Hemodynamic Effects, Myocardial Ischemia, and Timing of Tracheal Extubation with Propofol-Based Anesthesia for Cardiac Surgery

Abstract

Recent interest in earlier tracheal extubation after coronary artery bypass graft (CABG) surgery has focused attention on the potential benefits of a propofol-based technique. We randomized 124 patients (34 with poor ventricular function) undergoing CABG surgery to receive either a propofol-based (5 mg [center dot] kg^-1 [centered dot] h^-1 prior to sternotomy, 3 mg [center dot] kg^-1 [center dot] h (^-1) thereafter; n = 58) or enflurane-based (0.2%-1.0%, n = 66) anesthetic. Induction of anesthesia consisted of fentanyl 15 micro g/kg and midazolam 0.05 mg/kg intravenously in both groups. The enflurane group received an additional bolus of fentanyl 5 micro g/kg prior to sternotomy and fentanyl 10 micro g/kg with midazolam 0.1 mg/kg at commencement of cardiopulmonary bypass (CPB). Patients receiving propofol were extubated earlier (median 9.1 h versus 12.3 h, P = 0.006), although there was no difference in time to intensive care unit (ICU) discharge (both 22 h, P = 0.54). Both groups had similar hemodynamic changes throughout (all P > 0.10), as well as metaraminol (P = 0.49) and inotrope requirements (P > 0.10), intraoperative myocardial ischemia (P = 0.12) and perioperative myocardial infarction (P = 0.50). The results of this trial suggest that a propofol-based anesthetic, when compared to an enflurane-based anesthetic requiring additional dosing of fentanyl and midazolam for CPB, can lead to a significant reduction in time to extubation after CABG surgery, without adverse hemodynamic effects, increased risk of myocardial ischemia or infarction.

(Anesth Analg 1997;84:12-9)

Fundamental principles of anesthesia for coronary artery bypass graft (CABG) surgery include maintaining hemodynamic stability and minimizing myocardial ischemia ^[1-3]. The most widely used anesthetic technique incorporates large doses of short-acting opioids, which minimize the hemodynamic changes after induction, endotracheal intubation, and sternotomy ^[4,5]. This technique mandates and obligatory period of postoperative ventilation in an intensive care unit (ICU) environment. The requirement for ICU care may be reduced with an inhalation-based anesthetic technique ^[6], but patients with impaired left ventricular (LV) function may not tolerate inhaled anesthetic-induced reduction in myocardial contractility.

Recent publications have described the use of propofol for anesthesia during CABG surgery ^[7-16]. The major benefit of propofol, when compared to a large-dose opioid technique, is shorter recovery time leading to a significant reduction in time to tracheal extubation ^[13-17], and the potential to reduce nurse staffing levels and length of stay in the ICU ^[13,14,17]. This can reduce costs associated with CABG surgery ^[6,18]. Unfortunately, there is scant information in patients with impaired LV function ^[14,15], particularly whether propofol results in more hypotension and myocardial ischemia in this higher risk group. There is also no evidence demonstrating superiority of a propofol-based technique when compared to an inhalation-based technique in such patients. We therefore conducted this prospective, randomized trial.

Methods

After Ethics Committee approval, we approached all elective CABG surgical patients and obtained written, informed consent. Patients were excluded if concurrent valvular surgery was planned or if they were considered "very high risk," defined as a clinical severity score greater than 9 ^[19], since such patients were more likely to require continued sedation and length of stay in the ICU. Patients were also excluded if they had a preexisting left bundle-branch block or pacemaker in situ preventing ST-segment diagnosis of perioperative myocardial ischemia.

Two different anesthetic techniques were compared, using a prospective, randomized design. Blinding was not practically possible, so the end points were predetermined and objective. It should be noted that at our institution, ICU management occurs independent of the anesthesia department and decisions regarding ventilator weaning and extubation are made separately and without consultation. To reduce bias, all details of the anesthetic were retained until after tracheal extubation, so that ICU staff were blinded to group identity.

Anesthetic Technique

Medications, including nitrates, beta-adrenergic blockers, and calcium-channel blockers, were continued until the time of operation. Patients received a standard premedication of oral temazepam 10-20 mg, intramuscular morphine 5-10 mg or papaveratum 10-20 mg, hyoscine 0.2-0.4 mg, and oxygen delivered via face mask at 5 L/min.

Patient monitoring consisted of a five-lead electrocardiogram (ECG), pulse oximetry, capnography, invasive arterial pressure (via radial or brachial artery), and pulmonary artery pressure (via right internal jugular vein). The ECG and pulse oximeter were placed on arrival in the operating room. Correct ST-segment monitoring was confirmed, with identification of the isoelectric line and J point. The systemic and pulmonary arterial catheters were inserted with local anesthesia and light intravenous (IV) sedation (as required) prior to induction of anesthesia.

Both groups had an identical induction technique, consisting of IV midazolam 0.05 mg/kg and fentanyl 15 micro g/kg. Endotracheal intubation and muscle relaxation followed administration of IV pancuronium 0.12 mg/kg. If further doses of muscle relaxants were required, then IV vecuronium 2-4 mg was used. Maintenance of anesthesia differed.

The propofol group had a propofol infusion initiated after induction at a rate of 5 mg [center dot] kg^-1 [center dot] h^-1 and then reduced to 3 mg [center dot] kg^-1 [center dot] h^-1 after sternotomy. Propofol was delivered via the opaque (VIP) lumen of the pulmonary artery catheter, and was disconnected before transfer to the ICU, in order to maintain blinding in the ICU. Incremental boluses of IV propofol 20 mg were administered if the patient's mean systemic blood pressure (mBP) exceeded 85 mm Hg.

The enflurane-FM group received enflurane at a vaporizor concentration of 0.2%-1.0%, with a further bolus of IV fentanyl (5 micro g/kg) administered 1-2 min before sternotomy, and another bolus of IV fentanyl (10 micro g/kg) with midazolam (0.1 mg/kg) at the commencement of cardiopulmonary bypass (CPB). The enflurane concentration was increased if the patient's mBP exceeded 85 mm Hg. Enflurane was not administered during CPB.

Patients were ventilated with a tidal volume of 10 mL/kg, adjusting respiratory rate to an end-tidal carbon dioxide of 32-34 mm Hg; positive endexpiratory pressure was not used. CPB was standardized, with a crystalloid prime, membrane oxygenator, moderate hypothermia (27-32 degrees C), and alpha-stat pH management. Patients remained intubated and mechanically ventilated for transfer to the ICU after completion of surgery (without reversal of neuromuscular blockade). Patients were weaned from mechanical ventilation as soon as they responded to verbal stimuli, rewarming was complete, hemodynamic stability had been established, and blood loss was satisfactory (<100 mL/h). Analgesia consisted of a nurse-controlled morphine sulfate infusion ^[20]. Extubation occurred when the patient was awake and cooperative, with a respiratory rate between 10 and 20 breaths/min and satisfactory arterial blood gas tensions. Standard ICU discharge criteria were used, but eventual discharge often depended on availability of a general ward bed.

Inotrope and Vasoconstrictor Administration

In this study hypotension was defined as a mBP less than 65 mm Hg or a decrease in mBP greater than 20% of the preinduction control value. Administration of inotrope and vasopressor drugs were dictated by the following protocol:

1. For hypotension, IV fluids were administered until pulmonary capillary wedge pressure (PCWP) > 12 mm Hg.

2. If hypotension continues, but cardiac index (CI) >2.4 L [centered dot] min^-1 [center dot] m^-2, IV metaraminol 250 micro g was administered.

3. If CI <2.4 L [center dot] min^-1 [center dot] m^-2, fluids were administered to PCWP >12 mm Hg.

4. If hypotensive, CI <2.4 L [center dot] min^-1 [center dot] m^-2, and PCWP >12 mm Hg, epinephrine was administered as an IV infusion commencing at 20 ng [center dot] kg^-1 [center dot] min^-1 and increasing in 10 ng [center dot] kg¹ [center dot] min¹ steps until these hemodynamic goals have been achieved.

5. If perfusion pressure <40 mm Hg occurred during CPB, IV metaraminol 1 mg was administered incrementally.

Measurements

Hemodynamics (mBP, heart rate, central venous pressure, PCWP, CI, systemic vascular resistance index, pulmonary vascular resistance index, and LV stroke work index were determined within 2 min of the following time periods: 1) preinduction; 2) 5 min postinduction; 3) 5 min poststernotomy; 4) 10 post-CPB; 5) poststernal closure; 6) ICU admission; and 7) 4 h post-ICU admission.

Cardiac output was measured using the thermodilution method (10 mL of room-temperature saline at end-exhalation, averaged over three measurements within 10% of each other). ST-segment analysis and derived indices were calculated using the Hewlett-Packard Component Monitoring System (HP M1176A; Hewlett-Packard GmBH, Hamburg, Germany). Threelead (II, aVL and V5) ST-segment analysis was used for the detection of intraoperative myocardial ischemia and was confirmed directly with the raw ECG waveform. Ischemia was defined as ST-segment depression >1 mm or elevation >2 mm at 60 ms after the J point, persisting for at least 2 min. Blood was taken at 36 h for CK-MB isoenzyme concentration and a 12-lead ECG performed on admission to ICU and daily for 3 days after surgery. Acute myocardial infarction was diagnosed if new Q waves appeared in at least two ECG leads, as detected by an independent and blinded cardiologist, and creatine kinase-MB fraction >5%.

Total amounts of intraoperative metaraminol and epinephrine administered, time to extubation (after admission to ICU), and ICU discharge were recorded. All patients were interviewed by a research nurse within 3 days of surgery and queried about any recall of intraoperative events.

Sample Size Calculation and Statistics

A preliminary estimate of sample size was based on an expected 33% improvement for the propofol group over the enflurane-FM group, with an average time to extubation of 12 h and a SD of 5 h. With a Type I error of 0.01 and a Type II error of 0.05, the required number was calculated at 125 patients (Clinical Trials Design Program V1.0; Biosoft, Cambridge, UK). Because recovery time and complications after CABG surgery are related to LV function, we randomized patients after stratification according to the surgeon's angiographic assessment of contractility (1 = normal, 2 = mild impairment, 3 = moderate impairment, 4 = severe impairment, 5 = gross impairment), to maximize equality of both groups. Randomization was determined by a Table ofrandom numbers.

Continuous data were first assessed for normality using the Kolmogorov-Smirnov test, then analyzed using unpaired, two-tailed Student's t-test or repeated measures analysis of variance, as appropriate. Non-normally distributed data were analyzed using the Mann-Whitney U-test and Spearman rank correlation (r). Proportions were analyzed by chi squared with Yates' correction, or Fisher's exact test, as appropriate. Multivariate regression was used to identify patient, surgical, and anesthetic factors associated with earlier extubation. Normally distributed data is presented as mean (SD); skewed data is presented as median (interquartile range). All statistical analyses were performed using SPSS/PC + V4.0 (SPSS Inc., Chicago, IL). No adjustment was made for multiple testing, although a P value of less than 0.01 was considered significant.

Results

Of 138 patients approached, a total of 129 patients (103 male, 26 female) were enrolled in this study, with subsequent exclusion of 5 patients (1 patient undergoing reoperation suffered a cardiac arrest after sternotomy in which the right ventricle was lacerated and a previous graft divided, 1 had left bundle-branch block, 2 had concurrent surgery [unplanned LV aneurysmectomy, valve replacement], and 1 patient [enflurane-FM group] with severe LV impairment was unable to be weaned from CPB). This left 124 patients with an average age of 64 (SD 10) yr, clinical severity score of 2.1 (SD 2), aortic cross-clamp time of 73 (SD 26) min and duration of surgery of 4.3 (SD 1.0) h. Sixty-five patients had either moderate or severe impairment of LV contractility on angiographic assessment, or a preinduction cardiac index less than 2.2 L [center dot] min^-1 [center dot] m^-2. Two patients (1.6%) experienced perioperative myocardial infarction. No patient reported intraoperative awareness. There were 66 patients in the enflurane-FM group and 58 patients in the propofol group (Table 1), all of whom survived to hospital discharge. The groups were well matched for several perioperative risk factors, including age, clinical severity score ^[19], myocardial contractility, preinduction hemodynamics, and duration of surgery. Six patients in the enflurane-FM group, and seven patients in the propofol group had a clinical severity score of 6 to 9. Despite stratification and randomization, there were more women in the enflurane-FM group and those in the propofol group had a longer aortic cross-clamp time.

Table 1: Patient and Perioperative Characteristics

Both groups had similar hemodynamic changes throughout the perioperative period, with no significant differences between groups at any time period (Table 2). There was no difference between the groups for vasoconstrictor requirements (P = 0.49) or inotropic support (to wean from CPB, P = 0.096; on admission to ICU, P = 0.23) (Table 3). Both groups had a similar rate of intraoperative myocardial ischemia and postoperative myocardial infarction (Table 3).

Table 2: Perioperative Hemodynamic Changes for Enflurane-FM Group (n = 66) and Propofol Group (n = 58)

Table 3: Outcome After CABG Surgery

There was a significant difference in the time to extubation, with patients in the propofol group having a median time of 9.1 h, compared with those in the enflurane-FM group of 12.3 h (P = 0.006) (Table 3 and Figure 1). There was no difference in the time to ICU discharge (P = 0.54) (Table 3).

Figure 1: Extubation time curve in the intensive care unit (ICU). The propofol group had a significant reduction in time to extubation, particularly in the first 18 h after admission to ICU. Lee-Desu D statistic, P = 0.006. FM = fentanyl-midazolam.

Women tended to be older (67 yr versus 63 yr, P = 0.08) and rated as higher risk with the clinical severity score (2.5 versus 2.0, P = 0.03), but did not exhibit a significant difference in median time to extubation in ICU (14 h versus 11 h, P = 0.14). Subgroup analyses of the patients with severe impairment of LV contractility on angiographic assessment (which equates to an ejection fraction of less than 30%), or a preinduction cardiac index of less than 2.2 L [center dot] min^-1 [centered dot] m^-2, did not demonstrate any significant differences between anesthetic groups for hemodynamic changes (Table 4), nor time to ICU discharge. There was, however, a shorter median extubation time (propofol group 10.6 h, enflurane-FM group 13.2 h, P = 0.049). There was an association between preoperative risk (as measured by the clinical severity score) and time to extubation (r = 0.25, P = 0.006), and also time to discharge from ICU (r = 0.21, P = 0.02).

Table 4: Perioperative Hemodynamic Changes in Patients with Poor Left Ventricular Function^a

After adjustment for patient age, gender, and LV function, multivariate regression analysis identified useful predictors of time to extubation as post-CPB epinephrine infusion rate (P < 0.001), duration of surgery (P = 0.02), type of anesthetic (P = 0.02), and preoperative clinical severity score (P = 0.04). Predictors of discharge time from ICU were epinephrine infusion rate on admission to ICU (P = 0.006) and aortic cross-clamp time (P = 0.02). There was no difference in time to extubation between cases performed during the morning or afternoon (median time = 10 h versus 12 h, respectively, P = 0.09).

Discussion

We have demonstrated a significant reduction in the time to extubation after CABG surgery, in patients receiving a propofol-based anesthetic compared to an enflurane-based anesthetic requiring additional dosing of fentanyl and midazolam for CPB. Although this has been demonstrated previously in smaller studies, where propofol was compared to a large-dose opioid technique, it has not been shown when compared to an inhalation-based (enflurane) anesthetic. More importantly, there have been no previous studies looking at the comparative effects in a wider group of cardiac surgical patients, particularly those with impaired LV function (half our patients had moderate to severe impairment of LV contractility, or a preinduction cardiac index less than 2.2 L [center dot] min^-1 [center dot] m^-2). Our study population included patients with unstable angina requiring preoperative IV nitrate and heparin therapy, patients who had undergone previous CABG surgery, and patients with left main coronary artery disease or recent myocardial infarction. This study therefore more closely represents current cardiac surgical practice, and despite the concerns of propofol being an unsuitable drug in this broader population, our study firmly supports its use. Since this study excluded patients at highest risk (clinical severity score greater than 9), conclusions cannot be extended to such patients.

Earlier extubation was not associated with adverse hemodynamic effects, although there was a trend for more patients in the propofol group to be treated with epinephrine. This, in part, may be due to the protocol guiding management of hypotension. Because propofol has been shown to cause hypotension ^[9,12,21], the risk of myocardial ischemia has remained a concern for anesthesiologists ^[7,8,14-16]. Mora et al. ^[16] found no difference between their four groups, with a 10% to 20% incidence of intraoperative myocardial ischemia (detected by Holter ST-segment changes). Using intraoperative ST-segment analysis, a less sensitive measure of ischemia, we found no evidence of increased myocardial ischemia or postoperative myocardial infarction in propofol-treated patients, supporting the safety of propofol in a diverse range of patients undergoing CABG surgery. However, demonstration of a 50% change in postoperative myocardial infarction rate (with an incidence of 2%), would require a much larger study (in excess of 5000 patients).

The extubation time curve (Figure 1) illustrates the differential effect of the anesthetic techniques studied and suggests that choice of anesthetic has little impact after 18-20 hours, when other factors, such as intrinsic respiratory insufficiency and inotrope dependence, dictate ventilator weaning and extubation. Although we could not demonstrate any difference in ICU discharge times, this probably reflects discharge policy and practice in our ICU. This has also been noted by other authors ^[16,22]. It is logical that earlier extubation could be associated with earlier transfer to an environment where staffing ratios are reduced, such as a high-dependency unit ^[13,18]. This would have substantial cost benefits.

Several recent studies have investigated methods which may allow more cost-effective, earlier extubation and ICU discharge, followed by earlier hospital discharge ^{[13,17,23,24]}. Some of these studies should be interpreted with caution, as they were conducted in a selective (low-risk) population, or at centers with shorter aortic cross-clamp times and duration of surgery. Our institution, like most tertiary referral centers, is confronted with an increasingly higher risk population.

Hall et al. ^[14], in a small study of three groups of patients with reduced ventricular function (n = 60), compared a propofol induction and maintenance technique with sufentanil induction and propofol maintenance, or sufentanil induction and enflurane maintenance. They found similar hemodynamic changes in all three groups, but some differences in time to extubation and ICU discharge. Patients maintained with propofol had evidence of less myocardial ischemia than those maintained with enflurane, although this only occurred immediately after induction, and all groups were otherwise comparable. Bell et al. ^[15] compared propofol to a large-dose fentanyl technique in 39 patients with poor ventricular function and demonstrated earlier extubation and ICU discharge, with similar hemodynamic changes throughout. Nevertheless, it is axiomatic that large-dose fentanyl will prolong recovery time, and the question that remained unanswered was whether propofol, when compared to a balanced anesthetic approach which included a volatile drug, could shorten recovery time without adversely affecting hemodynamics in a higher risk population ^[6,25]. Mora et al. ^[16], in patients with normal LV function (four groups, n = 90), could not demonstrate any difference between a propofol-based anesthetic compared with enflurane (although they demonstrated a shorter extubation time when compared to a large-dose fentanyl or thiopentone anesthetic).

Although inhalation-based techniques have been shown to assist early extubation, their use is typically limited to patients with preserved myocardial function because of anesthetic-induced myocardial depression ^[1-3,5]. As confirmed in this study, patients with reduced LV function can be safely managed with a combined technique using a lower concentration of volatile drug (to reduce myocardial depression) and less opioid (to reduce recovery time). Despite this, we have shown that propofol-treated patients can be extubated faster.

An acknowledged deficiency of our study was that the investigators were not blinded to treatment allocation. Nevertheless, all end points were predetermined and objective, with ICU management occurring independent of the investigators. We retained details of intraoperative drug treatment until after extubation, so that ICU staff were unaware of group allocation, and so decisions were minimally biased. Another deficiency is that adverse hemodynamic changes were able to be treated, so that our ability to demonstrate differences in hemodynamic outcomes was reduced. Nevertheless, use of vasoconstrictors and inotropic drugs did not differ, nor did changes in blood pressure and systemic vascular resistance. Drug administration and interventions used in this study were dictated by protocol. Future anesthetic management could include individualization of dosages, allowing further reduction in opioid administration and the benefit of further reduction of time required to ventilate patients postoperatively ^[26].

There has been increased interest in "fast-tracking" after cardiac surgery, with the emphasis on earlier extubation and ICU discharge. Naturally, economic pressures are driving this interest, but improved patient outcomes can also occur. An anesthetic technique that can safely facilitate this process in most patients presenting for cardiac surgery is welcome, and a propofol-based anesthetic appears to fulfill these requirements. There are also other suggested benefits of propofol anesthesia, including lower rates of nausea and vomiting ^[27] and cerebral protection ^[11,28].

This study demonstrates that most patients undergoing CABG surgery can be safely anesthetized with a propofol-based technique, without increased hypotension, intraoperative myocardial ischemia or postoperative myocardial infarction, compared to an enflurane-based technique requiring additional dosing of fentanyl and midazolam for CPB, resulting in a significant reduction in time to extubation after surgery.

We would like to thank the nursing staff in the operating theaters, cardiothoracic ICU (Ward 1F) and postoperative ward (Ward 2F) for their cooperation and interest in this trial. We would also like to acknowledge the cooperation of the cardiac surgeons, and thank Dr. Meroula Richardson (cardiologist) for her expert assistance with the ECG reporting. The Graseby 3400 syringe pumps were provided by ICI Pharmaceuticals (Australia).