Valvular Regurgitation

Background. The effects of mechanical left ventricular assist on the nonassisted right ventricle have not been fully elucidated. Current information indicates that the right ventricle benefits from a lower left atrial pressure; however, ventricular septal shifting and increased venous return caused by left ventricular assist impair right ventricular function. Acute intraoperative alterations in mitral and tricuspid valve regurgitation (MR and TR, respectively) may occur as a result of mechanical left ventricular assist but have not yet been documented. Methods and Results. Eight patients undergoing implantation of a left ventricular assist device (LVAD)

D uring the past three decades, a variety of ventricular assist devices have been designed to supplement or replace left ventricular function. Ventricular assist devices are currently used for temporary circulatory support pending recovery of the patient's natural heart or pending cardiac transplantation, and left ventricular assist devices (LVADs) capable of chronic support are under development. It is anticipated that chronically implantable LVADs will provide an alternative to transplantation for the treatment of end-stage cardiac failure.
One of the issues central to the development and use of LVADs is the effect of mechanical left ventricular assist on the nonassisted right ventricle. Several experimental studies have examined systolic and diastolic ventricular interactions in the setting of left ventricular assist,'-8 and the early experiences with a number of LVADs have been published.9-15 Mitral and tricuspid valve function before and 4 to 110 days after LVAD implantation have been described.'6 However, to date, the acute effect of left ventricular assist on mitral and tricuspid valve function has not been investigated.
This report describes echocardiographic measurements of ventricular size and valvular regurgitation that From the Department of Surgery, University of Alabama at Birmingham.
Correspondence to William L. Holman, MD, Department of Surgery, University of Alabama at Birmingham, University Station, Birmingham, AL 35294.
were made during the implant of LVADs as a bridge to transplantation in eight patients. Comparisons were made between preimplant and postimplant ventricular dimensions and valvular regurgitation, and correlations were tested between the severity of valvular regurgitation and echocardiographic measurements of right and left ventricular size.

Methods
All eight patients had been accepted for cardiac transplantation before LVAD implantation. The patients were entered in a clinical ventricular assist device trial using LVADs from either Thoratec Laboratories Corp, Berkeley, Calif, or Thermo Cardiosystems Inc, Woburn, Mass. Each patient met the specific criteria of the trial for profound heart failure (Appendix). After obtaining informed patient consent, the eight LVADs were implanted between August 27, 1990, and February 14, 1992. The only ventricular assist device available for the first three implants was the Thoratec device. Institutional and Federal Drug Administration approval for the use of the Thermo Cardiosystems LVAD was received on June 1, 1991. After this date, all patients considered for LVAD support were randomized to receive one device or the other. Note that biventricular Thoratec assist devices were used in patients with intractable ventricular arrhythmias or severe right ventricular failure, as defined by a mean right atrial pressure of .20 mm Hg. Nine patients in our total series of 17 Thoratec and Thermo Cardiosystems ventricular assist device patients had biventricular assist devices.
All of the Thermo Cardiosystems LVADs were placed from the left ventricular apex to the aorta. In two Thoratec LVADs, the left atrium was cannulated from the right lateral aspect of the atrium between the right superior and inferior pulmonary veins. In the other two Thoratec LVADs, the apex of the left ventricle was cannulated ( Table 1). The Thoratec LVAD outflow grafts were anastomosed to the ascending aorta. All LVADs were initially run in the fixed-rate mode but were switched to the variable-rate mode as soon as possible (ie, "automatic" pumping for the Thermo Cardiosystems drive console; "volume" pumping for the Thoratec drive console). In the variable-rate mode, the LVAD cycles asynchronously with regard to the patient's intrinsic cardiac activity. During postimplant echocardiographic evaluation, the Thermo Cardiosystems and Thoratec devices were run in the variable-rate mode with relatively short ejection times (250 to 300 milliseconds).
Informed consent was obtained from all patients before transesophageal echocardiography. Transesophageal echocardiographic examinations were performed using a 5.0-MHz phased-array transducer (Hewlett-Packard Sonos-1000, Andover, Mass; Aloka 870 Color Doppler System, Wallingford, Conn). Standard transesophageal imaging planes17'18 were obtained in all patients. These planes included four chamber and transgastric short-axis views. Pulsed Doppler interrogation of the left and/or right superior pulmonary veins near their entrance into the left atrium was also performed in six patients. In two patients, adequate preimplant transesophageal echocardiographic examinations were not obtained during surgery; therefore, closed chest transthoracic studies obtained before LVAD implantation were used for the comparisons. Transthoracic echocardiographic studies were performed with a 2.5 -MHz phased-array transducer (Hewlett-Packard Color Doppler Imaging System). The transthoracic imaging included recordings of parasternal long-axis, short-axis, and apical four-chamber views.
During the transthoracic and transesophageal color Doppler studies, the color gain setting was optimized by gradually increasing the gain until static background noise first appeared.19'20 If mitral or tricuspid regurgitation was noted, the maximum regurgitant jet area was measured by planimetry.21 According to the measurement system of Yoshida et al,21 a regurgitant jet of <4 cm2 represents mild regurgitation, 4.0 to 7.0 cm2 represents moderate regurgitation, and >7.0 cm2 represents severe regurgitation. At the University of Alabama at Birmingham (UAB) a slightly different scale is used. A jet of <6 cm2 is associated with mild valve regurgitation, a jet of 6.0 to 10.0 cm2 is associated with moderate regurgitation, and a jet of >10 cm2 is associated with severe regurgitation (unpublished data, N.C. Nanda). In this study, the UAB scale was used to describe valvular regurgitation as mild, moderate, or severe.
Left and right ventricular end-diastolic and endsystolic dimensions were obtained before and after LVAD implantation from the four-chamber view. The end-diastolic dimension represents the maximum distance between the endocardial surface of the ventricular septum and the ventricular free wall; the end-systolic dimension represents the minimum distance between the endocardial surface of the ventricular septum and the ventricular free wall. These distances were measured at a point 3 cm below the mitral or tricuspid annulus. Pulmonary vein flow velocities were obtained by pulsed Doppler interrogation. The maximum amplitudes of the pulmonary vein systolic and diastolic waveforms were measured for subsequent comparisons.
All patients were off cardiopulmonary bypass with the chest open at the time of the postimplant echocardiogram. Each patient's blood volume was adjusted according to the operating surgeon's judgment of attaining optimal LVAD function without raising the mean right atrial pressure above clinically acceptable limits. At the time of postimplant echocardiography, negative pump pressure had not been activated in the Thoratec device.
Statistical analysis of the data was performed using SAS-PC software (SAS, Inc; Cary, NC). Within-group comparisons were made with a paired Student's t test (contained in the "Means" procedure of SAS-PC). Pearson's coefficients and P values for correlations were computed using the "Corr" procedure of SAS-PC. Subgroup comparisons (Tables 5 and 6) were made using an analysis of variance procedure (least-squares means test in the "General Linear Models" procedure of SAS-PC).

Results
The eight LVAD patients included one patient with acute myocarditis, four patients with idiopathic dilated cardiomyopathy, and three patients with ischemic heart disease. All patients had profound heart failure (Appendix) despite maximal conventional medical therapy. The patient with myocarditis was taken to the operating room on an emergency basis because of severe hypotension. One patient with ischemic heart disease was transferred to this institution on a centrifugal pump (Medtronic Bio-Medicus Inc, Eden Prairie, Minn) LVAD after surgery for acute myocardial infarction. After evaluation and acceptance for cardiac transplantation, the centrifugal pump was replaced with a Thoratec LVAD. Other demographic and baseline hemodynamic data are contained in Table 1.
All patients survived LVAD implantation. One patient required mediastinal reexploration for delayed tamponade followed by a second reexploration for removal of an infected expanded polytetrafluoroethylene membrane that had been used to reconstruct the pericardium. Another patient experienced a cerebral air embolism from the LVAD; however, he ultimately made a full neurological recovery. All patients required catecholamine support during the immediate postimplant period; however, the right atrial pressures of the eight patients were maintained at <20 mm Hg (range, 11 to 20 mm Hg). Catecholamines were discontinued within the first 5 postimplant days in all patients. Temporary mechanical right ventricular assist or prostaglandin infusions were not used.
There was an improvement in circulatory function noted immediately after LVAD placement ( Table 2). The mean systolic blood pressure rose from 84±4 to 111±7 mm Hg. The mean cardiac index rose from 1.8±0.1 to 2.8±0.2 L. minm m 2. This circulatory improvement was accompanied by a fall in the mean left atrial pressure from 26±2 to 5 ± 1 mm Hg and a rise in the mean right atrial pressure from 10±2 to 16±1 mm Hg.
Transesophageal echocardiography demonstrated a decrease in left ventricular size and an increase in right ventricular size after initiating left ventricular assist (Table 3). The preimplant to postimplant changes were significantly (ie, P<.05) different for the left ventricular end-diastolic dimension, the left ventricular end-systolic dimension, and the right ventricular end-systolic dimension. The postimplant increase in right ventricular enddiastolic dimension bordered on significance (P=.051).
The determination of valvular regurgitant jet area by echocardiography ( Table 3 The four-chamber view of the heart suggested that there was a leftward displacement of the septum after initiating LVAD pumping (Fig 5A and SB); however, this could not be accurately quantitated. Specifying septal position at various times in the cardiac cycle was difficult because the septum waved between the right and left ventricular cavities like a passively moving partition rather than an actively contracting muscular structure.
Correlations were sought between ventricular dimensions and regurgitant jet areas (Table 4). Significant correlations were discovered between the preimplant and postimplant dimensions of the right ventricle and the size of the tricuspid regurgitant jet as well as between the change in the right ventricular dimensions and the change in the tricuspid regurgitant jet. Interestingly, similar consistently significant correlations were not found for the left ventricular dimensions and the mitral regurgitant jet.
Subgroups of patients were compared according to the patient's pathology (ischemic cardiomyopathy, idiopathic dilated cardiomyopathy, or myocarditis), pump inflow cannulation site (left atrium or left ventricular apex), and type of LVAD (Thermo Cardiosystems or Thoratec) (Tables 5 and 6). Analysis according to cardiac pathology did not demonstrate any significant differences in ventricular dimensions or valvular regur-   The left atrium was cannulated in patients considered to have potential for cardiac recovery. In this series, one patient with acute myocarditis and one patient with an acute infarction had ventricular assist device cannulations of the left atrium. These two patients tended to have smaller ventricular dimensions  Circulation Vol 88, No  and valvular regurgitant jet areas than the other six patients, who were judged to have irreversible endstage cardiac dysfunction. An analysis was performed according to the type of LVAD used (Thermo Cardiosystems or Thoratec device). No consistent trends or significant differences were noted between groups. Discussion Controversy persists regarding the ability of an LVAD to provide adequate circulatory support in patients with an intermediate degree of right ventricular failure. Right ventricular function is difficult to quantitate precisely in the setting of severe left ventricular failure because of the increased afterload imposed on the right ventricle by elevated left ventricular diastolic pressure as well as the diminished volume stress on the right ventricle that results from poor left-sided systolic function. The inability to accurately quantitate right ventricular function in the setting of severe left ventricular failure, in turn, makes it difficult to predict the response of the right ventricle to the acute decrease in afterload and increase in preload that result from mechanical left ventricular assist.
Predicting the function of the right ventricle after LVAD implantation in patients with an intermediate degree of right ventricular failure has been the subject of several experimental studies that examined systolic and diastolic ventricular interactions'-8 as well as reports that described clinical results of LVADs used as a bridge to transplantation.9-'5 Currently available experimental evidence indicates that anatomic ventricular interactions play a relatively minor role in the normal heart subjected to mechanical left ventricular assist.78 However, in experimental models of congestive heart failure, leftward septal shifting due to left ventricular unloading and volume loading of the right ventricle due to increased venous return appear to be detrimental to right ventricular function. 1"6 Clinical studies of LVADs"1-'5 document a low incidence of right ventricular failure following LVAD implantation. However, temporary right ventricular assist devices have been necessary to sustain some patients immediately after LVAD implantation, and the mortality in this group of patients is higher than in the group that did not require temporary right-sided assist.13 '14 The present study documents the effect of left ventricular assist on mitral and tricuspid valve regurgitation. Some degree of mitral and tricuspid insufficiency was present in seven of the eight patients before surgery. One patient, who had acute myocarditis, had no detectable mitral valve regurgitation before or after LVAD placement. The other seven patients had predominantly mitral valve regurgitation (ie, mitral jet area>tricuspid jet area) before LVAD implant. The severity of preimplant mitral valve regurgitation was correlated with preimplant left ventricular size, and the severity of preimplant tricuspid valve regurgitation was correlated with preimplant right ventricular size. These findings suggest that annular dilatation and chordal tension were responsible for the observed valvular insufficiency and that shifts in pressure and volume loads resulting from left ventricular assist could alter the severity of this insufficiency.
The mitral regurgitant jet area and the left ventricular dimensions decreased in all seven of the patients with preimplant mitral valve regurgitation after initiating LVAD pumping. The increase in the pulmonary vein systolic flow measurement, although of borderline significance, corroborates the decrease in mitral valve regurgitation observed after LVAD implant. Neither the changes in left ventricular dimensions nor the absolute postimplant left ventricular dimensions were significantly correlated with the postimplant mitral regurgitant jet area. This lack of correlation may be related to the small number of patients in the study.
Right ventricular dimensions increased, and the severity of tricuspid valve regurgitation increased after initiating left ventricular assist in six of the eight patients. In contrast to the left ventricle, the postimplant right ventricular end-systolic and end-diastolic dimensions were significantly correlated with the postimplant tricuspid regurgitant jet area. Moreover, the preimplant to postimplant change in right ventricular dimensions was significantly correlated with the preimplant to postimplant change in tricuspid regurgitant jet area. It is possible that, in the six patients with a postimplant increase in tricuspid regurgitation, the volume load imposed on the right ventricle by increased venous return and leftward shifting of the ventricular septum acutely dilated the tricuspid annulus and increased The right atrial pressure increased significantly, and the left atrial pressure decreased significantly after LVAD implantation; however, clinical evidence of right ventricular failure did not develop in these eight pa- tients during 70 to 279 days of left ventricular assist. The clinical importance of acutely increased tricuspid valve regurgitation and decreased mitral valve regurgitation that was noted in this study after LVAD implantation remains uncertain; however, it is possible that progressively more severe tricuspid valve regurgitation could impair forward flow to an LVAD. It is also possible that falling pulmonary vascular resistance23,12 and improved right ventricular function due to increased coronary perfusion22 during left ventricular assist will ultimately lead to a decrease in right ventricular dimensions and a decrease in tricuspid valve regurgitation, with subsequent improvement in right ventricular function. Since multiple factors affect right ventricular performance during left ventricular mechanical assist,23 the question of long-term outcome in LVAD patients with moderate to severe tricuspid valve regurgitation is best answered by continued clinical trials of LVADs.