Dexmedetomidine, a highly selective alpha 2-adrenergic agonist, increases perioperative hemodynamic stability in healthy patients but decreases blood pressure and heart rate. The goal of this study was to evaluate, in a preliminary manner, the hemodynamic effects of perioperatively administered dexmedetomidine in surgical patients at high risk for coronary artery disease.
Twenty-four vascular surgery patients received a continuous infusion of placebo or one of three doses of dexmedetomidine, targeting plasma concentrations of 0.15 ng/ml (low dose), 0.30 ng/ml (medium dose), or 0.45 ng/ml (high dose) from 1 h before induction of anesthesia until 48 h postoperatively. All patients received standardized anesthesia and hemodynamic management. Blood pressure, heart rate, and Holter ECG were monitored; additional monitoring included continuous 12-lead ECG preoperatively, anesthetic concentrations and myocardial wall motion (echocardiography) intraoperatively, and cardiac enzymes postoperatively.
Preoperatively, there was a decrease in heart rate (low dose 11%, medium dose 5%, high dose 20%) and systolic blood pressure (low dose 3%, medium dose 12%, high dose 20%) in patients receiving dexmedetomidine. Intraoperatively, dexmedetomidine groups required more vasoactive medications to maintain hemodynamics within predetermined limits. Postoperatively, demedetomidine groups had less tachycardia (minutes/monitored hours) than the placebo group (placebo 23 min/h; low dose 9 min/h, P = 0.006; medium dose 0.5 min/h, P = 0.004; high dose 2.3 min/h, P = 0.004). Bradycardia was rare in all groups. There were no myocardial infarctions or discernible trends in the laboratory results.
Infusion of dexmedetomidine up to a targeted plasma concentration of 0.45 ng/ml appears to benefit perioperative hemodynamic management of surgical patients undergoing vascular surgery but required greater intraoperative pharmacologic intervention to support blood pressure and heart rate.
Key words: Dexmedetomidine: hemodynamics. Dose-effect. Heart: coronary artery disease. Sympathetic nervous system, alpha2-adrenergic agonist: dexmedetomidine.
SURGICAL and postoperative stress evoke an endocrine response that manifests as stimulation of the hypothalamus-pituitary-adrenal axis, renin-angiotensin axis, and the sympathetic nervous system. [1–3] Stimulation of the sympathetic nervous system increases the levels of circulating plasma norepinephrine and epinephrine, increasing blood pressure and heart rate [1,2] and the incidence of postoperative complications.  The hyperdynamic changes predispose the myocardium to ischemia, especially in the patient population with decreased reserve for coronary blood flow. Perioperative ischemia is associated with a significant increase in postoperative morbidity and mortality. [5,6] Attenuating the perioperative stress response could decrease the incidence of myocardial ischemia and thereby reduce the incidence of perioperative morbidity and mortality in patients at high risk for myocardial ischemia.
Several clinical studies suggest that alpha2-adrenergic agonists might be effective in blunting the perioperative stress response [7–9] and that clonidine may have perioperative antiischemic effects.  Dexmedetomidine is an alpha2-adrenergic agonist with a 10-fold greater alpha2/alpha1-receptor selectivity than clonidine.  In healthy volunteers, dexmedetomidine decreases circulating catecholamines by up to 90% and, like clonidine, has antinociceptive and sedative effects. [12–14] In healthy surgical patients, dexmedetomidine increases hemodynamic stability, decreases anesthetic requirements, and blunts the hyperdynamic response to intubation. [15–18] The sympatholysis also results in potentially adverse clinical effects, such as a decrease in blood pressure and bradycardia. Such hemodynamic changes might not be tolerated by patients with vascular disease or severe myocardial disease.
Thus far, dexmedetomidine has been administered only to healthy volunteers and healthy surgical patients. Therefore, to perform a preliminary evaluation of the feasibility and effects of perioperative administration of dexmedetomidine in high-risk surgical patients, we studied three consecutively increasing doses of an infusion of dexmedetomidine in vascular surgery patients, a population with a high incidence of coronary artery disease (CAD) who might benefit significantly from increased perioperative hemodynamic stability.
With approval from our Human Research Committee and written informed consent, we studied 25 patients with or at high risk for CAD who were scheduled for vascular surgery at the San Francisco Veterans Affairs Medical Center. Study entry criteria included one or more of the following: a history of classic angina pectoris; a history of myocardial infarction; electrocardiographic (ECG) evidence of Q waves typical of infarction without a history; CAD detected by angiography; or the presence of two or more risk factors for CAD, including cigarette smoking, treatment for hypertension, treated diabetes mellitus, or hypercholesterolemia (> 240 mg/dL). Excluded from study were patients with unstable angina, uninterpretable preoperative ECGs (left bundle branch block), patients taking clonidine or tricyclic antidepressants, and those who did not receive the study drug continuously for at least the first 24 postoperative hours. Cardiac medications were continued until the night of surgery.
The study was a double-blind, randomized, dose-escalation trial using three different doses of dexmedetomidine and placebo. Twenty-four patients were divided into three groups of eight to form low-, medium-, and high-dose test groups, each having six patients who received dexmedetomidine and two who received placebo. Thus, six patients received placebo during the study. The number of patients used in this study was not based on power calculations. Study began with the low-dose group, and once this dose was determined to be tolerable, proceeded to the medium-dose group, then, after the same determination, the high-dose group. Initially, 25 patients were enrolled in the study, but one (high-dose group) was excluded when dexmedetomidine was discontinued within 24 h of administration to permit an emergent return to surgery.
Dexmedetomidine was administered by a computer-controlled infusion pump (CCIP) targeting plasma concentrations of 0.15 ng/ml (low-dose), 0.30 ng/ml (medium-dose), and 0.45 ng/ml (high-dose). STAN-PUMP software (Steve Shafer, Stanford University, Palo Alto, CA) was used to run the infusion pump (Harvard Apparatus 22, Harvard Apparatus, South Natick, MA). The STANPUMP software updated the infusion rate at 10-s intervals using dexmedetomidine pharmacokinetic data  to allow drug delivery to targeted plasma concentrations. The infusion rate data were stored in the laptop computer, which was used to run the STANPUMP program. To study the effect of dexmedetomidine in awake and anesthetized patients, infusion was begun 1 h before induction of anesthesia and continued throughout the intraoperative period and for 48 h postoperatively. The average amount of dexmedetomidine infused was 2.64 micro gram/kg (range 2.30–3.75 micro gram/kg), 5.31 micro gram/kg (range 4.40–5.97 micro gram/kg), and 8.03 micro gram/kg (range 5.57–9.87 micro gram/kg) for the low-, medium-, and high-dose groups, respectively. Patients were not permitted to ambulate during study drug infusion.
The night before surgery, patients received 2 mg lorazepam orally. In the operating room, study drug infusion was started 1 h before induction of anesthesia. Patients then breathed oxygen while anesthesia was induced with alfentanil (up to 30 micro gram/kg) and thiopental (up to 3 mg/kg). Vecuronium (0.1 mg/kg) was administered to achieve muscle relaxation before tracheal intubation and as needed thereafter. Anesthesia was maintained with a continuous alfentanil infusion of 0.5 micro gram *symbol* kg sup -1 *symbol* min sup -1 and 70% N2O in oxygen. Isoflurane (0–2%) was administered to maintain hemodynamics within predetermined limits. Isoflurane and alfentanil were discontinued 10 min before the anticipated end of surgery. Nitrous oxide was discontinued at the end of surgery, and residual muscle relaxation was antagonized using neostigmine (60 micro gram/kg) and glycopyrrolate (8 micro gram/kg).
Baseline blood pressure and heart rate were determined by averaging three noninvasive measurements taken at least 5 min apart the night before surgery. Arterial cannula was inserted the morning of surgery, and hemodynamic monitoring was started at least 5 min before the study drug infusion. During the study drug infusion, preceding induction of anesthesia, hemodynamic changes were treated only if clinically judged to be detrimental to the patient's health.
Intraoperative Hemodynamic Management
Systolic blood pressure and heart rate were maintained within predetermined limits. Intraoperatively, bradycardia was defined as heart rate < 40 beats/min. tachycardia as greater or equal to 20% increase in heart rate. hypertension as greater or equal to 20% increase in systolic blood pressure, and hypotension as greater or equal to 20% decrease in systolic blood pressure, from preoperative noninvasive baseline values.
After establishing an adequate level of anesthesia (lack of movement, sweating, lacrimation) with isoflurane, tachycardia was treated by administration of esmolol, bradycardia by administration of glycopyrrolate, and hypertension by increasing the inspired isoflurane concentration. Hypotension initially was treated by decreasing the inspired isoflurane concentration; if persistent, by administration of phenylephrine. Inadequate level of anesthesia (spontaneous movement, sweating, lacrimation), as determined by the anesthesiologist, was treated by increasing the inspired isoflurane concentration. Arterial hemoglobin oxygen saturation (SpO2) less than 95% was managed by increasing the inspired oxygen concentration.
Postoperative Hemodynamic Management
Postoperatively, tachycardia was defined as heart rate greater or equal to 100 beats/min, bradycardia as heart rate < 40 beats/min, hypertension as systolic blood pressure > 160 mmHg, and hypotension as systolic blood pressure < 90 mmHg.
Postoperative tachycardia was treated by administration of esmolol, bradycardia with glycopyrrolate, hypertension with nitroprusside, and hypotension with phenylephrine. SpO2of less than 95% was treated by increasing the inspired oxygen concentration.
Postoperatively, routine medications (including oral nitrates) were restarted as clinically indicated. Intravenous nitroglycerin was used only to treat documented (12-lead ECG) myocardial ischemia.
Clinical Data Collection
Arterial blood pressure (systolic, diastolic, and mean) and heart rate were continuously measured during the 1-h infusion period preceding induction of anesthesia, intraoperatively, and for 60 h postoperatively, using a Marquette 7000 hemodynamic monitor (Marquette Electronics, Milwaukee, WI). Arterial blood pressure was measured via a radial artery cannula connected to a Transpac II transducer (Abbott Laboratories, North Chicago, IL) zeroed 5 cm posterior to the sternum. Hemoglobin oxygen saturation was measured noninvasively using an Ohmeda Biox 3700 pulse oximeter (Ohmeda, Luisville, CO) with the probe placed on a distal phalanx. End-tidal isoflurane concentrations were measured using a Chemetron Medspect mass spectrometer (PPG, Kansas City, KS). Hemodynamic and end-tidal isoflurane data were recorded at 1-min intervals from the clinical monitors through an automated data acquisition system (ARKIVE Series 2000, Diatek, San Diego, CA). SpOsub 2 data were recorded every minute intraoperatively and hourly postoperatively.
Blood samples for analysis of hematologic and biochemical values and urine for urinalysis were collected preoperatively and on postoperative days 1, 5, and 14 or on discharge from the hospital, whichever was sooner.
The samples were analyzed at the central laboratory of the San Francisco Veterans Affairs Medical Center Hospital.
A 12-lead ECG was obtained and reviewed before surgery and on postoperative days 1, 2, 3, and 5, then weekly, on discharge, and whenever clinically indicated. Cardiac enzymes were measured before surgery and every 12 h for 72 h postoperatively. Total CK levels were analyzed from all samples. CK-MB was analyzed in samples having a total CK above normal limits (235 U/l). All adverse cardiac outcomes (cardiac death, myocardial infarction, unstable angina, congestive heart failure, and life-threatening dysrhythmia) were recorded. Myocardial infarction was defined by the presence of new Q waves on the standard 12-lead ECG and/or positive myocardial enzymes. Positive myocardial enzymes required total CK-MB > 100 ng/ml for the first 12 postoperative hours and CK-MB > 70 ng/ml thereafter.
All intraoperative and postoperative medications, intravenous fluids, urinary output, and blood loss were recorded. All adverse events were recorded.
Holter ECG Monitoring
ECG (Holter) recordings were obtained continuously for at least 8 h preoperatively, during the 1-h drug infusion preceding induction of anesthesia, intraoperatively, and for 96 h postoperatively using a three-channel AM Holter ECG recorder (Marquette, series 8500). Three bipolar leads, CM5, CC5, and ML, were used. Before study, ECG tracings were obtained with the patient in the supine, upright, and left and right lateral decubitus positions. All recordings were analyzed by a technician and verified by a study physician blinded to patient treatment group.
After exclusion of all abnormal QRS complexes, the ST segment was trended continuously for the three leads. The baseline ST-segment level was defined as the average ST segment during a stable period (usually 15–60 min) preceding each ischemic episode. An ischemic episode was defined as: (1) horizontal or down-sloping ST-segment depression from baseline greater or equal to 1 mm lasting at least 1 min or (2) ST-segment elevation from baseline greater or equal to 2 mm measured at the J point. ST-segment depression was measured 60 ms after the J point unless that point fell within the T wave, in which case the point of ST-segment depression measurement was shortened to a minimum of J plus 40 ms. If T-wave changes occurred with ST depression. J-point depression from baseline of at least 1 mm also was required.
For each episode, the maximum ST-segment change from baseline and duration of ischemic episode were measured. Because the varying amount of time intraoperatively that a given patient is monitored can result in differences in the absolute amount of ischemia detected, an ischemic burden (total ischemic min/total h monitored) was computed for each patient.
Continuous 12-Lead ECG
The ECG was continuously monitored during the 1-h preinduction drug infusion using a microcomputer-augmented continuous 12-lead system (MAC II/ST exercise stress system, Marquette). After entry into the operative holding area but before entry into the operating room, patients were fitted with 10 silver/silver chloride ECG electrodes. A modified Mason-Likar lead configuration was used, with the leg electrodes placed midway between the anterior and posterior iliac spines. The six precordial leads were carefully placed in the standard positions in all patients. The frequency response, operating characteristics, and method of analysis of this system have been reported previously. .
After tracheal intubation, a 5.0-MHz 64-element miniaturized phased-array transesophageal echocardiography (TEE) transducer probe (GE RT6800, Milwaukee, WI) was inserted. The left ventricular short-axis view, at the level of the midpapillary muscles, was continuously recorded on half-inch VHS videotape. A study technician was present throughout surgery to maintain the optimal image and position. The probe was removed as close to completion of skin closure as possible
Tapes were edited to obtain samples for analysis of 65-s duration every 15 min and at the following predetermined times: 5 min before and at skin incision; 5 min before, at onset of, and 5 min after clamping and unclamping of major blood vessels; and 5 min before, at onset of, offset of, and 5 min after an increase in systolic blood pressure > 150 mmHg or a decrease in heart rate < 50 beats/min. Edited tapes were analyzed by a blinded investigator: The short-axis cross-sectional image was divided into four segments using the papillary muscles as the guide. Segmental wall motion was assessed only if > 70% of its entire endocardial outline was visible throughout systole and diastole. Wall motion was graded as: 0 = normal, 1 = mild hypokinesis, 2 = severe hypokinesis with myocardial thickening, 3 = akinesis, and 4 = dyskinesis. Image position and quality also were graded.
After all TEE images were scored, baseline wall motion for each quadrant was determined using the most normal wall motion in each quadrant at any time (minimum 1-min duration) during TEE monitoring. When regional wall motion deteriorated, the following criteria were used to define an echocardiographic ischemic episode: (1) deterioration in wall motion greater or equal to 2 grades and (2) deterioration from akinesis to dyskinesis (in a segment with normal wall thickness). Apparent deterioration in wall motion accompanied by a major shift in image position (i.e., from midpapillary to basal short-axis) was not considered significant (to avoid including preexisting wall-motion abnormalities at different anatomic levels).
Onset of an ischemic episode began when a significant change first occurred in one or more quadrants and ended when the last quadrant normalized by at least one grade toward baseline. Interexamination and intraobserver variabilities in our laboratory have been reported previously. .
Postoperative analgesia was provided by intravenous morphine sulfate, delivered by a patient-controlled analgesia (PCA) pump. The initial PCA setting was a 1-mg bolus dose with a lock-out interval of 6 min. For inadequate analgesia, additional 2-mg doses of morphine were administered intravenously as needed. If analgesia remained inadequate after the additional 2-mg bolus doses, the PCA dose was increased in increments of 0.5 mg.
Analgesia was assessed using a visual analog scale (VAS), comprised of a 100-mm horizontal line with one pole representing “no pain” and the other “worst pain imaginable.” The scale was administered every 4 h postoperatively for the first 48 h, as long as the patient was awake. Patients assessed their pain at rest and rated the severity of the worst pain since the last assessment.
To determine the intraoperative isoflurane requirements, the curve for the area under the end-tidal isoflurane concentration versus time was integrated and divided by the intraoperative time. Hemodynamic response to the 1-h dexmedetomidine infusion preceding induction of anesthesia was analyzed by comparing the values (5-min average) for systolic blood pressure and heart rate immediately before infusion with those obtained 1 h after the start of infusion. The response to intubation was analyzed by comparing the hemodynamic values immediately preceding intubation to the highest values obtained within 5 min after intubation. Hemodynamic response to emergence from anesthesia was analyzed by comparing preoperative noninvasive baseline values for systolic blood pressure and heart rate with the highest values obtained after discontinuation of anesthesia, before leaving the operating room.
The percent change in rate-pressure product (HR*SBP) in response to the 1-h dexmedetomidine infusion preceding induction of anesthesia was calculated by comparing the rate-pressure product (5-min average) at the end of the 1-h infusion to the rate-pressure product immediately preceding the study drug infusion. The percent change in rate-pressure product in response to intubation was calculated by comparing the highest rate-pressure product within 5 min after intubation to the rate-pressure product immediately preceding (5-min average) the study drug infusion and in response to emergence by comparing the highest rate-pressure product after discontinuation of anesthesia, before leaving the operating room, to the rate-pressure product immediately preceding (5-min average) the study drug infusion.
Nonparametric tests were used to analyze the data unless otherwise indicated. Kruskal-Wallis test was used to analyze the overall comparisons among the four groups. When a significant difference was found, the placebo group was compared to the dexmedetomidine groups using the Wilcoxon two-sample test without correction for multiple comparisons. The percent changes in rate-pressure product were compared between the dexmedetomidine groups and the placebo group using one-way analysis of variance and Dunnett's post hoc test. Descriptive statistics were used for demographic and ischemia data because of the small number of patients in each group. Data are reported as the mean plus/minus SD. P < 0.05 identified statistical significance. All statistical calculations were performed rising PC SAS 6.04 (SAS Institute, Cery, NC) software.
The clinical and demographic data were similar for all groups (Table 1). All 24 patients were men. The clinical course of the patient who was withdrawn from study was uneventful until his emergent return to surgery for revascularization. The induction doses of alfentanil, thiopental, and the intraoperative isoflurane requirements for the dexmedetomidine groups were not different from the placebo group (Table 1).
Both heart rate and systolic blood pressure decreased in response to the 1-h dexmedetomidine infusion (Figure 1). Compared with the placebo group, the decrease in heart rate was significant for the low-dose (P = 0.037) and high-dose (P = 0.004) groups (Figure 1), and the decrease in systolic blood pressure was significant for the medium-dose (P = 0.01) and high-dose (P = 0.004) groups (Figure 1). In the high-dose group, the heart rate decreased from 79 plus/minus 13 beats/min to 64 plus/min 15 beats/min (20% decrease), and the systolic blood pressure decreased from 168 plus/minus 27 mmHg to 133 plus/minus 27 mmHg (20% decrease). In the high-dose group, we observed an initial 2–11-mmHg increase in systolic blood pressure, followed by a decrease at 5 min after starting infusion.
Heart rate and systolic blood pressure increased in all patients in response to intubation. On the average, the heart rates increased by 32%, 22%, 43%, and 21% for the placebo, low-, medium-, and high-dose groups, respectively (P = NS). The respective increases in the systolic blood pressure were 51%, 35%, 26%, and 31%(P = NS).
During emergence from anesthesia, on the average, the heart rate values increased by 29%, 19%, 31%, and 0% above preoperative noninvasive baseline values and the systolic blood pressures increased by 27%, 31%, 9%, and 18% above preoperative noninvasive baseline values for the placebo, low-, medium-, and high-dose groups, respectively (P = NS).
The rate-pressure product decreased in response to the 1-h preinduction dexmedetomidine infusion (Figure 2). Compared to the placebo group, the decrease was significant for the medium-dose (P < 0.05) and high-dose (P < 0.05) groups. In response to intubation, the rate-pressure product of the dexmedetomidine groups were not different from placebo group (P = 0.19;Figure 2). The rate-pressure product increased in response to emergence from anesthesia (Figure 2), the high-dose group being significantly different from the placebo group (P < 0.01).
Postoperatively, during dexmedetomidine infusion, the mean heart rates of the high-dose group were less compared to the placebo group (P = 0.036). None of the other mean heart rate or systolic blood pressure values of the dexmedetomidine groups differed significantly from those in the placebo group (Table 2).
The amount of bradycardia was negligible in all groups (Figure 3). Intraoperatively, the total amount of tachycardia was 0.56 min/monitored h in the high-dose group and 8.71 min/monitored h in the placebo group (P = NS;Figure 3). Postoperatively, during study drug infusion, the dexmedetomidine groups had significantly less tachycardia than the placebo group (P = 0.006 placebo vs. low-dose, P = 0.004 placebo vs. medium-dose, P = 0.004 placebo vs. high-dose;Figure 3). This difference was not maintained after discontinuation of dexmedetomidine, after which the amount of tachycardia increased. No trends could be detected in the amount of hypertension or hypotension in the dexmedetomidine groups versus placebo group (Figure 4).
There were no significant changes in blood pressure or heart rate during the first 12 h after dexmedetomidine infusion (Table 2).
SpO2during dexmedetomidine infusion did not differ significantly from values in the placebo group. None of the patients had SpO2less than 85%. Intra- and postoperatively, the SpOsub 2 was less than 90% < 0.05% of the time in all groups.
Intraoperatively, most patients received phenylephrine to maintain systolic blood pressure within predetermined limits: 4 of 6 in the placebo group; five of six in the low-dose group, six of six in the medium-dose group, and six of six in the high-dose group (Table 3). Anticholinergic drugs were required to treat bradycardia intraoperatively in 44% of the patients receiving dexmedetomidine and none of the patients in the placebo group (Table 3). The total amounts of vasoactive medications used for each group did not differ. Intraoperative intravenous fluid and blood-product requirements, urinary output, and blood loss also were similar among the four study groups. Postoperatively, during the study drug infusion, vasoactive drug requirements were minimal. Most interventions occurred in the placebo and low-dose groups, which required esmolol to control heart rate (Table 4). The amount of intravenous fluids and urinary output for each group did not differ postoperatively.
Holter ECG Monitoring
Fourteen of the 24 patients (58%; three subjects in the placebo group, four in the low-dose group, four in the medium-dose group, and three in the high-dose group) had a total of 87 ischemic episodes (Table 5). Three episodes (3%) occurred before study drug infusion, two intraoperatively (2%), 29 postoperatively during the study drug infusion (33%), and 53 during the 48 h after the study drug infusion (61%). Of all ischemic episodes, only 8% occurred in the medium-dose group and 9% in the high-dose group, compared to 29% in the placebo group. Of the two intraoperative ischemic episodes, one occurred during induction of anesthesia (in the high-dose group), and one during emergence from anesthesia (in the medium-dose group). Both were associated with significant increases in heart rate. Neither of these patients had postoperative ischemic episodes.
Postoperatively, the amount of myocardial ischemia was higher than preoperatively or intraoperatively for all groups (Table 5).
Of the 87 postoperative ischemic episodes, 82 were associated with heart rates greater than 90 beats/min at some point during the episode. All ischemic episodes were associated with heart rates greater than 70 beats/min.
Eight of 24 patients (33%) developed echocardiographic evidence of ischemia during the intraoperative period (two in the placebo, three in the low-dose, one in the medium-dose, and two in the high-dose groups). There was no difference in the severity of echocardiographic ischemia as measured by mean change in wall-motion score among the four study groups. None of the episodes of wall-motion abnormalities were associated with significant ST-segment depression.
During the 1-h dexmedetomidine infusion preceding induction of anesthesia, no severe bradycardia, new conduction defects, sinus pauses, new dysrhythmias, or ischemic episodes were detected from the continuous 12-lead ECG or Holter recordings.
No patient had a myocardial infarction according to the 12-lead ECG and CK-MB data.
Postoperatively, there were no cardiac-related mortality, episodes of unstable angina, sinus pauses, or life-threatening dysrhythmias. Radiologic signs of increased pulmonary interstitial fluid developed in four patients. In one patient (placebo group), the increase in pulmonary fluid content was attributed to excessive administration of intravenous fluids and, in another patient (medium-dose group), to pneumonia. The other two patients (medium- and high-dose groups), both with a history of congestive heart failure, were clinically judged to have signs of congestive heart failure. Three patients were treated postoperatively with intravenous nitroglycerin by the attending clinician: two patients in the placebo group received nitroglycerin during the study drug infusion, one for an ST-segment depression and one for an elevated CK-MB level; and one patient in the high-dose group received nitroglycerin after the study drug infusion for shortness of breath and onset of atrial fibrillation.
No discernible trends were noted in any of the biochemistry, hematology, or urinalysis results for the dexmedetomidine groups relative to the placebo group.
Sedation and Analgesia
After the 1-h infusion preceding induction of anesthesia, all patients in the medium- and high-dose groups fell asleep but were easily arousable. During the second postoperative day, there was no clinically observable sedation from the study drug. Postoperative VAS pain scores were similar among groups, and postoperative morphine requirements did not differ.
Ten patients had adverse events: three in the placebo group, and one, three, and three, respectively, in the low-, medium-, and high-dose groups. The ten patients had a total of 13 adverse events (Table 6). None of these events was life-threatening, and all could be accounted for by the patient's underlying disease (atrial fibrillation, transient ischemic attack) or complications expected secondary to the surgical procedures (pneumothorax, cellulitis).
Given as a premedication to healthy patients, dexmedetomidine blunts the hemodynamic response to intubation, increases hemodynamic stability, decreases anesthetic requirements, and decreases the level of circulating catecholamines by up to 90%. [15–18,22] These effects may benefit vascular surgery patients documented to have a high prevalence of CAD.  The sympatholysis caused by dexmedetomidine has potential to increase the incidence of bradycardia and hypotension. [17,22,24] The current study evaluated, in a preliminary manner, the effects of dexmedetomidine in high-risk patients undergoing vascular surgery. This study also is the first to administer dexmedetomidine as a continuous perioperative infusion over a 2-day period to cover the duration of most perioperative stress and hemodynamic lability.  Our results suggest that a perioperative dexmedetomidine infusion to a targeted plasma concentration of 0.45 ng/ml can be used in high-risk vascular surgical patients if other drugs are given to offset the depression of heart rate and blood pressure.
In accordance with the studies in healthy patients, we found that dexmedetomidine decreased blood pressure by 20% in vascular surgery patients. [15,16,26] Because the present study focused on vascular surgery patients with or at risk for CAD, one concern was that the decrease in sympathetic tone might result in an intolerably large decrease in blood pressure. However, there were no clinically adverse effects due to the decrease in blood pressure in any of our patients. Postoperatively, there was no dose-dependent decrease in blood pressure during the study drug infusion. It is possible that the postoperative stress response was sufficient to overcome the blood pressure-lowering effect of dexmedetomidine or that there is tachyphylaxis to the effects of dexmedetomidine.
Patients who are dependent on a high level of sympathetic activity or have a reduced myocardial function might not be able to tolerate the decrease in sympathetic tone by dexmedetomidine. Whether their sympathetic tone is decreased to the same extent by dexmedetomidine as in patients with normal sympathetic tone is not known. The current study did not have patients in this category, and future studies need to verify the safety of dexmedetomidine in these patients.
High dexmedetomidine plasma levels may increase blood pressure due to the effect of dexmedetomidine on peripheral alpha2receptors. Consistent with two previous reports, [18,26] we observed an initial increase in the systolic blood pressure with our high-dose group (0.45 ng/ml) within 2–3 min of the start of infusion. Because the increase in blood pressure is related to high peak plasma dexmedetomidine concentrations, it may be circumvented by slower administration of dexmedetomidine.
A high incidence of bradycardia and use of anticholinergic drugs have been reported in several studies using intravenous or intramuscular dexmedetomidine in healthy surgical patients. We administered dexmedetomidine for 1 h before induction of anesthesia to assess the effect of dexmedetomidine on myocardial conduction in unanesthetized, unstimulated patients. Although dexmedetomidine gradually decreased heart rate, there were no sudden episodes of bradycardia, new conduction blocks, or new dysrhythmias. Intraoperatively, 8 of 18 patients receiving dexmedetomidine required anticholinergic drugs for bradycardia (HR < 40 beats/min). Up to a 33% incidence of postoperative bradycardia has been reported with the use of dexmedetomidine in healthy surgical patients.  The low incidence of postoperative bradycardia (1 of 24 patients) in our study population could be due, in part, to differences in the sympathetic and parasympathetic tones or vagal responsiveness between our patients with preexisting vascular disease and young, healthy, surgical patients, or to an effect of postoperative stress sufficient to overcome some of the heart rate-lowering effect of dexmedetomidine.
The results of the current study indicate that, in vascular surgery patients, dexmedetomidine blunts the increase in rate-pressure product in response to emergence from anesthesia. However, the present study cannot demonstrate conclusively that dexmedetomidine attenuates the hemodynamic changes in response to emergence, because the variables (HR, SBP) used to calculate the rate-pressure product were not significantly affected by dexmedetomidine.
Intraoperatively, the patients receiving dexmedetomidine required frequent use of vasopressors and anticholinergic medications to maintain systolic blood pressure and heart rate within the predetermined limits. There were no adverse effects from the frequent intraoperative use of phenylephrine and no further need for blood pressure support postoperatively, suggesting that any decrease in postoperative sympathetic activity with dexmedetomidine was not significant enough to cause postoperative hypotension.
Recent studies of other alpha2agonists suggest that these agents may be effective therapies for postoperative myocardial ischemia.  Our study was limited by a small number of patients in which only a few patients in each group had myocardial ischemia. Our small sample size limited the statistical power to discriminate among the groups. Further studies in a larger number of patients will be needed to evaluate whether alpha2agonists have antiishemic effects, as has been suggested by preliminary studies using clonidine.
Anesthetic Requirements and Sedation
Aho et al. reported that a continuous intraoperative dexmedetomidine infusion can decrease the requirements for isoflurane by up to 90% in healthy patients.  Our intraoperative use of alfentanil and nitrous oxide provided sufficient anesthesia for our vascular surgery patients, such that isoflurane requirements were low in all groups. Therefore, we cannot evaluate the potential reductive effect of dexmedetomidine on anesthetic requirements in vascular surgery patients. A study with minimal background anesthesia will be required to achieve this.
Several studies have reported dose-dependent sedative effects with dexmedetomidine.  During the 1-h dexmedetomidine infusion preceding induction of anesthesia, the patients in our medium- and high-dose groups fell asleep but were easily arousable. Although the dexmedetomidine infusion had a sedative effect before induction, sedation was not observable the day after surgery. This is consistent with recent findings of tachyphylaxis to the anesthetic effects of dexmedetomidine in rats. .
Our study is limited by the small number of patients. This study was designed as a dose-finding pilot study in high-risk surgical patients and, as such, provides useful information on the safety and hemodynamic effects of three different infusion doses of dexmedetomidine in surgical patients with preexisting vascular disease. This study was not designed to detect all adverse cardiac outcomes nor does it have enough power to draw definite conclusions on the safety and anti-ischemic effects of dexmedetomidine. A large-scale study is necessary to validate these preliminary results and, based on our ischemia data, should be conducted using a targeted 0.45 ng/ml plasma dexmedetomidine concentration. Whether a greater dexmedetomidine concentration will provide additional benefits without excessive side effects remains to be determined.
Administration of vasoactive drugs based on hemodynamic values limits the interpretation of hemodynamic data. In addition to the hemodynamic values, the hemodynamic variability was reflected in the quantity of vasoactive medications required to treat changes in hemodynamics that exceeded our protocol-defined limits.
We infused dexmedetomidine for 48 h postoperatively, when the natural incidence of myocardial ischemia is greatest.  A briefer infusion might have been as effective and should be investigated. Although a longer infusion period might benefit some patients, it would not be appropriate for those who could not ambulate and would increase their risk of pulmonary embolism.
The lack of concordance between intraoperative TEE and Holter ECG detection of ischemia is consistent with the findings of previous studies comparing the sensitivity of echocardiography with ECG in patients undergoing angioplasty. [28,29] Although regional contractile abnormalities are more sensitive than ST-segment changes for the early detection of myocardial ischemia, the use of TEE is limited intraoperatively mainly to patients who receive general anesthesia. Thus, the potentially high-risk periods of anesthetic induction and emergence cannot be monitored by TEE. However, Holter ECG is relatively insensitive in detecting ischemic changes in the posterior and lateral quadrants of the left ventricle,  where 56% of the echocardiographic ischemic changes occurred in our patients. Multiplane TEE and multilead ECG monitoring ultimately may enhance the sensitivity of each monitoring modality. Whether the true incidence of perioperative ischemia equals the sum of ECG and TEE ischemia remains to be determined. .
We did not measure cardiac output and pulmonary artery pressure. However, the effects of dexmedetomidine on these parameters and on the baroreflex should be studied to further validate the safety of dexmedetomidine in patients with vascular disease.
Finally, we have studied only vascular surgery patients and only males. Further studies should include other surgical patient populations with or at risk for CAD as well as females to verify that these results are applicable to the high-risk cardiac population in general.
The hemodynamic effects of dexmedetomidine in vascular surgery patients appear to be similar to those in healthy volunteers. A dose of 0.45 ng/ml appeared to be most effective in blunting hemodynamic responses to perioperative stress but required greater intraoperative pharmacologic intervention to support blood pressure and heart rate. Further studies in a larger number of high-risk patients will be conducted to verify these preliminary results.
The authors thank the Vascular Surgery Department at the VA Medical Center, for their cooperation; Long Ngo, M.S., and Ida M. Tateo, M.S., for the statistical analysis; and Winifred von Ehrenburg, Ph.D., for her editorial assistance.