Download Predictors Of Outcomes In Patients With Ventricular Assist Devices

PREDICTORS OF OUTCOMES IN PATIENTS ON VENTRICULAR ASSIST
DEVICES
by
AMELIA K BOEHME
GERALD McGWIN, COMMITTEE CHAIR
NITA A. LIMDI
SALPY V. PAMBOUKIAN
EMILY B. LEVITAN
RUSSELL GRIFFIN
A DISSERTATION
Submitted to the graduate faculty of The University of Alabama at Birmingham,
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
BIRMINGHAM, ALABAMA
2014
Copyright by
Amelia Katharine Boehme
2014
PREDICTORS OF OUTCOMES IN PATIENTS ON VENTRICULAR ASSIST
DEVICES
AMELIA K BOEHME
EPIDEMIOLOGY
ABSTRACT
Ventricular assist devices (VAD) are used as a bridge to transplant and to increase
the quality of life in advanced heart failure patients. The use of the VAD is not without
risks itself, with an increased risk of thromboembolism, hemorrhage and death.
Therefore, we aimed to (1) evaluate predictors of thromboembolism; (2) design a score to
predict hemorrhages in VAD patients; and (3) evaluate the change in kidney function
post-VAD implant and to investigate the relationship between kidney function and
mortality in VAD patients using data collected from the University of Alabama at
Birmingham Mechanical Circulatory Support Clinic.
Over the 51.3 person-years of follow-up for the 115 participants, a total of 23 first
thromboembolic events were encountered (Incidence Rate (IR) 4.5 events /10 patient
years, 95% CI 29.1-66.2). There was an increased risk of thromboembolism with Early
lactate dehydrogenase elevation, and estimated glomerular filtration rate <30 prior to
VAD implantation while there was a decreased risk with good anticoagulation control.
Out of 115 patients, total of 36 patients experienced a hemorrhage during the first
year (Incidence Rate 70 per 100 person years (95%CI 50-96). Patients with a VAD
bleeding risk score ≥3 have the highest risk of hemorrhage compared to the 3 fold
increased risk of hemorrhage for patients with a HAS-BLED score of ≥ 3 and the 2.5 fold
increased risk with an extended HAS-BLED score of ≥3.
iii In the 228 patients, kidney function improves post implant with improvement in
kidney function maintained over 1 year for all patients except those with CKD Stage 5 at
baseline. Age at implant was the only statistically significant predictor of sustained
improvement in kidney function. Regardless of baseline CKD stage, most patients
experience an improvement in CKD stage after VAD implantation. Despite improvement
in kidney function post VAD implant, patients with CKD stage 3b, 4 or 5 prior to VAD
are at an increased risk of mortality post-VAD implantation.
In conclusion, this study identifies predictors of thromboembolism, creates a
VAD specific bleeding risk score and highlights how pre-VAD kidney function is an
important risk factor for mortality post-VAD.
Keywords: Ventricular assist device, mechanical circulatory support, thromboembolism,
hemorrhage, kidney function
iv DEDICATION
To Nate
v
ACKNOWLEDGEMENTS
I would like to take this opportunity to thank all of my colleagues, mentors,
family and friends for their support over the past four years. I am not able to mention
each of you individually, but rest assured, your support and kind words were greatly
appreciated.
First, I want to thank my research mentor, Dr. Nita Limdi. Without her support,
guidance and patient approach this project would not have become what it is today. I
have learned so much more than I could have ever anticipated working with Dr. Limdi
and I will forever be grateful for how it has shaped me as an epidemiologist. I also thank
my committee chair Dr. Gerald McGwin for his guidance on how to approach a scientific
question. I would like to thank Dr. Salpy Pamboukian and Dr. James George for their
guidance on the clinical implications of this research and for teaching me about this
unique patient population. I would like to thank Drs. Emily Levitan and Russell Griffin
for all of their feedback, support, and flexibility during this dissertation process. I have
learned so much from you both, in and out of the classroom. Finally, I would like to
thank the American Heart Association for grant funds used to complete this project.
I would like to say thank you to all my friends and family. To all of the
PhD students and post-doctoral fellows at UAB, thank you for the friendship, laughter
and camaraderie. Mom, Dad and my siblings, thank you for your love, support, laughter
and help through this process. Finally, I would like to thank my son Nate, to whom this
is dedicated.
vi
TABLE OF CONTENTS
Page
ABSTRACT.....................................................................................................................iii
DEDICATION................................................................................................................. v
ACKNOWLEDGEMENTS.............................................................................................vi
LIST OF TABLES...........................................................................................................viii
LIST OF FIGURES ........................................................................................................x
INTRODUCTION ...........................................................................................................1
PREDICTORS OF THROMBOEMBOLIC EVENTS IN VENTRICULAR ASSIST
DEVICE PATIENTS .......................................................................................................7
PREDICTORS OF HEMORRHAGE IN VENTRICULAR ASSIST DEVICE
PATIENTS.......................................................................................................................37
IMPROVEMENT IN KIDNEY FUNCTION AFTER VENTRICUALR ASSIST
DEVICE IMPLANTATION AND ITS INFLUENCE ON THROMBOEMBOLISM,
HEMORRHAGE, AND MORTALITY ..........................................................................64
SUMMARY CONCLUSIONS........................................................................................91
GENERAL LIST OF REFERENCES .............................................................................94
APPENDIX: IRB APPROVAL.......................................................................................100
vii
LIST OF TABLES
Table
Page
PREDICTORS OF THROMBOEMBOLIC EVENTS IN VENTRICULAR ASSIST
DEVICE PATIENTS
1
Demographic and Clinical Characteristics for the Entire Cohort Stratified by
Thromboembolism.....................................................................................................27
2
Anticoagulation Control for the Entire Cohort and Stratified by Thromboembolic
Events using an INR range of 2-3 and an INR range of 1.8-3.2................................29
3
Incidence of Thromboembolic Events in the First Year after Implantation ..............30
4
Association of Elevated LDH within the first month post VAD implantation and Risk
(Hazard ratio (HR) and 95% Confidence interval (95%CI)) of
Thromboembolism.....................................................................................................31
PREDICTORS OF HEMORRHAGE IN VENTRICULAR ASSIST DEVICE
PATIENTS
1
Baseline Characteristics of VAD Patients with and without a Hemorrhage ............58
2
Anticoagulation Control for the Entire Cohort ..........................................................60
3
Association of Bleeding Risk Scores with Risk (Hazard Ratio (HR) and 95%
Confidence Interval (95%CI)) of Hemorrhage ..........................................................61
IMPROVEMENT IN KIDNEY FUNCTION AFTER VENTRICUALR ASSIST
DEVICE IMPLANTATION AND ITS INFLUENCE ON THROMBOEMBOLISM,
HEMORRHAGE, AND MORTALITY
1
Demographic and Clinical Characteristics for the Entire Cohort Stratified by Kidney
Function prior to VAD Implantation .........................................................................83
2
Association of Kidney Function at Baseline with Risk (Hazard Ratio (HR) and 95%
Confidence Interval (CI)) of Thromboembolism and Hemorrhage...........................85
viii
Table
3
Page
Prior VAD Studies that Investigate Kidney Function post VAD Implantation and the
Relationship between Kidney Function and Mortality ..............................................86
ix
LIST OF FIGURES
Figure
Page
PREDICTORS OF THROMBOEMBOLIC EVENTS IN VENTRICULAR ASSIST
DEVICE PATIENTS
1
Inclusion and Exclusion Criteria Used to Establish the Cohort...........................33
2
Change in LDH Over Time for the Entire Cohort and Stratified by
Thromboembolism...............................................................................................34
3a Kaplan Meier Curve Illustrating Time to Thromboembolism Stratified by
Elevated LDH at VAD Implantation ...................................................................35
3b Kaplan Meier Curve Illustrating Time to Thromboembolism Stratified by
Elevated LDH at any Point during the First Month Post VAD Implantation......36
PREDICTORS OF HEMORRHAGE IN VENTRICULAR ASSIST DEVICE
PATIENTS
1
Inclusion and Exclusion Criteria Used to Establish the Cohort ..........................61
2
Comparing HAS-BLED to the Extended HAS-BLED to the VAD Bleeding Risk
Score in the Entire VAD Population....................................................................62
IMPROVEMENT IN KIDNEY FUNCTION AFTER VENTRICUALR ASSIST
DEVICE IMPLANTATION AND ITS INFLUENCE ON THROMBOEMBOLISM,
HEMORRHAGE, AND MORTALITY
1
Change in Kidney Function over 1 year follow-up from VAD Implantation..... 86
2a Relative Risk Estimates for Mortality Stratified by Baseline Kidney Function 88
2b Adjusted Relative Risk Estimates for Mortality Stratified by Baseline
CKD Stage ......................................................................................................... 89
x
INTRODUCTION
Advanced Heart Failure
Heart failure is one of the leading causes of morbidity and mortality in the US
with approximately 5.7 million Americans currently living with heart failure.(1, 2)
Treatment of HF ranges and is used to decrease symptom and disease progression as well
as increase patient survival.(3, 4) The best treatment options for end stage heart failure
are a heart transplant or ventricular assist device (VAD).(4, 5) VADs are mechanical
pumps, either pulsatile or continuous flow, that aide in ventricular function and assist in
circulation.(6) Older device designs were based on pulsatile flow, in which the pump
filled and emptied, with newer devices based on continuous flow principles, where blood
is constantly being propelled into the circulation.(3) The VAD was originally developed
in the 1960s as a bridge to recovery. The use of the VAD increased when the strategy for
device implantation expanded to include use as a temporary bridge to transplantation.
The expanded indication for VAD use increased the survival rates of advanced heart
failure patients and improved the outcomes of patients post transplantation.(6-18) While
heart transplantation is the gold standard in treating end stage HF, there are fewer than
3000 donor organs available per year. This facilitated the evolvement of VAD therapy to
include destination therapy, which is a permanent treatment and alternative to cardiac
transplantation as a long-term solution for treating heart failure. (3) The use of VADs as
a destination therapy increases the quality of life for end stage heart failure patients not
eligible for transplantation.(6, 19-24) VAD technology has been rapidly evolving over
1 the last few years with improvements in device durability; with these improvements
increasing the duration that patients with end-stage systolic heart failure can be supported
on VADs. This coupled with the increasing prevalence of end stage heart failure means
that VAD populations will increase, thereby increasing the survival rates of end stage
heart failure patients.
Mortality in VAD Patients
Despite the increased survival rates and increase in quality of life, there is still a
considerable amount of morbidity and mortality associated with VAD use.(20) VAD use
has increased the functional outcomes in advance heart failure patients, however there is
still an increased risk of mortality.(7, 25-28) Previous studies have estimated
approximately 50% of destination therapy VAD patients survive their first year, reducing
the risk of mortality by 48% compared to patients who are only on medication
management. Furthermore, approximately 23% of destination therapy VAD patients
survive to two years.(20) The current literature has focused on comparing VAD mortality
rates between VAD patients and medical management patients. Previous studies have
not been powered for sub-analysis, but have shown that younger people with VADs
placed are less likely to experience a mortality event than older people.(20)
Risk of Thromboembolism in VAD Patients
The increasing improvement in survival of VAD patients has led to recognition of
complications associated with VADs with approximately 5-11% of VAD patients
experiencing a thromboembolic event. (20, 27, 29-35) Thromboembolism is associated
with all VADs due to blood flowing over a non-biologic surface causing an increase in
platelet activation, thereby requiring combination antiplatelet/anticoagulant therapy to
2 reduce this risk.(3) Despite this intense antithrombotic regimen, VAD patients remain at
high risk for thromboembolic events such as pump thrombosis, venous thromboembolism
and stroke.(36) Risk factors for thromboembolism in VAD patients include impaired
renal function, increased BMI, pre-operative atrial fibrillation, and international
normalized ratio (INR) measurements below 1.5.(37, 38),(39) In addition to these risk
factors, VAD patients are at risk for hemolysis due to their device. Hemolysis is an early
warning sign for pump thrombosis, with elevated markers of hemolysis, associated with
increased risk of pump thrombosis.(40) One such marker, lactate dehydrogenase (LDH),
when elevated has been recently identified as a risk factor for pump thrombosis.(40, 41)
Risk of Hemorrhage in VAD Patients
In addition to thromboembolism, VAD patients are at increased risk of
hemorrhagic events with 15-60% of VAD patients experiencing a hemorrhagic event.(20,
27, 29-35). The combination antiplatelet/anticoagulant therapy predisposes a patient to
higher risks of hemorrhagic neurological complications.(42-47) Gastrointestinal
hemorrhagic events are problematic as well not only from the AC therapy, but also from
hypoperfusion of the bowel wall due to the continuous flow of blood created by the
VAD.(42, 48, 49) The risk for hemorrhagic events due to antiplatelet/anticoagulant
therapy and acquired von Willebrand disease are an increasing issue.(3, 45, 47, 50)
Previous studies have shown that after VAD implantation platelet activation is increased.
Institution of combination antiplatelet/anticoagulant therapy contributes to platelet
function restoration and clot mechanics restoration in patients with VADs. Additionally
acquired von Willebrand disease may ensue due shear stress from the device unfolding
large von Willebrand factor (v-WF) multimers and causing enzymatic activation,
3 resulting in breakdown of the multimer.(51-54) The levels of platelet inhibition
(measured by PFTs) and the level of anticoagulation (measured by INR and PTTR) have
not been well studied in VAD populations. Generally, device manufacturers recommend
INR ranges, with clinical decision-making at the discretion of the treating physicians.
There are not well-established algorithms for initiation and monitoring of
antiplatelet/anticoagulant therapy. Because of the ongoing risks associated with platelet
dysfunction there is a great need for the determination of the ideal regimen of thrombosis
prophylaxis.(52)
Kidney Function in VAD Patients
While chronic hemodialysis is a contraindication to VAD implantation, decreased
cardiac output induced renal insufficiency is not a contraindication with these patients
experiencing improved renal function post VAD placement. (6) This contributes to
patients experiencing decreased kidney function prior to VAD placement due to the
progression of heart failure with improvement in renal function after VAD
implantation.(3, 55) (56, 57) (58) Additionally, in order to be transplant eligible patients
must have a creatinine clearance of at least 50 ml/min.(58) Patients with VAD implants
demonstrate an improvement in kidney function following implantation. (55) However,
the bulk of the evidence is based on assessment of kidney function in the early (7-30
days) post-implant period. Smaller, previous studies have determined that the presence
of renal comorbidities prior to VAD placement have an impact on outcomes post VAD
placement, but the risk of subsequent redevelopment of renal dysfunction after VAD
placement is not well documented.(55, 59) Once the VAD is placed renal function begins
4 to improve, but pre-implant renal function remains a primary predictor of post implant
renal failure in addition to adverse events.(56, 57)
Implications for current research
The trade-off of decreased mortality rates for increased adverse events is
acceptable given the nature of end stage HF, but the frequency of adverse events
experienced by VAD patients creates an imminent need to understand and further
improve patient care to decrease these events. The influence of clinical factors and
changing end-organ function on outcomes in VAD patients needs to be addressed.
The research presented in this dissertation represents important contributions to
the literature on the topic of risk of thromboembolism, risk of hemorrhage and the
influence of kidney function on mortality post-VAD.
In the first manuscript we evaluate the association between traditional risk factors,
biomarkers of hemolysis and previously developed thromboembolism risk scores with
risk of thromboembolism in VAD patients. We assessed the traditional risk factors
kidney function prior to VAD implantation and level of anticoagulation control. Prior
studies investigated the relationship between the hemolysis biomarker LDH and pump
thrombosis. We investigated the relationship between LDH and risk of any type of
thromboembolism while accounting for the relationship between anticoagulation control
and kidney function. Finally, we investigated whether previously developed risk scores,
both continuous and categorized, were associated with thromboembolism in this high-risk
population.
In the second manuscript we evaluate the association between traditional risk
factors and previously developed hemorrhage risk scores with the risk of hemorrhage in
5 VAD patients. We assessed traditional risk factors including, but not limited to, age at
implant, history of gastrointestinal bleed prior to implant, history of atrial fibrillation
prior to implant and history of hyperlipidemia prior to implant. In addition to the
investigation into the association between previously developed risk scores, we
developed a VAD specific risk score for predicting hemorrhage post-implant.
Finally, in the third manuscript we assess change in kidney function over time
post-VAD implant, predictors of kidney function improvement post-implant, and the
association between kidney function prior to VAD and thromboembolism, hemorrhage
and mortality. Change in kidney function over time was assessed as a continuous change
in function as well as a categorical change in kidney disease.
6 PREDICTORS OF THROMBOEMBOLIC EVENTS IN VENTRICULAR ASSIST
DEVICE PATIENTS
by
AMELIA K BOEHME, SALPY V. PAMBOUKIAN, JAMES F. GEORGE,
JAMES K KIRKLIN, JOSE TALLAJ, CHRISLY DILLON, EMILY B. LEVITAN,
RUSSELL GRIFFIN, T. MARK BEASLEY, GERALD McGWIN, NITA A. LIMDI
In preparation for Circulation: Heart Failure
Format adapted for dissertation
7
Background: Ventricular assist device patients (VAD) are at increased risk for
thromboembolism. Biomarkers of hemolysis, such as lactate dehydrogenase (LDH), and
poorly controlled INR values have been identified as predictors of thromboembolism.
Methods and Results: Patients aged 19 years and older who had a continuous flow VAD
placed from 2006-2012 were included in this study (N=115). We assessed LDH
elevation (≥ 600 IU/L) on the day of VAD implant prior to the implant, and early LDH
elevation defined LDH elevation at VAD implantation, 14 days post implantation or 1
month post implantation; LDH measurements after a thromboembolism were excluded.
Over the 51.3 person-years of follow-up, a total of 23 first thromboembolic events were
encountered (Incidence Rate (IR) 4.5 events /10 patient years, 95% CI 29.1-66.2).
Patients with elevated LDH at VAD implantation had an increased risk for
thromboembolism (HR 4.72, 95% CI 1.44-15.4, p=0.0103). In a multivariable model
there was an increased risk of thromboembolism with Early LDH elevation (HR 4.95,
95%CI 1.69-14.4, p=0.003), and estimated glomerular filtration rate <30 prior to VAD
implantation (HR 4.74, 95%CI 1.12-20.1, p=0.0346) while there was a decreased risk
with good anticoagulation control, more than 60% of their time post VAD in the target
INR range of 1.8-3.2 (HR 0.30, 95%CI 0.10-0.86, p=0.0247).
Conclusion: Our study is the first to highlight the association between pre-VAD LDH
and post VAD thromboembolism. Furthermore, this study details the increased risk of
thromboembolism due to early LDH elevation after accounting for additional risk factors
in VAD patients.
Key Words: Ventricular assist device, thromboembolism, lactate dehydrogenase,
mechanical circulatory support
8
Introduction
Heart failure (HF) affects approximately 5.7 million Americans and is one of the
leading causes of morbidity and mortality in the US.(1-3) Although medical
management is the mainstay of therapy, heart transplantation or ventricular assist device
(VAD) implantation are increasingly used among patients with end stage HF.(4, 5)
While VAD implantation improves survival among those with end-stage HF,
approximately 5-11% of VAD patients experience a thromboembolic event within the
first year of implantation, with reports as high as 30-50%, despite advances in device
technology.(3, 6-11)_ENREF_1 To mitigate the increased risk of thromboembolism,
VAD patients are usually treated with dual antithrombotic therapy of an anticoagulant
(warfarin) and an antiplatelet (aspirin, dipyridimole, or clopidogrel or a combination).(3)
Despite this intense antithrombotic regimen, VAD patients remain at high risk for
thromboembolic events such as pump thrombosis, venous thromboembolism and
stroke.(12) These factors all contribute to the morbidity and mortality recognized in
VAD patients. The increasing incidence of thromboembolism, particularly pump
thrombosis, has been highlighted in recent reports.(13, 14)
Consistent with other at-risk populations, risk factors for thromboembolism in
VAD patients include impaired renal function, increased BMI, pre-operative atrial
fibrillation, and international normalized ratio (INR) measurements below 1.5.(14,
15),(16) In populations at high risk for thromboembolism, such as patients with atrial
fibrillation, the CHADS2 and CHA2DS2-VASc scores have been used to identify those at
greatest risk of thromboembolism.(17, 18) While these risk factors and scores are
informative in other populations requiring chronic anticoagulation, additional risk factors
9
need to be considered in VAD patients. As hemolysis frequently precedes pump
thrombosis, biomarkers associated with hemolysis can serve as an early warning for
impending pump thrombosis.(19) Elevated lactate dehydrogenase (LDH), a sensitive
marker for hemolysis, has been evaluated to assess the increased risk of pump
thrombosis.(20) However, previous investigations have been limited to cross-sectional
assessments.
At our institution, the routine measurement of LDH facilitates our presentation of
the change in LDH over a 1 year follow-up and enables evaluation of its association
thromboembolic events.
Methods
Study Setting and Inclusion and Exclusion
This study was conducted at the University of Alabama at Birmingham (UAB)
under the approval of the Institutional Review Board. Patients aged 19 years and older
who had a continuous flow VAD (HeartMate II (Thoratec Corporation, Pleasanton, CA)
or HeartWare (HeartWare Inc, Framingham, MA) device) placed at UAB from 20062012 were included in this study. Patients were followed for 1-year post VAD
implantation to assess outcomes. Loss to follow up is minimal since both inpatient and
outpatient care occurs at UAB with routine clinic visits at least monthly.
Data collection
10
For all patients, a detailed baseline (pre-VAD) clinical phenotype including
demographics (e.g. gender, race, ethnicity etc.), medical history before VAD (e.g. history
of medical conditions, surgeries prior to VAD implant etc.), medications (e.g.
antithrombotic medications and medications that influence thromboembolic events), and
laboratory assessments (e.g. coagulation factors, liver function tests, kidney function tests
etc.) was documented by a research assistant through retrospective medical record
review. Post-VAD documentation included medications, laboratory assessments and
outcomes. These were collected through medical records using definitions established by
the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS)
Registry.(21)
LDH Monitoring and Assay
Lactate dehydrogenase was measured in VAD patients on the day of VAD
implantation prior to the implant, and then 2 weeks ± 2 days, 1 month ± 2 days, 3 months
± 7 days, 6 months ± 7 days, 9 months ± 7 days and 12 months ± 7 days after VAD
implantation. All samples were collected from a peripheral blood draw with special
attention to specimen handling to avoid hemolysis. LDH levels were assessed by an
enzymatic rate method, (22) that measures the change in absorbance at 340 nanometers
(Beckman Coulter DCX 800 pro with SYNCHRON system, Beckman Coulter, Inc.,
Brea, CA) which is directly proportional to the activity of LDH in the sample. For adults,
the reference range at UAB hospital laboratories is 120-240 IU/L for LDH. Based on the
sensitivity and specificity of LDH as a predictor of thromboembolism, we defined
11
elevated LDH as ≥ 600 IU/L, which is consistent with prior research on LDH levels in
VAD populations.(20, 23)
Additional Risk Factors for Thromboembolism
We also evaluated the influence of diabetes, atrial fibrillation, anticoagulation
control, decreased kidney function, and thromboembolism risk scores on risk of
thromboembolism. Factors with p-values <0.2 were included in multivariable
analysis.(24)
Among patients on warfarin, the proportion of time spent in target INR range
(PTTR) is used as a measure of anticoagulation control. Poor anticoagulation control
(low PTTR) is a risk factor for thromboembolism and hemorrhage. For patients on
warfarin, the PTTR was estimated for each patient using the Rosendaal linear
interpolation method.(25) This method, which assumes a linear relationship exists
between two consecutively measured INR values, allows one to allocate a specific INR
value to each day for each patient. Time in target range for each patient was assessed by
the percentage of interpolated INR values within the target range of 2.0–3.0 after
attainment of first INR in target range. We also assessed proportion of time spent in
extended target range (PTTRe; INR1.8-3.2) since deviations beyond this range usually
trigger dose adjustments to minimize the increased risk of thrombosis and hemorrhage.
PTTRe was categorized into Good control>60%, Moderate control ≥50<60% and Poor
control <50%.
Kidney function was assessed through the estimated glomerular filtration rate
(eGFR) using the Chronic Kidney Disease Epidemiology Collaboration Formula (CKD-
12
EPI)(26) which incorporates gender, race, age, and serum creatinine (Scr; mg/dL).
Kidney function was characterized with eGFR >60 ml/min/1.73 m2 considered to be
no/mild kidney disease (CKD stage 1, 2); those with eGFR =30-59 ml/min/1.73 m2 (CKD
stage 3) considered to be moderate kidney disease and those with eGFR<30 ml/min/1.73
m2 (or on dialysis) considered to be severe kidney disease (CKD stage 4, 5).(27, 28)
The thromboembolism risk scores, CHADS2 and CHA2DS2-VASc, were assessed.
The CHADS2 risk score is a clinical prediction score ranging from 0- 6 for estimating the
risk of thromboembolism in patients with atrial fibrillation. The score consists of 1 point
for congestive heart failure, 1 point for hypertension, 1 point for age ≥75, 1 point for
diabetes and 2 points for prior history of stroke, transient ischemic attack or
thromboembolism. The CHA2DS2-VASc risk score is an updated version of the
CHADS2 score to include additional stroke risk factors in the clinical prediction score.
This updated scoring mechanism ranges from 0-10 and consists of 1 point for congestive
heart failure, 1 point for hypertension, 2 points for age ≥ 75, 1 point for diabetes, 2 points
for prior history of stroke/transient ischemic attack/thromboembolism, 1 point for
vascular disease, 1 point for age 65-74, and 1 point for female sex. The CHADS2 and
CHA2DS2-VASc score were evaluated as continuous variables and categorized into
dichotomous risk groups.(17, 18) The CHADS2 score was categorized into moderate risk
(CHADS2 ≥ 1) compared to high risk (CHADS2 ≥ 2). The CHA2DS2-VASc score was
categorized as moderate risk (CHA2DS2-VASc ≤3) compared to severe risk (CHA2DS2VASc ≥4).
Definition of Outcome
13
The primary outcome of interest was any thromboembolic event defined as pump
thrombosis requiring hospitalization, pulmonary embolus, deep vein thrombosis and
ischemic stroke. Pump thrombosis was defined as a clinically documented pump
thrombosis that required hospitalization for either medical management or pump
exchange. Ischemic stroke was defined as a clinically documented ischemic stroke event
diagnosed by a neurologist. DVT and pulmonary embolism were defined as a clinically
documented DVT or PE event. As patients may experience multiple events within the 1year follow-up period, only the first event was included in the analysis because the risk
factors for a second event are heavily influenced by the first event. A sub analysis was
conducted assessing early thromboembolism and late thromboembolism. An early
thromboembolic event was defined as thromboembolism occurring within the first 30
days post-VAD implant. A late thromboembolic event was defined as defined as
thromboembolism occurring after 30 days post-VAD implant.
Statistical Analysis
The χ2 test of independence was used to assess group differences for baseline
demographic categorical variables and Wilcoxon Rank Sum for continuous variables.
We evaluated whether elevated LDH at VAD-implantation was associated with risk of
thromboembolism and early thromboembolism; and whether elevated LDH at 1-month
was associated with risk of late thromboembolism. As early elevation of LDH (within the
first month) is considered a risk factor for pump thrombosis, patients exhibiting elevation
14
at VAD implantation, 14 days post implantation or 1 month post implantation were
categorized as Early LDH Elevation. 14 Only LDH measurements prior to early
thromboembolic events were included. Patients without this measurement were excluded
from the analyses (N=18).
Repeated measures ANOVA was used to assess change in LDH over time among
patients who experienced vs. did not experience thromboembolism. Kaplan-Meier curves
with log rank tests were used to assess the influence of LDH on time to the
thromboembolic event. Cox-proportional hazard modeling was conducted to evaluate the
influence of LDH on risk of thromboembolism after adjustment for other factors. Log-log
survival plots were used to assess the proportional hazards assumption. Patients were
censored at the time of the first hemorrhage event, explantation (due to death, recovery or
transplant) or at 1 year after VAD implantation. All statistical analyses were conducted
using SAS version 9.3 (SAS Institute, Cary, NC) at a non-directional alpha of 0.05.
Results
Of the sample of 127 patients with continuous flow VAD implants at treated at
UAB between 2006 and 2012, a total of 115 patient the HeartMate II device (Thoratec
Corporation, Pleasanton, CA) or the HeartWare (HeartWare Inc, Framingham, MA) were
included in this study (Figure 1). Twelve patients implanted outside UAB were excluded
as early data on anticoagulant controls and warfarin dosing was not available.. The mean
age at implant was 52 years (±14.6), and the majority of patients were male (78.3%),
white (67.8%) and implanted as a bridge to transplant (56.6%; Table 1).
15
Of the 115 patients in this study, 23 (20%) experienced a thromboembolism
during the first year. A total of 11 thromboembolisms occurred within the first 30 days
post-VAD implant with 12 thromboembolisms occurring more than 30 days post-VAD
implant. Seven (30%) of the 23 experienced a second thromboembolic event and 7
(30%) of the 23 experienced a hemorrhage after the thromboembolic event within the
first year. Patients experiencing thromboembolic events within the first year had a lower
frequency of diabetes (21.7% vs 43.5%, p=0.031), atrial fibrillation (17.4% vs 42.4%,
p=0.016) and a higher frequency of decreased kidney function (eGFR <30
ml/min/1.73m2) at VAD implantation (19.1% vs 6.7%, p=0.0136).
The 115 patients contributed 624.5 months (51.3 person years) of follow-up time.
. Patients were seen at least once a month, with an average of 1.4 (± 0.96) visits per
patient per month (Table 2). Patients with a thromboembolism were seen more
frequently each month compared to those who did not have a thromboembolism (2.7
times vs 1.8 times, p<0.0001) in the pre-thromboembolism time period.
Anticoagulation control, measured as percent time in target range, was similar between
patients with and without thromboembolism regardless of whether the standard INR
range was used or the extended INR range. At the time of thromboembolic event, the
average INR was 2.4 (SD 0.91). INR was sub-therapeutic among 4 patients at the time of
the event.
A total of 56 patients had LDH measured at VAD implantation, and 7 (12.5%) of
these patients had elevated LDH. There were 97 patients who had an Early LDH
measurement, within the first month of VAD implantation, with 17 (17.5%) of these
16
patients having an elevated LDH measurement.
As illustrated in Figure 2, the mean
LDH increases over the first14 days. Patients who experienced a thromboembolism in
the first year post implant continue to have increased mean LDH levels at 1 month,
compared to patients without a thromboembolic event. Patients with a thromboembolic
event continued to have a higher mean LDH than those without a thromboembolic event,
and this pattern remains throughout the follow-up after the thromboembolic event
(p=0.0028).
Association of LDH and Thromboembolism
Over the 51.3 person-years of follow-up (Table 3), a total of 23 first
thromboembolic events were encountered (Incidence Rate (IR) 4.5 events /10 patient
years, 95% CI 29.1-66.2). Patients with elevated LDH at VAD implantation experienced
a shorter time to thromboembolism (p=0.0046; Figure 3a), and an almost 5 fold increase
in risk for thromboembolism (Hazard Ratio (HR) 4.72, 95% Confidence Interval (CI)
1.44-15.4, p=0.0103). Furthermore, elevated LDH at VAD implantation was associated
with an increased risk of early thromboembolism (HR 5.55, 95%CI 1.32-23.3, p=0.0193).
There was an association with elevated LDH at VAD implantation and late
thromboembolism however this association was not statistically significant (HR 3.37,
95%CI 0.37-30.3, p=0.28). The small sample of LDH measurements at VAD
implantation limited our ability to adjust for additional relationships.
Patients with Early LDH Elevation experienced a shorter time to thromboembolic
events (Figure 3b; p=0.0198). There was an increased risk of thromboembolism
17
associated with Early LDH Elevation, with this association remaining after adjusting for
kidney function at baseline and percent time in target INR range 1.8-3.2 post implant (HR
4.95, 95%CI 1.69-14.4, p=0.003; Table 4). Early LDH Elevation was associated with
late thromboembolisms (HR 6.38, 95%CI 1.32-30.8, p=0.0211).
Additional Risk Factors for Thromboembolism
Significant differences in baseline kidney function, percent time spent in target
INR range of 1.8-3.2, history of atrial fibrillation, and history of diabetes between those
with and without thromboembolism were further evaluated.
Patients with an eGFR <30 prior to VAD implantation had an increased risk for
thromboembolism after adjusting for Early LDH Elevation and PTTRe (HR 4.74, 95%CI
1.12-20.1, p=0.0346; Table 4). Patients who have moderate anticoagulation control,
PTTRe 50-60%, had a reduced risk of thromboembolism, even after adjusting for Early
LDH Elevation and kidney function at baseline (HR 0.09, 95%CI 0.01-0.86, p=0.0363;
Table 4). Additionally, patients who have good anticoagulation control, PTTRe ≥60%,
had a reduced risk of thromboembolism after adjusting for Early LDH Elevation and
kidney function at baseline (HR 0.30, 95%CI 0.10-0.86, p=0.0247; Table 4).
Neither diabetes (HR 0.44, 95%CI 0.16-1.19, p=0.11), nor atrial fibrillation (HR
0.36, 95%CI 0.12-1.05, p=0.06) were significantly associated with thromboembolism.
CHADS2 as a continuous variable was not significantly associated with
thromboembolism (HR 0.82, 95%CI 0.50-1.34, p=0.42); a CHADS2 score ≥2was not
associated with the risk of thromboembolism (HR 0.70, 95%CI 0.29-1.71, p=0.44).
18
CHA2DS2-VASc as a continuous variable was not a statistically significant risk factor for
thromboembolism (HR 0.91, 95%CI 0.63-1.31, p=0.61); the categorized CHA2DS2VASc score ≥ 4 was not associated with the risk of thromboembolism (HR 0.89, 95%CI
0.39-2.02, p=0.78).
Secondary analysis of LDH as a Predictor of Ischemic Stroke and Pump Thrombosis
During the 51.3 years of follow-up the average time to first ischemic stroke was
118 days among the 6 patients with ischemic stroke (IR 11.7 events/100 patient years,
95% CI 4.7-24.3). Patients with LDH elevation at time of VAD implant were at an
increased risk of ischemic stroke (HR 19.8, 95%CI 1.79-218, p=0.0148). Patients with
Early LDH Elevation were at an increased risk of ischemic stroke (HR 6.6, 95%CI 1.3732.8, p=0.0211). There were 6 pump thrombosis events during the 51.3 years of followup (IR 11.7 events/100 patient years; 95% CI 4.7-24.3) with an average time to a first
pump thrombosis 64 days. Patients with LDH elevation at time of VAD implant had an
increased risk of pump thrombosis (HR 18.7, 95%CI 1.68-208, p=0.0172). Moreover,
patients with Early LDH Elevation were at an increased risk of pump thrombosis (HR
9.6, 95%CI 1.6-57.5, p=0.0134). However, because so few events occurred, no
adjustments to these models were possible.
Discussion
19
The concerning increase in the rate of thromboembolism in VAD patients with
newer devices illustrates the need for clinical factors that can identify patients at high risk
of a thromboembolic event.(14) The ability to detect a subgroup at high risk could enable
clinicians to more stringently monitor and potentially prevent thromboembolism in these
patients. The current study demonstrates that early LDH elevation is associated with an
increased risk of thromboembolism, ischemic stroke and pump thrombosis. Our study is
the first to highlight the association between pre-VAD LDH and post VAD
thromboembolism, indicating non-VAD related LDH elevation may identify a high risk
subgroup.
There is a dynamic hemostatic environment post VAD implant with considerable
variability in coagulation and fibrinolysis markers.(29) Despite this, the biomarker LDH
has been identified as a sensitive predictor of hemolysis, an indicator of pump
thrombosis.(20, 23) Furthermore, elevations of LDH in and of itself have been shown to
be associated with increased embolic events in VAD patients.(20) Our study is the first to
assess LDH prior to VAD implant as well as early LDH elevation as a primary predictor
of thromboembolism, ischemic stroke and pump thrombosis. We illustrate that early
LDH elevation is a predictor of not only pump thrombosis, but also all thromboembolic
events including ischemic stroke. The mean LDH remains higher over time for patients
with a thromboembolic event than for patients without a thromboembolic event, even
after the event occured. This finding suggests that some patients may have a higher risk
of hemolysis and subsequent thromboembolism due to cell and platelet injury from other
conditions in addition to the VAD itself. The association between elevated LDH at VAD
implant and early thromboembolism as well as thromboembolism at any time suggests
20
that elevated LDH could be a marker of tissue breakdown and fibrin clots affecting
platelets prior to VAD implantation.(30) This phenomenon, when combined with the
VAD could contribute to increased hemolysis thereby increasing the risk of
thromboembolism. Furthermore, consistent with previous studies, elevated LDH at 1month post implant is associated with late thromboembolic events.(14)
Inadequate anticoagulation has been identified as a primary risk factor for
thromboembolism, with the risk of thrombotic events increasing with INR values below
1.5.(14, 15) However, the findings from the current study indicate that some of these
thromboembolic events are occurring when the patients are adequately anticoagulated
based on current guidelines. Our study found that only 4 people had an INR below 1.5
leading up to and at the time of thromboembolic event, with 2 of these patients
experiencing their thromboembolic event within 2 days after VAD implant while being
treated with heparin. There is no difference in percent time below INR target range,
percent time in target INR range or percent time above target INR range leading up to
their thromboembolic event between patients who have a thromboembolism and patients
who do not have a thromboembolism. However, patients who spend more percent time
in the target INR range of 1.8-3.2 have a reduced risk of thromboembolism. Clinically
non-significant fibrin deposits on the pump can occur during periods of poor
anticoagulation and remain even once anticoagulation control has been achieved. These
minor deposits could dislodge at a later time regardless of if a patient achieved
anticoagulation control leading to a thromboembolic event. This suggests that more
emphasis should be place on early good anticoagulation control to prevent the buildup of
these fibrin deposits, thereby reducing the risk of thromboembolism.
21
The current study highlights that in addition to LDH and anticoagulation control,
kidney function is a strong predictor of thromboembolism. Kidney function is a known
risk factor for thromboembolism in other patient populations, with an increased risk of
thromboembolism for patients with an eGFR < 30.(31) Chronic kidney disease is
associated with an increased procoagulant profile, with these patients having higher
levels of fibrinogen, Factor VIII and vWF.(32) Patients who have a VAD placed typically
regain kidney function after implantation due to reversal of the cardiorenal syndrome;
however, the damage from the decreased kidney function in addition to the increased
procoagulant profile, could contribute to the increasing the risk of thromboembolism.
Other risk factors for thromboembolism in VAD patients include history of
diabetes, increased BMI and a history of atrial fibrillation;(14, 16) however, these factors
are not statistically significantly associated with thromboembolism in this study.
Furthermore, we assessed the CHADS2 and CHA2DS2-VASc scores as possible
predictors of thromboembolism in this high-risk patient population.(17, 18) These scores
are clinical prediction scores used in estimating the risk of stroke in patients with atrial
fibrillation. There was no relationship between these scores and risk of
thromboembolism in VAD patients despite these patients having high CHADS2 and
CHA2DS2-VASc scores. In this population, the increased risk of thromboembolism from
traditional risk factors may be overwhelmed by the hypercoagulable state induced by the
device.
Our study was not without limitations. We only included patients with HeartMate
II and HeartWare devices to reduce patient heterogeneity and increase the relevance to
22
current clinical practice. Although the UAB clinic treats a large number of HeartMate II/
HeartWare patients for a single center, the number of thromboembolic events was low,
thereby decreasing statistical power. We did not have enough information on serum free
hemoglobin over time to cross-reference serum free hemoglobin with LDH as an
exposure. Since this was a retrospective study we were limited by the data in the medical
records. LDH was not routinely collected as part of standard of medical care until
recently, thus many patients in the sample were missing LDH measurements at various
time points.
Our study is the first to highlight the association between pre-VAD LDH and post
VAD thromboembolism. Futhermore, we show that early LDH elevation, in addition to
kidney function prior to VAD implantation and post-VAD anticoagulation control, are
significant risk factors for thromboembolism. We found that the risk for
thromboembolism was 5 fold higher for patients with early LDH elevation, 5 fold higher
for patients with an eGFR <30 prior to VAD implantation and decreased for patients with
adequate anticoagulation control. Prospective studies are needed to further evaluate these
associations in larger VAD populations.
23
Reference List
1.
Roger VL, Go AS, Lloyd-Jones DM, Adams RJ, Berry JD, Brown TM, et al.
Heart disease and stroke statistics--2011 update: a report from the American Heart
Association. Circulation. 2011;123(4):e18-e209.
2.
Jessup M, Albert NM, Lanfear DE, Lindenfeld J, Massie BM, Walsh MN, et al.
ACCF/AHA/HFSA 2011 survey results: current staffing profile of heart failure
programs, including programs that perform heart transplant and mechanical
circulatory support device implantation: a report of the ACCF Heart Failure and
Transplant Committee, AHA Heart Failure and Transplantation Committee, and
Heart Failure Society of America. Circ Heart Fail. 2011;4(3):378-87.
3.
Wilson SR, Givertz MM, Stewart GC, Mudge GH, Jr. Ventricular assist devices
the challenges of outpatient management. J Am Coll Cardiol. 2009;54(18):164759.
4.
Haeck ML, Hoogslag GE, Rodrigo SF, Atsma DE, Klautz RJ, van der Wall EE, et
al. Treatment options in end-stage heart failure: where to go from here? Neth
Heart J. 2011.
5.
Torpy JM, Lynm C, Golub RM. JAMA patient page. Heart failure. JAMA.
2011;306(19):2175.
6.
Dang NC, Naka Y. Perioperative pharmacotherapy in patients with left ventricular
assist devices. Drugs Aging. 2004;21(15):993-1012.
7.
Frazier OH, Rose EA, Oz MC, Dembitsky W, McCarthy P, Radovancevic B, et
al. Multicenter clinical evaluation of the HeartMate vented electric left ventricular
assist system in patients awaiting heart transplantation. J Thorac Cardiovasc Surg.
2001;122(6):1186-95.
8.
Meyns B, Sergeant P, Herijgers P, Flameng W, Vanhaecke J, Van Cleemput J, et
al. Bridging procedures to heart transplantation. World J Surg. 2002;26(2):212-7.
Rose EA, Gelijns AC, Moskowitz AJ, Heitjan DF, Stevenson LW, Dembitsky W,
et al. Long-term use of a left ventricular assist device for end-stage heart failure.
N Engl J Med. 2001;345(20):1435-43.
9.
10.
Deng MC. The challenges of generating evidence to guide mechanical circulatory
support-based management of advanced heart failure. Eur Heart J.
2005;26(10):953-5.
11.
Tsukui H, Abla A, Teuteberg JJ, McNamara DM, Mathier MA, Cadaret LM, et al.
Cerebrovascular accidents in patients with a ventricular assist device. J Thorac
Cardiovasc Surg. 2007;134(1):114-23.
24
12.
John R, Kamdar F, Liao K, Colvin-Adams M, Miller L, Joyce L, et al. Low
thromboembolic risk for patients with the Heartmate II left ventricular assist
device. J Thorac Cardiovasc Surg. 2008;136(5):1318-23.
13.
Starling RC, Moazami N, Silvestry SC, Ewald G, Rogers JG, Milano CA, et al.
Unexpected abrupt increase in left ventricular assist device thrombosis. N Engl J
Med. 2014;370(1):33-40.
14.
Kirklin JK, Naftel DC, Kormos RL, Pagani FD, Myers SL, Stevenson LW, et al.
Interagency Registry for Mechanically Assisted Circulatory Support
(INTERMACS) analysis of pump thrombosis in the HeartMate II left ventricular
assist device. The Journal of heart and lung transplantation : the official
publication of the International Society for Heart Transplantation. 2014;33(1):1222.
15.
Boyle AJ, Russell SD, Teuteberg JJ, Slaughter MS, Moazami N, Pagani FD, et al.
Low thromboembolism and pump thrombosis with the HeartMate II left
ventricular assist device: analysis of outpatient anti-coagulation. The Journal of
heart and lung transplantation : the official publication of the International Society
for Heart Transplantation. 2009;28(9):881-7.
16.
Stulak JM, Deo S, Schirger J, Aaronson KD, Park SJ, Joyce LD, et al.
Preoperative atrial fibrillation increases risk of thromboembolic events after left
ventricular assist device implantation. The Annals of thoracic surgery.
2013;96(6):2161-7.
17.
Gage BF, van Walraven C, Pearce L, Hart RG, Koudstaal PJ, Boode BS, et al.
Selecting patients with atrial fibrillation for anticoagulation: stroke risk
stratification in patients taking aspirin. Circulation. 2004;110(16):2287-92.
18.
Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk
stratification for predicting stroke and thromboembolism in atrial fibrillation
using a novel risk factor-based approach: the euro heart survey on atrial
fibrillation. Chest. 2010;137(2):263-72.
19.
Shah P, Mehta VM, Cowger JA, Aaronson KD, Pagani FD. Diagnosis of
hemolysis and device thrombosis with lactate dehydrogenase during left
ventricular assist device support. J Heart Lung Transplant. 2013.
20.
Cowger JA, Romano MA, Shah P, Shah N, Mehta V, Haft JW, et al. Hemolysis: a
harbinger of adverse outcome after left ventricular assist device implant. The
Journal of heart and lung transplantation : the official publication of the
International Society for Heart Transplantation. 2014;33(1):35-43.
21.
Interagency Registry for Mechanically Assisted Circulatory Support
(INTERMACS) Official Website. Available from: http://www.intermacs.org/.
25
22.
Amador E, Dorfman LE, Wacker WE. Serum Lactic Dehydrogenase Activity: An
Analytical Assessment of Current Assays. Clin Chem. 1963;12:391-9.
23.
Shah P, Mehta VM, Cowger JA, Aaronson KD, Pagani FD. Diagnosis of
hemolysis and device thrombosis with lactate dehydrogenase during left
ventricular assist device support. The Journal of heart and lung transplantation :
the official publication of the International Society for Heart Transplantation.
2014;33(1):102-4.
24.
Bender R, Lange S. Adjusting for multiple testing--when and how? J Clin
Epidemiol. 2001;54(4):343-9.
25.
Rosendaal FR, Cannegieter SC, van der Meer FJ, Briet E. A method to determine
the optimal intensity of oral anticoagulant therapy. Thromb Haemost.
1993;69(3):236-9.
26.
Valente MA, Hillege HL, Navis G, Voors AA, Dunselman PH, van Veldhuisen
DJ, et al. The Chronic Kidney Disease Epidemiology Collaboration equation
outperforms the Modification of Diet in Renal Disease equation for estimating
glomerular filtration rate in chronic systolic heart failure. Eur J Heart Fail.
2014;16(1):86-94.
27.
Levey AS, Coresh J, Balk E, Kausz AT, Levin A, Steffes MW, et al. National
Kidney Foundation practice guidelines for chronic kidney disease: evaluation,
classification, and stratification. Ann Intern Med. 2003;139(2):137-47.
28.
Huang SM, Temple R, Xiao S, Zhang L, Lesko LJ. When to conduct a renal
impairment study during drug development: US Food and Drug Administration
perspective. Clin Pharmacol Ther. 2009;86(5):475-9.
29.
Majeed F, Kop WJ, Poston RS, Kallam S, Mehra MR. Prospective, observational
study of antiplatelet and coagulation biomarkers as predictors of thromboembolic
events after implantation of ventricular assist devices. Nat Clin Pract Cardiovasc
Med. 2009;6(2):147-57.
30.
Kumar R, Beguin S, Hemker HC. The effect of fibrin clots and clot-bound
thrombin on the development of platelet procoagulant activity. Thromb Haemost.
1995;74(3):962-8.
31.
Wattanakit K, Cushman M, Stehman-Breen C, Heckbert SR, Folsom AR. Chronic
kidney disease increases risk for venous thromboembolism. J Am Soc Nephrol.
2008;19(1):135-40.
Dubin R, Cushman M, Folsom AR, Fried LF, Palmas W, Peralta CA, et al.
Kidney function and multiple hemostatic markers: cross sectional associations in
the multi-ethnic study of atherosclerosis. BMC Nephrol. 2011;12:3.
32.
26
Table 1. Demographic and Clinical Characteristics for the Entire Cohort Stratified
by Thromboembolism
All Patients
(N=115)
Age at Implant
Body Mass Index at
Implant
Male
Black Race
Status in the first year
Deceased
Transplanted
Recovered
Etiology of Heart Failure
Ischemic
Idiopathic
Indication for VAD
Bridge to transplant
Bridge to candidacy
Bridge to recovery
Destination
Comorbid Conditions
Diabetes
Right Ventricular
Dysfunction
Coronary Artery
Disease
Myocardial Infarction
Hypertension
Hyperlipidemia
Atrial Fibrillation
Ventricular
Tachycardia
Thromboembolism
(N=23)
Mean (SD)
52± 15
28 ± 8
No
Thromboembolism
(N=92)
Mean (SD)
53 ±14
29±7
N (%)
90 (78.3%)
37 (32.2%)
N (%)
74 (80.4%)
29 (31.5%)
N (%)
16 (69.6%)
8 (34.8%)
19 (16.5%)
23 (20.0%)
4 (3.5%)
13 (14.1%)
20 (21.7%)
3 (3.3%)
6 (26.1%)
3 (13.0%)
1 (4.3%)
60 (53.1%)
45 (39.8%)
50 (55.5%)
35 (38.8%)
10 (43.5%)
10 (43.5%)
64 (56.6%)
2 (1.8%)
1 (0.9%)
46 (40.7%)
49 (54.5%)
2 (2.2%)
0
39 (43.3%)
15 (65.2%)
0
1 (4.3%)
7 (30.4%)
45 (39.1%)
28 (24.4%)
40 (43.5%)
21 (22.8%)
5 (21.7%)
7 (30.4%)
0.031
0.45
46 (40.0%)
37 (40.2%)
9 (39.1%)
0.92
23 (20.0%)
71 (61.7%)
43 (37.4%)
26 (22.6%)
19 (16.5%)
19 (20.7%)
59 (64.1%)
44 (47.8%)
39 (42.4%)
45 (48.9%)
4 (17.4%)
12 (52.2%)
8 (34.8%)
4 (17.4%)
10 (43.5%)
0.73
0.29
0.26
0.016
0.64
Mean (SD)
49 ±15
29±7
0.20
0.49
0.26
0.76
0.50
0.25
0.14
Concurrent Meds at VAD Implantation
92 (100%)
23 (100%)
Warfarin 115 (100%)
113
(98.2%)
90
(97.8%)
23 (100%)
Aspirin
67
(58.3%)
53
(57.6%)
14
(60.9%)
Plavix
6 (5.2%)
5 (5.4%)
1 (4.3%)
Colchicine
70 (60.9%)
54 (58.7%)
16 (69.6%)
Amiodarone
Estimated Glomerular Filtration rate (eGFR ml/min/1.73m2) prior to VAD implantation
47 (42.3%)
40 (44.4%)
7 (33.3%)
eGFR ≥60
54 (48.7%)
44 (48.9%)
10 (47.6%)
eGFR 30 to 59
10 (9.0%)
6 (6.7%)
4 (19.1%)
eGFR < 30
p-value
27
0.48
0.78
0.84
0.34
0.014
Estimated glomerular filtration rate (eGFR ml/min/1.73m2) was calculated using the
Chronic Kidney Disease Epidemiology Collaboration Formula (CKD-EPI).
Continuous variables were tested using Wilcoxon Rank Sum test and categorical
variables were tested using a Chi-squared test
28
Table 2. Anticoagulation Control for the Entire Cohort and Stratified by
Thromboembolic Events using an INR range of 2-3 and an INR range of 1.8-3.2
All
No
Thromboembolism
Thromboembolism
Mean(SD)
92
p-value
Number of Patients
115
23
Number of Visits
4902
4311
591
0.003
Total Follow Up (months)*
624.5
569.6
54.9
<0.0001
Follow Up Months/patient
5.4 ± 4.8
8.4 ±4.5
2.4 ±2.8
<0.0001
Number of
Visits/patient/month
2.1 ± 1.2
1.8 ±0.91
2.7 ± 1.4
<0.0001
Anticoagulation control for INR Range of 2-3
Percent Time Below Range
41.6 ± 28.4
41.8±25.9
42.7±36.7
0.46
Percent Time In Range
42.9 ± 22.5
43.6 ±20.4
42.2±28.9
0.95
Percent Time Above Range
15.8 ± 13.9
15.1±12.1
18.5±19.4
0.93
Anticoagulation Control for INR Range of 1.8-3.2
Percent Time Below Range
30.3 ± 28.5
29.1 ±26.9
35.5 ±34.5
0.77
Percent Time In Range
59.2 ± 26.4
61.1±24.1
54.6±32.2
0.61
Percent Time Above Range
11.1 ± 11.5
10.5±9.7
13.8±16.9
0.92
*Follow-up is calculated as time from VAD to first event, or in the circumstance of no
event the end of study period (1 year), explantation, transplantation or death
Continuous variables were tested using Wilcoxon Rank Sum test
29
Table 3. Incidence of Thromboembolic Events in the First Year after Implantation
All cause TE
Entire Sample (N=115)
TE Type
Ischemic Stroke
Pump Thrombosis
DVT
TIA
PE
Mediastinal Clot
N
Follow-up*
Incidence Rate (IR)
(95%Confidence Interval
(CI)) per 10 person years
23
51.3 years
4.5 (2.9-6.6)
6
6
5
2
2
2
51.3 years
51.3 years
51.3 years
51.3 years
51.3 years
51.3 years
1.2 (0.5-2.4)
1.2 (0.5-2.4)
1.0 (0.4-2.2)
0.4 (0.07-1.3)
0.4 (0.07-1.3)
0.4 (0.07-1.3)
*Follow-up is calculated as time from VAD to first event, or in the circumstance of no
event the end of study period (1 year), explantation, transplantation or death
30
Table 4. Association of Elevated LDH within the first month post VAD
implantation and Risk (Hazard ratio (HR) and 95% Confidence interval (95%CI))
of Thromboembolism
Crude Analyses
No Early LDH Elevation (<600 IU/L)
Early LDH Elevation(≥600 IU/L)
Hazard
Ratio
ref
95% Confidence
Interval
ref
p-value
2.99
1.14-7.88
0.027
ref
2
Estimated Glomerular Filtration rate (eGFR ml/min/1.73m ) prior to VAD implantation
eGFR ≥60
ref
ref
ref
eGFR 30 to 59
1.56
0.59-4.12
0.36
eGFR < 30
3.65
1.06-12.5
0.040
ref
ref
ref
Moderate Control
0.15
0.02-1.20
0.07
Good Control
0.37
0.16-0.87
0.0231
Hazard
Ratio
ref
95% Confidence
Interval
ref
p-value
4.95
1.69-14.4
0.003
Anticoagulation Control
Poor Control
Adjusted Analyses*
No Early LDH Elevation (<600 IU/L)
Early LDH Elevation(≥600 IU/L)
ref
Estimated Glomerular Filtration rate (eGFR ml/min/1.73m2) prior to VAD implantation
eGFR ≥60
ref
ref
ref
eGFR 30 to 59
1.12
0.37-3.37
0.83
eGFR < 30
4.74
1.12-20.1
0.035
ref
ref
ref
Moderate Control
0.09
0.01-0.86
0.0363
Good Control
0.30
0.10-0.86
0.0247
Anticoagulation Control
Poor Control
Patients exhibiting elevation at VAD implantation, 14 days post implantation or 1 month
post implantation were categorized as Early LDH Elevation. No LDH measurements that
occurred after a thromboembolic event were included in the categorization of elevated
31
LDH within the first month post-VAD implant. The proportion of time spent in target
INR range of 1.8-3.2 (PTTRe) was used as a measure of anticoagulation control. PTTRe
was categorized into Good control>60%, Moderate control ≥50<60% and Poor control
<50%.
*The full model includes Early LDH Elevation, eGFR at VAD implantation and
anticoagulation control
Cox Proportional Hazards Models were used to assess the relative risk of
thromboembolism. The proportional hazards assumptions were met.
32
Figure 1. Inclusion and Exclusion Criteria Used to Establish the Cohort
127 HeartMate II/HeartWare
patients implanted from
2006-2012
115 patients with
Anticoagulation medications
and lab values available for
enrollment in this study
23 pa&ents experienced a primary thromboembolic event 7 pa&ents go on to have a second thromboembolism within the first year 7 pa&ents go on to have a hemorrhage as their second event within the first year 33
12 patients who were
implanted at different
facilities were excluded due
to missing information on
anticoagulation at the time of
VAD implant
92 pa&ents did not have a primary thromboembolic event 9 pa&ents do not have a second event within the first year Figure 2. Change in LDH Over Time for the Entire Cohort and Stratified by
Thromboembolism
p=0.0028
Differences in changes in LDH over time stratified by thromboembolism were assessed
using repeated measures ANOVA
34
Figure 3a. Kaplan Meier Curve Illustrating Time to Thromboembolism Stratified
by Elevated LDH at VAD Implantation
p=0.005
35
Figure 3b. Kaplan Meier Curve Illustrating Time to Thromboembolism Stratified
by Early LDH Elevation
p=0.020
36
PREDICTORS OF HEMORRHAGE IN VENTRICULAR ASSIST DEVICE
PATIENTS
by
AMELIA K BOEHME, SALPY V. PAMBOUKIAN, JAMES F. GEORGE,
JAMES K KIRKLIN, JOSE TALLAJ, CHRISLY DILLON, EMILY B. LEVITAN,
RUSSELL GRIFFIN, T. MARK BEASLEY, GERALD McGWIN, NITA A. LIMDI
In preparation for the Journal of Heart and Lung Transplantation
Format adapted for dissertation
37
ABSTRACT
Background: Ventricular assist device (VAD) patients are at increased risk for
hemorrhagic events due to intensive antithrombotic therapies and device-induced platelet
shearing.
Methods: Patients aged 19 years and older who have had a continuous flow device placed
at UAB from 2006-2012 were included in this study. This study developed a VAD
Bleeding risk score and compared it with the HAS-BLED hemorrhagic risk score.
Results: A total of 36 patients experienced a hemorrhage during the first year (Incidence
Rate 70 per 100 person years (95%CI 50-96). The VAD Bleeding Risk score ranges
from 0-5 and included age greater than 65 (1 point), history of gastrointestinal bleed prior
to VAD implantation (1 point), ischemic etiology of heart failure (1 point), history of
hyperlipidemia prior to implantation (1 point), and eGFR <30 prior to VAD implantation
(1 point), with a Harrell’s adjusted c-statistic of 0.789. The VAD Bleeding risk score has
a better c-statistic than the HAS-BLED score, 0.789 compared to 0.661. Patients with a
VAD bleeding risk score ≥3 have the highest risk of hemorrhage (HR 4.61, 95%CI 2.378.99, p<0.0001) compared to the 3-fold increased risk of hemorrhage for patients with a
HAS-BLED score of ≥ 3 (HR 3.18, 95%CI 0.96-10.5, p=0.06).
Conclusion: Our study is the first to create a VAD-specific bleeding risk score. We
found that the simple, newly developed VAD Bleeding risk score can identify patients at
risk for developing a hemorrhage.
Key Words: Ventricular assist device, hemorrhage, mechanical circulatory support
38
Introduction
Approximately 5.7 million Americans are living with heart failure (HF), a leading
cause of morbidity and mortality, with the prevalence continuing to increase in the US.(13) The best treatment options for end-stage advanced HF is heart transplantation or
placement of a ventricular assist device (VAD),(4, 5) with the use of VAD therapy
increasing as a long-term solution for treating end-stage HF. (6-12) The design of VADs
has changed through the years; newer devices (HeartMate II and HeartWare) maintain
continuous flow, with blood constantly being propelled into the circulation, while older
devices produce pulsatile flow.(3)
VAD implantation improves the life expectancy and quality of life for advanced
HF patients; however, this treatment option has an increased risk of both
thromboembolism and hemorrhage. To mitigate the risk of thromboembolism, VAD
patients are treated with dual antithrombotic therapy, including an anticoagulant
(warfarin) and an antiplatelet medication (aspirin, dipyridomole, clopidogrel, or a
combination agent).(3) Dual antithrombotic therapy predisposes these patients to even
higher risks of hemorrhagic complications.(13-18) Further complicating the coagulation
abnormalities of these complex patients, continuous-flow devices have been found to
result in an acquired form of von Willebrand syndrome, resulting in a higher risk for
hemorrhage.(19) Due to the increase in the number of VAD implants, the risk for
hemorrhagic events due to antithrombotic therapy and acquired von Willebrand disease is
becoming an increasingly pressing issue.(3, 16, 18, 20)
39
The risk of bleeding while on chronic anticoagulant therapy has been well studied
in other populations. The development of the novel bleeding risk score HAS-BLED
offers clinicians a method to establish the bleeding risk of patients with atrial fibrillation
on chronic anticoagulation.(21) Specific components in the HAS-BLED score, such as
age, have also been identified as risk factors for post-implantation bleeding in the VAD
population.(22) The HAS-BLED score, however, has not been validated in this high risk
population on chronic anticoagulation. Prior research has shown that risk factors for
hemorrhagic events in VAD patients, such as gender and etiology of heart failure, are not
included in the HAS-BLED score.(22)
We present data on predictors of hemorrhagic events in continuous flow VADs
for 115 patients implanted with HeartMate II and HeartWare devices. This study
evaluates risk factors for hemorrhagic events and the utility of the HAS-BLED score in
this high risk population. Furthermore, we developed a simple hemorrhagic risk score to
identify patients at highest risk for a hemorrhagic event while implanted with a VAD.
Methods
Study Setting and Inclusion and Exclusion
This study was conducted at the University of Alabama at Birmingham (UAB)
under the approval of the Institutional Review Board. Patients aged 19 years and older
who had a continuous flow VAD (HeartMate II (Thoratec Corporation, Pleasanton, CA)
or HeartWare (HeartWare Inc, Framingham, MA) device) placed at UAB from 20062012 were included in this study. Patients were followed for 1-year post VAD
40
implantation to assess outcomes. Loss to follow up is minimal since patients receive both
inpatient care and routine monthly outpatient clinic visits at UAB.
Data collection
For all patients, a detailed baseline (pre-VAD) clinical phenotype was
documented by a research assistant through retrospective medical record review. These
data included demographics (e.g. gender, race, ethnicity etc.), medical history before
VAD (e.g. history of medical comorbidities, surgeries prior to VAD implant etc.),
medications (e.g. antithrombotic medications and medications that influence
thromboembolic events), and laboratory assessments (e.g. coagulation factors, liver
function tests, kidney function tests etc.). Post-VAD documentation included
medications, laboratory assessments and outcomes. These were collected through
medical records using definitions established by the Interagency Registry for
Mechanically Assisted Circulatory Support (INTERMACS) Registry.(23)
Definition of Exposures
We evaluated the influence pre-existing history of hypertension, atrial fibrillation,
and decreased kidney function as well as type of post-VAD anticoagulation control on
risk of hemorrhage. Factors with p-values <0.2 were included in the multivariable
analysis.(24)
Oral anticoagulation is a risk factor for a hemorrhagic event in VAD patients.(22)
To explore this, anticoagulation control was assessed through the proportion of time
spent within target International Normalized Ratio (INR) range, which is referred to as
proportion of time in target range (PTTR). PTTR was estimated for each patient using
41
the Rosendaal linear interpolation method,(25) which assumes a linear relationship exists
between two consecutively measured INR values. This allows a specific INR value to be
allocated each day for every patient. Time in target range for each patient was assessed
by the percentage of interpolated INR values within the target range of 2.0–3.0 after
attainment of first INR in target range. We also assessed proportion of time spent in
extended target range (PTTRe; INR1.8-3.2) since deviations below or above this range
usually trigger dose adjustments to minimize risk of thrombosis and hemorrhage,
respectively. PTTRe was categorized into “good” control (>60%), “moderate” control
(≥50, but <60%), and “poor” control (<50%).
Kidney function was assessed through the estimated glomerular filtration rate
(eGFR) using the Chronic Kidney Disease Epidemiology Collaboration Formula (CKDEPI)(26) which incorporates gender, race, age, and serum creatinine (Scr; mg/dL).
Kidney function was characterized with eGFR >60 ml/min/1.73 m2 considered to be
no/mild kidney disease (Stage 1, 2), those with eGFR =30-59 ml/min/1.73 m2 considered
to be moderate kidney disease (Stage 3), and those with eGFR<30 ml/min/1.73 m2 (or on
dialysis) considered to be severe kidney disease (Stage 4, 5).(27, 28) Patients with
kidney dysfunction are known to experience additional platelet dysfunction, making this
measure germane to this study.
In addition to these individual-level risk factors, the hemorrhage risk score, HASBLED, was assessed as both a continuous variable and categorized into a dichotomous
risk group (low risk HAS-BLED<3 vs. HAS-BLED ≥3).(21) The score was calculated
with giving one point each for the following components: age at implant >65, pre-VAD
kidney function of CKD stage 4 or 5, impaired pre-VAD liver function (cirrhosis or AST
42
>3 times the upper limit normal), history of hypertension, major bleeding event prior to
VAD implantation, alcohol abuse, illicit drug use, history of stroke prior to VAD implant,
antithrombotic use, and percent time in target INR range post-VAD. The only
components of the score that use post-VAD information were antithrombotic treatments
and percent time in target range <60%. We evaluated the traditional HAS-BLED score
with the post-VAD INR range of 2.0-3.0 as well as an extended HAS-BLED (eHASBLED) score, where percent time in range post-VAD was calculated on the INR range of
1.8-3.2.
Definition of Outcomes
All hemorrhagic events were documented during the one-year follow-up. Since
patient with VADs only receive their care at UAB, the study ascertainment of
complications was robust. Hemorrhagic events included intracranial hemorrhage,
gastrointestinal (GI) bleeding, mediastinal bleeding requiring surgical intervention, or an
episode requiring transfusion of greater than four units of packed red blood cells
(pRBCs). As patients may experience multiple events within the 1-year follow-up
period, only the first event was included in the analysis. For patients who did not
experience a hemorrhagic event, follow-up time was censored at end of the study (1 year)
or explantation/transplantation/death (if earlier than 1 year).
Considering the risk of bleeding due to the surgery for the implantation of the
VAD itself, a secondary analysis was done assessing predictors of late hemorrhages,
defined as a hemorrhagic event occurring greater than 30 days post-VAD implant.
43
Statistical Analysis
The χ2 test of independence or Wilcoxon Rank Sum test was used to assess
baseline group differences for categorical variables and continuous variables,
respectively. Univariable logistic regression and Cox Proportional Hazards models were
used to assess the association between baseline characteristics and risk of hemorrhage to
establish which variables (p≤0.2) should be considered for development of a VAD
specific bleeding risk score. All variables that met the p≤0.2 were evaluated for the final
prediction model using logistic regression models. Model performance was assessed
using the c-statistic, corrected for optimism.(29) To correct for optimism, reduce bias
and internally validate the model we calculated the estimated decrease in the c-statistic
that would be expected in an independent dataset using the 0.632 bootstrap method.(30)
This option was chosen over the traditional split-sample modeling because the bootstrap
resampling technique has been shown to reduce bias and produce a stable and efficient
estimate of predictive accuracy when compared to other methods.(30) For this study we
created 100 datasets through bootstrap sampling with replacement. The models were fit
in each dataset and the average difference in the c-statistic between the bootstrap dataset
(the derivation dataset) and the original dataset (the validation dataset). This value
represents the expected optimism in the c-statistic calculated in this study. The points
assigned to the variables in the score were assigned based on the beta coefficients from
the logistic regression models. A cut point of the VAD bleeding risk score was
established based on sensitivity and specificity of the dichotomized score in predicting
hemorrhage.
44
The Cox-proportional hazard model was used to determine the influence of HASBLED, the extended HAS-BLED and the newly developed VAD bleeding risk score on
the risk of first hemorrhagic event. Log-log survival plots were used to assess the
proportional hazards assumption. A sub-analysis was done using Cox-proportional
hazards models to assess the newly developed VAD bleeding risk score on the risk of late
hemorrhage. All statistical analyses were conducted using SAS version 9.3 (SAS
Institute, Cary, NC) at a non-directional alpha of 0.05.
Results
A total of 115 patients who were implanted with HeartMate II device (Thoratec
Corporation, Pleasanton, CA) or HeartWare (HeartWare Inc, Framingham, MA) devices
at UAB between 2006 and 2012 were included in this study (Figure 1). The mean age at
implant was 52(±14.6 years, and the majority of patients were male (78.3%), white
(67.8%) and implanted as a bridge to transplant (56.6%) (Table 1).
A total of 36 patients experienced a hemorrhage during the first year (Incidence
Rate 70 per 100 person years (95%CI 50-96). Of these, 12 (33.3%) experienced a second
hemorrhage and 7 (19.4%) had a thromboembolism after the primary hemorrhage. There
were 3 (8.3%) with intracranial hemorrhage, 25 (69.4%) with GI bleeding, 6 (16.7%)
with mediastinal bleeding, and 2 (5.6%) required greater than 4 units of pRBCs. Patients
who had a hemorrhage within the first year had a higher percentage of ischemic HF
etiology (75% vs 41.8%, p=0.0124), hyperlipidemia (69.4% vs. 34.2%, p=0.0004), atrial
fibrillation (52.8% vs 30.4%, p=0.0213) and a higher age (60.9±10.1 vs. 50.3±14.8 years,
45
p=0.0002). Patients with a hemorrhage also had a higher frequency of moderate kidney
function (CKD Stage 3, eGFR 30-59; 60% vs 43.4%, p=0.0135) prior to VAD
implantation.
The 115 participants contributed 624.5 months (51.3 person years) of follow-up
time, an average 5.4 (± 4.8) months per patient. Patients were seen at least once a month,
an average of 1.4± 0.96 visits per patient per month (Table 2). The mean time to first
hemorrhage was 2.8± 3.1 months. Patients with a hemorrhage were seen more
frequently per month compared to those who did not have a hemorrhage (2.8±1.2 times
vs 1.7±0.9 times, p<0.0001).
Anticoagulation control, measured as percent time in
target range, was similar between patients with and without hemorrhage regardless of
which INR range is used.
Individual Predictors of Hemorrhage
Ischemic etiology of heart failure prior to VAD implantation had a higher risk of
hemorrhage (Hazard Ratio (HR) 2.77, 95%CI 1.30-5.89, p=0.0082). Patients with a
history of hyperlipidemia (HR 2.93, 95%CI 1.44-5.97, p=0.0030), atrial fibrillation (HR
1.91, 95%CI 0.99-3.69, p=0.05), age older than 65 at time of VAD implant (HR 2.51,
95%CI 1.28-1.85, p=0.0003) and history of a GI bleed prior to VAD (HR 3.81, 95%CI
1.66-8.75, p=0.0016) had a significantly higher risk of hemorrhage. Moderate decrease
in kidney function prior to VAD (CKD Stage 3, eGFR 30-59) increased the risk of
hemorrhage (HR 1.96, 95%CI 0.97-4.00, p=0.06); however, this association was not
statistically significant. There was no association between severely decreased kidney
46
function prior to VAD (eGFR< 30) and risk of hemorrhage (HR 1.08, 95%CI 0.24-4.84,
p=0.92).
Risk of Hemorrhage by Bleeding Risk Scores
Of the variables included in the development of the VAD bleeding risk score, the
variables that were selected for use in the final scoring system using Harrell’s optimism
bootstrapping method included age greater than 65 (1 point), history of GI bleed prior to
VAD implantation (1point), ischemic etiology of HF (1 point), history of hyperlipidemia
prior to implantation (1 point), and eGFR <30 prior to VAD implantation (1 point). The
VAD bleeding risk score ranged from 0-5. This score had the lowest Akaike Information
Criterion (118.6) with a Harrell’s adjusted c-statistic of 0.789. The sensitivity and
specificity of the dichotomized VAD Bleeding Risk Score <3 compared to VAD
Bleeding Risk Score ≥ 3 for predicting hemorrhage was 45.7% and 93.4%, respectively.
The positive predictive value and the negative predictive value of the dichotomized VAD
Bleeding Risk Score <3 compared to VAD Bleeding Risk Score ≥ 3 for predicting
hemorrhage was 76.2% and 78.9%.
The HAS-BLED, eHAS-BLED, and newly developed VAD Bleeding risk score
all predicted hemorrhagic complications. However, the VAD score predicted
hemorrhage better in VAD patients than the HAS-BLED or eHAS-BLED score (Figure
2) with an AUC of 0.789 ( vs. 0.661 and 0.655, respectively) (Table 3). The proportional
hazards assumption was met for each dichotomous score. Patients with a VAD bleeding
risk score ≥3 had the highest risk of hemorrhage (HR 4.61, 95%CI 2.37-8.99, p<0.0001)
compared to the 3 fold increased risk of hemorrhage for patients with a HAS-BLED score
47
≥ 3 (HR 3.18, 95%CI 0.96-10.5, p=0.06) and the 2.5 fold increased risk with an extended
HAS-BLED score ≥3 (HR 2.52, 95%CI 1.11-5.74, p=0.0275).
The newly developed VAD Bleeding risk score predicted hemorrhage better in
VAD patients than the HAS-BLED or extended HAS-BLED score (Figure 2) with an
AUC of 0.789. A higher bleeding risk score placed VAD patients at a higher risk of
hemorrhage (Table 3). The proportional hazards assumption was met for each
dichotomous score. The sensitivity and specificity of the VAD Bleeding risk score is
45.7% and 93.4% with a PPV of 76.2% and an NPV of 78.9%. The sensitivity and
specificity of the HAS-BLED score was 89.3% and 30.2% with a PPV of 36.2% and an
NPV of 86.4%. The sensitivity and specificity of the eHAS-BLED score was 71.4% and
50.8% with a PPV of 39.2% and an NPV of 80.0%.
Sub analysis of Late Hemorrhage
There were 23 (63.9%) hemorrhages that occurred 30 days post-VAD implant.
The newly developed VAD Bleeding Risk score as a continuous variable was predictive
of hemorrhages occurring 30 days post-VAD implant (HR 1.66, 95%CI 1.21-2.27,
p=0.0015). Patients with a VAD Bleeding Risk score ≥3 had an almost 5-fold increase in
risk of a hemorrhage that occurred 30 days post-VAD (HR 4.64, 95%CI 2.03-10.6,
p=0.0003) compared to those with a VAD Bleeding Risk score <3.
Discussion
Hemorrhage remains one of the most frequently reported complications post
VAD, and prevalence estimates range from 15-60%. (8, 9, 31-38) The ability to predict
which patients are at greatest risk for a hemorrhagic event can enable clinicians to
48
identify a high risk subgroup and alter anticoagulation therapy accordingly. To our
knowledge, this is the first study to develop a VAD-specific bleeding risk score and
assess the HAS-BLED score in a VAD population.
The VAD bleeding risk score is a simple and easy to use metric that can provide
clinicians information on risk of developing a hemorrhage post-VAD. This score ranges
from 0-5 and is predictive of hemorrhage in patients with a c-statistic of 0.789. When
dichotomized, a score of 3 or greater has a sensitivity and specificity of 45.7% and
93.4%, respectively, and a positive predictive value of 76.2% and an negative predictive
value of 78.9%.We acknowledge that by setting the VAD bleeding risk score threshold
relatively high, some patients with low VAD bleeding risk scores will be missed. Other
factors not included in the VAD bleeding risk score can contribute to patients with a low
risk score going on to have a hemorrhage. The VAD bleeding risk score is similar to the
HAS-BLED and extended HAS-BLED score, but it includes simple components that are
more VAD-specific in the calculation. Each component of the VAD bleeding risk score
is itself a risk factor for hemorrhagic events in this population. In testing the
predictability of the score, not all risk factors that place patients at risk for a hemorrhage
were included in the risk score due to their low predictive capabilities when included.
The VAD bleeding risk score is consistent with prior work that has shown risk factors for
hemorrhage in VAD patients include age, ischemic cardiomyopathy, and history of
gastrointestinal bleeding as pre-VAD risk factors implicated in post-VAD bleeding.(22,
39-42) Regardless of the population, advanced age is a significant factor in hemorrhage
in patients on chronic anticoagulation therapies. (41, 43) (44) Prior studies have assessed
predictors of hemorrhage, and age has consistently been shown to be the most important
49
determinant of whether a patient will experience GI bleeding.(41) Of interest, while in
the univariable analysis of risk factors moderate kidney dysfunction prior to VAD was a
significant risk factor for hemorrhage, however, the Harrell optimism score found that
severe kidney dysfunction prior to VAD was better in the overall score for predicting
hemorrhage. Other risk factors that were not included in the VAD bleeding risk score,
such as hypertension or increased BMI, could explain why some patients with a lower
VAD bleeding risk score experienced a hemorrhagic complication.
The bleeding rate in this population is higher than what has been reported in other
VAD populations (70 per 100 person years compared to 63 per 100 person years) and
even higher than other chronic anticoagulated populations (5.7 per 100 patient-years).(45,
46) A HAS-BLED score ≥3 and an extended HAS-BLED score ≥3 place patients at 3
and 2.5 times, respectively, the risk of hemorrhage compared to patients with lower
scores. The HAS-BLED score uses established risk factors in other populations on
chronic warfarin therapy to predict patients at highest risk for bleeding while on
antithrombotic therapy with a c-statistic of 0.91.(21) However, the propensity for
hemorrhagic complications in VAD patients cannot be attributed to anticoagulation alone
since they experience a higher rate of hemorrhagic complications compared to patients on
anticoagulation for other indications.(40) The components of the VAD bleeding risk
score emphasize the importance of pre-implant risk factors on the risk of hemorrhage
post-implant. Furthermore, the score is predictive of hemorrhages that occur 30 days
post-VAD. Hemorrhages are more likely to be ascribed to the device or the
antithrombotic regimen rather than complications of surgical implantation of the VAD.
50
When compared to non-VAD patients on anti-thrombotic therapy, VAD patients
are at a higher risk for hemorrhage.(22, 46, 47) The increased risk for gastrointestinal
bleeding in VAD patients is multifactorial and could be due to antithrombotic therapies,
development of arteriovenous malformations (AVMs), Heyde syndrome, or mucosal
ischemia.(45) VAD devices may predispose patients to develop AVMs by decreasing
intraluminal pressure due to the continuous flow VADs, thereby increasing the
sympathetic tone and dilating the mucosal veins, leading to the formation of AVMs.(40,
48, 49) These AVMs are predisposed to bleeding due to anticoagulation and deviceinduced stress. (49) Another potential reason for the increased risk of hemorrhage is that
continuous-flow devices have been found to result in acquired von Willebrand syndrome,
which further complicate the coagulation profiles of these patients.(19) The increased
risk of thromboembolism necessitates the use of antithrombotic therapies that increase
the risk of hemorrhagic complications.(50)
Our study was not without limitations. We only included patients with a
HeartMate II and HeartWare devices to reduce bias and increase generalizability to the
newer devices currently being implanted. While this is a large number of HeartMate II/
HeartWare patients at a single center, the numbers of events are low and decrease the
statistical power. We were not able to assess specific types of hemorrhagic events. Since
this was a retrospective study, we were limited by the data available in the medical
records. The VAD bleeding risk score is limited to first events and was not designed to
predict a hemorrhage in patients who had a thromboembolism prior to their hemorrhage.
We did not have an independent population to validate our risk score, although we did
correct for the expected overestimate of the c-statistic that results from testing its
51
properties in the population in which it was developed. Prospective, multicenter studies
are needed to validate the VAD bleeding risk score.
This complex patient population is at very high risk of hemorrhagic events due to
the device itself as well as the therapies needed to prevent thromboembolism. Our study
is the first develop a VAD specific bleeding risk score and to compare it to the HASBLED score, which was designed to assess risk of bleeding in patients on chronic oral
anticoagulation.
We found that patients with a VAD bleeding risk score of ≥3 are at 4
times the risk of hemorrhage than patients with a lower VAD bleeding risk score. Further
studies are needed to validate this score in other VAD populations.
52
Reference List
1.
Roger VL, Go AS, Lloyd-Jones DM, Adams RJ, Berry JD, Brown TM, et al. Heart
disease and stroke statistics--2011 update: a report from the American Heart
Association. Circulation. 2011;123(4):e18-e209.
2.
Jessup M, Albert NM, Lanfear DE, Lindenfeld J, Massie BM, Walsh MN, et al.
ACCF/AHA/HFSA 2011 survey results: current staffing profile of heart failure
programs, including programs that perform heart transplant and mechanical
circulatory support device implantation: a report of the ACCF Heart Failure and
Transplant Committee, AHA Heart Failure and Transplantation Committee, and
Heart Failure Society of America. Circ Heart Fail. 2011;4(3):378-87.
3.
Wilson SR, Givertz MM, Stewart GC, Mudge GH, Jr. Ventricular assist devices
the challenges of outpatient management. J Am Coll Cardiol. 2009;54(18):164759.
4.
Haeck ML, Hoogslag GE, Rodrigo SF, Atsma DE, Klautz RJ, van der Wall EE, et
al. Treatment options in end-stage heart failure: where to go from here? Neth Heart
J. 2011.
5.
Torpy JM, Lynm C, Golub RM. JAMA patient page. Heart failure. JAMA.
2011;306(19):2175.
6.
Goldstein DJ, Oz MC, Rose EA. Implantable left ventricular assist devices. N Engl
J Med. 1998;339(21):1522-33.
7.
Hosenpud JD, Bennett LE, Keck BM, Boucek MM, Novick RJ. The Registry of
the International Society for Heart and Lung Transplantation: seventeenth official
report-2000. The Journal of heart and lung transplantation : the official publication
of the International Society for Heart Transplantation. 2000;19(10):909-31.
8.
Rose EA, Gelijns AC, Moskowitz AJ, Heitjan DF, Stevenson LW, Dembitsky W,
et al. Long-term use of a left ventricular assist device for end-stage heart failure. N
Engl J Med. 2001;345(20):1435-43.
9.
Jaski BE, Kim JC, Naftel DC, Jarcho J, Costanzo MR, Eisen HJ, et al. Cardiac
transplant outcome of patients supported on left ventricular assist device vs.
intravenous inotropic therapy. The Journal of heart and lung transplantation : the
official publication of the International Society for Heart Transplantation.
2001;20(4):449-56.
10.
Stehlik J, Edwards LB, Kucheryavaya AY, Aurora P, Christie JD, Kirk R, et al.
The Registry of the International Society for Heart and Lung Transplantation:
twenty-seventh official adult heart transplant report--2010. The Journal of heart
and lung transplantation : the official publication of the International Society for
Heart Transplantation. 2010;29(10):1089-103.
53
11.
Birks EJ, Tansley PD, Hardy J, George RS, Bowles CT, Burke M, et al. Left
ventricular assist device and drug therapy for the reversal of heart failure. N Engl J
Med. 2006;355(18):1873-84.
12.
Slaughter MS, Rogers JG, Milano CA, Russell SD, Conte JV, Feldman D, et al.
Advanced heart failure treated with continuous-flow left ventricular assist device.
N Engl J Med. 2009;361(23):2241-51.
13.
Rogers JG, Butler J, Lansman SL, Gass A, Portner PM, Pasque MK, et al. Chronic
mechanical circulatory support for inotrope-dependent heart failure patients who
are not transplant candidates: results of the INTrEPID Trial. J Am Coll Cardiol.
2007;50(8):741-7.
14.
Lazar RM, Shapiro PA, Jaski BE, Parides MK, Bourge RC, Watson JT, et al.
Neurological events during long-term mechanical circulatory support for heart
failure: the Randomized Evaluation of Mechanical Assistance for the Treatment of
Congestive Heart Failure (REMATCH) experience. Circulation.
2004;109(20):2423-7.
15.
Pae WE, Connell JM, Boehmer JP, Korfer R, El-Banayosy A, Hetzer R, et al.
Neurologic events with a totally implantable left ventricular assist device:
European LionHeart Clinical Utility Baseline Study (CUBS). The Journal of heart
and lung transplantation : the official publication of the International Society for
Heart Transplantation. 2007;26(1):1-8.
16.
Slaughter MS, Sobieski MA, Gallagher C, Dia M, Silver MA. Low incidence of
neurologic events during long-term support with the HeartMate XVE left
ventricular assist device. Tex Heart Inst J. 2008;35(3):245-9.
17.
Thomas CE, Jichici D, Petrucci R, Urrutia VC, Schwartzman RJ. Neurologic
complications of the Novacor left ventricular assist device. The Annals of thoracic
surgery. 2001;72(4):1311-5.
18.
Tsukui H, Abla A, Teuteberg JJ, McNamara DM, Mathier MA, Cadaret LM, et al.
Cerebrovascular accidents in patients with a ventricular assist device. J Thorac
Cardiovasc Surg. 2007;134(1):114-23.
19.
Crow S, Chen D, Milano C, Thomas W, Joyce L, Piacentino V, 3rd, et al.
Acquired von Willebrand syndrome in continuous-flow ventricular assist device
recipients. The Annals of thoracic surgery. 2010;90(4):1263-9; discussion 9.
20.
Geisen U, Heilmann C, Beyersdorf F, Benk C, Berchtold-Herz M, Schlensak C, et
al. Non-surgical bleeding in patients with ventricular assist devices could be
explained by acquired von Willebrand disease. Eur J Cardiothorac Surg.
2008;33(4):679-84.
54
21.
Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. A novel userfriendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients
with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138(5):1093-100.
22.
Boyle AJ, Jorde UP, Sun B, Park SJ, Milano CA, Frazier OH, et al. Pre-operative
risk factors of bleeding and stroke during left ventricular assist device support: an
analysis of more than 900 HeartMate II outpatients. J Am Coll Cardiol.
2014;63(9):880-8.
23.
Interagency Registry for Mechanically Assisted Circulatory Support
(INTERMACS) Official Website. . Available from: http://www.intermacs.org/.
24.
Bender R, Lange S. Adjusting for multiple testing--when and how? J Clin
Epidemiol. 2001;54(4):343-9.
25.
Rosendaal FR, Cannegieter SC, van der Meer FJ, Briet E. A method to determine
the optimal intensity of oral anticoagulant therapy. Thromb Haemost.
1993;69(3):236-9.
26.
Valente MA, Hillege HL, Navis G, Voors AA, Dunselman PH, van Veldhuisen
DJ, et al. The Chronic Kidney Disease Epidemiology Collaboration equation
outperforms the Modification of Diet in Renal Disease equation for estimating
glomerular filtration rate in chronic systolic heart failure. Eur J Heart Fail.
2014;16(1):86-94.
27.
Levey AS, Coresh J, Balk E, Kausz AT, Levin A, Steffes MW, et al. National
Kidney Foundation practice guidelines for chronic kidney disease: evaluation,
classification, and stratification. Ann Intern Med. 2003;139(2):137-47.
28.
Huang SM, Temple R, Xiao S, Zhang L, Lesko LJ. When to conduct a renal
impairment study during drug development: US Food and Drug Administration
perspective. Clin Pharmacol Ther. 2009;86(5):475-9.
29.
Miao Y, Cenzer I, Kirby K, Boscardin WJ, editors. Estimating Harrell's Optimism
on Predictive Indices Using Bootstrap Samples. SAS Global Forum; 2013.
30.
Steyerberg EW, Harrell FE, Jr., Borsboom GJ, Eijkemans MJ, Vergouwe Y,
Habbema JD. Internal validation of predictive models: efficiency of some
procedures for logistic regression analysis. J Clin Epidemiol. 2001;54(8):774-81.
31.
Dang NC, Naka Y. Perioperative pharmacotherapy in patients with left ventricular
assist devices. Drugs Aging. 2004;21(15):993-1012.
32.
Frazier OH, Rose EA, Oz MC, Dembitsky W, McCarthy P, Radovancevic B, et al.
Multicenter clinical evaluation of the HeartMate vented electric left ventricular
55
assist system in patients awaiting heart transplantation. J Thorac Cardiovasc Surg.
2001;122(6):1186-95.
33.
Goldstein DJ, Beauford RB. Left ventricular assist devices and bleeding: adding
insult to injury. The Annals of thoracic surgery. 2003;75(6 Suppl):S42-7.
34.
Himmelreich G, Ullmann H, Riess H, Rosch R, Loebe M, Schiessler A, et al.
Pathophysiologic role of contact activation in bleeding followed by
thromboembolic complications after implantation of a ventricular assist device.
ASAIO J. 1995;41(3):M790-4.
35.
Meyns B, Sergeant P, Herijgers P, Flameng W, Vanhaecke J, Van Cleemput J, et
al. Bridging procedures to heart transplantation. World J Surg. 2002;26(2):212-7.
36.
Minev PA, El-Banayosy A, Minami K, Kortke H, Kizner L, Korfer R. Differential
indication for mechanical circulatory support following heart transplantation.
Intensive Care Med. 2001;27(8):1321-7.
37.
Vitali E, Lanfranconi M, Bruschi G, Russo C, Colombo T, Ribera E. Left
ventricular assist devices as bridge to heart transplantation: The Niguarda
Experience. J Card Surg. 2003;18(2):107-13.
38.
Pereira NL, Chen D, Kushwaha SS, Park SJ. Discontinuation of antithrombotic
therapy for a year or more in patients with continuous-flow left ventricular assist
devices. Interact Cardiovasc Thorac Surg. 2010;11(4):503-5.
39.
Uriel N, Pak SW, Jorde UP, Jude B, Susen S, Vincentelli A, et al. Acquired von
Willebrand syndrome after continuous-flow mechanical device support contributes
to a high prevalence of bleeding during long-term support and at the time of
transplantation. J Am Coll Cardiol. 2010;56(15):1207-13.
40.
Crow S, John R, Boyle A, Shumway S, Liao K, Colvin-Adams M, et al.
Gastrointestinal bleeding rates in recipients of nonpulsatile and pulsatile left
ventricular assist devices. J Thorac Cardiovasc Surg. 2009;137(1):208-15.
41.
Demirozu ZT, Radovancevic R, Hochman LF, Gregoric ID, Letsou GV, Kar B, et
al. Arteriovenous malformation and gastrointestinal bleeding in patients with the
HeartMate II left ventricular assist device. The Journal of heart and lung
transplantation : the official publication of the International Society for Heart
Transplantation. 2011;30(8):849-53.
42.
Morgan JA, Paone G, Nemeh HW, Henry SE, Patel R, Vavra J, et al.
Gastrointestinal bleeding with the HeartMate II left ventricular assist device. The
Journal of heart and lung transplantation : the official publication of the
International Society for Heart Transplantation. 2012;31(7):715-8.
56
43.
Stulak JM, Lee D, Haft JW, Romano MA, Cowger JA, Park SJ, et al.
Gastrointestinal bleeding and subsequent risk of thromboembolic events during
support with a left ventricular assist device. The Journal of heart and lung
transplantation : the official publication of the International Society for Heart
Transplantation. 2014;33(1):60-4.
44.
Mant J, Hobbs FD, Fletcher K, Roalfe A, Fitzmaurice D, Lip GY, et al. Warfarin
versus aspirin for stroke prevention in an elderly community population with atrial
fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study,
BAFTA): a randomised controlled trial. Lancet. 2007;370(9586):493-503.
45.
Islam S, Cevik C, Madonna R, Frandah W, Islam E, Nugent K. Left ventricular
assist devices and gastrointestinal bleeding: a narrative review of case reports and
case series. Clin Cardiol. 2013;36(4):190-200.
46.
Levine MN, Raskob G, Beyth RJ, Kearon C, Schulman S. Hemorrhagic
complications of anticoagulant treatment: the Seventh ACCP Conference on
Antithrombotic and Thrombolytic Therapy. Chest. 2004;126(3 Suppl):287S-310S.
Kushnir VM, Sharma S, Ewald GA, Seccombe J, Novak E, Wang IW, et al.
Evaluation of GI bleeding after implantation of left ventricular assist device.
Gastrointest Endosc. 2012;75(5):973-9.
47.
48.
Letsou GV, Shah N, Gregoric ID, Myers TJ, Delgado R, Frazier OH.
Gastrointestinal bleeding from arteriovenous malformations in patients supported
by the Jarvik 2000 axial-flow left ventricular assist device. The Journal of heart
and lung transplantation : the official publication of the International Society for
Heart Transplantation. 2005;24(1):105-9.
49.
Cappell MS, Lebwohl O. Cessation of recurrent bleeding from gastrointestinal
angiodysplasias after aortic valve replacement. Ann Intern Med. 1986;105(1):54-7.
50.
Kurien S, Hughes KA. Anticoagulation and bleeding in patients with ventricular
assist devices: walking the tightrope. AACN Adv Crit Care. 2012;23(1):91-8.
57
Table 1. Baseline Characteristics of VAD Patients with and without a Hemorrhage All Patients
(N=115)
Age at Implant
BMI at VAD Implantation
52 (14.6)
28 (7.6)
No Hemorrhage
(N=79)
Mean (SD)
50.3 (14.8)
29.4 (6.6)
N (%)
59 (74.6%)
27 (34.2%)
Hemorrhage
(N=36)
p-value
60.9 (10.1)
28.2 (5.5)
0.0002
0.45
Male
90 (78.3%)
31 (86.1%)
0.17
Black Race
37 (32.2%)
10 (27.8%)
0.50
Status in the first year
0.10
Deceased 19 (16.5%)
11 (13.9%)
8 (22.2%)
Transplanted 23 (20.0%)
19 (24.1%)
4 (11.1%)
Recovered
4 (3.5%)
4 (5.1%)
0
Etiology of Heart Failure
0.012
Ischemic 60 (53.1%)
33 (41.8%)
27 (75%)
Idiopathic 45 (39.8%)
38 (48.1%)
7 (19.4%)
Indication for VAD
0.54
Bridge to transplant 64 (56.6%)
47 (59.5%)
17 (47.2%)
Bridge to candidacy
2 (1.8%)
1 (1.3%)
1 (2.8%)
Bridge to recovery
1 (0.9%)
1 (1.3%)
0
Destination 46 (40.7%)
29 (36.7%)
17 (47.2%)
Comorbid Conditions
Diabetes 45 (39.1%)
27 (34.2%)
18 (50%)
0.11
Right Ventricular 28 (24.4%)
21 (26.7%)
7 (19.4%)
0.41
Dysfunction
Coronary Artery Disease 46 (40.0%)
30 (37.9%)
16 (44.4%)
0.51
Myocardial Infarction 23 (20.0%)
12 (15.2%)
11 (30.6%)
0.06
Hypertension 71 (61.7%)
45 (56.9%)
26 (72.2%)
0.12
Hyperlipidemia 52 (45.2%)
27 (34.2%)
25 (69.4%)
0.0004
Atrial Fibrillation 43 (37.4%)
24 (30.4%)
19 (52.8%)
0.021
Ventricular Tachycardia 55 (47.8%)
38 (48.1%)
17 (47.2%)
0.93
Concurrent Meds at VAD Implantation
Warfarin 115 (100%)
79 (100%)
36 (100%)
Aspirin 113 (98.2%)
79 (100%)
34 (94.4%)
0.09
Clopidogrel 67 (58.3%)
42 (53.2%)
25 (69.4%)
0.10
Colchicine
6 (5.2%)
3 (3.8%)
3 (8.3%)
0.20
Amiodarone 70 (60.9%)
49 (62.0%)
21 (58.3%)
0.71
Estimated Glomerular Filtration rate (eGFR ml/min/1.73m2) prior to VAD
0.014
Implantation
eGFR ≥60 47 (42.3%)
35 (46.1%)
12 (34.3%)
eGFR 30 to 59 54 (48.7%)
33 (43.4%)
21 (60%)
eGFR < 30
10 (9.0%)
8 (10.5%)
2 (5.7%)
Estimated glomerular filtration rate (eGFR ml/min/1.73m2) was calculated using the Chronic
Kidney Disease Epidemiology Collaboration Formula (CKD-EPI). Continuous variables were
tested using Wilcoxon Rank Sum test and categorical variables were tested using a Chi-squared
test
58 Table 2. Anticoagulation Control for the Entire Cohort
All Patients
(N=115)
No
Hemorrhage
(N=79)
Mean(SD)
Hemorrhage
(N=36)
p-value
Number of Patients
115
79
36
Number of Visits
4902
3612
1298
<0.0001
Total Follow Up (months)*
624.5
523.5
100.9
<0.0001
Follow Up Months/patient
5.4 ± 4.8
6.6 ±4.9
2.8±3.1
<0.0001
Number of Visits/patient/month
2.1 ± 1.2
1.7±0.9
2.8±2.6
<0.0001
Anticoagulation control for INR Range of 2-3
Percent Time Below Range
41.6 ± 28.4
40.8±26.5
44.7±31.9
0.74
Percent Time In Range
42.9 ± 22.5
44.9±21.1
39.9±24.7
0.35
Percent Time Above Range
15.8 ± 13.9
16.5±13.6
14.2±14.6
0.21
Anticoagulation Control for INR Range of 1.8-3.2
Percent Time Below Range
30.3 ± 28.5
29.2±25.5
32.9 ±34.8
0.84
Percent Time In Range
59.2 ± 26.4
61.2±23.2
56.6±31.3
0.81
Percent Time Above Range
11.1 ± 11.5
11.9±11.8
9.3±10.8
0.12
*Follow up time is designated as time to thromboembolism, hemorrhage, death,
transplant, withdrawal or end of study (1-year)
Continuous variables were tested using Wilcoxon Rank Sum test
59 Table 3. Association of Bleeding Risk Scores with Risk (Hazard Ratio (HR) and
95% Confidence Interval (95%CI)) of Hemorrhage
Hazard Ratio
95% Confidence
Interval
p-value
HASBLED
1.62
1.13-2.33
0.0088
eHASBLED
1.73
1.21-2.49
0.0028
VAD Bleeding Risk Score
1.66
1.29-2.15
0.0001
HASBLED ≥3
3.18
0.96-10.5
0.06
eHASBLED ≥3
2.52
1.11-5.74
0.0275
VAD Bleeding Risk Score ≥ 3
4.61
2.37-8.99
<0.0001
Scores Assessed as Continuous Variables
Scores Assessed as Dichotomous Variables
Hazard Ratios were generated through Cox Proportional Hazards Models. The
proportionality assumption was met for the dichotomized score. The proportional
hazards assumption was met for the assessment of the dichotomized score variable
60 Figure 1. Inclusion and Exclusion Criteria Used to Establish the Cohort
241 VAD patients enrolled in the
Cohort from 2002-2012
127 HeartMate II/HeartWare
patients
115 patients with
Anticoagulation medications and
lab values available for
enrollment in this study
7 of these pa&ents go on to have a second thromboembolism 79 of these pa&ents did not have a thromboembolic or hemorrhagic event 12 of these pa&etns go on to have a hemorrhage as a secondary event 61 12 patients who were implanted
at different facilities were
excluded due to no detailed
information on anticoagulation
at the time of VAD implant
23 of these pa&ents expereinced a thromboembolism, changing their INR range to 2.5-­‐3.5 were excluded 92 Pa&ents met the inclusion/
exclusion criteria 36 Pa&ents experienced a primary hemorrhage Excluded 114 patients with
Pulsatile devices
7 pa&ents go on to have a thromboembolism 8 pa&ents go on to have a second hemorrhage Figure 2. Comparing HAS-BLED to the Extended HAS-BLED to the VAD Bleeding
Risk Score in the Entire VAD Population
62 ROC curves were generated through logistic regression analyses using the Harrell
Optimism prediction method
63 IMPROVEMENT IN KIDNEY FUNCTION AFTER VENTRICULAR ASSIST
DEVICE IMPLANTATION AND ITS INFLUENCE ON THROMBOEMBOLISM,
HEMORRHAGE, AND MORTALITY
by
AMELIA K BOEHME, SALPY V PAMBOUKIAN, JAMES F GEORGE, CHRISLY
DILLON, JAMES K KIRKLIN, JOSE TALLAJ, EMILY B LEVITAN, RUSSELL
GRIFFIN, GERALD McGWIN, MICHAEL ALLON, NITA A LIMDI
In Preparation for the Clinical Journal of American Society of Nephrology
Format adapted for dissertation
64
ABSTRACT
Background and Objectives: To examine improvement in kidney function post-implant
of a Ventricular Assist Device (VAD), identify predictors of kidney function
improvement, and examine the whether kidney function improvement is associated with
thromboembolism, hemorrhage, and mortality.
Design, Setting, Participants and Measurements: Patients (aged 19 or older) with
either a continuous or pulsatile flow VAD placed from 2002-2012 at a single center were
included (N=228). Kidney function was defined using chronic kidney disease (CKD)
stages: Stage 1 (glomerular filtration rate (eGFR)≥90 ml/min/1.73m2), Stage 2 (eGFR 6090 ml/min/1.73m2), Stage 3a (eGFR 45-59 ml/min/1.73m2), Stage 3b (eGFR 30-44
ml/min/1.73m2), Stage 4 (eGFR 15-30 ml/min/1.73m2) and Stage 5 (eGFR<15
ml/min/1.73m2).We assessed change in kidney function at: pre-implantation, 14 days, 1,
3, 6, 9 and 12 months post-implantation. Improvement in kidney function was defined as
an improvement that resulted in a CKD stage change to one of lesser severity. Baseline
kidney function and improvement in kidney function were assessed as risk factors for
thromboembolism, hemorrhage and mortality.
Results: Kidney function improves post implant with improvement in kidney function
maintained over 1 year for all patients except those with CKD stage 5 at baseline. Age at
implant was the only predictor of sustained improvement in kidney function (OR 0.93,
95% CI 0.90-0.96, p<0.0001). The adjusted risk of death was 3 times higher for patients
with baseline CKD stage 3b (HR 3.32, 95%CI 1.10-9.98, p=0.03), and 4 times higher for
baseline CKD stage 4 (HR 4.07, 95%CI 1.27-13.1, p=0.02) and stage 5 (HR 4.01, 95%CI
65
1.17-13.7, p=0.03). Sustained improvement in kidney function was not associated with a
lower risk of death.
Conclusions: Most patients experience an improvement in CKD stage after VAD
implantation. However, despite these improvements in kidney function post-VAD
implant, it was renal function at the time of VAD implantation that was associated with
mortality.
66
Introduction
The prevalence of chronic kidney disease (CKD) is increasing,(1) especially in
patients with cardiovascular disease.(2) Thirty to forty percent of patients with heart
failure (HF) have CKD,(2) and the prevalence increases further (up to 63%) among
advanced HF patients.(3) (4) In approximately 25% of advanced HF patients, chronic
cardiac dysfunction is responsible for progressive CKD.(5, 6) CKD is shown to be a
stronger predictor of mortality among HF patients compared to cardiac function.(7-10)
Treatment strategies for HF range from lifestyle changes and medical management to
heart transplant or ventricular assist device (VAD) depending on the severity of HF.(11,
12) Among patients with end-stage HF, VAD implantation is increasingly used as a
bridge to recovery, bridge to transplant, or as destination therapy. Patients with VAD
implants demonstrate an improvement in kidney function following implantation. (13)
However, the bulk of the evidence is based on assessment of kidney function in the early
(7-30 days) post-implant period. In a recent report of eighty-three VAD recipients, kidney
function, assessed by glomerular filtration rate (eGFR), improved post-implant and this
improvement was sustained at 180 days.(14) Additionally, Kirklin et al illustrated
sustained improvement in blood urea nitrogen (BUN) and creatinine concentrations up to
2 years after VAD implantation.16 However no studies have evaluated whether
improvements in kidney function are associated with better clinical outcomes.
Herein we assess improvement in kidney function post-VAD implantation, whether
the improvement in kidney function is sustained for a period of one year, identify
predictors of kidney function improvement, and examine the whether kidney function
67
improvement is associated with a decrease in risk of thromboembolic events,
hemorrhagic events, and all-cause mortality.
Materials and Methods
Study Setting
The study enrolled patients who received a ventricular assist device (VAD) from
2002-2012 at the University of Alabama at Birmingham under the approval of the
Institutional Review Board.
Inclusion and Exclusion Criteria
Patients aged 19 years and older who had a VAD (continuous-flow or pulsatile
flow) placed at UAB from 2002-2012 were included in this study. All patients received
post-implant medical care at the implanting center. Most patients received inpatient care
for 2-4 weeks post-implantation. After discharge, all patients received care as outpatients
at monthly intervals.
Of the 241 VAD patients eligible, 13 patients were implanted at another
institution and then received follow-up care at UAB. These individuals were excluded
because their baseline kidney function assessment at time of implantation was not
available. This resulted in 228 patients in our final study sample.
Data collection
For all patients, information including demographic variables (age at implant,
self-reported race, ethnicity etc.), medical history prior to VAD (comorbid conditions,
heart failure etiology etc.), and laboratory assessments were collected at baseline (pre68
VAD implantation). Clinical data, including medications and laboratory assessments,
were collected from each post-VAD clinic visit or hospitalization period during the 1year follow-up period. Patients were followed for 1-year post-VAD implantation or until
explantation of the device, transplant or death prior to 1 year. All adverse events were
defined using definitions from the Interagency Registry for Mechanically Assisted
Circulatory Support (INTERMACS) Registry.(15)
Definition of Exposure
Kidney function was assessed based on estimated glomerular filtration rate
(eGFR) using the Chronic Kidney Disease Epidemiology Collaboration Formula (CKDEPI)(16) which incorporates gender, race, age, and serum creatinine (Scr; mg/dL). We
assessed change in kidney function from VAD implantation at the following six timepoints: pre-implantation, 14 ± 2 days, 1 month ± 2 days, 3 months ± 7 days, 6 months ± 7
days, 9 months ± 7 days and 12 months ± 7 days post-implantation.
We present CKD stages as recommended by the Kidney Disease Outcomes
Quality Initiative guidelines: The groups were Stage 1 (eGFR≥90 ml/min/1.73m2), Stage
2 (eGFR 60-90 ml/min/1.73m2), Stage 3a (eGFR 45-59 ml/min/1.73m2), Stage 3b (eGFR
30-44 ml/min/1.73m2), Stage 4 (eGFR 15-30 ml/min/1.73m2) and Stage 5 (eGFR<15
ml/min/1.73m2). (17)
Improvement in kidney function was defined as an improvement in eGFR that
resulted in a CKD stage change to one of lesser severity. For example, consider two
patients with eGFR of 36 at baseline. One patient’s eGFR improves to 52ml/min while
the second patient’s eGFR improves to 39ml/min. While eGFR improved in both
69
patients, in the first patient the improvement would reclassify the CKD stage from Stage
3b to Stage 3a.
We also assessed if improvements in eGFR were sustained over the 1-year
follow-up period post implantation. Sustained improvement was defined as improvement
in CKD stage over baseline and improvement in CKD stage that was sustained over the
1-year follow-up period. Patients with CKD Stage 1 or 2 at baseline who maintained
kidney function within the pre-VAD CKD stage were included in the sustained
improvement category as they did not have a decrease in kidney function post VAD.
Definition of Outcomes
Outcomes of interest included thromboembolic events, hemorrhagic events, and allcause mortality. Outcomes were determined as a documented event through review of
medical records. Thromboembolic events included ischemic stroke, transient ischemic
attack, pulmonary embolus, deep vein thrombosis, pump thrombosis requiring
hospitalization, and mediastinal clot requiring surgical removal. Hemorrhagic events
included intracranial hemorrhage, pericardial bleeding, intraarticular bleeding,
gastrointestinal bleed, greater than 2 units of packed red blood cells administered at one
time, and mediastinal bleeding requiring surgical intervention. Mortality was defined as
death from any cause.
Statistical Analysis
70
The χ2 test of independence was used to assess group differences for categorical
variables and ANOVA/Student’s t-test or Kruskal-Wallis/Wilcoxon Rank Sum where
appropriate for continuous variables. Logistic regression models were used to assess
predictors of sustained improvement. Cox proportional hazards models were used to
evaluate the association between kidney function and thromboembolism, hemorrhage,
and death. Log survival plots were used to assess the proportional hazards assumption.
For mortality analysis, follow-up time was censored at time of transplant, explant
(removal of the VAD device), or death. For analysis of thromboembolic or hemorrhagic
event, the follow-up time was censored at the time of first event; repeated events were not
considered. For patients who did not experience a thromboembolic or hemorrhagic event,
follow-up time was censored at end of the study (1 year) or explantation (if earlier than 1
year). All tests were performed using SAS version 9.2 (SAS Institute, Cary, NC) at a
non-directional alpha level of 0.05.
Results
Overall Population
Of the 241 VAD patients who were treated at UAB between 2002 and 2012,
thirteen patients implanted at an outside hospital were excluded because baseline kidney
function information was not available. Of the 228 patients (72% men, 71% White)
included in the analysis, the mean age at implant was 52 years (14.6 SD), and the
majority (57%) received VAD implants as a bridge to transplantation (57 %). At baseline,
35.5% participants had Stage 1 or 2 CKD (eGFR≥60 ml/min/1.73m2), 23% had Stage 3a
71
CKD (eGFR 45-59 ml/min/1.73m2), 23% had Stage 3b CKD (eGFR 30-44
ml/min/1.73m2), 13% had Stage 4 CKD (eGFR 15-29 ml/min/1.73m2) and 5% had Stage
5 CKD (eGFR <15 ml/min/1.73m2). Baseline characteristics of the cohort, stratified by
kidney function, are presented in Table 1. No significant differences were observed
between CKD stages and comorbidities except patients with continuous flow devices
who had less severe kidney dysfunction (p=0.001), and those with history of diabetes
(p=0.02), and history of hypertension (p=0.01) who had more baseline kidney
dysfunction.
Change in Kidney Function over Time
Within the first 14 days, a majority (56.7%; n=130) of VAD recipients
demonstrated meaningful improvement in kidney function compared to baseline kidney
function (Figure 1). Regardless of fluctuations in eGFR over time, early improvement in
kidney function was sustained for a majority of patients (n=127) for the duration of the 1year follow up, except those with Stage 5 CKD at baseline (n=11). Although patients
with Stage 5 CKD showed improvement in kidney function over the first 3 months, the
improvement was not sustained. There was no difference in improvement in CKD stage
after implantation between continuous and pulsatile flow devices.
Increasing age at implantation was inversely associated with sustained
improvement in kidney function. For each year increase in age at implant, the odds of
having sustained improvement in kidney function decreased by 5% (OR 0.95, 95% CI
0.92-0.97, p<0.0001). This association remained after adjusting for baseline CKD stage,
72
gender, race, history of diabetes and history of hypertension (OR 0.93, 95% CI 0.90-0.96,
p<0.0001). Patients who sustained kidney function improvement over time were more
likely to receive a heart transplant than those who did not sustain kidney function over
time (28% vs. 16%, p=0.05).
Kidney Function and Thromboembolism
There were 39 thromboembolic events over a follow-up time of 101 person-years
with an incidence rate of 3.9 per 10 person years (95% CI 2.8-5.2). Neither baseline
kidney function (Table 2), nor sustained improvement in kidney function at 1 year (data
not shown; HR 0.98, 95% CI 0.54-1.78, p=0.95) were associated with the risk of
thromboembolic events.
Kidney Function and Hemorrhage
There were 58 hemorrhagic events over a follow-up time of 101 person-years
with an incidence rate of 5.7 per 10 person years (95% CI 4.4-7.4). There was no
statistically significant association with baseline kidney function and risk of hemorrhagic
events (Table 2). There was no significant reduction in the risk of hemorrhage for those
who sustained kidney function improvement over the follow-up as compared to those
who did not sustain improvement (HR 0.75, 95% CI 0.46-1.22, p=0.24).
Kidney Function and Death
73
A total of 80 deaths occurred over a follow-up time of 151 person-years with an
incidence rate of 5.3 per 10 person-years (95% CI 4.2-6.6). The risk of death was higher
amongst patients with compromised kidney function at baseline (Figure 2a). Patients
with CKD stage 3b (HR 3.12, 95%CI 1.16-9.17, p=0.039), CKD stage 4(HR 3.94, 95%CI
1.29-12.0, p=0.02) and CKD stage 5 (HR 7.24, 95%CI 2.18-24.1, p=0.001) had a
significantly higher risk of mortality compared to those with CKD stage 1. This
association persisted, even after adjustment for device type, age at implant, history of
diabetes, and history of hypertension (Figure 2b). The risk of death was 3 times higher
for CKD stage 3b (HR 3.32, 95%CI 1.10-9.98, p=0.03), and 4 times higher for CKD
stage 4 (HR 4.07, 95%CI 1.27-13.1, p=0.02) and Stage 5 (HR 4.01, 95%CI 1.17-13.7,
p=0.03) compared to those with CKD stage 1. Improvement in kidney function that was
sustained for 1-year was not associated with a lower risk of death (HR 1.03, 95%CI 0.581.84, p=0.91).
Discussion
Our study demonstrates that VAD implantation is associated with clinically
meaningful improvement in kidney function that is sustained over a 1- year. However,
improvement in kidney function did not influence the risk of thromboembolism,
hemorrhage or mortality. We found that kidney function at the time of implantation
increased the risk of mortality, with patients in the more advanced stages of kidney
dysfunction having the highest risk of death.
74
Kidney function has been shown to improve after VAD implantation, but the
extent and duration of improvement has not been well characterized. The limited followup in prior reports has enabled demonstration of short-term kidney function
improvement. Only two studies have explored kidney function over a longer time period
(Table 3). Our study fills this gap by investigating long-term kidney function with
detailed information on the change in CKD stage and eGFR after VAD implantation.
Hasin et al illustrated improvement in eGFR post implant to 6 months in 83 VAD
patients, and Kirklin et al illustrated improvement in serum creatinine and BUN up to 2
years post implant in 4917 VAD patients.(14, 18)(14, 18)14, 16, 17, 18, 22 However, both of
these studies assessed kidney function over time as continuous measures. Assessing
changes in kidney function that reclassifies the CKD stage provides a clinically
meaningful metric. Our findings confirm previous observations that renal function
improves after VAD implantation and demonstrates that the improvements are clinically
meaningful and sustained.
The greatest improvements in kidney function are realized early, in the first 14
days post implant, as organ perfusion is increased. Although improvement in kidney
function are sustained, early gains in eGFR are attenuated after 14 days. Therefore gain in
kidney function after 14 days may more closely represent the organ capacity that was
previously masked due to poor kidney perfusion in these patients with poor cardiac
function.
Decreased kidney function is the most common contraindication to heart
transplantation.(19) Previous studies have shown that among patients with HF receiving
heart transplants, those maintained on VADs (vs. inotrope support) have better kidney
75
function at time of transplantation.(20) Therefore our results, wherein 55.8% of VAD
patients demonstrating sustained improvement in kidney function are of particular
importance for three reasons. First, it suggests the potential benefit of VAD implantation
over inotrope support, a hypothesis being investigated in REVIVE-IT study. Second,
kidney function improvement (or lack thereof) post-VAD implantation facilitates
assessment of organ capacity that was previously masked due to poor kidney perfusion in
these patients with poor cardiac function. Third, sustained improvement in kidney
function improves the candidacy for receiving a heart transplant. The latter is supported
by our findings wherein patients who demonstrated sustained kidney function
improvement over time were more likely to than those who did not sustain kidney
function over time.
In our study, neither baseline kidney function nor improvement in kidney function
was associated with thromboembolism or hemorrhage. This is contrary to reports from
non-VAD patient populations that have demonstrated that kidney function improvement
is associated with a decrease in risk of thromboembolism and hemorrhage.(21-24) The
lack of association in our study could be due to two reasons. First, patients who
demonstrated sustained improved kidney function on VAD support were more likely to
be transplanted, and therefore censored from further analysis. Second, unlike non-VAD
patient populations, patients with VADs have a different risk profile for these events due
to the complex nature of the device itself and its effects on the coagulation system and the
vasculature. Given these considerations, along with the baseline risk in patients with
cardiovascular disease in general, improvement in kidney function alone, even if
sustained over a one year period, may not be enough to outweigh the risk associated with
76
the VAD device itself or mitigate the risk associated with longstanding cardiovascular
disease. Moreover, the one-year follow-up time may have limited our ability to detect
associations between kidney function and thromboembolism or hemorrhage.
Decreased kidney function at prior to VAD implant was associated with higher
mortality consistent with reports in advanced heart failure patients. Although
improvement in kidney function decreases mortality in non-VAD patients, our findings
did not demonstrate decrease in mortality among VAD recipients. This may be explained
the higher and long standing disease burden in patients with advanced heart failure.
Although VAD implantation improves kidney function and reverses the cardiorenal
syndrome, the improvements do not attenuate the risk of mortality.
While previous studies have illustrated the influence of pre-VAD kidney function
on post-VAD mortality, these studies have primarily assessed eGFR, creatinine and BUN
as continuous measures, and their subsequent association with mortality. Consistent with
previous studies, poor kidney function pre-VAD is associated with a higher mortality
rate.(18) This is the first study to investigate CKD stage pre-VAD implantation and the
association with mortality post-VAD implantation. Furthermore, we categorized CKD
stage 3 into stage 3a (eGFR 45-59 ml/min) and stage 3b (eGFR 30-44 ml/min) as Levey
et al have demonstrated clinically meaningful differences in mortality among the two
groups.(25) To our knowledge, this report is the first to report that VAD patients stage 3b
CKD have a higher mortality, compared to those with stage 3a CKD.
The study has several strengths including a large (n=228) sample size, a racially
diverse population, uniformity of care at a single institution, detailed clinical
documentation, and ascertainment of clinically relevant events with minimal loss (n=4;
77
patient transferred care out-of-state) to follow-up. Kidney function was assessed at
multiple time points, and improvement in kidney function is defined using a clinically
relevant rubric. However we recognize its limitations such as lack of ascertainment of
biomarkers. We recognize that this is a single center study. Therefore independent
validation of our findings in larger datasets is needed to confirm our findings.
Regardless of baseline kidney function, most patients experience an improvement
in kidney function after VAD implantation. Regardless of the improvement in kidney
function post VAD implant, risk of mortality is driven by baseline kidney function.
Patients with CKD stage 3b, 4 or 5 prior to VAD are ay an increased risk of mortality
post-VAD implantation. This has important implications on patient selection during
evaluation for VAD therapy. Further research is needed to establish whether patients with
decreased kidney function due to advancing heart failure would benefit from earlier
implantation of a VAD to reduce the mortality associated with advancing kidney
dysfunction.
78
Reference List
1.
Coresh J, Selvin E, Stevens LA, Manzi J, Kusek JW, Eggers P, Van Lente F and
Levey AS. Prevalence of chronic kidney disease in the United States. Jama.
2007;298:2038-47.
2.
Adams KF, Jr., Fonarow GC, Emerman CL, LeJemtel TH, Costanzo MR,
Abraham WT, Berkowitz RL, Galvao M and Horton DP. Characteristics and
outcomes of patients hospitalized for heart failure in the United States: rationale,
design, and preliminary observations from the first 100,000 cases in the Acute
Decompensated Heart Failure National Registry (ADHERE). Am Heart J.
2005;149:209-16.
3.
Smith GL, Lichtman JH, Bracken MB, Shlipak MG, Phillips CO, DiCapua P and
Krumholz HM. Renal impairment and outcomes in heart failure: systematic
review and meta-analysis. J Am Coll Cardiol. 2006;47:1987-96.
4.
U.S. Renal Data System, USRDS 2013 Annual Data Report: Atlas of Chornic
Kidney Disease and End-Stage Renal Disease in the United States. 2013.
5.
Ronco C, Haapio M, House AA, Anavekar N and Bellomo R. Cardiorenal
syndrome. J Am Coll Cardiol. 2008;52:1527-39.
6.
McAlister FA, Ezekowitz J, Tarantini L, Squire I, Komajda M, Bayes-Genis A,
Gotsman I, Whalley G, Earle N, Poppe KK and Doughty RN. Renal dysfunction
in patients with heart failure with preserved versus reduced ejection fraction:
impact of the new Chronic Kidney Disease-Epidemiology Collaboration Group
formula. Circ Heart Fail. 2012;5:309-14.
7.
Hillege HL, Girbes AR, de Kam PJ, Boomsma F, de Zeeuw D, Charlesworth A,
Hampton JR and van Veldhuisen DJ. Renal function, neurohormonal activation,
and survival in patients with chronic heart failure. Circulation. 2000;102:203-10.
8.
McClellan WM, Flanders WD, Langston RD, Jurkovitz C and Presley R. Anemia
and renal insufficiency are independent risk factors for death among patients with
congestive heart failure admitted to community hospitals: a population-based
study. J Am Soc Nephrol. 2002;13:1928-36.
9.
McAlister FA, Ezekowitz J, Tonelli M and Armstrong PW. Renal insufficiency
and heart failure: prognostic and therapeutic implications from a prospective
cohort study. Circulation. 2004;109:1004-9.
10.
Hillege HL, Nitsch D, Pfeffer MA, Swedberg K, McMurray JJ, Yusuf S, Granger
CB, Michelson EL, Ostergren J, Cornel JH, de Zeeuw D, Pocock S and van
Veldhuisen DJ. Renal function as a predictor of outcome in a broad spectrum of
patients with heart failure. Circulation. 2006;113:671-8.
79
11.
Jessup M, Albert NM, Lanfear DE, Lindenfeld J, Massie BM, Walsh MN and
Zucker MJ. ACCF/AHA/HFSA 2011 survey results: current staffing profile of
heart failure programs, including programs that perform heart transplant and
mechanical circulatory support device implantation. J Card Fail. 2011;17:349-58.
12.
Roger VL, Go AS, Lloyd-Jones DM, Adams RJ, Berry JD, Brown TM, Carnethon
MR, Dai S, de Simone G, Ford ES, Fox CS, Fullerton HJ, Gillespie C, Greenlund
KJ, Hailpern SM, Heit JA, Ho PM, Howard VJ, Kissela BM, Kittner SJ, Lackland
DT, Lichtman JH, Lisabeth LD, Makuc DM, Marcus GM, Marelli A, Matchar
DB, McDermott MM, Meigs JB, Moy CS, Mozaffarian D, Mussolino ME, Nichol
G, Paynter NP, Rosamond WD, Sorlie PD, Stafford RS, Turan TN, Turner MB,
Wong ND and Wylie-Rosett J. Heart disease and stroke statistics--2011 update: a
report from the American Heart Association. Circulation. 2011;123:e18-e209.
13.
Demirozu ZT, Etheridge WB, Radovancevic R and Frazier OH. Results of
HeartMate II left ventricular assist device implantation on renal function in
patients requiring post-implant renal replacement therapy. The Journal of heart
and lung transplantation : the official publication of the International Society for
Heart Transplantation. 2011;30:182-7.
14.
Hasin T, Topilsky Y, Schirger JA, Li Z, Zhao Y, Boilson BA, Clavell AL,
Rodeheffer RJ, Frantz RP, Edwards BS, Pereira NL, Joyce L, Daly R, Park SJ and
Kushwaha SS. Changes in renal function after implantation of continuous-flow
left ventricular assist devices. J Am Coll Cardiol. 2012;59:26-36.
15.
De Schryver EL, van Gijn J, Kappelle LJ, Koudstaal PJ and Algra A. Nonadherence to aspirin or oral anticoagulants in secondary prevention after
ischaemic stroke. J Neurol. 2005;252:1316-21.
16.
Valente MA, Hillege HL, Navis G, Voors AA, Dunselman PH, van Veldhuisen
DJ and Damman K. The Chronic Kidney Disease Epidemiology Collaboration
equation outperforms the Modification of Diet in Renal Disease equation for
estimating glomerular filtration rate in chronic systolic heart failure. Eur J Heart
Fail. 2014;16:86-94.
17.
Levin A and Stevens PE. Summary of KDIGO 2012 CKD Guideline: behind the
scenes, need for guidance, and a framework for moving forward. Kidney Int.
2014;85:49-61.
18.
Kirklin JK, Naftel DC, Kormos RL, Pagani FD, Myers SL, Stevenson LW,
Givertz MM and Young JB. Quantifying the effect of cardiorenal syndrome on
mortality after left ventricular assist device implant. The Journal of heart and
lung transplantation : the official publication of the International Society for
Heart Transplantation. 2013;32:1205-13.
80
19.
Kirklin JK, Naftel DC, Kormos RL, Stevenson LW, Pagani FD, Miller MA,
Ulisney KL, Baldwin JT and Young JB. Third INTERMACS Annual Report: the
evolution of destination therapy in the United States. J Heart Lung Transplant.
2011;30:115-23.
20.
Bank AJ, Mir SH, Nguyen DQ, Bolman RM, 3rd, Shumway SJ, Miller LW,
Kaiser DR, Ormaza SM and Park SJ. Effects of left ventricular assist devices on
outcomes in patients undergoing heart transplantation. Ann Thorac Surg.
2000;69:1369-74; discussion 1375.
21.
Wattanakit K, Cushman M, Stehman-Breen C, Heckbert SR and Folsom AR.
Chronic kidney disease increases risk for venous thromboembolism. J Am Soc
Nephrol. 2008;19:135-40.
22.
Folsom AR, Lutsey PL, Astor BC, Wattanakit K, Heckbert SR and Cushman M.
Chronic kidney disease and venous thromboembolism: a prospective study.
Nephrol Dial Transplant. 2010;25:3296-301.
23.
Attallah N, Yassine L, Fisher K and Yee J. Risk of bleeding and restenosis among
chronic kidney disease patients undergoing percutaneous coronary intervention.
Clin Nephrol. 2005;64:412-8.
24.
Limdi NA, Beasley TM, Baird MF, Goldstein JA, McGwin G, Arnett DK, Acton
RT and Allon M. Kidney function influences warfarin responsiveness and
hemorrhagic complications. J Am Soc Nephrol. 2009;20:912-21.
25.
Levey AS, de Jong PE, Coresh J, El Nahas M, Astor BC, Matsushita K,
Gansevoort RT, Kasiske BL and Eckardt KU. The definition, classification, and
prognosis of chronic kidney disease: a KDIGO Controversies Conference report.
Kidney Int. 2011;80:17-28.
26.
Borgi J, Tsiouris A, Hodari A, Cogan CM, Paone G and Morgan JA. Significance
of postoperative acute renal failure after continuous-flow left ventricular assist
device implantation. The Annals of thoracic surgery. 2013;95:163-9.
27.
Iwashima Y, Yanase M, Horio T, Seguchi O, Murata Y, Fujita T, Toda K,
Kawano Y and Nakatani T. Serial changes in renal function as a prognostic
indicator in advanced heart failure patients with left ventricular assist system. The
Annals of thoracic surgery. 2012;93:816-23.
28.
Hasin T, Marmor Y, Kremers W, Topilsky Y, Severson CJ, Schirger JA, Boilson
BA, Clavell AL, Rodeheffer RJ, Frantz RP, Edwards BS, Pereira NL, Stulak JM,
Joyce L, Daly R, Park SJ and Kushwaha SS. Readmissions after implantation of
axial flow left ventricular assist device. J Am Coll Cardiol. 2013;61:153-63.
81
29.
Butler J, Geisberg C, Howser R, Portner PM, Rogers JG, Deng MC and Pierson
RN, 3rd. Relationship between renal function and left ventricular assist device
use. The Annals of thoracic surgery. 2006;81:1745-51.
30.
Genovese EA, Dew MA, Teuteberg JJ, Simon MA, Bhama JK, Bermudez CA,
Lockard KL, Winowich S and Kormos RL. Early adverse events as predictors of
1-year mortality during mechanical circulatory support. The Journal of heart and
lung transplantation : the official publication of the International Society for
Heart Transplantation. 2010;29:981-8.
31.
Ma L, Fujino Y, Matsumiya G, Sawa Y and Mashimo T. Renal function with left
ventricular assist devices: the poorer the preoperative renal function, the longer
the recovery. Medical science monitor : international medical journal of
experimental and clinical research. 2008;14:CR621-7.
32.
Russell SD, Rogers JG, Milano CA, Dyke DB, Pagani FD, Aranda JM, Klodell
CT, Jr., Boyle AJ, John R, Chen L, Massey HT, Farrar DJ and Conte JV. Renal
and hepatic function improve in advanced heart failure patients during
continuous-flow support with the HeartMate II left ventricular assist device.
Circulation. 2009;120:2352-7.
33.
Sandner SE, Zimpfer D, Zrunek P, Rajek A, Schima H, Dunkler D, Grimm M,
Wolner E and Wieselthaler GM. Renal function and outcome after continuous
flow left ventricular assist device implantation. The Annals of thoracic surgery.
2009;87:1072-8.
34.
Sandner SE, Zimpfer D, Zrunek P, Dunkler D, Schima H, Rajek A, Grimm M,
Wolner E and Wieselthaler GM. Renal function after implantation of continuous
versus pulsatile flow left ventricular assist devices. The Journal of heart and lung
transplantation : the official publication of the International Society for Heart
Transplantation. 2008;27:469-73.
35.
Singh M, Shullo M, Kormos RL, Lockard K, Zomak R, Simon MA, Bermudez C,
Bhama J, McNamara D, Toyoda Y and Teuteberg JJ. Impact of renal function
before mechanical circulatory support on posttransplant renal outcomes. The
Annals of thoracic surgery. 2011;91:1348-54.
82
Table 1. Demographic and Clinical Characteristics for the Entire Cohort Stratified by Kidney Function prior to VAD
Implantation
Age
Body Mass Index
Male
Black Race
Status within the First Year
Deceased
Transplanted
Recovered
Stage 1
eGFR ≥90
(N=32)
Stage 2
eGFR 6089
(N=49)
Stage 3A
Stage 3B
eGFR 45eGFR 3059
44
(N=53)
(N=53)
Mean (SD)
51.7 ±13.4
56.2±14
27.1±7.8
26.7±5.8
N (%)
Stage 4
eGFR 1530
(N=30)
Stage 5
eGFR <15
(N=11)
p-value
43.2 ±15.9
30.9 ±7.7
51.1 ±15.6
27.6±8.8
56.8±11.4
30.3±8.2
52.6±12.6
25.4±4.6
0.002
0.11
25 (78)
14 (44)
39 (80)
16 (33)
35 (66)
16 (30)
41 (77)
10 (19)
20 (67)
6 (20)
7 (64)
2 (18)
0.49
0.14
0.06
4 (12)
10 (31)
0 (0)
11 (22)
12 (25)
1 (2)
17 (32)
8 (15)
2 (4)
19 (36)
4 (8)
0 (0)
14 (47)
7 (23)
0
8 (73)
2 (18)
0
11 (34)
21 (66)
18 (37)
31 (63)
25 (47)
28 (53)
23 (43)
30 (57)
19 (63)
11 (37)
11 (100)
0
8 (25)
16 (50)
8 (25)
28 (57)
14 (28)
7 (14)
25 (47)
23 (44)
5 (9)
32 (60)
13 (25)
8 (15)
17 (57)
10 (33)
3 (10)
6 (55)
3 (27)
2 (18)
23 (72)
31 (66)
31 (58)
25 (48)
13 (43)
8 (73)
Device Type
Pulsatile Flow
Continuous Flow
Etiology of Heart Failure
Ischemic
Idiopathic
Other
Indication for VAD
Bridge to transplant
0.001
0.17
0.07
Bridge to candidacy
Bridge to recovery
Destination
Comorbid Conditions
Diabetes
Chronic Obstructive Pulmonary
Disease
Congestive Heart Failure
Peripheral Vascular Disease
Gout
Right Ventricular Dysfunction
Coronary Artery Disease
History of Myocardial Infarction
Hypertension
Pulmonary Hypertension
Hyperlipidemia
Atrial Fibrillation
Ventricular Tachycardia
1 (3)
0
8 (25)
3 (6)
0
13 (28)
3 (6)
0
19 (36)
3 (6)
2 (4)
22 (42)
5 (17)
1 (3)
11 (37)
2 (18)
1 (9)
0
9 (28)
2 (6)
11 (22)
1 (2)
24 (45)
2 (4)
27 (51)
5 (9)
13 (43)
1 (3)
2 (18)
0
0.019
0.51
27 (84)
2 (6)
3 (9)
6 (19)
8 (25)
5 (16)
15 (47)
12 (37)
8 (25)
10 (31)
11 (34)
44 (90)
2 (4)
9 (18)
14 (29)
13 (26)
8 (16)
28 (57)
7 (14)
19 (39)
16 (33)
19 (39)
43 (81)
2 (4)
5 (9)
8 (15)
22 (41)
11 (21)
38 (72)
13 (24)
22 (41)
19 (36)
28 (53)
44 (83)
1 (2)
4 (8)
14 (26)
17 (32)
14 (26)
27 (51)
10 (19)
20 (38)
18 (34)
16 (30)
24 (80)
1 (3)
5 (17)
5 (17)
14 (47)
8 (27)
10 (33)
6 (20)
15 (50)
10 (33)
16 (53)
6 (55)
0
2 (18)
3 (27)
2 (18)
2 (18)
4 (36)
1 (9)
1 (9)
0
3 (27)
0.15
0.90
0.50
0.52
0.21
0.74
0.014
0.16
0.14
0.34
0.11
Kidney function is presented as CKD stages as recommended by the Kidney Disease Outcomes Quality Initiative guidelines using
estimated glomerular filtration rate (eGFR) from the Chronic Kidney Disease Epidemiology Collaboration Formula (CKD-EPI).
Table 2. Association of Kidney Function at Baseline with Risk (Hazard Ratio (HR)
and 95% Confidence Interval(CI)) of Thromboembolism and Hemorrhage
Hazard Ratio
Thromboembolic Events
Stage of Kidney Disease at Baseline
Stage 1 (eGFR≥90)
Stage 2 (eGFR 60-89)
Stage 3a (eGFR 45-59)
Stage 3b (eGFR 30-44)
Stage 4 (eGFR 15-29)
Stage 5 (eGFR <15)
Hemorrhagic Events
Stage of Kidney Disease at Baseline
Stage 1 (eGFR≥90)
Stage 2 (eGFR 60-89)
Stage 3a (eGFR 45-59)
Stage 3b (eGFR 30-44)
Stage 4 (eGFR 15-29)
Stage 5 (eGFR <15)
95 % Confidence
Interval
p-value
ref
0.82
1.47
1.75
1.52
0.94
ref
0.23-2.94
0.47-4.62
0.55-5.58
0.41-5.68
0.10-8.38
ref
0.77
0.51
0.35
0.53
0.95
ref
1.55
1.71
2.04
0.99
n/a
ref
0.56-4.32
0.63-4.67
0.73-5.67
0.26-3.70
n/a
ref
0.40
0.29
0.17
0.99
n/a
Kidney function is presented as CKD stages as recommended by the Kidney Disease
Outcomes Quality Initiative guidelines using estimated glomerular filtration rate (eGFR)
from the Chronic Kidney Disease Epidemiology Collaboration Formula (CKD-EPI). The
groups are Stage 1 (eGFR≥90 ml/min/1.73m2), Stage 2 (eGFR 60-90 ml/min/1.73m2),
Stage 3a (eGRF 45-59 ml/min/1.73m2), Stage 3b (eGFR 30-44 ml/min/1.73m2), Stage 4
(eGFR 15-30 ml/min/1.73m2) and Stage 5 (eGFR<15 ml/min/1.73m2). Follow-up is
calculated as time from VAD to first event or end of study period (1 year or time until
explantation/transplantation) for thromboembolic events and hemorrhagic events.
Follow-up time for incidence of death is calculated as time to death or end of study
period (1 year or explantation/transplantation. Cox proportional hazards models were
used to assess the risk of thromboembolism and hemorrhage. The proportional hazards
assumption was met.
85 Table 3. Prior VAD Studies that Investigate Kidney Function post VAD
Implantation and the Relationship between Kidney Function and Mortality
Authors
Sample
Size
4,917
Definition of Kidney
Function
eGFR <30 for severe
kidney disease and
eGFR 30-60 for
moderate kidney
disease
Borgi
(2013)26
100
Acute Renal Failure
(ARF)
Iwashi
ma et al
(2012)27
110
Change in estimate
dglomerular filtration
rate (eGFR) from
baseline to 2 weeks
Assessed at
baseline and 2
weeks
Hasin et
al
(2012)28
83
eGFR
eGFR
Butler
et al
(2006)29
220
Renal function as
defined by creatinine
clearance
Assessed before
VAD, and 1,3
and 6 months
after VAD
Assessed 1 month
post VAD
Genove
se et al
(2010)30
163
Acute Renal Failure
(ARF)
Assessed 60 days
post implantation
One year
mortality
Ma et al
(2008)31
28
Renal function as
measured through
creatinine
Length of
recovery
Russell
et al
(2009)32
Sandner
et al
(2009)33
309
Renal function as
defined by BUN and
creatinine
Pre-VAD eGFR
Assessed at
baseline and 1
month post
implant
Assessed at
baseline and 6
months post VAD
Assessed prior to
VAD
Sandner
et al
(2008)34
92
Renal function as
defined by creatinine
and eGFR
Assessed at
baseline and 3
months
Renal function,
and mortality at 3
months
Singh et
al
(2011)35
116
Creatinine clearance
prior to VAD
Assessed prior to
VAD
Creatinine
clearance 1
month after VAD
implant
Kirklin
(2013)18
86
Assessment of
Kidney Function
Assessed at
baseline, BUN
and creatinine
were assessed up
to 4 years after
implantation
Assessed at
baseline and 7
days
86 Outcome of
Interest
Death, change in
serum creatinine
and BUN
Key Findings
Change in KF
and its influence
on 1- year
mortality
Mortality at 2
years
Postoperative ARF is
associated with mortality
All-cause
mortality, and
disease specific
mortality at 1
year
Renal function as
defined by BUN
and creatinine
Mortality postVAD to 6 months
Pre-implant renal
dysfunction is related to
higher mortality post VAD
implant.
Impaired renal function as
well as renal function that
does not improve with VAD
placement are predictors of
mortality
Renal function improves
after VAD implantation
Poor renal function at
baseline is associated with
worse outcomes. Patients
with improving renal
function after VAD
placement experience
improved outcomes
The presence of some
adverse events increased the
risk of mortality, namely
renal events, respiratory
events, bleeding events and
reoperation
Pre-VAD renal function is
predictive of post-VAD renal
function and length of stay in
the ICU
Implantation of a VAD
improves renal function
Patients with pre-VAD renal
dysfunction have poorer
outcomes than patients
without pre-VAD renal
dysfunction. Regardless of
pre-VAD renal status, renal
function improves after
VAD outcome
There is no difference
between continuous flow and
pulsatile flow devices on
renal function
VAD use improves renal
function after implantation
Figure 1. Change in Kidney Function over 1 year follow-up from VAD Implantation
VAD
(N=228)
Stage 1
Stage 2
Stage 3a
Stage 3b
Stage 4
Stage 5
N
32
49
53
53
30
11
%
14
21
23
23
13
5
14
Days
(N=200)*
1
Month
(N=188)
3
Months
(N=158)§
6
Months
(N=136)†
9 Months
(N=121)‡
N
93
40
32
23
11
1
N
72
51
30
21
12
2
N
54
49
24
25
5
1
N
31
44
22
25
11
3
N
24
36
30
23
6
2
%
46
20
16
11
5
1
**
%
38
27
16
11
6
1
%
34
31
15
16
3
1
%
23
32
16
18
8
2
12
Months
(N=104)
¥
%
20
30
25
19
5
2
N
22
31
25
15
10
1
%
21
30
24
14
10
1
Kidney function is presented as CKD stages as recommended by the Kidney Disease
Outcomes Quality Initiative guidelines using estimated glomerular filtration rate (eGFR)
from the Chronic Kidney Disease Epidemiology Collaboration Formula (CKD-EPI). The
groups are Stage 1 (eGFR≥90 ml/min/1.73m2), Stage 2 (eGFR 60-90 ml/min/1.73m2),
Stage 3a (eGRF 45-59 ml/min/1.73m2), Stage 3b (eGFR 30-44 ml/min/1.73m2), Stage 4
(eGFR 15-30 ml/min/1.73m2) and Stage 5 (eGFR<15 ml/min/1.73m2)
87 *
A total of 28 patients was not included in this assessment due to 2 patients being
transplanted, 1 patient recovered and 25 patients died
**
§
An additional12 patients were not included in this assessment due to patient death
An additional30 patients were not included in this assessment due to 4 patient
withdrawals, 2 patients recovered, 11 patients had a transplant and 13 patients died
†
An additional 22 patients were not included in this assessment due to 10 patient deaths,
11 patients were transplanted and 1 patient withdrew
‡
An additional 15 patients were not included in this assessment due to 5 patient deaths
and 10 transplants
¥
An additional 17 patients were not included in this assessment due to 8 patient deaths
and 9 transplants
88 Figure 2a. Relative Risk Estimates for Mortality Stratified by Baseline Kidney
Function
Kidney function is presented as CKD stages as recommended by the Kidney Disease
Outcomes Quality Initiative guidelines using estimated glomerular filtration rate (eGFR)
from the Chronic Kidney Disease Epidemiology Collaboration Formula (CKD-EPI). The
groups are Stage 1 (eGFR≥90 ml/min/1.73m2), Stage 2 (eGFR 60-90 ml/min/1.73m2),
Stage 3a (eGRF 45-59 ml/min/1.73m2), Stage 3b (eGFR 30-44 ml/min/1.73m2), Stage 4
(eGFR 15-30 ml/min/1.73m2) and Stage 5 (eGFR<15 ml/min/1.73m2). Cox proportional
hazards models were used to assess the risk of death. The proportional hazards
assumption was not met, however Cox proportional hazards models are robust to
moderate deviations in proportionality and the deviations in this model were minor.
89 Figure 2b. Adjusted Relative Risk Estimates for Mortality Stratified by baseline
CKD Stage
Kidney function is presented as CKD stages as recommended by the Kidney Disease
Outcomes Quality Initiative guidelines using estimated glomerular filtration rate (eGFR)
from the Chronic Kidney Disease Epidemiology Collaboration Formula (CKD-EPI). The
groups are Stage 1 (eGFR≥90 ml/min/1.73m2), Stage 2 (eGFR 60-90 ml/min/1.73m2),
Stage 3a (eGRF 45-59 ml/min/1.73m2), Stage 3b (eGFR 30-44 ml/min/1.73m2), Stage 4
(eGFR 15-30 ml/min/1.73m2) and Stage 5 (eGFR<15 ml/min/1.73m2). Cox proportional
hazards models were used to assess the adjusted risk of death. The proportional hazards
assumption was met.
90 SUMMARY CONCLUSIONS
Heart failure is a major concern in the United States with the prevalence
increasing.(1,2) The treatment options for heart failure range with the degree of severity.
Medication and lifestyle changes can aid in controlling symptoms and decrease disease
progression for those with mild and moderate heart failure, however the gold standard
treatment for advanced heart failure remains the heart transplant. The use of the
ventricular assist device (VAD) as a treatment option for advanced heart failure patients
is increasing. While the VAD has increased the survival rate and quality of life there is
still a considerable amount of morbidity and mortality associated with VAD use.(20)
VAD patients are at increased risk for thromboembolism, hemorrhage and death.
A retrospective analysis of patients under the care of the Mechanical Circulatory
Support Clinic at UAB was conducted to assess predictors of thromboembolism. We
assessed traditional risk factors for thromboembolism, clinical factors and the biomarker
lactate dehydrogenase (LDH). In this study we found an association between pre-VAD
LDH and post-VAD thromboembolism. Furthermore, we show that early LDH elevation
within the first month post-VAD implant, in addition to kidney function prior to VAD
implantation and post-VAD anticoagulation control, are significant risk factors for
thromboembolism. We found that the risk for thromboembolism was 5 fold higher for
patients with early LDH elevation, 5 fold higher for patients with poor kidney function
91 prior to VAD implantation and a decreased risk of thromboembolism for patients with
adequate anticoagulation control.
A retrospective analysis of patients under the care of the Mechanical Circulatory
Support Clinic at UAB was conducted to create a VAD specific bleeding risk score and
compare it to the previously developed HAS-BLED risk score, a score designed to
predict hemorrhage in other anticoagulated patients. We found that the HAS-BLED
score is not as predictive in the VAD population as it is in other chronic anticoagulation
populations. We found that patients with a VAD bleeding risk score of ≥3 are at 4 times
the risk of hemorrhage than patients with a lower VAD bleeding risk score. Further
studies are needed to validate this score in other VAD populations.
A retrospective analysis of patients under the care of the Mechanical Circulatory
Support Clinic at UAB was conducted to assess the change in kidney function over time
post-VAD implant, and evaluate whether change in kidney function post-VAD is
associated with thromboembolism, hemorrhage and death. Regardless of baseline CKD
stage, most patients experience an improvement in CKD stage after VAD implantation.
These results support the theories that while heart failure induced hypoperfusion of the
kidneys may cause kidney dysfunction, the dysfunction can be partially reversed once
perfusion of the kidneys is restored. Regardless of the improvement in kidney function
post VAD implant, patients with CKD stage 3b, 4 or 5 prior to VAD are at an increased
risk of mortality post-VAD implantation. Our study further illustrates the importance of
kidney function in heart failure patients; and how baseline kidney function remains a
strong predictor of mortality post-VAD regardless of improvements due to VAD
implantation. Further research is needed to establish whether patients with decreased
92 kidney function due to advancing heart failure would benefit from earlier implantation of
a VAD to reduce the mortality associated with advancing kidney dysfunction.
These studies identify risk factors for thromboembolism, hemorrhage and
mortality in VAD patients. Identifying risk factors for these outcomes is paramount in
this high risk patient population. The identification of subgroups with significant risk
factors can be identified and managed more closely to possibly prevent
thromboembolism and hemorrhage and reduce the mortality rate associated with VAD
use. In conclusion, in these three studies, we established the biomarker LDH is
associated with risk of thromboembolism, established a VAD Bleeding Risk score to
identify VAD patients at greatest risk of hemorrhage, and identified pre-VAD kidney
function as a strong risk factor for mortality post-VAD.
93 GENERAL LIST OF REFERENCES
1.
Roger VL, Go AS, Lloyd-Jones DM, Adams RJ, Berry JD, Brown TM, et al.
Heart disease and stroke statistics--2011 update: a report from the American Heart
Association. Circulation. 2011 Feb 1;123(4):e18-e209.
2.
Jessup M, Albert NM, Lanfear DE, Lindenfeld J, Massie BM, Walsh MN, et al.
ACCF/AHA/HFSA 2011 survey results: current staffing profile of heart failure
programs, including programs that perform heart transplant and mechanical
circulatory support device implantation: a report of the ACCF Heart Failure and
Transplant Committee, AHA Heart Failure and Transplantation Committee, and
Heart Failure Society of America. Circ Heart Fail. 2011 May;4(3):378-87.
3.
Wilson SR, Givertz MM, Stewart GC, Mudge GH, Jr. Ventricular assist devices
the challenges of outpatient management. J Am Coll Cardiol. 2009 Oct
27;54(18):1647-59.
4.
Haeck ML, Hoogslag GE, Rodrigo SF, Atsma DE, Klautz RJ, van der Wall EE, et
al. Treatment options in end-stage heart failure: where to go from here? Neth
Heart J. 2011 Nov 15.
5.
Torpy JM, Lynm C, Golub RM. JAMA patient page. Heart failure. JAMA. 2011
Nov 16;306(19):2175.
6.
Goldstein DJ, Oz MC, Rose EA. Implantable left ventricular assist devices. N
Engl J Med. 1998 Nov 19;339(21):1522-33.
7.
Frazier OH, Benedict CR, Radovancevic B, Bick RJ, Capek P, Springer WE, et al.
Improved left ventricular function after chronic left ventricular unloading. Ann
Thorac Surg. 1996 Sep;62(3):675-81; discussion 81-2.
8.
Mancini DM, Beniaminovitz A, Levin H, Catanese K, Flannery M, DiTullio M, et
al. Low incidence of myocardial recovery after left ventricular assist device
implantation in patients with chronic heart failure. Circulation. 1998 Dec
1;98(22):2383-9.
9.
Oosterom L, de Jonge N, Kirkels J, Klopping C, Lahpor J. Left ventricular assist
device as a bridge to recovery in a young woman admitted with peripartum
cardiomyopathy. Neth Heart J. 2008 Dec;16(12):426-8.
10.
Simon MA, Kormos RL, Murali S, Nair P, Heffernan M, Gorcsan J, et al.
Myocardial recovery using ventricular assist devices: prevalence, clinical
characteristics, and outcomes. Circulation. 2005 Aug 30;112(9 Suppl):I32-6.
94 11.
Frazier OH, Rose EA, Macmanus Q, Burton NA, Lefrak EA, Poirier VL, et al.
Multicenter clinical evaluation of the HeartMate 1000 IP left ventricular assist
device. Ann Thorac Surg. 1992 Jun;53(6):1080-90.
12.
Miller LW, Pagani FD, Russell SD, John R, Boyle AJ, Aaronson KD, et al. Use of
a continuous-flow device in patients awaiting heart transplantation. N Engl J Med.
2007 Aug 30;357(9):885-96.
13.
Pennington DG, McBride LR, Peigh PS, Miller LW, Swartz MT. Eight years'
experience with bridging to cardiac transplantation. J Thorac Cardiovasc Surg.
1994 Feb;107(2):472-80; discussion 80-1.
14.
Catanese KA, Goldstein DJ, Williams DL, Foray AT, Illick CD, Gardocki MT, et
al. Outpatient left ventricular assist device support: a destination rather than a
bridge. Ann Thorac Surg. 1996 Sep;62(3):646-52; discussion 53.
15.
Korfer R, El-Banayosy A, Arusoglu L, Minami K, Korner MM, Kizner L, et al.
Single-center experience with the thoratec ventricular assist device. J Thorac
Cardiovasc Surg. 2000 Mar;119(3):596-600.
16.
McBride LR, Naunheim KS, Fiore AC, Moroney DA, Swartz MT. Clinical
experience with 111 thoratec ventricular assist devices. Ann Thorac Surg. 1999
May;67(5):1233-8; discussion 8-9.
17.
Moskowitz AJ, Weinberg AD, Oz MC, Williams DL. Quality of life with an
implanted left ventricular assist device. Ann Thorac Surg. 1997 Dec;64(6):17649.
18.
Portner PM, Oyer PE, Pennington DG, Baumgartner WA, Griffith BP, Frist WR,
et al. Implantable electrical left ventricular assist system: bridge to transplantation
and the future. Ann Thorac Surg. 1989 Jan;47(1):142-50.
19.
Hosenpud JD, Bennett LE, Keck BM, Boucek MM, Novick RJ. The Registry of
the International Society for Heart and Lung Transplantation: seventeenth official
report-2000. J Heart Lung Transplant. 2000 Oct;19(10):909-31.
20.
Rose EA, Gelijns AC, Moskowitz AJ, Heitjan DF, Stevenson LW, Dembitsky W,
et al. Long-term use of a left ventricular assist device for end-stage heart failure.
N Engl J Med. 2001 Nov 15;345(20):1435-43.
21.
Jaski BE, Kim JC, Naftel DC, Jarcho J, Costanzo MR, Eisen HJ, et al. Cardiac
transplant outcome of patients supported on left ventricular assist device vs.
intravenous inotropic therapy. J Heart Lung Transplant. 2001 Apr;20(4):449-56.
95 22.
Stehlik J, Edwards LB, Kucheryavaya AY, Aurora P, Christie JD, Kirk R, et al.
The Registry of the International Society for Heart and Lung Transplantation:
twenty-seventh official adult heart transplant report--2010. J Heart Lung
Transplant. 2010 Oct;29(10):1089-103.
23.
Birks EJ, Tansley PD, Hardy J, George RS, Bowles CT, Burke M, et al. Left
ventricular assist device and drug therapy for the reversal of heart failure. N Engl
J Med. 2006 Nov 2;355(18):1873-84.
24.
Slaughter MS, Rogers JG, Milano CA, Russell SD, Conte JV, Feldman D, et al.
Advanced heart failure treated with continuous-flow left ventricular assist device.
N Engl J Med. 2009 Dec 3;361(23):2241-51.
25.
Frazier OH. First use of an untethered, vented electric left ventricular assist device
for long-term support. Circulation. 1994 Jun;89(6):2908-14.
26.
Frazier OH, Rose EA, McCarthy P, Burton NA, Tector A, Levin H, et al.
Improved mortality and rehabilitation of transplant candidates treated with a longterm implantable left ventricular assist system. Ann Surg. 1995 Sep;222(3):32736; discussion 36-8.
27.
Frazier OH, Rose EA, Oz MC, Dembitsky W, McCarthy P, Radovancevic B, et
al. Multicenter clinical evaluation of the HeartMate vented electric left ventricular
assist system in patients awaiting heart transplantation. J Thorac Cardiovasc Surg.
2001 Dec;122(6):1186-95.
28.
Oz MC, Argenziano M, Catanese KA, Gardocki MT, Goldstein DJ, Ashton RC, et
al. Bridge experience with long-term implantable left ventricular assist devices.
Are they an alternative to transplantation? Circulation. 1997 Apr 1;95(7):1844-52.
Dang NC, Naka Y. Perioperative pharmacotherapy in patients with left ventricular
assist devices. Drugs Aging. 2004;21(15):993-1012.
29.
30.
Goldstein DJ, Beauford RB. Left ventricular assist devices and bleeding: adding
insult to injury. Ann Thorac Surg. 2003 Jun;75(6 Suppl):S42-7.
31.
Himmelreich G, Ullmann H, Riess H, Rosch R, Loebe M, Schiessler A, et al.
Pathophysiologic role of contact activation in bleeding followed by
thromboembolic complications after implantation of a ventricular assist device.
ASAIO J. 1995 Jul-Sep;41(3):M790-4.
32.
Meyns B, Sergeant P, Herijgers P, Flameng W, Vanhaecke J, Van Cleemput J, et
al. Bridging procedures to heart transplantation. World J Surg. 2002
Feb;26(2):212-7.
96 33.
Minev PA, El-Banayosy A, Minami K, Kortke H, Kizner L, Korfer R. Differential
indication for mechanical circulatory support following heart transplantation.
Intensive Care Med. 2001 Aug;27(8):1321-7.
34.
Vitali E, Lanfranconi M, Bruschi G, Russo C, Colombo T, Ribera E. Left
ventricular assist devices as bridge to heart transplantation: The Niguarda
Experience. J Card Surg. 2003 Mar-Apr;18(2):107-13.
35.
Pereira NL, Chen D, Kushwaha SS, Park SJ. Discontinuation of antithrombotic
therapy for a year or more in patients with continuous-flow left ventricular assist
devices. Interact Cardiovasc Thorac Surg. 2010 Oct;11(4):503-5.
36.
John R, Kamdar F, Liao K, Colvin-Adams M, Miller L, Joyce L, et al. Low
thromboembolic risk for patients with the Heartmate II left ventricular assist
device. J Thorac Cardiovasc Surg. 2008 Nov;136(5):1318-23.
37.
Kirklin JK, Naftel DC, Kormos RL, Pagani FD, Myers SL, Stevenson LW, et al.
Interagency Registry for Mechanically Assisted Circulatory Support
(INTERMACS) analysis of pump thrombosis in the HeartMate II left ventricular
assist device. J Heart Lung Transplant. 2014 Jan;33(1):12-22.
38.
Boyle AJ, Russell SD, Teuteberg JJ, Slaughter MS, Moazami N, Pagani FD, et al.
Low thromboembolism and pump thrombosis with the HeartMate II left
ventricular assist device: analysis of outpatient anti-coagulation. J Heart Lung
Transplant. 2009 Sep;28(9):881-7.
39.
Stulak JM, Deo S, Schirger J, Aaronson KD, Park SJ, Joyce LD, et al.
Preoperative atrial fibrillation increases risk of thromboembolic events after left
ventricular assist device implantation. Ann Thorac Surg. 2013 Dec;96(6):2161-7.
40.
Cowger JA, Romano MA, Shah P, Shah N, Mehta V, Haft JW, et al. Hemolysis: a
harbinger of adverse outcome after left ventricular assist device implant. J Heart
Lung Transplant. 2014 Jan;33(1):35-43.
41.
Shah P, Mehta VM, Cowger JA, Aaronson KD, Pagani FD. Diagnosis of
hemolysis and device thrombosis with lactate dehydrogenase during left
ventricular assist device support. J Heart Lung Transplant. 2014 Jan;33(1):102-4.
42.
Rogers JG, Butler J, Lansman SL, Gass A, Portner PM, Pasque MK, et al.
Chronic mechanical circulatory support for inotrope-dependent heart failure
patients who are not transplant candidates: results of the INTrEPID Trial. J Am
Coll Cardiol. 2007 Aug 21;50(8):741-7.
43.
Lazar RM, Shapiro PA, Jaski BE, Parides MK, Bourge RC, Watson JT, et al.
Neurological events during long-term mechanical circulatory support for heart
failure: the Randomized Evaluation of Mechanical Assistance for the Treatment
97 of Congestive Heart Failure (REMATCH) experience. Circulation. 2004 May
25;109(20):2423-7.
44.
Pae WE, Connell JM, Boehmer JP, Korfer R, El-Banayosy A, Hetzer R, et al.
Neurologic events with a totally implantable left ventricular assist device:
European LionHeart Clinical Utility Baseline Study (CUBS). J Heart Lung
Transplant. 2007 Jan;26(1):1-8.
45.
Slaughter MS, Sobieski MA, Gallagher C, Dia M, Silver MA. Low incidence of
neurologic events during long-term support with the HeartMate XVE left
ventricular assist device. Tex Heart Inst J. 2008;35(3):245-9.
46.
Thomas CE, Jichici D, Petrucci R, Urrutia VC, Schwartzman RJ. Neurologic
complications of the Novacor left ventricular assist device. Ann Thorac Surg.
2001 Oct;72(4):1311-5.
47.
Tsukui H, Abla A, Teuteberg JJ, McNamara DM, Mathier MA, Cadaret LM, et al.
Cerebrovascular accidents in patients with a ventricular assist device. J Thorac
Cardiovasc Surg. 2007 Jul;134(1):114-23.
48.
Crow S, John R, Boyle A, Shumway S, Liao K, Colvin-Adams M, et al.
Gastrointestinal bleeding rates in recipients of nonpulsatile and pulsatile left
ventricular assist devices. J Thorac Cardiovasc Surg. 2009 Jan;137(1):208-15.
49.
Letsou GV, Shah N, Gregoric ID, Myers TJ, Delgado R, Frazier OH.
Gastrointestinal bleeding from arteriovenous malformations in patients supported
by the Jarvik 2000 axial-flow left ventricular assist device. J Heart Lung
Transplant. 2005 Jan;24(1):105-9.
50.
Geisen U, Heilmann C, Beyersdorf F, Benk C, Berchtold-Herz M, Schlensak C, et
al. Non-surgical bleeding in patients with ventricular assist devices could be
explained by acquired von Willebrand disease. Eur J Cardiothorac Surg. 2008
Apr;33(4):679-84.
51.
Bonaros N, Mueller MR, Salat A, Schima H, Roethy W, Kocher AA, et al.
Extensive coagulation monitoring in patients after implantation of the MicroMed
Debakey continuous flow axial pump. ASAIO J. 2004 Sep-Oct;50(5):424-31.
52.
Radovancevic R, Matijevic N, Bracey AW, Radovancevic B, Elayda M, Gregoric
ID, et al. Increased leukocyte-platelet interactions during circulatory support with
left ventricular assist devices. ASAIO J. 2009 Sep-Oct;55(5):459-64.
53.
Houel R, Mazoyer E, Boval B, Kirsch M, Vermes E, Drouet L, et al. Platelet
activation and aggregation profile in prolonged external ventricular support. J
Thorac Cardiovasc Surg. 2004 Aug;128(2):197-202.
98 54.
Wilhelm CR, Ristich J, Kormos RL, Wagner WR. Monocyte tissue factor
expression and ongoing complement generation in ventricular assist device
patients. Ann Thorac Surg. 1998 Apr;65(4):1071-6.
55.
Demirozu ZT, Etheridge WB, Radovancevic R, Frazier OH. Results of HeartMate
II left ventricular assist device implantation on renal function in patients requiring
post-implant renal replacement therapy. J Heart Lung Transplant. 2011
Feb;30(2):182-7.
56.
McCarthy PM, Savage RM, Fraser CD, Vargo R, James KB, Goormastic M, et al.
Hemodynamic and physiologic changes during support with an implantable left
ventricular assist device. J Thorac Cardiovasc Surg. 1995 Mar;109(3):409-17;
discussion 17-8.
57.
Sandner SE, Zimpfer D, Zrunek P, Rajek A, Schima H, Dunkler D, et al. Renal
function and outcome after continuous flow left ventricular assist device
implantation. Ann Thorac Surg. 2009 Apr;87(4):1072-8.
58.
Bourge RC, Naftel DC, Costanzo-Nordin MR, Kirklin JK, Young JB, Kubo SH,
et al. Pretransplantation risk factors for death after heart transplantation: a
multiinstitutional study. The Transplant Cardiologists Research Database Group. J
Heart Lung Transplant. 1993 Jul-Aug;12(4):549-62.
59.
Teuteberg JJ, Lewis EF, Nohria A, Tsang SW, Fang JC, Givertz MM, et al.
Characteristics of patients who die with heart failure and a low ejection fraction in
the new millennium. J Card Fail. 2006 Feb;12(1):47-53.
99 APPENDIX A
INSTITUTIONAL REVIEW BOARD APPROVAL
100 101