyperbaric cylinder, or possible ischemia of the splinted foot, by finding no effect of splinting in a pilot study of two normobaric animals anesthetized with isoflurane. Another possible factor was the probable presence of hypercarbia in these spontaneously ventilating animals. We reported previously that hypercarbia potentiates F-wave depression during isoflurane anesthesia. [9].

Nitrous oxide may depress F-wave activity out of proportion to its potency to cause surgical immobility or depression of spinal motor neurons. In the current study, rats exposed to hyperbaric nitrous oxide demonstrated tremor and spontaneous movements, yet, in some cases, did not respond to noxious stimuli. Hyperbaric nitrous oxide reportedly increased motor tone and caused rigidity in humans. [10]F-wave amplitude enhancement is usually apparent in pathologic states associated with increased muscle tone (e.g., in spastic patients), [15,17,18]however, F-wave depression also may result. [25]Decreased F-wave amplitude in these hyperexcitable states is presumed secondary to a relatively high degree of depolarization of the motor neuron. Because the membrane potential rests closer to the threshold required to generate an action potential, when an antidromic impulse reaches the ventral horn, there is decreased latency in generation of the recurrent action potential on the somato-dendritic tree. If the latency of the recurrent action potential is decreased enough (< [nearly equal] 0.7 msec), it will reach the initial segment of the motor axon while it is still refractory. [18]Therefore, although the soma of more motor neurons may generate recurrent potentials, few of these waves traverse the initial axonal segment to depolarize the innervated muscle and become detectable F waves. Accordingly, we propose that hyperbaric nitrous oxide leads to similar circumstances by rendering hyperexcitable the motor neurons whose recurrent impulse output is relatively blocked, thereby diminishing the observable F-wave response. Preliminary results indicate that 75% N²O depolarizes neurons,***** an action opposite to other inhaled anesthetics. [26-28]Noting that relatively small doses of nitrous oxide added to isoflurane did not increase, but rather reliably decreased, F-wave amplitude, we propose two alternative hypotheses: (1) a nonlinear dose-response curve with an increase in nitrous oxide potency between 532 mmHg (normobaric 70% N²O) and 745 mmHg (29% N²O at 35 PSIg), or (2) that nitrous oxide at any pressure may possess both excitatory and inhibitory effects on motor neurons and that isoflurane may block the excitatory effect and unmask nitrous oxide-mediated depression.

Acute central nervous system tolerance to nitrous oxide may be demonstrated by electroencephalographic patterns [29]or subjective pain scores. [30]We considered the possibility that tolerance phenomena might occur in the spinal cord and could complicate interpretation of the current observations. We did not observe evidence of tolerance, possibly because we allowed time at steady state (greater or equal to 20 m) for equilibration, which may have been sufficient, based on prior literature for the tolerant state to develop before the first measurement.

Another limitation of the current study is the lack of data that provide F-wave responses during low partial pressures of nitrous oxide in the absence of isoflurane. These data cannot be collected in intact animals because of the problems of awareness and pain, but may, in the future, be collected from decerebrate animals.

F-wave amplitude depression is an indirect indicator of the excitability of spinal motor neurons. In the absence of abnormal excitatory phenomena, this depression can, currently, only be classified as an association with increasing anesthetic dose. Nitrous oxide in combination with potent inhaled anesthetics enhances this depression, but when delivered alone, at increased partial pressures, appears to increase motor neuron excitability. We speculate, however, that depression of neuronal excitability may be a core mechanism of general anesthetics, and although different anesthetic agents may cause this depression by different subcellular mechanisms, depression of motor neurons may explain, in large part, surgical immobility.

The authors thank Ken Lee, for assistance in the laboratory; Winifred von Ehrenburg, for editorial assistance; Michael Laster, D.V.M., and Edmond I. Eger II, M.D., for assistance with the hyperbaric exposure system; and Michael Halsey, Ph.D., for suggestions.

*Yamamori Y, Kishikawa K, Collins JG: Halothane reduction of afferent input to the spinal cord in decerebrate, spinally transected rats. Soc Neurosci Abstr 1992;18:132.

**Collins JG, Ota K, Yanagidani T, Hinds A, Cannan S: Effects of nitrous oxide on spinal dorsal horn neuronal responses to low-intensity receptive field stimulation in cats. Soc Neurosc Abstracts 1993;19:1197.

***Smith NT, Hoff BH, Rampil IJ, Sasse FJ, Flemming DC: Does thiopental or N2O disrupt the EEG during enflurane? (abstract). ANESTHESIOLOGY 1979; 51:4.

****Using an Excel (Microsoft) spreadsheet written by J. Feiner and D. Fisher. The spreadsheet is available for downloading from "http://ira-mac.ucsf.edu/NK.xl".

*****Buck LT, Bickler PE, Rampil IJ: Nitrous oxide depolarizes pyramidal neurons in turtle cerebral cortex (abstract). Anesth Analg 1996;82:S52.