Direct Coronary Vasomotor Effects of Sevoflurane and Desflurane in In Situ  Canine Hearts

This study was conducted with compliance of the institutional research committee. Experiments were performed on 21 mongrel dogs of either gender (weight range, 20.5–34.5 kg). Anesthesia was induced with an intravenous bolus injection of thiopental 20 mg/kg and maintained by continuous intravenous infusion of fentanyl and midazolam at rates of 12 μg · kg⁻¹· h⁻¹and 0.6 mg · kg⁻¹· h⁻¹, respectively. Additional intravenous bolus injections of fentanyl were given as necessary to maintain heart rate at approximately 130 beats/min. After tracheal intubation, the lungs were mechanically ventilated (Air Shields, Inc., Hatboro, PA), with fraction of inspired oxygen (FI^O2) equal to 1.0. The volume and rate of the ventilation were established to maintain arterial partial pressure of carbon dioxide (P^CO2) at a physiologic level (30–40 mmHg). Partial pressure of oxygen P^O2, P^CO2, and p  H of coronary arterial and venous blood samples (discussed in Experimental Measurements) were measured electrometrically (Blood Gas Analyzer model 413; Instrumentation Laboratories, Lexington, MS). Muscle paralysis was obtained using an intravenous injection of vecuronium bromide 0.1 mg/kg with supplements at 0.05 mg · kg⁻¹· h⁻¹to facilitate mechanical ventilation. Body temperature was maintained at 38°C with a heating pad. Lactated Ringer’s solution was administered continuously intravenously at a rate of 5 ml · kg⁻¹· h⁻¹to compensate for fluid losses. Heparin 400 U/kg with supplementation was used for anticoagulation.

After a left-sided thoracotomy in the fourth intercostal space, the LAD was cannulated just distal to its first major diagonal branch and perfused via  an extracorporeal system, as described in detail previously. 15–17In brief, this system consisted of two pressurized reservoirs that served as alternate sources of blood for the LAD. The volatile anesthetic-free blood reservoir was supplied with blood withdrawn directly from the left femoral artery, and the volatile anesthetic-equilibrated blood reservoir was supplied with blood from the right femoral artery that was first pumped into a hollow-fiber oxygenator (Capriox 300 series; Terumo Corp., Tokyo, Japan). The oxygenator was supplied with a 95% O²–5% CO²gas mixture that passed through a calibrated vaporizer providing either sevoflurane or desflurane. Blood was recirculated for at least 15 min through the extracorporeal oxygenator to ensure complete equilibration at the desired anesthetic concentration.

The LAD perfusion line was equipped with a heat exchanger to maintain the temperature of coronary perfusate at 38°C, an ultrasonic transit-time flow transducer (Transonic System Inc., Ithaca, NY) to measure CBF, and a port for collecting samples of coronary perfusate. Coronary perfusion pressure (CPP) was measured through a small-diameter tube positioned at the orifice of the perfusion cannula. To avoid hypovolemia in the experimental animal, lactated Ringer’s solution was infused intravenously during priming of the perfusion system.

Measurements of aortic, left atrial, and left ventricular pressures; left ventricular dP/dt max; and heart rate were obtained using standard methods. 15A continuous record of these variables was obtained on a physiologic recorder (model 2800; Gould, Cleveland, OH).

Measurements of MV^O2were obtained in the LAD bed by applying the Fick principle. The anterior interventricular vein was cannulated in a retrograde direction to obtain blood samples. This vein has been shown to drain exclusively the LAD perfusion territory. 18The venous cannula was allowed to drain freely into a beaker to prevent venous stagnation and interstitial edema. The coronary venous blood was returned intermittently to the dog to maintain isovolemic conditions. At specified times during the study, 1-ml coronary arterial and venous blood samples were obtained for determination of the local arteriovenous oxygen difference. CBF was constant during the period of blood sampling, which satisfied the requirement of the Fick principle for steady state conditions. Hemoglobin concentration and percent hemoglobin saturation of these samples were measured with a CO-Oximeter (model 482, Instrumentation Laboratories) and used to calculate oxygen bound to hemoglobin, assuming an oxygen-carrying capacity for hemoglobin of 1.39 ml O²/g. The amount of oxygen dissolved in the blood was computed (oxygen dissolved = 0.003 ml O²· 100 ml blood⁻¹· mmHg⁻¹) and added to the bound component to calculate total oxygen content. MV^O2(in ml · min⁻¹· 100 g⁻¹) was calculated as the product of the coronary arteriovenous oxygen difference and CBF at the time that blood samples were taken. EO²(in percentage) was calculated by dividing arteriovenous oxygen difference by arterial oxygen content.

Measurements of SS, an index of regional myocardial contractile function, were obtained by sonomicrometry. 19A pair of ultrasonic crystals was implanted into the LAD-perfused myocardium to a depth approximating the subendocardium. Location and functionality of the crystals were verified by segmental lengthening during a brief (30-s) occlusion. The crystals were oriented so that they were parallel with the anticipated direction of myocardial fibers in the subendocardium. Changes in distance between the crystals were recorded from measurements of ultrasonic transit time between the crystals (Triton Technology, San Diego, CA). The end-diastolic and end-systolic lengths were identified by the beginning of a rapid increase in the left ventricular pressure just before isovolumetric contraction and the maximum rate of decrease of left ventricular systolic pressure (−dP/dt min), respectively. Percent SS was calculated using the following formula:

%SS =[(EDL − ESL)/EDL]× 100

where EDL = end-diastolic length; and ESL = end-systolic length.

The concentration of sevoflurane and desflurane in the coronary arterial blood was evaluated using a modification of the equilibration method described by Yamamura et al.  20Briefly, a 2-ml sample of blood was obtained from the LAD perfusion tubing using a airtight glass syringe and introduced into a 5-ml glass vial. The vial was placed in a constant-temperature chamber at 38°C for 30 min. After equilibration, 100 μl of the gas in the vial was introduced into a gas chromatograph (model 5890; Hewlett Packard, Wilmington, DE) equipped with a flame ionization detector, and the area under the curve was measured. Anesthetic concentration in blood was determined by means of a calibration curve derived from appropriate standards. The data are presented in both units of mg/100 ml and mM to facilitate comparisons with previous studies. 9,12 

The dogs in this series were divided randomly into two equal groups according to whether sevoflurane or desflurane was studied. After at least 45 min of recovery from surgical preparation, initial control measurements for CBF, MV^O2, and SS were obtained during perfusion from the volatile anesthetic-free blood reservoir with CPP at 80 mmHg. This level of CPP was maintained throughout the study. The LAD was then switched to the volatile anesthetic–equilibrated blood reservoir, and at the peak increase in CBF (approximately 3–5 min after switching reservoirs), measurements were repeated. Perfusion was then returned to the volatile anesthetic-free blood reservoir, and at least 20 min was allowed for recovery. This protocol, with each exposure to a volatile anesthetic immediately preceded by a control period, was followed as the vaporizer setting for sevoflurane was varied from 1.2, 2.4, and 4.8% and that for desflurane was varied from 3.6, 7.2, and 14.4% (equivalent to 0.5, 1.0, and 2.0 minimum alveolar concentration [MAC] for dogs). 21,22The order of exposure to the various volatile anesthetic concentrations was randomized. Adenosine was infused into the LAD perfusion tubing at 8 mg/min (1 ml/min) to assess the vasodilator reserve of each preparation. 19We showed in preliminary studies that infusion of the saline vehicle at this rate had no effect on CBF.

The changes in CBF by 1.0 MAC sevoflurane and desflurane (administered as described under series 1) were evaluated in the same dogs before, during, and after (recovery) an intracoronary infusion of glibenclamide (Sigma Chemical Co., St. Louis, MO) 100 μg/min. 11At least 30 min was allowed after discontinuation of glibenclamide administration before the recovery response to the anesthetic was assessed. Glibenclamide was dissolved with 10 mM NaOH under gentle heat and diluted to a concentration of 100 μg/ml in isotonic saline, which permitted intracoronary infusions at the rate of 1 ml/min. The protocol adopted for the glibenclamide infusions was based on our previous investigation, 11which indicated that these infusions markedly inhibited the coronary vasodilator responses to the K^ATPchannel-opening drug cromakalim while permitting a full recovery of these responses within 30 min of stopping glibenclamide, and that they had no effect on the coronary vasodilation by the K^ATPchannel-independent drugs sodium nitroprusside and acetylcholine or on systemic hemodynamic parameters. We also showed that intracoronary infusions of the vehicle for glibenclamide alone had no effect on CBF or related parameters. 11 

After the multiple administrations of sevoflurane (before, during, and after glibenclamide), it was removed from the oxygenator-supplied circuit by recirculating the blood for at least 20 min with the oxygenator provided with anesthetic-free gas. After complete washout of sevoflurane was confirmed by gas chromatography, desflurane was added to the oxygenator-supplied circuit, and its coronary effects were also evaluated in the absence and presence of glibenclamide. The order of the sevoflurane and desflurane trials was randomized. The interval between the two glibenclamide infusions in each preparation was at least 90 min. Adenosine, as described in series 1, was used to evaluate coronary vasodilator reserve. Series 2 did not include measurements of MV^O2or SS. Coronary arterial blood samples were obtained for measurements of sevoflurane and desflurane concentration during the administrations of these agents before and during (but not after) glibenclamide.

At the completion of each experiment, Evans blue dye was injected into the LAD perfusion tubing to delineate the LAD perfusion field. The heart was stopped with KCl and removed, and the dyed tissue was excised and weighed so that CBF could be expressed on a per-100-g basis. The average weight of the LAD perfusion field was 28 ± 8 g.

The Student t  test for paired samples was used to assess the difference of values during sevoflurane, desflurane, and adenosine administration relative to their respective control values. 23Hemodynamic responses were normalized on the basis of percentage change from control to facilitate comparisons at equivalent MAC for sevoflurane and desflurane. A two-way analysis of variance (ANOVA) for repeated measurements was used to make these comparisons. 23A one-way ANOVA was used to evaluate the effects of sevoflurane and desflurane before, during, and after administration of glibenclamide. 23,Post hoc  comparisons were made using the Student t  test with the Bonferroni correction. 23Control values for systemic hemodynamic parameters for the sevoflurane and desflurane groups were compared using the Student t  test for unpaired samples. 23Data are expressed as mean ± SD. All comparisons were two-sided. P < 0.05 was considered significant throughout the study.

Preanesthetic control values for systemic and cardiac parameters did not differ significantly. Therefore, for the sake of simplicity and brevity, the pooled means for all controls are presented in tables 1 and 2.

Figure 1is a representative tracing showing the effect of 2.0 MAC desflurane on CBF, SS, and systemic hemodynamic parameters. At “A,” perfusion of the LAD was switched to the desflurane-equilibrated reservoir, and at “B,” perfusion was returned to desflurane-free reservoir. After a short delay (corresponding to the time necessary for the desflurane-equilibrated blood to reach the vascular bed), desflurane caused an approximately threefold increase in CBF reflecting a proportional decrease in coronary vascular resistance, and it converted SS to segmental lengthening and reduced left ventricular dP/dt max. Return to the desflurane-free reservoir reversed these effects. Intracoronary desflurane did not affect aortic pressure or heart rate. This was a consistent finding throughout the study (table 1). There were no significant differences between the control values for systemic hemodynamic parameters in the sevoflurane and desflurane groups, with the exception that mean aortic pressure and left ventricular dP/dt max were approximately 15% higher in the sevoflurane group.

Table 2presents the concentration-related effects of sevoflurane and desflurane on CBF and related cardiac parameters. The control values for CBF, MV^O2, and SS were moderately higher in the sevoflurane group versus  the desflurane group; the control values for EO²in the two groups did not differ. Both anesthetics caused concentration-related increases in CBF. When viewed on the basis of percentage increase from the respective control value, the increases in CBF caused by equivalent MAC of sevoflurane and desflurane were comparable (fig. 2). Adenosine caused fivefold to sixfold increases in CBF in both groups (from 97 ± 31 to 509 ± 164 ml · min⁻¹· 100 g⁻¹in the sevoflurane group and 95 ± 23 to 711 ± 112 ml · min⁻¹· 100 g⁻¹in the desflurane group). The increases in CBF at the highest concentrations of sevoflurane (4.8%) and desflurane (14.4%) were 52 ± 24% and 57 ± 17% of the adenosine-induced responses, respectively.

Although 0.5 and 1.0 MAC sevoflurane did not affect MV^O2, 2.0 MAC sevoflurane reduced MV^O2. Desflurane did not alter MV^O2at any concentration. The combination of an increased CBF and either unchanged or decreased MV^O2-reduced EO²during administration of sevoflurane or desflurane caused concentration-dependent decreases in EO²(range equal to −20% to −75%). The decreases in EO²were similar for sevoflurane and desflurane at equivalent MAC.

Both sevoflurane and desflurane caused decreases in SS. At each MAC equivalent, the decreases in SS by sevoflurane were less than those by desflurane (0.5 MAC, −10 ± 11 vs. −106 ± 48%; 1.0 MAC, −28 ± 15%vs. −165 ± 57%; 2.0 MAC, −101 ± 62%vs. −250 ± 93%). The 0.5 and 1.0 MAC sevoflurane reduced SS, and 2.0 MAC sevoflurane completely abolished SS. Conversely, 0.5 MAC desflurane completely abolished SS, and 1.0 and 2.0 MAC desflurane converted SS to segmental lengthening.

Table 1shows that the coronary arterial concentrations for sevoflurane and desflurane varied directly with the percentage provided by the vaporizer.

Glibenclamide infusion itself caused approximately 30% reversible decreases in CBF; the control (preanesthetic) values for CBF were 96 ± 17, 71 ± 12, and 88 ± 15 ml · min⁻¹· 100 g⁻¹before, during, and after glibenclamide administration, respectively. Figure 3shows that glibenclamide blunted, in a reversible manner, the increases in CBF caused by 1.0 MAC sevoflurane and desflurane. The arterial blood concentrations for the anesthetics were similar before and during glibenclamide: Sevoflurane, 8.4 ± 1.2 and 8.1 ± 1.4 mg/100 ml (equivalent to 0.42 ± 0.06 and 0.40 ± 0.07 mM); desflurane, 31.1 ± 4.1 and 30.1 ± 5.7 mg/100 ml (equivalent to 1.85 ± 0.25 and 1.79 ± 0.34 mM). Control values for systemic hemodynamic parameters in series 2 were comparable to those in series 1 (table 1); the intracoronary administrations of the volatile anesthetics (and of glibenclamide) did not alter these values. Adenosine caused fourfold increases in CBF (104 ± 20 to 440 ± 44 ml · min⁻¹· 100 g⁻¹) in the dogs in series 2.

We have used this model of regional coronary perfusion in canine hearts to evaluate the direct coronary vascular effects of volatile anesthetics 15–17,24and to clarify the role of mechanisms potentially responsible for these effects, such as nitric oxide 25and the ATP-sensitive potassium channels. 11The baseline values for EO²in the cannulated LAD bed were moderately lower than those usually found in anesthetized dogs with an intact coronary circulation. 26This suggests modest vasodilation in the control preparation, probably as a result of dilators released from blood cells within the extracorporeal circuit. 27Nevertheless, this model shows pronounced vascular responsiveness to endothelium-dependent vasodilators (acetylcholine), nitric oxide donors (sodium nitroprusside), and K^ATPchannel openers (cromakalim). Moreover, vasodilator reserve is appreciable, as shown by the better than fourfold increases in CBF during adenosine infusion in the current study.

In previous studies using the regional coronary perfusion preparation, we demonstrated that the coronary vasodilative effects of volatile anesthetics were blunted when arterial blood concentration was increased gradually 24or when exposure of the coronary circulation to the anesthetic was prolonged. 25These findings implied a tendency for coronary vascular smooth muscle to adapt to the relaxing effects of these agents. Such vascular adaptation would, of course, complicate interpretation of studies in which multiple administrations of volatile anesthetics were compared in the same heart. To obviate this factor, we used only abrupt, relatively brief exposures of the LAD to blood previously equilibrated with various concentrations of sevoflurane and desflurane.

Infusion of crystalloid during priming of the extracorporeal perfusion system produced modest reductions in hematocrit. The animals were ventilated on 100% oxygen to increase arterial P^O2and thus minimize the concomitant decreases in arterial oxygen content. Previous studies have shown increases in coronary vascular resistance during hyperoxia, and that this response may be related to K^ATPchannel closure. 28Although we cannot rule out that hyperoxia limited the vasodilative effects of sevoflurane and desflurane, it is highly unlikely that this factor would have altered our conclusions. First, the values for coronary arterial P^O2were similar under control conditions and during administration of the anesthetics. Second, our previous study, 11conducted with arterial P^O2at similarly elevated levels, showed marked concentration-related coronary vasodilation in response to the K^ATP-channel opener cromakalim.

General anesthesia was necessary in this open-chest canine preparation. A balanced anesthetic technique using fentanyl and midazolam was used for several reasons. First, such techniques have been shown by clinical investigators to be free of significant effects on cardiac function. 29Second, Flacke et al.  30reported that the combination of fentanyl and diazepam (another benzodiazepine) caused no additional cardiac depression in dogs anesthetized with enflurane after elimination of cardiac sympathetic drive with a spinal block. Finally, Reves et al.  31found that excessive doses of fentanyl and diazepam together were necessary to depress isolated perfused rat hearts and that this effect occurred in a strictly additive fashion.

A CPP of 80 mmHg was used to perfuse the LAD bed throughout our study. Because the values for mean aortic pressure were, on average, modestly higher, the possibility of incoming collateral flow cannot be ruled out. However, our previous findings in the same canine model showed that collateral flow made only a negligible (approximately 3%) contribution to overall perfusion of LAD-dependent myocardium when an aorta–LAD pressure gradient of 40–50 mmHg was created, 32which suggests that this factor was not important under the conditions of the current study.

In our preparation, only the volatile anesthetic-equilibrated blood reservoir was supplied via  a circuit equipped with an oxygenator. Thus, it was necessary to rule out that the oxygenator itself and/or the recirculation protocol used to equilibrate the blood with the volatile anesthetics contributed to the observed increases in CBF and decreases in SS. This was accomplished in previous studies using two protocols. First, we demonstrated that blood, which was recirculated for 15 min through a membrane oxygenator not supplied with a volatile anesthetic, had no effects in the LAD bed. 33Second, we showed that turning off the vaporizer while maintaining perfusion from the oxygenator-supplied reservoir caused return of CBF to baseline. 15,17 

The ability of SS measurements to reflect changes in myocardial contractility is limited by variations in heart rate and in loading conditions of the heart. 34However, the constant values for heart rate and for indices of afterload (aortic pressure) and preload (left atrial pressure) during intracoronary administration of the anesthetic suggest that this methodologic limitation does not apply to the current study.

To facilitate clinically relevant comparisons, the coronary and myocardial effects of sevoflurane and desflurane were compared on the basis of corresponding MAC values. It should be kept in mind that the MAC value for sevoflurane (2.4%) is one third that of desflurane (7.2%); thus, these comparisons were not a reflection of the relative pharmacologic potency of these agents in the coronary circulation.

Both sevoflurane and desflurane caused concentration-dependent coronary vasodilation. When viewed on a MAC basis, these effects were similar for the two anesthetics. The extent of coronary vasodilation caused by the highest concentration of the sevoflurane or desflurane (4.8% and 14.4%, respectively, corresponding to 2.0 MAC) was impressive, being equivalent to approximately 50% of that achievable with adenosine. Because these agents either did not change or caused decreases in local MV^O2(reflecting a direct negative inotropic effect), EO²decreased markedly. These decreases in EO²indicated an uncoupling of coronary oxygen supply from myocardial oxygen demands, which is the hallmark of a coronary vasodilative drug. 35 

Mechanical forces in the left ventricular wall compress the coronary arteries during systole, causing a physical impediment to blood flow. 35This effect shows transmural variation (i.e. , subendocardium > subepicardium). Under normal conditions (intact vasomotor tone), metabolic vasodilative mechanisms operate during diastole to compensate for the systolic limitation to subendocardial flow, thus preserving a uniform distribution of flow across the ventricular wall over the cardiac cycle. Some, but not all, studies have shown that when coronary vasomotor tone is fixed with a maximally dilating infusion of adenosine and peak systolic left ventricular pressure is held constant, CBF varies inversely with the level of myocardial contractility. 35Thus, we cannot rule out the possibility that reduced extravascular compressive forces secondary to decreased or completely abolished regional contractile activity contributed to the increases in CBF caused by sevoflurane and desflurane. The role of this factor may be more prominent during desflurane administration because of its stronger negative inotropic effect.

We used the same canine model and experimental approaches to evaluate the effects of halothane, isoflurane, and enflurane in the coronary circulation. 15–17The lack of concurrence with the current study notwithstanding, these studies can provide insight into how the direct coronary vasodilative effects of sevoflurane and desflurane compare with those of the other widely used volatile anesthetics. The findings indicate that the coronary vasodilative effects of sevoflurane and desflurane are similar to those of halothane and enflurane but smaller than that of isoflurane. It is difficult to compare our findings to other in vivo  studies assessing the relative coronary vasodilative effects of the volatile anesthetics, 2–8because these studies administered the anesthetics in the inspired gas; thus, the resultant changes in CBF were the summation of several interacting factors, including the direct relaxing action of the anesthetics on the vascular smooth muscle tone, and the influence of indirect mechanisms, including metabolic mechanisms secondary to reduced cardiac work demand, pressure–flow autoregulation, and time-dependent vascular adaptation.

Glibenclamide inhibited profoundly the increases in CBF caused by sevoflurane and desflurane, implying a significant role for the K^ATPchannels in the responses. These findings are consistent with those obtained previously in studies of the coronary dilating effects of other volatile anesthetics (i.e. , isoflurane, enflurane, and halothane) in open-chest dogs 11and swine 13and in vitro  in porcine coronary arterioles 12and crystalloid-perfused rat hearts arrested with tetrodotoxin. 14Recent evidence suggests that K^ATPchannels are present in vascular smooth muscle cells and in vascular endothelial cells. 36Opening of K^ATPchannels in the vascular smooth muscle cells shifts the membrane potential closer to the K⁺reversal potential. Hyperpolarization then inhibits calcium entry through voltage-dependent calcium channels, leading to reduced vasomotor tone. Opening of the endothelial K^ATPchannels can reduce vascular smooth muscle tone either by production and release of second messengers (e.g. , nitric oxide) or by direct transduction of electrical potentials. The current findings do not distinguish between the contribution of the K^ATPchannels in vascular smooth muscle and in the endothelial cells to the volatile anesthetic-induced coronary vascular relaxation.

Several potential mechanisms can be proposed to explain K^ATPchannel–mediated coronary vasodilation by the volatile anesthetics.

• 1. The volatile anesthetics may interact directly with the K^ATPchannels.

• 2. A reduction in ATP concentration within the vascular smooth muscle cell may cause an opening of the K^ATPchannels.

• 3. Prostacyclin may be released (perhaps from the vascular endothelium), which in turns open the K^ATPchannels, perhaps via  a G protein–mediated pathway. This mechanism was suggested by studies indicating that glibenclamide inhibited coronary vasodilation caused by exogenous prostacyclin or iloprost (the stable analog of prostacyclin) in saline-perfused hearts. 37 

• 4. An adenosine receptor may be activated, leading to an opening of the K^ATPchannel via  a G protein. 38 

• 5. An interaction may occur with phosphorylating enzymes, such as protein kinase C, which in turn may regulate activity of the K^ATPchannel.

Further investigations are necessary to determine which, if any, of the these mechanisms are involved in the opening of the K^ATPchannels in the coronary circulation by the volatile anesthetics.

The ability of sevoflurane and desflurane (and of the volatile anesthetics in general) to cause cardiac depression is well established. 11,16,17,39–41The effect seems to be due to the combination of reduced Ca²⁺current through the sarcolemmal membrane and depletion of Ca²⁺from the sarcoplasmic reticulum. 42,43The current findings indicate that sevoflurane is a less potent negative inotrope than desflurane. Comparison of these results to those obtained previously in the same canine model shows that sevoflurane has a negative inotropic effect approximating that of halothane and isoflurane, whereas desflurane has negative inotropic effect approximating that of enflurane. 16,17 

The current findings contrast with results obtained in autonomically blocked, chronically instrumented dogs in which myocardial contractility was assessed using the slope of the regional preload recruitable stroke work relation (a relatively heart rate– and load-independent index), indicating that inspired sevoflurane and desflurane had equivalent negative intropic effects. 39–41We have no definitive explanation for this apparent discrepancy, but it is likely attributable to methodological differences, including those related to the index of contractility, route of administration of the volatile anesthetics, and the absence or presence of a background anesthetic.

We observed that 1.0 and 2.0 MAC desflurane caused systolic segmental lengthening, reflecting passive myocardial stretch in response to the rise in intraventricular pressure (fig. 1). This phenomenon is usually associated with regional myocardial ischemia, but it can occur when a negative inotrope administered directly into a branch of the left main coronary artery is potent enough to completely arrest segmental contractile function. 44The surrounding normal zones “pull” on the noncontracting myocardium during systole, causing it to lengthen. The ability of intracoronary desflurane to cause systolic lengthening in our model should not be construed as evidence that this same dose administered systemically would stop the heart from contracting. Under the latter condition, the entire heart would be exposed to the negative inotropic effect of desflurane; thus, there would be no tethering influence of normally contracting myocardium on depressed myocardium. Furthermore, mechanisms activated during the systemic administration of desflurane (e.g. , reduction in cardiac afterload, baroreceptor-mediated arousal of the sympathoadrenal system 1) would be expected to mitigate the direct negative inotropic effect of the anesthetic.

The major determinants of myocardial oxygen demand are myocardial contractility, wall tension, and heart rate. Because mean aortic pressure, mean left atrial pressure, and heart rate were constant, a decline in MV^O2might be expected during volatile anesthetic-induced reductions in local contractility. Why this did not occur is uncertain, but it may reflect an ability of the volatile anesthetics to directly reduce the efficiency of ATP production by the mitochondria. Further studies using sophisticated in vitro  techniques are needed to test this possibility. Another factor (especially during desflurane administration) may have been the paradoxically high level of MV^O2in the pharmacologically arrested bulging myocardial segments. 44 

The current study showed that abrupt intracoronary administrations of sevoflurane and desflurane caused concentration-dependent coronary vasodilation accompanied by a negative inotropic effect in in situ  canine hearts. An opening of the K^ATPchannels played a prominent role in the coronary vasodilation. The net effect of inspired sevoflurane or desflurane on CBF depends on to what extent their direct coronary vasodilative effects are counteracted by moderating factors, including time-dependent vasculature adaptation and metabolic mechanisms secondary to a reduced cardiac workload.

The authors thank Derrick L. Harris, B.S. for expert technical assistance.