A randomised controlled trial of dexmedetomidine for suspension laryngoscopy


Correspondence to: X. Ruan
Email: xc_ruan@hotmail.com


We randomly allocated 80 patients to intravenous dexmedetomidine (0.25, 0.5, or 1 μg.kg−1) or placebo 15 min before anaesthetic induction. Dexmedetomidine 0.5 and 1.0 μg.kg−1 significantly reduced the mean (95% CI) propofol effect-site concentrations by 0.83 (0.63–1.03) μg.ml−1, p = 0.001 and 1.29 (1.12–1.46) μg.ml−1, p = 0.0003 at intubation, by 1.05 (0.85–1.25 μg.ml−1, p = 0.0006 and 1.33 (1.15–1.51) μg.ml−1, p = 0.0002 when surgery started, and by 0.59 (0.39–0.79) μg.ml−1, p = 0.030 and 0.72 (0.57–0.87) μg.ml−1, p = 0.004 on completion of surgery, respectively. Patients’ tracheas were extubated sooner after 0.5 and 1.0 μg.kg−1 dexmedetomidine, by 5.36 (2.39–8.32) min, p = 0.009 and 7.37 (3.24–11.51) min p = 0.003, respectively. Tachycardic responses to intubation were present in five placebo patients and no dexmedetomidine patients. Bradycardia was treated after dexmedetomidine in six patients: five after 1.0 μg.kg−1; and one after 0.25 μg.kg−1. Single-dose dexmedetomidine can reduce anaesthetic requirements, with both desirable and undesirable haemodynamic effects.

Suspension laryngoscopy under general anaesthesia can cause marked increases in heart rate and blood pressure. The ability of adjuvant drugs to limit these changes has been assessed, including selective β1-adrenergic antagonists-blockers [1], calcium channel blockers [2], local anaesthetics and glucocorticoids [3]. These adjuvant drugs therefore reduce the utility of haemodynamic variables to indicate whether patients are adequately anaesthetised [4, 5].

α2-Adrenergic agonists, including dexmedetomidine, have sedative, anaesthetic-sparing and sympatholytic characteristics [6–8]. Dexmedetomidine by intravenous infusion has been used to maintain general anaesthesia [9–11]. The sedative properties of dexmedetomidine could conceivably delay the recovery from anaesthesia, particularly after continuous infusion [12, 13]. A single dose may reduce this problem: a single dose of 0.5 μg.kg−1 dexmedetomidine decreased thiopental requirements and improved recovery from anaesthesia after brief gynaecological procedures [14]. However, a single dose might not be sufficient for more stressful surgeries. Consequently, we designed this study to examine the effects of different doses of dexmedetomidine on level of anaesthesia and haemodynamic values during suspension laryngoscopy and subsequent recovery.


In this randomised, blinded and placebo-controlled trial, we compared three doses of dexmedetomidine: 0.25; 0.5; and 1.0 μg.kg−1. The study was approved by the institutional Ethics Committee. Written informed consent was obtained before enrolling the patients in the study. We assessed the following outcomes: propofol and remifentanil requirements to maintain anaesthetic depth; haemodynamic responses to intubation and extubation; time to breathe, extubation and consciousness; and patients’ satisfaction.

Patients were 20–60 years old, of ASA physical status 1-2, scheduled for elective suspension laryngoscopy under general anaesthesia between June 2011 and January 2012. We excluded patients with: suspected difficult airways; hypertensive histories; morbid obesity (body mass index ≥ 35 kg.m−2); heart rate < 45 bpm; or second or third degree heart block. We also excluded pregnant or nursing women, or pre-menopausal women without reliable contraception, patients taking α-methyldopa, clonidine or other α2-adrenergic agonists.

When patients arrived in the operating room, standard monitoring (ECG, blood pressure and pulse oximetry) and Narcotrend® electroencephalography (software version 4.3; MonitorTechnik, Bad Bramstedt, Germany) were applied and recorded continuously. The Narcotrend monitor was connected to the patient’s forehead according to the manufacturer’s instructions. Studies comparing Narcotrend and Bispectral Index (BIS®) monitors showed Narcotrend indices ‘D0’ to ‘E1’ were equivalent to BIS values 56 to 35 [15, 16].

Patients were randomly allocated by computer-generated random numbers to receive one out of four intravenous infusions over 10 min: placebo; or dexmedetomidine 0.25, 0.5 and 1.0 μg.kg−1 (Hengrui Med, Jiangsu, China). Infusions were commenced 15 min before the induction of anaesthesia. A nurse who was not involved in any other sections of the study obtained the numbered opaque envelopes subsequently and prepared the treatment agents by diluting dexmedetomidine or placebo with sodium chloride 0.9% into a volume of 10 ml. Enrolment of patients, preparation of opaque envelopes, administration of drugs, and collection and analysis of data were performed by doctors and nurses who were blinded to the study treatments.

Fifteen minutes after the study infusion was started, propofol was infused to a target plasma concentration of 2.5 μg.ml−1 using DiprifusorTM software (version 2.0, Graseby® 3500 anaesthesia pump; Smiths Medical, Watford, UK). The infusion rate was adjusted to achieve a Narcotrend index between ‘D0’ and ‘E1’, which was maintained until the end of surgery with plasma concentration increments of 0.5 μg.ml−1. A remifentanil infusion was then started to achieve a target plasma concentration of 3.0 ng.ml−1, derived from the Minto pharmacokinetic model [17]. After loss of consciousness, oxygen was given by facemask ventilation. Patients received 0.6 mg.kg−1 rocuronium, the trachea was intubated 1 min later, and the lungs were ventilated to an end-tidal carbon dioxide partial pressure of 4.0–4.7 kPa. Immediately after intubation, the remifentanil plasma concentration was reduced to 2.0 ng.ml−1, and then adjusted to maintain the systolic blood pressure within 25% of the pre-operative value and the heart rate to less than 90 bpm. To maintain the targeted Narcotrend stages and stable haemodynamics, we adjusted the concentrations of anaesthetics and used ephedrine when the systolic blood pressure fell by more than 25% of pre-operative values. Neuromuscular blockade was maintained with 0.15-mg.kg−1 increments of rocuronium when the first twitch in a train-of-four response was observed. Propofol and remifentanil infusions were stopped simultaneously when the operative laryngoscope was removed.

When clinically indicated, ephedrine 5 mg or atropine 0.5 mg was administered intravenously. We recorded episodes of hypotension (systolic blood pressure below 70% of the baseline value for more than 1 min), hypertension (systolic blood pressure above 130% of the baseline value for more than 1 min), bradycardia (heart rate < 40 bpm for more than 10 s) and tachycardia (heart rate > 100 bpm for more than 10 s). We recorded propofol effect-site concentrations when: the Narcotrend index reached ‘D0’; before inserting the operative laryngoscope; and on removal of the laryngoscope.

Residual neuromuscular block was reversed with neostigmine 2 mg and atropine 0.5 mg, and the patient was transferred to the recovery room. The patients’ lungs were ventilated with a manual resuscitation device with supplemental oxygen during transfer and with synchronised intermittent mandatory ventilation in recovery. Artificial ventilation was stopped when adequate spontaneous ventilation was confirmed: respiratory rate > 8 breaths.min-1, SpO2 > 90% on room air for at least 5 min. The trachea was extubated when patients met these criteria and squeezed the anaesthetist’s hand. The times from stopping anaesthetic infusions to adequate ventilation, consciousness and extubation were recorded.

An observer recorded episodes of shivering, nausea, vomiting, hypotension, hypertension, tachycardia, bradycardia, respiratory complications and analgesic requests for 2 h after extubation. Respiratory depression was defined as SpO2 < 90 % or < 8 breaths.min−1. Pain scored above 50 mm on a 0–100 mm visual analogue scale was treated with intravenous parecoxib 40 mg, whereas nausea or vomiting was treated with intravenous ondansetron 4 mg. Episodes of hypotension, hypertension, bradycardia and tachycardia were also treated. During recovery, effect-site concentrations of propofol were recorded on adequate spontaneous ventilation, on emergence and on extubation. On discharge from recovery, patients were asked to rate their satisfaction with the anaesthesia and the surgery that they had received as highly satisfactory, acceptable or unacceptable. All patients were interviewed about intra-operative recall on the first postoperative day. The senior author, who was not involved in the conduct of the anaesthesia and was blinded to the treatment administered, recorded patient satisfaction and conducted the follow-up interview.

The sample size was based on our previous experience in a similar patient population in which the mean (SD) time to extubation was 22 (2) min. To detect a 2-min difference in mean extubation time compared with 22 min, assuming a common SD of 2 min, we calculated that we would need 19 patients per group to achieve 80% power (α 5%). Assuming a 5% dropout rate, the final sample size was set at 80 patients (20 patients per group). Sex, ASA physical status and number of patients were analysed using the Kruskal–Wallis test to perform an overall comparison among groups. The other demographic, surgical and anaesthetic variables were analysed using one-way ANOVA. The numbers of patients with cardiovascular adverse episodes and high satisfaction were analysed with the Kruskal–Wallis test and post-hoc Mann–Whitney tests with Bonferroni adjustments. A p value < 0.05 was considered statistically significant.


Figure 1 shows the patient flow through the study. Baseline characteristics were similar in the four groups (Table 1).

Figure 1.

 Patient flow in the study.

Table 1. Characteristics and surgical data of patients receiving placebo or dexmedetodine 0.25, 0.5 and 1.0 μg.kg−1. Values are mean (SD) or number.
 Placebo (n = 20)0.25 μg.kg−1 (n = 20)0.5 μg.kg−1 (n = 20)1.0 μg.kg−1 (n = 20)
Age; years45 (9)45 (11)42 (13)44 (12)
Weight; kg58 (13)59 (10)63 (11)60 (10)
Height; cm162 (12)157 (14)162 (10)155 (11)
ASA status; 1/217/316/418/217/3
Duration of anaesthesia; min36 (4)36 (7)41 (8)39 (7)
Duration of surgery; min19 (5)19 (6)25 (7)20 (4)

The mean (SD) target concentration of propofol in the placebo group was 3.60 (0.33) μg.ml−1 at intubation, 3.58 (0.51) μg.ml−1 at the start of surgery and 3.28 (0.42) μg.ml−1 on completion of surgery. Dexmedetomidine 0.25, 0.5 and 1.0 μg.kg−1 reduced the mean (95% CI) target concentrations of propofol at these times by: 0.65 (0.43–0.87) μg.ml−1, p = 0.008, 0.83 (0.63–1.03) μg.ml−1, p = 0.001 and 1.29 (1.12–1.46) μg.ml−1, p = 0.0003; by 0.57 (0.34–0.79) μg.ml−1, p = 0.013, 1.05 (0.85–1.25) μg.ml−1, p = 0.0006 and 1.33 (1.15–1.51) μg.ml−1, p = 0.0002; and by 0.09 (−0.11 to 0.29) μg.ml−1, p = 0.846, 0.59 (0.39–0.79) μg.ml−1, p = 0.030 and 0.72 (0.57–0.87) μg.ml−1, p = 0.004, respectively (Fig. 2). Respective reductions in remifentanil concentrations were 0.27 (0.21–0.33) ng.ml−1, p = 0.035, 0.64 (0.52–0.76) ng.ml−1, p = 0.009 and 0.73 (0.51–0.95) μg.kg−1, p = 0.0002 at the start of surgery; and 0.09 (0.08–0.11) ng.ml−1, p = 0.613, 0.59 (0.51–0.67) ng.ml−1, p = 0.005 and 0.72 (0.60–0.84) ng.ml−1, p = 0.0006 on completion of surgery (Fig. 3).

Figure 2.

 Mean predicted propofol effect-site concentrations after placebo (inline image), dexmedetomidine 0.25 μg.kg−1 (inline image), dexmedetomidine 0.5 μg.kg−1 (inline image) and dexmedetomidine 1.0 μg.kg−1 (inline image). Error bars represent SD. TCI, target-controlled infusion.

Figure 3.

 Mean predicted remifentanil effect-site concentrations after placebo (inline image), dexmedetomidine 0.25 μg.kg−1 (inline image), dexmedetomidine 0.5 μg.kg−1 (inline image) and dexmedetomidine 1.0 μg.kg−1 (inline image). Error bars represent SD. TCI, target-controlled infusion.

The mean (SD) times in the placebo group for adequate ventilation, emergence and extubation were 15.2 (1.8) min, 20.9 (2.4) min and 26.2 (3.7) min, respectively. The respective mean (95% CI) reduction in these times after dexmedetomidine 0.25, 0.5 and 1.0 μg.kg−1 were 0.70 (0.36–1.04) min, 2.72 (1.52–3.92) min and 4.13 (2.18–6.08) min, p = 0.0006; −0.6 (−1.8 to 0.6), 2.58 (0.82–4.34) min and 1.82 (1.03–2.61) min, p = 0.58; and 0.23 (−0.35 to 0.81) min, 5.36 (2.39–8.32) min and 7.37 (3.24–11.51) min, p = 0.00002 (Fig. 4). There were no significant differences in the effect-site concentrations of either propofol or remifentanil between the treatment groups and the placebo group at the return of spontaneous breathing and at extubation (Fig. 5).

Figure 4.

 Mean recovery times after placebo (inline image), dexmedetomidine 0.25 μg.kg−1 (inline image), dexmedetomidine 0.5 μg.kg−1 (inline image) and dexmedetomidine 1.0 μg.kg−1 (inline image). Error bars represent SD.

Figure 5.

 Mean predicted effect-site concentrations of (a) propofol and (b) remifentanil after placebo (inline image), dexmedetomidine 0.25 μg.kg−1 (inline image), dexmedetomidine 0.5 μg.kg−1 (inline image) and dexmedetomidine 1.0 μg.kg−1 (inline image). Error bars represent SD. TCI, target-controlled infusion.

On leaving the recovery, 11 patients were highly satisfied in the placebo group, compared with 15, 18 and 19 patients with increasing doses of dexmedetomidine. No patient was unhappy.

Table 2 reports the rates of haemodynamic observations. One patient in the placebo group and two after 0.25 μg.kg−1 needed postoperative analgesia (p = 0.29). Two patients from both the placebo and dexmedetomidine 1.0 μg.kg−1 groups were treated for nausea and vomiting (p = 0.57). Postoperative shivering and respiratory depression were not reported. No patient recalled intra-operative awareness.

Table 2. Patients with cardiovascular adverse episodes in four groups. Values are number of patients.
 Placebo0.25 μg.kg−10.5 μg.kg−11.0 μg.kg−1
Bradycardic episode
Tachycardic episode
 After induction5000
Hypotensive episode
Hypertensive episode


We found that single-dose dexmedetomidine reduced anaesthetic requirements when guided by the Narcotrend index, with mixed effects on both haemodynamic and recovery profiles. Episodes of bradycardia after 1.0 μg.kg−1 dexmedetomidine suggest that a dose of 0.5 μg.kg−1 might be preferable.

When given by rapid intravenous injection, both dexmedetomidine and clonidine can exert a biphasic effect on arterial blood pressure, starting with transient hypertension followed by a longer lasting reduction in blood pressure [11, 18]. However, we did not find monophasic reduction or biphasic responses in our dexmedetomidine groups. We used relatively small doses at low infusion rates (up to 0.1 μg.kg−1.min−1), whereas others have used infusions of up to 1.0 μg.kg−1.min−1. However, we measured blood pressure every 5 min, so we may have missed these effects.

Anaesthetic regimens that limit hypertensive tachycardic responses to peri-operative stress might avoid myocardial ischaemia [19]. However, these unwanted cardiovascular responses may be replaced by severe bradycardia. Ingersoll-Weng et al. reported a case of cardiac arrest related to the use of dexmedetomidine [20]. The authors suggested factors that might have predisposed to cardiac arrest, including increased basal vagal tone, administration of pyridostigmine and the dose of dexmedetomidine (a 1.0-μg.kg−1 loading dose followed by an infusion of 0.2 μg.kg−1.h−1). Our results support their suggestion that dexmedetomidine at a dose of 0.5 μg.kg−1 is preferable to 1.0 μg.kg−1. The moderate analgesic effects of dexmedetomidine may peak at a dose of 0.5 μg.kg−1 [21].

Dexmedetomidine can decrease isoflurane requirements by up to 90% [9, 22]. Dexmedetomidine 0.63 μg.kg−1 reduced the dose of propofol required to induce anaesthesia [23]. Similarly, we observed that dexmedetomidine exerted a dose-dependent reduction in the concentrations of propofol required to induce and maintain general anaesthesia. Although the utility of processed EEG for assessment of the dexmedetomidine-induced sedation has been well documented [24–27], a recent study [28] found that dexmedetomidine increased intra and inter-individual variability of BIS and Entropy values in volunteers sedated with midazolam or remifentanil. The large variability calls into question the use of processed EEG to assess sedation levels. More recently, Kaskinoro et al. [29] demonstrated that BIS and Entropy were not able to differentiate consciousness from unconsciousness induced by dexmedetomidine, propofol and sevoflurane. However, these processed EEG monitors, including the Narcotrend in this study, are not designed to measure the level of sedation, as Kaskinoro et al. [29] had done, but were developed to measure the level of hypnosis during surgical anaesthesia.

Dexmedetomidine can blunt the cardiovascular responses to emergence and extubation, a finding we replicated [10, 30]. However, a dexmedetomidine dose of 0.25 μg.kg−1 did not affect the recovery profiles. The recovery effect-site concentrations of both propofol and remifentanil were comparable between groups: we think that a sympatholytic property, rather than sedative or analgesic properties, account for these differences in recovery between dexmedetomidine doses.

Dexmedetomidine causes less respiratory depression than other sedatives or narcotics [6, 7, 31]. In our study, infusions of dexmedetomidine did not significantly alter either SpO2 or end-tidal carbon dioxide partial pressure. In addition, tracheal extubation in patients receiving dexmedetomidine was possible a few minutes earlier than in patients receiving placebo.

Our study is limited by its small scale and relatively healthy participants. The effects of dexmedetomidine may differ with disease, drugs and other surgical stimuli. We relied upon calculated target concentrations, rather than measuring plasma concentrations. However, the pharmacokinetic models we used are well established. Finally, the differences that we recorded in patient outcomes might be considered clinically insignificant.

In conclusion, dexmedetomidine given before induction of anaesthesia reduced anaesthetic requirements, altered peri-operative haemodynamic stability, increased the speed of recovery and increased patient satisfaction. Furthermore, when used for suspension laryngoscopy, our data support a dexmedetomidine dose of 0.5 μg.kg−1 in preference to 1.0 μg.kg−1 to reduce the risk of bradycardia.


The authors thank Dr Li-ming Zhang from University of Pittsburgh School of Medicine for scientific discussion and critical review of the manuscript.

Competing interests

This study is partially supported by Natural Science Foundation of Guangdong Province (grant S2011010000587) and Key Project from Guangzhou Health Bureau (grant 20121A021007). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.