|Home | About | Journals | Submit | Contact Us | Français|
Neuraxial anaesthesia produces a sedative and anesthetic-sparing effect. Recent evidence suggests that spinal cord anaesthesia modifies reticulo-thalamo-cortical arousal by decreasing afferent sensory transmission. We hypothesized that epidural anaesthesia produces sensory deafferentation-dependent sedation that is associated with impairment of brainstem transmission. We used brainstem auditory evoked potentials (BAEP) to evaluate reticular function in 11 volunteers. Epidural anaesthesia was induced with 2% 2-chloroprocaine. Hemodynamic and respiratory responses, sensory block level, sedation depth and BAEP were assessed throughout induction and resolution of epidural anaesthesia. Sedation was evaluated using verbal rating score (VRS), observer's assessment alertness/sedation (OAA/S) score, and bispectral index (BIS). Prediction probability (PK) was used to associate sensory block with sedation, as well as BIS with other sedation measures. Spearman rank order correlation was used to associate block level and sedation with the absolute and interpeak BAEP latencies. Sensory block level significantly predicted VRS (PK = 0.747), OAA/S score (PK = 0.748) and BIS. Bispectral index predicted VRS and OAA/S score (PK = 0.728). The latency of wave III of BAEP significantly correlated with sedation level (rho = 0.335, P < 0.01) and sensory block (rho = 0.394, P < 0.01). The other BAEP parameters did not change during epidural anaesthesia. Hemodynamic and respiratory responses remained stable throughout the study. Sedation during epidural anaesthesia depends on sensory block level and is associated with detectable block-dependent alterations in the brainstem auditory evoked responses. Sensory deafferentation may reduce CNS alertness through mechanisms related to brainstem neural activity.
Subarachnoid and epidural anaesthesia reduce hypnotic requirements for midazolam1, 2, thiopental3, and isoflurane4. For example, high thoracic lidocaine epidural anaesthesia produces a 50% MAC-sparing effect for sevoflurane5, while bupivacaine spinal anaesthesia is associated with sedation that increases as a function of block height2, 6. These findings were confirmed by others who found that spinal anaesthesia is accompanied by significant sedation, although unrelated to the block height7.
The afferentation theory proposes that tonic sensory and muscle-spindle activity modulate cerebral activity and maintain a state of wakefulness8, 9. Activation of the EEG in animals after increasing muscle afferent activity8, 9 is associated with an increase in cerebral blood flow that exceeds metabolic demand9. On the other hand, temporary peripheral denervation decreases the excitability of cuneate nucleus in the brainstem10, while acute block of tonic retinal discharges produced synchronization of the cortical EEG, which was otherwise desynchronized11. More recent findings suggest that a spinal depressant action of isoflurane12 or lidocaine13 on ascending somatosensory transmission can modulate reticulo-thalamo-cortical arousal mechanisms. Accordingly, a decrease in tonic afferent input, as during neuraxial anaesthesia, would be expected to reduce the level of consciousness.
As with most cranial nerves14, secondary auditory neurons, travelling in the medial lemniscus toward the inferior colliculus, are anatomically and functionally related to the reticular formation15. These connections form the anatomical background for an extensive interaction between the auditory pathway and reticular formation16, 17-18. In addition, the nucleus of inferior colliculus, an obligatory station of the auditory pathway, receives somesthetic input from the spinothalamic tract and medial lemniscus19 ,20, while it appears to induce cortical activation by acting through both cholinergic and serotonergic systems21. Thus, neuronal events such as reduced afferent somatosensory transmission that could affect the excitability of reticular formation would change the character of the brainstem auditory response.
General anaesthetics22-25 and sedatives26 have a small but significant effect on brainstem auditory evoked potentials (BAEP) in humans. This effect is mainly associated with increased latencies of the auditory responses; however, its exact mechanism is not fully elucidated. We believe this impairment in brainstem conduction is partly the result of indirect anaesthetic-induced deafferentation of somatosensory transmission. We tested the hypothesis that epidural anaesthesia with chloroprocaine produces block-dependent sedation that is associated with an impairment of brainstem auditory evoked responses.
With approval from the University of Louisville Human Studies Research Committees and informed consent, we studied 11 healthy volunteers. None had taken any drug acting on the central nervous system for at least 24 hours before the study. Age was restricted to between 18 and 45 years.
Volunteers fasted for at least 8 hours before the study. All the standard monitors, the bispectral index (BIS) of the EEG (A-2000, XP platform, Aspect Medical Systems, Inc., Newton, MA, USA), and brainstem auditory evoked potentials (BAEP) monitoring were applied to the participating volunteers, who were resting supine in a warm and quiet environment. An 18-g catheter was inserted into an antecubital vein of the non-dominant arm, and a bolus of 500 ml of Ringer's lactate solution was administered.
An epidural catheter was inserted into the L3-4 interspace and a 15-ml bolus of 2% 2-chloroprocaine was administered after a 15-minute control period. Additional 10-ml boluses were given every 10 minutes until a sensory block near T5 was produced or a total chloroprocaine dose of 1 g was given. The extension of the block was assessed at 10-minute intervals till its total regression and the clinical recovery of all the sensory functions. Volunteers were maintained normothermic throughout the study using forced-air covers.
Bispectral index (BIS), observer's assessment alertness/sedation (OAA/S) score, verbal rating score (VRS) for sedation, BAEP and block level were evaluated (in that order) and recorded 10 minutes before (baseline) and 10 minutes after the epidural injection of 2-chloroprocaine, and at 10-minute intervals thereafter. An investigator blinded to the extent of the sensory block performed all the assessments. The the trial did not exceed 5 hours.
Demographic and morphometric characteristics of the volunteers, as well as the total dose of chloroprocaine administered during the trial were recorded. Heart and respiratory rate, end-tidal CO2, and pulse oximeter saturation were measured continuously and recorded at 10-min intervals. Blood pressure was determined non-invasively every 5 minutes and recorded at 10-min intervals with the other measurements. Tympanic membrane temperature was measured at 30-min intervals throughout the study.
Sensory block was evaluated at baseline and at 10-min intervals after the first epidural administration of chloroprocaine, using the response to ice and pinprick, not only bilaterally but also to the cephalad and caudal directions. The total number of blocked dermotomes was recorded for further analysis. At 10-minute intervals, BIS, sedation scores, BAEP, and block level were evaluated and recorded in that order. During the assessment of the electrophysiological indices, volunteers were advised to relax and keep their eyes closed.
Bispectral index data were gathered using four sensors arranged in frontal-temporal montage after mild abrasion of the skin. Impedance of the sensors was evaluated at 10-min intervals and kept less than 5 kΩ. The smoothing window for BIS was set at 30 sec. The volunteers were instructed to close their eyes and lowest BIS value in a 3-min monitoring period was recorded for analysis. During this observation period, special care was taken to maintain a signal quality index (SQI) higher than 50%. Sedation level was assessed by a blinded investigator using the responsiveness component of the OAA/S score27 (Table 1). The volunteers simultaneously evaluated their own degree of sleepiness using a verbal rating score for sedation/sleepiness (VRS, 0 = wide awake to 10 = asleep).
The Neurodiagnostics Laboratory of the University of Louisville Hospital participated in the evaluation of the brainstem auditory evoked potentials (BAEP). Brainstem responses were obtained from gold cup electrodes placed on the vertex (Cz) and at right and left ear lobes (A1, A2) with a ground electrode placed at mid-forehead (FPz). Two thousands broad-band binaural click stimuli were delivered by the auditory evoked potential module of a Nicolet Viking Select monitor via earphones inserted in the external aural canals (Nicolet Biomedical Inc., Madison, WI). A stimulus rate of 11.11 Hz was used at an intensity of 70 dB above the hearing threshold. Generated stimuli were 100-μs-long square waves of rarefaction polarity. The recording was triggered at stimulus onset and continued for 10 msec. The monitor used a 50- to 3,000-Hz band pass and a 100-ms time base. Each set of 2,000 stimuli was repeated and both traces were superimposed to ascertain the reproducibility of the waves. Masking was used in the non-test ear at 35 dB below the click intensity used in the test ear. This condition allowed well-defined waves to be obtained for measuring the absolute latencies and amplitudes of waves I, III, and V. Interpeak latencies I-V, I-III, and III-V were also obtained. The evaluations of BAEP and detection of the various waves were done by an experienced laboratory technician who was not aware of the study hypothesis, outcomes, or block level.
The volunteers' demographic and morphometric characteristics were summarized. Mean arterial pressure, heart rate, respiratory rate, core temperature, oxygen saturation, and end-tidal CO2 measurements were averaged over the study period and presented as means ± standard deviations for each volunteer separately.
To measure the strength of the association between level of sensory block and the different measures of sedation — VRS, OAA/S score, and BIS — the prediction probability PK, was calculated. PK, is the probability that an indicator correctly predicts the depth of sedation.28 The PK in this case is thus the estimate of the probability that sensory block level will correctly predict sedation depth as expressed by VRS, OAA/S, or BIS. An indicator with a perfect predictive ability has a PK value of 1.0, whereas an indicator that performs no better than chance has a PK value of 0.5. In addition, values of VRS and BIS at the same block level were compared in the same subject between the evolution and regression phase of epidural anaesthesia, using paired t-tests. The purpose of this analysis was to examine the possible confounding effect of systemically absorbed chloroprocaine on sedation depth.
The level of sedation expressed by VRS and BIS was graphically presented as a function of sensory block. The latter was categorized in five 5-dermotome groups (0, 1-5, 6-10, 11-15, 16-21).
The level of sensory block, VRS and OAAS scores, and BIS were correlated with the latencies and amplitudes of the different BAEP using the Spearman rank correlation test. Correlations were first calculated for each volunteer and then averaged across all volunteers. Since many correlations were calculated, we considered P < 0.01 to be statistically significant. In addition, for the latencies of the three main BAEP — I, III, and V — a graph was constructed to show the change in that parameter between baseline (no block) and maximum block extension (maximum block) for each volunteer. In the case that “maximum block” was maintained for more than one assessment, the latency data were averaged within the volunteers. Pseudo-paired t-tests were used to compare BAEP values between the above two block states.
Results are expressed as means ± standard deviations, unless otherwise noted; differences were considered statistically significant when P < 0.05, unless otherwise noted.
Volunteers' demographic and morphometric characteristics are summarized in Table 2, along with total chloroprocaine dose, core temperature, and circulatory and respiratory function. Volunteer No 3 (Table 2) had an episode of hypotension (systolic blood pressure of 75 mmHg) that lasted for about three minutes and resolved after an infusion of 500 ml of fluid. Otherwise, all hemodynamic and respiratory parameters remained essentially unchanged during the trial. Poor signal quality (SQI > 50 %) did not allow us to collect useful BIS data for further analysis in 3 of 11 volunteers.
Prediction probability, PK, estimates showed that the level of sensory block predicted sedation depth as measured by VRS and the OAA/S score (Table 3). The PK value of sensory block level to predict BIS was small, but still significantly different from 0.50 (0.50 indicates no association). Paired analysis showed that sedation depth, as evaluated by VRS and BIS, was the same between the evolution and regression of the block for the same block level. The level of sedation, expressed by VRS and BIS, is presented graphically as a function of sensory block in Figure 1.
Spearman rank correlation tests showed statistically significant associations between the VRS and block level and the latency of wave III from the BAEP (Table 4). The correlation of wave V latency did not reach statistical significance (P > 0.01). Graphical presentation and analysis also showed a significant increase in the latency of wave III at maximum block extension state compared to baseline (No Block, P < 0.05), whereas latencies of I and V remained unchanged (Fig. 2).
Experimental evidence suggest that various types of anaesthetics indirectly suppress cortical responses to painful or other ascending stimuli by an action on the spinal cord29. The mechanism for this effect is not completely understood, but compelling evidence exists that spinal cord anaesthesia, through a decrease in the ascending somatosensory transmission, depresses the activity of reticulo-thalamo-cortical mechanisms that regulate arousal12, 13. On the other hand, acute sensory deafferentation in various animal models is associated with reduced activity of the midbrain reticular formation and cerebral cortex10, 11. In humans, epidural and spinal anaesthetics depress consciousness and reduce anaesthetic requirements for different anaesthetic endpoints1-3, 5 7. However, resetting the brain's arousal system by deafferentation of sensory input implies a dose-dependent effect — which has not previously been clearly demonstrated.
Sensory block levelduring epidural anaesthesia significantly predicted sedation depth in our volunteers, as measured by both objective and subjective methods. Loss of afferent sensory input during epidural anaesthesia was associated with a downregulation of brainstem neural activity as this was evaluated using auditory evoked potentials (BAEP). The latency of BAEP III showed a significant correlation with the level of sensory block. Wave III is generated by neural elements around the nucleus of inferior colliculus, which receives somesthetic input from the spinothalamic tract and medial lemniscus19, 20, while it appears to induce cortical activation by acting through both cholinergic and serotonergic systems21. Auditory circuits in the brainstem interact with reticular nuclei providing the anatomical basis for the defence alertness reaction, sensory information control, and even sleep regulation17-18.
Taken into account the very limited BAEP response to general anaesthetics22-25, even at anaesthetic depths associated with EEG burst suppression30, we believe that the significant correlation of sensory block with the latency of brainstem potential III involves a selective effect of sensory deafferentation on the activity of the midbrain region inside and in the vicinity of inferior colliculus. Despite the fact that sensory input (including auditory stimulation) from the unblocked areas of the body might have counteracted any deafferentation effect on BAEP in unmedicated volunteers, brainstem conduction showed a significant positive correlation with the block level and sedation depth. These findings support the original hypothesis that sensory block affects, in a dose-dependent manner, brainstem transmission.
The effect of epidural anaesthesia on BAEP has been evaluated previously by others31 who found that lidocaine epidural anaesthesia prolonged the peak latencies and delayed brainstem transmission. However, a similar pattern of BAEP inhibition after intravenous lidocaine administration32 suggested that systemic absorption rather than anaesthetic block itself was responsible for the effect on brainstem neural activity. To avoid the potential effect of systemically absorbed anaesthetic on BAEP, we used chloroprocaine which is metabolized by plasma cholinesterase and has an elimination half-life ranging from 10 seconds to 3 minutes33, 34. Using multiple small epidural boluses enabled us to assess the sensory block level during its evolution, as well as its regression. Comparison of sedation depth at the same block between those two phases of the anaesthetic did not reveal any difference. Thus, despite our lack of a placebo-controlled experimental condition, we believe that repetitive assessments of the block level, during both evolution and regression of the sensory block, reduced the likelihood that systemic chloroprocaine or elapsed study time confounded our findings.
Since the activity of cholinesterase in the cerebrospinal fluid (CSF) is greately diminished compared to the plasma35, there still remains a theoretical possibility that part of the chloroprocaine absorbed into the CSF could flow in the cephalad direction and exert its effects directly on the brainstem. However, in that case, a rather homogeneous than selective effect on BAEP would be a more likely outcome. Our findings do not support this possibility.
Responsiveness of the volunteers was maintained at an OAA/S score of 4 or 5. These data are in agreement with previous findings7. Prediction probability analysis showed that sensory block level predicts OAA/S score with similar strength as VRS. PK is a nonparametric measure of association that is recommended as a performance measure for anaesthetic depth indicators and observed depth scales of any degree of coarseness or fineness28. It is important to understand that the PK statistic is not prediction as we typically think of it. Instead, PK is the probability that for any two data pairs selected from the data set, the order of one variable (e.g., block level) will correctly “predict” the order of the other variable (e.g., VRS). In a sense, the PK value is the proportion of times that the higher VRS score is paired with the higher block level. We believe that the use of self-reported sedation did not introduce bias in our results because the volunteers were not aware of the purpose of the study or the basic hypothesis. Consistent with this theory, PK values of sensory block level were identical for the VRS and OAA/S score.
The bispectral index predicted sedation depth as defined by the VRS and OAA/S score. The PK value of sensory block level for BIS was lower than for the other sedation depth indicators; however, it was significantly different from 0.50 which is the PK value indicating no association. This is a new finding and contrasts with results of a previous placebo-controlled trial which concluded that BIS was an insensitive indicator of sedation and was affected only 60 minutes after the injection of spinal anaesthetic7. This late sedation was ascribed to delayed cephalad spread of the local anaesthetic. On the other hand, in a non-randomized trial, the same investigators found that the largest deviation of BIS from the baseline value occurred 30 minutes after spinal injection and coincided with the highest block level7.
In conclusion, we found that sedation during epidural anaesthesia depends on the level of sensory block and is associated with detectable alterations in the brainstem auditory evoked responses. Sensory deafferentation during neuraxial anaesthesia may modulate CNS alertness through mechanisms related to brainstem neural activity.
Supported by NIH Grant GM 061655 (Bethesda, MD), the Gheens Foundation (Louisville, KY), the Joseph Drown Foundation (Los Angeles, CA), and the Commonwealth of Kentucky Research Challenge Trust Fund (Louisville, KY). Mallinckrodt Anesthesiology Products, Inc. (St. Louis, MO) donated the thermocouples we used. None of the authors has any personal financial interests related to this work.
We greatly appreciate the assistance of Ozan Ak?a, M.D. and Edwin B. Liem, M.D. We thank the Neurodiagnostics Laboratory at the University of Louisville for their technical assistance with brainstem auditory evoked potentials recording. We also thank Warren Smith, Ph.D., (California State University) who has provided the macro for calculation of the prediction probability (PK). And finally, we thank Harvey L. Edmonds, Ph.D., (University of Louisville) and Joseph F. Antognini, M.D., (University of California at Davis) for their critical comments, and Nancy Alsip, Ph.D., for her editorial assistance.