This article is accompanied by the following Invited Commentary:
Weibel S, Elia N, Kranke P. The transparent clinical trial: why we need complete and informative prospective trial registration. Eur J Anaesthesiol 2016; 33:72–74.
In patients undergoing general anaesthesia with volatile agents and opioids, about one-third are likely to suffer from postoperative nausea and vomiting (PONV).1 According to Apfel's risk stratification model, the incidence of PONV increases to over 60% in patients having three or more risk factors.2 Undoubtedly, PONV is unpleasant and even a mild episode can markedly decrease patient satisfaction, delay hospital discharge and increase the use of medical resources.3 Several strategies for the prevention of PONV have been described and incorporated into recent guidelines,4 but some risk factors are not modifiable.5
Despite being a key trigger for PONV, postoperative opioids are frequently used for the pain control. In an effort to reduce side effects related to opioid-based, intravenous (i.v.) patient-controlled analgesia (PCA), the addition of various adjuncts to the PCA solution has been tried and studied extensively,6–8 but their efficacy is uncertain in highly susceptible patients.9,10
Dexmedetomidine, a selective α2 adrenergic receptor agonist, has analgesic, sedative and sympatholytic properties, with a lack of respiratory depression.11 The pharmacokinetic profile, with an elimination half-life of 2 h, indicates a fast offset, but its clearance decreases with age resulting in a longer elimination half-life and context-sensitive half-life in the elderly.12 With its multiple beneficial effects, the systemic administration of dexmedetomidine in the perioperative period is gaining acceptance as a beneficial sedative and analgesic agent.13 A previous study assessing the efficacy of dexmedetomidine as an adjunct to a morphine-based i.v. PCA in gynaecological patients found a lower incidence of PONV and superior analgesia, despite using significantly less morphine.14 As yet, dexmedetomidine is only licensed for intensive care sedation and few data exist regarding the efficacy of dexmedetomidine in preventing postoperative PONV in highly susceptible patients using opioid-based PCA.
In this randomised, double-blind, placebo-controlled study, we evaluated whether adding dexmedetomidine to a fentanyl-based PCA drug mixture would reduce PONV in highly susceptible patients. The secondary objectives were to determine its analgesic efficacy and safety.
Materials and methods
Ethical approval for this study was provided by the 3rd Committee of Institutional review board of the Yonsei University Health System, Seoul, Korea on 7 September 2012, and the study commenced. Later, on 17 April 2013, the trial was registered at http://clinicaltrials.gov (NCT01840254). Written informed consent was obtained from all patients.
We recruited 108 patients, aged 20 to 65 years, scheduled for posterior lumbar spinal fusion (level 1 or 2), and who had at least three risk factors for PONV (female, nonsmoker, use of postoperative opioids). Exclusion criteria were American Society of Anesthesiologists physical status at least III, daily use of opioids for more than 1 week, uncontrolled hypertension, greater than first-degree atrioventricular block, cognitive impairment, use of psychiatric medications, drug or alcohol abuse and allergy to either fentanyl, or α-2 adrenergic receptor agonists, or NSAIDs.
Patients were allocated randomly to the control or dexmedetomidine group. Computer-generated randomisation was performed and the patient allocation was delivered in sealed, opaque, sequentially numbered envelopes. Before surgery, patients were asked to describe their nausea intensity on an 11-point verbal numerical rating scale (VNRS: 0, no nausea at all; 10, the most severe nausea imaginable) and their pain intensity on a 100 mm visual analogue scale (VAS: 0, no pain at all; 100, the most severe pain imaginable). Anaesthesia was induced with propofol (2 mg kg−1), remifentanil (1 μg kg−1) and rocuronium bromide (0.8 mg kg−1), and maintained with sevoflurane (end-tidal concentration of 1.5 to 2.5%) in 40% oxygen with air, and a continuous infusion of remifentanil (0.1 to 0.2 μg kg−1 min−1). All the patients received 4 to 8 ml kg−1 h−1 of Ringer's acetate solution (Plasma Solution A Inj.; CJ Pharma, Seoul, Korea) according to their preoperative fluid deficit, maintenance dose and intraoperative losses. In addition, 6% hydroxyethyl starch 130/0.4 (Voluven; Fresenius Kabi, Bad Homburg, Germany) was administered (up to 20 ml kg−1) to replace blood loss. The anaesthesia nurse, who was not involved in this study, opened the envelopes and prepared the drug/placebo mixtures for the PCA. Thirty minutes before the end of surgery, 0.5 μg kg−1 of dexmedetomidine (Precedex; Hospira, Seoul, Korea) or the equivalent volume of 0.9% normal saline was given i.v. over 10 min in the dexmedetomidine and control group, respectively. Thereafter, both groups received 0.5 μg kg−1 of fentanyl (Hana Pharm, Seoul, Korea) and 4 mg of ondansetron (Onseran; Yuhan Products, Seoul, Korea) and the PCA pump (Automed 3400; Acemedical, Seoul, Korea) was commenced.
The PCA regimen consisted of fentanyl 10 μg kg−1 and ketorolac 120 mg (Keromin; Hana Pharm), mixed with 0.9% normal saline to a total volume of 100 ml. In the dexmedetomidine group, in addition to the fentanyl and ketorolac, 10 μg kg−1 of dexmedetomidine (Precedex 100 μg ml−1; Hospira, Lake Forest, Illinois, USA) was added to the PCA solution and again the volume was made up to 100 ml with 0.9% normal saline. The PCA was programmed to deliver a 1 ml bolus on-demand, with a lockout interval of 15 min, and a background infusion rate of 2 ml h−1. In the dexmedetomidine group, this PCA programme thus allowed a background infusion of dexmedetomidine 0.2 μg kg−1 h−1 and a bolus of dexmedetomidine 0.1 μg kg−1. The PCA was used for the first 48 h postoperatively. During the study period after surgery, all patients received 0.3 mg of ramosetron (Nasea; Astellas, Tokyo, Japan) i.v. every 24 h, and one capsule of Mypol (Sungwon Adcock Parm, Seoul, Korea, a combination of acetaminophen 250 mg, ibuprofen 200 mg and codeine phosphate 10 mg) given every 6 h. When a patient reported a resting VAS pain score at least 40 mm, then 25 mg of pethidine (Demerol; Jeil Pharm, Seoul, Korea) was injected i.v. When the patients reported a VNRS nausea score at least 4, then 10 mg of metoclopramide (Macperan; Dongwha, Seoul, Korea) was injected i.v.
Noninvasive blood pressure (BP), heart rate (HR) and respiratory rate were monitored throughout the study period. Hypotension [mean arterial pressure (MAP) <60 mmHg] was treated with 100 ml of 6% hydroxyethyl starch 130/0.4 (Voluven; Fresenius Kabi, Bad Homburg, Germany) or ephedrine 6 mg i.v. Bradycardia (HR ≤45 beats min−1) was treated with atropine 0.5 mg i.v. The PCA infusion was stopped completely if hypoventilation (respiratory rate <10 breaths min−1) occurred or if severe PONV, hypotension or bradycardia persisted despite treatment. Anaesthesiologists, surgeons, patients, investigators and caregivers in the ward were blind as to the PCA drug mixture until the end of the study.
Patient's data including age, height, weight, smoking status, history of motion sickness or PONV, and baseline MAP, HR and respiratory rate were recorded. Perioperative data including duration of surgery, intraoperative fluid input (the sum of infused fluid and blood products) and output (the sum of urine output and blood loss), total dose of remifentanil, time to extubation after the completion of surgery and length of hospital stay were recorded. The presence and severity of nausea, vomiting and pain at rest and with movement (lifting a leg while lying in the supine position), the cumulative dose of PCA fentanyl and requirement for rescue analgesia and antiemetics were recorded at 1, 3, 6, 12, 24, 36, 48 h after surgery. The MAP, HR, respiratory rate, sedation score on a 5-point scale (0, fully awake; 1, drowsy/closed eyes; 2, asleep/easily aroused with a simple verbal command or light tactile stimulation; 3, asleep/arousable by strong physical stimulation; 4, unarousable)15 and any adverse events, including headache, dizziness and xerostomia, were recorded just after arrival in the postanaesthesia care unit and at 1, 3, 6, 12, 24, 36, 48 h after surgery.
Primarily, the cumulative dose of PCA fentanyl consumption, incidence of PONV and severity of nausea were compared between the dexmedetomidine and control groups during the first 48 h after surgery. With regard to the severity of nausea, changes in mean value of VNRS and the proportion of patients having mild (VNRS 1 to 3) versus moderate to severe (VNRS ≥4) were compared. Secondarily, pain intensities and additional opioid requirements were compared between the groups.
When the trial was eventually registered on ClinicalTrials.gov, the comparison of PONV between the groups was declared a secondary outcome measure by a mistake.
The sample size calculation was performed using Power Analysis and Sample Size 2008 software (NCSS). In our previous study,16 60% of female patients receiving fentanyl-based i.v. PCA following lumbar spinal surgery experienced PONV. We considered a 30% reduction in the incidence of PONV to be clinically relevant and the unilateral test computed that 49 patients were needed in each group at an α level of 0.05 and with power of 80%. We included 54 individuals per group to allow for a possible 10% drop-out.
Statistical analysis was performed using SAS version 9.2 (SAS Institute Inc., Cary, North Carolina, USA). To compare patient characteristics and operative data between the groups, independent t-tests or Mann-Whitney U tests for continuous variables, and Fischer exact tests or Chi-square tests for categorical variables were performed. The incidences of nausea, vomiting, differences in the proportions of nausea severity and adverse events were analysed by Fisher exact tests or Chi-square tests as appropriate. To compare changes in nausea and pain intensities and haemodynamic variables between the groups, a generalised linear mixed model was used with adjustment of the P value by Bonferroni correction. The doses of rescue antiemetics and analgesics were compared between the groups using the Mann-Whitney U test. A P value less than 0.05 was considered to be statistically significant. All results were expressed as mean ± standard deviation (SD), mean difference with 95% confidence interval (CI), median [interquartile range], number of patients (%) or odds ratio with 95% CI.
Of the 108 patients who were enrolled in the study between September 2012 and September 2013, the PCA was discontinued before the third postoperative hour in two patients with intractable PONV in the control group and in one patient with persistent hypotension in the dexmedetomidine group. The data from these three patients were excluded from analysis and 105 patients completed the study: 52 in the control group and 53 in the dexmedetomidine group (Fig. 1).
Patients’ characteristics and perioperative data including time to extubation and days to hospital discharge after surgery were comparable between the groups (Table 1).
With regard to the PONV (Table 2), patients in the dexmedetomidine group had a significantly reduced risk of postoperative nausea during the first 1 to 3 h compared with control group (odds ratio 0.32, 95% CI 0.13 to 0.77). The incidence of vomiting and the total amount of rescue antiemetics administered during 48 h were comparable between the groups. Comparisons of nausea severity between the groups among the patients with VNRS more than 0 are shown on Fig. 2. Differences in mean values of VNRS did not reach statistical significance (a), but the risk of experiencing a moderate to severe degree of nausea was significantly lower in the dexmedetomidine group at 12 h (odds ratio 0.28, 95% CI 0.12 to 0.67) (b).
The fentanyl consumption at each time interval was significantly greater in the control group up to 6 h (1 h: mean difference 4 μg, 95% CI 2.02 to 5.98, P = 0.003; 3 h: mean difference 8 μg, 95% CI 4.88 to 11.12, P < 0.001; 6 h: mean difference 7 μg, 95% CI 2.36 to 11.64, P = 0.005). The cumulative PCA fentanyl consumption was significantly greater in the control group up to 12 h (3 h: mean difference 11 μg, 95% CI 7.44 to 14.56, P < 0.001; 6 h: mean difference 20 μg, 95% CI 13.16 to 26.84, P < 0.001; 12 h: mean difference 28 μg, 95% CI 12.84 to 43.16, P = 0.002). The total amount of pethidine administered during 48 h was also greater in the control group (Table 3).
The VAS scores were not significantly different either at rest or on movement between the groups throughout the study period (Fig. 3).
Changes in haemodynamic variables from preoperative baseline to 48 h after surgery are shown in Fig. 4. When all time points were combined, the HR and MAP during the postoperative period were significantly higher in the control group, but there were no statistically significant differences between the groups in HR and MAP at each individual time point. Changes in respiratory rate were not significantly different between the groups throughout the study period. There was a trend towards a higher HR in the control group immediately after administration of the loading dose of dexmedetomidine [56 (SD 10) versus 60 (SD 9), P = 0.056)]in the operating theatre, but none of the patients required treatment. The levels of sedation were also comparable (data not shown).
Adverse events during the 48 h after surgery are summarised in Table 4. The incidences of hypotension, bradycardia, dizziness, headache and xerostomia were comparable between the groups. None of the patients developed hypoventilation or had a sedation score at least 3, and the proportions of patients with sedation score 0, 1 and 2 were similar between the groups throughout the study period.
In this randomised controlled study, dexmedetomidine added to a fentanyl-based PCA drug mixture provided beneficial effects in that it reduced the frequency and severity of nausea, the PCA fentanyl requirements and supplemental analgesic requirements in the early postoperative period in patients at high risk of developing PONV following lumbar spinal fusion. These benefits of dexmedetomidine were in addition to a prolonged and significant opioid-sparing effect.
Although postoperative nausea is often self-resolving, even a single mild episode can cause distress. As the severity of the nausea increases, so too does the potential for vomiting. The latter may be associated with more serious complications such as aspiration pneumonia and wound dehiscence. Thus, the importance of preventing PONV cannot be overemphasised. However, its prevention can prove to be very difficult after surgical procedures eliciting considerable postoperative pain, as these patients usually require opioid-based analgesic regimens. According to the Apfel simplified score,2 those patients receiving opioid-based PCA postoperatively who have more than three risk factors may experience intractable PONV, leading to premature cessation of the PCA.10 For these patients, aggressive and efficient preventive measures using a multimodal approach should be a high priority. In addition to the use of antiemetics, adjuncts that exert opioid-sparing effect have emerged as a feasible option.17
Dexmedetomidine is an α2 adrenergic receptor agonist exhibiting analgesic, sedative and sympatholytic actions without causing respiratory depression.12,13 A rise in the plasma concentrations of catecholamines is a known factor contributing to PONV and so, intuitively, dexmedetomidine might be expected to have antiemetic properties.18 Also, dexmedetomidine has been consistently reported as augmenting postoperative analgesia, suggesting a potential opioid-sparing effect.13 Indeed, its opioid-sparing effect has been demonstrated in gynaecologic patients, and this, on its own, contributes to a reduction in PONV.14 Proposed mechanisms of action for dexmedetomidine include the inhibition of nociceptive neurotransmission via activation of peripheral, spinal and supraspinal α2-adreno receptor,19 attenuation of the stress response and the affective-motivational components of pain,20 as well as the alleviation of hyperalgesia induced by intraoperative opioid infusions21 and surgical inflammation.22 Of interest, a synergism of α2 adrenergic receptor agonist and opioids has been demonstrated in neuropathic and inflammatory pain models.23 These latter features may be especially valuable in patients undergoing lumbar spinal surgery, as the nature of their pain is often associated with provocation or amplification of neuropathic and inflammatory pain pathways. However, previous evidence to support this is lacking.
In the present study, we investigated the efficacy of dexmedetomidine as an adjunct to a fentanyl-based PCA drug mixture in patients who, according to Apfel's criteria,2 are at high risk of PONV. The incidence of nausea reported at 3 h was more than 50% lower in the dexmedetomidine group than in the control group. This might be attributable to the initial loading dose of dexmedetomidine (0.5 μg kg−1) given near the end of surgery. Although there was no statistical significance, there was a trend towards a lower incidence of nausea in the dexmedetomidine group for up to 24 h and there was a reduction in the severity of nausea demonstrated in the dexmedetomidine group at the 12 h assessment. Our results suggest an interaction between fentanyl consumption and nausea and an opioid-sparing effect of dexmedetomidine. The time period when dexmedetomidine exerted its beneficial influence on nausea coincided with the time period when cumulative fentanyl consumption was significantly lower in the dexmedetomidine group. Conversely, the time period when the antiemetic efficacy of dexmedetomidine was no longer evident (after 24 h) was the time period during which there was no longer a statistically significant difference in the fentanyl consumption between the two groups.
Although it could be argued that the mean difference between the groups in the cumulative dose of fentanyl at 3 h was small (approximately 11 μg), it should be noted that patients in the dexmedetomidine group also received significantly less rescue analgesia compared with the control group. If the rescue doses of meperidine are included, the difference in opioid requirements is equivalent to 0.5 mg of morphine per hour. In the control group, there were more frequent PCA demands resulting in a significantly shorter time to completion of the PCA, as well as a greater need for rescue analgesia, both facts suggesting an opioid-sparing effect by dexmedetomidine. Despite the reduced use of opioids in the dexmedetomidine patients, there was a trend towards lower VAS scores in this group, perhaps again suggesting an additional analgesic effect from dexmedetomidine.
There are limitations in study. First, we included ketorolac in the PCA regimen of both groups. The different PCA consumption between the groups could have resulted in different plasma concentration of the ketorolac, and this may have had some effect in producing comparable pain scores in the groups as well as limiting the opioid-sparing effect of dexmedetomidine. Furthermore, all the patients received equal daily doses of Mypol (analgesic drug mixture) and ramosetron (antiemetic) throughout the study period. These additional analgesics and antiemetics might have reduced pain and nausea equally in both groups, hiding an opioid-sparing effect of dexmedetomidine. The reason we did not control those potential confounding factors was because we followed a multimodal approach for prophylaxis of PONV in our study cohort, regardless of randomisation.
The second limitation is the possible haemodynamic and respiratory instability. The most challenging issues facing clinicians in the postoperative use of dexmedetomidine are unwanted oversedation and haemodynamic instability. According to the literature, in the clinically relevant dose range, serious consequences arising from the sedative properties of dexmedetomidine have seldom been demonstrated and any sedation that did occur did so without respiratory depression.24 Despite our PCA regimen allowing an infusion rate of up to 0.6 μg kg−1 h−1, which is approaching the upper limit used for sedation in intensive care, in our current study, none of the patients receiving dexmedetomidine were deeply sedated nor did they have respiratory depression. However, the synergistic action of dexmedetomidine and fentanyl needs to be considered and the risk of respiratory depression should always be kept in mind with such combination therapy. There have also been reports of haemodynamic deterioration associated with dexmedetomidine,25 but these have been refuted by others.26 Because of the lack of evidence-based support for the off-label use of dexmedetomidine infusions in the non-ICU setting, we set the basal rate at 0.2 μg kg−1 h−1 with a maximum limit of 0.6 μg kg−1 h−1. This is below the manufacture's recommended dosage. Also, the 0.5 μg kg−1 initial loading dose that we administered has been shown to provide sufficient analgesia without the development of clinically significant hypotension or bradycardia in spontaneously breathing healthy adults.27 Nevertheless, HR and MAP were lower in the dexmedetomidine group, and although it did not attain statistical significance in the current study, the proportion of patients experiencing significant hypotension and bradycardia in the dexmedetomidine group was almost double that in the control group. The data from one patient in the dexmedetomidine group was excluded from analysis due to the development of persistent hypotension despite the administration of fluid and ephedrine and this necessitated early cessation of the PCA. The study was not sufficiently powered to address the occurrence of serious adverse events and the possibility that PCA with a dexmedetomidine-opioid mixture may endanger patients with limited cardiovascular reserve could not be ruled out from the result of this study.
In summary, in a group of patients susceptible to PONV undergoing lumbar spinal fusion, a loading dose of dexmedetomidine (0.5 μg kg−1) followed by a continuous infusion as an adjunct to a fentanyl-based PCA drug mixture was beneficial in terms of nausea, and this was accompanied by a significant reduction in opioid requirement without compromising pain management. However, considering its limited antiemetic efficacy and the potential risks, further studies are required before any firm conclusions on the safety and efficacy of a 48-h infusion of dexmedetomidine can be reached.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: none.
Conflicts of interest: none.
1. Cohen MM, Duncan PG, DeBoer DP, Tweed WA. The postoperative interview: assessing risk factors for nausea and vomiting. Anesth Analg
2. Apfel CC, Laara E, Koivuranta M, et al. A simplified risk score for predicting postoperative nausea and vomiting: conclusions from cross-validations between two centers. Anesthesiology
3. Hill RP, Lubarsky DA, Phillips-Bute B, et al. Cost-effectiveness of prophylactic antiemetic therapy with ondansetron, droperidol, or placebo. Anesthesiology
4. Gan TJ, Diemunsch P, Habib AS, et al. Consensus guidelines for the management of postoperative nausea and vomiting. Anesth Analg
5. Lichtor JL. Nausea and vomiting after surgery: it is not just postoperative. Curr Opin Anesthesiol
6. Elia N, Lysakowski C, Tramer MR, Phil D. Does multimodal analgesia with acetaminophen, nonsteroidal antiinflammatory drugs, or selective cyclooxygenase-2 inhibitors and patient-controlled analgesia morphine offer advantages over morphine alone? Meta-analyses of randomized trials. Anesthesiology
7. Clarke H, Pereira S, Kennedy D, et al. Adding Gabapentin to a multimodal regimen does not reduce acute pain, opioid consumption or chronic pain after total hip arthroplasty. Acta Anaesthesiol Scand
8. Carstensen M, Moller AM. Adding ketamine to morphine for intravenous patient-controlled analgesia for acute postoperative pain: a qualitative review of randomized trials. Br J Anaesth
9. Kim SH, Shin YS, Oh YJ, et al. Risk assessment of postoperative nausea and vomiting in the intravenous patient-controlled analgesia environment: predictive values of the Apfel's simplified risk score for identification of high-risk patients. Yonsei Med J
10. Song JW, Shim JK, Song Y, et al. Effect of ketamine as an adjunct to intravenous patient-controlled analgesia, in patients at high risk of postoperative nausea and vomiting undergoing lumbar spinal surgery. Br J Anaesth
11. Gerlach AT, Dasta JF. Dexmedetomidine: an updated review. Ann Pharmacother
12. Iirola T, Ihmsen H, Laitio R, et al. Population pharmacokinetics of dexmedetomidine during long-term sedation in intensive care patients. Br J Anaesth
13. Blaudszun G, Lysakowski C, Elia N, Tramer MR. Effect of perioperative systemic alpha 2 agonists on postoperative morphine consumption and pain intensity systematic review and meta-analysis of randomized controlled trials. Anesthesiology
14. Lin TF, Yeh YC, Lin FS, et al. Effect of combining dexmedetomidine and morphine for intravenous patient-controlled analgesia. Br J Anaesth
15. Malviya S, Voepel-Lewis T, Ludomirsky A, et al. Can we improve the assessment of discharge readiness? A comparative study of observational and objective measures of depth of sedation in children. Anesthesiology
16. Choi YS, Shim JK, Yoon do H, et al. Effect of ramosetron on patient-controlled analgesia related nausea and vomiting after spine surgery in highly susceptible patients: comparison with ondansetron. Spine
17. Costantini R, Affaitati G, Fabrizio A, Giamberardino MA. Controlling pain in the postoperative setting. Int J Clin Pharmacol Ther
18. Watcha MF, White PF. Postoperative nausea and vomiting. Its etiology, treatment, and prevention. Anesthesiology
19. Zhang X, Bai X. New therapeutic uses for an alpha 2 adrenergic receptor agonist - dexmedetomidine in pain management. Neurosci Lett
20. Kauppila T, Kemppainen P, Tanila H, Pertovaara A. Effect of systemic medetomidine, an alpha 2 adrenoceptor agonist, on experimental pain in humans. Anesthesiology
21. Zheng Y, Cui S, Liu Y, et al. Dexmedetomidine prevents remifentanil-induced postoperative hyperalgesia and decreases spinal tyrosine phosphorylation of N-methyl-d-aspartate receptor 2B subunit. Brain Res Bull
22. Mansikka H, Zhou L, Donovan DM, et al. The role of mu-opioid receptors in inflammatory hyperalgesia and alpha 2-adrenoceptor-mediated antihyperalgesia. Neuroscience
23. Ulger F, Bozkurt A, Bilge SS, et al. The antinociceptive effects of intravenous dexmedetomidine in colorectal distension-induced visceral pain in rats: the role of opioid receptors. Anesth Analg
24. Hall JE, Uhrich TD, Barney JA, et al. Sedative, amnestic, and analgesic properties of small-dose dexmedetomidine infusions. Anesth Analg
25. Gerlach AT, Murphy CV. Dexmedetomidine-associated bradycardia progressing to pulseless electrical activity: case report and review of the literature. Pharmacotherapy
26. Klinger RY, White WD, Hale B, et al. Hemodynamic impact of dexmedetomidine administration in 15,656 noncardiac surgical cases. J Clin Anesth
© 2016 European Society of Anaesthesiology
27. Cortinez LI, Hsu YW, Sum-Ping ST, et al. Dexmedetomidine pharmacodynamics: Part II: Crossover comparison of the analgesic effect of dexmedetomidine and remifentanil in healthy volunteers. Anesthesiology