The 15th Annual Scientific Meeting
HFSA
S83
congestive heart failure (CHF). Additionally, many small studies have postulated that low serum cholesterol levels may be harmful in patients with CHF. We attempt to correlate outcomes in patients admitted with acutely decompensated heart failure (ADHF) according to admission low-density lipoprotein (LDL) levels. Hypothesis: We hypothesize that patients with very low LDL levels will have worse outcomes than those with higher LDL levels in patients with any type and degree of ADHF. Methods: In this retrospective analysis, patients were included if they met the following criteria: first known heart failure-related hospitalization at our institution and a lipid panel within three months of admission. The main outcomes studied included time to death, time to left ventricular assist device (LVAD) implantation, time to heart transplant, and number of hospitalizations. Results: A total of 800 charts were analyzed and followed for a mean of 2.6 years. After adjusting for baseline characteristics, with admission LDL levels of greater than 100 had better survival rates when compared to patients with LDL levels of 70 and below (hazard ratio, 0.752; 95% CI, 0.574 to 0.985; P50.038). There was a trend which did not reach statistical significance when comparing those with LDL levels between 71 and 100 and those with LDL levels of 70 and below (hazard ratio, 0.826; 95% CI, 0.655 to 1.040; P50.104). There was no statistically significant difference between groups in time to heart transplant (n518, P50.380) or time to LVAD implantation (n514, P50.379). There was no difference in number of hospitalizations between groups (P50.574).
Conclusion: Hypoalbuminemia is common in HFpEF and is associated with significant mortality and morbidity especially within 1 year of onset of low albumin. Increased attention should be paid to low albumin in treating CHF especially HFpEF. Further investigation of pathophysiologic mechanisms underlying hypoalbuminemia in HFpEF is warranted.
267 Prognostic Value of Serial Clinical, Laboratory and Ecocardiographic Tests in Ambulatory Patients with Chronic Heart Failure Carlo Lombardi, Alessandra Manerba, Enrico Vizzardi, Giovanni Cuminetti, Silvia Bugatti, Valentina Carubelli, Valentina Lazzarini, Savina Nodari, Marco Metra, Livio Dei Cas; Cardiology, University of Brescia, Brescia, Italy Background: The clinical history of pts affected by chronic heart failure (CHF) remains burdened by a high incidence of hospitalizations. Serial controls of clinical conditions, laboratory exams and ecocardiographic parameters may prove useful to predict the worsening of the pts conditions and their forthcoming hospitalisation. However, the value of such clinical-instrumental tests remains controversial. Aim of the study was to analyzed the prognostic value of demographic variables and clinical, laboratory and ecocardiographic paramenters assessed during two periodical ambulatory visits. Methods: we included 232 pts with CHF (age 67 +/- 12 years, left ventricular ejection fraction [LVEF], 36 +/- 7%) evaluated during two periodical ambulatory visits (average interval between both assessments, (203 +/- 71 days, median, 201 days). Pts had optimal medical therapy. Results: During a 236 +/- 176 day follow-up, 49 (21%) pts died because of cardiac causes or were hospitalised to WHF. As compared to the others, those pts were older (71 +/- 8 vs 66 +/- 13 years; p 5 0.005), presented more severe clinical conditions (NYHA class, p ! 0.0001), higher blood urea nitrogen (BUN) values during the last visit (86 +/- 50 vs 64 +/- 39 mg/dl; p 5 0.001), greater BUN increase between the first and second ambulatory control(13 +/- 21 vs 2 +/- 17 mg/dl; p 5 0.002), lower LVEF during the second visit(p ! 0.0001), a LVEF decrease between both ambulatory visits (p 5 0.001), greater prevalence of severe mitral insufficiency (37 vs 13%, p 5 0.0001). All such variables had a lower predictive value when assessed during the first visit. At multivariate analysis, the NYHA class(p ! 0.0001) and LVEF during the last visit (P 5 0.018), but not their variations between two subsequent assessments, and the BUN variations between the first and second visit (p 5 0.068) had an independent predictive value for death or new hospitalisation due to HF. Conclusion: serial ambulatory controls of clinical, laboratory and instrumental variables allow improving the prediction of death or hospitalisation in HF pts. Our data showed that the severity of NYHA class and the value of FE maintain an important prognostic value irrespective of the changes between two periodical ambulatory controls, while the deterioration of BUN in two subsequent ambulatory controls showed an independent prognostic factor on multivariate analysis.
268 Optimal Low-Density Lipoprotein Levels in Patients with Acutely Decompensated Heart Failure Mark R. Kahn, Arzhang Fallahi, Greg Serrao, Gabriel Wagman, Timothy J. Vittorio; Zena and Michael A. Wiener; Cardiovascular Institute, Mount Sinai Medical Center, New York, NY Introduction: Lipid-lowering agents have been successful in reducing primary and secondary cardiovascular events and in decreasing mortality in patients with atherosclerotic heart disease. The two large randomized-controlled trials, GISSI-HF and CORONA, demonstrated no survival benefit with statin use in patients with chronic
Time to death after first hospitalization.
Conclusion: As shown in this retrospective study, maintaining higher levels of LDL may be beneficial in patients with ADHF, although further randomized trials are needed to make this conclusion.
269 Prognostic Scoring in Patients Requiring BIVAD as a Bridge to Cardiac Transplantation Richard K. Cheng1, Chi-Hong Tseng2, Richard Shemin3, W. Robb MacLellan1; 1 Cardiology, UCLA, Los Angeles, CA; 2General Internal Medicine, UCLA, Los Angeles, CA; 3Cardiothoracic Surgery, UCLA, Los Angeles, CA Introduction: Prior studies have identified risk factors for survival in patients with end-stage heart failure (HF) requiring LVAD support. However, patients with biventricular HF may represent a unique cohort. We evaluated patients requiring BIVAD support to generate a risk model to stratify patients at increased risk of death. Methods: We retrospectively evaluated a consecutive cohort of 113 adult, end-stage HF patients at UCLA who required BIVAD support between 1/2000-12/2009. A multivariate model was generated from 10 imputations of our original dataset. The primary outcome was survival to cardiac transplantation. Regression coefficients were used to generate a prognostic scoring system, with fit tested on ROC curves. A simplified scoring system was created by dichotomizing variables with CART. KaplanMeier analysis was used to estimate survival over time. Our final model was validated with bootstrapping. Appropriate IRB approval was obtained. Results: 66.4% of patients survived to transplant. All patients were INTERMACS level 1 or 2 and received Thoratec paracorporeal BIVAD as bridge-to-transplant. Female gender, dialysis, ECMO, low hematocrit, low platelets, low cholesterol, low LDL, low HDL, elevated troponin, low albumin, normal-high sodium, high AST, low cardiac output, ventilator use, and lack of pre-operative antibiotics were associated with increased risk of death. In multivariate modeling, one of the strongest predictors of survival was pre-operative antibiotic use; however, this was excluded from the final model since it can be controlled by the care provider. Our final predictive model based on age, gender, dialysis, cholesterol, ventilator, and albumin gave a c-statistic of 0.870. A simplified scoring system preserved a c-statistic of 0.844. Patients were divided into low (median survival 367 days) or high risk (17 days) groups. The LietzMiller model was applied to our cohort and found to have a c-statistic of 0.713. Other