The rate of adaptation of coronary blood flow in response to stepwise changes in heart rate (HR) has been extensively studied in dogs and goats to improve our understanding of the dynamics of coronary regulation processes and their pathophysiology and to obtain time constants for mathematical modeling of the coronary regulation. However, little is known about the dynamic characteristics of coronary flow adaptation in humans. In patients undergoing coronary artery surgery, we investigated the rate of coronary adaptation in response to stepwise changes in HR, in the awake and anesthetized states.
In 11 patients with stable coronary artery disease, arterial blood pressure, right atrial pressure, and coronary sinus blood flow, measured by continuous thermodilution, were calculated per beat. The ratio of beat-averaged arterial blood pressure minus right atrial pressure and coronary sinus blood flow was calculated to obtain an index of coronary resistance. The rate of change of coronary resistance index was quantified by t50, defined as the time required to establish 50% of the total change in coronary resistance index. Responses of coronary resistance index after HR changes, before and after induction of anesthesia, were compared. The anesthesia technique consisted of 100 micrograms.kg-1 fentanyl and 0.1 mg.kg-1 pancuronium bromide in combination with oxygen in air ventilation (FIO2 = 0.5).
In the awake situation, t50 values of the dilating and constricting responses, induced by an increase and a decrease in HR were 5.0 +/- 2.1 (SD) s (range 2.6-9.0 s) and 5.7 +/- 1.2 s (range 4.1-7.8 s), respectively. During fentanyl/pancuronium anesthesia, the rate of coronary flow adaptation was significantly slower, with t50 values of 10.2 +/- 2.1 s (range 7.7-13.1 s) after an HR step-up and 9.8 +/- 2.1 s (range 6.6-13.2 s) after an HR step-down. Compared to the awake situation, arterial blood pressure was significantly reduced during anesthesia, but coronary vascular resistance remained unchanged. This implies that the steady-state static regulation of coronary blood flow had not changed.
These preliminary data suggest that, in patients with coronary artery disease, the rate of change in coronary vascular resistance in response to pacing-induced changes in HR is mitigated by fentanyl/pancuronium anesthesia during positive pressure ventilation. A further qualification of our findings in a larger number of patients is warranted.
Key words: Analgesic opioid: fentanyl. Coronary artery disease. Coronary artery surgery. Heart rate changes. Muscle relaxant: pancuronium. Rate of coronary blood flow adaptation.
IN 1947, Eckenhoff et al. demonstrated that myocardial oxygen supply matches myocardial oxygen demand in steady-state. This finding was confirmed by a number of investigators, using different species of experimental animals. [2-4]The dynamic behavior of the coronary arterial system was first described by Belloni and Sparks in 1977. Using open-chest dogs, they calculated the time course of changes in coronary vascular resistance (CVR) in response to pacing-induced changes in heart rate (HR). Using dogs and goats, Dankelman et al. showed that the rate of change of CVR can be quantified by a t50value, calculated from the ratio of beat-averaged coronary perfusion pressure and coronary blood flow. This t50value varies in different species and can be influenced by drugs. [6-8]Neither in experimental animals nor in humans is it known whether there is a difference in the rate of coronary flow regulation during awake and anesthetized conditions, although the impact of anesthesia on the static relation between myocardial oxygen consumption (MVO2) and coronary blood flow is well documented. [9-13]The effect of a faster or slower coronary response to changes in HR and MVO2on physiologic and pathophysiologic processes is also unknown. It is conceivable that, especially in patients with coronary stenoses, too slow a response might lead to myocardial ischemia, whereas a faster response might preserve the ratio between myocardial oxygen supply and MVO2. This may be a simplification, because the mechanism involved in regulation of coronary blood flow is a complex process involving several factors, including metabolic, myogenic, and neurohumoral regulation and endothelial responses. [3,14-18]The process can be described as a high-order control system. These types of control systems can oscillate depending on phase shifts and amplification factors within the system. It is therefore conceivable that the mechanism(s) involved in the dynamics of flow regulation and pathologic processes, including coronary artery disease (CAD) and endothelial lesions, may interact with each other. Within such an interaction, a fast dynamic response of the coronary system (short t50value) may lead to unstable flow regulation, which in turn may be associated with imbalanced flow distribution or vasospasm. There is evidence that coronary vasospasm plays an important role in the etiology of perioperative myocardial ischemia, developing in patients with CAD undergoing surgery. [20,21]These ischemic episodes are significantly more prevalent in the period before induction of anesthesia (i.e., in the awake situation) than during deep surgical anesthesia. [22-28].
Measuring the response time of the human coronary system before and during anesthesia in the same patients is a first step in understanding the potential role of the dynamics of coronary flow adaptation in the pathophysiology of coronary perfusion in humans. Therefore, the current study was designed to measure the dynamic characteristics of coronary flow adaptation in response to stepwise changes in HR in awake and anesthetized patients with CAD, scheduled for coronary artery surgery.
Methods and Materials
Eleven patients with stable CAD and scheduled for elective coronary artery surgery gave informed consent to participate in this study, which had institutional approval. Excluded from the study were patients with the following conditions: left ventricular end diastolic pressure > 18 mmHg, left ventricular hypertrophy (assessed by electrocardiographic or echocardiographic criteria), ejection fraction < 45%, atrioventricular conduction defects, left main coronary stenosis, and unstable angina. Patients undergoing additional surgical procedures, e.g., valve replacement or aneurysmectomy, also were excluded.
On arrival in the operating room, electrocardiogram leads were connected. Leads II, III, and V5were continuously monitored (HP Merlin System, Hewlett-Packard, Boblingen, Germany). A wide-bore peripheral infusion and a 20-G radial artery cannula were inserted under local analgesia.
In awake patients, a triple-lumen pulmonary artery catheter (Baxter, Irvin, CA) and a coronary sinus thermodilution catheter (Wilton-Webster, Altadena, CA, type CCS-7U-90B) were introduced via the left subclavian vein. The coronary sinus catheter was advanced into the coronary sinus using image intensification fluoroscopy and injection of contrast medium, so that the external thermistor lay 1.5-2 cm from the ostium and there was no major side-branching vein in the vicinity. The coronary sinus catheter was connected to a Wheatstone bridge (Wilton-Webster). Coronary sinus thermodilution signals were recorded with a multichannel amplifier/recorder system. Catheter calibration factors provided by the manufacturer were used. The absence of right atrial admixture in coronary sinus blood was checked by injection of cold saline in the right atrium, while coronary sinus temperature curves were recorded simultaneously. [29,30]Under fluoroscopy, pacing (via coronary sinus catheter) was used for 10-30 s to ascertain the stability of the position of the tip of the coronary sinus catheter in relation to the surrounding anatomic structures and fluoroscopic landmarks. If the stability of the catheter could not be guaranteed at that time, the experiment was discontinued prematurely. In addition, the electrical threshold for pacing was determined. For the measurement of coronary sinus blood flow (CSBF) normal saline at room temperature was used as an indicator and infused into the coronary sinus at a rate of 45 ml *symbol* min sup -1 via a Mark IV infusion pump (Medrad Technology for People, Pittsburgh, PA). Infusion rates were verified by timed volume collection, and flow calculations reflected the indicator infusion rate that was used.
Calcium channel blockers and long-acting nitrates were given until the evening before surgery. beta-Adrenoreceptor blocking agents were continued until the morning of surgery.
Lorazepam (4-5 mg) was given for premedication 2 h before surgery.
In the operating room, while the patients breathed oxygen, 2 mg pancuronium bromide was given, followed by 100 micro gram *symbol* kg sup -1 fentanyl injected over 5 min. When the patient became unresponsive to commands, an additional dose of 6 mg pancuronium was given and ventilation was assisted and then controlled manually. After intubation of the trachea, the lungs were ventilated with air/oxygen (FIO2= 0.5). Ventilation was adjusted to maintain the end-tidal carbon dioxide concentration between 4 and 4.5%. In the first 15 min after induction of anesthesia, 250-500 ml of gelofusine (Vifor Medical SA, Crisier, Switzerland) was infused to maintain a stable hemodynamic situation after induction of anesthesia. Gelofusine is a gelatin solution containing, per 500 ml, 20 g of modified gelatin, 77 mmol Nitrogen sup +, and 63 mmol Chlorine sup -, with pH 7.1-7.7 and osmolality 279 mOsm/l.
After adequate instrumentation and a resting period of 20 min, the following measurements were performed in awake and anesthetized patients (Figure 1).
At the beginning of each series of measurements, a complete set of hemodynamic measurements, including pulmonary capillary wedge pressure and single bolus thermodilution cardiac output, was obtained, using injectate at room temperature. Cardiac output is calculated as the average of at least three measurements and reported as cardiac index. Blood samples from the radial artery and coronary sinus were drawn to determine plasma hemoglobin concentration, oxygen partial pressure (pO2), hemoglobin oxygen saturation (SatO2), and lactate concentration.
After completion of the intermittent measurements, continuous recording of CSBF, arterial blood pressure (ABP), right atrial pressure (RAP), pulmonary artery pressure, and electrocardiographic lead II was started at a sampling rate of 80 Hz and stored on a computer disk. When a steady-state in coronary sinus thermodilution signal was obtained after 10-15 s of registration, HR was abruptly increased by 25 beats/min by pacing via the coronary sinus catheter. After a recording period of 70 s, indicator infusion was discontinued. Pacing was continued at the same rate. Immediately after the completion of the continuous recordings, all intermittent measurements and blood sampling were repeated.
During pacing, continuous measurements were resumed. After a steady-state in CSBF signal was obtained (10-15 s), pacing was stopped. Recording of thermodilution signals, blood pressures, and electrocardiogram continued for a total duration of 70 s. A final series of hemodynamic measurements and blood sampling completed the measuring protocol in the awake patient. Anesthesia was induced following the technique described above. Twenty minutes after induction of anesthesia and tracheal intubation, the complete protocol was repeated.
Below, the increase and decrease in HR will be referred to as HR step-up and HR step-down, respectively.
Laboratory Analysis and Calculations
SatO2was measured by an OSM-II hemoxymeter (Radiometer, Copenhagen, Denmark) and pO2by an ABL-III (Radiometer). Lactate concentrations were measured using standard enzymatic techniques. .
Calculated hemodynamic parameters were obtained from measured hemodynamic signals using standard formulas: Equation 1where CI indicates cardiac index; CO, cardiac output; and BSA, body surface area.
Myocardial metabolic indexes were calculated according to standard formulas: Equation 2where cO2indicates oxygen content; Hb, plasma hemoglobin concentration; SatO2, hemoglobin oxygen saturation; cO2,art - cO2,ca, difference between arterial and coronary sinus oxygen content; MVo2, myocardial oxygen consumption; MLE, myocardial lactate extraction percentage; lactateart, lactate concentration in arterial blood; and lactatecs, lactate concentration in coronary sinus blood.
The response of the coronary vascular tree to pacing-induced stepwise changes in HR was analyzed as described by Dankelman et al. ABP, RAP, and CSBF were averaged per beat. The coronary resistance index (CRI) was calculated as the quotient of beat-averaged ABP-RAP and CSBF. The CRI is identical to CVR in steady-state. Under conditions in which flow and/or pressure vary so slowly that capacitance effects can be ignored, CRI reflects CVR. During fast dynamic changes in driving pressure or CSBF, CRI does not reflect CVR. [7,33]The first 2 s after an HR step were not included in the analysis of CRI, because after 2 s, the major capacitive effects are complete.
The rate of the coronary adaptation was quantitated by a t50value, which was defined as the time in seconds after an HR-step at which the change in CRI had reached 50% of its total change. Using a signal analysis program (386-Matlab, version 3.5j (1991), MathWorks, Natick, MA), this value was calculated by fitting a high-order polynomial to a part of the signal over a period of at least 10 s around the 50% value of the CRI (Figure 2).
To compare the time course of the response of the CRI to the different interventions, the CRI was normalized. The normalized index was calculated by averaging CRI over 8 s before the HR step, yielding CRIO. The normalized CRI is given by: Equation 3.
The normalized response of the CRI starts at unity, and as a result of the coronary regulation process, it decreases (HR step-up) or increases (HR step-down).
Data were analyzed using paired standard t tests. A value of P < 0.05 was considered significant. Results are reported as mean+/-SD or as percentage change+/-SD where applicable.
Characteristics of the Patients
Patients characteristics and preoperative chronic medication are shown in Table 1. The patients were comparable with respect to age, weight, and height. Although five patients had suffered a previous myocardial infarction and three had (treated) hypertension, there was no evidence of either impaired left ventricular function and dilatation or left ventricular hypertrophy. One patient (4) was not using chronic beta-adrenoreceptor blockers. Patients 3 and 5 were not using calcium entry blockers. Patients 2, 6, 9, and 10 were not using long-acting nitrates preoperatively.
Hemodynamic and Metabolic Characteristics
Systemic and coronary hemodynamic variables and myocardial metabolic data are listed (Table 2). Hemodynamic results obtained in the awake and anesthetized conditions are reported at baseline, after onset of pacing (HR step-up), and before and after discontinuation of pacing (HR step-down). The stepwise changes in HR tended to be larger during anesthesia, because after induction of anesthesia, HR decreased from 60 to 55 beats/min, whereas the rate of pacing that was used in the awake situation was not changed. All HR steps resulted in a change in ABP in the direction of the HR step, and during anesthesia, these changes in ABP were significantly larger, compared to the awake state. Awake, an HR step-up resulted in a blood pressure increase of 4 mmHg, whereas during anesthesia, this pressure increase was 14 mmHg. Similarly, an HR step-down resulted in a blood pressure decrease of 3 mmHg in the awake state and a blood pressure decrease of 15 mmHg in the anesthetized state. Compared to the awake situation, ABP decreased in all patients after induction of anesthesia.
After an HR step-up, mean CSBF increased by 46% in the awake state and by 46% in the anesthetized state, and after an HR step-down, mean CSBF decreased by 29% and 32%, respectively. Compared to the awake situation, CSBF decreased in all patients after induction of anesthesia. CVR decreased significantly after an HR step-up, remained at the same level during pacing, and increased after an HR step-down. These changes were in the same order of magnitude in the awake and anesthetized conditions (P > 0.39).
Before an HR step-up, during a pacing period (with CSBF at steady-state), and after an HR step-down (Figure 1), coronary sinus oxygen tension (pcsO2), MVo2, and myocardial lactate extraction percentage could be calculated. In the awake and anesthetized situations, pcsO2did not change in response to HR changes, although it increased 3-4 mmHg (P < 0.05) after induction of anesthesia.
MVO2increased significantly during pacing, and this change in MVO2was similar in the awake and anesthetized states. However, at baseline before the HR step-up and step-down, the level of MV sub O2was 1.3 ml O2*symbol* min sup -1 and 1.2 ml O2*symbol* min sup -1 less (P < 0.05) during anesthesia than in the awake state.
Myocardial lactate extraction percentage did not change during pacing and after induction of anesthesia and remained positive in all our patients, during all interventions. Furthermore, we did not observe electrocardiographic signs of myocardial ischemia in any patient at any time during the study period.
There was no significant change in arteriovenous oxygen content difference between the awake and anesthetized states. The arterial oxygen content increased from 8.0+/-0.9 mmol O2*symbol* l sup -1 awake to 8.5+/-1.0 mmol O2*symbol* l sup -1 during anesthesia (P < 0.05), while arterial oxygen supply decreased, due to a decreased coronary flow. The myocardial oxygen extraction percentage (M sub O2E) did not change in response to the HR steps but decreased after induction of anesthesia because of the increase in arterial oxygen content.
Dynamic Characteristics of Coronary Control
Typical results obtained by increasing HR are shown in Figure 3. In the awake state, CSBF increased to a new steady-state level, indicating coronary dilation. In the anesthetized state, this vasodilating response became slower, as is clear from the normalized response of CRI.
The response rates as induced by the different interventions of all individual patients were quantified by t50values (Table 3). Fentanyl anesthesia clearly reduced the rate of adaptation of the CRI. In the awake situation, t50values of the dilating and constricting responses, induced by an increase and decrease in HR were 5.0+/- 2.1 (SD) s and 5.7+/-1.2 s, respectively. During anesthesia, the rate of change of coronary flow adaptation was significantly decelerated with t50values of 10.2+/-2.1 s after an HR step-up and 9.8 +/-2.1 s after an HR step-down. The rate of the vasoconstricting response to a decrease in HR and the vasodilating response to an increase in HR were similar and this applied to the awake and anesthetized states.
The aim of the current study was to measure the rate of change of coronary blood flow in response to HR steps in awake and anesthetized patients with CAD. It was demonstrated that, in awake patients, the dynamic process of adaptation of coronary flow in response to stepwise changes in HR and MVO2takes place in 5.2+/-1.6 s. Fentanyl/pancuronium anesthesia significantly (P < 0.001) delayed this adaptation to 10.0+/-1.7 s.
In the current study, the coronary venous thermodilution technique was chosen, because this technique has been used successfully in hundreds of patients undergoing cardiac or noncardiac surgery. [34-36]There are a number of limitations to its use that are related to the physiology and anatomy of the coronary venous drainage and to the thermodilution technique. [29,30,37-42]The major anatomic problem is formed by the existence of extensive cardiac venous intercommunications, resulting in drainage of left anterior descending arterial blood through routes other than the great cardiac vein and the coronary sinus. [38,40,42]This may result in an underestimation of coronary blood flow. Because we focused on the dynamic adaptation of flow in response to HR steps and not on absolute blood flow values, these venous intercommunications and the potential underestimation of the absolute coronary blood flow are probably of no importance for this study.
Movement of the catheter in the coronary sinus may induce an important change in measured flow, presumably because of the varying contribution of venous tributaries. [37,41]Both pacing and changes in heart size after the induction of anesthesia are factors that may have influenced the position of the coronary sinus catheter. The effect of pacing on the stability of the tip of the coronary sinus catheter was ascertained, and we have no indication from the thermodilution recordings that shifts in catheter position after an HR step occurred.
Movement of the catheter after induction of anesthesia could have been partly responsible for the absolute reduction in CSBF. However, this does not affect the measurement of the rate of the coronary responses after an HR step. For the calculation of the t50values, the rate of change of CVR in response to an HR step was used, and therefore, absolute values of CSBF are of minor importance. Furthermore, absolute CVR was similar before each series of measurements, both awake and during anesthesia.
Possible Influence of the Extent of CAD on the Rate of Change of Coronary Flow
Comparison of our findings with healthy human data is not possible because, to the best of our knowledge, data describing the rate of adaptation of coronary flow in humans have not been reported. Table 4shows the previously reported t50values (recalculated) of the coronary arterial systems of dogs and goats and the t50values obtained in patients with CAD in the current study. Therefore, all t50values were calculated using the same analysis technique described in the methods section. It appears that the rate of coronary adaptation in anesthetized humans is similar to the rate of coronary adaptation in anesthetized goats. In dogs and goats, using a coronary perfusion system, the effects of constant pressure and constant flow perfusion could be studied. It was shown that the t50values of the coronary system were +/-50% larger at constant flow perfusion than at constant pressure perfusion. [6,7]Constant flow perfusion was induced by placing a clamp on the perfusion catheter, which may resemble the effect of a stenosis. All our patients had severe coronary lesions, although not left main coronary stenosis. Hence, in case the coronary stenoses in our patients had affected the results of the current study, the reported t50values are probably larger than the values we would have found in patients without coronary stenoses in the awake and anesthetized states.
Rate of Change of Myocardial Oxygen Consumption and Changes in Coronary Hemodynamics
A number of factors may have influenced the rate of change of the coronary vascular wall after induction of anesthesia. These factors include the rate of change of MVO2, the reduction in coronary perfusion pressure and CSBF, and the increase in coronary venous posub 2.
In our clinical experimental setup, we could not measure oxygen consumption of the myocardium continuously. Comparing the course of the rate pressure product as a measure of MVO2, after a sudden change in HR in the awake and anesthetized state, the change in rate pressure product after an HR step was always instant. This suggests an instant and similar change in MVO2, [43,44]both awake and during anesthesia. It therefore seems unlikely that the rate of change of MVO2played a role in the observed deceleration of coronary responses after induction of anesthesia.
The reductions in CSBF and ABP, and thus coronary perfusion pressure, associated with induction of anesthesia are factors that must be considered as a potential explanation for the reduction in the rate of change in CRI. Animal experiments have yielded evidence that the rate of coronary adaptation is relatively independent of the level of coronary blood flow and that a reduction in coronary perfusion pressure results in an increase and not a decrease in the rate of change of coronary resistance. That implies that, in our study, reported t50values during anesthesia might have been even longer if ABP would have remained unchanged after induction of anesthesia. Furthermore, it is unlikely that differences in CVR and HR at baseline have affected the results of our study. CVRs at baseline were similar in the awake and anesthetized states. Although there was a significant reduction in HR after induction of anesthesia (60 beats/min awake vs. 55 beats/min under anesthesia), this difference at baseline is within such a narrow range that it cannot have played a role.
Finally, the increase in coronary venous pO2after induction of anesthesia may have influenced the rate of coronary flow adjustment. Dole et al. demonstrated in anesthetized dogs that the strength of coronary autoregulation and its dynamic behavior is influenced by the level of coronary venous pO2. They concluded that there is a marked attenuation of autoregulation at venous pO2levels above 30 mmHg. In our study, the venous pO2was below the threshold level of 30 mmHg at all times, although there was a small but significant change in coronary venous pO2between the awake and anesthetized states.
Thus, no myocardial metabolic or coronary hemodynamic factors can clearly explain the reported increase in the rate of coronary flow adjustment after the induction of anesthesia.
Pharmacologic Effects Directly or Indirectly Induced by Anesthetic Agents
In the current study, fentanyl and pancuronium bromide were the only pharmacologic agents used for the induction and maintenance of anesthesia and muscle relaxation. Unfortunately, little is known about the impact of these agents on the multifactorial process of coronary flow regulation. Therefore, it is not possible to develop a specific pharmacologic mechanism explaining the effect of anesthesia on the dynamic response of human coronary flow regulation.
It is known from animal models that the rate of change of coronary resistance can be influenced by drugs. It has been demonstrated in goats with normal coronary arteries anesthetized with fentanyl that glibenclamide, an antidiabetic drug with blocking effects on adenosine triphosphate-dependent potassium channels in vascular smooth muscle tissue, can reduce the rate of change of coronary resistance by a factor of 4, without changing the steady-state relations of coronary hemodynamics. Glibenclamide has been shown to have an inhibiting effect on membrane hyperpolarization of the vascular smooth muscle cell. In contrast, fentanyl is associated with membrane hyperpolarization of neuronal cells via potentiation of the Potassium sup + channel current. [48,49]However, it is not known whether there is a direct effect of fentanyl on myocardial or coronary vascular smooth muscle membranes. An indirect effect of fentanyl, via reduction of circulating catecholamine concentrations and sympathetic activity, [50-53]appears to be an unlikely factor to explain our findings, because all patients (except for patient 4) were adequately beta-blocked both awake and during anesthesia. However, the involvement of coronary alpha-receptor constrictor mechanisms (in the presence of high levels of circulating catecholamines) competing with metabolic vasodilation cannot be excluded. Pancuronium bromide is a steroid-based nondepolarizing muscle relaxant with additional blocking effects on muscarinic receptors. [11,55]Muscarinic receptors are involved in the local release of endothelium-derived relaxing factor from endothelium. [56-58]This substance plays a role in the local regulation of coronary blood flow via an effect on flow-dependent dilatation. [59-61]Theoretically, this mechanism might affect our findings.
Clinical Implications of the Current Findings
The clinical implications of our findings remain mostly speculative, because the current study was not designed to address a specific clinical problem and because the concept of dynamic coronary flow adaptation has never been described before as a parameter of coronary function in humans. From the results of our study, one cannot conclude that the awake coronary responses were "normal" and the responses obtained under anesthesia were slow or the opposite, i.e., that the responses in awake patients with CAD were fast and the responses under anesthesia were normal. Thus far, no studies describing the dynamics of coronary flow in humans with normal coronary arteries have been reported, and therefore, we do not know the normal t50value of coronary flow adaptation in healthy humans.
The higher rate of flow adjustment found during our measurements in awake patients might imply that the coronary system can follow a high frequency of stimulation and, therefore, has little damping and a high probability of oscillation. Oscillation may result in unstable flow regulation. Although highly speculative, it is possible that unstable flow regulation in the presence of multiple rapid changes in HR and MVO2may lead to an imbalance in flow distribution or coronary spasm and, subsequently, myocardial ischemia.
Finally, the results of the current study suggest the existence of a specific mechanism for controlling the dynamics of coronary resistance. Because fentanyl in combination with pancuronium can influence this mechanism, it may be possible to identify or develop other compounds with a specific effect on the rate of coronary flow adjustment.
Fentanyl/pancuronium anesthesia did not change steady-state regulation of coronary blood flow in patients with CAD. However, our preliminary findings suggest that, in these patients, the rate of change of coronary flow adjustment is reduced by this anesthetic technique. A further qualification of our findings in a larger number of patients is warranted.
The authors thank Arjan Hoksbergen, Michiel van der Meer, and Aart Boekee, for their technical assistance; Jan Fiolet, Ph.D., and his team, for measuring the lactate concentrations; and Julian I. E. Hoffman, M.D., Ph.D., and Frits L. Meijler, M.D., Ph.D., for reviewing the manuscript.