Isoflurane and Desflurane Impair Right Ventricular–Pulmonary Arterial Coupling in Dogs

All experiments were approved by the Animal Ethics Committee of the Brussels Free University School of Medicine (Brussels, Belgium) and were done in accordance with the Guide of the Care and Use of Laboratory Animals  of the National Academy of Science.

The study included 14 mongrel dogs (mean weight, 25 kg) premedicated with 20 mg/kg intramuscular ketamine, anesthetized with 0.15 mg/kg intravenous midazolam and 0.5 μg/kg intravenous sufentanil followed by 0.15 mg · kg⁻¹· h⁻¹midazolam and 2 μg · kg⁻¹· h⁻¹sufentanil, and paralyzed with 0.2 mg/kg and 0.2 mg · kg⁻¹· h⁻¹intravenous pancuronium bromide. Additional sufentanil was given occasionally as to reverse increases of heart rate during thoracotomy, mainly at the time of rib spreader insertion. The dogs were ventilated with a fraction of inspired oxygen (Fio²) of 0.40, a PEEP of 5 cm H²O, and a tidal volume of 20 ml/kg. Other details of our preparation have been published previously.15,16A thermistor-tipped catheter (131H 7-French; Baxter Edwards, Irvine, CA) was inserted into a branch of the pulmonary artery, and a balloon catheter (Percor Stat-DL 10.5-French; Datascope, Paramus, NJ) was advanced into the inferior vena cava to produce a titratable decrease in cardiac output by reducing venous return. Temperature was maintained between 36° and 38°C with an electrical heating pad. Thrombus formation was prevented by 100 U/kg intravenous heparin given before insertion of the balloon catheter. A left lateral thoracotomy was performed, and a 16- to 20-mm nonconstricting ultrasonic flow probe (T206; Transonic Systems, Ithaca, NY) was positioned around the main pulmonary artery for measurement of instantaneous flow. A 5F manometer-tipped catheter (SPC 350; Millar Instruments, Houston, TX) was introduced through the right ventricle into the pulmonary artery, and its tip was positioned just distal to the flow probe. A second 5-French manometer-tipped catheter was positioned in the right ventricle. The chest was closed, pleural air was evacuated, and the lungs were expanded with several large insufflations.

Dogs were randomly assigned to receive isoflurane or desflurane that was administered at 0.5, 1.0, and 1.5 MAC with separate vaporizers. Inspired and expired anesthetic concentrations were measured by infrared spectrophotometry (Ultima II; Datex, Helsinki, Finland). After 30 min of postoperative stabilization, one group (n = 7) received isoflurane at target end-tidal concentrations of 0.7, 1.4, and 2.1%17for periods of 60 min. The other group (n = 7) received desflurane at target end-tidal concentrations of 3.6, 7.2, and 10.8%18for periods of 60 min. Actual end-tidal concentrations (mean ± SD) were 0.70 ± 0.05, 1.49 ± 0.09, and 2.13 ± 0.07% for isoflurane and 3.58 ± 0.12, 7.14 ± 0.09, and 10.78 ± 0.16% for desflurane. Different concentrations were delivered in random sequence. At each concentration, the animals were ventilated for 30 min in hyperoxia (Fio², 0.40) and for 30 min in hypoxia (Fio², 0.10), in a random sequence. Flow and pressures were collected at the end of each 30 min period, at steady state for calculations of impedance and RV–PA coupling, and during a brief flow reduction for determination of flow–pressure relations. At the end of the study, animals were killed by intravenous potassium chloride during deep anesthesia.

The instantaneous pressure and flow signals were sampled at 200 Hz and stored on a personal computer for off-line analysis. PA resistance was assessed by flow–pressure relations obtained by rapid flow reduction.15The relations looked rectilinear and were analyzed by linear regression. PA pressure values were interpolated at flows of 2 and 4 l · min⁻¹· m⁻²from individual regressions and were averaged to obtain composite flow–pressure plots. PA impedance was calculated from Fourier series expressions of pressure and flow.15Total resistance or 0 Hz impedance (Zo) and characteristic impedance (Zc) were derived from the impedance spectra. Zc was computed as the average of moduli between 2 and 15 Hz. PA compliance was calculated by the pulse pressure method.19Total work was computed as the integration of the instantaneous flow–pressure product over time, steady state work was computed as the mean flow by mean pressure product, and oscillatory work was computed as the difference between total and steady state work. RV contractility and RV–PA coupling were assessed from steady state RV pressure–volume curves using our single-beat method.16RV end-systolic elastance (Ees) was computed as the slope of the end-systolic pressure–volume line, PA effective elastance (Ea) was computed as the absolute slope of the end-systolic to end-diastolic line, and ventriculoarterial coupling efficiency was computed as the Ees:Ea ratio.16 

Data are expressed as mean ± SD. Results were first submitted to analysis of variance, allowing identification of the effects of isoflurane versus  desflurane and hyperoxia versus  hypoxia. The anesthetic dose–response relations were submitted to trend analyses, corrected by the Sidak-Holm procedure. Two-tailed P  values less than 0.05 were accepted as significant.

Isoflurane and desflurane caused similar dose-related decreases in cardiac output, systemic arterial pressure, and PA pressure (table 1). Flow–pressure curves were shifted to higher pressures (fig. 1). Impedance showed an increase in Zo and in Zc (table 2and fig. 2). Both anesthetics decreased steady state and oscillatory work and decreased the ratio of oscillatory to total work. Isoflurane and desflurane progressively increased Ea from approximately 0.85 to 1.45 mmHg/ml, decreased Ees from approximately 1.10 to 0.70 mmHg/ml, and markedly decreased Ees:Ea from approximately 1.50 to 0.50 (all P < 0.05).

Acute hypoxia decreased arterial oxygen tension and increased heart rate, cardiac output, and PA pressure (table 1). Flow–pressure curves were shifted to higher pressures (fig. 1). Hypoxia increased Zo and did not affect Zc (table 2and fig. 2). Hypoxia increased steady state and oscillatory work and decreased the ratio of oscillatory to total work. Hypoxia markedly increased Ea from 0.85 to 1.55 mmHg/ml, increased Ees from approximately 1.10 to 1.60 mmHg/ml, and decreased the Ees:Ea ratio from approximately 1.50 to 1.10 (table 2).

During hypoxia, isoflurane and desflurane again caused similar decreases in cardiac output, systemic arterial pressure, and PA pressure (table 1). Flow–pressure curves were shifted to higher pressures, but less than in hyperoxia (fig. 1). Impedance spectra showed an increase in Zo but no change in Zc (table 2and fig. 2). Both anesthetics decreased steady state and oscillatory work, with the ratio of oscillatory to total work decreasing nonsignificantly. Isoflurane and desflurane increased Ea nonsignificantly from approximately 1.55 to 1.90 mmHg/ml, decreased Ees from approximately 1.60 to 1.00 mmHg/ml, and markedly decreased Ees:Ea from approximately 1.10 to 0.50 (P < 0.05)

The current results show that isoflurane and desflurane exert similar and dose-related effects on PA hemodynamics, RV function, and RV–PA coupling efficiency. The results were obtained in dogs, after thoracotomy and during background anesthesia with midazolam and sufentanil, and therefore should be extended only with caution to other species or conditions. The number of experiments was limited and possibly insufficient to detect additional more subtle effects of isoflurane or desflurane.

Pulmonary vascular mechanics are commonly described by PA pressure and calculated pulmonary vascular resistance, but these variables are flow dependent and do not take into account the pulsatile nature of circulation. Flow–pressure relations better discriminate between passive (flow-induced) changes and active (tone-induced) changes in PA pressure and permit detection of more subtle changes in vascular tone.14,20Flow–pressure curves here were generated with transient flow reductions (< 10 s) to prevent cardiovascular responses due to sympathetic activation.15Pulmonary vascular impedance allows for the quantification of the three components that oppose blood ejection from the ventricle into the artery, i.e. , resistance due to small distal vessels, elastance due to large proximal vessels, and reflected waves that increase pressure and decrease forward flow.21In previous studies, isoflurane was found to cause some increase or no change in pulmonary vascular tone.10,17,20,22Heerdt et al.  11reported that PA resistance and elastance were increased with isoflurane, in contrast with aortic resistance that decreased and aortic elastance that remained unchanged. In the current study, isoflurane and desflurane reduced PA pressure. The reduction was clearly flow related, because flow–pressure curves showed an absence of change at lower doses and an increase in resistance at higher doses. These results again confirm the confounding effect of flow changes when interpreting changes in PA pressure or in pulmonary vascular resistance. Hydraulic work decreased together with flow and pressure, whereas oscillatory work (the proportion of work “wasted” into pulsatility) decreased at lower pressure. Impedance data mainly showed a dose-related increase in Zc. Zc may increase as a result of higher pressure, decreased diameter, or increased elastance.15,21Pressure and thus diameter decreased, so that the increase in Zc here may reflect the decrease in diameter or an increase in vessel elastance (higher stiffness). The unchanged compliance despite lower pressure suggests that vessel elastance did increase. Isoflurane and desflurane thus increased the afterload by increasing both the resistance and elastance of pulmonary vessels. The right ventricle can normally adapt to such increase in afterload and maintain cardiac output, so that the decrease in cardiac output observed with isoflurane and desflurane already suggests a concomitant decrease in RV contractility.

Right ventricular function in whole animals or humans is commonly assessed by RV ejection fraction or dP/dt, but both variables are critically dependent on preload and afterload.16,23,24Ventricular contraction is better investigated by ventricular pressure–volume curves and by end-systolic pressure–volume relations that are essentially load independent in physiologic ranges of preload and afterload.16,23In that approach, contractility is directly quantified by Ees, afterload by Ea, and ventriculoarterial coupling efficiency by the Ees:Ea ratio.23,25,26The ability of Ees and Ees:Ea to measure RV contractility and coupling efficiency, in normal conditions or during pulmonary hypertension and RV failure, has been shown in several studies.16,27–29Previous studies reported that halothane, enflurane, and isoflurane decreased RV contractility in a dose-related manner.7,9,11Isoflurane deteriorated RV function more than LV function because it decreased LV afterload but did not affect RV afterload.8,11In a previous study, we noted that dogs receiving propofol during pulmonary embolism tolerated the procedure, whereas those receiving isoflurane had RV failure and died.10In the current study, isoflurane and desflurane caused a marked and dose-dependent decrease in RV contractility. Unfortunately, the decrease in RV contractility occurred in addition to the above-mentioned increase in RV afterload. As a result of this combination, RV–PA coupling efficiency decreased by 65% to reach approximately 0.50 at the highest anesthetic dose. Ees:Ea values well below 1 denote an advanced degree of uncoupling and predict an inability to maintain flow.22Cardiac output actually decreased markedly, despite the normal baseline cardiovascular status of these experimental animals.

Hypoxia was associated with the expected pulmonary vasoconstriction. PA resistance increased, reflecting smooth muscle contraction in small distal vessels.20Zc did not change, indicating a balance between the opposite effects of higher pressure and higher diameter due to passive dilation of proximal arteries from distal vasoconstriction and possible effects of elastance changes.15RV afterload thus increased markedly. RV contractility also increased, as a result of adrenergic stimulation and of the Anrep effect.16Ees:Ea decreased moderately, and cardiac output was maintained. Previous studies have shown that the right ventricle can adapt to pulmonary hypertension by an increase in contractility (homeometric autoregulation) and maintain cardiac output.16,27,30,31The current study confirms that the decreased cardiac output observed with isoflurane and desflurane does not result from an isolated increase in RV afterload but also from a decrease in RV contractility.

Isoflurane and desflurane had different effects on the pulmonary circulation in hypoxia than in hyperoxia. Resistance increased somewhat less, suggesting a weakening of hypoxic pulmonary vasoconstriction (HPV). This change is consistent with previous reports that isoflurane preserves HPV at a lower dose but reduces HPV at a higher dose.17,20,22,32,33This reduction of HPV also explains why arterial carbon dioxide tension is lower with isoflurane or desflurane than at baseline. Elastance did not increase with isoflurane or desflurane in hypoxia, suggesting that anesthetics do not further affect elastance when proximal vessels are already dilated by HPV-induced hypertension and activated by hypoxia-induced adrenergic stimulation.15Isoflurane and desflurane had similar deleterious effects on Ea and Ees as in hyperoxia, resulting in a similar dramatic decrease in coupling efficiency.

We conclude (1) that isoflurane and desflurane have similar, dose-related effects on pulmonary hemodynamics and ventricular function in hyperoxic and hypoxic dogs and (2) that isoflurane and desflurane markedly impair RV–PA coupling efficiency, both by increasing RV afterload and by decreasing RV contractility.