In uence of propofol concentrations on multipulse transcranial motor evoked potentials

Transcription

1 British Journal of Anaesthesia 91 (4): 493±7 (2003) DOI: /bja/aeg211 In uence of propofol concentrations on multipulse transcranial motor evoked potentials N. Nathan 1 *, F. Tabaraud 2, F. Lacroix 1, D. MoulieÁs 3, X. Viviand 4, A. Lansade 1, G. Terrier 1 and P. Feiss 1 1 DeÂpartement d'anestheâsie ReÂanimation Chirurgicale, 2 Service d'exploration Fonctionnelle Neurologique, and 3 Service de Chirurgie PeÂdiatrique, CHU Dupuytren, 2 Avenue Martin Luther King, Limoges, France. 4 DeÂpartement d'anestheâsie ReÂanimation Chirurgicale, Hopital Nord, Boulevard P Dramard, Marseille, France *Corresponding author. nnathan.lim@invivo.edu Background. Motor evoked potentials can be affected by propofol anaesthesia. We studied how increasing target concentrations of propofol altered transcranial motor evoked potentials (tcmep) during scoliosis surgery. Methods. Fifteen patients undergoing surgery for scoliosis were anaesthetized with remifentanil and propofol without nitrous oxide or neuromuscular blocking agents (BIS<60). tcmep were elicited by transcranial electric multipulse stimulation of the motor cortex and recording of compound action potentials from the anterior tibialis muscle. tcmep were obtained before surgery with propofol target values set from 4 to 8 mg litre ±1, and then during surgery. Arterial propofol concentrations were measured for each tcmep recording. Results. Before surgery, increasing propofol reduced tcmep amplitude in a dose-dependent manner, with no effect on latency. During surgery, at equivalent propofol concentrations, tcmep were not statistically different from those obtained before surgery. In all except one patient, tcmep signals were present during the entire procedure. In this patient the loss of tcmep was unfortunately related to an anterior spinal cord lesion, which was con rmed by a wake-up test. Conclusion. We found that, although propofol had a dose-dependent effect on tcmep amplitude, anaesthesia could be maintained with remifentanil and propofol to allow recording and interpretation of tcmep signals. Br J Anaesth 2003; 91: 493±7 Keywords: anaesthetics i.v., propofol; analgesics opioid, remifentanil; monitoring, evoked potential; surgery, spinal Accepted for publication: May 29, 2003 Injury to the spinal cord is a serious risk during spinal surgery and can be detected by appropriate monitoring. Because injury to the spinal cord can occur only in the anterior parts, somatosensory evoked potential (SEP) monitoring may not detect injury of the anterior motor pathway. Loss of motor function can occur after spinal injury despite no changes in recording of SEP, 1 suggesting that this type of monitoring cannot detect motor damage. 23 Many centres advocate monitoring motor evoked potentials (MEP) to assess the motor pathway and avoid the need for a wake-up test. A report of paraplegia, despite normal MEP obtained by direct stimulation of the spinal cord, suggests that MEP should be measured using transcranial stimulation of the motor cortex (tcmep). 4 However, tcmep recordings are markedly affected by anaesthetics such as nitrous oxide, barbiturates and halogenated agents and also by neuromuscular blocking agents, hypothermia and variations of blood carbon dioxide. 56 Propofol in low doses decreases the amplitude of the early components of the SEP, 7 but its effects on MEP have not been fully investigated 589 so we Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2003

2 Nathan et al. Table 1 Patients details [mean (SD)] and amplitude of left and right tcmep at rst measurement [median (25th±75th percentiles)] Mean (SD) Median (25th±75th percentiles) Age (yr) 20.1 (13.3) 16 (14±18) Weight (kg) 56.5 (12.5) 56 (50±63) Height (cm) (11.7) 166 (158±170) Baseline left tcmep amplitude (mv) 181 (112±559) Baseline right tcmep amplitude (mv) 188 (130±462) studied the effects of increasing doses of propofol on MEP evoked by transcranial electric stimulation of the motor cortex. Patients and methods After local ethical committee approval, and with written informed consent, we recruited 15 patients undergoing surgery on the spine with evoked motor potential (MEP) monitoring. The day before surgery, MEP were obtained from all patients. After 1 mg kg ±1 hydroxyzine orally 2 h before surgery, anaesthesia was induced with remifentanil (1 mg kg ±1 i.v. followed by 0.25±0.4 mg kg ±1 min ±1 by infusion) and a target concentration infusion of propofol set to give 6±8 mg litre ±1 (Gepts' pharmacokinetic model). Measurements of propofol arterial concentrations showed large errors in predicted values in the rst three patients. We then used the Marsh's pharmacokinetic model (Diprifusor â, Astra-Zeneca Inc, UK) to target propofol concentrations. ECG, radial arterial and central venous pressure, pulse oximetry and BIS values (Aspect 1000 â, frontal montage, Aspect Medical System Inc., Framingham, MA, USA) were measured in all patients. After orotracheal intubation, the ventilator was set to maintain normocapnia and the propofol target concentration was reduced to 4 mg litre ±1. Oesophageal temperature was monitored and normothermia was maintained by forced air warming. No patient was given nitrous oxide or neuromuscular blocking agents during anaesthesia. We measured the latency and amplitude of tcmep before the patient was positioned or surgery started. The tcmep were elicited by transcranial electric multipulse stimulation (train-of- ve square wave stimulations, 2 ms interstimulus interval, 50 ms stimulus duration and 1000 ma intensity) of the motor cortex (Digitimer D185, Hertfordshire, UK). The compound action potentials were obtained from needle electrodes inserted in the anterior tibialis muscle and recorded by an EMG apparatus (Nicolet Spirit, Madison, WI, USA). tcmep amplitude and latency values were derived automatically by the system and did not depend on the experience of the operator. The same operator performed all the tcmep measurements. The rst measurements of tcmep were obtained at a 4 mg litre ±1 propofol target concentration. The target concentration was then increased by 1 mg litre ±1 increments until between 6 and 8 mg litre ±1. tcmep was recorded after an equilibration period in order to allow the predicted site-effects and arterial concentrations to become similar as shown by the Diprifusor â or computer calculations (for Gept's pharmacokinetics). The equilibration period allowed was 8±10 min before measuring each tcmep. This equilibration period was the same when we used the Marsh's pharmacokinetic model since the same K eo was used in the two models. For each increment of propofol target, arterial pressure, heart rate, PE CO2 and BIS values were obtained. Increases in target concentrations were discontinued when arterial pressure was decreased by >30% of pre-anaesthetic values, or if BIS values were <30, or if the raw EEG showed burst suppression indicating that anaesthesia was too deep. If the tcmep was not obtained at the initial 4 mg litre ±1 target concentration and if the BIS value was <60, the propofol target concentration was reduced by 1 mg litre ±1 decrements until the tcmep was detected. Slow increases were started again as soon as tcmep were obtained. If tcmep were not detected but the BIS value was >60, propofol concentrations were not reduced and a technical problem (such as disconnection of electrodes) was sought. During surgery, after obtaining a set of measurements, the propofol target concentrations were set to obtain a BIS<60. Transcranial tcmep measurements were repeated immediately after over-distraction and derotation of the spine. SEP were also monitored continuously during the entire surgical procedure. Each time tcmep were measured, we sampled arterial blood to measure propofol concentrations using high performance liquid chromatography (HPLC) with electrochemical detection and thymol as internal standard. The method is speci c for propofol as shown with a diode array detector and is linear between 0.02 and 10 mg litre ±1. 10 At the end of surgery, i.v. morphine was given to all except one patient, who was given intrathecal morphine. Results are presented as mean (SD) and compared using the Kruskall±Wallis test. Measured concentrations of propofol and the latency or the amplitude of the tcmep were related by linear regression. Statistical signi cance was assumed if P<0.05. Results None of the 15 patients recruited had any history of epilepsy. The study plan was not completed in four patients and these patients were excluded from statistical analysis. In two of these patients, MEP was only obtained at either one or two concentrations of propofol because of time constraints. In two other patients, tcmep could not be obtained before surgery although the apparatus was functioning perfectly. In one of these two patients (patient A), when the target concentration of propofol was reduced, the BIS increased and clinical awakening occurred. This patient had been given morphine 1 mg intrathecally. During surgery, tcmep reappeared about 4 h 30 min after the start of the procedure. After this, we did not give intrathecal morphine to patients. In the other patient (patient B), the initial 494

3 Propofol target concentration and motor evoked potentials Table 2 Time elapsed from induction (min), BIS, PE CO2 (mm Hg), oesophageal temperature ( C), systolic and diastolic arterial pressure (mm Hg) and heart rate (beats min ±1 ) during different predicted propofol concentrations before and during surgery. Data are mean (SD) Predicted Delay from BIS value PE CO2 Temperature Systolic arterial Diastolic arterial concentration induction (mm Hg) ( C) pressure pressure (mg litre ±1 ) (min) (mm Hg) (mm Hg) Heart rate (beats min ±1 ) Before surgery 4 57 (12) 37 (11) 29 (4) 36.3 (0.6) 98 (14) 55 (12) 70 (18) 5 66 (12) 35 (10) 29 (4) 36.2 (0.6) 101 (12) 57 (12) 64 (9) 6 72 (11) 32 (7) 28 (4) 36.2 (0.6) 99 (14) 56 (13) 65 (10) 7 82 (12) 23 (8) 28 (4) 36.1 (0.8) 99 (18) 58 (10) 64 (7) 8 92 (14) 23 (8) 30 (3) 36.1 (1.1) 95 (16) 57 (14) 72 (7) During surgery (50) 46 (16) 29 (2) 36.6 (0.2) 109 (20) 40 (21) 64 (15) (48) 39 (8) 28 (3) 36.3 (1.0) 96 (16) 60 (10) 81 (17) Table 3 Prediction error (%) of propofol concentration and its absolute value obtained before and during surgery. Data are mean (SD) Prediction error (%) Absolute prediction error (%) Before surgery (n=49) 32 (16) 32 (15) During surgery (n=19) 25 (29) 34 (19) 4 mg litre ±1 target concentration of propofol was reduced, but the BIS value was persistently <60. tcmep remained undetectable for 3 h. The arterial propofol concentration measured by HPLC was 18.9 mg litre ±1, which was four times greater than predicted by Gepts' pharmacokinetic model. As a consequence, we used the commercially available Marsh pharmacokinetic model (Diprifusor â ) for subsequent patients. Details of the 11 patients who completed the study are given in Table 1. Time from induction of anaesthesia, BIS values, PE CO2 and oesophageal temperature at each tcmep recording are shown in Table 2. Greater propofol target concentrations were associated with smaller BIS values (P<0.05, ANOVA) The predicted propofol concentrations of the two patients in whom Gepts' pharmacokinetic model was used, were calculated according to Marsh's pharmacokinetic model. Because the measured values were close to the concentrations predicted by this latter model, the data were kept in the analysis. The mean prediction errors (MPE) of propofol target-concentrations and their absolute values (MAPE) are presented in relation to their time of sampling and to the predicted concentration in Table 3. Baseline amplitudes and latencies of tcmep had considerable variations between subjects (Table 1), so we calculated changes in tcmep amplitude from baseline to compare data. In two out of 11 patients, the predicted propofol concentration had to be reduced to 3 mg litre ±1 to obtain tcmep and thereafter were increased by 1 mg litre ±1 increments. The mean amplitudes of tcmep, but not their latencies, were signi cantly reduced by increasing propofol concentrations during stage I (Table 4 and Fig. 1). A statistical correlation was found between measured propofol blood concentration and change from baseline of the tcmep amplitudes recorded on the left side (r=±0.68, P<0.001) and the right side (r=±0.53, P<0.001; Fig. 2). There was no statistical correlation between blood propofol concentration and tcmep latencies during stage I. The amplitude variation of tcmep [mean (SD)] obtained during surgery reached 34 (63) mv and 65 (86) mv on the left side when propofol concentrations were 3 and 4 mg ml ±1 respectively. The corresponding values obtained on the right side were 26 (54) mv and 30 (63) mvat the same propofol concentrations. These values were not statistically different from those obtained during surgery. However, in one patient, both SEP and tcmep disappeared at the time of rod xation. In this patient, the use of target concentration infusion of propofol with the combination of remifentanil allowed a rapid wake-up test, which con rmed an anterior spinal cord lesion and the patient unfortunately sustained complete paraplegia. No patients moved unexpectedly during surgery and anaesthesia. Discussion We found that propofol reduced tcmep amplitude in a dosedependent manner when using a multipulse stimulation with a constant intensity and number of stimuli. Consequently, an exact knowledge of the effects of increasing propofol concentration on MEP amplitude is needed to avoid the erroneous conclusion that changes in tcmep amplitude represent spinal cord injury during surgical manipulations. However, tcmep latencies were not in uenced by propofol concentrations commonly used to maintain a deep anaesthesia and remained stable during the whole anaesthesia. We could obtain tcmep using multipulse stimulation of the motor pathway, while controlling propofol anaesthesia with Marsh's pharmacokinetic model in all except two patients. The absence of detectable MEP in the patient who had received high dose intrathecal morphine is unexplained and deserves further study. In the other patient in whom tcmep were not recorded before surgery, the 18 mg litre ±1 propofol concentration suppressed the excitability of cortical and spinal motor neurones. Previous authors found that motoneurone excitability is markedly impaired when target propofol concentrations reach 9 mg litre ±1. 11 In 495

4 Nathan et al. Fig 1 Variations from baseline value of tcmep amplitudes (ms) before surgery when target concentrations of propofol are increased by 1 mg litre ±1 increments. Mean (SD), P<0.001 using Kruskall±Wallis tests. Table 4 Latency (ms) of tcmep obtained before (Phase I) and during surgery (Phase II). Data are mean (SD). *Other measurements were made but only at one concentration Propofol target concentrations (mg litre ±1 ) Right latency (ms) Left latency (ms) Before surgery 4(n=11) 26.2 (2.5) 26.2 (2.5) 5(n=11) 26.2 (1.5) 26.2 (2.2) 6(n=11) 26.1 (2.1) 26.1 (2.1) 7(n=7) 25.9 (2,1) 25.1 (2.1) 8(n=7) 26.3 (2.3) 26.3 (2.3) During surgery* 4(n=7) 27.1 (2.2) 26.8 (2.5) 3(n=5) (2.3) 27.0 (2.3) Fig 2 Correlation between propofol concentrations and variation from baseline of the amplitude of tcmep obtained on the right (A) and left (B) side before surgery. this patient, Gepts' pharmacokinetic model was used to control propofol concentrations. This lacked precision and caused abnormally high measured concentrations in one patient. Gepts' pharmacokinetic model was obtained in young adults undergoing sedation, and may not be suf ciently accurate to predict surgical concentrations. 12 Although the weight and age range of our patients was different from those for which the Marsh model was validated, we obtained a mean prediction error, and mean absolute values similar to those previously published. 13 Marsh's pharmacokinetic model tted the present clinical purpose well, as suggested by the absence of movement during surgery and by the low mean prediction error of the model after induction as well as after prolonged infusion. Indeed, the prediction error of propofol remained stable during the whole procedure suggesting that the observed change in tcmep was not a result of an error in predicted propofol concentration, but resulted from changes in spinal cord function. Therefore, the Marsh model is now used routinely in our institution for anaesthesia for scoliosis surgery with tcmep because it allows deep anaesthesia, stable propofol concentrations whatever the duration of infusion, and precise determination of propofol effects on tcmep. In this study, greater propofol concentration reduced tcmep amplitude in a dose-dependent manner, without affecting latency. In the clinical range of propofol concentrations, this effect is of moderate intensity, but may be severe when propofol is over dosed. This effect was observed after a delay to allow an equilibration of propofol between the plasma and effect-site compartments, indicated by the decrease in BIS values in response to propofol increase. These results are not affected by core temperature, FE CO2 or blood pressure, because their values remained stable when the dose±effect relationship was obtained. Because noxious stimuli do not modify myogenic motor evoked potentials during a propofol-based anaesthesia, our results are not likely to have been affected by differences in the depth of analgesia. 14 We found that it is possible to maintain anaesthesia with hypnotic doses of propofol (without nitrous oxide) and to record tcmep adequately. Other authors have used propofol for anaesthesia maintenance and simultaneous MEP measurements. However, in 496

5 Propofol target concentration and motor evoked potentials these studies, propofol was given at a constant infusion rate either alone, or to supplement a ketamine-based anaesthesia ±16 In such studies, MEP during propofol anaesthesia were of similar amplitude to those with an opioid±nitrous oxide-based anaesthetic technique, but less than during ketamine±nitrous oxide anaesthesia. 16 A dose± effect relationship was not investigated, and propofol was infused at a constant rate, which would have resulted in progressively increasing propofol concentrations, and a risk of decreasing MEP amplitudes. Indeed, doubling the propofol concentration from 0.7 mg litre ±1 to 1.4 mg litre ±1 may reduce MEP amplitude by 30±50%. 15 We did not observe such a profound effect of doubling propofol concentration, perhaps because nitrous oxide was omitted. Nitrous oxide 40±60% given during propofol infusion reduces MEP amplitude by 50±70%, 515 suggesting that nitrous oxide enhances the effects of propofol on the MEP. When nitrous oxide was combined with propofol, tcmep amplitude was three times less than during ketamine± nitrous oxide anaesthesia. 16 Multipulse stimulation allows quantitative monitoring of MEP by transcranial electric stimulation during anaesthesia with propofol The ease of obtaining tcmep in the present study may also be related to the use of highfrequency multipulse stimulation. Stimulation of the spinal cord in this way can overcome the conduction block caused by propofol at the alpha-motoneurone, which comes either from supraspinal inhibition or from decreased motoneurone excitability. 12 Multipulse stimulation simulates the physiological spinal I- and D-wave pattern and allows temporal summation of excitatory post-synaptic potentials. Multipulse stimulation creates additional D-waves that help to overcome the inhibition of cortical I-waves induced by propofol We found that tcmep amplitudes were attenuated in a dose-dependent manner by increasing propofol concentrations, whereas their latencies remained unchanged. Large concentrations of propofol can suppress tcmep recording. Thus, the effects of propofol on the tcmep must be known before surgery to allow correct interpretation of changes in tcmep. This can be done easily in current practice by limiting measurements to the clinical range for propofol, namely 2±4 mg litre ±1. Moreover, a combination of propofol and remifentanil without nitrous oxide or neuromuscular blocking agent can provide adequate anaesthesia and allows recording and interpretation of tcmep. Thus a wake-up test was not used in most of the patients, except in one, in whom a spinal cord lesion was diagnosed by the disappearance of tcmep and con rmed by the wake-up test. References 1 Machida M, Weinstein SL, Yamada T, Kimura J, Toriyama S. Dissociation of muscle action potentials and spinal somatosensory evoked potentials after ischemic damage of spinal cord. Spine 1988; 13: 1119±24 2 Owen JH, Laschinger J, Bridwell K, et al. Sensitivity and speci city of somatosensory and neurogenic-motor evoked potentials in animals and humans. Spine 1988; 13: 1111±18 3 Ashkenaze D, Mudiyam R, Boachie-Adjei O, Gilbert C. Ef cacy of spinal cord monitoring in neuromuscular scoliosis. Spine 1993; 18: 1627±33 4 Minahan RE, Sepkuty JP, Lesser RP, Sponseller PD, Kostuik JP. Anterior spinal cord injury with preserved neurogenic `motor' evoked potentials. Clin Neurophysiol 2001; 112: 1442±50 5 Pechstein U, Nadstawek J, Zentner J, Schramm J. Iso urane plus nitrous oxide versus propofol for recording of motor evoked potentials after high frequency repetitive electrical stimulation. Electroencephalogr Clin Neurophysiol 1998; 108: 175±81 6 Calancie B, Klose KJ, Baier S, Green BA. Iso urane-induced attenuation of motor evoked potentials caused by electrical motor cortex stimulation during surgery. J Neurosurg 1991; 74: 897±904 7 Maurette P, Simeon F, Castagnera L, Esposito J, Macouillard G, Heraut LA. Propofol anaesthesia alters somatosensory evoked cortical potentials. Anaesthesia 1988; 43 [Suppl.]: 44±5 8 Jellinek D, Platt M, Jewkes D, Symon L. Effects of nitrous oxide on motor evoked potentials recorded from skeletal muscle in patients under total anesthesia with intravenously administered propofol. Neurosurgery 1991; 29: 558±62 9 Kawaguchi M, Sakamoto T, Inoue S, et al. Low-dose propofol as a supplement to ketamine-based anesthesia during intraoperative monitoring of motor-evoked potentials. Spine 2000; 25: 974±9 10 Uebel RA, Wium CA, Hawtrey AO, Coetzee J. Electrochemical determination of 2,6-diisopropylphenol after high-performance liquid chromatography of extracts from serum. J Chromatogr 1990; 526: 293±5 11 Kerz T, Hennes HJ, Feve A, Decq P, Filipetti P, Duvaldestin P. Effects of propofol on H-re ex in humans. Anesthesiology 2001; 94: 32±7 12 Gepts E, Camu F, Cockshott ID, Douglas EJ. Disposition of propofol administered as constant rate intravenous infusions in humans. Anesth Analg 1987; 66: 1256±63 13 Coetzee JF, Glen JB, Wium CA, Boshoff L. Pharmacokinetic model selection for target controlled infusions of propofol. Assessment of three parameter sets. Anesthesiology 1995; 82: 1328±45 14 Inoue S, Kawaguchi M, Takahashi M, et al. Noxious stimuli do not modify myogenic motor evoked potentials by electrical stimulation during anesthesia with propofol-based anesthesia. Can J Anaesth 2003; 50: 86±91 15 vandongen EP, ter Beek HT, Aarts LP, et al. The effect of two low-dose propofol infusions on the relationship between sixpulse transcranial electrical stimulation and the evoked lower extremity muscle response. Acta Anaesthesiol Scand 2000; 44: 799± Inoue S, Kawaguchi M, Kakimoto M, et al. Amplitudes and intrapatient variability of myogenic motor evoked potentials to transcranial electrical stimulation during ketamine/n 2 O- and propofol/n 2 O-based anesthesia. J Neurosurg Anesthesiol 2002; 14: 213±17 17 Ubags LH, Kalkman CJ, Been HD, Drummond JC. Differential effects of nitrous oxide and propofol on myogenic transcranial motor evoked responses during sufentanil anaesthesia. Br J Anaesth 1997; 79: 590±4 18 Woodforth IJ, Hicks RG, Crawford MR, Stephen JP, Burke D. Depression of I waves in corticospinal volleys by sevo urane, thiopental, and propofol. Anesth Analg 1999; 89: 1182±7 497