Nesiritide for Decompensated Heart Failure Is Associated with Increased Risk of Renal Failure: A Meta-Analysis of Randomized Controlled Trials

S94 Journal of Cardiac Failure Vol. 12 No. 6 Suppl. 2006 302 Preoperative Screening Scale Predicts Successful Bridge to Transplantation among Chronic Congestive Heart Failure Patients Mark J. Russo1, Deborah D. Ascheim2, Heather Hussey1, Nicholas Dang1, Jonathan M. Chen1, Allan S. Stewart1, Michael Argenziano1, Donna M. Mancini2, Mehmet C. Oz1, Yoshifumi Naka1; 1Division of Cardiothoracic Surgery, College of Physicians and Surgeons, Columbia Univ, NY, NY; 2Division of Cardiology, College of Physicians and Surgeons, Columbia Univ, NY, NY

including tests of interaction with gender, is shown in the table below. Conclusion: Women with HF were at less risk for adverse events than men, but the apparent protective effect of female gender was significantly reduced by the presence of DM. All Patients (n 5 7788) Men vs. Women DM

Objective: To devise a preoperative risk score to predict successful bridge to transplant (BTT) following left ventricular assist device (LVAD) implantation among patients with chronic congestive heart failure (CCHF). Methods: Analysis included 132 CCHF patients (diagnosis $6 months) who underwent LVAD implant as a BTT at a single institution. The primary outcome measure was survival to transplantation; other measures included 1-year survival. Univariate and multivariable analyses were performed to determine the predictors of survival to transplant after LVAD insertion. Threshold analysis using receiver operating (ROCs) curves and stratum-specific likelihood ratios were used to determine cut points on continuous variables. Using the relative risks for each identified variable, a risk factor summation score was devised. In order to assess the predictive power of the model, ROCs curve were constructed to determine the area under the curve (AUC). Results: Patients risk was determined by assigning points based on the following scoring scheme: albumin !2.9 (1), hematocrit !27 (1), central venous pressure O16 (1), age $56 years old (1), CrCl !55.2 (2), female sex (2), previous cardiac surgery (2), PTO16 (2), and BMI #20.4 (3) Survival to transplant by risk score is as follows: 0e4 96.2% (n 5 79); 5e8 59.1% (n 5 44); $ 9 11.1% (n 5 9); the AUC was 0.87 (0.81e0.94). Likewise, this risk score is highly predictive of longer term survival with 1-year survival following LVAD implant by risk score as follows: 0e4 86.0% (n 5 57); 5e8 46.0% (n 5 37); $9 11.1% (n 5 9); the AUC was 0.80 (0.71e0.88) CONCLUSIONS: Pre-VAD implant patient characteristics are highly associated with survival to transplant as well as longer term survival. Because CCHF patients may undergo non-emergent VAD implant, this risk score could assist in patient selection, timing of implant, and pre-implant optimization of patients.

Non DM Interaction P Value

All-Cause Hospitalization

Death HR (95% CI)

HR (95% CI)

P

1.23 0.008 (1.06e1.43) 1.64 !0.001 (1.44e1.86) 0.003

P

1.03 (0.92e1.15) 1.28 (1.17e1.39)

0.598 !0.001

HF Hospitalization HR (95% CI) 1.00 (0.86e1.16) 1.28 (1.13e1.46)

0.002

P 0.982 !0.001 0.011

304 B-Type Natriuretic Peptide Levels Measured 12 Hours after Cardiac Surgery Correlate Best with Mortality Timothy Fitzgibbons1, Vladimir Birjiniuk2, Andreia Biolo1, Prasad Maddukuri1, Michael Crittenden2, Miguel Haime2, Gifford Lum2, Thomas Rocco2, Shukri Khuri2, Jacob Joseph2; 1Cardiology Dept., Boston University School of Medicine, Boston, MA; 2V.A. Boston Healthcare System, West Roxbury, MA Introduction/Hypothesis: Prior small studies have demonstrated that preoperative B-type natriuretic peptide (BNP) levels are correlated with outcomes. Since the perioperative state after cardiac surgery is associated with marked changes in cardiocirculatory homeostasis, we postulated that postoperative BNP levels may be better correlated with outcomes compared to preoperative levels. Methods: We analyzed clinical data of 192 male patients who underwent cardiac surgery at VA Boston Healthcare during the time period 2004e2005. Log-transformed BNP levels (preoperative, 3, 6, 12 hours and days 1e5 postoperative) were correlated with length of stay (LOS), intensive care unit (ICU) LOS, 6-month mortality, and 6-month hospital admission rate using Pearson’s correlation. Multiple linear and logistic regression analyses were utilized to correct for the following confounding factors: age, body mass index, creatinine, and left ventricular ejection fraction (LVEF). Results: Mean age of the patient group was 66 6 9.1 years; mean preoperative LVEF was 47 6 13 %. Preoperative BNP levels were independent predictors of total LOS and ICU LOS, while 12 hour postoperative BNP levels were independent predictors of ICU LOS and 6-month mortality (Table 1). Neither measurement was correlated with the 6-month admission rate. Conclusions: BNP levels measured at the 12-hour postoperative time point had the greatest correlation with postoperative outcomes compared to preoperative BNP levels. This may reflect the fact that 12 hours gives adequate time for return to cardiocirculatory homeostasis and that elevation of BNP levels at that time may indicate significant cardiovascular perturbation. Our results indicate that a single measurement of BNP level at the 12-hour postoperative time point may allow appropriate risk stratification of the cardiac surgical patient, and guide resource allocation to improve outcomes. Table 1. Correlations of BNP Levels With Outcomes PreOp BNP

303 Diabetes Reduces the Likelihood of More Favorable Outcomes in Women Than Men with Heart Failure: Retrospective Analysis of the DIG Trial Francois M. Alla1, Ahmad Y. Al-Hindi1, Craig R. Lee1, Todd A. Schwartz1, J. Herbert Patterson1, Jalal K. Ghali1, Kirkwood F. Adams, Jr.1; 1Heart Failure Program, University of North Carolina-Chapel Hill, Chapel Hill, NC Background: Previous studies suggest women with heart failure (HF) may be at less risk of death or hospitalization than men with this syndrome. Whether this advantage is modified by diabetes mellitus (DM) has not been well investigated. Methods: We studied the influence of a history of DM on the relative risk for adverse outcomes (death, the risk of hospitalization for worsening HF, and all-cause hospitalization) in men versus women with HF using data from the Digitalis Investigation Group trial (n 5 7788). The analysis included patients with reduced left ventricular systolic function (LVD) (defined as LVEF !5 45%) and preserved left ventricular systolic function (PEF) (defined as LVEF O 45%). Multivariable proportional hazards analysis determined if DM modified the relative risk of men versus women (overall and in LVD or PEF groups) for adverse outcomes. Results: Among patients with LVD (n 5 6800), women had better survival than men in DM (1.22, 95% confidence interval (CI) 1.04e1.44, p 5 0.015) and non DM groups (1.66, 95% CI 1.45e1.91, p ! 0.001), but the advantage was significantly less in DM (p 5 0.004 for interaction). In the subgroup with PEF (n 5 988), women did not have better survival than men in the DM group (1.42, 95% CI 0.90e2.25, p 5 0.136) but did in the non DM group (1.72, 95% CI 1.20e2.46, p 5 0.003, p 5 0.505 for interaction). The influence of DM on the relative risk of men versus women for study end points in all patients,

Total LOS ICU LOS 6 month mortality 6 month admissions

12 Hour PostOp BNP

Pearsons R

P value (multivariate analysis)

Pearsons R

P value (multivariate analysis)

0.34 0.36 0.09 0.04

0.024* 0.003* NS** NS**

0.19 0.33 0.27 0.15

NS* !.001* 0.007** NS*

*multiple linear regression, **multiple logistic regression.

305 Nesiritide for Decompensated Heart Failure Is Associated with Increased Risk of Renal Failure: A Meta-Analysis of Randomized Controlled Trials Prasanna Kumar Venkatesh1, Saurabh Kandpal1, Janos Molnar1, Rohit R. Arora1; 1 Cardiology, Chicago Medical School, Chicago, IL Background: Nesiritide (recombinant B-type natriuretic peptide, BNP) is approved for the treatment of acute decompensated heart failure (ADHF). There has been a concern of increase in serum creatinine with nesiritide use. We performed a meta-analysis to evaluate the risk of renal failure (RF) in patients treated with nesiritide for ADHF. Methods: Five large randomized controlled trials on nesiritide with available data on serum creatinine were included for analysis. The nesiritide dose used in the selected trials was 0.01e0.03 mcg/kg/min. Data on incidence of RF (defined as increase in serum creatinine O0.5 mg/dl at any time through 30 days) in the nesiritide and control arms were extracted from the selected trials and from the nesiritide

The 10th Annual Scientific Meeting database (Scios Inc.). Results: The pooled estimate for the relative risk (RR) of RF revealed a significantly higher risk of developing RF with nesiritide as compared to control [RR with 95% confidence interval (CI): 1.449 (1.122e1.871), p ! 0.005]. Conclusions: Supporting a previous analysis, our results indicate that nesiritide use in ADHF is associated with a significant risk of developing RF. Further analysis adjusted for baseline creatinine, rapidity of RF onset, urine output, nesiritide dose, duration of infusion, hypotension, baseline BNP levels and concurrent use of drugs affecting renin angiotensin aldosterone system may reveal subgroups in jeopardy. Large scale randomized controlled trials are required to conclusively validate these findings. Meta-analysis Showing Risk of Renal Failure With Nesitiride Use Trial

Relative Risk of Renal Failure (95% CI)

Mills et al Efficacy Comparative PRECEDENT VMAC Pooled Relative Risk

1.470 3.706 2.010 1.641 1.301 1.449

(0.532e4.059) (0.888e15.46)* (1.007e4.010)* (0.815e3.303) (0.940e1.800) (1.122e1.871)**




HFSA

S95

4 Cardiology, Cleveland Clinic, Cleveland, OH; 5Cardiology, Berkshire Medical Center, Berkshire, MA; 6Cardiology, Duke Medical Center, Durham, NC

Background: Get With The Guidelines (GWTG) is a national initiative of the AHA to improve the quality of heart failure (HF) care. One metric of quality is the percentage of patients who receive all guideline indicated care for which they were eligible. We studied trends in this complete care metric over the initial year of GWTG-HF. Methods: GWTG employs a collaborative model of care and a web-based patient management tool. We evaluated 18,516 HF patients treated at 97 GWTG-HF hospitals from 1/05e3/06. Five pre-specified performance measures were included: discharge instructions, measurement of LV function, ACEI/ARB use, beta blocker use, and smoking cessation counseling. The complete care measure was defined as the percentage of patients who received all the process measures for which they were eligible. Results: Patients were mean age 72.9 6 14.2 years, 50% female, 47.2% ischemic etiology, and mean LVEF 38.7 6 17.1. Overall in 64.6% of patients, there were no missed opportunities for care. This quality metric improved significantly over the 4 quarters following baseline from 60.3% to 68.6%, P ! 0.0001. Conclusions: Hospitals participating in GWTG-HF significantly improved evidence-based care of HF patients over time as reflected by this composite performance measure. However, with at least one missed opportunity for complete care in up to 35% of patients, further efforts are needed to enhance the reliability of HF care delivery.

*p! 0.05, **p! 0.005. Measure

306 Predictors of Mortality in a CRT-D Patient Population Teresa De Marco1, John Boehmer2, Mark Carlson3, Ramona Ruble4, Sudha Iyer4; 1 Med., UCSF, San Francisco, CA; 2Med. & Surg., Penn State Coll. Med., Hershey, PA; 3Medicine, Case Western Reserve Univ., Cleveland, OH; 4Clinicals, Guidant, St. Paul, MN Background: Discovering the baseline features that indicate which patients (pts) are at higher risk for death may help physicians to determine therapy options. Various baseline features have been identified in general heart failure (HF) population, but there are few analyses that identify variables predictive of death in CRT-D pts. Methods: Pts in the Device Evaluation of CONTAKÒ RENEWALÒ 2/4/4HE and EASYTRAKÒ 2: Assessment of Safety and Effectiveness in Heart Failure (DECREASE-HF) study were included in this retrospective analysis. Univariate and Multivariate Cox Proportional Hazards Regression (PHR) Analyses at baseline were done. Results: 342 of 360 pts enrolled were successfully implanted; 54 (16%) have died as of March 30, 2006. The baseline demographics & cardiac features were typical of CRT-D-indicated pts. Of the 41 variables analyzed by a univariate cox PHR model, 14 were found to be significant predictors, including 4 that were significant predictors of mortality by multivariate cox PHR. Conclusion: HF pts getting CRT-D, with LVEF ! 23%, GFR ! 60, on non-loop diuretic, are NYHA Class IV, or have afib may be at higher risk for death than similar HF pts without this profile, and should undergo early and aggressive treatment. Cox PHR for Time-to-Mortality Hazard Ratio (95% C.I.)

Variable Age O5 68 years LVEF ! 23% PR Interval O5 190 ms BUN O5 22 mg/dL Creatinine O5 1.2 mg/dL Peak VO2 ! 12.5 ml/kg/min VE/VCO2 Slope O5 36 Glomerular Filtration Rate GFR ! 60 Diuretic Non-loop Diuretic NYHA IV at Implant Atrial Tach History of AF Renal Disease

2.12 (1.21, 3.70) 2.01 (1.15, 3.51) 1.91 (1.09, 3.34)

P-value

Multivariate Hazard (95% C.I.) P-value

0.009 0.01 1.96 (1.10, 3.50) 0.02 -

0.022 -

7.13 (2.11, 24.12) 0.002 Not included* 5.79 (2.61, 12.82) !0.001 -

-

5.14 (2.39, 11.05) !0.001 Not included*

-

2.63 (1.41, 4.94)

-

0.003 Not included*

5.20 (2.61, 10.35) !0.001 4.12 (2.04, 8.33) 4.26 4.75 4.26 2.11 2.38 2.66

(1.04, (2.58, (1.04, (1.20, (1.31, (1.50,

17.48) 0.04 8.74) !0.001 3.08 (1.63, 5.84) 17.48) 0.04 3.08 (1.63, 5.84) 3.72) 0.01 4.32) 0.004 2.33 (1.26, 4.33) 4.73) 0.001 -

!0.001 0.001 0.001 0.007 -

*Due to missing data.

307 Reliability of Heart Failure Patient Care: Initial Results from the American Heart Association’s Get With The Guidelines (GWTG) Heart Failure Program Gregg C. Fonarow1, Kenneth A. LaBresh2, Clyde Yancy3, Nancy M. Albert4, Gray Elrodt5, Adrian F. Hernandez6, Li Lang6, Eric D. Peterson6, GWTG Steering Committee and Hospitals; 1Cardiology, UCLA Medical Center, Los Angeles, CA; 2 Cardiology, Mass Pro, Waltham, MA; 3Cardiology, UT Southwestern, Dallas, TX;

N Composite Measure of Complete Care

Baseline

Q1

Q2

Q3

Q4

P value, time trend

2408 60.3%

5898 64.0%

4447 66.1%

2907 65.3%

1234 68.6%

!0.0001

308 Sustained Versus Non-Sustained Improvement in Left Ventricular Systolic Function in Heart Failure Patients on Beta Blocker Therapy Marie Galvao1, Catherine A. Galvin1, Yunling Du2, Heather Trivedi1, Robert J. Ostfeld1; 1Medicine/Cardiology, Montefiore Medical Center, Bronx, NY; 2 Epidemiology and Population Health, Montefiore Medical Center, Bronx, NY Introduction: Beta blockers (BB) have been shown to improve left ventricular (LV) ejection fraction (EF) in patients with heart failure (HF) due to LV systolic dysfunction. Some do not sustain this improvement. Predictors of re-decline have not been well described. We compared characteristics of those who sustained this improvement to those who did not. Methods: All medical records from the Montefiore Medical Center HF clinic were retrospectively reviewed. Subjects were identified who had an initial LVEF #40% and subsequently improved their EF by $10% on BB: carvedilol (C) or metoprolol succinate (M). Subjects were excluded if an event 3 months (mo) prior to or anytime after BB initiation could have influenced EF, if pts were not on maximally tolerated HF therapy or if routine follow-up echocardiography was not available. Subjects were divided into 2 groups. Group R (re-decline) included patients who had a subsequent decrease of #10% in their EF. Group S (sustained) included all other patients. Baseline means and proportions were compared. Logistic regression was used. Results: Sixty-three pts met criteria. For groups R (n 5 12) and S (n 5 51), mean age (yr) was 53 (612), 54 (616); male (%) 50, 59; ischemic (%) 25, 14; C use (%) 33, 65 (p 5 ns for all groups), respectively. Average mo to improvement in EF for all pts was 21.7 (62.1). Average mo from improvement to re-decline in EF was 23.2 (64.8). On comparison of multiple clinical characteristics between groups R and S, only DM (83% vs. 42%, p 5 0.02) and higher NYHA class at time of improvement in EF (N-I) (p 5 0.008) differed. C use was associated with a trend toward a decreased rate of re-decline (p50.058). On multivariate analysis that included DM, BB type and N-I, only DM was associated with increased odds of re-decline (p 5 0.04). Multivariate analysis of the sub-group of pts with DM (n 5 31) that included BB type and N-I, C use was associated with decreased odds of re-decline (p 5 0.05). Conclusion: Re-decline in EF is not uncommon. In this small retrospective analysis, 19% of subjects experienced a re-decline in EF. DM and increased NYHA class at time of EF improvement are associated with a re-decline in EF. Carvedilol compared to metoprolol succinate use in pts with DM appears to be associated with decreased odds of re-decline in EF. Further study is warranted.

309 Feasibility Study for Spousal Caregivers Receiving a Telehealth Social Support Intervention Louise M. LaFramboise1, Bernice C. Yates1, Bunny Pozehl1, Susan A. Barnason1; 1 College of Nursing, University of Nebraska Medical Center, Omaha, NE Objective: Spousal caregivers with major caregiving responsibilities for seriously chronically ill persons may neglect healthy lifestyle behaviors and be vulnerable for poor health outcomes. Caregivers are more stressed and have mortality risks 63% higher than for non-caregivers. Research with heart failure (HF) caregivers has assessed burden but hasn’t evaluated interventions to diminish stress outcomes. The objective of this feasibility study was to examine the outcomes of an 8-week social support (SS) intervention on the health behaviors of coping and sleep, and the