Pulmonary and Left Ventricular Decompression by Artificial Pulmonary Valve Incompetence During Percutaneous Cardiopulmonary Bypass Support in Cardiac Arrest
Background In cardiac arrest, use of percutaneous cardiopulmonary bypass support (PCPS) may lead to left ventricular loading, with deleterious effects on the myocardium, and is often accompanied by an increase in pulmonary artery pressure. The present study was designed to assess the potential of artificially induced pulmonary valve incompetency to retrogradely decompress the left ventricle during PCPS in ventricular fibrillation.
Methods and Results Studies were performed using a standardized experimental animal model in sheep (n=12; body weight, 77 to 112 kg). When PCPS was used during fibrillation, an increase in left ventricular pressure (from 21.4±5.0 mm Hg after 1 minute to 28.4±9.5 mm Hg after 10 minutes of fibrillation) was observed in all animals, with a simultaneous increase in pulmonary artery pressure in 6 animals from 15.5± 3.8 to 24.3±5.4 mm Hg (group A). In these animals, artificial pulmonary valve incompetency, which was induced by a special “pulmonary valve spreading catheter,” led to effective decompression of both the pulmonary circulation (decrease in pulmonary artery pressure from 24.3 to 11.3 mm Hg) and the left ventricle (decrease in left ventricular pressure from 30.5 to 17.7 mm Hg). We simultaneously measured a decrease in the myocardial release of lactate (increase in arterial coronary-venous difference in lactate content from −0.01 to 0.14 mmol/L), demonstrating the myocardial protective effect of the procedure. In contrast, in 6 animals without an increase in pulmonary artery pressure during PCPS (group B), artificial pulmonary valve incompetency did not reduce left ventricular loading, which was probably because of competent mitral valves in these animals.
Conclusions In case of increasing pulmonary artery pressure during PCPS in cardiac arrest, artificial pulmonary valve incompetency might be a useful tool for effective pulmonary and retrograde left ventricular decompression.
In recent years, percutaneous insertion of a cardiopulmonary bypass support (PCPS) became feasible1 and was suggested for use in selected patients undergoing high-risk coronary angioplasty2 and in patients with cardiac arrest.3 However, using PCPS during experimental ventricular fibrillation, we regularly observed left ventricular pressure and, to a minor extent, volume loading, with deleterious effects to the myocardium.4 These data clearly demonstrated the need for active left ventricular decompression during PCPS in cardiac arrest.5 For use in interventional cardiology, nonsurgical methods of left ventricular venting need to be developed, which would have to be applicable even during external cardiocompression.
The left ventricular pressure rise during PCPS in ventricular fibrillation is frequently accompanied by an increase in pulmonary artery pressure.5 In this setting, artificially induced pulmonary valve incompetence might be useful to retrogradely decompress the left ventricle. We used a pulmonary valve spreading catheter in a standardized model of cardiac arrest and PCPS in sheep to assess the potential benefit of artificial pulmonary valve incompetence and to analyze its influence on myocardial metabolism.
Studies conformed to the guiding principles of the American Physiological Society.
We used a conventional open-chest preparation in a standardized experimental animal model4 5 6 in sheep (n=12; mean body weight, 97.4 kg; range, 77 to 112 kg). Anesthesia was achieved starting with thiopental 750 mg IV followed by continuous infusions of piritramide (1.45 mg · kg−1 · h−1), fentanyl hydrochloride (0.0037 mg · kg−1 · h−1), and midazolam (0.31 mg · kg−1 · h−1). Additionally, isoflurane was given as needed to maintain adequate anesthesia during surgical interventions such as thoracotomy. Artificial respiration was performed with a mixture of nitrous oxide and oxygen (70%:30%; Engström respirator). Arterial pH, Po2, and Pco2 were determined frequently to ensure adequacy of ventilation and a stable acid-base state (ABL 500, Radiometer).
Except for the pulmonary valve spreading catheter, all right and left ventricular catheters were placed by direct transmural puncture to avoid catheter-induced aortic or pulmonary valve regurgitation. The injection catheter for cardiac output measurements was placed into the right ventricle by cannulation of the free right ventricular wall. The thermistor for cardiac output measurements and the catheter-tip manometer for pressure recordings both were placed inside the left ventricle by use of a left ventricular apex cannula (21F OD). The stem of the pulmonary artery was cannulated 2 to 3 cm above the pulmonary valve to measure pulmonary artery pressures (Fig 1⇓).
Left ventricular pressures were measured with a catheter-tip manometer (PC-350, Millar Instruments). Aortic, central venous, and pulmonary artery pressures were measured via fluid-filled catheters (Statham P23 ID transducers, Gould).
Myocardial blood flow was measured by electromagnetic coronary sinus outflow measurements directing total coronary sinus blood to an external flow probe (501 D, Carolina Medical Electronics) with a Morawitz cannula.4 5 6 7 The hemiazygos vein, which in sheep normally drains into the coronary sinus, was blocked simultaneously by a Fogarty catheter placed inside the Morawitz cannula.4 6
Left ventricular volume was calculated as follows: To allow on-line wall motion measurements, two electromagnetic distance transducers were subepicardially sutured above the first branch of the left anterior descending coronary artery.6 From wall motion measurements at the left ventricular surface, left ventricular cavity volume (in milliliters) was approximated by postmortem calibration with a fluid-filled balloon.5 Via the left ventricular apex cannula, this balloon was positioned into the left ventricle. To avoid mismeasurements due to dislocation of parts of the balloon into the ascending aorta, the aortic valve was first closed with sutures. Assuming a spherical shape of the left ventricle, the corresponding wall stress (ς, dyn/cm2) was calculated from the left ventricular volume and the pressure registrations by the Laplace rule.5
The ECG (lead I, II, or III), pulmonary artery pressures, central venous pressures, myocardial blood flow, left ventricular wall motion measurements, aortic pressures, and left ventricular pressures were recorded continuously (Thermoprinter UD 2108, Rikadenki Electronics).
At baseline (without PCPS), cardiac output was measured by thermodilution technique (BN 6560, August Fischer KG) with the injection catheter placed in the right ventricle and the thermistor placed in the left ventricle.
Plasma electrolytes and concentrations of oxygen and lactate were determined in arterial, pulmonary arterial, central venous, and coronary venous blood.
Myocardial oxygen consumption (MV̇o2=MBF×ACVDo2, where ACVDo2 is arterial coronary-venous difference in oxygen content and MBF is myocardial blood flow) and arterial coronary-venous difference in lactate content were calculated.
The PCPS system, consisting of a centrifugal pump (Sarns 7850), a capillary membrane oxygenator (HF-5000, C.R. Bard Inc), a volume reservoir, and a heat exchanger, was connected to an 18F multihole venous suction catheter (C.R. Bard) and an 18F arterial perfusion catheter (C.R. Bard). The venous suction catheter was positioned near the right atrium via a jugular vein, and the arterial perfusion catheter was placed into the abdominal aorta via a femoral artery by use of a guide wire under fluoroscopic control. The PCPS system was primed with 1.5 L heparinized (5000 IU) electrolyte solution (Ringer’s lactate, B. Braun). PCPS flow rates (in L/min) were measured with a Doppler flow probe.
After control measurements during sinus rhythm, ventricular fibrillation was induced by electrical stimulation. Immediately after initiation of fibrillation, PCPS was started with maximum possible flow, and baseline measurements were performed after 1 minute (hemodynamics alone) and 10 minutes (hemodynamic measurements and collection of blood samples for metabolic measurements) of fibrillation.
Then artificial pulmonary valve insufficiency was induced with a catheter developed for this purpose (Figs 2⇓ and 3⇓). Parameters were remeasured during fibrillation with PCPS and pulmonary valve insufficiency after a hemodynamic “steady state” had been held for at least 10 minutes.
All data are expressed as the mean±SD. To assess statistical significance, an ANOVA for repeated measures was performed. Differences were considered statistically insignificant if P>.05.
PCPS During Ventricular Fibrillation
During fibrillation, PCPS reached maximum flow rates of 4.9±0.4 L/min, leading to mean aortic pressures of 49.7±6.9 mm Hg in the total group of 12 animals. Left ventricular pressures simultaneously rose from 21.4±5.0 mm Hg after 1 minute of fibrillation to 28.4±9.5 mm Hg after 10 minutes of fibrillation with PCPS (P<.05).
Left Ventricular Decompression by Artificial Pulmonary Valve Insufficiency
In 6 of 12 animals, the left ventricular pressure rise was accompanied by a simultaneous increase in pulmonary artery pressure from 15.5±3.8 to 24.3±5.4 mm Hg (group A; Table 1⇓) (this increase in pulmonary artery pressure was defined as at least 5 mm Hg). In these animals, artificial pulmonary valve insufficiency lowered the pulmonary artery pressure from 24.3±5.4 to 11.3±3.2 mm Hg (P=.0005) and led to a subsequent decrease in the left ventricular pressure from 30.5±4.6 to 17.7±4.2 mm Hg (P=.004) (Fig 4⇓), resulting in both a significant decrease in left ventricular wall stress (from 33 438±6608 to 18 149±4808 dyn/cm2; P=.001) and an increase in myocardial perfusion pressure (from 18.7±7.8 to 28.0±5.9 mm Hg; P=.042). The arterial coronary-venous difference in lactate content rose from −0.01±0.05 to 0.14±0.12 mmol/L, indicating a trend toward a reversal from myocardial release of lactate to lactate uptake during fibrillation (not significant).
In another 6 sheep, the increase in left ventricular pressures during fibrillation with PCPS (from 20.3±4.9 mm Hg after 1 minute of fibrillation to 26.2±11.4 mm Hg after 10 minutes of fibrillation) did not result in an increase in pulmonary artery pressures (group B; Table 2⇓). In these animals, artificially induced pulmonary valve incompetence did not influence left ventricular pressures, myocardial perfusion, and myocardial metabolism.
During cardiopulmonary bypass in open-heart surgery, venting of the fibrillating left ventricle has been established and regularly performed by transmural cannulation or via the aortic root during aortic cross-clamping. In contrast, during PCPS, the problem of performing effective left ventricular decompression has not been solved to date. In addition to left ventricular pressure loading with a consecutive decrease in myocardial perfusion pressures and subsequent damaging effects on the myocardium, we frequently observed an increase in pulmonary artery pressures during PCPS in experimental cardiac arrest.5 This was probably caused by simultaneous mitral valve incompetence, which sometimes may occur in the nonbeating heart. In this situation, pulmonary capillary pressure may increase to reach levels close to those of the systemic circulation, with the risk of severe pulmonary edema and consequent irreversible pulmonary damage.
Some authors previously reported that venting via the pulmonary artery appears to be an effective alternative method to decompress the left ventricle during cardiopulmonary bypass in cardiac surgery.8 9 10 In an experimental animal study, in a total of six lambs weighing 14.5 to 17 kg, Rossi et al11 used artificial pulmonary valve incompetency during PCPS to decompress the fibrillating left ventricle. In their experimental design, however, cardiac arrest had been induced with preexisting pulmonary valve insufficiency. Thus, the investigators could not make any control measurements during fibrillation before the induction of pulmonary valve incompetency. In addition, they did not measure left ventricular pressures at all. Thus, no reliable data concerning direct hemodynamic effects and no measurements of metabolic effects of this interesting new approach have been available so far.
As in previous work,5 in our present study an increase in pulmonary artery pressure was observed in 50% of the animals during PCPS in cardiac arrest. In these animals, artificially induced pulmonary valve incompetence during cardiopulmonary bypass allowed both pulmonary and, presumably as a result of mitral valve regurgitation, retrograde left ventricular decompression, with significant decrease in left ventricular pressure and wall stress. The observed subsequent increase in myocardial perfusion pressure, total myocardial blood flow, and myocardial oxygen supply (indicated by an increase in coronary venous oxygen saturation) and the resulting decrease in myocardial release of lactate demonstrated the myocardial protective effects of this procedure.
Compared with other methods of left ventricular venting conceivable during PCPS, such as transseptal left atrial cannulation or retrograde transaortic left ventricular decompression, this new approach with pulmonary valve spreading offers some advantages. First, there is no additional risk of fatal iatrogenic aortic valve regurgitation, as observed during attempts at direct transaortic catheter venting in preliminary experiments (unpublished observations; Fig 5⇓). Second, the spreading catheter can be placed easily and, in contrast to transseptal left atrial cannulation, may be applicable in humans during resuscitation. Third, after removal of the guide wire, this catheter allows pulmonary artery pressure measurements, thus allowing assessment of both the need for and the success of pulmonary decompression. Finally, the method may prevent impending pulmonary edema due to a marked reduction of elevated pulmonary artery pressure.
A limitation of this method of retrograde pulmonary left ventricular venting, however, is the inability to directly decompress the left ventricle. Thus, in case of competent mitral valves, as indicated by the lack of a simultaneous increase in pulmonary artery pressure (group B), we found no decompression of the left ventricle. As mentioned above, this phenomenon occurred in one half of our animals during PCPS in ventricular fibrillation. Similarly, Roach and Bellows12 suspected competent mitral valves to be the cause of the failure to decompress the left ventricle during surgery by direct pulmonary artery venting during total cardiopulmonary bypass in a patient who developed severe left ventricular distension.
In conclusion, the prevention of left ventricular and pulmonary damage may be crucial during the use of PCPS in patients with cardiac arrest. In case of an increased pulmonary artery pressure, artificially induced pulmonary valve regurgitation might be a sufficient method to avoid pulmonary edema and irreversible pulmonary damage. Moreover, the method appears to result in effective retrograde left ventricular venting during cardiac arrest. Thus, placement of a pulmonary valve spreading catheter should be considered during the application of PCPS in patients with cardiac arrest. In case of normal pulmonary artery pressures in the presence of competent mitral valves, additional echocardiographic monitoring may be necessary to recognize abrupt and severe left ventricular loading during PCPS. In this situation, the potential of other methods for left ventricular decompression, such as transseptal catheter venting and intermittent mechanical cardiocompression, needs to be examined in both experimental and clinical studies. In the future, direct retrograde transaortic left ventricular venting during PCPS possibly could be managed with a new percutaneous catheter-mounted transvalvular left ventricular assist device 14F in maximum OD and capable of producing flow rates of about 2 L/min.13
We thank Frauke Koopmann for her excellent technical assistance.
- Received November 9, 1994.
- Revision received November 29, 1994.
- Accepted December 3, 1994.
- Copyright © 1995 by American Heart Association
Vogel RA, Shawl F, Tommaso C, O’Neill W, Overlie P, O’Toole J, Vandormael M, Topol E, Kam Tabari K, Vogel J, Smith S, Freedmann R, White C, George B, Teirstein B. Initial report of the National Registry Of Elective Cardiopulmonary Bypass Supported Coronary Angioplasty. J Am Coll Cardiol. 1990;15:23-29.
Hering JP, Scholz KH, Schröder T, Ferrari M, Bock H, Figulla HR, Kreuzer H, Hellige G. Risk of left ventricular loading during percutaneous cardiopulmonary support in cardiac arrest. Coron Artery Dis. 1992;3:419-424.
Scholz KH, Hering JP, Schröder T, Uhlig P, Kreuzer H, Tebbe U, Ferrari M, Hellige G. Protective effects of the hemopump left ventricular assist device in experimental cardiogenic shock. Eur J Cardiothorac Surg. 1992;6:209-214.
Rossi F, Kolobow T, Foti G, Borelli M, Mandava S. Long term cardiopulmonary bypass by peripheral cannulation in a model of total heart failure: the decompression of the left heart through a percutaneous helical spring positioned within the lumen of the tricuspid and pulmonary artery valves. J Thorac Cardiovasc Surg. 1990;100:914-920.
Scholz KH, Figulla HR, Schweda F, Smalling RW, Hellige G, Kreuzer H, Aboul-Hosn W, Wampler RK. Mechanical left ventricular unloading during high risk coronary angioplasty: first use of a new percutaneous transvalvular left ventricular assist device. Cathet Cardiovasc Diagn. 1994;31:61-69.