Cardiovascular Effects of Propofol in Dogs with Dilated Cardiomyopathy 

All experimental procedures and protocols used in this investigation were reviewed and approved by the Animal Care and Use Committee of the Medical College of Wisconsin. All procedures conformed to the Guiding Principles in the Care and Use of Animals of the American Physiological Society and were performed in accordance with the Guide for the Care and Use of Laboratory Animals.

The surgical implantation of instruments has been described in detail previously. [23]Briefly, in the presence of general anesthesia and using aseptic techniques, a left thoracotomy was performed in conditioned mongrel dogs for placement of instruments for measurement of aortic, left atrial, and intrathoracic pressures (heparin-filled catheters); aortic blood flow (Doppler flow transducer); left anterior descending coronary artery blood flow velocity (ultrasonic flow transducer); subendocardial segment length (ultrasonic crystals); LV pressure (high fidelity, miniature micromanometer); and the peak rates of increases and decreases in LV pressure (dP/dt^maxand -dP/dt^min, respectively). A hydraulic vascular occluder was placed around the inferior vena cava for abrupt reduction of LV preload. Stainless steel pacing electrodes were sutured to the epicardial surface of the LV free wall. All instruments were firmly secured, tunneled between the scapulae, and exteriorized via several small incisions. The pericardium was left open wide, the chest wall was closed in layers, and the pneumothorax was evacuated by a chest tube. Each dog was fitted with a jacket to prevent damage to the instruments, catheters, and pacing electrodes.

All dogs received fentanyl for analgesia as needed after surgery. Dogs were allowed to recover for a minimum of 7 days before experimentation, during which time all were treated with intramuscular antibiotic agents (cephalothin at 40 mg/kg and gentamicin at 4.5 mg/kg) and trained to stand quietly in a sling during hemodynamic monitoring. Segment length and coronary blood flow velocity signals were monitored by ultrasonic amplifiers. End-systolic and end-diastolic segment lengths were measured 30 ms before LV -dP/dt^minand immediately before the onset of LV isovolumic contraction, respectively. Percent segment shortening (%SS) was determined using the equation %SS =(EDL - ESL)[center dot] 100 [center dot] EDL sup -1. Relative diastolic coronary vascular resistance was calculated as the quotient of diastolic arterial pressure to diastolic coronary blood flow velocity. An estimate of myocardial oxygen consumption, the pressure-work index, was determined using a previously validated formula. [28]Hemodynamic data were recorded continuously on a polygraph and simultaneously digitized and recorded on a computer.

After each dog (n = 7; mean +/- SEM weight, 25.3 +/- 0.4 kg) had recovered from surgery, the LV of the dog was paced continuously at rates between 220 and 240 beats/min as previously described. [23]Dogs were brought to the laboratory on each day after the initiation of pacing to monitor the development of pacing-induced cardiomyopathy. Pacing was discontinued during and restarted immediately after this brief period of daily hemodynamic monitoring and was discontinued for the duration of the experiment. All dogs were fasted overnight before experimentation. Fluid deficits were replaced with 0.9% saline (500 ml), and intravenous saline infusion was continued at 3 ml [center dot] kg sup -1 [center dot] h sup -1 for the duration of each experiment.

After instruments were calibrated, baseline systemic and coronary hemodynamics were recorded in the conscious state (sinus rhythm). Left ventricular and intrathoracic pressures and segment length waveforms also were recorded for later off-line analysis of LV systolic and diastolic function. Regional myocardial contractility was evaluated using a series of LV pressure-segment length diagrams generated by abrupt inferior vena cava constriction as previously described. [23]The slope of the regional preload recruitable stroke work (M^W) relation derived from these LV pressure-segment length diagrams was used to quantify myocardial contractility. The time constant of LV isovolumic relaxation (tau) was calculated using the derivative method. Segment-lengthening velocities during early ventricular filling (dL/dt sub E) and atrial systole (dL/dt^A) and their corresponding time-velocity integrals (TVI-E and TVI-A, respectively) were determined from the continuous dL/dt waveform. Early ventricular filling fraction was calculated as the fraction of total segment lengthening occurring during early ventricular filling. A regional chamber stiffness constant (K) was derived from LV pressure-segment length data between minimum ventricular pressure and the beginning of atrial systole using a monoexponential relation assuming a simple elastic model.

Left ventricular afterload was quantified with Zⁱⁿ(omega) spectra and interpreted using a three-element Windkessel model of the arterial circulation as previously described. [29]Briefly, digitized, steady-state aortic blood pressure and blood flow waveforms were transformed from the time to the frequency (omega) domain using power spectral analysis to determine Zⁱⁿ(omega). The autopower spectrums of aortic pressure (P^pp[omega]) and blood flow (P^ff[omega]) and the crosspower spectrum between aortic pressure and blood flow waveforms (P^pf[omega]) were determined using a Welch periodogram, and Zⁱⁿ(omega) was calculated as the ratio of P^pp(omega) to P^pf(omega). Each calculated Zⁱⁿ(omega) spectrum was corrected for the phase responses of the aortic pressure and blood flow transducers. Correlation of aortic pressure and blood flow waveforms at each frequency of the Zⁱⁿ(omega) spectrum was determined using magnitude squared coherence. Zⁱⁿ(omega) data with magnitude squared coherence values < 0.8 were discarded from the analysis. Characteristic aortic impedance (Z^C) was determined as the magnitude of Zⁱⁿ(omega) between 2 and 15 Hz. Total arterial resistance (R) was calculated as the difference between Zⁱⁿ(omega) at 0 Hz and Z^C. Total arterial compliance (C) was determined from aortic pressure and blood flow waveforms using a previously validated formula. [30] 

Anesthesia was induced with 5 mg/kg of intravenous propofol and maintained with propofol infusions at 25, 50, and 100 mg [center dot] kg sup -1 [center dot] h sup -1 administered in a sequential manner. The lungs of each dog were mechanically ventilated with oxygen-supplemented air (25% oxygen). Arterial blood gas tensions were maintained at levels obtained in the conscious state by adjustment of air and oxygen concentrations and respiratory rate throughout each experiment. Systemic and coronary hemodynamics and LV pressure, intrathoracic pressure, aortic pressure, aortic blood flow, and myocardial segment length waveforms were recorded under steady-state conditions and during inferior vena cava occlusion at end expiration after 15 min of equilibration at each propofol infusion rate. Thus, the effects of propofol on LV systolic and diastolic function and LV afterload were compared with data observed in the conscious state in chronically instrumented dogs with pacing-induced cardiomyopathy.

Statistical analysis of the data in the conscious state and during propofol anesthesia was performed by analysis of variance with repeated measures, followed by use of Student's t test with Duncan's adjustment for multiplicity. A probability value < 0.05 was considered statistically significant. All data are expressed as mean +/- SEM.

The hemodynamic effects of propofol in dogs with pacing-induced cardiomyopathy are summarized in Table 1. Propofol caused significant and dose-related decreases in mean arterial pressure, LV systolic and end-diastolic pressures, end-diastolic segment length, and end-systolic segment length (P < 0.05). Decreases in rate-pressure product, pressure-work index, and diastolic coronary vascular resistance were observed during infusion of the highest dose of propofol. Heart rate and diastolic coronary blood flow velocity were unchanged. Propofol caused dose-related reductions in M^W(67 +/- 5 during control to 34 +/- 4 mmHg during the 100 mg [center dot] kg sup -1 [center dot] h sup -1 dose;Figure 1(A)), +dP/dt^max, and %SS in cardiomyopathic dogs, which is consistent with depression of intrinsic myocardial contractility. Propofol also caused dose-related decreases in mean aortic blood flow (2.14 +/- 0.25 during control to 1.47 +/- 0.17 l/min during the 100 mg [center dot] kg sup -1 [center dot] h sup -1 dose;Table 1) and stroke volume. The lowest dose of propofol caused a decrease in tau (62 +/- 4 during control to 54 +/- 4 ms;Figure 1(B)); however, no changes in tau were observed during administration of higher doses. Decreases in K (2.05 +/- 0.54 during control to 0.33 +/- 0.06 min sup -1 during the 100 mg [center dot] kg sup -1 [center dot] h sup -1 dose;Figure 1(C)) occurred concomitant with the declines in end-diastolic segment length and LV end-diastolic pressure. Propofol decreased dL/dt^E(57 +/- 9 during control to 36 +/- 7 mm/s during the 100 mg [center dot] kg sup -1 [center dot] h sup -1 dose;Figure 2(A)) and TVI-E (4.8 +/- 0.6 during control to 3.2 +/- 0.5 mm during the 100 mg [center dot] kg sup -1 [center dot] h sup -1 dose;Figure 3(A)), which is consistent with a reduction in the rate and extent of early LV filling. In contrast, dL/dt^Aand TVI-A were unchanged (Figure 2(B) and Figure 3(B), respectively), and as a result, the ratios of dL/dt^Eto dL/dt^Aand TVI-E to TVI-A decreased (Figure 2(C) and Figure 3(C), respectively). Propofol also caused a reduction in early LV filling fraction (78.5 +/- 3.0 during control to 67.9 +/- 4.3% during the 50 mg [center dot] kg sup -1 [center dot] h sup -1* dose;Table 1). Propofol decreased R (3,850 +/- 470 during control to 3,020 +/- 270 dyn [center dot] s [center dot] cm sup -5 during the 100 [center dot] mg [center dot] kg sup -1 [center dot] h sup -1 dose;Figure 4(A)) and increased C (0.70 +/- 0.90 during control to 1.79 +/- 0.23 ml/mmHg during the 100 mg [center dot] kg sup -1 [center dot] h sup -1 dose;Figure 4(B)). Z^Cremained unchanged during administration of propofol (Figure 4(C)).

The current results indicate that 18 +/- 3 days of rapid LV pacing produces cardiovascular effects that appear very similar to those we have reported previously in the identical canine model [23,24]When compared with the baseline systemic hemodynamics observed in healthy, chronically instrumented dogs, [3,12]conscious dogs with pacing-induced cardiomyopathy demonstrated increases in baseline heart rate (underlying sinus rhythm), LV end-diastolic pressure, end-diastolic segment length, and end-systolic segment length; decreases in mean arterial and LV systolic pressures; and LV systolic and diastolic dysfunction in the current and previous [23,24]studies. Cardiac output and systemic vascular resistance remained unchanged during pacing of this duration. Thus, the current investigation examined the cardiovascular actions of propofol in a well-known, extensively characterized model of dilated cardiomyopathy [23–27]that has not decompensated into frank congestive heart failure.

These results in cardiomyopathic dogs indicate that propofol did not increase heart rate. These findings are similar to those observed in acutely instrumented dogs anesthetized with ketamine and fentanyl, [31]an experimental preparation characterized by pronounced depression of LV systolic and diastolic function. [32]The current results are also supported by previous studies [16–18]examining the cardiovascular effects of propofol anesthesia in patients with coronary artery disease and mildly depressed LV performance. In these studies, propofol did not change [17]or caused only modest increases in heart rate concomitant with a reduction in arterial pressure. [16,18]The lack of change in heart rate may reflect altered baroreceptor reflex activity and downregulation of cardiac beta sub 1 -adrenoceptors concomitant with basal increases in central sympathetic nervous system tone and withdrawal of parasympathetic nervous system activity associated with the development of heart failure. [33] 

Propofol also reduced calculated indices of myocardial oxygen consumption in dogs with pacing-induced LV dysfunction, results that may be attributed to favorable decreases in LV preload, afterload, and myocardial contractility during the maintenance of constant heart rate. Diastolic coronary blood flow velocity was unchanged, and diastolic coronary vascular resistance decreased despite simultaneous reductions in coronary artery perfusion pressure. When accompanied by decreases in indices of myocardial oxygen consumption, these findings suggest that propofol causes coronary vasodilation in the setting of LV dysfunction. This conclusion should be qualified, however, because myocardial metabolism was not measured directly in the current investigation. In addition, although the pressure-work index has been demonstrated to reflect myocardial oxygen consumption under anesthesia accurately [34]and during a variety of loading changes and contractile states in vivo, [35]this index of global myocardial oxygen consumption has not been validated specifically in the presence of global LV dysfunction.

Propofol decreased LV end-diastolic pressure and reduced chamber dimension in cardiomyopathic dogs. [3,12]These results indicate that venodilation and LV preload reduction are major hemodynamic consequences of propofol anesthesia in the setting of LV dysfunction. The findings are supported by previously described decreases in right atrial and pulmonary capillary occlusion pressures during induction or maintenance of propofol anesthesia in patients with coronary artery disease [16–18]or artificial hearts. [36]Reductions in mean aortic blood flow (i.e., cardiac output) and stroke volume observed in the current study were qualitatively similar to those observed in healthy dogs with normal LV function. [3]These findings suggest that LV preload reduction during propofol anesthesia in dogs with LV dysfunction does not compromise global LV performance adversely. In fact, the decreases in cardiac dimension and chamber pressures in diastole may be beneficial.

The effects of chronic LV pacing and propofol anesthesia on LV afterload in cardiomyopathic dogs were quantified with Zⁱⁿ(omega) spectra. In contrast to calculated systemic vascular resistance, Zⁱⁿ(omega) incorporates the frequency-dependent and viscoelastic properties of the arterial vasculature and provides a more complete description of LV afterload. Zⁱⁿ(omega) spectra were interpreted using a three-element Windkessel model of the arterial circulation that describes R, C, and Z^Cas physiologically meaningful properties of the arterial system. [37]In the current investigation, mean aortic blood flow, R, and Z^Cwere similar and C was somewhat increased compared with values seen in healthy dogs during administration of propofol at identical infusion rates. [3]Reductions in cardiac output and increases in systemic vascular resistance and R do not occur until late in the development of heart failure produced by chronic rapid LV pacing, probably because autoregulatory vasodilation resulting from reduced peripheral perfusion is balanced by enhanced neurohormonal activation. [38]R and Z^Calso have been shown to remain unchanged in patients with moderate congestive heart failure, [39]presumably by a similar mechanism. The relative increase in C observed in conscious dogs with LV dysfunction (e.g., 0.53 +/- 0.04 in healthy dogs [3]vs. 0.70 +/- 0.09 ml/mmHg in cardiomyopathic dogs) probably resulted from concomitant reductions in mean arterial pressure and not from direct alterations in aortic architecture during 3 weeks of rapid LV pacing. Changes in the interrelation between structural elements in the proximal aorta may contribute to reduced C late in the natural history of congestive heart failure, however. [40] 

In the current investigation, propofol reduced R in cardiomyopathic dogs. The current findings support previous reports describing propofol-induced reductions in systemic vascular resistance in patients with coronary artery disease [16–18]or artificial hearts [36]and during cardiopulmonary bypass. [41]Propofol also increased C in dogs with pacing-induced LV dysfunction. C is determined primarily by aortic compliance, [42]and these elastic properties of the proximal aorta allow it to efficiently store and redistribute LV ejection energy during systole and diastole, respectively. The magnitude of the alterations in R and C produced by propofol was similar in cardiomyopathic compared with healthy dogs [3]; however, propofol did not increase Z^Cin dogs with pacing-induced LV dysfunction. Although the resistance of the aorta to LV ejection makes a relatively small contribution to R, an increase in Z^C, such as that observed with propofol in healthy dogs, [3]may contribute to an attenuation of LV-arterial coupling and mechanical efficiency. [43]Thus, the current findings suggest that propofol produces beneficial decreases in resistance to LV ejection in the presence of LV dysfunction by favorably affecting the mechanical properties of the aorta and the arterial vasculature without adversely increasing Z^C. The current findings are qualitatively similar to those observed with the arterial vasodilator sodium nitroprusside in patients with congestive heart failure. [39,44] 

Propofol caused dose-related negative inotropic effects in dogs with pacing-induced LV dysfunction. The magnitude of propofol-induced depression of myocardial contractility was quantified with M^W, a relatively heart rate- and load-independent index of inotropic state derived from a series of LV pressure-segment length diagrams. Propofol caused qualitatively similar decreases in M^Win dogs in the presence and absence [12]of pacing-induced LV dysfunction. A 100-mg [center dot] kg sup -1 [center dot] h sup -1 infusion of propofol caused a 51 +/- 7% decrease in M^Win cardiomyopathic dogs in the current investigation compared with a 50 +/- 6% decrease in M^Wduring a 120-mg [center dot] kg sup -1 [center dot] h sup -1 infusion of propofol in healthy dogs in our previous study. [12]The current results contrast with the findings of a recent study in vitro [20]demonstrating that propofol causes more pronounced direct negative inotropic effects in failing myocardia; however, these findings [20]were obtained in isolated LV trabecular myocardia obtained from pigs with overt signs of congestive heart failure. Conversely, the current results are partially supported by previous findings [22]demonstrating that propofol does not alter contractile function in hamsters with congenital hypertrophic cardiomyopathy. The data of Hebbar et al. [20]and Riou et al. [22]emphasize that the effects of propofol on intrinsic myocardial contractility in models of cardiomyopathy in vitro remain somewhat controversial. Differences in basal autonomic nervous system activity between cardiomyopathic and healthy dogs also may have influenced the effects of propofol on myocardial contractility in the current investigation and cannot be completely excluded from the analysis. The mechanisms responsible for the negative inotropic effects of propofol in the failing heart are unknown. It is likely that the inhibition of transsarcolemmal calcium current [45]and L-type calcium channel function [46]observed with propofol in normal cardiac muscle also may play key roles in the contractile depression observed in the current and previous investigations. [20,21]These hypotheses remain to be tested in vitro, however.

The 25-mg [center dot] kg sup -1 [center dot] h sup -1 dose of propofol decreased tau in dogs with pacing-induced LV dysfunction despite a concomitant reduction in contractility. This result probably occurred as a consequence of decreases in end-diastolic segment length and LV end-diastolic pressure because relaxation of cardiac muscle has been shown to be preload dependent. [47]In contrast to the findings with the lowest dose, the 50- and 100-mg [center dot] kg sup -1 [center dot] h sup -1 doses of propofol did not change tau. Further reductions in LV preload and decreases in afterload observed with higher doses of this intravenous anesthetic were balanced by negative inotropic effects. Despite the modest improvement in tau observed with the 25-mg [center dot] kg sup -1 [center dot] h sup -1 dose of propofol, indices of the rate and extent of early LV filling (e.g., dL/dt^Eand TVI-E) decreased. In addition, propofol caused declines in early LV filling fraction and the dL/dt-E/A and TVI-E/A ratios because dL/dt^Aand TVI-A remained unchanged (Figure 2and Figure 3). These findings suggest that early LV filling may be compromised and the contribution of atrial systole to total LV filling accentuated by propofol in dogs with cardiomyopathy. The rate and extent of cardiac muscle lengthening during rapid LV filling, however, are also directly affected by several factors, including LV loading conditions, the left atrial-LV pressure gradient, and inotropic state. [48]Venodilators are known to reduce dL/dt^Eand TVI-E by decreasing the left atrial-LV gradient. [48]Despite the finding that the left atrial-LV gradient was not specifically quantified in the current investigation, it appears to be likely that propofol-induced reductions in LV preload probably played an important role in the reduction in dL/dt^Eand TVI-E produced by this intravenous anesthetic.

The current results with propofol are somewhat different than those obtained with isoflurane in a similar experimental model of pacing-induced LV dysfunction. [49]Isoflurane (1.1 minimum alveolar concentration) increased TVI-E/A and did not reduce dL/dt-E/A, suggesting an improvement in the pattern of LV filling concomitant with reductions in chamber dimension and LV end-diastolic pressure. [49]The previous results with isoflurane were obtained in dogs after only 10 days of rapid LV pacing, an experimental preparation characterized by decreases in the contribution of early LV filling and reductions in or reversal of the classical E to A ratio. [27]The current investigation was conducted in dogs after almost 3 weeks of pacing, during experimental conditions that display normalization of the early LV filling rate as LV compliance decreases. Thus, the influence of propofol on indices of LV filling was examined in a model with a greater severity of LV dysfunction than those encountered in our previous study with volatile anesthetics. [49]Decreases in K produced by propofol were accompanied by declines in LV end-diastolic pressure and chamber dimension. These findings indicate that propofol-induced venodilation caused the LV pressure-segment length diagram to shift to the left, causing the left ventricle to operate in a more compliant region of the end-diastolic pressure-length relation in propofol-anesthetized compared with conscious, cardiomyopathic dogs. Interpretation of these results require qualification, however, because the pericardium remained open after the surgical implantation of instruments and during the development of pacing-induced cardiomyopathy.

We [3,12]have previously demonstrated that the doses of propofol used in the current investigation produce reliable anesthesia in chronically instrumented dogs. Although plasma concentrations of propofol were not measured in this investigation, a previous study [50]demonstrated that infusions of 20 and 40 mg [center dot] kg sup -1 [center dot] h sup -1 produced plasma concentrations of 2–13 micro gram/ml in dogs. These plasma concentrations lie within the anesthetic range in humans. Thus, the 25- and 50-mg [center dot] kg sup -1 [center dot] h sup -1 infusions of propofol used in the current investigation may correlate with clinically relevant propofol concentrations. Direct comparison of the cardiovascular effects of this intravenous anesthetic between dogs with pacing-induced cardiomyopathy and humans with heart failure should be made with caution, however.

In summary, the current results indicate that propofol produces favorable reductions in LV preload, afterload, and K and preserves LV relaxation but also causes negative inotropic effects and impairs LV filling dynamics in dogs with dilated cardiomyopathy by chronic rapid LV pacing.

The authors thank Dave Schwabe, John Tessmer, and Rich Rys for technical assistance.