Portal de Periódicos da CAPES

Transesophageal Echocardiography in a Patient in Hemodynamic Compromise After Jarvik 2000 ™ Implantation: The Suckdown Effect

2008; Lippincott Williams & Wilkins; Volume: 107; Issue: 3 Linguagem: Inglês

10.1213/ane.0b013e3181806009

1526-7598

William J. Mauermann, Kent H. Rehfeldt, Soon J. Park,

Cardiac Arrest and Resuscitation

A 39-yr-old man presented with decompensated heart failure for placement of a Jarvik 2000 ™ (Jarvik Heart Inc., New York, NY) left ventricular (LV) assist device (LVAD) as a bridge to transplant. He was inotrope-dependent with a LV ejection fraction of 14% and severe right ventricular (RV) dysfunction. The device was placed at the LV apex with the outflow cannula anastomosed to the ascending aorta. He was weaned from cardiopulmonary bypass (CPB) to LVAD support without hemodynamic instability. However, his course was complicated by persistent nonsurgical bleeding. Nine hours after his initial operation he returned to the operating room in hemodynamic distress for emergent exploration. The patient was tachycardic to 140 bpm with a mean arterial blood pressure of 42–55 mm HG, despite support from the LVAD at setting 3 (approximately 10,000 revolutions per minute [RPM]). Considering the persistent postoperative bleeding, his hemodynamic compromise was likely due, in part, to hypovolemia despite aggressive resuscitation in the intensive care unit. Cardiac tamponade also needed to be excluded. Intraoperative transesophageal echocardiography (TEE) demonstrated no evidence of pericardial fluid or clot accumulation. However, his LV end-diastolic volume was markedly reduced with near collapse of the ventricular walls adjacent to the LVAD inflow site (Fig. 1, Video 1; please see video clip available at www.anesthesia-analgesia.org). Based on the TEE images, LVAD flow was transiently reduced to lower settings and additional resuscitative fluids were administered resulting in adequate LV filling (Fig. 2, Video 2), a mean arterial blood pressure of 75 mm Hg and a heart rate of 90 bpm. Of note, neither the pulmonary artery diastolic pressure nor central venous pressure changed significantly before, during, or after this episode. He was discharged from the hospital without any deficits and subsequently underwent successful orthotopic heart transplantation.Figure 1.: In this midesophageal four chamber view the left ventricle is severely under-filled. The left ventricular assist device pump (large arrow) is contacting the myocardium and inflow obstruction is likely present. Note the small diameter of the mitral annulus and the bowing of the ventricular septum into the left ventricle.Figure 2.: In this midesophageal four chamber view the left ventricle is now well filled and the left ventricular assist device pump (arrow) is ideally positioned to receive the left atrial inflow through the mitral valve.DISCUSSION TEE is essential in the perioperative management of patients undergoing placement of LVADs. In the pre-bypass period, the overall structure of the heart is assessed with careful attention to the competency of the aortic valve and the presence or absence of intracardiac shunts.1,2 After weaning from CPB, TEE is used to assess RV function, LVAD cannulae placement, intracardiac shunting, and LV filling.1,3 We report the use of TEE in the diagnosis and management of a patient in hemodynamic distress after placement of a Jarvik 2000 LVAD. The second generation, axial flow LVADs, (Jarvik 2000 and Heartmate II ™) represent a change in technology from the pulsatile devices. Axial flow devices use an impeller rotating at 6000 to 15,000 RPM with maximum outputs of up to 10 L/min, depending on the settings. Thus, the impeller of the device continues at its set speed regardless of inflow to the ventricular cannulae. Clearly, a LVAD is capable of providing hemodynamic support only if there is adequate inflow to the device. Delivery of blood to the LV cannulae is dependent on adequate intravascular volume, RV function, and a well positioned inflow cannulae. If hemodynamic compromise occurs, the position of the inflow cannula (Heartmate II) or pump (Jarvik 2000) in the LV, RV function, and LV filling may all be easily assessed by TEE using midesophageal windows.1 The aorta should be interrogated for signs of dissection. A careful exploration should be undertaken looking for thrombus in the left-sided chambers that may embolize and obstruct the LVAD. Obstruction of the device inflow can occur from thrombus or by collapse of the myocardial walls over the device inflow. Inflow cannula obstruction may be detected by turbulent flow with color Doppler imaging or by a peak velocity more than 2.3 m/s using continuous wave Doppler.1 We find this to be difficult in patients with a Jarvik 2000. With this device, the pump resides within the LV and the inflow is essentially directly into the pump. Although it is still possible to track inflow from the mitral orifice to the device with color Doppler, there is always significant aliasing from the mechanical artifact of the device. In our experience, neither continuous nor pulse wave Doppler evaluations of the Jarvik 2000 inflow have been possible due to reverberation artifact from the pump. This report demonstrates a phenomenon often seen when patients are initially weaned from CPB and onto support from the axial flow LVAD. The transition must be slow with meticulous attention paid to LV filling. LV filling may be assessed by TEE or direct left atrial pressure measurements. If the RPMs of the device are increased too quickly without adequate ventricular filling, a "suck down" effect quickly occurs whereby the ventricular walls begin to collapse and contact the LVAD inflow leading to obstruction. The ventricular septum bows into the LV and the RV assumes an unfavorable geometry leading to RV dysfunction and further worsening of volume delivery to the LV. Position of the LVAD cannula or pump, LV filling, and the ventricular septum are assessed with the midesopheageal four chamber view. If the midesophageal two chamber view is also used, LV filling and position of the LVAD inflow can be assessed in two planes. When a patient with a LVAD becomes hemodynamically unstable, the natural tendency is to increase the RPMs of the device in an effort to increase support. However, in the situation of inadequate volume delivery to the device, increasing the RPMs will worsen the clinical picture by causing LV wall collapse at the site of inflow and inducing RV dysfunction. The correct maneuver is to decrease the RPMs while restoring intravascular volume. Once the hyopvolemia is corrected, full support from the LVAD may be reinstituted slowly. It may be possible to diagnose LV "suckdown" and other causes of hemodynamic instability using transthoracic echocardiography. However, the postoperative surgical changes including drains and bandages, and the position of the device in the chest, make transthoracic imaging and interpretation difficult. We maintain a low threshold for using TEE to evaluate these patients in the immediate postoperative period, and we consider doing so in the operating suite where surgical therapy, if needed, can be swiftly instituted.