A simplified echocardiographic technique for detecting continuous-flow left ventricular assist device malfunction due to pump thrombosis

http://www.jhltonline.org

A simplified echocardiographic technique for detecting continuous-flow left ventricular assist device malfunction due to pump thrombosis Jerry D. Estep, MD,a Rey P. Vivo, MD,a,b Andrea M. Cordero-Reyes, MD,a Arvind Bhimaraj, MD,a Barry H. Trachtenberg, MD,a Guillermo Torre-Amione, MD, PhD,a,c Su Min Chang, MD,a Barbara Elias, RN,a Brian A. Bruckner, MD,a Erik E. Suarez, MD,a and Matthias Loebe, MDa From the aHouston Methodist DeBakey Heart & Vascular Center, Houston, Texas; the bMechanical and Circulatory Support and Heart Transplantation Program, UCLA Ahmanson Cardiomyopathy Center, UCLA, Los Angeles, California; and cCátedra de Cardiologia y Medicina Vascular, Tecnológico de Monterrey, Monterrey, Nuevo León, México.

KEYWORDS: heart failure; left ventricular assist device; thrombosis; echocardiography; cardiomyopathy

BACKGROUND: Malfunction of a continuous-flow left ventricular assist device (CF-LVAD) due to device thrombosis is a potentially life-threatening event that currently presents a diagnostic challenge. We aimed to propose a practical echocardiographic assessment to diagnose LVAD malfunction secondary to pump thrombosis. METHODS: Among 52 patients implanted with a CF-LVAD from a single center who underwent echocardiographic pump speed-change testing, 12 had suspected pump thrombosis as determined by clinical, laboratory, and/or device parameters. Comprehensive echocardiographic evaluation was performed at baseline pump speed and at each 1,000-rpm interval from the low setting of 8,000 rpm to the high setting of 11,000 rpm in 11 of these patients. RESULTS: Receiver operating characteristic curves and stepwise logistic regression analyses showed that the best diagnostic parameters included changes in the LV end-diastolic diameter (o0.6 cm), aortic valve opening time (o80 msec), and deceleration time of mitral inflow (o70 msec) from lowest to highest pump speed. One parameter was predictive of pump malfunction, with 100% sensitivity and 89% specificity, whereas 2 of 3 parameters increased the sensitivity to 100% and specificity to 95%. CONCLUSIONS: The 3 echocardiographic variables of measured changes in LV end-diastolic diameter, aortic valve opening time, and deceleration time of mitral inflow between the lowest (8,000 rpm) and highest pump speed settings (11,000 rpm) during echo-guided pump speed-change testing appear highly accurate in diagnosing device malfunction in the setting of pump thrombosis among patients supported with CF-LVAD. Further investigation is warranted to create and validate a prediction score. J Heart Lung Transplant 2014;33:575–586 r 2014 International Society for Heart and Lung Transplantation. All rights reserved.

Registry data show that more than 6,000 patients with advanced heart failure (HF) are currently supported by Reprint requests: Jerry D. Estep, MD, Methodist DeBakey Heart & Vascular Center, 6550 Fannin St, Smith Tower Ste 1901, Houston, TX 77030. Telephone: 713-441-2761. Fax: 713-790-2643. E-mail address: [email protected]

continuous-flow left ventricular assist devices (CF-LVADs), representing most of the devices implanted for mechanical circulatory support.1–4 Although 1-year and 2-year survival rates have progressively improved to approximately 80% and 70%, respectively, CF-LVAD use is not without risks.3–5 One such important post-implant complication is thrombus formation in the LVAD system that can result in hemolysis

1053-2498/$ - see front matter r 2014 International Society for Heart and Lung Transplantation. All rights reserved. http://dx.doi.org/10.1016/j.healun.2014.01.865

576

The Journal of Heart and Lung Transplantation, Vol 33, No 6, June 2014

and malfunction, with one or more clinical consequences, including HF, stroke, renal failure, and death. Thrombosis leading to LVAD failure is a medical emergency, yet management of this condition can be compounded by the challenge of making a definitive diagnosis.5 Although objective testing with different imaging techniques and management algorithms has been recently reported, only one report has been published on the role of echocardiography to define LVAD malfunction, and there remains a lack of consensus on the most practical imaging protocol to detect this potentially fatal condition.6–9 Two-dimensional and Doppler echocardiography can be useful in detecting thrombus formation and turbulent flow suggestive of cannula obstruction.10 However, its capacity to evaluate the inflow and outflow cannula is limited by acoustic windows/shadowing and cannula artifact. Multidetector cardiac computed tomography (CT) can better visualize cannulas and surrounding anatomy, yet it is unable to detect thrombus within the LVAD pump itself, is incapable of diagnosing LVAD malfunction, and involves some radiation and contrast exposure.10,11 Recently, Uriel et al12 demonstrated that their echocardiographic speed optimization “ramp” test protocol reliably detected LVAD thrombosis and malfunction. They found that in 9 patients with suspected device thrombosis, attenuated reduction in LV dimensions by echo with increasing LVAD pump speed and increased power detected by the LVAD console were diagnostic of flow obstruction secondary to LVAD-related thrombosis. However, the Columbia echo ramp-up protocol is partly limited by its use of 11 different pump speed measurements and the need to derive linear slope calculations to screen for a LV end-diastolic diameter (LVEDd) slope 4 –0.16. Our institution has used pump speed-change testing for several years to bracket the extent of pump speed support to screen for device malfunction, myocardial recovery, and detect pulmonary hypertension before heart transplantation. In addition, we perform it to detect optimal LV unloading based on LV size measurements and aortic valve (AV) opening in persistently symptomatic patients.13 The aims of this study were to






perform comprehensive echocardiography-guided rampup speed testing to determine the association between standard 2-dimensional (2D) and spectral Doppler echo parameters and LVAD malfunction, examine serologic markers of hemolysis with LVAD malfunction secondary to pump system thrombosis, and examine those parameters with the strongest association with pump thrombosis in an attempt to further define and improve the diagnosis of LVAD malfunction in this potentially life-threatening condition.

Methods The Methodist DeBakey Heart & Vascular Center Investigational Review Board approved this study.

Patient population Between January 2008 and July 2013, a HeartMate II (Thoratec Corp, Pleasanton, CA) CF-LVAD was implanted in 201 patients at

our center. We identified retrospectively from our LVAD database 12 patients who had suspected or confirmed pump thrombosis according to the following criteria:

1. suspected pump thrombosis with evidence of significant hemolysis, defined as a lactate dehydrogenase (LDH) level 42.5 upper limit of normal, associated decrease in baseline hemoglobin, hematuria, evidence of HF without another explanation, and acute renal failure and/or stroke and/or death; 2. confirmed pump thrombosis by direct visualization at the time of surgery or unequivocal evidence by cardiac CT. From the remaining 189 patients, we randomly selected 40 clinically stable patients for the control group who had an echo pump speed-change test as part of a prospective echo surveillance protocol. Patients in the control group who had a mitral valve annuloplasty ring (n ¼ 2), significant mitral annular calcification (n ¼ 1), or sub-optimal images (n ¼ 1) were excluded. None of the patients included in the study had greater than moderate aortic insufficiency or underlying atrial fibrillation.

Echo pump speed-change protocol An echo pump speed-change protocol was performed prospectively in all patients as part of surveillance testing for evaluation of myocardial recovery and pre-heart transplant pulmonary pressure assessment and for those with a change in clinical status (i.e., residual HF or suspected pump thrombosis). Complete transthoracic echocardiographic studies were performed by standard fashion and were reviewed by an independent reader blinded to the patient’s clinical outcomes (pump thrombosis vs control stable LVAD cohort). Images were initially acquired at each patient’s baseline pump speed, typically at 8,800 to 10,200 rpm. The LVAD pump speed was then decreased to the lowest setting of 8,000 rpm and successively ramped up at 1,000-rpm intervals to the highest setting of 11,000. At each interval, we allowed 2 minutes before acquiring a complete set of 2D and Doppler parameters. From the parasternal window, LVEDd, pulmonic annulus diameter, and right ventricular outflow tract (RVOT) velocity were measured per guidelines.14,15 RVOT stroke volume was derived as the RVOT cross-sectional area  RVOT time-velocity integral flow by pulsed-wave Doppler. Systemic cardiac output (CO) was calculated as RV stroke volume  heart rate. In addition, 2D echo and M-mode were used from the parasternal window to record AV function per institutional guidelines in patients on LVAD support.13 AV opening time was measured from M-mode images and averaged over 3 cardiac cycles. From the apical window, pulsed-wave Doppler was used to record mitral inflow at the level of the mitral valve leaflet tips. Doppler signals were analyzed for peak early (E) and late (A) diastolic velocities, E/A ratio, and deceleration time (DT) of mitral E velocity.16 Tissue Doppler was applied to measure mitral annular early (e0 ) velocities at the lateral and septal sides of the annulus.16 The resulting annular velocities by pulsed-wave Doppler were recorded for 3 to 5 cardiac cycles at a sweep speed of 100 mm/sec. E/e0 ratios were computed. Valvular regurgitation signals were recorded and interpreted in accord with standard recommendations.17 Inferior vena caval diameter and its collapse and hepatic venous flow were recorded in the sub-costal view.18 Pulmonary artery (PA) systolic pressure was derived using the modified Bernoulli equation as PA systolic pressure in mm Hg ¼ 4(v)2 of peak tricuspid regurgitation velocity in m/sec þ right atrial pressure in mm Hg. Right atrial pressure was

Estep et al.

Echo Assessment of LVAD Pump Thrombosis

estimated using the inferior vena caval diameter and its change with respiration and hepatic venous flow recorded in the sub-costal view.18 An echo pump speed-change protocol was not performed in patients with (1) aortic root thrombus (known or noted at baseline pump speed settings) and/or (2) a sub-therapeutic international normalized ratio (INR) o 1.8 without concomitant parental heparin.

Hemolysis screening protocol At our institution, we obtain LDH, haptoglobin, plasma free hemoglobin, complete blood count, liver function tests, urinalysis, prothrombin time, INR, and partial thromboplastin time as part of a surveillance protocol and with a change in clinical status and/or in the presence of changes in LVAD console parameters at baseline LVAD rpm support. We also track LVAD console findings for pulsatility index, flow, and power at the different pump speed settings.

Statistical analysis Continuous variables are presented as mean ⫾ standard deviation or median (range) and were compared using the Mann-WhitneyWilcoxon test with a 95% confidence interval (CI), whereas categoric variables are reported as number (%) and were compared using Fisher’s exact test. Multiple measures analysis of variance with the Tukey post hoc test was performed to determine changes at different pump speed settings, and the change (Δ) between the lowest and highest speeds was calculated. Receiver operating characteristic curves were constructed for each parameter Δ to determine the optimal threshold cutoff values, sensitivity, and specificity. A stepwise logistic regression analysis was then performed to examine the best sub-set of variables for pump thrombosis assessment. Values of p of o 0.05 were considered significant.

Results

577 post-operatively, within a range of 1 to 11 days. In 2 of these 4 patients, the device exchange represented the second device exchange, with the initial exchange not related to pump thrombosis, as detailed in Table 2. Nine of 12 patients who underwent device exchange for pump thrombosis survived the early post-operative period, with a median survival after exchange of 225 days (range, 36–309 days).

Echo pump speed-change protocol The echo pump speed-change protocol was performed in all patients without complications. Table 3 reports the echocardiographic changes from lowest to highest pump speed settings between patients with and without suspected or confirmed pump thrombosis. Compared with patients with a normally functioning LVAD, the study group patients had significantly different findings in 4 of the 13 echo parameters measured at the lowest (8,000 rpm) and highest (11,000 rpm) pump speeds: (1) LVEDd, (2) AV opening time, (3) right-sided CO, and (4) mitral inflow DT (Figures 1 and 2).

LV end-diastolic diameter In the presence of a normally functioning LVAD, the LVEDd decreases as pump speed is ramped up, a reflection of augmented LV unloading. Our analysis to detect differences using the interquartile range showed LVEDd was significantly reduced when the values obtained at 8,000 and 11,000 rpm were compared. We observed a mean reduction of 1.1 ⫾ 0.5 cm (low vs high) in patients with normally functioning devices and 0.1 ⫾ 0.2 cm among those with suspected or confirmed pump thrombosis and associated LVAD malfunction.

Patient population

AV opening time

Of the 201 patients who received a CF-LVAD implant, the final study population comprised 48 patients. The baseline characteristics of the 12 patients who met the criteria for suspected or confirmed pump thrombosis and the 36 stable control patients who did not are compared in Table 1. Additional characteristics and outcome of the 12 patients who met the criteria for suspected or confirmed pump thrombosis are provided in Table 2. One of the 12 patients did not have an echo pump speed-change test due to pump arrest and was excluded from the final echocardiographic analysis. The comparison groups did not differ significantly in demographic and clinical characteristics, including variables that may affect hemolysis in LVAD patients such as baseline pump speed and anti-platelet or anti-coagulant medications. Laboratory surrogates of hemolysis, including LDH, total bilirubin, and plasma hemoglobin, were higher among those in the study group than in the control group. The median time to event from LVAD implant was 122 days (range, 1–1,124 days); however, 4 of the 12 patients developed suspected or confirmed pump thrombosis early

AV opening is usually preserved at the lowest levels of pump support. Increasing pump speed in CF-LVADs can unload the LV up to a point in which the mean arterial pressure exceeds the LV systolic pressure. This precludes aortic ejection and AV opening. In patients with normal pump function, we observed a significant decrease in AV opening time from 147.8 ⫾ 78 msec at the lowest speed to 0, indicating persistent AV closure at the highest speed tested. In contrast, the expected response of progressive shortening of AV opening to increasing pump speed was absent in those with pump malfunction. In our cohort of stable patients, all 36 patients had complete AV closure at 11,000 rpm, which included 4 patients with an LV ejection fraction (LVEF) 4 40% at baseline levels of LVAD support.

Right-sided CO With CF-LVADs, right-sided CO is a surrogate of systemic blood flow generated by the LVAD pump and the native LV when the AV opens at least partially. Compared with stable

578 Table 1

The Journal of Heart and Lung Transplantation, Vol 33, No 6, June 2014 Baseline Characteristics of Study Population

Variablea Age, years Body mass index, kg/m2 Male Race Caucasian African American Hispanic Other Etiology of cardiomyopathy Ischemic Dilated Hypertension Diabetes mellitus Days from LVAD implant to echo acquisition Baseline pump speed, rpm Baseline echo parametersb LVEF, % LVEDd, cm LAVi, ml/m2 Aortic valve Open with every cycle Open partial Closed with every cycle Aortic valve insufficiency None/trace Mild/moderate Severe Mitral regurgitation None/trace Mild/moderate Severe Tricuspid regurgitation None/trace Mild/moderate Severe Mean blood pressure, mm Hgc Laboratory values International normalized ratiod Nadir hemoglobin, mg/dl Plasma hemoglobin, mg/dl Haptoglobin, mg/dl LDH,e IU/liter Total bilirubin, mg/dl Creatinine, mg/dl Urinalysis Macroscopic hematuria (% with large blood) Medications Aspirin Warfarin ACE inhibitors β-Blockers Aldosterone antagonists

Suspected/confirmed pump thrombosis cohort (n ¼ 12)

Control cohort (n ¼ 36)

53 ⫾ 12 30.4 ⫾ 4.5 8 (67)

57 ⫾ 10 29.9 ⫾ 5.5 29 (81)

7 (58) 3 (25) 0 (0) 2 (17)

17 (47) 11 (31) 4 (11) 4 (11)

7 (58) 5 (42) 8 (67) 3 (25) 122 (1–1,124) 9,055 ⫾ 614

22 (61) 14 (39) 20 (56) 12 (33) 217 (56–858) 9097 ⫾ 456

26 ⫾ 6 5.8 ⫾ 0.9 26.1 ⫾ 12.6

27 ⫾ 12 5.3 ⫾ 1.0 31 ⫾ 12.2

11 (92) 1 (8) 0 (0)

6 (17) 21 (58) 9 (25)

11 (92) 1 (8) 0 (0)

21 (58) 15 (42) 0 (0)

6 (50) 6 (50) 0 (0)

25 (69) 11 (31) 0 (0)

6 (50) 6 (50) 0 (0) 74 ⫾ 8.9

16 (44) 20 (56) 0 (0) 79 ⫾10

2.0 ⫾ 0.7 7.8 ⫾ 1.6 190 ⫾ 113 11.6 ⫾ 16 7680 ⫾ 6148 5.9 ⫾ 7.9 1.5 ⫾ 0.7

1.9 ⫾ 0.6 10.8 ⫾ 1.9 4.3 ⫾ 2.6 35 ⫾ 70 884 ⫾ 477 0.9 ⫾ 0.6 1.5 ⫾ 0.6

0.2 0.003 0.0003 0.3 0.001 0.0002 0.6

8 (67)

0 (0)

0.0001

p- value 0.3 0.6 0.4 0.6

0.9

0.7 0.7 0.6 0.2 0.7 0.1 0.4 0.05

0.05

0.3

0.7

5 9 5 9 6

(42) (75) (42) (75) (50)

17 31 18 30 17

(47) (86) (50) (83) (47)

0.5

0.9 0.4 0.7 0.7 0.9

ACE, angiotensin-converting enzyme; LAVi, left atrial volume index; LDH, lactate dehydrogenase; LVAD, left ventricular assist device; LVEDd, left ventricular end diastolic diameter; LVEF, left ventricular ejection fraction. a Continuous data are shown as mean ⫾ standard deviation or median (range) and categoric data as number (%). b Derived at baseline pump speed. c Obtained by Doppler if no pulse and/or no aortic valve opening noted and by cuff if pulse present and/or aortic valve opening noted. d At time of initial evaluation for pump thrombosis or at time of echo pump speed change for control cohort. e Normal reference value range: 300–600 IU/liter.

Estep et al. Table 2

Echo Assessment of LVAD Pump Thrombosis

579

Patients With Suspected/Confirmed Pump Thrombosis

Days from INR original before Age implant event (years) Sex to event

Peak Peak total Peak Nadir bilirubin creatinine LDH Hgb (mg/dl) (IU/liter) (mg/dl) (mg/dl)

Peak pump powera (watts)

34 57 58 59 33

M F M M F

86 118 1 438 2

1.5 2.2 1.6 2.3 1.3

6.4 6.2 6.6 7.3 7.1

12,548 19,590 900 5,342 NA

4.1 16.0 1.4 6.9 5.9

9.3 5.5 2.2 2.3 1.8

10.8 13.4 9.0 9.0 NA

48 63 50 57 55

M M F M M

63 64 64 6 11

2.3 4.0 1.9 1.5 1.5

9.7 7 8.5 8.0 9.1

5,184 18,063 11,735 440 912

2.1 27 9.6 2.1 1.0

1.3 1.1 1.7 1.2 0.6

10.1 10.8 9.7 17.7 NA

11 65 12 61

M F

1124 547

1.9 1.7

7.1 9.0

3,699 1,621

1.8 0.6

1.0 0.7

10.0 15

Pt 1 2 3d 4 5e 6 7 8 9 10

Defined eventb Pump clot Hemolysis and death Pump and OC clot Pump clot Pump clot and pump arrest Pump clot Hemolysis and death Pump clot Pump clot OC clot Bend relief disconnect, OC clot, pump arrest Pump clot Pump clot and death

Device exchangec (Yes/No) Yes No Yes Yes Yes Yes No Yes Yes Yes Yes No

F, female; Hgb, hemoglobin; INR, international normalized ratio; LDH, lactate dehydrogenase; M, male; NA, data not available; OC, outflow cannula, Pt, patient. a Derived from the left ventricular assist device console at 11,000 rpm. b 2 of 12 suspected and 10 of 12 confirmed. c All 12 received intensified anti-coagulation, 3 patients were not deemed surgical candidates, refer to text. d Defective percutaneous driveline and required pump exchange after 2 years of HeartMate II (Thoratec, Pleasanton, CA) support, after that he developed pump and OC clot. e VentrAssist research device (Ventracor Sydney, NSW, Australia) exchanged for Heart Mate II and then developed pump clot and pump arrest.

patients with normal device function who had an echoderived CO that increased with increasing pump speed (mean increase: 1.2 ⫾ 0.9 liters/min, low vs high pump speed), those with suspected or confirmed pump thrombosis and associated LVAD malfunction had no significant change in echo-derived CO during ramp-up testing.

Mitral inflow DT DT of mitral early filling velocity reflects the time interval for the maximum mitral gradient to decline to baseline during early diastole. It shortens in the setting of increased LV filling pressures and lengthens with enhanced LV unloading. With increasing pump speed, this expected response was seen in patients with normally functioning devices (mean increase: 88 ⫾ 53 msec) but was markedly blunted in those with suspected or confirmed pump thrombosis and associated LVAD malfunction. The receiver operating characteristic curves for the corresponding cutoff values and area under the curve (AUC) for each significant echo parameter are summarized in Table 4. AV opening time had the strongest specificity, followed by LVEDd, mitral inflow DT, and right-sided CO. The stepwise logistic regression analysis demonstrated that the best sub-set of parameters included LVEDd, AV opening time, and mitral inflow DT, with an adjusted R2 value of 80%, a Mallows’ Cp value of 3, and an S of 0.24. One parameter was predictive of pump malfunction with 100% sensitivity and 89% specificity, whereas 2 of 3 parameters increased the sensitivity to 100% and specificity to 95%. The feasibility results

indicated that all 3 echo parameters were reliably measured in 92% of the patients.

Other surrogates of pump thrombosis Associated LVAD console findings are reported in Table 5. Of the 7 laboratory parameters that were included in the analysis, those that were significantly associated with pump thrombosis were LDH 41,050 IU/liter (AUC ¼ 0.86; specificity, 86%; sensitivity, 80%; p ¼ 0.001), total bilirubin 4 2 mg/dl (AUC ¼ 0.90; sensitivity, 70%; specificity, 96%; p ¼ 0.0002) and plasma hemoglobin 4 67 mg/dl (AUC ¼ 0.97; sensitivity, 83%; specificity, 90%; p ¼ 0.0002).

Discussion Using a comprehensive transthoracic echocardiographic evaluation during pump speed-change testing, we identified 3 echo variables—LVEDd, AV opening time, and mitral inflow DT composed of the change (Δ) between the lowest and highest pump speed settings—that define LVAD device malfunction in the setting of pump thrombosis. Knowing the expected differences comparing values at these 2 settings may help the clinician make a more definitive and rapid diagnosis of LVAD device malfunction. Device thrombotic events have been documented in numerous case reports.6–8,19–23 Results from a few singlecenter and multicenter reports show pump thrombosis occurs in 2% to 4% of patients supported with CF-LVADs.24–27 Correspondingly, the estimated freedom from device exchange or death related to device malfunction

580

Table 3

Echocardiographic Parameters of Patients With and Without Suspected/Confirmed Pump Thrombosis Suspected/confirmed pump thrombosis cohort (n ¼ 11)

Variablea

8,000 rpm

9,000 rpm

10,000 rpm

11,000 rpm 0.9 16 52 1.4 10 25 1.7 56 2.2 1.3 1.6 4.2 6.3

0.9 0.7 0.5 0.8 0.5 0.8 0.8 0.7 0.8 0.6 0.5 0.2 0.9

Δ8–11,000 rpm 0.1 0.7 23.7 –0.2 11.6 –1.2 0.4 –16.7 –0.8 –0.3 –0.8 5.1 0

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.2 1.8 25.9 0.4 12.1 7.9 1.2 24.9 1.5 1.4 1.3 1.8 1

8,000 rpm 5.5 32.9 148 3.9 90.1 61.2 1.8 162 7.6 5.8 6.8 14.7 10.2

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.9 16 78 0.9 32 31 1.0 34 3.0 2.3 2.3 7.7 5.2

9,000 rpm 5.3 32.8 107 4.2 88.3 61.0 1.8 179 7.6 5.6 6.7 14.5 10.2

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

1.0 16 86 1 33 31 1.1 46 2.9 2.2 2.4 7.4 5.2

10,000 rpm 4.9 32.5 25 4.7 81.6 60.2 1.6 218 8.1 5.9 7.1 13.5 10.6

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

1.0 16 68 1.3 31 26 0.8 55 2.6 2.3 2.1 6.9 6.0

11,000 rpm 4.5 33.0 0.0 5.1 80.9 57.2 1.6 251 7.7 5.7 6.8 13.1 10.8

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

p-value

Δ 8-11,000 rpm p-value Δb

1.0 0.0009 1.1 ⫾ 0.5 o0.0001 16 0.9 0.2 ⫾ 0.9 0.2 0 o0.0001 182 ⫾ 49 o0.0001 1.4 o0.0001 -1.2 ⫾ 0.9 0.0009 26 0.5 10.6 ⫾ 13.4 0.8 23 0.9 4.1 ⫾ 13.8 0.2 0.8 0.7 0.2 ⫾ 0.5 0.4 59 o0.0001 –88 ⫾ 53 o0.0001 2.3 0.8 0.1 ⫾ 2.4 0.2 2.1 0.9 –0.1 ⫾ 1.6 0.7 2.1 0.8 –0.2 ⫾ 1.5 0.2 5.6 0.8 2.2 ⫾ 4.3 0.03 5.8 0.9 –0.2 ⫾ 1.1 0.2

A, mitral inflow late diastolic filling peak velocity; AV, aortic valve; CO, cardiac output; E, mitral inflow early diastolic filling peak velocity; e0 , early diastolic mitral annular velocity; LAVi, left atrial volume index; LVEDd, left ventricular end diastolic dimension; MV DT, mitral valve deceleration time; RAP, right atrial pressure. a Data are shown as mean ⫾ standard deviation. b The p-value compares 8,000 vs 11,000 rpm between control and pump thrombosis cohorts.

The Journal of Heart and Lung Transplantation, Vol 33, No 6, June 2014

LVEDd, cm 5.7 ⫾ 1.1 5.6 ⫾ 0.9 5.6 ⫾ 0.9 5.6 ⫾ LAVi, ml/m2 38.3 ⫾ 16 38.3 ⫾ 16 38.9 ⫾ 15 37.6 ⫾ AV opening time, msec 217 ⫾ 55 217 ⫾ 26 215 ⫾ 46 206 ⫾ Right-sided CO, liters/min 4.1 ⫾ 1.5 4.2 ⫾ 1.4 3.9 ⫾ 1.4 4.2 ⫾ E, cm/sec 103.7 ⫾ 16 100.8 ⫾ 18 95.6 ⫾ 17 97.1 ⫾ A, cm/sec 54.5 ⫾ 22 57.5 ⫾ 27 63.6 ⫾ 26 55.1 ⫾ E/A 2.6 ⫾ 1.6 2.5 ⫾ 1.8 2.1 ⫾ 1.6 2.4 ⫾ MV DT, msec 148 ⫾ 31 155 ⫾ 38 159 ⫾ 30 161 ⫾ e' lateral, cm/sec 8.5 ⫾ 3.4 7.9 ⫾ 1.7 8.4 ⫾ 1.7 9.0 ⫾ e' septal, cm/sec 5.6 ⫾ 1.8 6.2 ⫾ 1.9 5.4 ⫾ 1.8 6.4 ⫾ e' average, cm/sec 6.5 ⫾ 1.5 6.9 ⫾ 1.8 6.9 ⫾ 1.5 7.7 ⫾ E/e' 16.5 ⫾ 3.8 15.7 ⫾ 4.8 13.9 ⫾ 3.2 13.5 ⫾ RAP, mm Hg 11 ⫾ 5.5 11 ⫾ 5.5 11.0 ⫾ 5.0 11.3 ⫾

p-value

Control cohort (n ¼ 36)

Estep et al.

Echo Assessment of LVAD Pump Thrombosis

581

Figure 1 Echo detection of normal left ventricular assist device (LVAD) function. Changes are seen in LV dimension, aortic valve (AV) opening time, and mitral valve (MV) inflow deceleration time (DT) from the lowest to highest pump speed in a patient without LVAD dysfunction or clot. The double-sided red arrow defines LV end-diastolic diameter, the double-sided and single yellow arrow marks AV opening time, and the single red-arrow notes MV DT.

Figure 2 Echo detection of left ventricular assist device (LVAD) malfunction. Changes in LV end-diastolic diameter (LVEDd), aortic valve (AV) opening time, and mitral valve (MV) inflow deceleration time (DT) from lowest to highest pump speed in a patient with LVAD dysfunction and confirmed pump thrombosis. The double-sided red arrow defines LVEDd, the double-sided yellow arrow marks AV opening time, and the single red arrow notes MV DT.

582 Table 4

The Journal of Heart and Lung Transplantation, Vol 33, No 6, June 2014 Accuracy of Echocardiographic Parameters in Identifying Patients With Device Malfunction

Variable

Cutoffa

Sensitivity % (95% CI)

Specificity % (95% CI)

AUC (95% CI)

p-value

AV opening time, msec LVEDd, cm LVEDd, % change MV DT, msec Right-sided CO, liter/min

o80 o0.6 o 7.7 o70 o0.8

100 100 100 100 100

100 78 86 60 67

0.99 0.95 0.95 0.93 0.81

0.0001 0.0001 0.0001 0.001 0.01

(50–100) (50–100) (50–100) (42–100) (40–100)

(60–94) (61–91) (68–94) (40–80) (47–81)

(0.98–1.00) (0.89–1.00) (0.89–1.00) (0.81–1.00) (0.67–0.96)

AUC, area under the curve; AV, aortic valve; CI, confidence interval; CO, cardiac output; LVEDd, left ventricular end-diastolic diameter; MV DT, mitral valve deceleration time. a Echo values cutoff represents the between 8,000 and 11,000 rpm.

is 96% according to national registry data.4 More recently published and anecdotal reports suggest the incidence of pump thrombosis may be on the rise.9 The incidence of pump thrombosis with LVAD malfunction is 5.9% at our institution. In addition to increasing mortality risk, pump thrombosis leads to prolonged hospital readmissions.28–30 Our experience indicates that patients can present with symptoms of HF, hemolysis, and less frequently, with overt pump arrest (only 1 of 12 patients in our study group), which mirrors the experience reported at other centers.6–8,12,19,21–23 However, patients can be asymptomatic and present only with signs of hemolysis or pump console parameter changes. In this particular scenario, screening for LVAD device malfunction, which may not be clinically obvious, remains important. Pump thrombosis occurring late after implant and after bleeding events with interruption of anticoagulation treatment has been reported.28 Although there were no significant differences in duration of pump support or INR level among our study and control patients, 6 of the 12 patients (50%) with suspected or confirmed pump thrombosis had an INR o 1.8, which is our institutionally defined anti-coagulation lower limit goal. One patient had an INR of 4.0 in the context of a sepsis syndrome, reflecting the possibility of developing pump thrombosis and LVAD malfunction with a supra-therapeutic INR (Table 2). Most of our cases were accompanied by recognizable perturbations in device console parameters. Of 10 patients with available LVAD console parameters and suspected or confirmed pump thrombosis in our cohort, 7 (70%) had an increase in power 4 10 watts at 11,000 rpm. However, we and others have observed that suspected or confirmed pump thrombosis with pump malfunction may occur in the absence of overt device alarms.7,21 Although our LVAD console findings demonstrate relatively higher pump powers in those with suspected or confirmed pump thrombosis compared with the control group, one cannot rely on pump power surges alone to make the diagnosis. Because the clinical presentation is non-specific, making a diagnosis of pump thrombosis and LVAD malfunction can be challenging. Post-implant adverse events, such as RV failure, significant aortic regurgitation, residual left-sided HF due to inadequate LV unloading, or gastrointestinal bleeding, may account for similar symptoms. Laboratory findings consistent with hemolysis significantly increase the

level of suspicion for device thrombosis, but imaging is invariably needed to document associated LVAD malfunction in the absence of overt pump arrest. At present, the choice of imaging strategy varies by institution. Cardiac CT scanning has emerged as a useful tool for detecting clots and obstruction in the LVAD inflow and outflow cannula but is unable to visualize intra-pump thrombus.8–11 Cardiac CT with contrast was diagnostic of significant outflow cannula thrombosis in 2 of 5 patients with suspected thrombosis who underwent CT examination (Figure 3). Although still the principal imaging modality in CF-LVAD patients, routine transthoracic echocardiography at baseline levels of LVAD support is unable to illustrate the dynamic interface between native ventricular and valvular function and CF-LVAD support. Echocardiographic testing only at baseline LVAD pump speeds was non-diagnostic of LVAD malfunction in our patients with suspected or confirmed pump thrombosis. Performing a systematic pump speed-change test overcomes this limitation. As LVAD pump speed and support is ramped up from 8,000 rpm, a series of changes are generally observed: LV dimensions decrease, AV opening time decreases (i.e., AV opening becomes partial, intermittent, or the AV remains closed), right-sided CO increases, and surrogates of LV filling pressures (e.g., mitral inflow DT) decrease. This expected response was seen in our sub-group of stable control patients but was markedly blunted among those with suspected or confirmed pump thrombus with associated device malfunction. This finding reinforces the principle that any obstruction to normal blood flow within the LVAD system in the form of thrombosis or kinking of the outflow cannula disturbs the ability to augment LV unloading, as demonstrated in an echocardiography-guided ramp study. Our results are in concordance with some of the findings recently reported by Uriel et al.12 Foremost, our and their speed-change protocols were highly feasible, overall safe, and well tolerated by patients. The algorithms were generally similar, with a few caveats. Our finding of an LDH 4 1,050 IU/liter is similar to that reported by Uriel et al,12 with the best LDL cutoff of 1,103 IU/liter associated with 100% sensitivity and 93% specificity as an indicator for device thrombosis. Haptoglobin has been less useful as a parameter because it is not uncommonly low, as demonstrated in our cohort of

–1.7 ⫾ 0.5 0.2 ⫾ 1.1 –0.1 ⫾ 2.1 o0.0001 0.3 o0.0001 6.3 ⫾ 0.9 4.5 ⫾ 0.9 8.7 ⫾ 0.9 5.6 ⫾ 0.5 4.8 ⫾ 0.5 7.7 ⫾ 1.1 4.7 ⫾ 0.3 4.5 ⫾ 0.6 5.7 ⫾ 0.7 3.9 ⫾ 0.4 5.3 ⫾ 1.1 4.5 ⫾ 0.3 a Data b Data c Data d

are shown as mean ⫾ standard deviation. available for 10 of 12 patients. derived from 10 of 36 patients. The p-value compares 8,000 vs 11,000 rpm in control vs pump thrombosis cohorts.

–3.0 ⫾ 0.2 0.9 ⫾ 1.6 –4.7 ⫾ 0.9 o0.0001 0.4 o0.0001 7.1 ⫾ 0.1 3.8 ⫾ 1.5 11.6 ⫾ 2.9 5.1 ⫾ 0.6 3.4 ⫾ 1.4 8.2 ⫾ 3.4 Flow Pulsatility index Power

4.6 ⫾ 0.8 4.0 ⫾ 1.7 7.1 ⫾ 3.3

6.5 ⫾ 0.9 3.4 ⫾ 1.4 9.6 ⫾ 3.3

Δ 8 vs 11,000 rpm p-value 11,000 rpm 10,000 rpm 9,000 rpm 8,000 rpm 9,000 rpm Variablea

8,000 rpm

10,000 rpm

11,000 rpm

p-value

Δ8 vs 11,000 rpm

Control cohortc (n ¼ 36) Suspected/confirmed pump thrombosis cohortb (n ¼ 10)

Left Ventricular Assist Device Console Findings Table 5

o0.0001 0.1 o0.0001

Echo Assessment of LVAD Pump Thrombosis p-value Δd

Estep et al.

583 stable patients receiving CF-LVAD support. Distinct differences between our and the Uriel et al12 protocol include the following: We examined 13 vs 6 echo variables and studied these echo parameters at 4 vs 11 pump speed settings (i.e., they increased speeds to as high as 12,000 rpm and in increments of 400 rpm). Of the 47 patients in our cohort who tolerated the ramp testing up to 11,000 rpm, only 2 (4%) had an uncomplicated LVAD suction event. In contrast, Uriel et al12 reported in 37 patients that the mean speed at termination of the ramp test was 10,573 ⫾ 856 rpm, with 22 suction events. In their cohort of 9 patients with device malfunction, pump speed ramp testing revealed that an attenuated reduction in LVEDd (slope 4 –0.16) was the strongest echocardiographic measurement diagnostic of pump thrombosis or outflow cannula bend relief disconnect. Our results extend the Columbia study findings by showing that apart from measuring the change in LVEDd (o0.6 cm) or absolute percentage change, the addition of changes in AV opening time (o80 msec) and mitral inflow DT (o70 msec) potentially provides a more robust echo diagnostic criteria to detect device malfunction. It is important to recognize that our patients had a wide range of baseline LVEFs, which can influence the degree of AV opening at higher pump speeds. Overall, 5 patients had an LVEF 4 40%, 4 of these 5 patients were in the control cohort, and all 4 had complete AV closure at a pump speed of 11,000 rpm. Our findings have important clinical implications. First, a reliable and simplified echo study using easily measurable echocardiographic parameters independent of LVAD console parameter changes (pulsatility index and pump power changes are not included) can be performed. In our experience, a complete echo pump speed change optimization study can be more time consuming than a ramp test after the 3 major echo indices of LVEDd, AV opening time, and mitral inflow DT. Only 1 patient in our study cohort had moderate mitral regurgitation and only 1 had moderate aortic insufficiency at baseline levels of LVAD support. Given that the vast majority of patients did not have more than mild aortic insufficiency and/or mitral regurgitation, we did not observe significant reductions in valve regurgitation severity with pump speed-change testing. We believe increasing the pump speed in increments of 400 rpm remains important when optimizing the pump speed for subtle echo changes in AV opening, LV size, or mitral insufficiency severity. For those patients in whom device thrombosis is highly suspected, a simplified “rule-out” LVAD device malfunction protocol based on changes in LV size, AV opening time, and mitral inflow DT may be more helpful and efficient. Second, the protocol can be performed safely at the bedside or as an outpatient, obviating any need for patient transport (i.e., sending for cardiac CT), which may potentially permit the diagnosis of device malfunction at an earlier time. At our institution, the evaluation of these 3 variables by echo ramp-up testing is an additive diagnostic tool that does not replace clinical judgment and decision making. In addition to performing our echo pump speed-change protocol, we obtain a dedicated contrast CT study in keeping with recent recommendations to screen for outflow

584

The Journal of Heart and Lung Transplantation, Vol 33, No 6, June 2014

Figure 3 Clot seen in the inflow cannula (IC) and outflow cannula (OC) at the time of device exchange confirms pump thrombosis of the left ventricular assist device. (B) A computed tomography scan demonstrates the presence of pump thrombosis (*) in the same patient before device exchange. AA, ascending aorta; LV, left ventricle; RV, right ventricle.

cannula clot and kinking for patients with a serum creatinine r 1.5 mg/dl and glomerular filtration rate 4 60 ml/min/ 1.73 m2.9,31 Among those with clinical HF or renal failure secondary to pigment-induced nephropathy and laboratory evidence of hemolysis (LDH 4 1,000 IU/liter, decrement in hemoglobin, macroscopic or dipstick microhematuria in the absence of visible red cells by urine microscopy), anticoagulation is intensified and parenteral unfractionated heparin is given (goal partial thromboplastin time of 60–80 seconds). Anti-coagulation intensification includes changing to a full-dose aspirin daily and addition of dipyridamole 75 mg every 8 hours, while maintaining an INR level of 2 to 3. Glycoprotein IIb/IIIa inhibitors and thrombolytics are not used to treat suspected or confirmed pump thrombosis in our institution. For those patients with hemolysis without LVAD malfunction defined by echo (r1 parameter met), anti-coagulation is intensified, and patients are monitored in the hospital. Device exchange is performed those patients with persistent hemolysis and LVAD malfunction defined by echo (Z2 parameters met) despite augmentation in anti-coagulation. The technique used to perform device exchange was initially with a sternotomy; however, we use and now prefer the sub-costal approach. The average time from event to pump exchange in our cohort was 2 ⫾ 0.5 days. Of the 12 patients in our total cohort with device malfunction, 3 (25%) died without device exchange due to poor surgical candidacy: 1 had a large disabling ischemic stroke and renal failure, the second had severe multiorgan failure, including renal failure and sepsis syndrome, and the third patient rapidly developed multiorgan failure, including

cardiogenic shock and respiratory failure. Our observation is that pump thrombosis with hemolysis and LVAD malfunction detected by echo is best treated with early device exchange to minimize the development of progressive endorgan failure, stroke, or death. Study limitations include the following: 1. our retrospective case-control study design, which may have falsely increased the calculated sensitivity given the severity of disease in the cases; 2. a relatively small group of patients in the suspected or confirmed pump thrombosis group, including 2 patients without a direct post-mortem LVAD examination, albeit in the context of a highly suggestive clinical picture; 3. a limited number of cardiac cycles (3) used to quantify AV opening duration, which may be less accurate of a measurement compared with using a greater number of cardiac cycles in those patients with underlying intermittent AV opening; 4. our findings were derived from patients with the HeartMate II and may not be applicable to patients supported with other VADs; and 5. application of our echo findings (parameters and cutoff values) may not apply to patients presenting with less clinically severe pump thrombosis or those with an intermediate pre-test probability of pump thrombosis (i.e., hemolysis without renal failure and/or stroke and/or death or confirmed pump thrombosis). In conclusion, pump thrombosis carries significant morbidity and mortality in patients supported by CF-LVADs. Differences in transthoracic echo-derived indices, including

Estep et al.

Echo Assessment of LVAD Pump Thrombosis

LVEDd, AV opening time, DT of mitral inflow between the lowest (8,000 rpm) and highest pump speed setting (11,000 rpm) during an echo pump speed-change protocol, provides a high degree of accuracy in diagnosing LVAD device malfunction secondary to pump system thrombosis in this patient population. The detection of LVAD device malfunction is an important part of the algorithm to guide management. Pending validation in larger, multicenterbased studies, these echo parameters can potentially be easily used to detect LVAD malfunction and the need for device exchange in those patients on the more advanced end of the clinical spectrum of pump thrombosis. Importantly, our findings need to be confirmed and validated in patients with less clinically advanced pump thrombosis, including those with an intermediate pre-test probability for this potentially life-threatening post–CF-LVAD complication.

Disclosure statement None of the authors has a financial relationship with a commercial entity that has an interest in the subject of the presented manuscript or other conflicts of interest to disclose. J.D.E. has received consulting fees from Thoratec Corp. other authors have no conflicts of interest to report.

References 1. Hunt SA, Abraham WT, Chin MH, et al. 2009 focused update incorporated into the ACC/AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines: developed in collaboration with the International Society for Heart and Lung Transplantation. Circulation 2009;119:e391-479. 2. Taylor DO, Stehlik J, Edwards LB, et al. Registry of the International Society for Heart and Lung Transplantation: twenty-sixth official adult heart transplant report—2009. J Heart Lung Transplant 2009;28:1007-22. 3. Peura JL, Colvin-Adams M, Francis GS, et al. American Heart Association Heart Failure and Transplantation Committee of the Council on Clinical Cardiology; Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation; Council on Cardiovascular Disease in the Young; Council on Cardiovascular Nursing; Council on Cardiovascular Radiology and Intervention, and Council on Cardiovascular Surgery and Anesthesia. Recommendations for the use of mechanical circulatory support: device strategies and patient selection: a scientific statement from the American Heart Association. Circulation 2012;126:2648-67. 4. Kirklin JK, Naftel DC, Kormos RL, et al. Fifth INTERMACS annual report: risk factor analysis from more than 6,000 mechanical circulatory support patients. J Heart Lung Transplant 2013;32:141-56. 5. Feldman D, Pamboukian SV, Teuteberg JJ, et al. The 2013 International Society for Heart and Lung Transplantation Guidelines for mechanical circulatory support: executive summary. J Heart Lung Transplant 2013;32:157-87. 6. Paluszkiewicz L, Gürsoy D, Spiliopoulos S, et al. HeartMate II ventricular assist device thrombosis-an echocardiographic approach to diagnosis: can Doppler evaluation of flow be useful? J Am Soc Echocardiogr 2011;24:350.e1-4. 7. Bashir J, Cheung A, Kaan A, Kearns M, Ibey A, Ignaszewski A. Thrombosis and failure of a HeartMate II device in the absence of alarms. J Heart Lung Transplant 2011;30:1197-9. 8. Mishkin JD, Enriquez JR, Meyer DM, et al. Utilization of cardiac computed tomography angiography for the diagnosis of left ventricular assist device thrombosis. Circ Heart Fail 2012;5:e27-9.

585 9. Goldstein DJ, John R, Salerno C, et al. Algorithm for the diagnosis and management of suspected pump thrombosis. J Heart Lung Transplant 2013;32:667-70. 10. Estep JD, Stainback RF, Little SH, Torre G, Zoghbi WA. The role of echocardiography and other imaging modalities in patients with left ventricular assist devices. JACC Cardiovasc Imaging 2010;3: 1049-64. 11. Raman SV, Sahu A, Merchant AZ, Louis LB 4th, Firstenberg MS, Sun B. Noninvasive assessment of left ventricular assist devices with cardiovascular computed tomography and impact on management. J Heart Lung Transplant 2010;29:79-85. 12. Uriel N, Morrison KA, Garan AR, et al. Development of a novel echocardiography ramp test for speed optimization and diagnosis of device thrombosis in continuous-flow left ventricular assist devices: the Columbia ramp study. J Am Coll Cardiol 2012;60:1764-75. 13. Estep JD, Chang SM, Bhimaraj A, Torre-Amione G, Zoghbi WA, Nagueh SF. Imaging for ventricular function and myocardial recovery on nonpulsatile ventricular assist devices. Circulation 2012;125:2265-77. 14. Lang RM, Bierig M, Devereux RB, et al. Chamber Quantification Writing Group; American Society of Echocardiography’s Guidelines and Standards Committee; European Association of Echocardiography. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440-63. 15. Quinones MA, Otto CM, Stoddard M, Waggoner A, Zoghbi WA. Recommendations for quantification of Doppler echocardiography: a report from the Doppler Quantification Task Force of the Nomenclature and Standards Committee of the American Society of Echocardiography. J Am Soc Echocardiogr 2002;15:167-84. 16. Nagueh SF, Appleton CP, Gillebert TC, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography. Eur J Echocardiogr 2009;10:165-93. 17. Zoghbi WA, Enriquez-Sarano M, Foster E, et al. Recommendations for the evaluation of the severity of native valvular regurgitation with twodimensional and Doppler echocardiography. J Am Soc Echocardiogr 2003;16:777-802. 18. Rudski LG, Lai WW, Afilalo J, et al. Guidelines for the echocardiographic assessment of the right heart in adults: a report from the American Society of Echocardiography endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography. J Am Soc Echocardiogr 2010;23:685-713. 19. Al-Quthami AH, Jumean M, Kociol R, et al. Eptifibatide for the treatment of HeartMate II left ventricular assist device thrombosis. Circ Heart Fail 2012;5:e68-70. 20. Sims DB, Takayama H, Uriel N, Gillam LD, Naka Y, Jorde UP. Ventricular assist device-associated thrombus. Circulation 2011;124: e197-8. 21. Bhamidipati CM, Ailawadi G, Bergin J, Kern JA. Early thrombus in a HeartMate II left ventricular assist device: a potential cause of hemolysis and diagnostic dilemma. J Thorac Cardiovasc Surg 2010; 140:e7-8. 22. Meyer AL, Kuehn C, Weidemann J, et al. Thrombus formation in a HeartMate II left ventricular assist device. J Thorac Cardiovasc Surg 2008;135:203-4. 23. Thomas MD, Wood C, Lovett M, Dembo L, O’Driscoll G. Successful treatment of rotary pump thrombus with the glycoprotein IIb/IIIa inhibitor tirofiban. J Heart Lung Transplant 2008;27:925-7. 24. John R, Kamdar F, Eckman P, et al. Lessons learned from experience with over 100 consecutive HeartMate II left ventricular assist devices. Ann Thorac Surg 2011;92:1593-9. 25. Miller LW, Pagani FD, Russell SD, et al. HeartMate II Clinical Investigators. Use of a continuous-flow device in patients awaiting heart transplantation. N Engl J Med 2007;357:885-96. 26. Slaughter MS, Rogers JG, Milano CA, et al. HeartMate II Investigators. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009;361:2241-51.

586

The Journal of Heart and Lung Transplantation, Vol 33, No 6, June 2014

27. Moazami N, Mohacsi P, Mooney M, et al. The 2013 International Society for Heart and Lung Transplantation guidelines for mechanical circulatory support: executive summary. J Heart Lung Transplant 2013;32:157-87. 28. Hasin T, Marmor Y, Kremers W, et al. Readmissions after implantation of axial flow left ventricular assist device. J Am Coll Cardiol 2013;61:153-63. 29. Smedira NG, Hoercher KJ, Lima B, et al. Unplanned hospital readmissions after HeartMate II implantation: frequency, risk

factors, and impact on resource use and survival. JACC Heart Fail 2013;1:31-9. 30. Rogers JG, Milano CA. Ramping up evidence-based ventricular assist device care. J Am Coll Cardiol 2012;60:1776-7. 31. Chang SM, Kassi M, Abu Sharifeh TM, et al. Diagnostic utility and safety of contrast enhanced gated multislices cardiac computed tomography (CT) for evaluation of patients with left ventricular assist device (LVAD). Circulation 2012;126:A18981 [abstr].