Urine and Plasma Catecholamine and Cortisol Concentrations after Myocardial Revascularization : Modulation by Continuous Sedation

This study included 121 patients from the Kaiser Permanente Medical Center, San Francisco, California, enrolled as part of a multicenter, prospective, double-blind, open-label randomized study involving 351 patients that compared the safety and efficacy of propofol and midazolam sedation during the first 12 h after CABG surgery. [18] The protocol was approved by the Institutional Review Board of Kaiser Permanente Medical Center, San Francisco. Male or female patients older than 35 yr with at least 50% stenosis of the left main coronary artery or at least 70% stenosis of one or more of the major coronary arteries were eligible for inclusion. Exclusion criteria included (1) an ejection fraction of less than 25% before surgery, (2) requirement of intraaortic balloon pump therapy, (3) an evolving myocardial infarction (MI), (4) left bundle branch block or permanent ventricular pacemaker, and (5) hemodynamic instability at admission to the ICU. Hemodynamic stability was defined as a systolic blood pressure of more than 90 mmHg for 5 min without change in cardiovascular medications and without mechanical ventricular support.

Patients were permitted to receive 1–4 mg lorazepam orally on the eve of surgery and were premedicated before operation with morphine sulfate (as much as 0.2 mg/kg) and midazolam (as much as 0.1 mg/kg) given intramuscularly or intravenously.

Anesthesia was induced with sufentanil given by a computer-controlled infusion pump set initially for a target effect site concentration of 3 ng/ml. Sufentanil was maintained at the same target effect site until CPB, at which time the infusion was decreased to a target effect site of 1 ng/ml. At the termination of CPB, the infusion was decreased to a target effect site of 0.6 ng/ml. Midazolam was infused continuously at 0.5 micro gram [center dot] kg sup -1 [center dot] min sup -1 throughout the surgical procedure, with increases in increments of 0.1 micro gram [center dot] kg sup -1 [center dot] min sup -1 as needed. If an increase in anesthesia was required, the sufentanil target effect site was changed in increments of 0.5–1.0 ng/ml. Vecuronium was administered at induction to facilitate intubation and as required during operation to maintain muscle relaxation. Recovery of neuromuscular function was documented by the presence of four of four twitches on train-of-four neuromuscular testing before transport to the ICU.

Specific guidelines to control hemodynamics were prespecified. Before CPB, systolic blood pressure and heart rate were controlled to within 20% of preoperative values (determined by the mean of three sets of preoperative vital signs). Hypertension was treated with increases in anesthesia followed, if necessary, by an infusion of sodium nitroprusside. Hypotension was treated with a decrease in anesthesia, fluid administration, infusion of phenylephrine, or all three. Tachycardia was treated with increases in anesthesia, or an infusion of esmolol. By protocol, nitroglycerin use was reserved only for treatment of clinically detected electrocardiographic or echocardiographic evidence of myocardial ischemia (S-T segment changes or regional wall motion abnormalities) or elevations in pulmonary arterial pressures (diastolic pressure greater or equal to 20 mmHg or capillary occlusion pressure greater or equal to 18 mmHg). Nitroglycerin could not be administered prophylactically without evidence of ischemia. During CPB, the mean arterial pressure was maintained between 40 and 80 mmHg by changing anesthetic levels or administering phenylephrine or sodium nitroprusside.

After CPB, systolic blood pressure was to be maintained between 90 and 140 mmHg and heart rate between 60 and 100 beats per minute. Hemodynamic aberrations were treated similarly to those of the prebypass period. Bradycardia after CPB was treated using a temporary pacemaker.

Cardiopulmonary bypass was performed using a membrane oxygenator, hemodilution, and moderate hypothermia (26–28 degrees Celsius). Graft and anastomosis quality were assessed and categorized (excellent, good, fair, poor) by the surgeons, who were blinded to the study group and the results of the hormone analyses at the time of this assessment.

After surgery, patients were transported to the ICU, and when hemodynamically stable (as defined previously) they were randomized to receive continuous propofol infusion (CP group) or “standard” sedation (SC group) with intermittent boluses of midazolam and morphine. For those patients randomized to the CP group, propofol was administered by computer-assisted continuous infusion (using STANPUMP, Harvard Apparatus, So. Natick, MA), initiated at a target plasma concentration of 0.25 mg/ml and titrated to achieve and maintain a sedation score of 5 on a modified Ramsay sedation score (Table 1). Patients in whom an adequate sedation score could not be achieved received 1–4 mg morphine given intravenously every 15 min. Patients randomized to the SC group were given 1–4 mg midazolam administered intravenously every 15 min as needed for agitation. Patients inadequately sedated received 1–4 mg morphine intravenously every 15 min as needed for pain. In both sedation groups, the lungs were mechanically ventilated for a minimum of 12 h after aortic unclamping, after which patients were separated from mechanical ventilation and the trachea was extubated as clinically indicated.

All medications were recorded by dose, route, and time of administration, and classified as follows: analgesics-aspirin, acetaminophen, ketorolac; opioids-morphine, meperidine; antihypertensives: beta adrenergic blockers-labetalol, esmolol, propranolol, metoprolol; calcium channel blockers-diltiazem, nifedipine, verapamil; vasodilators-sodium nitroprusside, trimethopham, phentolamine; anti-ischemics-nitroglycerin (paste or infusion); cardiac stimulants: inotropic agents-dopamine, dobutamine, epinephrine, digoxin, calcium, isoproterenol; vasopressors-phenylephrine, norepinephrine, amrinone, milrinone; antidysrhythmics-lidocaine, procainamide. The incidence of use of each class of medication was evaluated for each study group. Medication interventions were defined as the number of bolus doses of medication or the number of changes in infusion doses during the sedation period. The number of medication interventions per patient between groups was analyzed.

After operation, heart rate and blood pressure were electronically recorded at 1-min intervals during the study period. Patients for whom hemodynamic data were unavailable for more than 25% of the study period (ICU entry to 12 h after aortic unclamping) were excluded from hemodynamic analysis. Data were analyzed for the presence or absence of hemodynamic episodes during the period of ICU sedation. A hemodynamic episode was defined as an occurrence of tachycardia, bradycardia, hypertension, or hypotension (as defined for the post-CPB period) that lasted for at least 4 min within any 5-min period. If the same hemodynamic alteration recurred within 5 min of a previous episode, it was considered to be a continuation of the same episode. Treatment of a hemodynamic episode was not considered in episode identification.

The presence of a hemodynamic episode was determined initially using validated software algorithms (developed by the Ischemia Research and Education Foundation), with identified episodes validated independently by two physicians blinded to treatment group assignment. If they disagreed, the episode was analyzed independently by a third physician and the discrepancy was resolved by consensus. Hemodynamic episodes were characterized by incidence, number of episodes per patient, average area under the curve per hour monitored-in patients with episodes, and average episode minutes per hour monitored-in patients with episodes. [18]

Given the inherent variability in plasma catecholamine concentrations, as well as their evanescent nature, both plasma concentrations and urine concentrations were analyzed. [19]

Urine for measurements of dopamine, epinephrine, norepinephrine, and cortisol was collected during the following time periods:(1) 12 h before surgery;(2) during the intraoperative period;(3) from ICU arrival for six subsequent hours;(4) from 6 to 12 h of their ICU stay;(5) from 12 h in the ICU until tracheal extubation; and (6) from tracheal extubation to 12 additional hours. Urine was stored on ice during the entire collection period and then separated into aliquots. Catecholamines and free cortisol were measured by high-performance liquid chromatography at Nichols Institute Reference Laboratories (San Juan Capistrano, CA).

Blood for measurements of plasma dopamine, epinephrine, norepinephrine, and cortisol was collected from the arterial catheter at the following time points:(1) in the operating room before induction;(2) ICU arrival; and (3) 8, 16, and 24 h after ICU arrival. The plasma samples were collected and stored on ice, and plasma catecholamines were measured by high-performance liquid chromatography at Nichols Institute Reference Laboratories. Plasma or serum cortisol was measured by TDX (Abbott Laboratories, Chicago, IL) fluorescent immunoassay at the Regional Laboratories for Kaiser Permanente Medical Center. Other measurements included creatine kinase with MB fraction on ICU arrival and every 8 h for 48 h; and electrocardiographs before operation, on ICU arrival, and on postoperative days 1 and 2.

Adverse cardiac outcomes were not end points for this trial because the study was not statistically powered to address such. However, outcomes were observed and included MI (Q wave and non-Q wave), congestive heart failure, left ventricular dysfunction, and cardiac death. [18] Myocardial infarction was defined by the protocol as (1) electrocardiographic evidence of infarction based on the Minnesota codes 1.1–2.12 (Q wave MI)[20];(2) creatine kinase with MB fraction isoenzyme concentrations greater than 70 ng/ml (non-Q wave MI); or (3) diagnosis of MI on autopsy without having met the previous criteria. All electrocardiographs were coded for new perioperative Q waves indicative of MI by two blinded electrocardiographers. [21] If coded positively, the Q wave findings were validated by a panel of three electrocardiographers blinded to treatment group.

Investigators, blinded to the study group, determined the distribution of urine and plasma hormones, and using prespecified statistical and clinical criteria excluded data considered to be outliers. At each time measurement point, for each hormone measured, the distribution of data points over the entire population was examined, and those values greater than three standard deviations from the mean were considered outliers. At any time point, the number of outliers ranged from 1–5% of the data points. Validation of this prespecified, blinded analysis, using data including rather than excluding outliers, revealed no significant difference in the results of this trial. The serial hormone measurements were treated as repeated-measures data in which the within-subject correlation is included in the estimation of the fixed effect. For this linear mixed-effects model, the fixed treatment effect (SC control versus CP propofol) is estimated. The random effect is the random variability that derives from within patients as well as between patients. The estimation of the variance components was obtained by using the method of restricted maximum likelihood. The structure of the covariance matrix, which contains both the within- and between-subject variability, is assumed to be compound symmetry, and the fixed treatment effect is estimated for each time period. The estimated mean difference between the two treatments (CP-SC) at each time is computed using the method of linear model contrast. The point estimate of the mean difference, its standard error, and the probability value are reported for each time point.

Unpaired t test, Fisher's exact test, and the chi-squared test were used to analyze the demographics and intraoperative and hemodynamics differences. All tests are two tailed, with the type I error set at 5%. The linear mixed-effects model was constructed and estimated using SAS's procedure Proc Mixed (SAS Institute, Cary, NC). [22] All data management and statistical analyses were performed using SAS Institute software. [23] The software Stata (Computing Resource Center, Santa Monica, CA) was used to examine the individual profile's repeated-measures data graphically. [24]

A total of 121 patients were enrolled in the study and randomized to treatment with either continuous propofol infusion (CP group) or intermittent boluses of midazolam or morphine (SC group). Patient characteristics of the two treatment groups are similar and are shown in Table 2and Table 3.

A total of 105 patients had evaluable hemodynamic data for the entire 12-h period. During the 12-h treatment period, the CP group (compared with the SC group) had a lower incidence of tachycardia (60% vs. 79%, respectively; P = 0.038) and hypertension (33% vs. 58%, respectively; P = 0.012), but a greater incidence of hypotension (81% vs. 49%, respectively; P = 0.001).

Results for each of four plasma hormones are listed in Table 4a-d, and for each of four urine hormones in Table 5a-d. Regarding plasma samples, the estimated mean difference between study groups (CP group-SC group) was determined for each of the four plasma hormones at the five sampling times described previously (see Table 4a-d). There was no significant difference between the two groups before operation for all of the hormones assayed. Nor was there any difference in the plasma hormone concentrations between the study groups after bypass, on arrival into the ICU, or before initiation of sedation for the four plasma hormones assayed. At 8 h in the ICU, during which time all patients were being sedated according to the protocol, the CP group had plasma cortisol concentrations of 8.7 micro gram/dl, which was 31% less than the SC group (P = 0.0004); plasma epinephrine was 69.2 pg/ml, which was 53% less than the SC group (P = 0.009); and a trend (P = 0.062) for plasma norepinephrine (at 118 pg/ml) to be less (by 28%) than the SC group. Plasma dopamine concentrations were not significantly different, although the values as shown in Table 7 tended to cluster at the low end of the range for both groups. At 24 h (that is, 12 h after drug discontinuation), when nearly all (87%) patients had sedation discontinued and their trachea's extubated, there was no significant difference between the CP and SC groups.

Regarding urine samples, there was no significant difference between the CP and SC groups before operation for urine cortisol, epinephrine, or norepinephrine levels (Tables 5a-d). Urine dopamine concentration was 20.3 micro gram, which was about 18% less in the CP group (P = 0.051). The estimated mean difference (CP - SC) was determined for each of the four urine hormones, during the six sampling periods, as described above. During operation, there was no significant difference between the groups for any of the four hormones. During the first 6 h in the ICU, during which time all patients were being sedated with the assigned regimen, the CP group demonstrated lower cortisol by 59% or 126.9 micro gram (P = 0.007), dopamine by 38% or 32.1 micro gram (P = 0.002), epinephrine by 45% or 3.3 micro gram (P = 0.009), and norepinephrine by 48% or 11.3 micro gram (P = 0.0004). During the second 6 h in the ICU, also during which time all patients were being sedated with the assigned regimen, the CP group demonstrated lower cortisol levels by 67% or 269.8 micro gram (P = 0.0001), dopamine by 32% or 28 micro gram (P = 0.009), epinephrine by 49% or 4 micro gram (P = 0.002), and norepinephrine by 45% or 11.1 micro gram (P = 0.005). During the period between 12 h after arrival into the ICU and extubation, urine dopamine (P = 0.007), epinephrine (P = 0.088), and norepinephrine (P = 0.001) were lower in the CP group than in the SC group. During the 12 h after extubation, only urine dopamine remained lower for the CP group by 20% or 28.5 micro gram (P = 0.008). Otherwise there was no significant difference between study groups.

As shown in Table 6, patients in the CP group received opioids less frequently (P = 0.001). There was no difference between the groups in their use of antihypertensive, anti-ischemic, or antiarrhythmic medications. The CP group received more cardiac stimulant medications (P = 0.016), including both inotropic drugs (P = 0.016) and vasopressors (P = 0.025).

Adverse outcomes included three myocardial infarctions in the SC group (two Q wave, one creatine kinase fraction) and two in the CP group (both Q wave) and one cardiac death. No episodes of unstable angina of heart failure were reported.

The results of our study suggest that in patients undergoing CABG surgery, continuous postoperative sedation, using propofol targeted to a controlled concentration, can attenuate postbypass increases in catecholamine and cortisol serum and urine concentrations for 12 h. During the first 6 to 8 h in the ICU, patients in the CP group had significantly lower plasma cortisol (31%) and epinephrine (53%) concentrations, and lower urine cortisol (59%), dopamine (38%), epinephrine (45%), and norepinephrine (48%) concentrations. These reductions generally persisted throughout the 12-h infusion period and then rapidly returned to the “control” SC group values when the propofol infusion was discontinued. The CP group experienced a 24% reduction in the incidence of postoperative tachycardia, a 43% reduction in hypertension, a 64% increase in hypotension, and a 49% reduction in opioid use, compared with the SC group, which were all consistent with significant suppression of sympathetic activity.

Mitigation of the postoperative endocrine response may be of clinical benefit in patients undergoing CABG surgery. Release of catecholamines, specifically norepinephrine, is associated with increases in systemic vascular resistance and arterial pressure, [7] and suppression of the sympathetic nervous system response could have beneficial effects on fibrinolysis, sensitivity of platelets to epinephrine, [25] left ventricular function, [26] and coronary artery vasoconstriction. [8] However, outcome data from cardiac surgical patients, relating adverse events to modulation of the postoperative endocrine response, are limited. Anesthetics, including propofol, have been shown both to suppress the intraoperative stress response [8,17,27–29] and intraoperative myocardial ischemia. [8,10,30–33] Previously we found that using intensive postbypass analgesia is associated with decreased severity of myocardial ischemia. [8] Similarly, treatment with clonidine, a preferential alpha²-adrenergic agonist, also can reduce opioid and anesthetic requirements, reduce plasma epinephrine and norepinephrine, and improve hemodynamics. [34,35] Regarding major, abdominal non-cardiac surgery, Hosoda et al. [36] has suggested that attenuation of the stress response is associated with improved tissue oxygenation in the immediate postanesthetic period after major abdominal surgery. Thus, for both types of surgery, development of new therapeutic approaches aimed at reducing the postoperative stress response may allow greater hemodynamic stability and consequent clinical benefit. This concept largely underlies the recent strategy for clinical trials to determine the efficacy of perioperative use of beta blockers and selective alpha²agonists, which may modulate the sympathetic response and afford hemodynamic stability in high-risk patients undergoing general anesthesia. [2,37,38]

In our current study, it is unclear whether the differences in the postoperative catecholamine and cortisol concentrations in the CP group were due to the significantly different doses of sedatives or analgesics received, the mode of administration, or properties specific to propofol per se. Previous CABG studies have shown that using large doses of opioids (fentanyl) does not adequately suppress the endocrine response. [5,7] Similarly, Moller et al. [39] reported that morphine reduced pain after cholecystectomy but had minimal effect on plasma cortisol concentrations and no effect on plasma catecholamine concentrations, suggesting that administration of opioids without sedative-hypnotics may not have a substantial effect on the endocrine response, despite improved pain relief. On the other hand, as mentioned before, we studied the postoperative course of 100 patients undergoing elective CABG randomized for ICU sedation with either intensive analgesia using sufentanil infusion or standard analgesia (intermittent morphine and midazolam), similar to that used in the control group of the present trial. [8] Although the incidence of ischemia was similar before bypass and extubation, the severity of ischemia (measured by the area under the curve) was significantly decreased in the sufentanil-sedated group, suggesting that intensive analgesia could affect ischemia, which may be a surrogate of adverse outcome. Although further study clearly is needed, it is likely that the endocrine response to surgery depends both on the type and depth of sedation and analgesia. This hypothesis, linking sympathetic activation and degree of sedation, is supported by the results of the present trial because patients given intensive, continuous sedation had significantly lower catecholamine and cortisol concentrations and received lesser doses of opioids than did control patients.

In patients undergoing CABG surgery, plasma epinephrine and norepinephrine serum concentrations have been shown to peak either at the time of aortic cross clamping or after operation on arrival in the ICU. [27,40,41] These results differed from our findings that showed that epinephrine and norepinephrine concentrations decreased transiently on arrival in the ICU compared with baseline and subsequently increased for the rest of the study period. These results may affect the current support for rapid emergence from anesthesia ("fast tracking") and warrant further consideration, especially in high-risk subgroups of patients. Clearly, for a substantial proportion of “low-risk” patients, rapid emergence and tracheal extubation may be safe and therefore suitable. Conversely, a group of “high-risk” patients probably exists in whom rapid emergence and tracheal extubation may further exacerbate this emergence stress phenomenon, leading to ischemia and perhaps irreversible adverse outcome. To date, few trials have appropriately examined this question and established the safety of the rapid emergence/tracheal extubation technique for specific high-risk subgroups of patients. The mandate, then, is to delineate such low-risk from high-risk groups, using appropriately designed trials, before decisions are made regarding the institution of new critical pathways.

The present study shows that modulation of the postoperative endocrine response is possible using intensive, but rapidly reversible, sedation with propofol. However, our study has several limitations. First, sedation was administered differently in the two groups. Because sedation with propofol was an unfamiliar sedation regimen for our ICU, it was necessary to create clear guidelines for its administration, including the use of Ramsay scores. The SC group regimen represented standard care in our institution (Kaiser Permanente), and Ramsay scores were not used for this group because we did not want to modify the behavior of the nursing staff administering sedation to this “control” group. Therefore the study cannot correlate level of sedation with changes in catecholamine and cortisol concentration, nor with changes in hemodynamics. This study design limits our ability to determine whether the hormonal attenuation found for the CP group is due to intrinsic beneficial properties of propofol, production of a controlled state of sedation, the technique of continuous infusion, or some combination of these. It is clear however, that because the use of inotropic drugs and vasopressors were more common in the CP group, the lower catecholamine concentrations found in this group could not be attributed to iatrogenic catecholamine administration.

Second, although we have shown the ability to mitigate postoperative catecholamine and cortisol responses associated with CABG surgery, conclusive data linking these reductions to cardiac outcomes (myocardial infarction, heart failure, death) are lacking. Our study was not statistically powered to discern such outcomes and a substantively larger outcome study would be necessary.

Third, we chose only four plasma and five urine sampling time points after operation, and although we found a significant difference between groups, and a good correlation between the two types of samples, too few samples were obtained to examine the relation between catecholamine and cortisol levels and hemodynamic changes in individual patients. Future studies will be needed to determine whether modulation of the endocrine stress response with propofol sedation is responsible for the hemodyamic effects we observed in propofol-treated patients.

Fourth, we studied patients with, at worse, mild to moderate left ventricular dysfunction (> 25%), and thus the results may not apply for patients with markedly reduced function (left ventricular dysfunction < 25%).

The results of our study show that the catecholamine and cortisol responses to CABG surgery are substantial and persistent, and that these responses can be modulated by continuous, intensive sedation using propofol. A relation to postoperative hemodynamics is suggested, with a lower incidence of tachycardia and hypertension and an increased incidence of hypotension in the propofol-treated group.

The authors thank Diane Beatty (of the Ischemia Research and Education Foundation) and Winifred Von Ehrenberg (an Ischemia Research and Education Foundation consultant) for editorial assistance.

The Multicenter Study of Perioperative Ischemia (McSPI) Research Group is a consortium of investigators from approximately 150 medical centers worldwide, and focuses on the problems of perioperative myocardial infarction, stroke, renal dysfunction, and other organ dysfunction, and the health economic implications of such diseases. The Ischemia Research and Education Foundation is a nonprofit foundation supporting multicenter international research in these areas and is closely affiliated with the McSPI investigators and their institutions. The coordinating analysis group for this trial consisted of the following persons: Study director: Dennis T. Mangano, Ph.D., M.D.; general analysis: Jerill Plunkett, M.D., Ahvie Herskowitz, M.D.; statistical analysis: Long Ngo, M.S., Elizabeth Li, M.S., Rong Ji, M.S., Catherine Ley, Ph.D., Alex Gaber, M.S.; electrocardiographic analysis: Uday Jain, Ph.D, M.D., Tatiana Titov, Ph.D, M.D., Vladimir Titov, Ph.D., M.D., Adam Zhang, M.D.; editorial administrative assistants: Diane Beatty, Brenda Xavier, Andre Marschalko, and Winifred von Ehrenburg.

The following institutions and investigators participated in the data collection phase of the trial: Kaiser Permanente Medical Center, San Francisco, California-Wayne Bellows, M.D., Nito Pineda, M.D., John Reeves, M.D., Gary Roach, M.D.

Five other institutions participated in the general multicenter trial: Cleveland Clinic Foundation, Cleveland, Ohio-Colleen Koch, M.D.; Emory University Hospital, Atlanta, Georgia-James G. Ramsay, M.D., Christina T. Mora, M.D.; University of Michigan, Ann Arbor, Michigan-Joyce Wahr, M.D.; Veterans Affairs Medical Center, Milwaukee, Wisconsin-Anil Aggarwal, M.D., David C. Warltier, M.D., Veterans Affairs Medical Center, San Francisco, California-Dennis T. Mangano, Ph.D., M.D.