JACC

March 10, 2009

ABSTRACTS - Cardiac Function and Heart Failure
3:30 p.m.

1024-206

Effects of the Selective PPAR Gamma Modulating
Angiotensin Receptor Blocker Irbesartan on Impaired
Insulin Sensitivity in Patients With Chronic Heart
Failure IRIS-HF: A Placebo-Controlled, Double Blinded,
Randomized Study

Wolfram Doehner, Cornelia Kenneke, Johanna Todorovic, Mathias Rauchhaus, Stephan
von Haehling, Stefan D. Anker, Charite University Medical School, Berlin, Germany

3:30 p.m.
1024-207

Relaxin, A Novel, Multiple Mechanism Vasodilator, for
the Treatment of Acute Heart Failure - The PreRELAXAHF Study

John R. Teerlink, Marco Metra, G. Michael Felker, Adriaan A. Voors, Beth Weatherley,
Elaine Unemori, Sam L. Teichman, Gad Cotter, Section of Cardiology, Veterans Affairs
Medical Center, University of California, San Francisco, CA, Section of Cardiovascular
Diseases, Department of Experimental And Applied Medicine, Univ of Brescia, Brescia,
Italy
Relaxin is a naturally-occurring peptide that is known to affect multiple vascular control
pathways. In preliminary studies, it reduced wedge pressure and increased cardiac
output.
Objective: PreRELAX-AHF is the pilot for a phase 3 trial evaluating safety and doseresponse to 48h of IV relaxin in patients with acute heart failure (AHF) and normal to high
systolic blood pressure (SBP).
Methods: Eligible patients were hospitalized for AHF with dyspnea, congestion on
chest X-ray, NTproBNP >1,400 pg/ml, SBP >125 mmHg and eGFR 30-75 mL/min.
Randomization (double-blind) was <16 h from presentation to placebo (PBO) or relaxin
10, 30, 100 or 250 mcg/kg/d. Endpoints of interest were dyspnea (by 7-pt Likert and
Visual Analog Scales [VAS]), clinical outcomes and renal function.
Results: In this interim analysis, 209 patients in 8 countries were followed for >14 days.
Baseline characteristics were balanced across groups. Improvement in dyspnea was
greater in the relaxin arms vs PBO, with greater efﬁcacy seen at 10, 30 and 100 mcg/kg/d
at all time points from 6h to Day 14. The largest effects were seen at 30 mcg/kg/d. Dyspnea
improvement by VAS was 96% greater in the 30 mcg group than placebo at 6h. The VAS
change from baseline over time was signiﬁcant (P=0.03 for 30 mcg vs PBO, for 6h to Day
14). Time to ﬁrst moderate or marked improvement by Likert was 2.7, 2.4, 1.8, 2.4 and
2.2 days in PBO, 10, 30, 100 and 250 dose groups (P<0.05, relaxin 30 vs PBO). Similar
trends were seen in measures of HF including greater weight loss (up to 1 kg difference),
lower IV loop diuretic use and lower incidence of in hospital worsening HF. Mean length of
stay was shorter in the active arms by 1-2.5 days, most pronounced at 30 mcg/kg/d. There
were no safety or tolerability issues, including the absence of symptomatic hypotension or
worsening of renal function. Follow-up to Day 180 will be presented.
Conclusion: In patients with AHF and normal to high SBP (>75% of AHF patients), early
IV relaxin administration for 48 hours may produce rapid and sustained improvement in
dyspnea and other heart failure measures, without evidence of symptomatic hypotension
or adverse effects on renal function. Larger studies are planned with relaxin to conﬁrm
these effects.

3:30 p.m.
1024-208

Does Endothelial Dysfunction Contribute to Exercise
Limitation in Heart Failure With Preserved Ejection
Fraction?

Barry A. Borlaug, Thomas P. Olson, Carolyn S.P. Lam, Kelly S. Flood, Bruce D. Johnson,
Margaret M. Redﬁeld, Amir Lerman, Mayo Clinic, Rochester, MN
Background Arterial stiffening and impaired vasodilation with exercise are frequently
observed in Heart Failure (HF) with preserved Ejection Fraction (HFpEF). While
endothelial dysfunction is common in systolic HF, its role in HFpEF remains uncertain.
Objectives To compare endothelial function and vascular stiffening in patients with
HFpEF and hypertension (HTN) and determine relationships with exercise performance.
Methods Age/gender matched subjects with HFpEF (n=21) and HTN (n=19) underwent
baseline assessment of peripheral reactive hyperemia index (RHI), a measure of
endothelium-dependent vasodilation, and central aortic pulse pressure (PP), an index
of vascular stiffness, followed by metabolic exercise testing (Ex). Systemic vascular
resistance (SVR) was determined from echo-Doppler at rest and Ex.
Results RHI and PP were similar in HFpEF and HTN, while Ex capacity was reduced
in HFpEF (*p<0.0001; Table). In combined analysis, $SVR with Ex was associated with
peak VO2 (r= -0.55, p<0.05). $SVR was lower in HFpEF than HTN (†p=0.003). In HFpEF,
peak VO2 was directly related to resting RHI (r= 0.50, p<0.05) and inversely to aortic PP
(r= -0.51, p<0.05), but neither of the latter was associated with $SVR.
Conclusions While patients with HFpEF display abnormal vasodilation with exercise,
resting indices of endothelial function are similar to HTN. Endothelial dysfunction and
vascular stiffening contribute to exercise limitation in HFpEF and may serve as novel
therapeutic targets.
HTN (n=19)

HFpEF (n=21)

Aortic Systolic Blood pressure (mmHg)

130±12

125±19

Aortic Pulse Pressure (mmHg)

54±12

59±22

Reactive Hyperemia Index

2.31±0.66

2.18±0.69

Percent with Endothelial Dysfunction (RHI<2)

28

42

Peak VO2 (ml/min*kg)

18.6±3.7

12.7±3.0*

Baseline SVR(DSC)

1780±400

1600±500

$SVR with Exercise(DSC)

-980±300

-620±350†

ACC.ORAL CONTRIBUTIONS
901

Ventricular Assist Devices: The Good, the
Bad and the Ugly
Monday, March 30, 2009, 8:00 a.m.-9:30 a.m.
Orange County Convention Center, Room W308A
8:00 a.m.
0901-3

Predictors of Stroke on Heartmate II Left Ventricular
Assist Device

Biswajit Kar, Sukhdeep S. Basra, Pranav Loyalka, Igor D. Gregoric, Reynolds Delgado,
Andrew Civitello, Roberta Bogaev, William Cohn, Antonius Attallah, Rohan Wagle, O. H.
Frazier, Texas Heart Institute, Houston, TX, University of Texas at Houston, Houston, TX
Background: The Heartmate II is an effective treatment alternative for patients with end
stage heart failure refractory to medical therapy. However, a relatively large proportion of
patients on Heartmate II develop stroke. We aimed to determine the predictors of stroke
in this patient population.
Methods and Materials: All patients (n=71) implanted with Heartmate II at our institute
were included in the study. Patients developing a stroke were classiﬁed as cases (n
= 18) with the remaining as controls (n=53). We speciﬁcally assessed the following
predictors of stroke: Non-Pulsatile ﬂow, Anticoagulation, Mean Arterial Pressure (MAP)
and Infections.
Results: The mean duration of support for cases and controls was 429.06 and 304.92
days respectively, with a median implant-stroke interval of 174 days. On statistical
analysis, Non Pulsatile Flow (O.R: 4.68, 72.2% vs. 37.7% in cases and controls; p =
0.039) and Bacteremia (O.R: 9.59, 88.9% vs. 51% in cases and controls ; p = 0.045)
were found to be signiﬁcantly associated with an increased risk of developing stroke. This
remained signiﬁcant even after adjusting for potential confounders like duration of support,
MAP, anticoagulation status, gender, and atrial ﬁbrillation using logistic regression in a
multivariate analysis. However, Anticoagulation (Median INR 1.3 vs. 1.55 in cases and
controls, p = 0.181) and MAP (Median 85.5 vs. 78.6 mm Hg in cases and controls; p =
0.123) were not signiﬁcantly associated with stroke. 53.8% of the cases suffered from
acute peristroke infections (positive cultures 15 days before upto 5 days after stroke).
Conclusions: Traditional risk factors of stroke like Diabetes, Hypertension, Hyperlipidemia,
Age, Smoking and Atrial Fibrillation are not associated with stroke in patients on Heartmate
II LVAD. Instead, Non-pulsatile ﬂow and Bacteremia are the important predictors of stroke
in patients on Heartmate II.

Cardiac Function and Heart Failure

Background: Impaired insulin sensitivity (Si) is common in chronic heart failure (CHF),
contributes to symptomatic status and independently predicts prognosis. It has been
suggested that the angiotensin II-receptor antagonist irbesartan can improve insulin
sensitivity via selective activation of the peroxisome proliferator-activated receptor
gamma (PPAR gamma). We aimed to assess the effect of irbesartan on impaired insulin
sensitivity in patients with CHF.
Methods: In a prospective, placebo-controlled, double-blinded, randomized single center
study we included 36 non-diabetic patients with stable ischemic CHF (age 63±9y, BMI
28.2±3.9kg/m2, peakVO2 16.6±4.8mL/kg/min all mean ±SD). Irbesartan (target dose
300mg/d) or placebo was given on top of standard optimum CHF therapy including ACE
inhibitor and beta-blockers for 16 weeks. Change of insulin sensitivity from baseline to
week 16 (primary endpoint) was assessed using the minimal modelling technique from
glucose and insulin proﬁles of a frequently sampled intravenous glucose tolerance test.
Results: At baseline both groups were similar for age, NYHA class, peak VO2, BMI, body
composition (DEXA scan), Si (p=0.2), and main clinical characteristics. Si was 2.51±1.58
min-1.μU.mL-1.104 in the study population, which is 30% lower than in healthy controls
of similar age (P<0.05). In the irbesartan treated group SI increased by 26% (p<0.001
within group), but it decreased by 15% in the placebo group (p=0.17 within group).
Change in SI from baseline was signiﬁcantly different between groups (mean difference
1.044 min-1.μU.mL-1.104; 95%CI 0.45 to 1.64, p=0.001). Treatment with irbesartan was
well tolerated. NYHA class, peak VO2 and body composition did not change but blood
pressure signiﬁcantly decreased on irbesartan (-5±2 mmHg, p<0.002).
Conclusion: Our study shows that 16 weeks of added therapy with irbesartan compared
to placebo signiﬁcantly improves impaired insulin sensitivity in non-diabetic patients with
chronic heart failure. Whether this metabolic effect of irbesartan translates into additional
clinical beneﬁts for heart failure patients should be tested in larger studies.

A163

 A164

ABSTRACTS - Cardiac Function and Heart Failure

JACC

8:15 a.m.
0901-4

Validation of the Model for End-Stage Liver Disease in
Predicting Left Ventricular Assist Device Mortality

Cardiac Function and Heart Failure

Jennifer Cowger Matthews, Todd F. Dardas, Jonathan W. Haft, Francis D. Pagani, Keith
D. Aaronson, University of Michigan Health System, Ann Arbor, MI
Background: Preoperative Model for End-Stage Liver Disease (MELD) scores have
previously been shown to predict mortality in LVAD recipients at our center. Our aim was
to validate the prior ﬁndings using a national LVAD database.
Methods: Preoperative MELD scores were calculated for subjects undergoing LVAD
support enrolled into the INTERMACS registry between 2006-2008. MELD score=
9.57(logeCreatinine) + 3.78(logeBilirubin) + 11.2(logeINR) + 6.43, with minimums for all
variables set at 1.0. Complete data for MELD score calculations were available on 324
(87%) of 372 subjects.
The primary outcome of interest was the relationship between preoperative MELD score
and operative death (death within 30 days of LVAD operation). Student’s t tests were
used to compare mean MELD scores amongst operative deaths and survivors with odds
ratios (OR, 95% CI) generated from logistic regression. MELD scores were dichotomized
at the previously published threshold of 17. Secondary outcomes included survival and
the hazard ratio for death at 6 months in high (MELD q17) versus low (MELD <17) strata,
calculated with Kaplan-Meier and Cox regression analyses, respectively.
Results: The mean±standard deviation preoperative MELD score for the entire cohort
was 15.2±5.8, with mean creatinine, bilirubin, and INRs of 1.6 ±0.9 mg/dL, 1.7±0.1 mg/
dL, and 1.4±0.02 sec., respectively. Operative deaths (n=19, 6%) had higher mean (95%
CI) MELD scores [17.9(14.2,21.6)] than survivors [15.0(14.4,15.7); p=0.036)], such that
the odds of operative death increased 50% [OR 1.5(1.1,2.1)] for each 5-unit increase in
MELD. At 6 months, survival for subjects with MELD scores q17 (n=120) was 67±5%
compared with 82±3% in subjects with lower scores (p=0.032). At 6 months, the hazard
ratio for death (n=56) in subjects with MELD scores q17 was 1.76 (1.04,2.98) times that
of those with lower scores (p=0.035).
Conclusion: Preoperative MELD scores, a marker of multisystem dysfunction, are
predictive of operative and 6-month LVAD mortality in INTERMACS and may serve as an
important preoperative tool in assessing LVAD risk.

March 10, 2009
8:45 a.m.

0901-6

Predictors of Improvement in Right Ventricular Function
After Placement of a Left Ventricular Assist Device

Barry A. Boilson, John A. Schirger, Basar Sareyyupoglu, Sudhir S. Kushwaha, Irina
Penev, Christopher GA McGregor, Richard C. Daly, Lucian A. Durham, III, Brooks S.
Edwards, Soon J. Park, Mayo Clinic, Rochester, MN
Background:End stage heart failure is increasing in prevalence and left ventricular assist
device (LVAD) technology being used increasingly as a bridge to transplantation and also
as destination therapy. Although LVAD placement is followed by left ventricular ofﬂoading
and increase in systemic cardiac output, persistent right ventricular (RV) dysfunction
is common.The aim of this study was to investigate the baseline characteristics which
contribute most to improvement in right ventricular dysfunction after LVAD placement.
Methods:All patients who underwent LVAD implantation at our institution between May
2003 and July 2008 were evaluated. The primary endpoint was improvement in RV
function measured by echo. Secondary endpoints were improvement in measures of
pulmonary hypertension by right heart catheterization and echo. Both bridge to transplant
and destination therapy patients were included. Both continuous ﬂow and pulsatile
devices were included.
Results:A total of 57 patients were studied. Mean duration of followup was 230.5 days.
Mean age was 58.4±12.3 years. Baseline mean peak VO2 was 10.3±2.1 ml/kg/min,
mean serum creatinine 1.5±0.5 mg/dl, and mean BNP 1688±1195 pg/ml. The most
powerful predictors of improvement in right ventricular function at last followup were
baseline severity of mitral valve regurgitation (MR) (p=0.0006, R=-0.51), tricuspid valve
regurgitation (TR) (p=0.01, R=0.39) and pulmonary vascular resistance (PVR) (p=0.006,
R=0.42). Similarly, improvement in pulmonary hypertension was predicted by baseline
MR (p=0.005, R=0.4), PCWP (p=0.002, R=0.44) and PVR (p<0.0001, R=0.53). Predictors
of change in PVR were baseline PA pressure (p=0.0002, R=0.5), MR severity (p=0.05,
R=0.28 )and PVR at baseline (p<0.0001, R=0.9)
Conclusions:Improvement in RV function improves most after LVAD implantation when it
occurs in the setting of severe MR. Dramatic ofﬂoading of the left ventricle and reduction in
MR severity in these patients post LVAD placement results in a signiﬁcant drop in RV afterload
and PVR which facilitates recovery of RV function. RV dysfunction present at baseline in the
absence of signiﬁcant MR may not improve to the same degree post LVAD placement.

8:30 a.m.
9:00 a.m.
0901-5

Early Decrease in Platelet Counts Following Left
Ventricular Assist Devices Implantation Is Part of the
Systemic Inﬂammatory Response Syndrome and a
Marker of Poor Long-Term Outcome.

Andrea Mignatti, Daniel B. Sims, Nir Uriel, Basil Ramlawi, Steve Holleran, Yoshifumi
Naka, Ulrich P. Jorde, Columbia University, New York, NY
Background: Alterations of platelet (plt) count have been described as part of the
systemic inﬂammatory response syndrome (SIRS) following open heart surgery. Several
studies suggest that SIRS can be modulated using perioperative steroid and/or statin
therapy. Plt behavior has not been studied after implantation of left ventricular assist
devices (LVADS).
Methods and results: Retrospective single center study of 114 consecutive HM I & II
implants performed at our institution for BTT or DT since 04/2004. For the purposes of this
study, success was deﬁned as being alive and/or transplanted 180 days post-op. Mean
plt count on day 3 (123 vs 91, p 0.009) and day 7 (196 vs 130 p 0.001) were signiﬁcantly
higher in success (n=87) compared to failure (n=27) pts. Overall success was 77 %.
However, success rate was 59% in pts whose plt count averaged < 100 between day 3
and 7 and 90% in those who averaged > 100 (Fig 1).Decrease in plt was weakly correlated
with pump time (R= 0.3 p 0.01) and pump time was not different between success and
failure pts (105 vs 111 min P=NS). Decrease in platelet count was signiﬁcantly correlated
with increase in WBC (median correlation coeff = -0.45 p<0.01) (Fig. 2).
Conclusion: Early plt count decrease following LVAD implant at least in part reﬂects
SIRS and the degree of plt decrease is a marker of long term outcome. Pharmacological
modulation of SIRS occurring in the early postoperative period should be investigated and
may improve long term outcomes of pts undergoing LVAD implantation.

0901-7

Ventricular Assist Device Removal After Cardiac
Recovery: Main Predictors for Long-Term Outcome
Without Transplantation or Assist Device ReImplantation

Michael Dandel, Yuguo Weng, Henryk Siniawski, Hans B. Lehmkuhl, Thorsten Drews,
Evgenij Potapov, Thomas Krabatsch, Christoph Knosalla, Roland Hetzer, Deutsches
Herzzentrum Berlin, Berlin, Germany
Background: There is increasing evidence for cardiac recovery which allows ventricular
assist device (VAD) removal. However, few chronic heart failure (HF) patients have been
weaned from VADs, most of them only recently. Thus, long-term (> 5 years) outcome data
after VAD removal are few. Now that our patient cohort with post-weaning stability for > 5
years has become larger, we focused on these patients, in order to obtain new information
for future weaning decisions.
Methods: Among 84 patients weaned from VADs since 3/1995 we selected only patients
(n = 36) who were weaned q 5 years ago, before 9/2003. We evaluated echocardiogtaphic
(ECHO) data recorded before VAD implantation and during pre-explantation “off-pump”
trials, HF duration before VAD implantation, duration of mechanical support, and stability
of unloading-induced recovery before and early after VAD removal.
Results: Of 36 evaluated patients, 33 (91.7%) had non-ischemic, and the other 3 ischemic
cardiomyopathy. Only 2 had BVADs; the other 34 had LVADs. Post-weaning 5- and 10-year
survival with native hearts was 76.5% and 70.6.%, respectively. Post-weaning survival for q
13 years was reached by 3 patients. During the ﬁrst 5 post-weaning years, HF recurred in 13
(36.1%) patients (9 underwent HTx, one received another LVAD and 3 died). Patients with q
5 years stability were younger, their history of HF and recovery time during unloading were
shorter and pre-weaning LV assessment revealed lower diameters, less altered geometry
and higher LVEF. For LVEF q 45% at end-diastolic diameter a 55mm the predictive value for
q 5 year was 88.9%. Early post-wening time course of LVEF also appeared predictive for
long-term stability. History of HF > 5 years and pre-weaning instability of unloading-induced
recovery appeared predictive for HF recurrence.
Conclusions: Long-term successful weaning from VADs is possible even in patients
with chronic HF and incomplete cardiac recovery. Pre-explantation off pump ECHO-data,
stability of recovery, duration of HF before VAD insertion and duration of VAD support
allow identiﬁcation of patients with the potential to remain stable for > 5 years. LVEF time
course early after-weaning facilitates prognostic assessment.