Secondary Logo

Journal Logo


Sedative, Amnestic, and Analgesic Properties of Small-Dose Dexmedetomidine Infusions

Hall, Judith E. MA, FRCA; Uhrich, Toni D. MS; Barney, Jill A. MS; Arain, Shahbaz R. MD; Ebert, Thomas J. MD, PhD

Author Information
doi: 10.1097/00000539-200003000-00035
  • Free


Dexmedetomidine has a relatively high ratio of α21-activity (1620:1 as compared with 220:1 for clonidine) (1) and, therefore, is considered a full agonist of the α2 receptor. This may result in more potent effects of sedation without unwanted cardiovascular effects from α1 receptor activation. The 2-h half-life of dexmedetomidine is nearly 4-fold shorter than that of clonidine (2), which increases the likelihood that a continuous infusion of dexmedetomidine might be useful for sedation. Dexmedetomidine also has minimum alveolar anesthetic concentration (MAC)-sparing properties (3), but its use as an anesthetic adjuvant has been complicated by persistent hypotension that has mandated IV fluid administration and vasopressor administration (4,5). In addition, its use in large doses is complicated by hypertension from α2 receptor-mediated vascular constriction (6).

We investigated two relatively small doses of dexmedetomidine that have few cardiovascular effects. The purpose of this research in volunteers was to evaluate the effects of dexmedetomidine on sedation, cognitive function, analgesia, and cardiorespiratory function to gain insight about the safety and efficacy of dexmedetomidine for its proposed use in postoperative sedation.


The study design included three sessions in which subjects received, in random order, either placebo (saline load and infusion), or a 6-μg · kg−1 · h−1 dexmedetomidine load followed by infusions of either 0.2 or 0.6 μg · kg−1 · h−1 dexmedetomidine (“small” and “moderate” doses, respectively). After institutional review board approval, the subjects provided written, informed consent and were screened. Inclusion criteria were men or women between the ages of 18 and 35 yr who were free of systemic disease. Exclusion criteria included a positive urine test for illicit drugs, alcohol, or pregnancy; exposure to an experimental drug within the past 3 mo or to dexmedetomidine within the previous year; history of serious adverse reaction or allergy to any drug; body weight greater than 25% of recommended for height; asthma that required therapy on a scheduled basis; inability to complete psychomotor tests; an abnormal electrocardiogram; or use of α2 agonists or antagonists.

The subjects had fasted for at least 8 h before the study. Before each study session, urine was tested again for absence of alcohol and drugs and, in women, to confirm a negative urine pregnancy test. In addition, complete blood count, electrolyte, and liver function tests were done before each research session to document normalcy. Cardiorespiratory monitoring included electrocardiogram for the determination of heart rate (HR) and monitoring ST segment in leads II and V, oscillometric blood pressure measurement (mean arterial pressure [MAP]), finger tip pulse oximetry for SpO2, and nasal cannula for ETCO2 monitoring and determination of respiratory rate. An IV catheter was placed for administration of dexmedetomidine, and sedation/hypnosis was monitored via bispectral electroencephalogram analysis (BIS). All monitors and catheters were placed on the nondominant hand, leaving the dominant hand free for written tests and the cold pressor test (CPT).

A 10-min rest period followed instrumentation, after which baseline measurements (HR, MAP, SpO2, CO2, respiratory rate, and BIS) were collected. This quiet period was followed by tests that assessed alertness and cognition during resting conditions and pain and hemodynamic responses to the CPT. After baseline testing, a 10-min initial dose of 6 μg · kg−1 · h−1 dexmedetomidine or saline (placebo trial) was given, followed by a 50-min infusion of 0.2 or 0.6 μg · kg−1 · h−1 dexmedetomidine or saline (placebo trial). Both the investigators and the subjects were blinded to the randomization schedule. Measurements and testing were repeated at the end of infusion and at 1 and 4 h after termination of the infusion. Additional sedation (Observer Assessment of Alertness/Sedation [OAA/S] and visual analog scale [VASSEDATION], see below) and cardiorespiratory measurements were obtained at 10 and 30 min after beginning the infusion, and 2 h after terminating the infusion. Further, BIS was recorded at 10-min intervals during the infusion, at 30-min intervals for the first 2 h of recovery, and every hour thereafter. In addition, BIS was averaged during the time of the various tests.

There were three measures of sedation. The first was objective, and it was derived from the processed electroencephalogram signal, which indicates state of wakefulness (bispectral index or BIS). BIS was recorded continuously throughout the study, and scores reported were taken before stimulation by the investigators. The second measure was assessment by an independent investigator using the OAA/S. The OAA/S rates a subject’s alertness by using a variety of categories, i.e., their responsiveness to verbal stimuli, speech, facial expression, and eyes. The same research investigator determined this value at all time points and in all subjects. Finally, each subject was asked to assess their level of sedation by sliding a movable indicator line between the two endpoints of a VAS (scale from 0–100, where 0 = “asleep” and 100 = “wide awake”; VASSEDATION).

Cognition was tested by using a digit symbol substitution test (DSST) and a memory recall test (MEM). The DSST is a timed test of psychomotor performance that presents the subject with a chart of eight rows of digits in random sequence. A key of the nine numerical digits and their corresponding symbols is provided. The subject’s task is to replace the digit with its corresponding symbol. A different test form is given at each time point. The number of correct substitutions within a 60-s period was tallied. The MEM test consisted of listening to a series of 16 unrelated words and immediately reciting as many as possible within 1 min. At the 4-h postinfusion time point, subjects performed a comprehensive memory test (CMEM). Rather than listening to a new word list, the CMEM test required them to recall as many of the words as possible from the three previous lists. The number recalled from each list was identified, and the total number recalled was tallied.

The CPT consisted of immersion of the subject’s hand into ice water for 1 min. HR and MAP were captured at the end of the 1-min period (from 45 to 60 s), throughout which the ice water was agitated. The subjects assessed their pain immediately after the cold exposure with another VAS (0, “no pain” to 100, “worst pain imaginable”; VASPAIN). There was at least 1 wk between dosing sessions.

Population statistics were mean ± SEM. Repeated measures analysis of variance (ReANOVA) was used to determine dose, time effects, and interaction terms (response of each dose over time) for each measured variable. When appropriate, the Scheffépost hoc test was performed. Because the CMEM test at 4 h was a single test (not repeated over time), a Student’s t-test was used to determine differences between groups. Significance was established at P < 0.05.


Eight healthy subjects, aged 23 to 31 yr (mean = 25.4 yr), were enrolled. Four men and three women completed the full study; one subject attended only one of the three sessions, and that subject’s data were not included in this analysis. The seven subjects had a mean weight of 72 kg (range 52–90 kg) and height of 171 cm (range 155–193 cm).

Sedation Measures

Both small and moderate doses of dexmedetomidine produced significant sedation (VASSEDATION) compared with placebo. However, sedation achieved after the 60-min infusion was not different between dexmedetomidine doses, increasing 61% and 76% in small and moderate dose groups, respectively (Figure 1). The OAA/S scores decreased 31% and 37%, and BIS scores decreased 31% and 36% after the 60-min dexmedetomidine infusion in small and moderate dose groups, respectively (Figure 1). ReANOVA indicated that, compared with placebo, dexmedetomidine had significant time effects and dose-response (interaction) terms for VASSEDATION, OAA/S, and BIS. Figure 2 shows the BIS scores before and after the subjects were aroused at the end of the infusion period to assess sedation and perform the various tests. Once stimulated (verbally and/or physically), the subjects’ alertness returned to baseline levels.

Figure 1
Figure 1:
Three measures of sedation taken before, during, and after dexmedetomidine infusion or placebo (saline). BIS is the bispectral index system (processed electroencephalogram).VASSEDATION is a visual analog scale rated by the volunteers as to their subjective sedation level with scale endpoints of 0 = asleep and 100 = wide awake. OAA/S is the Observer’s Assessment of Alertness/Sedation scale. Sedation was not different between the two dexmedetomidine doses. *Significant time effect established with post hoc analysis. †Significant dose effect. ¥Significant interaction term (P < 0.05).
Figure 2
Figure 2:
Bispectral index system (BIS; processed electroencephalogram) before (unstimulated) and after subjects were asked to perform various tasks including talking, rating alertness, a written cognitive test, a memory test, and a cold pressor test.

Cognitive Function

There was no significant dose effect between treatment groups for the MEM test, but there was a significant time effect reflected in the impaired recall at the end of 60-min infusion (post hoc testing) (Figure 3A). At 60-min infusion, there was a 37% and 46% reduction in word recall in the small and moderate infusion groups, respectively, compared with a 6% reduction in the placebo group. The recovery CMEM test at 4 h postinfusion was a test of total recall from the three previous word lists. No new list was presented. Recall in the placebo group was evenly distributed across all three lists (34%, 38%, and 32% for baseline, 60-min infusion, and 1 h recovery, respectively) (Figure 3b). However, recall of the word list presented at the end of the 60-min infusion was significantly reduced in the two dexmedetomidine groups (12% and 2% in the small and moderate dose groups, respectively, P < 0.05 compared with placebo).

Figure 3
Figure 3:
Memory testing of recalling lists of words heard before the administration of dexmedetomidine (baseline), at the end of the 1-h infusion, and at 1 h after termination of the infusion. A, Memory test (MEM) results were determined from the number of correctly recalled words immediately after hearing a 16-word list. At the end-of-infusion test, there were 37% and 46% reductions in recall in the small and moderate dose dexmedetomidine groups, respectively (time effect * =P < 0.05), compared with a 6% reduction in the placebo group. This memory impairment did not persist when evaluated 1 h into recovery. B, Comprehensive memory recall (CMEM) results were determined 4 h after terminating the infusions by quantifying the total number of words recalled from the three lists heard previously. Recalled words were then categorized into the list from which they were heard. Under the placebo condition, approximately the same number of words was recalled from each list (34%, 38%, and 32%). Fewer words were recalled from the list heard during the dexmedetomidine infusion (regardless of dose) than during the placebo trial (§P < 0.05 versus placebo). Dexmedetomidine infusions did not affect recall of words from the preinfusion baseline list.

There was no treatment effect for DSST, however there was a significant time effect and different responses (interaction) for the doses. Post hoc analysis showed differences between the dexmedetomidine infusion groups compared with the placebo at the 60-min infusion and 1-h recovery time points, but no differences in test results between dexmedetomidine infusion groups. At 60-min infusion, there were 28% and 41% decreases in performance of the DSST in the small and moderate dose dexmedetomidine groups, respectively, and after 1-h of recovery, responses remained 14% less in both dexmedetomidine groups compared with placebo (Figure 4). Performance returned to baseline at the 4-h recovery test.

Figure 4
Figure 4:
Psychomotor performance determined from the digital symbol substitution test indicated that dexmedetomidine reduced the number of correct substitutions at the end of infusion and 1 h into the recovery period. *Significant time effect established with post hoc analysis. ¥Significant interaction term. P < 0.05.


VASPAIN during CPT decreased significantly in both dexmedetomidine groups during the CPT at the 60-min infusion (approximately 30% lower than baseline) with some analgesia remaining up to the first hour of recovery (approximately 15% lower than baseline). Once again, both doses of dexmedetomidine were significantly different from placebo, but were not different from each other (Figure 5). ReANOVA revealed a different response pattern for VASPAIN, change in MAP, and change in HR. (Hemodynamic responses to the CPT are reported as the change from the steady state established just before the CPT.) The CPT-induced MAP increases at the 60-min infusion and 1-h recovery in the small and moderate dexmedetomidine groups were less than the placebo group response (Figure 5). The CPT MAP response was still significantly decreased 1 h after dexmedetomidine termination but had recovered to baseline at the 4-h recovery mark.

Figure 5
Figure 5:
Group responses to cold pressor test (CPT; hand immersed in ice water for 1 min). The heart rate and mean arterial pressure data presented are the difference (Δ) between the peak response (taken from 45–60 s into the test) and the data taken just before the CPT. Results from the visual analog scale for pain (VASPAIN) demonstrate that dexmedetomidine provided analgesia during and for 1 h after the infusion period. In addition, the mean blood pressure increase (but not the heart rate) was decreased with dexmedetomidine but not with placebo. *Significant time effect established with post hoc testing. †Significant dose effect. ¥Significant response over time (interaction term). P < 0.05.

Cardiorespiratory Function

We found small but significantly different MAP responses to dexmedetomidine versus placebo. When analyzed as a change from preinfusion baseline, MAP was significantly decreased at 60-min and 1- and 2-h recovery. No changes from baseline were noted with placebo infusions (Figure 6). There were significant, 20% and 16%, heart rate decreases from baseline during the 10-min initial dose in the small and moderate dose groups, respectively. The moderate dose, but not the small dose, of dexmedetomidine significantly increased HR compared with placebo (Figure 6).

Figure 6
Figure 6:
Cardiorespiratory variables before, during, and after dexmedetomidine or placebo infusion. Mean arterial pressure (MAP) decreased with dexmedetomidine at 60 min and at 1 and 2 h of recovery (*). There was a significant interaction (¥) term for MAP. There was a brief but significant bradycardia after the loading dose of dexmedetomidine (*) and a significant drug effect (†). There was a significant drug effect (†) on SpO2, although at no time did SpO2 fall below 95% in any group. Both ETCO2 and respiratory rate had significantly different responses over time (interaction term, ¥). Despite a significant time effect (*) on all three respiratory variables, post hoc testing failed to reveal where the differences were.

At no time did oxygen saturation decrease below 95% in any groups, although there was a statistically, but not clinically significant decrease in oxygen saturation in the small dose group compared with placebo (Figure 6). Respiratory function was well maintained. There were no dose effects on respiratory rate or ETCO2, however there were significant interaction terms (Figure 6). Although all three measured respiratory variables had significant time effects, post hoc analysis failed to reveal where these differences were.


The present randomized, double-blinded study demonstrated that both 0.2- and 0.6-μg · kg−1 · h−1 infusions of dexmedetomidine (small and moderate doses) produced significant sedation (VASSEDATION, OAA/S, BIS) that resolved two hours after terminating the infusions. Both infusions produced significant analgesia to the CPT and reduced performance on psychomotor tests. In addition, performance of the comprehensive memory test (four-hour recovery test, CMEM) demonstrated that the treatment groups recalled significantly fewer words from the “end of infusion” list than the placebo group. Importantly, retrograde memory impairment was not evident in the treatment groups, because recall of the word list given before dexmedetomidine was not affected by subsequent dexmedetomidine infusions. Cardiovascular stability and respiratory function were both well maintained and, at most measurement end points, not different from placebo. In addition, individuals were easily aroused from a sedative state and able to perform various tasks (which included verbal, motor, and cognitive skills).

There are a number of reasons that there is growing interest in the use of α2-adrenoceptor agonists as sedatives. One is the availability of a drug with a shorter half-life (dexmedetomidine). Secondly, in addition to sedation, there are other favorable effects of α2-adrenoceptor agonists, e.g., analgesia and maintained cardiorespiratory function. Finally, antagonists to the effects of α2-adrenoceptor agonists have been described that make quick reversal of sedation an option (7). These properties may prove useful for postoperative or intensive care unit sedation, but both safety and efficacy data are needed. We studied human volunteers in good health to establish some of the dose-effect relationships of relatively low concentrations of dexmedetomidine given as a continuous infusion. These concentrations were chosen based on our earlier work that evaluated considerably larger doses of dexmedetomidine. At larger concentrations, dexmedetomidine resulted in increases in blood pressure and decreases in HR and cardiac output (T. J. Ebert, written communication, 1999) (3). These effects have been ascribed to direct vascular effects from α2-adrenoceptor agonism combined with inhibition of cardiac sympathetic drive. We used doses of dexmedetomidine that were predicted to have few cardiovascular effects, but still might be sufficient to produce sedation and analgesia. In fact, the sedation measures were similarly increased with either dose of dexmedetomidine, and no clear advantage could be determined between doses for analgesia to the CPT.

Sedation and analgesia probably account for the MAC-sparing effects of this class of compound. Central α2-adrenoceptors in the locus ceruleus (8) and receptors in the dorsal horn of the spinal cord are likely involved in these effects (9). The MAC reduction from dexmedetomidine is much greater than with clonidine, presumably because of the greater specificity of dexmedetomidine for the α2-adrenoceptor. It reduces halothane requirements by up to 90% in rats (3,10), whereas clonidine reduces halothane requirements by 48% (11). We recorded BIS numbers as low as 37–49 (in two subjects) and an average BIS of 66 during the infusion period. During anesthesia administration, similar BIS numbers would be indicative of a state of moderate to deep sedation/hypnosis highly unlikely to be associated with recall (12). However, unlike general anesthesia, the volunteers were readily awakened from deep hypnosis simply by talking to them, and the BIS returned to awake levels. Despite this, the CMEM test, administered four hours after terminating the infusions, indicated that the word list presented to the volunteers at the end of either dose of dexmedetomidine was consistently difficult to recall. This was not the case in volunteers receiving placebo infusions.

The DSST was used to assess psychomotor performance, and the results are consistent with an earlier study (13). It was clear that, even though the dexmedetomidine-treated volunteers could be easily awakened to perform the testing, their performance was impaired. This persisted for at least one hour after terminating the dexmedetomidine infusions, but was restored to baseline at the four-hour postinfusion testing period. There were no differences in the degree of performance impairment or the time course of impairment of either the DSST or CMEM tests between the two infusion doses of dexmedetomidine.

A biphasic cardiovascular response has been described when dexmedetomidine is given as an IV bolus. A 1-μg/kg bolus results in a transient increase in blood pressure and a reflex decrease in heart rate (6,14). The increase in blood pressure began one minute after the bolus and was attributed to the direct effects of α2-adrenoceptor stimulation of vascular smooth muscle. After the transient increase in BP, a decrease in BP occurred, presumably caused by an inhibition of sympathetic outflow that overrode the direct effects of dexmedetomidine on the vasculature. Despite an initial dose over a 10-minute period, a transient, but nonsignificant increase in MAP (7%) was still observed and associated with a significant decrease in HR (16%–18%). This might be an unavoidable effect of infusing α2 agonists, because the time differential between directly binding to vascular receptors and diffusion into the central nervous system to decrease sympathetic outflow during IV infusions might be ever present. Earlier work with oral administration of dexmedetomidine in animals suggested that this time differential was eliminated, and as a result, hypertension was not seen (15).

The respiratory effects of dexmedetomidine have been greatly debated (16–19), but the consensus appears to be that dexmedetomidine is associated with little respiratory depression. This study confirmed a lack of a clinically significant respiratory effect. Belleville et al. (20) reported that dexmedetomidine could be associated with episodes of obstructive apnea, and this was increasingly common at doses of 1 and 2 μg/kg that were given for two minutes and presumably associated with a rapid increase in sedation. Obstructive apnea was not evident in our study. We cannot exclude the possibility that more rapid loading doses might cause irregular breathing or obstructive apnea, as described by Belleville et al. (20), but it is clear, based on receptor binding studies, that the α2 agonists should have little effect on respiration. Thus, an obstruction resulting in apnea is more likely related to the deep sedation and oral/pharyngeal anatomic events that are common to deep sleep.

We found that, in a small population of volunteers with healthy cardiovascular systems, small doses of dexmedetomidine provided sedation that could be easily reversed with verbal or physical stimuli. Dexmedetomidine also provided some analgesia and immediate (not retrograde) memory impairment. Cardiorespiratory variables were not significantly impaired. These properties might prove useful in a postoperative setting or in the intensive care unit.


1. Virtanen R, Savola JM, Saano V, Nyman L. Characterization of selectivity, specificity and potency of medetomidine as alpha 2-adrenoceptor agonist. Eur J Pharmacol 1988; 150:9–14.
2. Davies DS, Wing LMH, Reid JL, et al. Pharmacokinetics and concentration-effect relationships of intravenous and oral clonidine. Clin Pharmacol Ther 1977; 21:593–601.
3. Segal IS, Vickery RG, Walton JK, et al. Dexmedetomidine diminishes halothane anesthetic requirements in rats through a postsynaptic alpha-2 adrenergic receptor. Anesthesiology 1988; 69:816–23.
4. Jalonen J, Hynynen M, Kuitunen A, et al. Dexmedetomidine as an anesthetic adjunct in coronary artery bypass grafting. Anesthesiology 1997; 86:331–45.
5. Talke P, Jain U, Leung J, et al. Effects of perioperative dexmedetomidine infusion in patients undergoing vascular surgery. Anesthesiology 1995; 82:620–33.
6. Bloor BC, Ward DS, Belleville JP, Maze M. Effects of intravenous dexmedetomidine in humans, II: hemodynamic changes. Anesthesiology 1992; 77:1134–42.
7. Scheinin H, Aantaa R, Anttila M, et al. Reversal of the sedative and sympatholytic effects of dexmedetomidine with a specific alpha2-adrenoceptor antagonist atipamezole: a pharmacodynamic and kinetic study in healthy volunteers. Anesthesiology 1998; 89:574–84.
8. Guo TZ, Jiang JY, Buttermann AE, Maze M. Dexmedetomidine injection into the locus ceruleus produces antinociception. Anesthesiology 1997; 84:873–81.
9. De Kock M, Crochet B, Morimont C, Scholtes JL. Intravenous or epidural clonidine for intra- and postoperative analgesia. Anesthesiology 1993; 79:525–31.
10. Vickery RG, Sheridan BC, Segal IS, Maze M. Anesthetic and hemodynamic effects of the stereoisomers of medetomidine, an alpha 2-adrenergic agonist, in halothane-anesthetized dogs. Anesth Analg 1988; 67:611–5.
11. Kaukinen S, Pykkö K. The potentiation of halothane anaesthesia by clonidine. Acta Anaesthesiol Scand 1979; 23:107–11.
12. Glass PS, Bloom M, Kearse L, et al. Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997; 86:836–47.
13. Mattila MJ, Mattila ME, Olkkola KT, Scheinin H. Effect of dexmedetomidine and midazolam on human performance and mood. Eur J Clin Pharmacol 1991; 41:217–23.
14. Kallio A, Scheinin M, Koulu M, et al. Effects of dexmedetomidine, a selective alpha 2-adrenoceptor agonist, on hemodynamic control mechanisms. Clin Pharmacol Ther 1989; 46:33–42.
15. Sagen J, Proudfit HK. Evidence for pain modulation by pre- and postsynaptic noradrenergic receptors in the medulla oblongata. Brain Res 1985; 331:285–93.
16. Benhamou D, Veillette Y, Narchi P, Ecoffey C. Ventilatory effects of premedication with clonidine. Anesth Analg 1991; 73:799–803.
17. Ooi R, Pattison J, Feldman SA. The effects of intravenous clonidine on ventilation. Anaesthesia 1991; 46:632–3.
18. Penon C, Ecoffey C, Cohen SE. Ventilatory response to carbon dioxide after epidural clonidine injection. Anesth Analg 1991; 72:761–4.
19. Zornow MH. Ventilatory, hemodynamic and sedative effects of the alpha 2 adrenergic agonist, dexmedetomidine. Neuropharmacology 1991; 30:1065–71.
20. Belleville JP, Ward DS, Bloor BC, Maze M. Effects of intravenous dexmedetomidine in humans. Anesthesiology 1992; 77:1125–33.
© 2000 International Anesthesia Research Society