Identify Heart Failure Patients Research Articles

Association between echocardiographic parameters and abnormal spirometry among chronic heart failure patients

Abstract Background A significant proportion (10-40%) of heart failure (HF) patients have concomitant pulmonary disease which have negative effects on cardiovascular outcome of patients. Because of the similarities in symptoms between HF and pulmonary disease, the prevalence of concomitant pulmonary disease may be underestimated. Early detection manage accordingly may improve outcomes of HF patients. We aimed to study clinical and echocardiographic parameters identifying HF patients with concomitant pulmonary disease. Methods This cross-sectional study included adults (age≥18 years) who were diagnosed with HF. The echocardiogram, arterial blood gas and pulmonary function test (PFT) were performed. The patients were classified into normal and abnormal (obstructive and restrictive) PFT patterns. The primary outcome was the association between echocardiographic parameters and abnormal PFT. The receiver operating characteristic (ROC) curve was performed to evaluate diagnostic performance of echocardiographic parameter in detecting abnormal PFT. Results A total of 57 patients was enrolled, of which 32 patients (56.14%) had normal PFT, 14 patients (24.56%) had obstructive PFT, and 11 patients (19.30%) had restrictive PFT. The clinical characteristics were not different between groups. ROC curve demonstrated that only left atrial volume index (LAVI) and main pulmonary artery (MPA) diameter had significant diagnostic accuracy to predict abnormal PFT with AUC of 0.72 (95%CI 0.57-0.87) and 0.66 (95%CI 0.51-0.80) respectively. Using combination of LAVI (cutoff 40mm2/m2) and MPA diameter (cutoff 22.0 mm) had a sensitivity of 95.7% (95%CI 78.05-99.89%) and specificity of 41.94% (95%CI 24.55-60.92%) in detection patients with abnormal PFT. Conclusion Almost half of HF had concomitant abnormal PFT. The combination of LAVI and MPA diameter can identify patients with abnormal PFT with high negative predictive value and should be considered as screening parameters in HF patients.

Association of osteopontin and shrunken pore syndrome with hospitalization for renal failure in acute heart failure patients

Abstract Background Shrunken pore syndrome (SPS) emerges as a distinctive manifestation of glomerular filtration dysfunction mainly manifested by the impairment of moderate-sized molecular filtration, distinguished by a variance in renal filtration between cystatin C and creatinine. SPS exhibits correlations with adverse prognosis of various cardiovascular diseases. A decline in renal function correlates with a poor prognosis among individuals with heart failure (HF). Osteopontin (OPN), a protein involved to bone mineralization, has been linked to deteriorating renal function and heightened risk for adverse outcomes in patients with chronic kidney disease. Purpose To explore if OPN and SPS are associated with hospitalization for renal failure in patients with acute HF. Methods Cystatin C and creatinine were analyzed in 395 hospitalized HF patients, of which 324 also had osteopontin (OPN) analyzed using a proximity extension assay. SPS was calculated using the CKD-EPI Equations for Glomerular Filtration Rate and defined as a cystatin C-based estimated glomerular filtration rate (eGFR) <60% of the creatinine based eGFR. Hospitalization for renal failure was extracted from regional registries using ICD10 codes N17-N19. Logistic and Cox regression models were deployed to explore a) OPN’s association with SPS; b) OPN’s association with hospitalization for renal failure; and c) SPS’s association with hospitalization for renal failure, adjusting for age, sex, systolic blood pressure and use of renin-angiotensin-aldosterone system (RAAS) inhibitors. Results The study population had a mean age of 74.8 (±12.0) years, 32.4% were women and 104 (26.5%) patients presented with SPS (Table 1). During the median follow-up period of 71 (interquartile range 56-88) months, 35 (8.9%) patients were hospitalized for renal failure. OPN was associated with the prevalence of SPS (OR 1.66; 95%CI 1.08-2.27; p=0.002), as well as with higher risk of hospitalization due to renal failure (HR 2.71; 95%CI 1.61-4.56; p<0.001). Further, also prevalent SPS was associated with higher risk of hospitalization due to renal failure (HR 2.44; 95%CI 1.14-5.24; p=0.022). Conclusions Here, we demonstrate that higher OPN levels are associated with a higher prevalence of SPS and an elevated risk of hospitalization due to renal failure among patients with HF. Additionally, the presence of SPS was associated with a higher risk of renal failure-related hospitalizations. These findings highlight the potential utility of OPN and SPS as biomarkers for identifying heart failure patients at risk of renal function decline, underscoring the importance of early risk stratification and targeted interventions in this population.

Assessing stroke and mortality risk in heart failure: the CHA2DS2-VASc score's prognostic value in patients with and without atrial fibrillation - a systematic review and meta analysis

Abstract Background Heart failure (HF) is a common cardiovascular disorder that is linked to substantial morbidity and mortality. The CHA2DS2-VASc score, originally designed to assess the risk of stroke in patients with atrial fibrillation, has demonstrated potential as a prognostic indicator for adverse outcomes in patients with heart failure. This systematic review and meta-analysis aimed to assess the predictive accuracy of the CHA2DS2-VASc score in determining the occurrence of stroke and increased mortality rates in patients with heart failure. Methods A thorough search of electronic databases, such as PubMed, EMBASE, and the Cochrane Library, was performed to find pertinent papers published until the search cutoff date. Included studies examined the correlation between the CHA2DS2-VASc score and the likelihood of stroke or death in patients with heart failure. The assessment of the quality of the included studies was conducted using recognized criteria, and data were collected for quantitative synthesis. Findings The meta-analysis encompassed four studies (45864 patients), revealing a consistent and substantial correlation between elevated CHA2DS2-VASc scores and heightened risks of both stroke and mortality in heart failure patients. More precisely, patients with CHA2DS2-VASc scores ≥4 had a greater risk of stroke compared to those with scores <4 (odds ratio 0.38, 95% confidence interval 0.33-0.43, p-value < 0.00001). The combined data showed homogeneity (p-value = 0.24, I2 = 29%). Similarly, patients with CHA2DS2-VASc scores ≥4 had a higher risk of mortality compared to those with scores <4 (OR 0.49, 95% CI 0.30-0.80, P = 0.05). However, the combined data showed significant heterogeneity (P < 0.00001, I2 = 96%). Conclusion The CHA2DS2-VASc score is a useful predictive tool for identifying heart failure patients who are hospitalized and at a high risk of both stroke and high mortality. The results of this study provide evidence for the practicality of using the CHA2DS2-VASc score for risk stratification in patients with heart failure. Additional investigation is necessary to confirm these findings and examine the incorporation of the CHA2DS2-VASc score into risk assessment algorithms for tailored patient management.Figure:Risk of strokeFigure:Risk of mortality

Machine Learning For Risk Prediction After Heart Failure Emergency Department Visit or Hospital Admission Using Administrative Health Data.

Patients visiting the emergency department (ED) or hospitalized for heart failure (HF) are at increased risk for subsequent adverse outcomes, however effective risk stratification remains challenging. We utilized a machine-learning (ML)-based approach to identify HF patients at risk of adverse outcomes after an ED visit or hospitalization using a large regional administrative healthcare data system. Patients visiting the ED or hospitalized with HF between 2002-2016 in Alberta, Canada were included. Outcomes of interest were 30-day and 1-year HF-related ED visits, HF hospital readmission or all-cause mortality. We applied a feature extraction method using deep feature synthesis from multiple sources of health data and compared performance of a gradient boosting algorithm (CatBoost) with logistic regression modelling. The area under receiver operating characteristic curve (AUC-ROC) was used to assess model performance. We included 50,630 patients with 93,552 HF ED visits/hospitalizations. At 30-day follow-up in the holdout validation cohort, the AUC-ROC for the combined endpoint of HF ED visit, HF hospital readmission or death for the Catboost and logistic regression models was 74.16 (73.18-75.11) versus 62.25 (61.25-63.18), respectively. At 1-year follow-up corresponding values were 76.80 (76.1-77.47) versus 69.52 (68.77-70.26), respectively. AUC-ROC values for the endpoint of all-cause death alone at 30-days and 1-year follow-up were 83.21 (81.83-84.41) versus 69.53 (67.98-71.18), and 85.73 (85.14-86.29) versus 69.40 (68.57-70.26), for the CatBoost and logistic regression models, respectively. ML-based modelling with deep feature synthesis provided superior risk stratification for HF patients at 30-days and 1-year follow-up after an ED visit or hospitalization using data from a large administrative regional healthcare system.

Implementation of a heart failure disease management program in the rural southern United States: a best practice implementation project.

Heart failure affects thousands of patients annually, often resulting in hospitalization, emergency room visits, and decreased quality of life. The scientific evidence indicates that disease management programs using a multidisciplinary approach improve outcomes for heart failure patients. The overarching aim of this evidence implementation project was to establish a disease management program for heart failure patients. Specifically, the project sought to promote compliance with best practices for the early identification of heart failure exacerbation symptoms to prevent emergency room visits and hospitalizations, thereby reducing morbidity and mortality. This project was guided by the JBI Evidence Implementation Framework, which is grounded in an audit, feedback, and re-audit strategy. A baseline audit was conducted, revealing the need for a heart failure disease management program. A policy was developed to implement a disease management program and virtual relationships were established to create a multidisciplinary team. A follow-up audit was conducted to determine whether the implemented interventions closed the gap in practice identified by the baseline audit. The baseline audit revealed 0% compliance with best practices for the management of patients with heart failure, while the follow-up audit revealed 93.3% compliance. This project enhanced best practices by implementing a protocol in a rural health clinic to ensure that all heart failure patients were enrolled in a disease management program. Recommendations include using the electronic health record to quickly identify heart failure patients and their status in a disease management program. Chart audits should be performed every 6 months to ensure the program's sustainability and to determine heart failure patients' enrollment status. http://links.lww.com/IJEBH/A270.

Assessing Stroke and Mortality Risk in Heart Failure: The CHA2DS2-VASc Score's Prognostic Value in Patients With and Without Atrial Fibrillation: A Meta-Analysis.

The CHA2DS2-VASc [congestive heart failure, hypertension, age (≥75 years earns 2 points, 65-74 years earns 1 point), diabetes mellitus, prior stroke, transient ischemic attack, or thromboembolism (2 points), vascular disease (eg, prior myocardial infarction, peripheral artery disease), and female sex category] score has demonstrated potential as a prognostic indicator for adverse outcomes in patients with heart failure (HF). This systematic review and meta-analysis aimed to assess the predictive accuracy of the CHA2DS2-VASc score in determining the occurrence of stroke and mortality in HF patients. We did a thorough search of electronic databases until December 2023. Included studies examined the correlation between the CHA2DS2-VASc score and the likelihood of stroke or death in patients with HF. The meta-analysis showed a substantial correlation between elevated CHA2DS2-VASc scores and heightened risks of both stroke and mortality in HF patients. Patients with CHA2DS2-VASc scores ≥4 had a greater stroke risk than those with scores <4 (odds ratio, 0.38, 95% confidence interval, 0.33-0.43, P < 0.00001). Similarly, patients with CHA2DS2-VASc scores ≥4 had a higher mortality risk (OR, 0.49, 95% confidence interval, 0.30-0.80, P = 0.05). The CHA2DS2-VASc score is a useful predictive tool for identifying HF patients who are at a high risk of both stroke and mortality. Additional investigation is necessary to confirm these findings and examine the incorporation of the CHA2DS2-VASc score into risk assessment algorithms for tailored patient management.

Admission prioritization of heart failure patients with multiple comorbidities.

The primary objective of this study was to enhance the operational efficiency of the current healthcare system by proposing a quicker and more effective approach for healthcare providers to deliver services to individuals facing acute heart failure (HF) and concurrent medical conditions. The aim was to support healthcare staff in providing urgent services more efficiently by developing an automated decision-support Patient Prioritization (PP) Tool that utilizes a tailored machine learning (ML) model to prioritize HF patients with chronic heart conditions and concurrent comorbidities during Urgent Care Unit admission. The study applies key ML models to the PhysioNet dataset, encompassing hospital admissions and mortality records of heart failure patients at Zigong Fourth People's Hospital in Sichuan, China, between 2016 and 2019. In addition, the model outcomes for the PhysioNet dataset are compared with the Healthcare Cost and Utilization Project (HCUP) Maryland (MD) State Inpatient Data (SID) for 2014, a secondary dataset containing heart failure patients, to assess the generalizability of results across diverse healthcare settings and patient demographics. The ML models in this project demonstrate efficiencies surpassing 97.8% and specificities exceeding 95% in identifying HF patients at a higher risk and ranking them based on their mortality risk level. Utilizing this machine learning for the PP approach underscores risk assessment, supporting healthcare professionals in managing HF patients more effectively and allocating resources to those in immediate need, whether in hospital or telehealth settings.

Renal Congestion in Heart Failure: Insights in Novel Diagnostic Modalities.

Heart failure is increasingly prevalent and is estimated to increase its burden in the following years. A well-reported comorbidity of heart failure is renal dysfunction, where predominantly changes in the patient's volume status, tubular necrosis or other mechanical and neurohormonal mechanisms seem to drive this impairment. Currently, there are established biomarkers evaluating the patient's clinical status solely regarding the cardiovascular or renal system. However, as the coexistence of heart and renal failure is common and related to increased mortality and hospitalization for heart failure, it is of major importance to establish novel diagnostic techniques, which could identify patients with or at risk for cardiorenal syndrome and assist in selecting the appropriate management for these patients. Such techniques include biomarkers and imaging. In regards to biomarkers, several peptides and miRNAs indicative of renal or tubular dysfunction seem to properly identify patients with cardiorenal syndrome early on in the course of the disease, while changes in their serum levels can also be helpful in identifying response to diuretic treatment. Current and novel imaging techniques can also identify heart failure patients with early renal insufficiency and assess the volume status and the effect of treatment of each patient. Furthermore, by assessing the renal morphology, these techniques could also help identify those at risk of kidney impairment. This review aims to present all relevant clinical and trial data available in order to provide an up-to-date summary of the modalities available to properly assess cardiorenal syndrome.

The Prognostic Significance of Secondary Mitral Regurgitation in Heart Failure Patients with Varying Estimated Pulmonary Artery Systolic Pressure.

Limited research has been conducted to investigate the impact of secondary mitral regurgitation (MR) in heart failure (HF) patients with different levels of estimated pulmonary artery systolic pressure (ePASP). A total of 468 patients suffering from HF and secondary MR were enrolled and categorized into non-severe and severe MR groups based on the degree of MR. The primary endpoint of the study was a composite of cardiovascular death and a first-heart-failure hospitalization. The secondary endpoints were the primary outcomes, individually. The outcomes of the two groups were compared. Patients were further classified based on whether their ePASP was 50 mmHg or 50 mmHg. Subsequently, the outcomes of the non-severe and severe MR groups were compared within each ePASP category. In a median (SD) follow-up of 694 (410) days, severe MR was associated with higher risk for primary endpoints in patients with heart failure, especially in those with ePASP 50 mmHg. In patients with ePASP 50 mmHg, the prognostic value of severe MR was diminished. Assessment of the severity of MR can identify heart failure patients who are at greater risks for poor clinical outcomes. Additionally, the prognostic value of secondary MR was more pronounced in patients with elevated ePASP.

Abstract 19056: Tailoring Strategies to Improve Implementation of a Heart Failure Admission Decision Aid: A Qualitative Study

Introduction: High variability in heart failure (HF) readmissions persists despite national efforts and incentives. A standardized care path, CodeHF, was developed for use in the emergency department (ED) to better identify HF patients warranting admission. CodeHF was implemented in Kansas City in 2018. Despite positive patient and hospital-level outcomes, use of CodeHF has lagged. Hypothesis: A qualitative study was conducted to understand barriers and facilitators behind the implementation of CodeHF with the goal of tailoring strategies to increase usage and scalability to other hospitals. Methods: Semi-structured interview guides were developed based on the Tailored Implementation in Chronic Diseases (TICD) list. 21 stakeholder interviews were conducted and thematic coding of interview transcripts was performed. Codes were matched to a relevant TICD determinant and implementation strategy from the Expert Recommendation for Implementing Change project. A tailored recommendation was created for each strategy. Results: Analyses revealed 10 key themes: 1) Quality of Evidence: Skepticism of evidence; 2) Liability: Concern of CodeHF use in litigation; 3) Technical Difficulties: Challenges and errors during roll-out; 4) Ownership: Lack of ED ownership and champions; 5) Buy-In: Lack of ED buy-in due to the external project origin; 6) Understanding of Tool: Confusion of purpose and intended users; 7) Utility of Tool: Doubts of value in practice; 8) Exceptions to pathway: Concerns about inappropriate use; 9) Admitting Physician Awareness: Lack of recognition of CodeHF by admitting physicians; 10) Follow-up: Appreciation for rapid follow-up for discharged patients. Conclusions: Despite evidence supporting the positive impact of CodeHF, several implementation barriers and few facilitators were identified. Presenting updated evidence, partnering with new champions, educating providers on clinical and legally protective benefits, spreading awareness about technical solutions, inviting ED providers to lead redesign/relaunch, encouraging initial and ongoing educational meetings at all levels, emphasizing the follow-up benefit, and educating cardiologists and hospitalists are potential strategies for overcoming these barriers.

Understanding the Burden of 30-Day Readmission in Patients With Both Primary and Secondary Diagnoses of Heart Failure: Causes, Timing, and Impact of Co-Morbidities

Although efforts to reduce 30-day readmission rates have mainly focused on patients with heart failure (HF) as a primary diagnosis at index hospitalization, patients with HF as a secondary diagnosis remain common, costly, and understudied. This study aimed to determine the incidence, etiology, and patterns of 30-day readmissions after discharge for HF as a primary and secondary diagnosis and investigate the impact of co-morbidities on HF readmission. The National Readmission Database from 2014 to 2016 was used to identify HF patients with a linked 30-day readmission. Patient and hospital characteristics, admission features, and Elixhauser-related co-morbidities were compared between the 2 groups. Readmitted patients in both groups were younger, male, with lower household income, higher mortality risk, and higher hospitalization costs. Over 60% of readmissions were for reasons other than HF, and greater than 1/3 had more than 2 readmissions within 30 days, with a median time to readmission of 12 days. Both cohorts had high readmission rates and high rates of readmission for causes other than HF. Our findings suggest that efforts to reduce 30-day readmission rates should be extended to patients with secondary HF diagnosis, with surveillance extending to 2 weeks postdischarge to identify patients at risk.

Multi-organ sonographic assessment of congestion has incremental prognostic value in outpatients with heart failure irrespective of left ventricular ejection fraction

Abstract Funding Acknowledgements Type of funding sources: None. Background Congestion is a key driver of symptoms, disease progression and prognosis for patients with heart failure (HF), irrespective of left ventricular ejection fraction (LVEF). However, clinically overt congestion is often undiagnosed unless severe. Recent research has shown that ultrasonographic quantification of systemic venous (i.e. inferior vena cava), renal venous and lung (B-lines) congestion is feasible and identifies HF patients with a worse prognosis. Purpose Whether a simultaneous multi-organ evaluation of congestion by ultrasound may improve risk stratification in HF is unknown. In the present study, we assessed signs of ultrasonographic congestion in the inferior vena cava, lungs and kidneys in HF patients across the LVEF spectrum and investigated their association with pathophysiology, clinical characteristics and prognosis. Methods We prospectively enrolled 326 patients with a prior clinical diagnosis of HF attending a routine follow-up visit. We also enrolled 102 consecutive patients with cardiovascular risk factors (including hypertension, type II diabetes mellitus or chronic ischaemic heart disease) and no history of HF; these latter patients had NT-proBNP levels &amp;lt;125 ng/L, LVEF &amp;gt;50%, and were not taking loop diuretics. All patients underwent a complete clinical assessment, blood and urine tests, and an ultrasound evaluation. Ultrasound congestion was defined as inferior vena cava ≥21 mm, highest tertile of lung B-lines or discontinuous renal venous flow. Results The final population consisted of 310 HF patients (median age 77 years, median NT-proBNP 1037 ng/L, 51% with LVEF &amp;lt;50%), and 101 subjects without HF. Among the 224 HF patients (72%) with no clinical signs of congestion, 95 (42%) had at least one sign of congestion by ultrasound (p&amp;lt;0.0001). HF patients with ≥2 ultrasound signs were older, had greater neurohormonal activation (i.e., higher NT-proBNP; Figure 1), lower urinary sodium concentration, and larger left atria despite similar LVEF. During a median follow-up of 13 (interquartile range 6–15) months, 77 HF patients (19%) died or were hospitalized for HF. HF patients without ultrasonographic evidence of congestion had a similar outcome to subjects without HF (reference; HR 1.02, 95% CI 0.86 – 1.35), while those with ≥2 ultrasound signs had the worst outcome (HR 26.7, 95% CI 12.4–63.6), even after adjusting for multiple clinical variables and NT-proBNP (Figure 2). Adding multi-organ assessment of congestion by ultrasound to a clinical model, including NT-proBNP, provided a 28% net reclassification improvement (p=0.03). Conclusion Simultaneous assessment of pulmonary, systemic venous and kidney congestion by ultrasound identifies a high prevalence of sub-clinical congestion associated with poor outcomes in HF, irrespective of LVEF. The speed of investigation, along with the widespread availability and relatively low cost of ultrasound, makes this protocol easy to implement for real-time assessment of congestion.

Beyond exercise oscillatory ventilations: the prognostic impact of loop gain in heart failure.

Exercise oscillatory ventilation (EOV) is a strong prognostic marker in patients with heart failure (HF) and left ventricular (LV) dysfunction. This phenomenon can be explained through a single quantitative measurement of ventilatory instability, the loop gain. Therefore, we aimed to evaluate whether loop gain could be a better tool than subjective EOV evaluation to identify HF patients with a higher risk of major cardiovascular complications. This was a single-center retrospective study that included patients with left ventricular ejection fraction (LVEF) ≤ 50% consecutively referred for cardiopulmonary exercise testing (CPET) from 2016-2020. Loop gain was measured through computational evaluation of the minute ventilation graph. Of the 250 patients included, the 66 that presented EOV also had higher values of loop gain, when compared to patients without EOV. Those with both EOV and higher loop gain had more severe HF, with higher NT-proBNP and VE/VCO2 slope as well as lower peak VO2 and LVEF. On multivariable analysis, loop gain was strongly correlated with the composite endpoint of cardiovascular death, urgent heart transplantation, urgent left ventricular assist device implantation or HF hospitalization, even after correcting for peak VO2, LVEF, VE/VCO2 slope and NT-proBNP. Presence of EOV was not prognostically significant in this analysis. Loop gain is an objective parameter that quantifies ventilatory instability and showed to have a strong prognostic value in a cohort of patients with HF and LVEF ≤ 50%, outperforming the classification of EOV.

A Prognostic Model Based on Clinical Biomarkers for Heart Failure in Adult Patients Following Allogeneic Hematopoietic Stem Cell Transplantation

Heart failure (HF) is an uncommon but serious cardiovascular complication after allogeneic hematopoietic stem cell transplantation (allo-HSCT). Unfortunately, knowledge about early mortality prognostic factors in patients with HF after allo-HSCT is limited, and an easy-to-use prognostic model is not available. This study aimed to develop and validate a clinical-biomarker prognostic model capable of predicting HF mortality following allo-HSCT that uses a combination of variables readily available in clinical practice. To investigate this issue, we conducted a retrospective analysis at our center with 154 HF patients who underwent allo-HSCT between 2008 and 2021. The patients were separated according to the time of transplantation, with 100 patients composing the derivation cohort and the other 54 patients composing the external validation cohort. We first calculated the univariable association for each variable with 2-month mortality in the derivation cohort. We then included the variables with a P value <.1 in univariate analysis as candidate predictors in the multivariate analysis using a backward stepwise logistic regression model. Variables remaining in the final model were identified as independent prognostic factors. To predict the prognosis of HF, a scoring system was established, and scores were assigned to the prognostic factors based on the regression coefficient. Finally, 4 strongly significant independent prognostic factors for 2-month mortality from HF were identified using multivariable logistic regression methods with stepwise variable selection: pulmonary infection (P=.005), grade III to IV acute graft-versus-host disease (severe aGVHD; P=.033), lactate dehydrogenase (LDH) >426 U/L (P=.049), and brain natriuretic peptide (BNP) >1799 pg/mL (P=.026). A risk grading model termed the BLIPS score (for BNP, LDH, cardiac troponin I, pulmonary infection, and severe aGVHD) was constructed according to the regression coefficients. The validated internal C-statistic was .870 (95% confidence interval [CI], .798 to .942), and the external C-statistic was .882 (95% CI, .791-.973). According to the calibration plots, the model-predicted probability correlated well with the actual observed frequencies. The clinical use of the prognostic model, according to decision curve analysis, could benefit HF patients. The BLIPS model in our study can serve to identify HF patients at higher risk for mortality early, which might aid designing timely targeted therapies and eventually improving patients' survival and prognosis.

A Novel Tropical Geometry-Based Interpretable Machine Learning Method: Pilot Application to Delivery of Advanced Heart Failure Therapies.

A model's interpretability is essential to many practical applications such as clinical decision support systems. In this article, a novel interpretable machine learning method is presented, which can model the relationship between input variables and responses in humanly understandable rules. The method is built by applying tropical geometry to fuzzy inference systems, wherein variable encoding functions and salient rules can be discovered by supervised learning. Experiments using synthetic datasets were conducted to demonstrate the performance and capacity of the proposed algorithm in classification and rule discovery. Furthermore, we present a pilot application in identifying heart failure patients that are eligible for advanced therapies as proof of principle. From our results on this particular application, the proposed network achieves the highest F1 score. The network is capable of learning rules that can be interpreted and used by clinical providers. In addition, existing fuzzy domain knowledge can be easily transferred into the network and facilitate model training. In our application, with the existing knowledge, the F1 score was improved by over 5%. The characteristics of the proposed network make it promising in applications requiring model reliability and justification.

Hyponatremia Is a Powerful Predictor of Poor Prognosis in Left Ventricular Assist Device Patients.

Serum sodium is an established prognostic marker in heart failure (HF) patients and is associated with an increased risk of morbidity and mortality. We sought to study the prognostic value of serum sodium in left ventricular assist device (LVAD) patients and whether hyponatremia reflects worsening HF or an alternative mechanism. We identified HF patients that underwent LVAD implantation between 2008 and 2019. Hyponatremia was defined as Na ≤134 mEq/L at 3 months after implantation. We assessed for differences in hyponatremia before and after LVAD implantation. We also evaluated the association of hyponatremia with all-cause mortality and recurrent HF hospitalizations. There were 342 eligible LVAD patients with a sodium value at 3 months. Among them, there was a significant improvement in serum sodium after LVAD implantation compared to preoperatively (137.2 vs. 134.7 mEq/L, P < 0.0001). Patients with and without hyponatremia had no significant differences in echocardiographic and hemodynamic measurements. In a multivariate analysis, hyponatremia was associated with a markedly increased risk of all-cause mortality (HR 3.69, 95% CI, 1.93-7.05, P < 0.001) when accounting for age, gender, co-morbidities, use of loop diuretics, and B-type natriuretic peptide levels. Hyponatremia was also significantly associated with recurrent HF hospitalizations (HR 2.11, 95% CI, 1.02-4.37, P = 0.04). Hyponatremia in LVAD patients is associated with significantly higher risk of all-cause mortality and recurrent HF hospitalizations. Hyponatremia may be a marker of ongoing neurohormonal activation that is more sensitive than other lab values, echocardiography parameters, and hemodynamic measurements.

Incidence of heart failure following exposure to a protein kinase inhibitor, a French population-based study.

Pharmacovigilance signals of heart failure (HF) following exposure to protein kinase inhibitors (PKIs) have been detected in recent years. Our aim was to identify the PKIs most frequently associated with the development of HF. Using the French National Healthcare Database, all patients newly exposed to a PKI between January 2011 and June 2014 were followed up for 18 months. Specific hospitalization diagnosis and long-term HF-related disease codes were used to identify HF patients. HF incidence rate ratios (IRRs) were measured and adjusted hazard ratios (aHRs) were estimated using a Cox model. Sensitivity analyses were performed to limit the potential indication and competitive risk bias. Thirteen PKIs were studied. Among the 49 714 new PKI users registered during the study period, the mean IRR of HF was 3.38 per 100 person-years, with a median time to onset of 155 days. We found a significant increase in the incidence of HF for six medicinal products: pazopanib (aHR = 2.42, 95% confidence interval [CI] 1.67-3.52), dasatinib (aHR = 2.22, 95% CI 1.42-3.44), ruxolitinib (aHR = 2.11, 95% CI 1.69-2.64), crizotinib (aHR = 1.71, 95% CI 1.07-2.72), everolimus (aHR = 1.45, 95% CI 1.26-1.67) and vemurafenib (aHR = 1.37, 95% CI 1.01-1.86). Sensitivity analyses were consistent with our primary analysis. The current study provides knowledge on HF following exposure to a PKI. Additional studies could confirm these results for dasatinib, everolimus, pazopanib and ruxolitinib, and particularly for the two medicinal products with results slightly above the significance threshold, namely, crizotinib and vemurafenib, in our sensitivity analyses.

Abstract 10390: Using a Trial Modelling Framework to Emulate a Target Trial Evaluating Revascularisation Strategies in Patients With Heart Failure

Introduction: We developed a novel trial modelling framework which utilises real world data to emulate a target trial and estimate key parameters to support trial planning. We applied this framework using the UK Hospital Episode Statistics (HES), which captures all admissions at NHS hospitals in England, and reported the findings of an emulated target trial evaluating revascularisation strategies in heart failure (HF) patients with coronary artery disease. Methods: Phenotyping algorithms derived from the target trial protocol were applied to HES, to identify HF patients who had coronary artery bypass grafting (CABG) or percutaneous transluminal coronary angioplasty (PTCA). Trial participants are often different from the target trial population due to voluntary basis of participation. The HES-derived cohort was matched on baseline features (age, gender, comorbidities) to an existing trial (STICH) which targeted the same high-risk population, so that the matched cohort mimics a likely trial population. Treatment effects were estimated using instrumental variable analysis (IVA) and meta-learning models to adjust for confounding. Results: A total of 13,519 HF patients undergoing revascularisation during April 2009 and March 2015 met the eligibility criteria. This HF patient cohort was matched with the STICH cohort, resulting in 2046 patients in the study cohort. The 5-year composite outcome of all-cause mortality or cardiovascular hospitalisation was observed in 51.1% of HF patients who had CABG and 70.0% in those who had PTCA. IVA showed that CABG was associated with a lower risk of the composite outcome (risk difference -16.2%, 95%CI: -20.6% to -11.8%). This was consistent with the treatment effects derived from other machine learning models. Our analyses also identified a trend in greater treatment effects being observed in the matched cohort than unmatched cohort. This suggests that a clinical trial may over-estimate treatment effects when compared to the actual treatment benefit observed in clinical practice. Conclusions: The framework provides a structured process to emulate a trial and anticipate findings of the trial. Our analyses showed that CABG is associated with a greater long-term benefit in HF patients requiring revascularisation.

Abstract 14400: Predicting In-Hospital Mortality Among Patients Admitted With Heart Failure: A Machine Learning Approach

Introduction: Patients hospitalized with heart failure (HF) experience high in-hospital mortality. Machine learning predictive models have the potential to identify heart failure patients at risk of adverse outcomes. Methods: Using the electronic medical record (EMR) at a large academic medical center, we identified 49,457 patients who were admitted with heart failure from 2011 to 2022. 71% were admitted with HF with reduced ejection fraction (HFrEF, EF≤40%) and 29% with HF with mildly reduced or preserved ejection fraction (HFmrEF/HFpEF, EF &gt;40%). Using patient demographics, comorbidities, admission medications, labs, and vitals throughout the hospitalization, we constructed random forest machine learning models to predict in-hospital mortality. We compared model performance to that of logistic regression utilizing the same variables. Results: The median age (25th-75th percentile) of the cohort was 72 (59-83) years, and 45% of the group was female. 71% patients were admitted with HFrEF and 29% patients were admitted with HFmrEF/HFpEF. For the overall cohort, our machine learning model was able to predict in-hospital mortality with a sensitivity (SN) of 91%, a specificity (SP) of 85%, and an area under the curve (AUC) of 0.95 (Figure 1). For those with HFrEF, our model demonstrated a SN of 91%, SP of 87% and AUC of 0.95, and for those with HFmrEF/HFpEF, a SN of 87%, a SP of 89%, and an AUC of 0.94. For all three cohorts, our models outperformed logistic regression. The top predictors of in-hospital mortality were BNP, BUN, and Systolic BP for those admitted with HFrEF, and sex, creatinine, and diastolic BP for those with HFmrEF/HFpEF. Conclusions: Machine learning models built from EMR data have the ability to accurately predict in-hospital mortality among patients admitted for HF. Such prediction tools, when used in the appropriate context, may help flag hospitalized heart failure patients at high risk for in-hospital mortality.

Prevalence and Risk Factors of Cardiac Amyloidosis in Heart Failure: A Systematic Review and Meta-Analysis.

Heart failure (HF) is one of the leading causes of mortality worldwide. Heart failure is also one of the most common presentations of cardiac amyloidosis (CA). Contemporary epidemiological data of CA in HF patients is lacking. Hence, this systematic review and meta-analysis was conducted to determine the prevalence of amyloidosis in HF patients, and to clarify the risk factors of concomitant CA and HF. A systematic review and meta-analysis was performed. Studies were retrieved from Medline, EMBASE, Scopus and Cochrane library. The search was not restricted in time, type or language of publication. The prevalence of CA in HF grouped according to diagnostic techniques and risk factors of CA with HF was analysed. Eleven (11) studies were included, involving 3,303 patients. The pooled prevalence of CA in HF was 13.7%. The overall prevalence of CA in HF with preserved ejection fraction was 15.1%, and that of HF with reduced ejection fraction was 11.3%. The main factors associated with the diagnosis of CA in HF included older age, males, raised NT pro-BNP, increased interventricular septal thickness in diastole, apical sparing, and reduced left ventricular systolic function. A high index of clinical suspicion is required to identify HF patients with CA. Supportive investigations may be helpful when clinically correlated. A considerable proportion of HF patients have CA and certain risk factors may be helpful in increasing suspicion of CA in HF.