Baseline Renal and Hepatic Function Predict Mortality in VA-ECMO

S134 Journal of Cardiac Failure Vol. 25 No. 8S August 2019 of extravascular lung water. In this context, the purpose of this research was to assess the dynamic features of B-line measurements during changes in position in patients with heart failure. Methods and Results: The study population consisted of individuals who were nearing completion of in-hospital decongestion for acute decompensated heart failure, Screened patients met the inclusion criteria of having no more than grade 1 peripheral edema and JVP<10, who were off intravenous diuretic medication and were free of orthopnea. We employed the method of Jambrik et al. to quantitate B-lines using 16-sectors per lung (max 5 per sector). The patients underwent Bline quantitation after sitting in an upright, seated position for 1 hour with measurements repeated after reclining in a supine position for 20 minutes. In the eight patients studied, the average B-line score at Baseline was 48§5 (SEM) and increased to 71§ 5.9 while supine (p<0.0001). After one-hour returning to a seated position, the B-line score decreased to 61§5.1, and did not return to the baseline level in any patient. Conclusions: These preliminary results underscore the dramatic and rapid effects of patient positioning upon B-line scores among patients with heart failure. These results have implications for the use of lung ultrasound for patient triage in the emergency department, decisions about timing of hospital discharge and for potential applications related to experimental therapeutics.

367 The Effect of Nurse Navigation Program on 30 Day Heart Failure Readmissions Meenaz Bachoo1, Kristen Tecson2; 1Baylor Scott & White The Heart Hospital Plano, Plano, TX; 2Baylor Scott & White Heart and Vascular, Dallas, TX Background: In May 2016, Baylor Scott and White The Heart Hospital Plano (BSWTHHP), Texas hired a Nurse Navigator to reduce 30 day Heart Failure (HF) readmissions. The Nurse Navigation Program (NNP) consisted of a combination of interactions with a nurse navigator, enrollment in remote Telemonitoring service, and HF education from the multidisciplinary team. We sought to evaluate the impact of this program on 30 day HF readmissions. Methods: We performed a retrospective chart review for consecutive patients who admitted to BSWTHHP for HF from July 4, 2014 - June 30, 2018. We created a flag for hospital discharges occurring on or after October 1, 2016 to separate the data into a pre- and post-program timeframe. We assessed the relation between the intervention and 30 day readmissions via ChiSquare test and performed multivariable logistic regression to assess the relation while adjusting for other patient characteristics. Results: There were 1248 records meeting criteria for analysis, 560 (44.9%) of which occurred during the intervention time period. There were no statistically significant differences in patient characteristics across study time periods, but the 30 day readmission rate was significantly lower for those who were hospitalized during the intervention period (3% vs 11%, p < 0.001). After adjusting for insurance status and age, the effect of the intervention remained clear; those who were admitted prior to the collaborative discharge protocol were approximately 4 times more likely to readmit than those who were admitted after the initiation of the collaborative discharge protocol (odds ratio = 4.0, 95% CI 2.2 - 7.1, p < 0.001). Conclusions: In this single center experience, we found that implementing a specialized HF discharge program was associated with a significant reduction in 30 day readmissions. Larger, randomized trials may be warranted to further examine efficacy.

available donor hearts, LVADs are being used with increased frequency to treat advanced heart failure. Use of the educational intervention we developed may improve important patient outcomes.

369 ICD Codes for Stress Cardiomyopathy in Administrative Databases Have High Positive Predictive Values Anusha G. Bhat, Kevin White, Kyle Gobeil, Tara Lagu, Peter Lindenauer, Quinn Pack; Baystate Medical Center, Springfield, MA Introduction: Stress Cardiomyopathy (SCM) gained an International Classification of Disease code at the 9th- Clinical Modulation (ICD- 9 CM) in 2006. Since then, more than a dozen studies have used these ICD codes for SCM, but no prior study has assessed the reliability of these codes, suggesting that prior reports could have under or overestimated the incidence of this emerging heart failure syndrome. Objective: To determine the positive predictive value (PPV) of the ICD 9/10 code to detect a clinical diagnosis of SCM. Methods: We conducted a retrospective cohort study of all stress cardiomyopathy cases seen at a large tertiary medical center from 20102016. We used ICD-9/10 to identify all adult patients hospitalized with incident SCM. We then used chart review to confirm documentation of SCM, which was confirmed by a cardiologist. We recorded the use of echocardiograms and cardiac angiograms; finally, we determined if patients had a true “incident” clinical manifestation of SCM based on chart review. This determination included documentation of clinical judgement by the clinician, electrocardiogram, biochemical markers, echocardiogram, and cardiac angiographic findings of SCM. We trended the incidence over time using Cochrane Armitage trend test and calculated PPV for the diagnosis of SCM. Results: We identified a total of 592 patients hospitalized with SCM ICD-9/10 codes; 242 (41%) consisting of primary SCM codes. Upon chart review, twelve patients (2.2%) had an ICD diagnosis of SCM without any clinical documentation in the chart, which resulted in an overall PPV of 98.0%; 100% PPV for primary, whereas 96.6% for secondary SCM as all the above mentioned 12 patients were coded as secondary SCM. Among primary ICD codes for SCM, 98% had active disease, while 87% of secondary codes had active manifestations of SCM. Primary SCM had higher rates of cardiac catheterization (88% versus 53%; p<0.001) and echocardiogram (95.9% vs 88.6%; p<0.002) as compared with secondary diagnosis. Over the study period, the incidence of SCM increased significantly, particularly among those with secondary SCM. (X2= 22.2, Cochrane -Armitage p < 0.0001, Figure). Conclusions: We found that ICD codes for SCM have a high PPV of 98%, which was higher when sample was limited to principle diagnosis of SCM thus, proving good reliability of these ICD codes in identifying SCM. Although our study was not set up to determine sensitivity or specificity of these codes, the high PPV increases our confidence in prior studies of SCM and suggests that future studies using these codes in administrative databases will likely produce valid results.

368 The Effect of an Educational Video on Device-Related Concerns in a SingleCenter Left Ventricular Assist Device Population Devan Costelle, Jessica Harman, Debra Moser; University of Kentucky, Lexington, KY Introduction: Left ventricular assist device (LVAD) implantations are increasing across the country as a treatment modality for patients with end-stage heart failure. With the increasing prevalence of patients with these devices, there is an urgent need for improved education addressing lifestyle considerations and concerns. The purposes of this study were to evaluate the prevalence of various concerns among patients with an LVAD and evaluate the effect of a short educational video intervention on reducing those concerns in a single-center adult LVAD population. Hypothesis: We hypothesize that through the implementation of a short, educational video focused on specific device-related concerns, patients will report lower levels of these device-related concerns. Methods: A 15-point LVAD Concerns Scale was created to evaluate the prevalence of specific device-related concerns. An educational video was created to address the concerns presented in the LVAD Concerns Scale. Data collection took place from November 2018 to February 2019. A cross-sectional, pre-test post-test study design was used to both identify various device-related concerns that exist among patients with an LVAD and examine the effectiveness of an educational intervention on reducing patients’ device-related concerns. Results: Of the available patients with an LVAD who met inclusion criteria, 30 were enrolled to participate. The highest concerns were related to the LVAD abruptly malfunctioning, being a burden to their partner or family, and having no warning the device may malfunction. The average total score pre-video was 22 § 14, and post-video the mean total score was 17 § 13 (p-value = 0.002). Six individual questions saw statistically significant reductions in scores post-video. Conclusions: Implementation of an educational intervention tailored to address patients’ specific LVAD related concerns resulted in a statistically significant decrease in those concerns. With the disparity between the number of patients awaiting heart transplantation and the number of

Figure. Trends in Incidence of Primary and Secondary Diagnosis Documentation of SCM from 2010 to 2016.

370 Baseline Renal and Hepatic Function Predict Mortality in VA-ECMO Matthew S. Delfiner1, Kyle Peters2, Yoshiya Toyoda2, Eman Hamad2; 1Temple University Hospital, Philadelphia, PA; 2Lewis Katz School of Medicine at Temple University, Philadelphia, PA Background: Venoarterial Extracorporeal Membrane Oxygenation (VA-ECMO) provides mechanical circulatory support for patients in cardiogenic shock, though risk stratification for potential ECMO patients is inconsistent. We seek to determine if baseline organ dysfunction and post-ECMO cannulation organ dysfunction predict mortality at 30 days of diagnosis of shock. Hypothesis: Patients with renal and hepatic dysfunction at baseline have higher 30 day mortality than those with normal organ function. Methods: All VA-ECMO patients at a single urban quaternary care

The 23rd Annual Scientific Meeting  HFSA center from August 2016 to November 2018 were collected into a database. The electronic medical record was queried for demographic information and laboratory values concerning kidney and liver function. All variables were compared between patients who died while on ECMO and patients who survived to 30 days. Continuous variables were analyzed with t-test and categorical variables compared with chi-square analysis and risk ratios. Results: There were 89 patients on VA-ECMO during the study period. Of these, 58 patients (65%) died while on ECMO and 31 survived (35%). There was no difference in age, sex, and race. The baseline mean creatinine (Cr) for patients who survived was 1.49 mg/dL compared to 2.38 mg/dL for those in the deceased group (p<0.001). There was no difference in the change in Cr after cannulation between the two groups. The mean baseline alanine aminotransferase (ALT) in the survived group was 188 U/L and was 357 U/L in the deceased group (p=0.089); the survived subjects had a mean baseline aspartate aminotransferase (AST) of 146 U/L while the deceased subjects was 424 U/L (p=0.011). While on ECMO, the ALT in the survived and deceased populations were 119 U/L and 545 U/L, respectively (p = 0.02). The AST trended similarly with a mean of 184 U/L for survived subjects and 1136 U/L for deceased subjects (p<0.001). The baseline total bilirubin (T. bili) for those who survived was 1.28 mg/dL and for deceased was 1.72 mg/dL (p=0.008). While on ECMO, the T.bili for the survived population was 2.05 mg/dL compared to 5.72 mg/dL for the deceased (p<0.001). The baseline mean lactate in the survival group was 7.8 mmol/L and 5.5 mmol/L in the deceased group (p=0.093); while cannulated the survived group had a mean lactate of 4.14 mmol/L compared to 9.98 mmol/L in the deceased group (p=0.003). See Figure 1 for mortality risks. Conclusions: Baseline renal and hepatic function predicts ECMO mortality, as does worsening hepatic function while cannulated. By using these trends to risk stratify patients, physicians can more appropriately decide who will most benefit from this procedure.

S135

GI bleeding. This is a single center pre/post evaluation of GI bleeding with use of Octreotide IM 20mg LAR given once a month for a minimum of 6 months. LVAD patients were enrolled after recurrent GI bleeding was diagnosed despite discontinuation of aspirin and reduction of INR goal. To date eight patients have been initiated on our octreotide injection protocol. Results: Despite removal of aspirin and reduction in INR goal, these 8 patients suffered recurrent GI bleeding. With the initiation of octreotide injections, the number of GI bleeds was decreased by one half, with 16 bleeds prior to initiation and seven bleeds post. Six of the bleeds occurred in one patient, another patient re-bled once after a year of therapy. Six patients have not rebled since initiation of octreotide. Treatment time ranged from 2-14 months with an average of 6 months. Though all patients remain off aspirin, all patients have been able to maintain anticoagulation with an INR target in the range of 1.5-2.5 with only one thrombotic event (TIA) diagnosed in this cohort since the initiation of octreotide. Only one patient stopped treatment after two doses due to intolerable diarrhea. Discussion: In our cohort, IM octreotide was well tolerated and there was a noticeable decrease in recurrent GI bleeding. Despite our small numbers, these results are encouraging, as it has led to a decreased need for blood transfusions, GI procedures and hospitalization due to GI bleeding.

372 Women are More Likely to be Physically Frail Compared with Men in Heart Failure Quin E. Denfeld1, Shirin O. Hiatt1, S. Albert Camacho1, James O. Mudd2, Jonathan Q. Purnell1, Mary Davis1, Kerri Winters-Stone1, Christopher S. Lee3; 1Oregon Health & Science University, Portland, OR; 2Providence Sacred Heart Medical Center, Spokane, WA; 3Boston College, Chestnut Hill, MA Introduction: Physical frailty is highly prevalent in heart failure (HF); however, little is known about the role of age, gender, ejection fraction phenotype (HF with reduced ejection fraction (HFrEF) vs. HF with preserved ejection fraction (HFpEF)), and HF etiology (ischemic vs. non-ischemic) in explaining physical frailty in HF. Hypothesis: There is a higher prevalence of physical frailty in adults with HF who are  65 years, are women, have HFpEF, and have non-ischemic HF etiology. Methods: This was an analysis of data collected from a study of adults with HF. Physical frailty was measured with the Frailty Phenotype Criteria: unintentional weight loss, weakness, slowness, physical exhaustion, and low physical activity. Physical frailty was determined by the number of criteria met; those who met 0-2 criteria were not physically frail, and those who met 3-5 were considered physically frail. Comparative statistics were used to compare prevalence of physical frailty between adults  65 years vs. < 65 years, women vs. men, those with HFrEF vs. HFpEF, and those with ischemic etiology vs. non-ischemic etiology. Logistic regression was used to identify significant predictors of physical frailty. Results: The average age of the sample (n = 65) was 62.0§16.6 years, and 35.3% were female. A majority (84.6%) had HFrEF, and most (76.9%) had non-ischemic etiology. Physical frailty was identified in 43% of the sample. Adults  65 years were more likely to be physically frail than those < 65 years (x2 = 4.77, p = 0.029). Women were more likely to be physically frail than men (x2 = 10.19, p = 0.001). Those with HFpEF were more likely to be physically frail than those with HFrEF (x2 = 6.57, p = 0.015). There were no significant differences in etiology (x2 = 0.004, p = 0.95). In a multivariate logistic regression model, gender was the only significant predictor of physical frailty. Women were over 4 times more likely to be physically frail compared with men (OR [95% CI] = 5.5 [1.57, 19.38]). Conclusions: In HF, women have higher rates of physical frailty compared with men, irrespective of age, ejection fraction phenotype, and etiology, indicating that women with HF may be at increased risk for poor outcomes related to physical frailty.

373 A Descriptive Analysis of ReDS Technology across the Continuum of Care Lisa Curran1, Kelly Peck2, Daniel Bensimhon1; 1Cone Health, Greensboro, NC; 2Triad Health Network, Greensboro, NC

371 Octreotide: Prevention of Recurrent Gastrointestinal Bleeding in LVAD Patients Lisa Curran, Erika Nicolsen, Molly Reece, Sarah Herbert, Daniel Bensimhon, Stephanie Dixon, Dalton McLean; Cone Health, Greensboro, NC Background: Gastrointestinal (GI) bleeding is a common complication of continuous flow Left Ventricular Assist Device (LVAD) therapy leading to frequent hospitalizations, invasive GI procedures, and reduction of antithrombotic therapies. There is a need for ongoing investigation of GI bleed prevention in LVAD patients in order to understand mechanisms, predictors and optimal medical therapy. Currently, very few pharmacological methods have been found to provide reliable decreased GI bleeding with minimal side effects. Octreotide has shown efficacy in treating and preventing GI and variceal bleeding through decreasing splanchnic blood flow. Our goal was to reduce burdensome GI bleeding events while maintaining maximally tolerated anticoagulation therapy to prevent thrombotic events. Methods: The Advanced Heart Failure Team at Moses Cone Hospital has placed over 70s LVADs since inception, May 2013. These patients are a combination of HeartMate II, III and 2 Heart Ware devices. While most LVAD patients are maintained on warfarin with INR goal 2-2.5 and aspirin 81mg daily without difficulty, we have had eight patients with recurrent

Background: Heart failure (HF) affects over 6 million patients in the US and is associated with a significant health and economic burden. Volume overload is a major determinant in the symptoms and quality of life in patients with HF and is the most common cause for HF admissions. The degree of pulmonary congestion can be difficult to assess on standard exam. The Remote Dielectric Sensing (ReDS) Wearable System is an FDA-approved device developed by Sensible Medical Innovations that allows clinicians and other health care providers to quickly and accurately measure lung fluid in patients, noninvasively. Using the ReDS technology our center sought to directly measure the degree of pulmonary congestion in HF patients at various points of care. Methods: The Advanced Heart Failure Program at Moses Cone Hospital in Greensboro, NC is a multidisciplinary program that provides care for nearly 2,000 HF patients both on an inpatient and outpatient basis. We also partner with the Triad Health Network (a large next-gen ACO) and local home health agencies to provide HF care in our surrounding communities. With these partners, we developed an initiative to measure pulmonary congestion at multiple points of care across the continuum. Using the ReDS technology, lung fluid was measured on 108 patients at HF hospital discharge, 190 patients at return office visits to the HF clinic, 105 patients in the home health setting, 58 HF patients presenting to the Emergency Department for increased dyspnea/swelling SOB and 144 HF patients at a community PCP office.