Journal of the American College of Cardiology
© 2005 by the American College of Cardiology Foundation
Published by Elsevier Inc.
Vol. 45, No. 9, 2005
ISSN 0735-1097/05/$30.00
doi:10.1016/j.jacc.2005.01.037
Left Ventricular Assist Device Malfunction
An Approach to Diagnosis by Echocardiography
Steven C. Horton, MD, FACC,†‡ Reza Khodaverdian, MD,* Peter Chatelain, BS RDCS,†
Marsha L. McIntosh, RDCS,† Benjamin D. Horne, MSTAT, MPH,* Joseph B. Muhlestein, MD,†‡
James W. Long, MD, PHD*‡
Salt Lake City, Utah
A protocol using transthoracic echocardiography was designed to diagnose the common
malfunctions of patients on chronic support with a left ventricular assist device (LVAD).
BACKGROUND Mechanical circulatory support, primarily with a LVAD, is increasingly used for treatment of
advanced heart failure as a bridge to transplant and for long-term treatment of heart failure.
The LVAD dysfunction is a recognized complication. To date, no studies have defined the
role of transthoracic echocardiography in evaluating long-term mechanical complications of
chronic LVAD support.
METHODS
Transthoracic echocardiography was used in a protocol designed to detect the common types
of mechanical malfunction. Patients were followed up with serial echocardiograms, and
clinical validations were made with findings from a catheter-based protocol and inspection at
the time of cardiac transplant or corrective surgery.
RESULTS
Thirty-two patients with 44 LVADs were followed up during a four-year period using this
protocol that correctly identified 11 patients with inflow valve regurgitation, 2 with
intermittent inflow conduit obstruction, 1 with severe kinking of the outflow graft, and 9 with
new insufficiency of the native aortic valve.
CONCLUSIONS As LVAD use for end-stage heart failure becomes widespread, and durations of support are
extended, dysfunction will be increasingly prevalent. Transthoracic echocardiography provides a practical method to accurately identify the causes of mechanical dysfunction with
patients on chronic LVAD support. (J Am Coll Cardiol 2005;45:1435– 40) © 2005 by the
American College of Cardiology Foundation
OBJECTIVES
Mechanical circulatory support, primarily with a left ventricular assist device (LVAD), is used increasingly to treat
advanced heart failure. The LVADs are able to bridge
patients with end-stage heart failure to cardiac transplantation (1,2). The Randomized Evaluation of Mechanical
Assistance for the Treatment of Congestive Heart Failure
(REMATCH) trial demonstrated that patients with New
York Heart Association functional class IV congestive heart
failure, who were ineligible for heart transplantation, received a longer-term survival benefit from the LVAD
compared with optimal medical therapy (3). Thus, LVAD
technology has also proven effective for long-term heart
failure treatment, referred to as destination therapy. Given
the substantial number of patients with end-stage heart
failure, it appears that LVAD usage is likely to increase.
The LVAD malfunction is an important cause of morbidity and mortality. Device failure was the second most
common cause of death in the REMATCH trial; at 24
months’ post-implant, 35% of patients suffered device failure (3). As cardiologists provide care for more LVAD
patients, it is important that they be able to troubleshoot a
malfunctioning device.
Diagnosis of LVAD component malfunction remains a
challenge. Diagnostic studies have not been standardized. A
From the *Division of Utah Artificial Heart Program and the †Department of
Cardiology, LDS Hospital, and the ‡University of Utah School of Medicine, Salt
Lake City, Utah.
Manuscript received September 19, 2004; revised manuscript received January 1,
2005, accepted January 11, 2005.
Downloaded From: https://content.onlinejacc.org/ on 09/30/2016
systematic catheter-based approach for the diagnosis of
LVAD system malfunction has been reported but not one
principally utilizing echocardiography (4).
Transesophageal echocardiography (TEE) is ideal for
defining LVAD dysfunction in both the pre-operative and
intra-operative setting (5–7). However, no studies have used
echocardiography, especially the less invasive transthoracic
echocardiography (TTE), to evaluate the long-term mechanical complications of discharged patients with chronic
LVAD support.
Therefore, we prospectively followed up with patients
discharged from the hospital with the HeartMate LVAD
(Thoratec Corp., Pleasanton, California) and performed
serial examinations with TTE to see if this noninvasive test
could identify the common types of LVAD dysfunction.
METHODS
We studied 35 patients (30 male; age 52 years [range 20 to
77 years]) undergoing implantation of the HeartMate VE
or XVE (Thoratec Corp.) LVAD between September 1999
and October 2003 at the LDS Hospital, of which, 26 were
as a bridge to transplant and 9 as destination therapy.
Forty-five LVADs were implanted with nine patients receiving a second LVAD and one patient receiving four.
Three patients underwent transplantation before echocardiography could be performed, 2 died after receiving repeat
LVAD implants, and 32 were followed with serial TTEs
(Table 1).
1436
Horton et al.
Diagnosis of LVAD Dysfunction by Echocardiography
Abbreviation and Acronyms
IVR
⫽ inflow valve regurgitation
LV
⫽ left ventricle
LVAD
⫽ left ventricular assist device
REMATCH ⫽ Randomized Evaluation of Mechanical
Assistance for the Treatment of
Congestive Heart Failure
TEE
⫽ transesophageal echocardiography
TTE
⫽ transthoracic echocardiography
VTI
⫽ velocity time integral
The mechanical functioning of the Thoratec Corp.
LVAD is described in Figure 1. Patients beyond the early
postoperative period were evaluated routinely with TTE,
generally every three months, and more frequently if mechanical dysfunction was suspected. Studies were performed
with the Hewlett-Packard Sonos 5500 (Andover, Massachusetts) and a S3 transducer. Standard transthoracic windows evaluated native heart function and anatomy, and
special measurements were performed of the LVAD components including the inflow cannula, outflow graft, and
native aortic valve, which can malfunction with the chronically indwelling LVAD.
In a properly aligned inflow cannula (Fig. 2), intracavitary
flow was considered normal if it was laminar and unidirectional. Inflow valve regurgitation (IVR) was defined as turbulent flow originating at the inflow cannula during LVAD
ejection (Fig. 3). A semiquantitative value was assigned
based on the area of turbulent flow seen within the left
ventricle (LV). Pulsed Doppler flow patterns delineated the
timing, direction of cannula flow, and flow variation relative
to the native cardiac cycle (Fig. 4).
Inflow valve regurgitation was quantified further by
assessing the flow through the outflow graft. Right parasternal views were used. Peak velocities, velocity time integral
(VTI), and the outflow graft diameter were measured (Fig. 5).
The stroke volume within the outflow graft was calculated
as the product of the area of the outflow graft and the VTI.
The stroke volumes of stable LVAD patients were compared with those in patients with IVR.
Inflow cannula obstruction was defined as interrupted
flow at the mouth of the inflow cannula occurring during
LVAD diastole. Outflow valve regurgitation was defined as
retrograde flow seen within the outflow graft occurring
during LVAD diastole. Outflow graft distortion was defined by an acceleration of Doppler velocities proximal in
the graft compared with the values measured more distally.
The native aortic valve was observed in the parasternal views
for thickening, systolic opening, and the presence of aortic
insufficiency by color flow Doppler.
Twelve (38%) of the patients followed up with echocardiography underwent 17 angiographic evaluations for suspected LVAD dysfunction. Echocardiograms were performed within a 30-day window of the angiographic studies.
All patients had their findings correlated at the time of
Downloaded From: https://content.onlinejacc.org/ on 09/30/2016
JACC Vol. 45, No. 9, 2005
May 3, 2005:1435–40
corrective surgery. Eighteen patients with stable functioning
LVADs underwent cardiac transplantation, and LVAD
components were examined at that time.
Statistical analysis Data for discrete variables are presented
as percentages with sample sizes, and data for continuous
variables are presented as mean with standard deviations or
mean with range in the case of time periods. For tests of
significance, the chi-square test was used for discrete variables and the t test was used for continuously distributed
variables.
RESULTS
Image quality. A total of 244 TTEs were performed with
71 TTEs utilizing the protocol. Of these studies, the inflow
conduit was imaged adequately in 68 (96%). In all patients,
the outflow graft velocities were obtained, but in one case
(2%), the outflow graft diameter could not be measured.
IVR. Eleven of the 42 LVADs (26%) had findings of
IVR, with 3 of the LVADs developing new IVR during the
study period. By the end of the study period, 9 of the 11
LVADs had progressed to having severe IVR. Dopplerechocardiography identified all eight patients with IVR by
angiography and confirmed by surgery to have severely
deformed inflow valves. Eighteen patients (56%) found to
be IVR-free on echocardiography had successful cardiac
transplantation. The inflow valves were inspected at explant.
Seventeen patients had normal inflow valves, and one had
only a minor gap between two inflow valve leaflets. Thus,
absence of IVR was correctly identified in 100% of cases.
Pulsed Doppler at the inflow cannula found significant
variability of the IVR flow relative to the native cardiac
cycle. The IVR waveforms were denser with higher velocities when they occurred during native left ventricular (LV)
diastole (Fig. 4).
Table 2 depicts the differences found in patients with
IVR compared with patients without IVR. Normal function
is associated with an outflow graft peak velocity of about 2.1
m/s and a calculated stroke volume of approximately 76 cc.
Outflow graft velocities, VTI, and stroke volume were all
significantly reduced in patients with IVR. Consistent with
a decompressed heart, LV diastolic dimensions were generally normal in patients without IVR, and significantly
dilated in patients with IVR.
With a poorly contracting native LV and with a normally
functioning LVAD allowing for decompression, the aortic
valve would be expected to open sparingly. Inflow valve
regurgitation was associated with frequent aortic valve
opening (65%) compared with 19% in patients without
IVR.
Outflow valve regurgitation. No cases of outflow valve
regurgitation were found in the 71 TTEs. Absence of
outflow regurgitation was confirmed in the 15 angiographic
studies, and no deformities of the outflow valve were seen at
surgery.
JACC Vol. 45, No. 9, 2005
May 3, 2005:1435–40
Horton et al.
Diagnosis of LVAD Dysfunction by Echocardiography
1437
Figure 3. Color flow Doppler showing inflow valve regurgitation.
Figure 1. Diagram of Thoratec left ventricular assist device (LVAD). The
Thoratec HeartMate LVAD initiates support at the left ventricular apex
with a cannula allowing blood to flow across a 25-mm porcine valve, into
the LVAD pumping chamber. After the pump fills, it ejects a volume of 83 cc.
Blood is ejected across a second 25-mm porcine valve into an outflow graft
with a distal anastomosis into the ascending aorta. The pump rhythm is
independent of native cardiac rhythm.
Inflow conduit obstruction. Two patients (6%) had significant obstruction to flow at the inflow conduit. In both
cases, the usual laminar LVAD diastolic inflow into the
apical cannula became intermittently interrupted (Fig. 6). In
both cases, the intermittent obstruction was not clinically
compromising.
Outflow graft distortion. One patient (3%) was diagnosed
with distortion of the LVAD outflow graft. Imaging of the
Figure 2. Properly oriented inflow cannula at the apex of the left ventricle.
Downloaded From: https://content.onlinejacc.org/ on 09/30/2016
outflow graft in the right parasternal view showed initial
normal velocities of 2.2 m/s, but with sitting forward
velocities increased to 4.96 m/s. At angiography, the outflow graft showed an acute angle bend that was confirmed at
the time of surgery.
Native aortic valve disease. Of the 28 patients without
aortic insufficiency before LVAD placement, 9 patients
(32%) had mild, insignificant, aortic insufficiency during
follow-up.
DISCUSSION
Transthoracic echocardiography correctly characterizes a
normally functioning LVAD and identifies multiple types
of LVAD dysfunction:
Normally functioning LVAD. This study shows that the
inflow cannula and its orientation within the LV as well as the
outflow graft can be imaged. A stable functioning LVAD is
generally associated with a normal-sized LV. Flow into the
apical cannula during LVAD filling is unidirectional and
Figure 4. Pulsed Doppler showing attenuation of inflow valve regurgitation flow during left ventricular (LV) systole and increased flow during LV
diastole. IVR ⫽ inflow valve regurgitation; LVAD ⫽ left ventricular assist
device.
1438
Table 1. Patient Demographics
Indication
1
2
3
1
2
4
Pre-Tx
DT
DT
Normal LVAD function
Normal LVAD function
Severe IVR and new AI
13
257
352
4
5
5
6
7
8
DT
Pre-Tx
DT
Normal LVAD function
Normal LVAD function
TDS
Moderate to severe IVR
681
54
755
531
N/A
N/A
Severe IVR
391
385
13
None performed
Normal LVAD function
Normal LVAD function
Normal LVAD function and new AI
Normal LVAD function
Mild IVR and AI
Normal LVAD function
Mild IVR
Normal LVAD function
Moderate to severe IVR and new AI
66
238
128
25
819
196
370
19
812
N/A
N/A
18
19
20
Pre-Tx
Pre-Tx
Pre-Tx
Pre-Tx
DT
Pre-Tx
Pre-Tx
Pre-Tx
DT
14
25
Pre-Tx
Normal LVAD function
443
Normal LVAD function
New AI, severe IVR
None performed
Normal LVAD function
Inflow cannula obstruction
Normal LVAD function
Severe IVR and new AI
639
512
Pre-Tx
Normal LVAD function
Severe IVR and new AI
Normal LVAD function
Severe IVR and new AI
Normal LVAD function
Pump malfunction and new AI
Normal LVAD function
Severe IVR and new AI
355
639
648
451
519
543
107
572
Pre-Tx
DT
Normal LVAD function
Normal LVAD function
Outflow graft distortion
412
729
421
9
6
7
8
9
10
11
12
15
16
17
18
19
13
14
15
16
17
28
30
34
35
36
38
39
24
41
42
43
44
50
51
52
53
25
26
58
59
60
20
21
22
23
DT
Pre-Tx
Pre-Tx
Pre-Tx
DT
Pre-Tx
Pre-Tx
Pre-Tx
Pre-Tx
Echo Findings
Downloaded From: https://content.onlinejacc.org/ on 09/30/2016
LVAD Duration
(d)
103
118
569
715
Time to AI
(d)
Time to IVR
(d)
205
N/A
N/A
205 to IVR
349
91
256
157
N/A
781
N/A
216
N/A
355 to mild;
470 to severe
N/A
N/A
378
N/A
139
638
54
N/A
275 to mild;
715 to severe
N/A
335
N/A
163
N/A
552
583
N/A
312 to mild;
533 to severe
N/A
N/A
Documentation/Clinical Correlation
Normal LVAD at inspection at time of transplant
Died of sepsis
Severe IVR at angiography and deformed inflow
valve at corrective surgery
N/A
Normal LVAD at inspection at time of transplant
Outflow graft kink and secondary pump failure
Severe IVR and deformed inflow valve at
corrective surgery
Severe IVR at angiography and deformed inflow
valve at corrective surgery
Died immediately postoperatively
Normal LVAD at inspection at time of transplant
Normal LVAD at inspection at time of transplant
Inspection at time of transplant
Inspection at time of transplant
N/A
Inspection at time of transplant
Angiography and surgical inspection
Inspection at time of transplant
Severe IVR at angiography and deformed inflow
valve at corrective surgery
Mechanical pump failure. Normal inflow valve at
angiography and corrective surgery
Patient alive and clinically stable
Angiography and surgical inspection
Expired immediately post-operatively
Normal LVAD at inspection at time of transplant
Normal LVAD at inspection at time of transplant
Normal LVAD at inspection at time of transplant
Severe IVR at angiography and deformed inflow
valve at corrective surgery
N/A
Severe IVR at angiography and surgical inspection
Inspection at time of transplant
Angiography and surgical inspection
Inspection at time of transplant
Angiography and surgical inspection
Normal LVAD at inspection at time of transplant
Severe IVR at angiography and deformed inflow
valve at corrective surgery
Normal LVAD at inspection at time of transplant
Normal LVAD at inspection at time of transplant
Acute angle bend in outflow graft at angiography
and surgical inspection
JACC Vol. 45, No. 9, 2005
May 3, 2005:1435–40
Study
No.
Horton et al.
Diagnosis of LVAD Dysfunction by Echocardiography
Patient
No.
Patient
No.
Study
No.
27
28
29
30
31
32
61
62
67
68
71
Indication
Pre-Tx
DT
Pre-Tx
Pre-Tx
Pre-Tx
Pre-Tx
Echo Findings
LVAD Duration
(d)
Normal LVAD function
Normal LVAD function
Inflow cannula obstruction
Normal LVAD function
Normal LVAD function and AI
Normal LVAD function
Normal LVAD function
265
105
834
170
210
671
207
Time to Al
(d)
Time to IVR
(d)
N/A
N/A
N/A
N/A
N/A
Documentation/Clinical Correlation
Died of sepsis
Normal LVAD at inspection at time of transplant
Prolapsing papillary muscle found at autopsy
Normal LVAD at inspection at time of transplant
Normal LVAD at inspection at time of transplant
N/A
Normal LVAD at inspection at time of transplant
JACC Vol. 45, No. 9, 2005
May 3, 2005:1435–40
Table 1 Continued
Average time to new AI ⫽ 297 d; average time to new IVR ⫽ 341 d.
AI ⫽ aortic insufficiency; DT ⫽ destination therapy; IVR ⫽ inflow valve regurgitation; LVAD ⫽ left ventricular assist device; Pre-Tx ⫽ pre-transplant.
Horton et al.
Diagnosis of LVAD Dysfunction by Echocardiography
1439
Figure 5. (A) Outflow graft imaged from the right parasternal view. (B)
Pulsed Doppler of the outflow graft showing normal flows.
laminar. Normal peak velocities and their associated stroke
volumes within the outflow graft were defined. Infrequent
opening of the aortic valve is seen in the stable LVAD,
consistent with nonejecting decompressed LV.
Figure 6. Pulsed Doppler of inflow cannula. Arrows indicate periods of
obstruction to flow from compression of the interventricular septum during
left ventricular systole.
Downloaded From: https://content.onlinejacc.org/ on 09/30/2016
1440
Horton et al.
Diagnosis of LVAD Dysfunction by Echocardiography
JACC Vol. 45, No. 9, 2005
May 3, 2005:1435–40
Table 2. Comparison of Patients With and Without IVR
Outflow graft peak velocity (m/s)
Outflow graft VTI (cm)
Outflow graft SV (cc)
LV diastolic size (mm)
AoV opening
Angiographically proven IVR
Free of IVR by angiography
Patients With
IVR
Patients Without
IVR
Significance
(p Value)
Threshold
for IVR
Sensitivity
Specificity
1.60 ⫾ 0.30
23.7 ⫾ 3.7
51.1 ⫾ 7.4
61 ⫾ 6
65%
100% sensitivity by
color Doppler
0% by color Doppler
2.12 ⫾ 0.39
35.9 ⫾ 7.9
76.5 ⫾ 15.1
49 ⫾ 10
19%
0% by color Doppler
⬍0.0001
⬍0.0001
⬍0.0001
⬍0.0001
0.0003
NA
ⱕ1.8
⬍30
⬍65
⬍55
NA
NA
89%
96%
100%
88%
NA
NA
84%
98%
97%
67%
NA
NA
NA
NA
NA
NA
100% IVR-free by
color Doppler
AoV ⫽ aortic valve; IVR ⫽ inflow valve regurgitation; LV ⫽ left ventricular; SV ⫽ stroke volume; VTI ⫽ velocity time integral.
IVR. Inflow valve regurgitation can be caused by a torn
cusp or commissural dehiscence of the prosthetic (porcine)
valve. This valve is under high mechanical stress as it
opposes high pump chamber pressures. The rigid pumping
chamber is unlike biological systems, which are compliant.
Hypertension and outflow graft twisting increase afterload
to the LVAD and may lead to inflow valve regurgitation (8).
Inflow valve regurgitation is the most common cause of
LVAD dysfunction and was associated with a nondecompressed, dilated LV, and frequent opening of the aortic
valve. Color flow Doppler directly visualizes IVR as a
turbulent flow originating at the inflow cannula during the
LVAD ejection period. Outflow graft flows are reduced
significantly in LVADs with IVR.
Pulsed Doppler of the inflow cannula shows significant
variability of the IVR flow in relation to the native cardiac
cycle Ejection of the LVAD into the low pressures of LV
diastole allows for a larger volume of IVR. When the
LVAD ejects during LV systole, the LV pressures are
higher, and a smaller regurgitant volume of IVR is ejected.
Native aortic valve distortion. The native aortic valve may
become fused and cause aortic stenosis or insufficiency (9).
Connelly et al. (10) examined hearts of patients with LVADs
and found commissural fusion was more common in patients
with VE HeartMates than pneumatic HeartMates (p ⬍
.0002). Hence, with effective LV decompression, aortic valve
opening is minimized leading to commissural fusion.
This observational study is limited by its moderate
number of patients but reflects the literature regarding the
types of LVAD malfunctions that have been reported. Our
practice has been to perform catheterization before corrective surgery to confirm LVAD dysfunction identified with
TTE. In cases of inadequate visualization with TTE, we
recommend TEE. All patients undergo TEE at the time of
corrective surgery.
CONCLUSIONS
Given the number of patients with end-stage heart failure
and the survival benefit from the LVAD, the number of
Downloaded From: https://content.onlinejacc.org/ on 09/30/2016
patients with long-term implants will likely grow. Component failure is an important cause of morbidity and mortality. This paper describes a protocol utilizing TTE that
correctly diagnosed a spectrum of common malfunctions in
patients with LVADs stabilized beyond the early postoperative period. TTE provides diagnostic imaging in
patients with LVADs confirmed by angiographic studies
and surgical inspection. Its diagnostic accuracy may eliminate the need for more invasive diagnostic procedures.
Reprint requests and correspondence: Dr. Steven C. Horton,
324 10th Avenue, Suite 206, Salt Lake City, Utah 84103. E-mail:
schorton@msn.com.
REFERENCES
1. Frazier OH, Rose EA, McCarthy P, et al. Improved mortality and
rehabilitation of transplant candidates treated with a long-term
implantable left ventricular assist system. Ann Surg 1995;222:327–
38.
2. Bank A, Sajad H, Nguyen D, et al. Effects of left ventricular assist
devices on outcomes in patients undergoing heart transplantation. Ann
Thorac Surg 2000;69:1369 –75.
3. Rose EA, Gellins AC, Moskowitz AJ, et al. Long-term use of a left
ventricular assist device for end-stage heart failure. N Engl J Med
2001;345:1435– 43.
4. Horton S, Khodaverdia R, Powers A, et al. Left ventricular assist
device malfunction: a systematic approach to diagnosis. J Am Coll
Cardiol 2004;43:1574 – 83.
5. Scalia G, McCarthy P, Savage R, Smedira N, Thomas J. Clinical
utility of echocardiography in the management of ventricular assist
devices. J Am Soc Echocardiogr 2000;13:754 – 63.
6. Ferns J, Dowling R, Bhat G. Evaluation of a patient with left
ventricular assist device dysfunction. ASAIO J 2001;696 – 8.
7. Akosah K, Song A, Guerraty A, Mohanty P, Paulsen W. Echocardiographic evaluation of patients with a left ventricular device. ASAIO
J 1998;44:M624 –7.
8. Poirier V. Inflow valve incompetence. J Congestive Heart Fail Circ
Support 2001;2:23–5.
9. Rose AG, Park SJ, Bank AJ, Miller LW. Partial aortic valve fusion
induced by left ventricular assist device. Ann Thorac Surg 2000;70:
1270 – 4.
10. Connelly J, Abrams J, Klima T, Vaughn W, Frazier O. Acquired
commissural fusion of aortic valves in patients with left ventricular
assist devices. J Heart Lung Transplant 2003;22:1291–5.