Comparison of 3 Methods to Induce Acute Pulmonary Hypertension in Pigs

Abstract

Large animal models for acute pulmonary hypertension (PHT) show distinct differences between species and underlying mechanisms. Two embolic procedures and continuous infusion of a stable thromboxane A₂ analogue (U46619) were explored for their ability to induce PHT and their effects on right ventricular function and pulmonary and systemic circulation in 9 pigs. Injection of small (100 to 200 µm) or large (355 to 425 µm) polystyrene beads and incremental dosage (0.2 to 0.8 µg kg⁻¹ min⁻¹) of U46619 all induced PHT. However, infusion of U46619 resulted in stable PHT, whereas that after bead injection demonstrated a gradual continuous decline in pressure. This instability was most pronounced with small beads, due to right ventricular failure and consecutive circulatory collapse. Furthermore, cardiac output decreased during U46619 infusion but increased after embolization with no relevant differences in systemic pressure. This result was likely due to the more pronounced effect of U46619 on pulmonary resistance and impedance in combination with limited effects on pulmonary gas exchange. Coronary autoregulation and adaption of contractility to afterload increase was not impaired by U46619. All parameters returned to baseline values after infusion was discontinued. Continuous infusion of a thromboxane A2 analogue is an excellent method for induction of stable, acute PHT in large animal hemodynamic studies.

Systematic investigation of the pathophysiology of acute pulmonary hypertension (PHT), especially adaption of the right ventricular function in response to increased afterload, requires valid animal models with conclusions that are transferable to humans. In addition, the availability of such models would promote the evaluation of treatment options for pulmonary vasodilatation and inotropic support of the right ventricle. The various models reported in the literature can be classified by animal size, developmental period, and techniques. Due to cardiac dimensions and basic regulatory principles, sophisticated and transferable hemodynamic measurements require large animals such as dogs, pigs, and goats, and differences in vasoconstrictory responsiveness and adaption to hypoxia between these species and humans must be taken into account.^9,16,21 Chronic models of PHT in large animals are used less frequently than acute models and typically are induced through injection of monocrotaline pyrrole,⁶ surgical creation of an aortopulmonary shunt,²² or pulmonary banding.⁵ Techniques for the induction of acute PHT can be weighed in light of their underlying mechanisms, side effects, stability, and reversibility. Exposure to hypoxia ⁶ and repeated embolism^{9,17, 23} are used more frequently than are constriction of the pulmonary artery or infusion of the stable thromboxane A₂ analogue U46619.⁵ Whereas hypoxia mediates vasoconstriction by means of endothelin 1, serotonin, and the inhibition of voltage-gated potassium channels in smooth muscle cells,⁷ embolic procedures reduce the vascular cross-sectional area and increase concentrations of thromboxane A2.¹⁹ The size of injected particles positively correlates with the degree of hypoxia⁹ and inversely correlates with induction of thromboxane A2 production, thereby resulting in PHT and circulatory collapse.¹⁹ These mechanisms influence the stability of PHT, cardiac function, and sympathetic tone as a consequence of hypoxia. A leading advantage of transient occlusion, constriction of the pulmonary artery, and infusion of U46619 is that the resulting PHT is reversible. Compared with embolic procedures, proximal occlusion of the pulmonary artery induced different grade of afterload increase for the right ventricle, whereas U46619 may have systemic and coronary vasoconstrictory effects, thus causing negative inotropy.¹¹ The design of a study involving a PHT model therefore is influenced not only by the animal and technique selected but also by the underlying mechanisms of the technique and the sensitivity of the resulting PHT to drug intervention.

To study the effects of volatile anesthetics on right ventricular function during acute PHT, we aimed to develop a large animal model with stable increased afterload over several hours and minimal direct effects on cardiac function. We tested embolization techniques with different sizes of microbeads and the infusion of U46619. We favored pigs over dogs and goats because of the thickness of the arteriolar vascular muscle layer and the degree of collateral ventilation, which thus make the sensitivity of the pulmonary vasculature of swine more representative of that in humans.

Materials and Methods

All experimental procedures and protocols were reviewed and approved by the local animal care committee as well as the governmental animal care office (No. 50.203.2-AC 38, 4/05; Landesamt für Natur, Umwelt, und Verbraucherschutz Nordrhein-Westfalen, Recklinghausen, Germany) and are in accordance with The Guide for the Care and Use of Laboratory Animals.¹³ Nine farmbred German Landrace pigs, weighing 31.4 ± 2.2 kg, were examined by a veterinarian on arrival at our facility. This examination included behavior, motor ability, skin, hair, cardiopulmonary auscultation. None of the animals showed any symptoms of disease or elevated temperature. After an acclimation period of 5 d and overnight fasting, they received an intramuscular premedication of 4 mg kg⁻¹ azaperone. Animals were anesthetized with an intravenous injection of 3 mg kg⁻¹ propofol, tracheally intubated, and maintained under anesthesia by continuous infusion of 15 mg kg⁻¹ h⁻¹ thiopental, which did not induce cardioprotective actions, without muscle relaxants. Adequate anesthesia was accompanied by lack of withdrawal reactions and stable heart rate and blood pressure during surgical stimulation; if needed, a bolus of 50 mg thiopental was given. The lungs were ventilated with 21% O₂ by using a tidal volume of 10 mL kg⁻¹, an end-expiratory pressure of 7 cm H₂O, and a rate of 20 to 24 min⁻¹ as needed to keep the end-tidal partial CO₂ tension between 36 and 42 mm Hg (PhysioFlex, Draeger, Luebeck, Germany). Basic monitoring (S/5, Datex–Ohmeda, Helsinki, Finland) included electrocardiography, pulse oximetry, and invasive femoral arterial pressure measurement.

Surgical instrumentation.

After surgical dissection of the right cervical vessels, a central venous line was placed through the external jugular vein, and an 8-French introducer sheet was placed in the common carotid artery. Median sternotomy and longitudinal pericardiotomy were performed in each pig to place a perivascular transit-time–flow probe (MA20PAX, Transonic Systems Europe, Maastricht, The Netherlands) around the pulmonary artery; the probe was connected to a flow meter (T402PB, Transonic) to measure cardiac output). A 6-French catheter with a solid-state pressure sensor (CA61000PL, CD Leycom, Zoetermeer, The Netherlands) was placed through a stab wound into the right ventricular outflow tract, such that the sensor was in the main pulmonary artery, 3 to 4 cm distal to the pulmonary valve; the sensor was connected to a pressure interface (Sentron, CD Leycom). A 7-French multisegment pressure–volume conductance catheter with 12 electrodes, 10-mm spacing, and a pressure sensor between the fifth and sixth electrodes (SPR570-7, Millar Instruments, Houston, TX) was placed via a small incision into the right ventricular apex along the long axis, with the tip of the catheter in the right ventricular outflow tract, just below the pulmonary valve. The volume arm of the catheter was connected to a signal processor for dual-field measurement (Sigma 5DF, CD Leycom), and the pressure sensor was connected to an electronic pressure interface (PCU2000, Millar Instruments). In U46619-treated animals, the right coronary artery was dissected 2 cm behind its origin for placement of an additional flow probe (MA 2.5 PSB, Transonic).

A 7-French balloon catheter was placed in the inferior vena cava through the right femoral vein for short-term preload reduction. After instrumentation, animals were covered with a circulating-air warming blanket to maintain normothermia (38.5 °C) and allowed to recover from surgery for 120 min.

Data acquisition and analysis.

All signals from conductance and pressure catheters and flow probes were recorded after digitization at 500 Hz and stored (Conduct 2000, CD Leycom) for offline analysis with custom software for conductance data (Circlab 2004, Paul Steendijk, Leiden University Medical Center, Leiden, The Netherlands) and impedance calculations (LabView 8.5, National Instruments, Austin, TX). Mean values of parameters were calculated over 20-s intervals recorded during apnea and maintenance of positive end-expiratory airway pressure.

Heart rate was calculated from the electrocardiogram, the mean pressure values were calculated from signals for systemic arterial and pulmonary pressure, and the mean flow values were calculated from signals for cardiac output and coronary blood flow over time. The volume derived from conductance catheter data was calibrated at each measurement time point by blood conductivity, parallel conductance as determined by injection of hypertonic saline and the ratio α between the stroke volume derived from conductance catheter data and pulmonary flow probe.^4,20

End-diastolic volume was defined at the time point of the R wave in the electrocardiogram, and end-systolic volume was defined at the point of the maximal ratio of end-systolic pressure and volume.¹⁵ Right ventricular function was calculated from the right ventricular ejection fraction and the maximal first derivative of right ventricular pressure. The preload recruitable stroke work, a load-independent parameter, was determined as the slope of the linear relation between stroke work and end-diastolic volume during gradual temporary preload reduction.¹² This loading intervention also was used to describe ventricular afterload by the slope and intercept of composite flow–pressure curves. The mean systemic arterial pressure was calculated from individual regressions at cardiac output values of 1.5, 2.5, and 3.5 L min⁻¹. Pulmonary vascular impedance was calculated from the Fourier series expressions for pressure and flow. The total resistance and characteristic impedance were derived from the average of moduli between 2 and 15 Hz.⁴ Characteristic impedance was determined by the ratio between inertance and compliance of the proximal pulmonary tree,¹⁷ whereas total resistance was influenced primarily by peripheral vasoconstriction. Total hydraulic power was calculated as the integral of the pressure–flow product over time. The proportionate amount of oscillatory power was calculated as difference between steady power (the product of the mean pulmonary arterial pressure and cardiac output) and total work divided by total work. An increase in oscillatory power indicates a reduction in mechanical efficiency, due to wastage of power in generating oscillations rather than forward flow.⁴ Arterial partial O₂ and CO₂ tensions were measured (ABL500, Radiometer, Copenhagen, Denmark) simultaneously.

Study protocol.

After baseline hemodynamic parameters were acquire (t = 0 min), PHT was induced in 9 pigs (3 per group) within 30 min by means of repeated intravenous injection of small (100 to 200 µm; Microparticles GmbH, Berlin, Germany) or large (355 to 425 µm; Polysciences, Warrington, PA) polystyrene beads every 5 to 10 min or by increasing dosages (0.2, 0.4, 0.5, 0.6, 0.7, 0.8 µg kg⁻¹ min⁻¹) of continuous infusion of the stable thromboxane A₂ analogue U46619 (1 µg mL⁻¹; Cayman Chemical, Ann Arbor, MI) to achieve a mean pulmonary arterial pressure of 40 mm Hg. Subsequent hemodynamic data were acquired every 45 min for 3 h (t = 45, 90, 135, 180 min). Data regarding recovery from PHT induced by U46619 were acquired at 45 min after discontinuing the infusion.

Statistical analysis.

Data (mean ± 1 SD) are presented as plots (Prism 5.01, GraphPad Software, La Jolla, CA). Linear regressions were used to correlate coronary blood flow and heart rate with mean pulmonary arterial pressure by calculating Pearson coefficients for U46619-treated animals; P values less than 0.05 were considered statistically significant.

Results

Intravenous injection of small (100- to 200-µm) beads led to right ventricular failure in our pigs, which required cardiac resuscitation in 2 of 3 cases without reestablishment of a stable circulation. No severe problems occurred in the pigs in which large (355- to 425-µm) beads or U46619 were used. To reach the target mean pulmonary arterial pressure (40 mm Hg), 1 to 2 g of polystyrene beads or 0.6 to 0.8 µg kg⁻¹ min⁻¹ U46619 were needed. Whereas U46619 infusion produced a stable PHT without development of tachyphylaxia for 180 min, each bolus injection of beads progressively decreased the mean pulmonary arterial pressure (Figure 1). This decrease was faster after injection of small beads (a decrease of 10 mm Hg within 20 min) than after large beads (a decrease of 10 mm Hg within 90 min; Figures 1 and 2 A).

Figure 1.

Representative initial time course of mean pulmonary arterial pressure after injection of small (100 to 200 μm) beads and during increasing dosages of U46619.

Figure 2.

Time course of hemodynamic variables during PHT initiated during the first 30 min. Recovery after U46619 infusion (rec) was measured for 45 min after discontinuation of the drug infusion.

All 3 interventions led to an increase in heart rate as a compensatory mechanism, without remarkable differences between groups (Figure 2 B). Cardiac output increased after injection of beads (both sizes) but decreased during infusion of U46619 (Figure 2 C). Mean systemic arterial pressure remained stable within the first 90 min of PHT and decreased thereafter only after injection of large beads (Figure 2 D). End-diastolic volume decreased in the U46619 group and showed a transient increase within both other groups (Figure 2 E). In all groups, the maximal rate of right ventricular pressure rise increased continuously over time during PHT induction, with the highest values in the large-bead group (Figure 2 F), but the right ventricular ejection fraction decreased only during U46619 infusion (Figure 2 G). In contrast, the load-independent contractility index increased in all groups (Figure 2 H).

Compared with baseline values obtained before induction of PHT, injection of beads (both sizes) increased the slopes of pressure–flow curves without affecting the intercept at zero flow after 90 min of PHT. A steeper slope in association with an upward shift occurred in the U46619-treated animals (Figure 3). This result correlated with a clearly higher increase in total resistance after U46619 infusion compared with bead injection (Figure 4 A). The increase in characteristic impedance reached a comparable level in all groups after 180 min of PHT; only the initial increase after 45 min demonstrated differences, with the highest values in the U46619 group (Figure 4 B). Total work increased over time, with initial peaks in both bead-injection groups, whereas the proportion of oscillatory work decreased in the U46619 group (Figure 4D). All described parameters returned to baseline values in the U46619-treated animals after drug infusion was discontinued.

Figure 3.

Pressure–flow plots obtained during rapid preload reduction. Curves were characterized by slope and pressure intercept at zero flow. Data are displayed as mean ± 1 SD after regression analysis of results from individual animals at 90 min after induction of PHT.

Figure 4.

Time course of (A, B) impedance and (C, D) right ventricular power during pulmonary hypertension. Recovery after U46619 infusion (rec) was measured for 45 min after discontinuation of the drug infusion.

The 3 methods of inducing PHT in pigs influenced pulmonary gas exchange in different ways. The largest drop in oxygenation with the largest increase in CO₂ tension occurred after injection of large beads. Small beads showed initially the same effect, but values moved toward baseline over time. Infusion of U46619 did not influence CO₂ tension but impaired oxygenation (Figure 5 A and B). Ventilatory parameters such as tidal volume and airway pressure did not vary over time or between groups. In the U46619 group, coronary blood flow increased with increased heart rate and mean pulmonary arterial pressure, as demonstrated by their significant (P < 0.05) correlation (Figure 6 A and B).

Figure 5.

Time course of (A) oxygenation and (B) carbon dioxide tension during pulmonary hypertension. Recovery after U46619 infusion (rec) was measured for 45 min after discontinuation of the drug infusion.

Figure 6.

Correlation between coronary blood flow and (A) heart rate and (B) mean pulmonary arterial pressure. Values were obtained during stepwise increase of U46619 dosage.

Discussion

In general, the present study demonstrated that pulmonary embolism with microbeads did not produce stable PHT in pigs and was difficult to manage, whereas U46619 infusion led to well-controlled stable PHT. Particle size as well as U46619 infusion at a mean dosage of 0.6 to 0.8 µg kg min⁻¹ will produce differential effects on the pulmonary and systemic circulation and thus provoke variable right ventricular adaptive mechanisms. Microbeads, particularly the smaller (100 to 200 µm) ones, were associated with hemodynamic instability and spontaneous recovery. This outcome might be explained by an increase in endogenous production of thromboxane A₂ in early stages¹⁹ and activation of neurohumoral reflexes.²³

U46619 infusion produced the largest increase in peripheral pulmonary resistance and greatest decrease in proximal compliance, as demonstrated by total resistance, characteristic impedance, and pressure–flow plots. The effects of the large (325 to 455 µm) beads seemed to be more pronounced than those of the small beads, thus confirming the greater relevance of remote reflexes in small-particle embolism. A previous study¹⁴ demonstrated that vasoconstriction could be counterbalanced by efferent sympathetic pathways.

As an adaption to afterload elevation, heart rate and contractility (maximal slope of ventricular pressure increase, slope of preload recruitable stroke work) increased during PHT. The decreased right ventricular ejection fractions in the U46619 group might be explained by the pronounced afterload dependency of this parameter. As reflected by the higher cardiac output values, the initial increase in total energy was more distinct after embolism than during U46619 infusion. Although pulsatility (characteristic impedance) was higher in the U46619 group, less energy was wasted during pulmonary circulation as demonstrated by the proportion of oscillatory work. This finding confirms previous suggestions that reduced compliance and increased characteristic impedance improve coupling between the right ventricle and hypertensive pulmonary circulation.¹⁷

The differences in the response of cardiac output to PHT might simply reflect differences in systemic vascular resistance. Whereas U46619 induces systemic vasoconstriction, pulmonary embolism presumably stimulated the release of vasodilatory agents (prostacyclin, nitric oxide, atrial natriuretic peptide), which might attenuate this effect¹⁹ as well as evoke respiratory acidosis, another mechanism for systemic vasodilation.¹

Relevant differences in sympathetic tone as a cause for the different responses to PHT among groups likely would be indicated by differences in heart rates. Therefore, sympathetic activation by hypoxia was not considered relevant in this setting and might be counterbalanced by chemoreceptor stimulatory effects.⁴ However, pO₂ values below 60 mm Hg, as obtained during hypoxia induced experimental PHT and sometimes during pulmonary embolism (lowest value in this study, 53 mm Hg), increase myocardial inotropy, perfusion, and oxygen consumption, which effects are relevant in the design and execution of hemodynamic studies.¹⁸

Coronary perfusion pressure and vasoconstrictor agents might also have important influences on hemodynamic effects during PHT. Coronary autoregulation will maintain flow to coronary perfusion pressures of 40 mm Hg, before both flow and subsequent contractility drop.⁸ In the current study, coronary perfusion pressures did not drop below 40 mm Hg (data not shown) and negative inotropic effects did not occurred. Even during U46619 infusion, coronary regulation was not impaired (Figure 6).

Temporary acute animal models of PHT enable the description of direct cardiac effects of anesthetic drugs, vasodilators, and inotropic agents on functional reserve, compensatory mechanisms, and their protective or adverse effects. Anesthetics, like xenon, that do not influence myocardial contractility, might be superior to isoflurane or propofol in this regard. These findings will be relevant for the narcotic regime and treatment of thromboxane-A₂–mediated PHT, as a consequence of respiratory distress syndrome, sepsis, protamine infusion, and use of bone cement in orthopedic surgery.¹⁹ With the current model, the direct effects on the non adapted right ventricle could be studied, whereas the pig represents the best similarity with humans. The pulmonary vascular responsiveness regulability, coronary anatomy and perfusion characteristics, and absence of collateral ventilation in pigs favor them as a PHT model system as compared with dogs, sheep, and goats.^16,21 In addition, maintenance of anesthesia with barbiturates allows the investigation of cardioprotective properties of drugs during PHT, as described for myocardial ischemia reperfusion injury.¹⁰

During experimental PHT, numerous factors influence various hemodynamic responses, with measurable differences in the pulmonary circulatory system and right ventricular adaption. These differences could be demonstrated between species, triggers of PHT and their underlying pathophysiologic mechanisms. For hemodynamic studies, the stability and controllability of induced PHT are important features. Despite the small experimental population in the current study, the potential advantage of U46199 for induction of PHT was demonstrated by its feasibility, lack of direct myocardial effects, and reversibility.