Predictors Of Gastrointestinal Bleeding Research Articles

Predictors of gastrointestinal bleeding in patients with acute coronary syndrome and the optimal duration of dual antiplatelet therapy.

This study aims to estimate the risk factors of gastrointestinal (GI) bleeding in patients with acute coronary syndrome (ACS) and to evaluate the optimal duration of dual antiplatelet therapy (DAPT). We enrolled 1266 patients with ACS in a telephone follow-up program to determine whether any of the patients were hospitalized for GI bleeding. We collected baseline data, laboratory tests, electrocardiograms, and echocardiography covering all ACS patients. Multivariable regression was performed to adjust for confounders and predictors of GI bleeding. At the same time, the optimal duration of DAPT for ACS patients was evaluated. A total of 1061 ACS patients were included in the study. After 13-68 months, 48 patients (4.5%) were hospitalized for GI bleeding. The risk of GI bleeding was significantly increased in patients treated with DAPT for more than 18 months (hazard ratio 12.792, 5.607-29.185, P < 0.01). Receiver Operating Characteristic curve showed that the duration of DAPT using a cutoff of 14.5 months resulted in a sensitivity of 66.7% and a specificity of 77%. In patients with ACS, DAPT time are the main risk factors of GI bleeding. The optimal duration of DAPT is 14.5 months.

S831 Outcomes, Trends, and Predictors of Gastrointestinal Bleeding in Patients Undergoing Left Ventricular Assist Device Placement: A Nationwide and Propensity Matched Analysis of United States

S831 Outcomes, Trends, and Predictors of Gastrointestinal Bleeding in Patients Undergoing Left Ventricular Assist Device Placement: A Nationwide and Propensity Matched Analysis of United States

#4768 25 HYDROXYVITAMIN D ASSOCIATES WITH GASTROINTESTINAL BLEEDING IN DIALYSIS

Abstract Background and Aims The INSPIRE collaborative group built a prediction model that showed serum 25 hydroxyvitamin D concentration (25OH Vit D) may be a predictor of gastrointestinal bleeding (GIB) hospitalizations in dialysis patients [1]. Unexpectedly, 25OH Vit D levels in the target range (i.e., &amp;gt;30 ng/mL) were most associated with GIB hospitalization risk. We aimed to further investigate the association between GIB episodes and 25OH Vit D levels, considering the onset of a GIB comorbidity or a GIB hospitalization event in the next 180 days. Method Analysis used data from adult (age ≥18 years) dialysis patients in the United States who had one or more 25OH Vit D lab measurement between 01-Jan-2018 to 31-Mar-2021. We identified GIB episodes (comorbidity or hospitalization observations) that were within ≤180 days after a 25OH Vit D measurement. GIB episodes and types were based on ICD-10 diagnosis code clusters defined by the United States Healthcare Cost and Utilization Project [2]. We calculated the percent of GIB comorbidities that onset, or GIB hospitalizations that occurred, by categories of 25OH Vit D levels (&amp;lt;15, ≥15 to &amp;lt;30, ≥30 to &amp;lt;50, ≥50 to &amp;lt;100, or ≥100 ng/mL). Results Among 347,165 adult dialysis patients who had a 25OH Vit D lab measurement(s), 3.2% (n = 11,038) experienced ≥1 GIB comorbidity and 6.7% (n = 23,389) experienced ≥1 GIB hospitalization. In total, 16.5% (n = 4,873) of these patients had ≥1 GIB comorbidity and hospitalization. GIB episodes were within ≤180 days after a 25OH Vit D measurement for 5,558 GIB comorbidity and 21,862 GIB hospitalization observations. We found patients with 25OH Vit D levels that were &amp;lt;15 ng/mL or ≥50 ng/mL exhibited the lowest GIB comorbidity and GIB hospitalization rates (Figure 1A &amp; B). Patients with 25OH Vit D levels that were between ≥15 ng/mL and &amp;lt;50 ng/mL had the highest GIB comorbidity and GIB hospitalization rates. This observation was consistently observed for all GIB types, including upper, lower, and unspecified GIB episodes. Conclusion Findings showed serum 25OH Vit D levels are associated with 180-day GIB episodes in dialysis patients. The highest GIB rates were found in patients with a 25OH Vit D level between 15–50 ng/mL versus all other levels. This finding was consistent for the onset of a of GIB comorbidity and the occurrence of a GIB hospitalization, and in both cases was associated with 2-fold higher GIB rates. Outside the kidney failure population, 25OH vitamin D levels in the range of 30–100 ng/mL have been shown to be associated to higher GIB risk in warfarin users [3]. These findings may question recommendations in kidney disease that suggest repletion of 25OH Vit D levels to ≥30 ng/mL, which has been a controversy for more than a decade. Nonetheless, further analyses are needed and should consider adjustments for patient factors (e.g. age, sex, medications) and competing risks (non-GIB hospitalization events).

Predicting factor analysis of gastrointestinal bleeding complication among hospitalised ischemic stroke patients

Background: Stroke is a common cardiovascular disease and the biggest cause of death and disability in the world. Treatment in affected patients aims to stabilise their condition and prevent further deterioration. One of the therapies in stroke patients is the administration of antiplatelet which may also cause adverse reactions in the form of gastrointestinal bleeding. Objectives: This study aims to analyse the factors that can predict the emergence of gastrointestinal bleeding in hospitalised Ischemic stroke patients. Methods: A cohort study was conducted within a population of hospitalised patients diagnosed with Ischemic stroke. Patients' data from the hospital stroke registry was analysed. The significance of factors that can predict gastrointestinal bleeding incidence and the main predictor was determined. Results: The results showed that out of the 96 patients included, 49 experienced gastrointestinal bleeding. Factors that had significant relationships with the incidence of gastrointestinal bleeding were symptoms of decreased consciousness with OR 9.03; 95% CI: 3.361 – 24.22, and loading dose OR 5.769; 95% CI: 1.191 – 27.941. In addition, the main predicting factor was symptoms of decreased consciousness with a p &lt; 0.05 (&lt;0.010), and OR value of 2.165 with 95% CI: 3.177 – 23.906. Conclusion: The occurrence of decreased consciousness can be a predictor of gastrointestinal bleeding.

Predictors of gastrointestinal bleeding in patients following left ventricular assist deviceimplantation: a systematic review and meta-analysis.

Aim: Our study aims to providea more holistic understanding of the available data and predictive risk factors for gastrointestinal bleed (GIB). Materials &methods: We searched MEDLINE, Embase, Cochrane Central Register of Controlled Trialsand Web of Science Core Collection and calculated relative riskand meta-regression was utilized to evaluate for risk factors in order to assess the effect of covariates. Results: Our meta-analysis reported a pooled prevalence rate of GIB of 24.4%. Meta-regression analysis did not yield a statistically significant association between GIBand risk factors, including age, gender, hypertension, chronic kidney diseaseand diabetes. Conclusion: Studies investigating larger sample sizes are required for conclusive findings.

Incidence and Predictors of Gastrointestinal Bleeding Following Mesenteric Revascularization

Postoperative gastrointestinal bleeding (GIB) following mesenteric revascularization when performed either open (OR) or endovascularly (ER) have been clinically observed but not reported. The aim of the study is to assess the incidence and predictors of GIB in patients undergoing mesenteric revascularization.

Outcomes, Trends, and Predictors of Gastrointestinal Bleeding in Patients Undergoing Transcatheter Aortic Valve Implantation (from the National Inpatient Sample)

Major bleeding has been identified as one of the most common complications after transcatheter aortic valve implantation (TAVI) with some suffering gastrointestinal bleeding (GIB). This study aimed at assessing the incidence and predictors of GIB after TAVI in the United States. We performed a retrospective analysis of data from the National Inpatient Sample database from 2011 to 2018. A total of 216,023 hospitalizations for TAVI were included. Of the included patients, 2,188 (1%) patients had GIB, whereas 213,835 (99%) patients did not have GIB. The presence of arteriovenous malformation was associated with the highest odds of having a gastrointestinal bleed (odds ratio (OR) 24.8, 95% confidence interval (CI) 17.13 to 35.92). Peptic ulcer disease was associated with an eightfold increased risk of bleeding (OR 8.74, 95% CI, 6.69 to 11.43) followed closely by colorectal cancer (OR 7.89, 95% CI, 5.33 to 11.70). Other comorbidities that were associated with higher propensity-matched rates of GIB were chronic kidney disease (OR 1.27,95% CI, 1.14 to 1.41), congestive heart failure (OR 1.18, 95% CI,1.06 to 1.32), liver disease (OR1.83, 95% CI,1.53 to 2.19), end-stage renal disease (OR 2.08,95% CI, 1.75 to 2.47), atrial fibrillation (OR1.63,95% CI, 1.49 to 1.78), and lung cancer (OR 2.80, 95% CI,1.77 to 4.41). Patients with GIB had higher propensity-matched rates of mortality than those without GIB, (12.1% vs 3.2%, p <0.01). Patients with GIB had a higher median cost of stay ($68,779 vs $46,995, p <0.01) and a longer length of hospital stay (11 vs 3 days, p <0.01). In conclusion, health care use and mortality are higher in hospitalizations of TAVI with a GIB. Baseline comorbidities like peptic ulcer disease, chronic kidney disease, liver disease, atrial fibrillation and, colorectal cancer are significant predictors of this adverse event.

Predictors of gastrointestinal bleeding in adult ICU patients in the SUP-ICU trial.

In previous studies of predictors of gastrointestinal (GI) bleeding in the intensive care unit (ICU), most patients received pharmacological stress ulcer prophylaxis (SUP). We aimed to assess associations between potential predictors of clinically important GI bleeding (CIB) and overt GI bleeding in adult ICU patients, while considering the effect and potential interaction with use of SUP. We included 3291 acutely admitted adult ICU patients with risk factors for GI bleeding randomized to SUP (pantoprazole) or placebo in the SUP-ICU trial. We used logistic regression models adjusted for allocation to SUP to estimate associations between 23 potential predictors and CIB (primary outcome) and overt GI bleeding (secondary outcome). Furthermore, we assessed associations between potential predictors and both outcomes in each allocation group and assessed potential interaction with allocation to SUP. Increasing SAPS II and SOFA scores, use of circulatory support and renal replacement therapy were associated with increased risk of CIB and overt GI bleeding; chronic lung disease was associated with increased risk of overt GI bleeding. Results for the remaining potential predictors were compatible with both no difference or increased and decreased risks. We found no strong evidence for any interaction between treatment allocation and any potential predictors. In adult ICU patients at risk of GI bleeding, severity of illness, use of circulatory support and renal replacement therapy were associated with higher odds of CIB, with no strong evidence of interaction with SUP.

Glasgow-Blatchford Score Predicts Post-Discharge Gastrointestinal Bleeding in Hospitalized Patients with Heart Failure.

Background: The Glasgow-Blatchford Score (GBS) is one of the most widely used scoring systems for predicting clinical outcomes for gastrointestinal bleeding (GIB). However, the clinical significance of the GBS in predicting GIB in patients with heart failure (HF) remains unclear. Methods and Results: We conducted a prospective observational study in which we collected the clinical data of a total of 2236 patients (1130 men, median 70 years old) who were admitted to Fukushima Medical University Hospital for acute decompensated HF. During the post-discharge follow-up period of a median of 1235 days, seventy-eight (3.5%) patients experienced GIB. The GBS was calculated based on blood urea nitrogen, hemoglobin, systolic blood pressure, heart rate, and history of hepatic disease. The survival classification and regression tree analysis revealed that the accurate cut-off point of the GBS in predicting post-discharge GIB was six points. The patients were divided into two groups: the high GBS group (GBS > 6, n = 702, 31.4%) and the low GBS group (GBS ≤ 6, n = 1534, 68.6%). The Kaplan–Meier analysis showed that GIB rates were higher in the high GBS group than in the low GBS group. Multivariate Cox proportional hazards analysis adjusted for age, malignant tumor, and albumin indicated that a high GBS was an independent predictor of GIB (hazards ratio 2.258, 95% confidence interval 1.326–3.845, p = 0.003). Conclusions: A high GBS is an independent predictor and useful risk stratification score of post-discharge GIB in patients with HF.

Readmissions Rates After Myocardial Infarction for Gastrointestinal Bleeding: A National Perspective.

Gastrointestinal (GI) bleeding is one most common complications of acute myocardial infarction (AMI). We aimed to determine the incidence, in-hospital outcomes, associated healthcare burden and predictors of GI bleeding within 30days after AMI. Data were extracted from Nationwide Readmission Database 2010-2014. Patients were included if they had a primary diagnosis of ST or non-ST elevation myocardial infarction. Exclusion criteria were admissioned in December, aged less than 18years and a diagnosis of type-2 MI. The primary outcome was 30-day readmission with upper or lower GI bleeding. Secondary outcomes were in-hospital mortality, etiology of bleeding, in-hospital complications, procedures, length of stay, and total hospitalization charges. Independent predictors of readmission were identified using multivariate logistic regression analysis. Out of the 3,520,241 patients discharged with ACS, 10,018 (0.3%) were readmitted with GI bleeding within 30days of discharge. 60% had lower GI bleeding. Most common sources suspected were GI cancers in 17% and hemorrhoidal bleeding in 10%. In hospital mortality rate for readmission was 3.6%. Independent predictors of readmission were age, Charlson comorbidity score, history of chronic kidney disease, GI tumor, inflammatory bowel disease and artificial heart valve. Type of treatment for AMI had no impact on readmission. Patients readmitted had higher rates of shock (adjusted odds ratio, 1.48, 95% CI 1.01-3.72). In the first nationwide study, 30-day incidence of GI bleeding after AMI is 0.3%. GI bleeding complicating AMI carries a substantial in-hospital mortality and cost of care.

Enteric tube placement in patients with esophageal varices: Risks and predictors of postinsertion gastrointestinal bleeding.

Background and AimEnteric tube (ET) placement is approached with caution in patients with esophageal varices (EV) due to concern of causing variceal bleeding. Data are limited on rates and predictors of gastrointestinal bleeding (GIB) in these patients. This study aims to assess the rate and predictors of bleeding from EV after ET placement.MethodsWe performed a retrospective chart review on patients requiring ET access with known EV. Inclusion criteria were age >18 with endoscopically proven EV who required ET placement. Patients who were admitted with, or developed a GIB prior to placement of ET were excluded, as were patients admitted for liver transplantation. Primary outcome was incidence of GIB within 48 h of tube placement. Secondary outcome was a >2 g/dL drop in hemoglobin within 48 h of placement without evidence of bleed. Statistical analysis was performed using Fischer's exact test, Mann–Whitney U test, and univariate logistic regression model.ResultsA total of 75 patients were included in the analysis. The most common etiology of cirrhosis was alcohol (44%). The most common location of EV was in the lower third of the esophagus (61%). The primary outcome was observed in 11 (14.6%) patients. The secondary outcome was found in eight (10.6%) patients. On univariate analysis, GIB was associated with higher MELD‐Na (P = 0.026) and EV located in the lower third of the esophagus (P = 0.048).ConclusionET placement in patients with EV is associated with low risk of bleeding. Elevated MELD‐Na and lower EV location conferred a higher risk of bleeding after ET placement.

Mo1256 – Prevalence and Predictors of Gastrointestinal Bleeding in Patients with Temporary Mechanical Circulatory Support Devices: A Nationwide Perspective

Mo1256 – Prevalence and Predictors of Gastrointestinal Bleeding in Patients with Temporary Mechanical Circulatory Support Devices: A Nationwide Perspective

Predictors of gastrointestinal bleeding in older persons taking nonsteroidal anti-inflammatory drugs: Results from the FDA adverse events reporting system.

Older adults continue to take nonsteroidal anti-inflammatory drugs (NSAIDs) to manage chronic pain. The study's purpose was to identify predictors of gastrointestinal (GI) bleeding in older adults taking NSAIDs. A secondary analysis of the 2016 Food and Drug Administration's Adverse Events Reporting System data was conducted with 1,347 cases aged 65 years and older with an NSAID as the primary suspect for an adverse drug event (ADE). Data included age, sex, NSAID, multiple NSAID use, rivaroxaban, warfarin, clopidogrel, cardiovascular drug (proxy for cardiovascular disease), diabetes drug (proxy for diabetes mellitus), and primary adverse drug response. Aspirin was the primary suspect NSAID in 72.5% of cases. Rivaroxaban was taken in 67.9% of cases. Logistic regression was conducted to predict GI bleed versus other NSAID-related ADEs with age, sex, cardiovascular medication, diabetes medication, warfarin, clopidogrel, concurrent NSAID use, aspirin, and rivaroxaban as predictors. Aspirin, rivaroxaban, and concurrent NSAID were significant predictors of GI bleed. Gastrointestinal bleed risk versus other ADE risk increased by 39.77 times when taking aspirin, rivaroxaban, and another NSAIDs concurrently. Results support reduced NSAID use by older adults, especially aspirin, and avoidance of rivaroxaban in older persons taking NSAIDs.

Predictors of gastrointestinal bleeding in older persons taking nonsteroidal anti-inflammatory drugs: Results from the FDA adverse events reporting system.

Journal of the American Association of Nurse Practitioners 31(3):p 214-215, March 2019. | DOI: 10.1097/JXX.0000000000000222

Gastrointestinal Bleeding During Hospitalization for Transcatheter Aortic Valve Replacement

Introduction: Gastrointestinal bleeding (GIB) has been shown to be the most common bleeding complication post transcatheter aortic valve replacement (TAVR), likely due to multiple factors such as use of antiplatelets agents and anticoagulants, advanced age, chronic kidney disease, and prior history of GI disease. Yet there is a paucity of literature regarding the incidence and risk factors for GIB in this setting. The aim of this study was to determine the incidence and risk factors for GIB immediately following TAVR at UC Davis Medical Center (UCDMC). Methods: We retrospective reviewed medical records of 482 patients who were hospitalized for TAVR at UCDMC between January 1, 2012 to June 30, 2017. The incidence of both upper and lower GIB, and potential risk factors (age, antiplatelet use, chronic liver and kidney disease, Hb level, MCV, PPI use) during the same hospitalization were reviewed. The patterns of gastroenterology consultation, endoscopic findings and therapies were described. Multivariate regression analysis was done to identify independent risk factors for GIB. Results: Mean age was 81.5 years and 50% were males; 23% had CKD stage 3 or worse, 9.5% had history of prior GIB. Mean baseline Hb, MCV and platelet count were 11.9 g/dl, 91.6fl and 191000/ml, respectively. Dual antiplatelet therapy was used in 196 (40.6%) patients. Anticoagulation (mainly with warfarin) was used in 141 (29%) patients. PPIs and H2 receptor blockers were used in 167 (34.6%) and 20 (4%) patient, respectively. Incidence of GIB was 1.4% (7/482) with 5 cases of upper GIB, 1 small intestinal GIB and 1 lower GIB. Endoscopy (2 EGDs, 2 colonoscopies and one video capsule endoscopy) was performed in 5 patients, and showed esophagitis and angiodysplasia (1 in stomach, 1 in jejunum and 1 in cecum) in 2 and 3 patients respectively. Successful endoscopic intervention with argon plasma coagulation was performed in 3 patients with upper and lower angiodysplasia. 1 patient with a deep esophageal ulceration was not a candidate for endoscopic intervention. On multivariate regression, only a low hemoglobin level on admission was associated with increased likelihood of GIB (OR=1.53, CI 1.03-2.27). Conclusion: GIB is rare immediately post- TAVR. Angiodysplasia was the most common etiology in our series of 482 elderly patients, 23% of whom had advanced CKD. When indicated endoscopy is safe and successful in achieving hemostasis. Baseline anemia which is likely due to a preexisting GI lesion was a predictor of GIB in this setting.

Enteric Tube Placement in Patients With Known Esophageal Varices: Risks and Predictors of Post-Insertion Gastrointestinal Bleeding

Introduction: Placement of enteric tubes in cirrhotic patients with portal hypertension is often approached with caution due to the risk of rupture and hemorrhage of esophageal varices (EV). However, the incidence and predictors of gastrointestinal bleeding (GIB) with enteric tube placement in patients with known esophageal varices has not been well described. Methods: We conducted a retrospective review of patients with known EV admitted from January 2010 to January 2017. Patients with endoscopy proven EV and nasogastric or orogastric tube placement confirmed by imaging were included. Patients admitted with GIB were excluded. Demographic data, etiology of liver disease, grade and location of EV, use of beta blocker, previous EV bleed or banding, history of encephalopathy, presence of ascites, baseline labs and MELD score were collected. The primary endpoint was incidence of hematemesis, melena, or hematochezia after enteric tube insertion or during hospitalization. The secondary endpoint was hemoglobin (Hb) drop by more than 2 g/dl within 24 hours of enteric tube placement; patients who underwent liver transplantation during hospitalization were excluded from this analysis. Results: Out of 63 patients, 64% were male and mean age was 58. Hepatitis C was the most common cause of cirrhosis (37%) followed by alcohol related cirrhosis (25%). Forty-four percent of patients had grade 1 EV and 71% of patients had EV located in the lower third of the esophagus. Thirty percent had previous EV bleed while 24% had previous EV banding. The most common indication for admission was altered mental status. Baseline laboratory values prior to enteric tube placement were Hb 9.05 g/dl, INR 2.01, platelets 82.3 109/L. Average MELD score prior to enteric tube placement was 24.3. Over 80% of patients had nasogastric tubes with 40% of all tubes inserted by physicians. The primary outcome of hematemesis occurred in 8 patients (13%) while melena/hematochezia occurred in 9 patients (14%). The secondary outcome of Hb drop, evaluated in 46 patients, was observed in 6 patients (13%). On univariate analysis using logistic regression model, MELD score was the only significant predictor of GIB (P=0.04). Conclusion: Post-enteric tube placement GIB occurred in 27% of patients with known EV. MELD score was the only predictor of GIB. Larger studies are needed to further evaluate the risk and predictors of bleeding after enteric tube placement in patients with esophageal varices.Table: Table. Patient DemographicsTable: Table. Characteristics of Enteric Tube Placement

Gastrointestinal Bleeding During the Index Hospitalization for Mechanical Circulatory Support Devices Implantation, a Nationwide Perspective.

Gastrointestinal bleeding (GIB) is a common adverse event after mechanical circulatory support device implantation. However, the majority of the reported data were obtained from small single-center studies. Our aim was to study the prevalence and predictors of GIB during the index hospitalization of mechanical circulatory support devices implantation using a nationwide database. Nationwide inpatient sample (2009-2011) was used to perform a retrospective cross-sectional study. Adult patients with discharge diagnosis codes of congestive heart failure and procedure codes of left-ventricular assist device (LVAD) or intra-aortic balloon pump (IABP) implantation or orthotopic heart transplant (OHT, reference group) were identified. Our outcome was GIB during the index hospitalization when the device was implanted. Predictors that achieved statistical significance on the univariate analysis were included in a multivariable logistic-regression analysis. A total of 87,462 patients were included, 87% of the patients received an IABP, 6% received LVAD, and 5% underwent OHT. Prevalence of GIB was 8, 5, and 3% among those who had LVAD, IABP implantation, and OHT recipients, respectively (p<0.001). Patients who underwent LVAD implantation had twofold increase in the prevalence of GIB (OR 2.1, 1.7-2.5, p<0.001) when using IABP or OHT groups as a reference. This increase in the prevalence was not demonstrated among IABP recipients on a multivariate level. Prevalence of GIB was higher among LVAD compared to OHT and IABP recipients and could occur as early as the index admission of the device implantation.

Tu1362 Temporal Relationship and Predictors of Gastrointestinal Bleeding in Recipients of Non-Pulsatile Left Ventricular Assist Devices

Tu1362 Temporal Relationship and Predictors of Gastrointestinal Bleeding in Recipients of Non-Pulsatile Left Ventricular Assist Devices

Abstract 19925: Novel Pre-implant Characteristics Associated With 12 Month Risk of Gastrointestinal Bleeding Among Patients Supported With Continuous Flow Left Ventricular Assist Device Therapy

Introduction: Continuous-flow Left Ventricular Assisted Device (CF-LVAD) therapy is increasingly utilized for patients with end stage heart failure. Among the most common and unpredictable complications after CF-LVAD is gastrointestinal bleeding (GIB). Hypothesis: We hypothesize that pre-implant characteristics are associated with GIB post-implant of a CF-LVAD. The aim of this analysis is to identify novel pre-implant factors that influence risk of post-implant GIB. Methods: All CF-LVAD implants between January 2006 and December 2014 among patients who survived more than 15 days were included and followed for 12 months. Primary event was GIB and patients were censored at time of re-implant, heart transplant, or death. Student’s t-test was used to compare continuous variables and chi-square test for categorical variables. Cox Proportional Hazards model was used to identify univariate and multivariable models predicting GIB. Results: Among the total 257 patients included, 65 (25.3%) were identified as having a GIB. Baseline differences and their independent univariate hazard ratio (HR) are noted in Table 1. Using stepwise selection and developing a multivariable model, prior GI abnormalities (HR=2.12, p&lt;0.01), prior percutaneous coronary intervention (PCI: HR=2.65, P&lt;0.01) and chronic obstructive pulmonary disease (COPD: HR=1.73, P&lt;0.01) remained statistically significant predictors of GIB (Table 2). Conclusions: We describe novel risk factors (history of GI abnormalities, prior PCI, COPD) as predictors for developing GIB post CF-LVAD implantation. We advocate a closer examination of these risk factors when evaluating patients for CF-LVAD therapy.

Abstract 18299: Device Selection: Does Ventricular Size Matter in LVAD Recipients?

Introduction: Axial and centrifugal-flow left ventricular assist devices (LVAD) have varied response to changes in loading conditions. Whether ventricular size modulates the response to LVAD therapy between the axial flow Heartmate II (HMII) and centrifugal flow Heartware (HVAD) is unknown. Hypothesis: Adverse event and survival rates between devices are affected by recipient pre-op left ventricular end-diastolic diameter (LVEDD). Methods: Data on 1,149 continuous-flow LVAD recipients from May 2004 to May 2015 was pooled from the Mechanical Circulatory Support Research Network (MCSRN). The effect of pre-op LVEDD on the instantaneous hazard of death and adverse events was evaluated between HMII (n=864) and HVAD (n=283). Results: The average LVEDD was 7.0cm and did not differ between the device recipients. Patients with a smaller LVEDD (≤ 6.3cm) were more likely female 39% v. 24% (p &lt; 0.0001) and to have an ischemic cardiomyopathy (30% v. 24%, p = 0.045). For every 1cm increase in pre-op LVEDD, the adjusted hazard ratio (HR) for mortality increased with HVAD (HR 1.31, 95% CI 0.94-1.82, p = 0.005) and decreased with HMII (HR 0.78, 95% CI 0.67-0.92, p = 0.005), Figure 1a. Other predictors of mortality included age (HR 1.31, 95% CI 1.12 - 1.53 per 10y, p &lt; 0.001) and bilirubin (HR 1.15, 95% CI 0.94-1.42 per mg/dL, p = 0.027). LVEDD by device type interaction was not a significant predictor of gastrointestinal bleeding or pump thrombosis. The risk of stroke increased with LVEDD for HVAD and decreased for HMII, but this did not reach statistical significance, Figure 1b. Ischemic etiology of heart failure (HR 1.7 95% CI 1.14 - 2.55, p = 0.01) and HVAD device use (HR 1.47 95% CI 1.00 - 2.17, p = 0.03) were significant predictors of stroke. Conclusions: Our data suggest important differences exist in patient outcomes between HMII and HVAD based on pre-op ventricular size. Studies are needed to determine if these differences in outcome are due to device or flow-specific differences in patient support.