- Email: [email protected]
Single-sweep cortical somatosensory evoked potentials: N20 and evoked bursts in sevoflurane anaesthesia
Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
Single-sweep cortical somatosensory evoked potentials: N20 and evoked bursts in sevoflurane anaesthesia V. Ja¨ntti a ,*, E. Sonkaja¨rvi b, S. Mustola c, S. Rytky a, P. Kiiski a, K. Suominen a a
Department of Clinical Neurophysiology, Oulu University Hospital, Oulu, Finland b Department of Anaesthesiology, Oulu University Hospital, Oulu, Finland c South Carelian Central Hospital, Lappeenranta, Finland Accepted for publication: 18 December 1997
Abstract Cortical evoked responses to median nerve stimulation were recorded from 21 subjects during sevoflurane anaesthesia at the level of burst suppression in EEG. The N20/P22 wave had the typical form of a negative wave postcentrally, and positive precentrally. The amplitude exceeded 4 mV in all patients, making it easily visible without averaging on the low-amplitude suppression. These results show that two kinds of somatosensory evoked potential can be studied without averaging during EEG suppression in deep anaesthesia. One is the localised N20/P22 wave, which is seen regularly during suppression after stimuli with intervals exceeding 1 s. The other is the burst, involving the whole cortex, which is not evoked by every stimulus. We suggest that somatosensory evoked potentials can be monitored during sevoflurane-induced EEG suppression, and often can be evaluated reliably from a couple of single sweeps with stimulation interval exceeding 1 s. The enhancement of early cortical components of SEP, their adaptation to repeated stimuli, and the disappearance of later polysynaptic components during EEG suppression, give new possibilities to study the generators of SEP and the different effects of anaesthetics. 1998 Elsevier Science Ireland Ltd. Keywords: EEG; Anaesthesia; Somatosensory evoked potentials (SEP); Burst suppression; Reactivity
1. Introduction Somatosensory evoked potentials (SEP) are widely used for diagnostic purposes as well as monitoring during surgery. Due to the relatively low amplitude of the response it is usually extracted from the ongoing EEG with an averaging technique. However, it should be visible in the ongoing EEG, as the amplitude of most SEP components exceeds the noise level of modern EEG amplifiers. It has been shown that with filtering techniques single-sweep SEPs can be extracted from ongoing EEG (Nishida et al., 1993). Another approach is to suppress pharmacologically the ongoing EEG. This is readily achieved during general anaesthesia with volatile anaesthetics which produce burst suppression, such as isoflurane, desflurane or sevoflurane * Corresponding author. Department of Clinical Neurophysiology, Oulu University Hospital, P.O. Box 22, FIN-90221 Oulu, Finland. Tel.: +358 8 3154525; fax: 358 8 3154544; e-mail: [email protected]
0168-5597/98/$19.00 1998 Elsevier Science Ireland Ltd. All rights reserved PII S0168-5597 (98 )0 0005-7
(Osawa et al., 1994; Hoffman and Edelman, 1995). The N20 wave has been shown to be recordable during isoflurane-induced EEG suppression (Porkkala et al., 1994). Vandesteene et al. (1993) showed that the P22 wave is actually enhanced during isoflurane anaesthesia. During isoflurane-induced burst suppression, somatosensory, auditory and photic stimuli readily produce bursts in EEG with a latency of approximately 300 ms (Hartikainen et al., 1995). Therefore, it has been suggested that the threat of damage to peripheral nerves could be detected by monitoring the burst responses to stimulation of a peripheral nerve (Hartikainen et al., 1996). Sevoflurane is a new volatile anaesthetic which has many properties favouring its use in general anaesthesia over the isoflurane anaesthesia we used in previous studies (Scholz et al., 1996). The present study was undertaken to see whether cortical short-latency evoked potentials could be detected without averaging in sevoflurane anaesthesia. We also wanted to study the bursts induced by electrical stimulation of the
EEP 97656
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
median nerve and whether any other cortical components could be detected between the N20 wave and burst. Part of the results have been presented at the Progress Report Meeting of the Finnish Society of Clinical Neurophysiology in February, 1997 and at the International Congress of Clinical Neurophysiology, Florence, August, 1997 (Ja¨ntti et al., 1997).
321
bandwidth 0.2–1000 Hz. Due to the limited number of channels, different recording montages were used for different patients. Electrodes were located at scalp positions Fp1, F3, FC3, C3, CP3, P3, O1, F4, C4, P4, O2, ear electrode A2 (ipsilateral to stimulus) and extracerebral electrodes at neck C7 or the shoulder contralateral to stimulation. A custom-made analysis program was used to mark burst and suppression onsets and classify responses. This program also reconfigured montages and averaged selected sweeps.
2. Methods and materials Twenty-one ASA 1-2 patients undergoing routine surgery (age 22–52 years, mean 35 years, 16 male) were studied. The study was approved by the Oulu University Hospital Ethics Committee and the patients gave their informed consent. Anaesthesia was induced by mask with sevoflurane in 100% oxygen. The anaesthesia was deepened to a level of steady burst suppression which was achieved at 1.5–2.5 MAC. At this level the patients were intubated and anaesthesia was continued with sevoflurane in 40% oxygen in air. Four patients were given muscle relaxation with a precurarisation dose of vecuronium bromide and short-acting depolarising muscle relaxant succinyl choline, because they had spontaneous ventilation. The percentage of suppression for the whole analysis time was between 38% and 74%. Electrical stimulation was applied to the right median nerve in 20 patients and to the left in one. The intensity was adjusted to 3 times sensory threshold, enough to produce movement of the thumb and the duration was 0.2 ms. Trains of stimuli at frequencies 20, 10, 5, 4, 3, 2 and 1 Hz were applied in addition to single stimuli. Due to limited recording time before starting the operation, different stimulation trains were studied in different patients. EEG was recorded with a Nicolet Viking IV P electromyograph, using the IOM program. Eight channels were recorded: one was used to monitor EEG on screen to detect the burst suppression pattern, one recorded the electric stimulation at the wrist, and 6 were used to record EEG continuously to hard disk at a sampling frequency of 1000 Hz,
3. Results Two kinds of response could be seen after stimulation of the median nerve without averaging. The first is the shortlatency waveform, regularly seen after single stimuli. The second is the burst, occasionally following the stimulus (Fig. 1). No cortical activity was visible between the two responses, not even after averaging several responses (Fig. 2). Stimuli delivered during suppression evoked a waveform, which was negative parietally, and positive centrally and frontally, contralateral to stimulation (Fig. 2). In 6 patients only frontal and central electrodes were applied and the negative peak was not seen. The maximum amplitude of the positive wave was 12 mV and negative 10 V with ear reference ipsilateral to the stimulus. The average latency of the peak was 26 ms, SD 1.8 ms with P3-C3 or O1-C3. In all patients, responses exceeding 4 mV were seen. Therefore, this response was visible in all patients without averaging, optimally in parietal to frontal or central derivation (Fig. 3B). With ear reference ipsilateral to stimulus the widespread P14, preceding the N20 wave was also visible after averaging (Fig. 2). Some of the single stimuli evoked a burst (Fig. 1). The percentage of stimuli evoking bursts in individual patients ranged from 0% to 54%, mean 17%. The N20/P22 was similar in amplitude and latency after those stimuli which evoked bursts and those which did not. Trains of stimuli also induced the N20/P22 wave repeatedly. However, at 20 Hz the waveform decreased quickly in
Fig. 1. Two types of response to right median nerve stimulation are seen, without averaging. Upper trace: recording from P3-C3. Lower trace: stimulus marker. Note the short-latency response, N20, and the long-latency burst.
322
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
Fig. 2. Averaged (n = 104) SEP to stimulation of the left median nerve, reference left ear. Note that the precentral positive wave peaks later than the negative parietal peak, probably reflecting the P22 generator. Also note the widespread P14 wave.
amplitude after the first stimulus of the train. A stimulus frequency of 5 Hz, typically used in evoked potential monitoring, caused a gradual decrease in amplitude to a fraction of the first response. Even with the low rate of 1 Hz, a 30% decrease in amplitude from the first stimulus to the second was seen (Fig. 4).
Hz. Vandesteene et al. (1993) showed that the topographic distribution of the enhanced P22 did not differ from the distribution while awake. This suggests that despite the increase in latency, the N20/P20 and P22 generators are the same during anaesthesia and while awake. This, together with our results, suggests that studying SEP topography and adaptation to stimulation at different frequencies during anaesthesia could help to distinguish between the different generators of SEP. Furthermore, this could give information about the mode of action of different anaesthetics, as has been shown in animal studies by Angel (1993). This could also give information on the mechanism of burst suppression. Interestingly, isoflurane, enflurane and sevoflurane all produce burst suppression and enhance the P22. Halothane, which does not produce the burst suppression pattern, suppresses the P22 (Vandesteene et al., 1991). We hypothesise that enhancement of P22 and N20/P20 during anaesthesia reflects an increased tendency to synchronisation or disinhibition which is, in the end, counteracted by suppression. Failure of the suppression mechanism might be the reason for epileptic seizures, occasionally occurring during enflurane anaesthesia (Ja¨ntti and Yli-Hankala, 1990). The patient in our series who had the highest amplitude N20 (Fig. 1) had occasionally very sharp waves in bursts, resembling epileptiform patterns, despite no history of epileptic seizures. In isoflurane anaesthesia at the level of EEG burst suppression the N20 wave is a single monophasic wave with no following waves. This can still be recorded during contin-
4. Discussion Our results show that during sevoflurane-induced EEG suppression a high amplitude N20/P22 wave can be regularly recorded in all our patients. Sevoflurane anaesthesia induces a striking increase in the P22 + P20 potential recorded over the precentral scalp region as well as the N20 potential recorded over the postcentral region. Due to the very low amplitude of ongoing EEG and the lack of EMG, these are visible in single sweeps without averaging. After only a few stimuli, the repetitive waveform is confirmed from superimposed sweeps and a reliable estimate of amplitude and latency achieved. This should be available in approximately 10 s with the optimal technique. Due to the negative postcentral potential and widespread almost synchronous positive potential in central and frontal leads, the optimal recording derivation for monitoring purposes is P3– C3 (P4–C4). An increase in P22 amplitude during isoflurane and enflurane anaesthesia, but not during halothane anaesthesia, has been reported by Vandesteene et al. (1991). They did not report an increased N20 amplitude, which was considerable in our study with sevoflurane (Fig. 4). Vandesteene et al. (1991) did not, however, use such high MAC concentrations as we did, neither did they correlate their findings with ongoing EEG, and they used the stimulation frequency 3
Fig. 3. (A) Averaged right median nerve SEP. Upper trace: patient awake (n = 77). Lower trace, during EEG suppression (n = 45). Recording from P3-C3. Note the increase of the N20 wave and the disappearance of later waves. (B) Single-sweep somatosensory evoked potential to stimulation of left median nerve. Recording from P4-C4. Responses to 82 successive stimuli are superposed. Note the stability of the waveform, each response is distinctly visible from the baseline.
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
323
Fig. 4. Adaptation of the N20 wave to repeated stimuli to right median nerve. Grand average from 6 patients. Recording from P3-C3. Note the rapid decrease in the amplitude of the response after the sharp stimulus artefact spike at 20 Hz stimulation. Repeated stimuli even at the frequency of 1/s still cause a minor decrease in amplitude.
uous EEG suppression (Porkkala et al., 1994). The subcortical P14, recorded with a nasopharyngeal electrode, is only minimally affected by isoflurane (Porkkala et al., 1997). Our results show that the stimulation frequency used by Porkkala et al. (1994), 5/s, gives during sevoflurane-induced suppression a significantly lower-amplitude response than can be achieved with stimulus intervals exceeding 1 s. Hence, the decrease in SEP N20 amplitude with increasing concentrations of volatile anaesthetics reflects the increased adaptation to repeated stimuli at the stimulus frequencies usually applied, and at very low stimulation rates the response is actually enhanced. We hypothesise that this is due to decreasing inhibition mediated by gamma-aminobutyric acid with increasing concentrations of sevoflurane. It has been suggested that N20 is generated by excitatory postsynaptic potentials and the following positivity by inhibitory postsynaptic potentials in the primary somatosensory area (Wikstro¨m et al., 1996). Increasing concentrations of isoflurane (Porkkala et al., 1997) and sevoflurane first cause disappearance of long and middle latency components and then a decrease in the positive wave following N20, particularly at a stimulation frequency of 5 Hz, leaving a monophasic N20. Suppressed EEG is not synonymous with electrical silence (Prior and Maynard, 1986). While most of the cortical cells are silent during suppression, most thalamic cells remain active (Steriade et al., 1994). Low amplitude theta activity is seen during isoflurane-induced suppression (Rosner and Clark, 1973), 13 Hz spindles up to 100 mV are seen on the negative side of the positive suppression level in propofol-induced suppression (Ja¨ntti et al., 1993), pathological patterns such as alpha coma pattern (Zaret, 1985) and focal or generalised epileptiform activity can be seen during EEG suppression (Ja¨ntti et al., 1994). These patterns seem to occur spontaneously. The somatosensory short-latency response described in this paper represents a cortical
response to an external stimulus, as does the stimulusinduced burst. As in a healthy brain, the focal N20 wave can reach 10 mV during suppression, it is obvious that the amplitude criterion alone cannot be sufficient for definition of a burst suppression pattern, and that other features such as the DC shift should be among the criteria (Ja¨ntti et al., 1993). Three-second 20 Hz runs of somatosensory, photic or auditory stimuli and 3 s runs of vibration applied at the palm evoke bursts at the onset or the end of stimulation (Yli-Hankala et al., 1993; Hartikainen et al., 1995). The evoked bursts are strongly adapting: regularly-occurring stimuli with frequencies as low as 1/10 s, which is the respiration frequency we have used in previous studies, only rarely evoke bursts. The different, and often very repeatable, waveform, of burst onset with different modes of stimuli suggests stimulus specificity. In this experiment, all single stimuli produced N20 waves which were similar in waveform and amplitude both when followed by a burst and when not. No cortical activity was detected between the N20 wave and the burst. This suggests that the onset of burst is determined at the subcortical level and does not involve cortical activity after the N20 wave. We had, however, a very limited number of channels, and more detailed topographic analysis is necessary to study the patterns of cortical activation at burst onset. Our results show that two kinds of response to median nerve stimulation can be recorded without averaging during sevoflurane-induced suppression. The short-latency N20/ P20 and partly-overlapping P22 can be recorded, particularly with short derivations across the central sulcus after every stimulus. The other response is the burst, but it does not follow every stimulus. Even the short-latency components adapt to repeated stimuli, as shown in Fig. 4. The differences in adaptation and the lack of later, polysynaptic, components should together give new possibilities for char-
324
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
acterisation of the generators of short-latency components, as well as the different effects of general anaesthetics. Acknowledgements The skilful technical assistance of Mr. Mika Mustonen, Rt. EEG Tech., is gratefully acknowledged. References Angel, A. Central neuronal pathways and the process of anaesthesia. Br. J. Anaesth., 1993, 71: 148–163. Hartikainen, K.M., Rorarius, M., Pera¨kyla¨, J.J., Laippala, P.J. and Ja¨ntti, V. Cortical reactivity during isoflurane burst-suppression anesthesia. Anesth. Analg., 1995, 81: 1223–1228. Hartikainen, K.M., Rorarius, M. and Ja¨ntti, V. Monitoring the integrity of somatosensory pathways with evoked electroencephalographic bursts. Anesth. Analg., 1996, 83: 354–358. Hoffman, W.E. and Edelman, G. Comparison of isoflurane and desflurane anesthetic depth using burst suppression of the electroencephalogram in neurosurgical patients. Anesth. Analg., 1995, 81 (4): 811–816. Ja¨ntti, V. and Yli-Hankala, A. Correlation of instantaneous heart rate and EEG suppression during enflurane anaesthesia: synchronous inhibition of heart rate and cortical electrical activity?. Electroenceph. clin. Neurophysiol., 1990, 76: 476–479. Ja¨ntti, V., Yli-Hankala, A., Baer, G.A. and Porkkala, T. Slow potentials of EEG burst suppression pattern during anaesthesia. Acta Anaesthesiol. Scand., 1993, 37: 121–123. Ja¨ntti, V., Eriksson, K., Hartikainen, K. and Baer, G.A. Epileptic EEG discharges during burst suppression. Neuropediatrics, 1994, 25: 271– 273. Ja¨ntti, V., Sonkaja¨rvi, E., Porkkala, T., Mustola, S., Rytky, S. and Suominen, K. Short latency somatosensory evoked potentials P14 and N20 during isoflurane and sevoflurane induced EEG suppression. Abstract P-19-8. Electroenceph. clin. Neurophysiol., 1997, 103: 105. Nishida, S., Nakamura, M. and Shibasaki, S. Method for single-trial recording of somatosensory evoked potentials. J. Biomed. Eng., 1993, 15: 257–262.
Osawa, M., Shingu, K., Murakawa, M., Adachi, T., Kurata, J., Seo, N., Murayama, T., Nakao, S. and Mori, K. Effects of sevoflurane on central nervous system electrical activity in cats. Anesth. Analg., 1994, 79: 52– 57. Porkkala, T., Ja¨ntti, V., Kaukinen, S. and Ha¨kkinen, V. Somatosensory evoked potentials during isoflurane anaesthesia. Acta Anaesthesiol. Scand., 1994, 38: 206–210. Porkkala, T., Ja¨ntti, V., Kaukinen, S. and Ha¨kkinen, V. Median nerve somatosensory evoked potentials during isoflurane anaesthesia. Can. J. Anaesth., 1997, 44: 963–968. Prior, P.F. and Maynard, D.M. Monitoring Cerebral Function: Long-term Monitoring of EEG and Evoked Potentials. Elsevier, Amsterdam, 1986. Rosner, B.S. and Clark, D.L. Neurophysiologic effects of general anesthetics: II. Sequential regional actions in the brain. Anesthesiology, 1973, 39: 59–81. Scholz, J., Bischoff, P., Szafarczyk, W., Heetel, S. and Schulte, J. Sevofluran im Vergleich zu Isofluran bei ambulanten Operationen. Ergebnisse einer multizentrischen Studie. Anaesthetist, 1996, 45 (Suppl. 1): 63–70. Steriade, M., Amzica, F. and Contreras, D. Cortical and thalamic cellular correlates of electroencephalographic burst-suppression. Electroenceph. clin. Neurophysiol., 1994, 90: 1–16. Vandesteene, A., Nogueira, M.C., Mavroudakis, N., Defevrimont, M., Brunko, E. and Zegers de Beyl, D. Synaptic effects of halogenated anesthetics on short-latency SEP. Neurology, 1991, 41: 913–918. Vandesteene, A., Mavroudakis, N., Defevrimont, M., Brunko, E. and Zegers de Beyl, D. Topographic analysis of the effects of isoflurane anesthesia on SEP. Electroenceph. clin. Neurophysiol., 1993, 88: 77– 81. Wikstro¨m, H., Huttunen, J., Korvenoja, A., Virtanen, J., Salonen, O., Aronen, H. and Ilmoniemi, R.J. Effects of interstimulus interval on somatosensory evoked magnetic fields (SEFs): a hypothesis concerning SEF generation at the primary sensorimotor cortex. Electroenceph. clin. Neurophysiol., 1996, 100: 479–487. Yli-Hankala, A., Ja¨ntti, V., Pyykko¨, I. and Lindgren, L. Vibration stimulus induced EEG bursts in isoflurane anaesthesia. Electroenceph. clin. Neurophysiol., 1993, 87 (4): 215–220. Zaret, BS. Prognostic and neurophysiological implications of concurrent burst suppression and alpha patterns in the EEG of post-anoxic coma. Electroenceph. clin. Neurophysiol., 1985, 61: 199–209.
Single-sweep cortical somatosensory evoked potentials: N20 and evoked bursts in sevoflurane anaesthesia V. Ja¨ntti a ,*, E. Sonkaja¨rvi b, S. Mustola c, S. Rytky a, P. Kiiski a, K. Suominen a a
Department of Clinical Neurophysiology, Oulu University Hospital, Oulu, Finland b Department of Anaesthesiology, Oulu University Hospital, Oulu, Finland c South Carelian Central Hospital, Lappeenranta, Finland Accepted for publication: 18 December 1997
Abstract Cortical evoked responses to median nerve stimulation were recorded from 21 subjects during sevoflurane anaesthesia at the level of burst suppression in EEG. The N20/P22 wave had the typical form of a negative wave postcentrally, and positive precentrally. The amplitude exceeded 4 mV in all patients, making it easily visible without averaging on the low-amplitude suppression. These results show that two kinds of somatosensory evoked potential can be studied without averaging during EEG suppression in deep anaesthesia. One is the localised N20/P22 wave, which is seen regularly during suppression after stimuli with intervals exceeding 1 s. The other is the burst, involving the whole cortex, which is not evoked by every stimulus. We suggest that somatosensory evoked potentials can be monitored during sevoflurane-induced EEG suppression, and often can be evaluated reliably from a couple of single sweeps with stimulation interval exceeding 1 s. The enhancement of early cortical components of SEP, their adaptation to repeated stimuli, and the disappearance of later polysynaptic components during EEG suppression, give new possibilities to study the generators of SEP and the different effects of anaesthetics. 1998 Elsevier Science Ireland Ltd. Keywords: EEG; Anaesthesia; Somatosensory evoked potentials (SEP); Burst suppression; Reactivity
1. Introduction Somatosensory evoked potentials (SEP) are widely used for diagnostic purposes as well as monitoring during surgery. Due to the relatively low amplitude of the response it is usually extracted from the ongoing EEG with an averaging technique. However, it should be visible in the ongoing EEG, as the amplitude of most SEP components exceeds the noise level of modern EEG amplifiers. It has been shown that with filtering techniques single-sweep SEPs can be extracted from ongoing EEG (Nishida et al., 1993). Another approach is to suppress pharmacologically the ongoing EEG. This is readily achieved during general anaesthesia with volatile anaesthetics which produce burst suppression, such as isoflurane, desflurane or sevoflurane * Corresponding author. Department of Clinical Neurophysiology, Oulu University Hospital, P.O. Box 22, FIN-90221 Oulu, Finland. Tel.: +358 8 3154525; fax: 358 8 3154544; e-mail: [email protected]
0168-5597/98/$19.00 1998 Elsevier Science Ireland Ltd. All rights reserved PII S0168-5597 (98 )0 0005-7
(Osawa et al., 1994; Hoffman and Edelman, 1995). The N20 wave has been shown to be recordable during isoflurane-induced EEG suppression (Porkkala et al., 1994). Vandesteene et al. (1993) showed that the P22 wave is actually enhanced during isoflurane anaesthesia. During isoflurane-induced burst suppression, somatosensory, auditory and photic stimuli readily produce bursts in EEG with a latency of approximately 300 ms (Hartikainen et al., 1995). Therefore, it has been suggested that the threat of damage to peripheral nerves could be detected by monitoring the burst responses to stimulation of a peripheral nerve (Hartikainen et al., 1996). Sevoflurane is a new volatile anaesthetic which has many properties favouring its use in general anaesthesia over the isoflurane anaesthesia we used in previous studies (Scholz et al., 1996). The present study was undertaken to see whether cortical short-latency evoked potentials could be detected without averaging in sevoflurane anaesthesia. We also wanted to study the bursts induced by electrical stimulation of the
EEP 97656
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
median nerve and whether any other cortical components could be detected between the N20 wave and burst. Part of the results have been presented at the Progress Report Meeting of the Finnish Society of Clinical Neurophysiology in February, 1997 and at the International Congress of Clinical Neurophysiology, Florence, August, 1997 (Ja¨ntti et al., 1997).
321
bandwidth 0.2–1000 Hz. Due to the limited number of channels, different recording montages were used for different patients. Electrodes were located at scalp positions Fp1, F3, FC3, C3, CP3, P3, O1, F4, C4, P4, O2, ear electrode A2 (ipsilateral to stimulus) and extracerebral electrodes at neck C7 or the shoulder contralateral to stimulation. A custom-made analysis program was used to mark burst and suppression onsets and classify responses. This program also reconfigured montages and averaged selected sweeps.
2. Methods and materials Twenty-one ASA 1-2 patients undergoing routine surgery (age 22–52 years, mean 35 years, 16 male) were studied. The study was approved by the Oulu University Hospital Ethics Committee and the patients gave their informed consent. Anaesthesia was induced by mask with sevoflurane in 100% oxygen. The anaesthesia was deepened to a level of steady burst suppression which was achieved at 1.5–2.5 MAC. At this level the patients were intubated and anaesthesia was continued with sevoflurane in 40% oxygen in air. Four patients were given muscle relaxation with a precurarisation dose of vecuronium bromide and short-acting depolarising muscle relaxant succinyl choline, because they had spontaneous ventilation. The percentage of suppression for the whole analysis time was between 38% and 74%. Electrical stimulation was applied to the right median nerve in 20 patients and to the left in one. The intensity was adjusted to 3 times sensory threshold, enough to produce movement of the thumb and the duration was 0.2 ms. Trains of stimuli at frequencies 20, 10, 5, 4, 3, 2 and 1 Hz were applied in addition to single stimuli. Due to limited recording time before starting the operation, different stimulation trains were studied in different patients. EEG was recorded with a Nicolet Viking IV P electromyograph, using the IOM program. Eight channels were recorded: one was used to monitor EEG on screen to detect the burst suppression pattern, one recorded the electric stimulation at the wrist, and 6 were used to record EEG continuously to hard disk at a sampling frequency of 1000 Hz,
3. Results Two kinds of response could be seen after stimulation of the median nerve without averaging. The first is the shortlatency waveform, regularly seen after single stimuli. The second is the burst, occasionally following the stimulus (Fig. 1). No cortical activity was visible between the two responses, not even after averaging several responses (Fig. 2). Stimuli delivered during suppression evoked a waveform, which was negative parietally, and positive centrally and frontally, contralateral to stimulation (Fig. 2). In 6 patients only frontal and central electrodes were applied and the negative peak was not seen. The maximum amplitude of the positive wave was 12 mV and negative 10 V with ear reference ipsilateral to the stimulus. The average latency of the peak was 26 ms, SD 1.8 ms with P3-C3 or O1-C3. In all patients, responses exceeding 4 mV were seen. Therefore, this response was visible in all patients without averaging, optimally in parietal to frontal or central derivation (Fig. 3B). With ear reference ipsilateral to stimulus the widespread P14, preceding the N20 wave was also visible after averaging (Fig. 2). Some of the single stimuli evoked a burst (Fig. 1). The percentage of stimuli evoking bursts in individual patients ranged from 0% to 54%, mean 17%. The N20/P22 was similar in amplitude and latency after those stimuli which evoked bursts and those which did not. Trains of stimuli also induced the N20/P22 wave repeatedly. However, at 20 Hz the waveform decreased quickly in
Fig. 1. Two types of response to right median nerve stimulation are seen, without averaging. Upper trace: recording from P3-C3. Lower trace: stimulus marker. Note the short-latency response, N20, and the long-latency burst.
322
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
Fig. 2. Averaged (n = 104) SEP to stimulation of the left median nerve, reference left ear. Note that the precentral positive wave peaks later than the negative parietal peak, probably reflecting the P22 generator. Also note the widespread P14 wave.
amplitude after the first stimulus of the train. A stimulus frequency of 5 Hz, typically used in evoked potential monitoring, caused a gradual decrease in amplitude to a fraction of the first response. Even with the low rate of 1 Hz, a 30% decrease in amplitude from the first stimulus to the second was seen (Fig. 4).
Hz. Vandesteene et al. (1993) showed that the topographic distribution of the enhanced P22 did not differ from the distribution while awake. This suggests that despite the increase in latency, the N20/P20 and P22 generators are the same during anaesthesia and while awake. This, together with our results, suggests that studying SEP topography and adaptation to stimulation at different frequencies during anaesthesia could help to distinguish between the different generators of SEP. Furthermore, this could give information about the mode of action of different anaesthetics, as has been shown in animal studies by Angel (1993). This could also give information on the mechanism of burst suppression. Interestingly, isoflurane, enflurane and sevoflurane all produce burst suppression and enhance the P22. Halothane, which does not produce the burst suppression pattern, suppresses the P22 (Vandesteene et al., 1991). We hypothesise that enhancement of P22 and N20/P20 during anaesthesia reflects an increased tendency to synchronisation or disinhibition which is, in the end, counteracted by suppression. Failure of the suppression mechanism might be the reason for epileptic seizures, occasionally occurring during enflurane anaesthesia (Ja¨ntti and Yli-Hankala, 1990). The patient in our series who had the highest amplitude N20 (Fig. 1) had occasionally very sharp waves in bursts, resembling epileptiform patterns, despite no history of epileptic seizures. In isoflurane anaesthesia at the level of EEG burst suppression the N20 wave is a single monophasic wave with no following waves. This can still be recorded during contin-
4. Discussion Our results show that during sevoflurane-induced EEG suppression a high amplitude N20/P22 wave can be regularly recorded in all our patients. Sevoflurane anaesthesia induces a striking increase in the P22 + P20 potential recorded over the precentral scalp region as well as the N20 potential recorded over the postcentral region. Due to the very low amplitude of ongoing EEG and the lack of EMG, these are visible in single sweeps without averaging. After only a few stimuli, the repetitive waveform is confirmed from superimposed sweeps and a reliable estimate of amplitude and latency achieved. This should be available in approximately 10 s with the optimal technique. Due to the negative postcentral potential and widespread almost synchronous positive potential in central and frontal leads, the optimal recording derivation for monitoring purposes is P3– C3 (P4–C4). An increase in P22 amplitude during isoflurane and enflurane anaesthesia, but not during halothane anaesthesia, has been reported by Vandesteene et al. (1991). They did not report an increased N20 amplitude, which was considerable in our study with sevoflurane (Fig. 4). Vandesteene et al. (1991) did not, however, use such high MAC concentrations as we did, neither did they correlate their findings with ongoing EEG, and they used the stimulation frequency 3
Fig. 3. (A) Averaged right median nerve SEP. Upper trace: patient awake (n = 77). Lower trace, during EEG suppression (n = 45). Recording from P3-C3. Note the increase of the N20 wave and the disappearance of later waves. (B) Single-sweep somatosensory evoked potential to stimulation of left median nerve. Recording from P4-C4. Responses to 82 successive stimuli are superposed. Note the stability of the waveform, each response is distinctly visible from the baseline.
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
323
Fig. 4. Adaptation of the N20 wave to repeated stimuli to right median nerve. Grand average from 6 patients. Recording from P3-C3. Note the rapid decrease in the amplitude of the response after the sharp stimulus artefact spike at 20 Hz stimulation. Repeated stimuli even at the frequency of 1/s still cause a minor decrease in amplitude.
uous EEG suppression (Porkkala et al., 1994). The subcortical P14, recorded with a nasopharyngeal electrode, is only minimally affected by isoflurane (Porkkala et al., 1997). Our results show that the stimulation frequency used by Porkkala et al. (1994), 5/s, gives during sevoflurane-induced suppression a significantly lower-amplitude response than can be achieved with stimulus intervals exceeding 1 s. Hence, the decrease in SEP N20 amplitude with increasing concentrations of volatile anaesthetics reflects the increased adaptation to repeated stimuli at the stimulus frequencies usually applied, and at very low stimulation rates the response is actually enhanced. We hypothesise that this is due to decreasing inhibition mediated by gamma-aminobutyric acid with increasing concentrations of sevoflurane. It has been suggested that N20 is generated by excitatory postsynaptic potentials and the following positivity by inhibitory postsynaptic potentials in the primary somatosensory area (Wikstro¨m et al., 1996). Increasing concentrations of isoflurane (Porkkala et al., 1997) and sevoflurane first cause disappearance of long and middle latency components and then a decrease in the positive wave following N20, particularly at a stimulation frequency of 5 Hz, leaving a monophasic N20. Suppressed EEG is not synonymous with electrical silence (Prior and Maynard, 1986). While most of the cortical cells are silent during suppression, most thalamic cells remain active (Steriade et al., 1994). Low amplitude theta activity is seen during isoflurane-induced suppression (Rosner and Clark, 1973), 13 Hz spindles up to 100 mV are seen on the negative side of the positive suppression level in propofol-induced suppression (Ja¨ntti et al., 1993), pathological patterns such as alpha coma pattern (Zaret, 1985) and focal or generalised epileptiform activity can be seen during EEG suppression (Ja¨ntti et al., 1994). These patterns seem to occur spontaneously. The somatosensory short-latency response described in this paper represents a cortical
response to an external stimulus, as does the stimulusinduced burst. As in a healthy brain, the focal N20 wave can reach 10 mV during suppression, it is obvious that the amplitude criterion alone cannot be sufficient for definition of a burst suppression pattern, and that other features such as the DC shift should be among the criteria (Ja¨ntti et al., 1993). Three-second 20 Hz runs of somatosensory, photic or auditory stimuli and 3 s runs of vibration applied at the palm evoke bursts at the onset or the end of stimulation (Yli-Hankala et al., 1993; Hartikainen et al., 1995). The evoked bursts are strongly adapting: regularly-occurring stimuli with frequencies as low as 1/10 s, which is the respiration frequency we have used in previous studies, only rarely evoke bursts. The different, and often very repeatable, waveform, of burst onset with different modes of stimuli suggests stimulus specificity. In this experiment, all single stimuli produced N20 waves which were similar in waveform and amplitude both when followed by a burst and when not. No cortical activity was detected between the N20 wave and the burst. This suggests that the onset of burst is determined at the subcortical level and does not involve cortical activity after the N20 wave. We had, however, a very limited number of channels, and more detailed topographic analysis is necessary to study the patterns of cortical activation at burst onset. Our results show that two kinds of response to median nerve stimulation can be recorded without averaging during sevoflurane-induced suppression. The short-latency N20/ P20 and partly-overlapping P22 can be recorded, particularly with short derivations across the central sulcus after every stimulus. The other response is the burst, but it does not follow every stimulus. Even the short-latency components adapt to repeated stimuli, as shown in Fig. 4. The differences in adaptation and the lack of later, polysynaptic, components should together give new possibilities for char-
324
V. Ja¨ntti et al. / Electroencephalography and clinical Neurophysiology 108 (1998) 320–324
acterisation of the generators of short-latency components, as well as the different effects of general anaesthetics. Acknowledgements The skilful technical assistance of Mr. Mika Mustonen, Rt. EEG Tech., is gratefully acknowledged. References Angel, A. Central neuronal pathways and the process of anaesthesia. Br. J. Anaesth., 1993, 71: 148–163. Hartikainen, K.M., Rorarius, M., Pera¨kyla¨, J.J., Laippala, P.J. and Ja¨ntti, V. Cortical reactivity during isoflurane burst-suppression anesthesia. Anesth. Analg., 1995, 81: 1223–1228. Hartikainen, K.M., Rorarius, M. and Ja¨ntti, V. Monitoring the integrity of somatosensory pathways with evoked electroencephalographic bursts. Anesth. Analg., 1996, 83: 354–358. Hoffman, W.E. and Edelman, G. Comparison of isoflurane and desflurane anesthetic depth using burst suppression of the electroencephalogram in neurosurgical patients. Anesth. Analg., 1995, 81 (4): 811–816. Ja¨ntti, V. and Yli-Hankala, A. Correlation of instantaneous heart rate and EEG suppression during enflurane anaesthesia: synchronous inhibition of heart rate and cortical electrical activity?. Electroenceph. clin. Neurophysiol., 1990, 76: 476–479. Ja¨ntti, V., Yli-Hankala, A., Baer, G.A. and Porkkala, T. Slow potentials of EEG burst suppression pattern during anaesthesia. Acta Anaesthesiol. Scand., 1993, 37: 121–123. Ja¨ntti, V., Eriksson, K., Hartikainen, K. and Baer, G.A. Epileptic EEG discharges during burst suppression. Neuropediatrics, 1994, 25: 271– 273. Ja¨ntti, V., Sonkaja¨rvi, E., Porkkala, T., Mustola, S., Rytky, S. and Suominen, K. Short latency somatosensory evoked potentials P14 and N20 during isoflurane and sevoflurane induced EEG suppression. Abstract P-19-8. Electroenceph. clin. Neurophysiol., 1997, 103: 105. Nishida, S., Nakamura, M. and Shibasaki, S. Method for single-trial recording of somatosensory evoked potentials. J. Biomed. Eng., 1993, 15: 257–262.
Osawa, M., Shingu, K., Murakawa, M., Adachi, T., Kurata, J., Seo, N., Murayama, T., Nakao, S. and Mori, K. Effects of sevoflurane on central nervous system electrical activity in cats. Anesth. Analg., 1994, 79: 52– 57. Porkkala, T., Ja¨ntti, V., Kaukinen, S. and Ha¨kkinen, V. Somatosensory evoked potentials during isoflurane anaesthesia. Acta Anaesthesiol. Scand., 1994, 38: 206–210. Porkkala, T., Ja¨ntti, V., Kaukinen, S. and Ha¨kkinen, V. Median nerve somatosensory evoked potentials during isoflurane anaesthesia. Can. J. Anaesth., 1997, 44: 963–968. Prior, P.F. and Maynard, D.M. Monitoring Cerebral Function: Long-term Monitoring of EEG and Evoked Potentials. Elsevier, Amsterdam, 1986. Rosner, B.S. and Clark, D.L. Neurophysiologic effects of general anesthetics: II. Sequential regional actions in the brain. Anesthesiology, 1973, 39: 59–81. Scholz, J., Bischoff, P., Szafarczyk, W., Heetel, S. and Schulte, J. Sevofluran im Vergleich zu Isofluran bei ambulanten Operationen. Ergebnisse einer multizentrischen Studie. Anaesthetist, 1996, 45 (Suppl. 1): 63–70. Steriade, M., Amzica, F. and Contreras, D. Cortical and thalamic cellular correlates of electroencephalographic burst-suppression. Electroenceph. clin. Neurophysiol., 1994, 90: 1–16. Vandesteene, A., Nogueira, M.C., Mavroudakis, N., Defevrimont, M., Brunko, E. and Zegers de Beyl, D. Synaptic effects of halogenated anesthetics on short-latency SEP. Neurology, 1991, 41: 913–918. Vandesteene, A., Mavroudakis, N., Defevrimont, M., Brunko, E. and Zegers de Beyl, D. Topographic analysis of the effects of isoflurane anesthesia on SEP. Electroenceph. clin. Neurophysiol., 1993, 88: 77– 81. Wikstro¨m, H., Huttunen, J., Korvenoja, A., Virtanen, J., Salonen, O., Aronen, H. and Ilmoniemi, R.J. Effects of interstimulus interval on somatosensory evoked magnetic fields (SEFs): a hypothesis concerning SEF generation at the primary sensorimotor cortex. Electroenceph. clin. Neurophysiol., 1996, 100: 479–487. Yli-Hankala, A., Ja¨ntti, V., Pyykko¨, I. and Lindgren, L. Vibration stimulus induced EEG bursts in isoflurane anaesthesia. Electroenceph. clin. Neurophysiol., 1993, 87 (4): 215–220. Zaret, BS. Prognostic and neurophysiological implications of concurrent burst suppression and alpha patterns in the EEG of post-anoxic coma. Electroenceph. clin. Neurophysiol., 1985, 61: 199–209.