Patient Population

We retrospectively analyzed data from 63 consecutive patients (52 men, 11 women) who had been supported for ≥90 days with continuous-flow HeartMate II LVADs between February 2007 and January 2010. All subjects were in NYHA functional class I or II at follow-up, with no echocardiographic or clinical evidence of device malfunction, no need for intravenous inotropic support or diuretics, and no other significant illness requiring hospitalization at the time of baseline (90 days after LVAD surgery) or 6-month echocardiographic examination. All 63 patients underwent comprehensive echocardiographic examinations preoperatively and at 90 days after LVAD surgery. Thirty-nine patients underwent follow-up echocardiographic examinations 180 days after surgery.

Echocardiographic Measurements

Two-dimensional transthoracic echocardiography was performed in a standard manner using the Sonos 5500 (Philips Medical Systems, Andover, MA), Sequoia 512 (Siemens Medical Solutions USA, Inc., Mountain View, CA), or Vivid 7 (GE Medical Systems, Milwaukee, WI).

LV diameters and interventricular septal and posterior wall width were measured from the parasternal short-axis view using two-dimensional or two-dimensionally guided M-mode echocardiography of the left ventricle at the papillary muscle level using the parasternal short-axis view from the trailing edge of the septum to the leading edge of the posterior wall, as recommended. Ejection fraction was calculated using the Quinones method, assuming an akinetic apex, and LV mass was calculated using the method proposed by Devereux et al. , using the same LV dimensions made with M-mode or two-dimensional echocardiography as described above.

Valvular regurgitation was qualitatively assessed using color Doppler according to the guidelines of the American Society of Echocardiography (normal or trivial = 1, mild = 2, moderate = 3, and severe = 4). Right ventricular (RV) chamber size and function were measured according to the guidelines of the American Society of Echocardiography. RV size and systolic function were evaluated by measuring RV end-diastolic area (four-chamber view), RV end-systolic area (four-chamber view), and tricuspid annular end-systolic diameter (four-chamber view) and calculating RV fractional area change. RV function was qualitatively graded using a four-point grading system (normal, mild, moderate, or severe) using all apical views and the RV inflow, parasternal long-axis, parasternal short-axis, and subcostal views. Care was taken to obtain a true, nonforeshortened apical or para-apical four-chamber view, oriented to obtain the maximum RV dimension before making the RV estimation. RV function was then assessed using two other methods: two-dimensional lateral tricuspid annular motion and the RV index of myocardial performance (RIMP), as previously described by others ( Figures 1 and 2 , Videos 1 and 2 ; view video clips online). Tricuspid annular diameter was measured at end-systole as well. The RV outflow tract (RVOT) velocity-time integral (VTI) was measured from pulsed-wave Doppler, and RVOT diameter was measured from the parasternal short-axis view at the level of the great arteries. Total RV output (the sum of LVAD output and the flow ejected through the LV outflow tract) was calculated using the formula (RV outflow diameter/2) ² × π × RV outflow VTI. Mean pulmonary artery pressure was estimated using the modified Bernoulli formula: [4 × (peak systolic tricuspid regurgitation [TR] velocity at end-expiration) ² ] + right atrial pressure estimated by the inferior vena cava diameter as well as its response to inspiration, as previously described. Measurements of the LVAD components were performed, including the inflow cannula, outflow graft flow velocities, and native aortic valve opening status ( Figures 3–5 ).

(A) Parasternal long-axis view showing severely dilated left ventricle (end-diastolic dimension, 85 mm). (B) Apical four-chamber view showing a short E-wave deceleration time, suggesting very high LV end-diastolic pressure. (C) RV ejection time was measured from pulsed-wave Doppler of the RVOT flow, and tricuspid valve flow time was measured from the interval between the onset and cessation of TR flow using continuous-wave Doppler echocardiography. RIMP was calculated as (TR flow time − RV ejection time)/RV ejection time. Note the very short RV ejection time, suggesting low RV output, and the TR signal, suggesting prolonged positive and negative dP/dt. The calculated RIMP is increased, suggesting that the right ventricle is spending most of the time in isovolumic contraction and relaxation and not in ejection (inefficiency). (D) End-diastolic diameter decreased and the septal and posterior walls thickened. (E) E-wave deceleration time lengthened, suggesting a decrease in LV end-diastolic pressure. Note also the plateau in mid-diastolic flow velocity resulting from the continuous suction created by the LVAD throughout diastole. (F) RV ejection time is longer, suggesting improved RV output. TR flow time is shorter and RIMP is smaller, suggesting improved RV efficiency (more time in ejection and less time spent as isovolumic contraction and relaxation). Note that the TR signal is symmetric and steeper, suggesting improved positive and negative dP/dt.

(A) Right ventricular end-diastolic area (RVEDA) and (B) right ventricular end-systolic area (RVESA) were traced before LVAD therapy (see Video 1 ). Right ventricular fractional area change (RVFAC) was calculated as (RVEDA − RVESA)/RVEDA. (C,D) RVEDA and (B) RVESA were traced 3 months after LVAD implantation (see Video 2 ). (C) RVEDA was unchanged, RVESA decreased, and calculated RVFAC improved after LVAD implantation. LA , Left atrium; LV , left ventricle; RA , right atrium.

(A) Interrogation of the inflow cannula (see Video 3 ). Inflow cannula and its orientation within the left ventricular apex visualized on the apical four-chamber view, aligned with the left ventricular inflow tract, not touching any wall. Note the properly aligned inflow cannula with laminar and unidirectional flow from the left ventricle (LV) to the device (red). (B) Pulsed-wave Doppler assessment showing laminar, low-velocity forward flow without regurgitation with a pulsatile inflow pattern because the pump inflow originates from the beating LV, resulting in periodic changes in flow throughout the cardiac cycle, reaching a maximum during systole and a minimum during diastole. Maximal systolic and diastolic velocities were recorded, and the difference between them was calculated. The difference between systolic and diastolic inflow velocities is related to left ventricular contraction in a geometry-independent way. (C) Interrogation of the outflow cannula (OC) in the high left parasternal long-axis view showing the end-to-side anastomosis of the OC to the mid ascending aorta (see Video 4 ). (D) Interrogation of the OC in the low right parasternal view (positioned in the sixth to seventh intercostal space) so that the color Doppler shows the typical continuous color flow pattern toward the transducer (red) coming from the pump toward the proximal OC (see Video 5 ). (E) We then continue to follow the OC along the right side of the sternum as it transverses upward toward the ascending aorta (see Video 6 ). In the midsternum, the cannula is parallel to the transducer, and the color flow transforms from red to blue. (G) We continue to follow the OC to the high parasternal area (see Video 7 ). The flow will change direction away from the transducer (blue), bending toward the ascending aorta. This is usually the best view to measure flow velocity, with the pulsed-wave sample (the angle of interrogation will be optimal) positioned ≥1 cm proximal to the aortic anastomosis and the flow moving away from the transducer. AO , Aorta; Asc , ascending; RA , right atrium; RPA , right pulmonary artery; RV , right ventricle.

Pulsed-wave Doppler of the outflow cannula (1) and M-mode echocardiography of the aortic valve (2) in three patients (A–C) with normal LVAD function. The status of aortic valve opening and aortic flow pulsatility depend on pump speed and output, LV contractility, preload, and afterload conditions. Reduction in LVAD speed with reduced unloading of the left ventricle or increased native LV contractility will increase LV systolic pressure above aortic pressure allowing pulsatile flow (A1) and aortic valve opening during every cycle (A2) . Increasing the speed to further unload the ventricle or reduced native heart contraction will result in reduced systolic LV pressure, decreased outflow pulsatility (B1) , and intermittent aortic valve opening (B2) . Increasing speed to maximum and further unloading of the left ventricle or markedly decreased native heart contractility will result in nearly continuous LVAD aortic flow (C1) and a permanently closed aortic valve (C2) .

Although pump flow can be estimated by the device controller, we prefer to measure it directly during the echocardiographic examination. (A) First, we measure the outflow cannula diameter in the right parasternal view, calculating its surface area. (B) Second, we measure the outflow cannula flow integral by pulsed-wave Doppler 1 cm proximal to its anastomosis to the aorta. This can be done in the low right parasternal view (flow toward the transducer), high right parasternal view (flow away from the transducer), or high left parasternal view (flow away from the transducer). We usually measure the integral of flow of three successive cycles and divide the number by 3 to receive the flow for one cycle. We then calculate the pump flow using the equation LVAD output = (cannula diameter/2) ² × π × cannula flow time-velocity integral (TVI) × heart rate (HR). (C) For total cardiac output estimation, we measure the RVOT diameter in the short-axis view on the level of large vessels and calculate its surface area. (D) Second, we measure the RVOT flow integral by pulsed-wave Doppler in the same view. We then calculate the total cardiac output by multiplying these measurements by the heart rate.

LVAD Settings

We routinely use an echocardiographically guided “ramped speed study” to set the optimal pump speed before discharge. We generally use a fixed speed setting that falls midway between the minimum and maximum speeds on the basis of the surgeon’s preference and changes in ventricular dimensions, the position of the interatrial septum and interventricular septum, and the frequency of aortic valve opening ( Figure 4 ).

Tricuspid Repair and Replacement

Candidates for LVAD implantation are evaluated using echocardiography to assess native TR. Those with at least moderate TR (jet area/right atrial area ≥20% or vena contracta >4.5 mm) need to undergo tricuspid valve repair or replacement. For tricuspid valve repair, the valve was inspected for pathology. If TR was associated with annular dilatation, a flexible annuloplasty ring (either complete or incomplete) was implanted to tighten the tricuspid annulus. If TR was related to entrapped leaflets from defibrillator or pacemaker leads, the responsible portion of the leads was freed from the leaflets and tucked into the commissure between the septal and posterior leaflets within a modified Kay annuloplasty stitch. If uncorrectable TR pathology was noted (too many adhesions, significant valvular tenting, severe TR such that repair was not likely to hold), we proceeded with tricuspid valve replacement using a tissue valve.

Aortic Valve Stitch

Candidates for LVAD implantation were evaluated using echocardiography to assess for native aortic valve insufficiency. Those with aortic insufficiency due to significant structural defects in the aortic valve underwent aortic valve replacement or patch closure of the valve. However, the majority of patients with aortic insufficiency were corrected with a simple coaptation stitch placed at the central portion of the aortic cusps, as previously described. The ascending aorta was cross-clamped to arrest the heart. A lateral aortotomy incision was made, long enough to complete the anastomosis between the outflow graft and the aorta. An aortic root retractor was placed, and the native aortic valve was inspected through the incision. The cusps of the native aortic valve were carefully examined to make sure that no other structural problems could be causing the aortic insufficiency. If the central aortic insufficiency was due to poor coaptation, a simple coaptation stitch using pledget-supported 5-0 Prolene sutures at the center easily mended this problem. Using this approach, the aortic valve was still able to open for ejection, even though the effective orifice area of the aortic valve was diminished. After placement of the coaptation stitch, the anastomosis between the outflow graft and the aorta was completed in the usual fashion.

Interobserver and Intraobserver Variability

Interobserver variability was assessed by comparing the readings made by another independent echocardiographer in 15 randomly selected patients. Intraobserver variability was determined by having the first observer who measured the data in all patients remeasure the velocities in 15 patients ≥3 month apart. The degrees of interobserver and intraobserver variability were determined using the Bland-Altman method and the within-subjects coefficient of variation. The within-subjects coefficient of variation (calculated as ratio of the standard deviation of the measurement difference to the mean value of all measurements) provides a scale-free, unitless estimate of variation expressed as a percentage. We measured the intraobserver and interobserver reproducibility for the inflow cannula velocity, the outflow cannula velocity, and the outflow cannula VTI, and the variations in measurements were expressed as ratios.

Statistical Analysis

Data are presented as mean ± SD, as percentages, or as medians with first and third quartiles, as appropriate. Comparisons between patients before and after LVAD implantation were performed using the paired Student’s t test or Wilcoxon’s signed-rank test, as appropriate. To analyze potential factors that influence the progression in aortic regurgitation (AR) (dichotomized into worse or stable AR by comparing AR grade at the baseline, 3 month and 6 month evaluations), univariate analyses on the basis of logistic regression were performed on various parameters (clinical, hemodynamic, and echocardiographic). All P values were two sided, and values < .05 were considered statistically significant. All data were analyzed using JMP version 8.0 (SAS Institute Inc., Cary, NC).