OBJECTIVE: Artefact detection is an essential feature of automatic EEG monitoring systems used in anaesthesia. Clinical experience indicates that Narcotrend monitoring (MonitorTechnik, Bad Bramstedt, Germany, version 4.0) excludes more EEG epochs because of artefacts than bispectral index monitoring (BIS, Aspect Medical Systems, Newton, MA, version XP). Whether this increased exclusion of epochs is justified has not been investigated yet. METHODS: Eighteen adult patients undergoing radical prostatectomy were investigated. Induction of anaesthesia was performed with a fentanyl bolus and a propofol infusion. Additionally, following intubation patients received 15 ml bupivacaine 0.5% epidurally. After a waiting period of 45 min depth of anaesthesia was varied two times by increasing and decreasing propofol concentrations. Narcotrend index, BIS values and calculated propofol effect site concentrations were automatically recorded at intervals of 5 s. We tested the hypothesis whether exclusion of artefacts detected by the Narcotrend monitor would possibly improve the prediction probability of the BIS monitor, justifying the necessity of artefact suppression. RESULTS: Simulated propofol effect site concentrations ranged from 2 microg/ml to 6 microg/ml. The Narcotrend monitor excluded a significantly higher percentage of epochs because of artefact detection (12.6 + 1.0%) than the BIS monitor (0.4 +/- 0.1%). The performance of BIS as an indicator of predicted propofol effect site concentrations did not differ when including (P(K) = 0.86 +/- 0.05) or excluding (P(K) = 0.85 +/- 0.04) the data pairs where Narcotrend monitor but not BIS monitor indicated an artefact. Artefacts were evenly distributed over the investigated range ofpropofol effect site concentrations. CONCLUSION: Exclusion of data pairs that were detected as artefacts by Narcotrend but not by BIS did not change the performance of bispectral index as an indicator of propofol effect site concentration.
OBJECTIVE: Artefact detection is an essential feature of automatic EEG monitoring systems used in anaesthesia. Clinical experience indicates that Narcotrend monitoring (MonitorTechnik, Bad Bramstedt, Germany, version 4.0) excludes more EEG epochs because of artefacts than bispectral index monitoring (BIS, Aspect Medical Systems, Newton, MA, version XP). Whether this increased exclusion of epochs is justified has not been investigated yet. METHODS: Eighteen adult patients undergoing radical prostatectomy were investigated. Induction of anaesthesia was performed with a fentanyl bolus and a propofol infusion. Additionally, following intubation patients received 15 ml bupivacaine 0.5% epidurally. After a waiting period of 45 min depth of anaesthesia was varied two times by increasing and decreasing propofol concentrations. Narcotrend index, BIS values and calculated propofol effect site concentrations were automatically recorded at intervals of 5 s. We tested the hypothesis whether exclusion of artefacts detected by the Narcotrend monitor would possibly improve the prediction probability of the BIS monitor, justifying the necessity of artefact suppression. RESULTS: Simulated propofol effect site concentrations ranged from 2 microg/ml to 6 microg/ml. The Narcotrend monitor excluded a significantly higher percentage of epochs because of artefact detection (12.6 + 1.0%) than the BIS monitor (0.4 +/- 0.1%). The performance of BIS as an indicator of predicted propofol effect site concentrations did not differ when including (P(K) = 0.86 +/- 0.05) or excluding (P(K) = 0.85 +/- 0.04) the data pairs where Narcotrend monitor but not BIS monitor indicated an artefact. Artefacts were evenly distributed over the investigated range ofpropofol effect site concentrations. CONCLUSION: Exclusion of data pairs that were detected as artefacts by Narcotrend but not by BIS did not change the performance of bispectral index as an indicator of propofol effect site concentration.