Secondary Logo

Journal Logo

The relative effects of dexmedetomidine and propofol on cerebral blood flow velocity and regional brain oxygenation

A randomised noninferiority trial

Farag, Ehab; Kot, Michael; Podolyak, Attila; Argalious, Maged; Deogaonkar, Milind; Mascha, Edward J.; Xu, Zekun; Katzan, Irene; Ebrahim, Zeyd

European Journal of Anaesthesiology (EJA): November 2017 - Volume 34 - Issue 11 - p 732–739
doi: 10.1097/EJA.0000000000000662

BACKGROUND Dexmedetomidine constricts cerebral blood vessels without a concomitant reduction in cerebral metabolic oxygen consumption. Its safety as a sedative in patients with neurological diseases thus remains uncertain.

OBJECTIVE Our primary objective was to test the hypothesis that dexmedetomidine is noninferior to propofol as regards cerebral blood flow (CBF) velocity and brain oxygenation.

DESIGN Unblinded randomised trial.

SETTING Cleveland Clinic Hospital, Cleveland, from November 2010 to July 2013.

PATIENTS Forty-four patients scheduled for insertion of a deep-brain stimulating electrodes.

INTERVENTIONS Patients were randomised to receive either dexmedetomidine or propofol sedation during deep-brain stimulating electrode insertion.

MAIN OUTCOME MEASURES Intraoperative CBF velocity was measured with transcranial Doppler, and brain oxygenation was assessed with near-infrared spectroscopy. Noninferiority of dexmedetomidine to propofol was defined as a less than 20% difference in means.

RESULTS Twenty-three patients were given dexmedetomidine and 21 propofol. Baseline characteristics and operative management were similar in each group. Dexmedetomidine was noninferior to propofol on both CBF and brain oxygenation, confirming our primary hypothesis. For cerebral flood flow, the estimated ratio of means (dexmedetomidine/propofol) was 0.94 [90% CI: 0.84 to 1.05], P = 0.011 for noninferiority. For brain oxygenation, the estimated ratio of means was 0.99 [90% CI: 0.96 to 1.02], P < 0.001 for noninferiority. Superiority was not found for either primary outcome. Dexmedetomidine provided deeper sedation than propofol, with a difference of medians of 1 [90% CI: 0 to 2], P < 0.001 on the Observer's Assessment of Alertness/Sedation scale. No significant differences were observed in pulsatility index, cerebral perfusion pressure, number of hypertensive or apnoeic episodes.

CONCLUSION Regional brain oxygenation and CBF velocity are comparably preserved during dexmedetomidine and propofol sedation. Thus, the use of dexmedetomidine in patients with movement disorders appears reasonable.

TRIAL REGISTRATION The trial was registered at (NCT 01200433).

From the Department of General Anesthesiology, Cleveland Clinic (EF, MK, AP, ZE), Department of Neurosurgery, Ohio State University (MD), Department of Quantitative Health Sciences and Department of Outcomes Research, Cleveland Clinic (EJM, ZX), Cerebrovascular Center, Cleveland Clinic (IK)

Correspondence to Ehab Farag, MD, FRCA, Department of Anesthesiology, Cleveland Clinic, 9500 Euclid Ave., E-30, Cleveland, OH 44195, USA Tel: +1 216 444 7550; fax: +1 216 444 9247; e-mail:

Published online 19 June 2017

Back to Top | Article Outline


Deep-brain stimulation (DBS) improves motor symptoms of movement disorders such as Parkinson disease, essential tremor and dystonia. Several clinical trials are currently assessing the effects of DBS for neuropsychiatric disorders including depression, obsessive-compulsive disorder, epilepsy and chronic pain syndromes.1,2 The procedure involves implanting neurostimulation electrodes in subcortical targets, including the basal ganglia and diencephalon. The subthalamic nucleus and globus pallidum pars interna are commonly targeted for the management of the cardinal motor manifestations of Parkinson's disease, whereas the ventral intermedius nucleus of the thalamus is commonly targeted for managing tremor. Although DBS electrodes can be implanted under general anaesthesia,3 the most common technique is to implant them under conscious sedation. Keeping patients conscious during the procedure allows operators to use intraoperative responses to refine stereotactic targeting and assess the effects of macrostimulation on motor symptoms and possible side effects.

Propofol is the most commonly used sedative for deep-brain stimulating electrode insertion,4 but dexmedetomidine appears to be a reasonable alternative and may facilitate evaluation of intraoperative physiology. The advantages of dexmedetomidine over sedatives, such as benzodiazepines, propofol and opioids, include maintenance of normal sleep patterns, limited impairment of ventilation and good patient responsiveness,5 all of which facilitate communication between the surgical team and patient. However, the use of dexmedetomidine in patients with neurological diseases, especially those in neuro-intensive care, remains controversial because dexmedetomidine decreases cerebral blood flow (CBF) without decreasing the cerebral metabolic rate to oxygen (CMRO2).6,7 For example, dexmedetomidine reduces CBF without influencing CMRO2 in normoxic8 and hypoxic dogs.9 Limited human data suggest that dexmedetomidine impairs cerebral autoregulation both in volunteers10 and in patients with septic shock.11 Uncoupling of CBF and CMRO2 might promote adverse outcomes in patients with neurological diseases who are given dexmedetomidine, including those having deep-brain stimulator electrode implants.

We therefore questioned whether dexmedetomidine and propofol influence CBF velocity and brain oxygenation comparably. Specifically, we tested the hypothesis that dexmedetomidine is noninferior to propofol regarding both CBF velocity and regional brain oxygenation in the absence of other sedatives or anaesthetics during deep-brain stimulator electrode insertion for patients with movement disorders. Our secondary aims were to compare the effects of dexmedetomidine and propofol on sedation [assessed with the modified Observer's Assessment of Alertness/Sedation (OAA/S) scale],12 arterial oxygenation, the number of hypertensive episodes and the number of apnoeic episodes during insertion of deep-brain stimulating electrodes.

Back to Top | Article Outline


Ethical approval for this study (Ethical Committee N 10-715) was provided by the Ethical Committee (Institutional Review Board) of Cleveland Clinic, Cleveland, Ohio, USA (Chairperson Dr Alan Lichtin M.D.) on 03/09/2010, and written consent was received from each participating patient.

With Institutional Review Board approval and written consent from each participating patient, we conducted an unblinded randomised trial of the effects of dexmedetomidine and propofol on CBF and brain oxygenation in patients having a DBS inserted. The study started on 30 November 2010.

We included adults less than 80 years old who were American Society of Anesthesiology physical status 1 to 3 and were having DBS electrodes implanted for Parkinson's disease or essential tremor management. Patients were excluded if they had one or more of the following: history of isolated dystonia without Parkinson's disease, severe heart failure with an ejection fraction less than 30%, history of renal failure with serum creatinine concentration more than 2 mg dl−1, allergies to alpha−2 agonists or propofol or current use of alpha−2 agonists.

Patients were randomised 1 : 1 to dexmedetomidine or propofol sedation using random-sized permuted blocks without stratification. Randomisation was based on computer-generated codes that were kept in sequentially numbered opaque envelopes until just before use. To avoid bias in the assessment of patients, one investigator was responsible for preparing the study drug and programming the infusion pump, whereas a second was responsible for observing the patient and collecting the data were thus blinded to study drug allocation.

In patients assigned to dexmedetomidine, the drug was infused to target plasma concentrations of 0.8 to 1.0 ng ml−1, titrated to patient comfort: this corresponds to an infusion rate of about 0.4 to 0.5 μg kg−1 h−1, which is the commonly used sedation dose in clinical practice. Dexmedetomidine was administered to the desired target plasma concentration via a Harvard Apparatus Model 22 syringe pump controlled by STANPUMP software (Steve Shafer, Stanford University, Palo Alto, CA, USA) was used to run the infusion pump (Harvard Apparatus 22, Harvard Apparatus, South Natick, MA, USA).13 This software is freely available (, accessed January 2011) and uses a three compartment pharmacokinetic model and first-order intercompartmental transfer calculations to estimate input into and elimination from the central compartment.14 Patients assigned to propofol were given an infusion to achieve a plasma concentration of 2 to 3 μg ml−1, titrated to the patient's comfort: this corresponds to approximately 40 μg kg−1 min−1. The target plasma concentration range of propofol to achieve an adequate level of sedation was chosen according to a previous study.14 Target propofol levels were based on the pharmacokinetic model of Schnider et al.,15 again using STANPUMP software.

After electrode insertion and before intraoperative microelectrode testing, dexmedetomidine was reduced to target a plasma concentration of 0.4 ng ml−1 (corresponding to about 0.2 μg kg−1 h−1), whereas propofol was discontinued. Between microelectrode testing and macrostimulation, neither drug was given, so as to avoid interference with neurological–physiological testing. Between the time from insertion of the permanent electrodes to completion of the procedure the target plasma concentration of dexmedetomidine was set to 0.8 to 1.0 ng ml−1 and the propofol infusion was set to target a plasma concentration of 2 to 3 μg ml−1. The drug infusions were turned off at the end of the procedure and measurements in the post-anaesthesia care unit (PACU) took place after at least 1 h, when the patients were wide awake.

Patients were given supplemental oxygen at a rate of 2 to 3 l min−1 via a nasal cannula. Periods of SpO2 less than 94% were recorded and treated by increasing the supplemental O2 or decreasing sedative administration. Apnoeic episodes (absence of the end-tidal PCO2 waveform for more than 15 s) or two recordings of SBPs at least 140 mmHg within 5 min were treated as clinically appropriate. The hypertension was treated usually with 5 to 10 mg boluses of labetalol. If labetalol was insufficient, 10 to 20 mg of hydralazine was given as necessary to reduce blood pressure to an acceptable range.

Demographic and morphometric characteristics were recorded. Patients were monitored as per the American Society of Anesthesiologists recommendations: an arterial catheter was inserted only if clinically indicated.

CBF velocity in the middle cerebral artery (MCA) was measured with transcranial Doppler ultrasonography (TCD, pulsed wave 2-MHz Nicolet Pioneer TC 2020). The artery was viewed through the transtemporal window 45 to 65 mm below the skin surface. TCD flow velocities were measured before starting the procedure, at the highest and lowest doses of propofol or dexmedetomidine, and in the PACU. Cerebral perfusion pressure (CPP) was calculated from MCA flow velocity every time CBF was measured using the formula:


Pulsatility index is a measure of the variability of blood velocity in a vessel and was calculated as the difference between the peak systolic and minimum diastolic velocities divided by the mean velocity during the cardiac cycle.17

Near-infrared spectroscopy (NIRS), an index of brain oxygenation, was continuously monitored noninvasively at the capillary–venous level of brain tissue with a Fore-Sight (Casmed, Branford, Connecticut, USA) monitor and the result recorded at 2-s intervals. NIRS measures regional haemoglobin oxygen saturation (rSO2). Normal values in healthy young men breathing air at one atmosphere are 71 ± 6% (SD).18 An rSO2 less than 50 to 55% at baseline or a reduction of more than 20% are thought to represent serious cerebral ischaemia.19,20 We therefore considered a relative difference between groups in mean rSO2 of 20% (i.e. a noninferiority delta for the ratio of means of 0.80) as our noninferiority threshold.

All the measurements were recorded at five different times during the procedure as follows:

  1. The baseline measurement was performed on the morning of the procedure.
  2. The second measurement was at the first peak concentration of either dexmedetomidine or propofol.
  3. The third measurement was at the lowest concentration of either dexmedetomidine or propofol.
  4. The fourth measurement was performed when dexmedetomidine or propofol were recommenced after the implantation of the microelectrodes during the surgical closure period (second peak concentration).
  5. The fifth and last measurement was undertaken in the PACU to confirm a second baseline measurement.

Sedation was evaluated with the modified observer's assessment of alertness/sedation (OAA/S) scale, which ranges from 0 to 5 (0 = does not respond to noxious stimuli and 5 = responds to name spoken in normal tone).21

Back to Top | Article Outline


Primary outcomes

We tested the joint hypothesis22 that dexmedetomidine was noninferior to propofol both in terms of CBF velocity and brain oxygenation during DBS surgery. A relative difference of 20% or less was, a priori, deemed to be clinically unimportant. We therefore specified an a priori noninferiority ratio of means of 0.80 for each outcome, so that noninferiority would be claimed for an outcome if the mean for dexmedetomidine was not more than 20% lower (worse) than that of propofol.

Noninferiority on both outcomes was a priori deemed necessary and sufficient to claim dexmedetomidine comparable with propofol (i.e. superiority was not required on either outcome). Hence, no adjustment for having two primary outcomes was needed, and the overall type I error was maintained at 0.05 for the primary outcomes. Since we used one-tailed statistical tests, 90% confidence intervals were reported, with 5% alpha in the direction of interest. Following intention-to-treat, we assigned missing values in the outcome variable with the observed minimum (worst) value for the dexmedetomidine group and with the observed maximum (best) value for the propofol group. Both primary outcomes followed a log-normal distribution, so each analysis was conducted on the log-transformed scale.

The outcome variable for CBF was mean MCA velocity flow velocity (FVm), which was the average of velocities in the right and left MCAs recorded by transcranial Doppler at two points: first, at the first peak of study drug during the first infusion period (i.e. the second measurement in the above list); second, at the lower dose of dexmedetomidine or when the propofol was stopped (i.e. the third measurement period in the list above). We first assessed the treatment effect (i.e. difference in means) of dexmedetomidine versus propofol on log-transformed FVm data in a linear regression model, including baseline FVm as a covariate to gain precision. We then used the estimated treatment effect in a one-tailed test for noninferiority, as shown below.

Brain oxygenation values were estimated by NIRS and brain oxygenation was averaged across the first study drug infusion period. For analysis, the brain oxygenation outcome variable was the mean of the right and left sides of the brain oxygen saturations. We first assessed the treatment effect (i.e. difference in means) of dexmedetomidine versus propofol on log-transformed brain oxygenation data in a simple linear regression model. We then used the estimated treatment effect in a 1-tailed test for noninferiority, as explained below.

Noninferiority testing: Using the estimated treatment effects from the regression models (see above), we assessed noninferiority of dexmedetomidine to propofol on each outcome using 1-tailed noninferiority tests with an a-priori noninferiority ratio of means23 of means of 0.80. Thus, the alternative hypothesis was that the geometric mean (since on the log-scale) for dexmedetomidine was no more than 20% lower than that of propofol for each of CBF and brain oxygenation. The 1-tailed t test statistic was of the form



were the estimated least square means (from the respective model) of the log-transformed outcome for the dexmedetomidine and propofol groups, respectively, and SE was the estimated standard error for the treatment effect (log scale). The null hypothesis was rejected if T was significantly greater than zero at the 0.05 level.

Back to Top | Article Outline

Secondary outcomes

We evaluated the effect of dexmedetomidine and propofol on the pulsatility index, the CPP, the OAA/S, the number of hypertensive episodes during DBS surgery and the number of apnoeic episodes during DBS surgery. We defined the outcome variable for pulsatility index, CPP and OAA/S as the corresponding measurements at the first peak during DBS surgery.

Separate linear regression models were built (after log transformation) to estimate the effect of dexmedetomidine versus propofol on the pulsatility index and CPP while including baseline measurements as a covariate. Furthermore, the two groups were compared on OAA/S, number of hypertensive episodes and number of apnoeic episodes using student t test and Wilcoxon rank-sum test as appropriate.

For the statistical analysis on all five secondary outcomes, we conducted two-sided tests for superiority. A Bonferroni correction was used to adjust for multiple comparisons to control the overall type I error at 0.05 for secondary outcomes. Thus, 99% confidence intervals were reported; the significance criterion for the five secondary outcomes was P < 0.01.

Back to Top | Article Outline

Sample size considerations

The study was designed to include 44 patients as this provided a 90% power at the 0.05 significance level to detect noninferiority of dexmedetomidine compared with propofol using a noninferiority ratio of means of geometric means of 0.80 and, based on our experience and the literature, assuming a coefficient of variation of 25% for each of the two primary outcomes. Both outcomes were expected to follow a log-normal distribution. The actual power was somewhat higher since smaller coefficient of variations were observed than had been planned.

SAS software version 9.3 (SAS Institute, Cary, North Carolina, USA) was used for all analyses.

Back to Top | Article Outline


Of the 145 patients assessed for eligibility, 101 were excluded for failure to meet inclusion criteria (Fig. 1). Thus, 44 patients were randomised to either dexmedetomidine (n = 23) or propofol (n = 21) sedation. The groups were well balanced on demographic and baseline characteristics, Table 1.

Fig. 1

Fig. 1

Table 1

Table 1

Noninferiority of dexmedetomidine to propofol was found for both CBF velocity and regional brain oxygenation using the a priori noninferiority ratio of means of 0.8. Thus, mean values for dexmedetomidine were no more than 20% lower (worse) than those for propofol. Dexmedetomidine was not superior to propofol for either outcome (Fig. 2).

Fig. 2

Fig. 2

For the primary outcome of mean CBF velocity (FVm), there was one missing value in the dexmedetomidine group, which was assigned the observed minimum (worst) FVm value. The estimated ratio of geometric means (dexmedetomidine versus propofol) was 0.94 [90% CI: 0.84 to 1.05], P = 0.011 (significantly greater than the a-priori noninferiority ratio of 0.8, Table 2). In the dexmedetomidine group, the estimated mean (SD) change in FVm from the peak dose to the lowest dose was 0.5 ± 5.4 cm s−1 (peak versus lowest dose). CBFs along with mean and SBPs and end tidal CO2 at key times during surgery are presented in Table 3.

Table 2

Table 2

Table 3

Table 3

For the primary outcome of regional brain oxygenation, there were four missing values in the dexmedetomidine group, which were assigned the observed minimum (worst) brain oxygenation value. Accordingly, the estimated ratio of geometric means for brain oxygenation (dexmedetomidine versus propofol) was 0.99 [90% CI: 0.96 to 1.02], P < 0.001 (significantly greater than the a-priori noninferiority ratio of 0.8, Table 2).

In superiority tests on secondary outcomes the only significant difference between groups (using a significance criterion of 0.01 after Bonferroni correction for five outcomes) was in the OAA/S score, where the estimated difference in medians (dexmedetomidine minus propofol) was 1.0 [99% CI: 1.0 to 3.0], P < 0.001. Dexmedetomidine thus provided deeper sedation. There were no significant differences detected between the groups on pulsatility index (P = 0.91), CPP (P = 0.95), number of hypertensive episodes during DBS surgery (P = 0.02, i.e. NS after Bonferonni correction), or number of apnoeic episodes during DBS surgery (P > 0.99, Table 4).

Table 4

Table 4

Back to Top | Article Outline


Our primary result is that dexmedetomidine was noninferior to propofol as regards CBF velocities (CBFV) or regional brain oxygenation during DBS insertion. We also found that, with the dexmedetomidine dose we used, the patients were more deeply sedated as measured by OAA/S than those given propofol, indicating that dexmedetomidine was given at a dose that was at least physiologically comparable with propofol. There were no significant differences in the number of hypertensive or apnoeic episodes, and the end tidal CO2 and arterial oxygenation were comparable with each drug.

Our results are consistent with those of Drummond et al.12 who evaluated the relationship between CBFV and cerebral metabolic rate equivalent in six volunteers sedated with dexmedetomidine. Compared with the baseline (no drug) values, dexmedetomidine caused a dose-dependent reduction in CBFV of 18% and 32% when the plasma concentration was 0.6 and 1.2 ng ml−1, respectively, whereas CMR decreased by 26% and 41% at these same drug concentrations. Therefore, in that study,12 there was no consequent decrease in CBFV/CMR ratio. In addition, similar to our study, NIRS values were unchanged at all study intervals.12 Zornow et al.7 also found a reduction in CBFV in volunteers sedated with dexmedetomidine similar to that of Drummond et al.;12 however, unlike the current study, these authors7 did not measure CMR or regional brain oxygenation. The use of dexmedetomidine sedation in 20 patients with traumatic brain injury was associated with preservation of brain oxygenation.24 In addition, brain oxygenation was well preserved with dexmedetomidine infusion in five neurovascular surgical patients.25 However, in this latter investigation,25 the effect of dexmedetomidine on brain oxygenation was measured in the presence of volatile anaesthetics: brain oxygenation might thus have been protected by the substantial reduction in CMRO2 caused by volatile anaesthetics rather than dexmedetomidine per se.

In our study, dexmedetomidine provided deeper sedation than propofol without increasing apnoeic episodes. Also, dexmedetomidine did not increase the number of hypertensive episodes, a crucial factor because hypertension can provoke intracerebral haemorrhage during the microelectrode insertion.4 Dexmedetomidine provides limited analgesia, which might be beneficial during insertion of deep-brain stimulating electrodes.

Our study design was efficient in that we defined that dexmedetomidine would be noninferior (i.e. either better or no worse) to propofol only if it was noninferior on both mean CBF and brain oxygenation. Noninferiority deltas for each outcome were a-priori specified as a 20% reduction in the mean versus propofol, as smaller differences were not considered clinically important. We thus used a joint hypothesis test requiring noninferiority on both outcomes at the specified noninferiority deltas.19 As both outcomes were required to be significant (i.e. an intersection–union test), no adjustment to the significance criterion for multiple testing was required, thus allowing a smaller sample size. We also gained efficiency in not requiring superiority (for reasons elaborated above), although superiority could have been claimed, if found, since it is part of the noninferiority rejection region.

As with any noninferiority design, our conclusions depend on the a-priori noninferiority delta that was chosen. We chose a delta of 20% difference in means between dexmedetomidine and propofol based on our clinical experience and impressions that smaller differences would not be considered clinically relevant for brain oxygenation and CBF. Noninferiority was found on both outcomes. As always, more information is contained in confidence intervals for the treatment effects than the significant P values for noninferiority testing. CIs thus best characterise our results: these are consistent with dexmedetomidine being, on average, 4% worse to 2% better than propofol on brain oxygenation, and 16% worse to 5% better on mean CBF.

A limitation of our study is that we did not measure cerebral autoregulation: alterations in cerebral autoregulation may have affected the brain's response to the sedative medications. Furthermore, we were unable to directly measure CBF and it was estimated from the flow velocity measured by transcranial Doppler. The latter requires that various assumptions be met, specifically the perfusion territory, haemodynamic parameters and arterial vessel diameter must remain stable during measurements.26 It thus remains possible that dexmedetomidine causes vasoconstriction of small-calibre vessels distal to the MCA rather than constriction of the MCA itself.27 However, there was no difference in the Pulsatility Index between our two study groups. The Pulsatility Index is a measurement for downstream vascular resistance28 which is associated with angiographically diffuse intracranial vessel disease in haemodynamic stable patients29 and alterations of cerebral microcirculation in critically ill patients.30 Therefore it appears that dexmedetomidine and propofol have similar effects on small-calibre cerebral vessel resistance, suggesting in turn that there was no appreciable vasoconstriction even of distal vessels.

It is likely that overall hemodynamic conditions and brain injury both influence the effect of dexmedetomidine on cerebral autoregulation. Nevertheless, dexmedetomidine does not appear to have detrimental effects on cerebral perfusion, at least in healthy volunteers or patients with movement disorders who were haemodynamically stable.

In summary, dexmedetomidine was noninferior to propofol with regard to both regional brain oxygenation and CBF velocity. There were no significant differences between the sedatives on pulsatility index, CPP, number of hypertensive episodes or the number of apnoeic episodes. However, patients given dexmedetomidine were more sedated. Our results suggest that dexmedetomidine is a suitable and well tolerated sedative for awake deep-brain stimulator electrode insertion.

Back to Top | Article Outline

Acknowledgements relating to this article

Assistance with the study: none.

Financial support and sponsorship: Hospira company (research support and drug supply).

Conflicts of interest: AM has the following conflicts of interest to declare, none of which are directly related to this research: Consultant: Functional Neuromodulation and Spinal Modulation. Distribution rights from intellectual property: ATI, Cardionomics and Enspire. Fellowship support: Medtronic.

Presentation: part presentation at the American Society Meeting 2014, in New Orleans, and to the Association of University Anesthesiologists in Nashville, Tennessee 2015.

Back to Top | Article Outline


1. Machado A, Fernandez HH, Deogaonkar M. Deep brain stimulation: what can patients expect from it? Cleve Clin J Med 2012; 79:113–120.
2. Machado AG, Deogaonkar M, Cooper S. Deep brain stimulation for movement disorders: patient selection and technical options. Cleve Clin J Med 2012; 79 (suppl 2):S19–S24.
3. Starr PA, Markun LC, Larson PS, et al. Interventional MRI-guided deep brain stimulation in pediatric dystonia: first experience with the ClearPoint system. J Neurosurg Pediatr 2014; 14:400–408.
4. Khatib R, Ebrahim Z, Rezai A, et al. Perioperative events during deep brain stimulation: the experience at Cleveland Clinic. J Neurosurg Anesthesiol 2008; 20:36–40.
5. Venn RM, Hell J, Grounds RM. Respiratory effects of dexmedetomidine in the surgical patient requiring intensive care. Crit Care 2000; 4:302–308.
6. Prielipp RC, Wall MH, Tobin JR, et al. Dexmedetomidine-induced sedation in volunteers decreases regional and global cerebral blood flow. Anesth Analg 2002; 95:1052–1059.
7. Zornow MH, Maze M, Dyck JB, et al. Dexmedetomidine decreases cerebral blood flow velocity in humans. J Cereb Blood Flow Metab 1993; 13:350–353.
8. Karlsson BR, Forsman M, Roald OK, et al. Effect of dexmedetomidine, a selective and potent alpha 2-agonist, on cerebral blood flow and oxygen consumption during halothane anesthesia in dogs. Anesth Analg 1990; 71:125–129.
9. McPherson RW, Koehler RC, Kirsch JR, et al. Intraventricular dexmedetomidine decreases cerebral blood flow during normoxia and hypoxia in dogs. Anesth Analg 1997; 84:139–147.
10. Ogawa Y, Iwasaki K, Aoki K, et al. Dexmedetomidine weakens dynamic cerebral autoregulation as assessed by transfer function analysis and the thigh cuff method. Anesthesiology 2008; 109:642–650.
11. Kadoi Y, Saito S, Kawauchi C, et al. Comparative effects of propofol vs dexmedetomidine on cerebrovascular carbon dioxide reactivity in patients with septic shock. Br J Anaesth 2008; 100:224–229.
12. Drummond JC, Dao AV, Roth DM, et al. Effect of dexmedetomidine on cerebral blood flow velocity, cerebral metabolic rate, and carbon dioxide response in normal humans. Anesthesiology 2008; 108:225–232.
13. Talke P, Richardson CA, Scheinin M, et al. Postoperative pharmacokinetics and sympatholytic effects of dexmedetomidine. Anesth Analg 1997; 85:1136–1142.
14. Struys MRF, Jensen EW, Smith W, et al. Performance of the ARX-derived auditory evoked potential index as an indicator of anesthetic depth. Anesthesiology 2002; 96:803–816.
15. Schnider TW, Minto CF, Gambus PL, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88:1170–1182.
16. Belfort MA, Saade GR, Yared M, et al. Change in estimated cerebral perfusion pressure after treatment with nimodipine or magnesium sulfate in patients with preeclampsia. Am J Obstet Gynecol 1999; 181:402–407.
17. Ract C, Le Moigno S, Bruder N, et al. Transcranial Doppler ultrasound goal-directed therapy for the early management of severe traumatic brain injury. Intensive Care Med 2007; 33:645–651.
18. Kim MB, Ward DS, Cartwright CR, et al. Estimation of jugular venous O2 saturation from cerebral oximetry or arterial O2 saturation during isocapnic hypoxia. J Clin Monit Comput 2000; 16:191–199.
19. Carlson BW, Neelon VJ, Carlson JR, et al. Cerebral oxygenation in wake and during sleep and its relationship to cognitive function in community-dwelling older adults without sleep disordered breathing. J Gerontol A Biol Sci Med Sci 2011; 66:150–156.
20. Kampf S, Schramm PB, Klein KU. Transcranial doppler and near infrared spectroscopy in the perioperative period. Curr Opin Anaesthesiol 2013; 26:543–548.
21. Chua MV, Tsueda K, Doufas AG. Midazolam causes less sedation in volunteers with red hair. Can J Anaesth 2004; 51:25–30.
22. Mascha EJ, Turan A. Joint hypothesis testing and gatekeeping procedures for studies with multiple endpoints. Anesth Analg 2012; 114:1304–1317.
23. Laster LL, Johnson MF. Noninferiority trials: the ’at least as good as’ criterion. Stat Med 2003; 22:187–200.
24. Wang X, Ji J, Fen L, et al. Effects of dexmedetomidine on cerebral blood flow in critically ill patients with or without traumatic brain injury: a prospective controlled trial. Brain Injury 2013; 27:1617–1622.
25. Drummond JC, Sturaitis MK. Brain oxygenation during dexmedetomidine administration in surgical patients with neurovascular injuries. J Neurosurg Anesthesiol 2010; 22:336–341.
26. Kontos HA. Validity of cerebral arterial blood flow calculations from velocity measurements. Stroke 1989; 20:1–3.
27. Asano Y, Koehler RC, Kawaguchi T, et al. Pial arteriolar constriction to alpha 2-adrenergic agonist dexmedetomidine in the rat. Am J Physiol 1997; 272:H2547–H2556.
28. Bragoni M, Feldmann E. Tegeler CH, Babikian VL, Gomez CR. Transcranial Doppler indices of intracranial hemodynamics neurosonology. St. Louis: Mosby; 1996. 129–139.
29. Sharma VK, Tsivgoulis G, Lao AY, et al. Noninvasive detection of diffuse intracranial disease. Stroke 2007; 38:3175–3181.
30. Pierrakos C, Attou R, Decorte L, et al. Transcranial Doppler to assess sepsis-associated encephalopathy in critically ill patients. BMC Anesthesiol 2014; 14:45.
© 2017 European Society of Anaesthesiology