Effects of dexmedetomidine on performance of bispectral index as an indicator of loss of consciousness during propofol administration

OBJECTIVE: The performance of bispectral index (BIS) for the measurement of the sedative depth when dexmedetomidine is administered in propofol anaesthesia and sedation has not yet been established. This study evaluated the effects of adjunctive dexmedetomidine on the accuracy of BIS to predict loss of consciousness (LOC) and BIS values predicting LOC during propofol administration. METHODS: In this randomised, double-blind and placebocontrolled trial, 225 patients scheduled for general anaesthesia were assigned to one of three groups. Dexmedetomidine 0.5 and 1.0 μg kg–1 were intravenously infused for 15 minutes in the dexmedetomidine 0.5 and 1.0 μg kg–1 groups, and saline was infused in the control group. Propofol was administered as an effect-site target-controlled infusion after completion of dexmedetomidine infusion. Patients in each group were allocated to five subgroups in which the concentration of propofol was set at 0, 1, 2, 3 and 4 μg ml–1, respectively. Three minutes after propofol administration, the BIS values and Observer’s Assessment of Alertness/Sedation (OAA/S) scores were recorded. RESULTS: There were no significant differences in the prediction probability of BIS for detecting LOC in the three groups. At the time of LOC, BIS50 values were 71.1 and 71.4 in the dexmedetomidine 0.5 and 1.0 μg kg–1 groups, respectively, which were significantly larger than the BIS50 of 63.2 in the control group. CONCLUSIONS: The ability of BIS to predict LOC is not influenced by dexmedetomidine during propofol administration, but BIS values are enhanced at the time of LOC.


Introduction
Dexmedetomidine is an α 2 -adrenegic agonist, with sedative, analgesic and sympatholytic properties [1,2].It has been demonstrated that dexmedetomidine is an efficacious and safe adjuvant in general anaesthesia [3,4].When dexmedetomidine is administrated in propofol anaesthesia, it can exert an anaesthetic-sparing effect and stabilise intraoperative haemodynamics [5][6][7].The bispectral index (BIS) is a derived multifactorial electroencephalographic (EEG) parameter in which the outcome of the EEG analysis is represented a unitless value between 0 (isoelectric EEG) and 100 (alert and oriented).BIS monitoring is used by clinical anaesthetists to titrate anaesthetic agents in order to maintain loss of consciousness (LOC) and prevent intraoperative awareness [8,9].However, the performance of BIS for detecting LOC appears to depend on the anaesthetic regimen used.For example, BIS is considered to be a reliable method of monitoring sedative depth and can accurately predict LOC in propofol anaesthesia, but not nitrous oxide and ketamine anaesthesia.Furthermore, BIS values for detecting LOC change when other anaesthetics are used in combination with propofol [10][11][12][13][14][15].Dexmedetomidine, acting on the α 2 -adrenoceptors in the locus coeruleus, produces physiological sleep-like phenomena in the EEG and a characteristic rousable sedation, which is markedly different from that of other sedatives such as the gamma-aminobutyric acid (GABA) agonist propofol [16][17][18].It is unclear whether BIS can accurately predict LOC and whether BIS values predicting LOC change when dexmedetomidine and propofol are coadministrated.Recently, Kasuya et al. [19] showed that BIS could accurately predicte LOC with dexmedetomidine monoinfusion, but the cut-off BIS values for detecting LOC were lower than with propofol monoinfusion.We hypothesised that adjunctive dexmedetomidine changes BIS values for detecting LOC during propofol administration.The purpose of our study was to assess the effects of dexmedetomidine on the performance of BIS for detecting LOC during propofol administration.

Patient selection
The study was approved by the ethics committee of the Affiliated People's Hospital of Jiangsu.Written informed consent was obtained from 225 patients with ASAⅠor Ⅱ, aged 20-50 years and weighing 41-68 kg, who were scheduled for general anaesthesia.Exclusion criteria included body mass index >30 kg/m 2 , bradycardia, atrioventricular block, neurologic disorder and recent use of psychoactive medication, including alcohol.

Study protocol
Atropine 0.5 mg was injected intramuscularly as preanaesthetic medication.An 18G catheter was inserted into large forearm vein for fluid and drug administration.Lactated Ringer's solution was infused at a rate of 10 ml kg -1 h -1 and oxygen was delivered via facemask with a flow of 5 L min -1 .Heart rate, noninvasive blood pressure and pulse oxygen saturation were monitored at 3-minute intervals using Datex-Ohmeda S/5 monitor.BIS was derived from the frontal electroencephalogram and calculated by
an Aspect Vista TM monitor (version 3.2, Aspect Medical System, Inc.) using BIS sensor electrodes.Patients were randomly allocated to three groups according to computergenerated random codes that were enclosed in sequentially numbered opaque envelopes: a control group (dexmedetomidine 0 µg kg -1 ), a dexmedetomidine 0.5 µg kg -1 group and a dexmedetomidine 1.0 µg kg -1 group (75 subjects in each) (fig.1).Dexmedetomidine 0.5 and 1.0 µg kg -1 were intravenously infused for 15 minutes in the 0.5 and 1.0 µg kg -1 groups, respectively, and saline was infused in control group.After dexmedetomidine and saline infusion were completed, propofol was administered as an effect-site target-controlled infusion (TCI).Patients in each group were randomly allocated to five subgroups based on computer-generated random codes that were enclosed in other sequentially numbered opaque envelopes: a P 0 subgroup, a P 1 subgroup, a P 2 subgroup, a P 3 subgroup, a P 4 subgroup (15 subjects in each), for which the effective-site concentration of propofol was set at 0, 1, 2, 3 and 4 μg ml -1 , respectively (fig.1).After patients entered into operating room, the two allocation envelopes were opened by one anaesthetist who knew which group and sub-group the patients were assigned to and administered the corresponding dose of dexmedetomidine and concentration of propofol.The syringe pump and TCI pump were shielded from another anaesthetist who did not know the grouping and recorded the BIS values and Observer's Assessment of Alertness/Sedation (OAA/S) scale (table 1) three minutes after propofol administration.Patients were considered to be conscious if their OAA/S scores were 5, 4, or 3 and considered to be unconscious at scores of 2, 1, and 0. Propofol was administered using a commercially available TCI pump (the Slgo CP-600 Infusion System, China), which uses Marsh's pharmacokinetic model.This system displays predicted plasma and effect-site concentration.The effectsite concentration of propofol was computed to yield a time-to-peak effect of 1.6 minutes after bolus injection and clinically confirmed by Struys et al. [20].Therefore, we were able to obtain a steady-state effect-site concentration for propofol when recording BIS values and OAA/S scores.If mean arterial pressure was more than 130% or less than 70% of baseline value or systolic pressure was less than 90 mm Hg, urapidil and phenylephrine were administered.When heart rate was less than 55 bpm, atropine was used.Respiratory depression was treated with assisted ventilationvia facemask.The primary endpoint of the study was BIS values for detecting LOC and the secondary endpoint was the effect-site concentration of propofol for LOC.

Statistical analysis
Data was analysed using SPSS statistical package version 17.0 for Microsoft Windows.The significance level was set at 5% unless otherwise reported.BIS values were subjected to two-way analysis of variance (ANOVA; propofol × dexmedetomidine).When the ANOVA revealed a significant main effect or interaction between the main factors, differences in subgroup means were tested using Fisher's protected least-significant difference test.Spearman rankorder correlation was used for the correlation analysis of BIS and OAA/S.The ability of BIS to detect LOC for each

Original article
Swiss Med Wkly.2013;143:w13762 Swiss Medical Weekly • PDF of the online version • www.smw.ch group was evaluated using prediction probability (P k ), as developed by Smith et al. [21].A P k of 1 would mean that BIS can predict anaesthetic depth perfectly.A P k of 0.5 would mean BIS is useless for measuring anaesthetic depth.By applying Probit analysis, the BIS values and the effect-site concentration of propofol at which 50% (ED 50 ) and 95% (ED 95 ) of patients lost consciousness for each group were calculated.EC 50 and EC 95 represented the effect-site concentration of propofol at which 50% and 95% of patients lost consciousness.BIS 50 and BIS 95 represented the BIS values at which 50% and 95% of patients lost consciousness.Receiver operating characteristic (ROC) curve analysis was used to evaluate the capacity of BIS to predict LOC and obtain the cut-off BIS values for LOC at which the sum of sensitivity and specificity were highest.In addition, the cut-off BIS values resulting in 100% sensitivity were calculated.A Student t-test with Bonferroni correction was used to evaluate whether the correlation coefficient and area under ROC curve, P k and ED 50 and ED 95 were different between groups.The significance level for this analysis was set at 0.0167.

Results
The demographics of the 225 patients in the three groups and fifteen subgroups are shown in table 2. There were no relevant demographic differences found between groups or subgroups (p >0.05).As expected, propofol and dexmedetomidine significantly decreased BIS.A significant propofol × dexmedetomidine interaction indicated that dexmedetomidine exerted an antagonistic effect on BIS with propofol (p <0.01).Post hoc analysis indicated that dexmedetomidine 0.5 and 1.0 µg kg -1 decreased BIS values at every propofol concentration (fig.2).Scatter plots of BIS values and OAA/S scores in all groups are shown in figure 3.In all groups, a monotonic decrease in BIS was observed at decreasing OAA/S levels.The correlation coefficients (with standard errors) of BIS and OAA/S in the three groups were 0.931(0.012),0.903(0.025)and 0.939(0.013),respectively.There were no significant differences in correlation coefficient between the groups (p >0.0167).The ability of BIS to predict LOC as represented by the P k values is shown in table 3. P k values were similar in all groups (p >0.0167).The relationships between the probability of LOC and effect-site concentration or BIS are shown in figures 4 and

Original article
Swiss Med Wkly.2013;143:w13762 5. The EC 50 values of the dexmedetomidine 0.5 and 1.0 µg kg -1 groups were significantly lower than that of the control group, and the EC 50 of the dexmedetomidine 1.0 µg kg -1 group was significantly lower than that of the dexmedetomidine 0.5 µg kg -1 group (p <0.0167).The BIS 50 values of the dexmedetomidine 0.5 and 1.0 µg kg -1 groups significantly increased compared with that of the control group (p <0.0167) and there was no difference in BIS 50 between the dexmedetomidine 0.5 and 1.0 µg kg -1 groups (p >0.0167) (table 4). Figure 6 shows ROC curves for BIS to detect LOC.There was no significant difference in ROC curve areas between groups (p >0.0167) (table 3).Table 5 shows the cut-off BIS values for detecting LOC at which the sum of the sensitivity and specificity were the highest, and the cut-off values at which sensitivity was 100%.

Discussion
The current study demonstrated that the ability of BIS to predict LOC was not influenced by dexmedetomidine during propofol administration, but BIS values were enhanced at the moment of LOC.
The prediction probability (P k ) and area under ROC curve provide a good comparison of the ability of BIS to measure the degree of sedation between groups.In the three groups, The relationship between the probability of loss of consciousness (LOC) and the propofol effect-site concentration.
(DEX = dexmedetomidine.) P k and area under the curve for LOC were similar, indicating identical accuracy of BIS to predict LOC with and without adjunctive dexmedetomidine.P k and area under the curve were more than 0.9, which indicated BIS monitoring was an excellent indicator of LOC during dexmedetomidine and propofol coadministration.BIS thresholds for LOC depend on the combinations of different anaesthetics used [22].For example, when opioids were used in combination with propofol, BIS values for The relationship between the probability of loss of consciousness (LOC) and the bispectral index.(DEX = dexmedetomidine.) Receiver operating characteristic (ROC) curve for loss of consciousness detected by BIS.(DEX = dexmedetomidine.)detecting LOC increased [14,15].Although dexmedetomidine decreased BIS in this study, LOC was reached at higher BIS values when adding dexmedetomidine to propofol administration.BIS 50 for LOC was 63.2 during propofol administration, whereas BIS 50 increased to 71.1 and 71.4 after administering dexmedetomidine 0.5 and 1.0 µg kg -1 , respectively.One possible explanation is that dexmedetomidine obviously decreases the effect-site concentration of propofol for LOC.Dutta et al. [23] reported that with a maintained dexmedetomidine plasma concentration of 0.66 ± 0.080 ng ml -1 , which was reached with larger dexmedetomidine dose than used in our study, the propofol EC 50 levels required for Ramsay sedative scores of >4 and 5 were reduced by 64% and 68%, respectively.Peden et al. [24] showed a 30% decrease in the propofol EC 50 for LOC by dexmedetomidine 0.63 µg kg -1 premedication during propofol and alfentanil induction of anaesthesia.In our study, dexmedetomidine induced a leftward shift of the propofol dose-response curve for LOC and decreased the effect-site concentration of propofol in a dosedependent matter.Propofol EC 50 was 60% less in the presence of 1.0 µg kg -1 dexmedetomidine and 35% less with 0.5 µg kg -1 dexmedetomidine than without dexmedetomidine.Although the effect of dexmedetomidine on EC 95 did not reach the level of significance, a clear declining trend could be observed.Another possible reason is that BIS did not detect the sedative interaction of the anaesthetics [25,26].Dutta et al. [23] reported that dexmedetomidine and propofol produced an additive interaction for achieving sedative endpoints, whereas our study showed antagonism on BIS between dexmedetomidine and propofol.The inhibitory effects of adjunctive dexmedetomidine with propofol on BIS were less than the cumulative effects of dexmedetomidine and propofol on BIS.
Our results contradict the findings of Kasuya et al. [19], who reported that BIS values were lower during dexmedetomidine than propofol monoinfusion at comparable OAA/S scores and the cut-off values for LOC with dexmedetomidine were less than with propofol.Theoretically, BIS values for detecting LOC decrease when dexmedetomidine is used in conjunction with propofol.Our study did not show this, possibly because of the easily rousable condition characteristic of dexmedetomidine sedation.Subjects with low BIS values can be roused to a state where meaningful responses to calling and shaking stimulation used by OAA/S score are seen and are thus given a high OAA/S score.The ROC curve for LOC was obtained by calculating sensitivity and specificity at every possible BIS cut-off point and was used to select optimal cut-off values.The values lay at the elbow of the curve and best differentiated wakefulness and unconsciousness.Adjunctive dexmedetomidine slightly enhanced the optimal BIS cut-off values.However, the highest sum of sensitivity and specificity did no obviously change, which meant the addition of dexmedetomidine had no influence on the ability to discriminate LOC at optimal BIS values.Drummond [27] expressed the opinion that an indicator of anaesthetic depth should have a minimum of 100% sensitivity if clinicians seek a specific numeric threshold that can be interpreted to mean LOC.The BIS cut-off values for LOC at which sensitivity was 100% increased when adding dexmedetomidine.An increase in specificity at the level of 100% sensitivity was observed following adjunctive dexmedetomidine.One limitation of this study is that we did not measure the plasma concentration of propofol.However, in clinical practice we use a target-controlled infusion device to predict, rather than high-performance liquid chromatography to determine, propofol concentration.The concentration predicted by the target-controlled infusion device is well correlated with measured concentration.In addition, because dexmedetomidine causes cardiac output and renal blood flow reduction, and a redistribution of cardiac output [2,28], it possibly affects distribution and excretion of propofol and the accuracy of predicted propofol concentration.But, Dutta et al. [23]    The effects of dexmedetomedine on bispectral index at every propofol effect-site concentration.Bars represent mean ± SD. * Significant differences compared with control group at the same propofol effect-site concentration (p<0.01).# Significant differences between dexmedetomidine (DEX) 0.5 and 1.0 μg kg -1 (p<0.01).＆ Significant differences between dexmedetomidine 0.5 and 1.0 μg kg -1 (p<0.05).

Figure 1
Figure 1Disposition of the study patients.

Figure 3
Figure 3Raw data for the bispectral index at every Observer's Assessment of Alertness/Sedation Scale score.(DEX = dexmedetomidine.)

Figures (large format) Figure 1
Figures (large format)

Figure 3
Figure 3Raw data for the bispectral index at every Observer's Assessment of Alertness/Sedation Scale score.(DEX = dexmedetomidine.)

Figure 4
Figure 4The relationship between the probability of loss of consciousness (LOC) and the propofol effect-site concentration.(DEX = dexmedetomidine.)

Figure 5
Figure 5The relationship between the probability of loss of consciousness (LOC) and the bispectral index.(DEX = dexmedetomidine.)

Table 1 :
Responsiveness scores of the modified Observer's Assessment of Alertness/Sedation scale.

Table 2 :
Demographic characteristics of the study groups and subgroups.

Table 3 :
Area under the receiver operating characteristic (ROC) curve and prediction probability (Pk) for loss of consciousness (LOC) detected by bispectral index (BIS) in the three groups.

Table 5 :
showed dexmedetomidine did not have a significant influence on propofol pharmacokinetics.Another limitation is that we used only the subjective OAA/S test to identify LOC.The OAA/S test has potential disadvantages.It requires some extrinsic stimulation such as calling, prodding, or shaking to evaluate the neurological status, which can affect subject's sedative status if sedation is not deep, and results are dependent on the assessor.This study was financially supported by Department of Anesthesiology, the Affiliated People's Hospital of Jiangsu University.Cut-off BIS values for detecting loss of consciousness at maximum level of sensitivity and specificity and at a level of 100% sensitivity.Swiss Med Wkly.2013;143:w13762 Swiss Medical Weekly • PDF of the online version • www.smw.ch Funding/potential competing interests: