Dataset Information

Machine Learning Approach to Understand Worsening Renal Function in Acute Heart Failure.

ABSTRACT: Acute heart failure (AHF) is a common and severe condition with a poor prognosis. Its course is often complicated by worsening renal function (WRF), exacerbating the outcome. The population of AHF patients experiencing WRF is heterogenous, and some novel possibilities for its analysis have recently emerged. Clustering is a machine learning (ML) technique that divides the population into distinct subgroups based on the similarity of cases (patients). Given that, we decided to use clustering to find subgroups inside the AHF population that differ in terms of WRF occurrence. We evaluated data from the three hundred and twelve AHF patients hospitalized in our institution who had creatinine assessed four times during hospitalization. Eighty-six variables evaluated at admission were included in the analysis. The k-medoids algorithm was used for clustering, and the quality of the procedure was judged by the Davies-Bouldin index. Three clinically and prognostically different clusters were distinguished. The groups had significantly (p = 0.004) different incidences of WRF. Inside the AHF population, we successfully discovered that three groups varied in renal prognosis. Our results provide novel insight into the AHF and WRF interplay and can be valuable for future trial construction and more tailored treatment.

SUBMITTER: Urban S

PROVIDER: S-EPMC9687716 | biostudies-literature | 2022 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine Learning Approach to Understand Worsening Renal Function in Acute Heart Failure.

Urban Szymon S Błaziak Mikołaj M Jura Maksym M Iwanek Gracjan G Ponikowska Barbara B Horudko Jolanta J Siennicka Agnieszka A Berka Petr P Biegus Jan J Ponikowski Piotr P Zymliński Robert R

Biomolecules 20221102 11

Acute heart failure (AHF) is a common and severe condition with a poor prognosis. Its course is often complicated by worsening renal function (WRF), exacerbating the outcome. The population of AHF patients experiencing WRF is heterogenous, and some novel possibilities for its analysis have recently emerged. Clustering is a machine learning (ML) technique that divides the population into distinct subgroups based on the similarity of cases (patients). Given that, we decided to use clustering to fi ...[more]

PMID: 36358966

Similar Datasets

Project description:AimsKidney function changes dynamically during AHF treatment, but risk factors for and consequences of worsening renal function (WRF) at hospital admission are uncertain. We aimed to determine the significance of WRF at admission for acute heart failure (AHF).Methods and resultsWe evaluated a subgroup of 406 patients from The Acute Kidney Injury Neutrophil gelatinase-associated lipocalin Evaluation of Symptomatic heart failure Study (AKINESIS) who had serum creatinine measurements available within 3 months before and at the time of admission. Admission WRF was primarily defined as a 0.3 mg/dL or 50% creatinine increase from preadmission. Alternative definitions evaluated were a ≥0.5 mg/dL creatinine increase, ≥25% glomerular filtration rate decrease, and an overall change in creatinine. Predictors of admission WRF were evaluated. Outcomes evaluated were length of hospitalization, a composite of adverse in-hospital events, and the composite of death or HF readmission at 30, 90, and 365 days. Biomarkers' prognostic ability for these outcomes were evaluated in patients with admission WRF. One-hundred six patients (26%) had admission WRF. These patients had features of more severe AHF with lower blood pressure, higher BUN, and lower serum sodium concentrations at admission. Higher BNP (odds ratio [OR] per doubling 1.16-1.28, 95% confidence interval [CI] 1.00-1.55) and lower diastolic blood pressure (OR 0.97-0.98, 95% CI 0.96-0.99) were associated with a higher odds for the three definitions of admission WRF. The primary WRF definition was not associated with a longer hospitalization, but alternative WRF definitions were (1.3 to 1.6 days longer, 95% CI 1.0-2.2). WRF across definitions was not associated with a higher odds of adverse in-hospital events or a higher risk of death or HF readmission. In the subset of patients with WRF, biomarkers were not prognostic for any outcome.ConclusionsAdmission WRF is common in AHF patients and is associated with an increased length of hospitalization, but not adverse in-hospital events, death, or HF readmission. Among those with admission WRF, biomarkers did not risk stratify for adverse events.

Project description:AimsWorsening renal function (WRF) often develops during heart failure (HF) treatment. However, prognostic implications of WRF in acute HF remain controversial, and risk stratification of WRF is challenging. Although the fibrosis-4 index (FIB-4) was initially established as a liver fibrosis marker, recent studies show that high FIB-4 is associated with venous congestion and poor prognosis in acute HF. This study aimed to evaluate whether FIB-4 could identify prognostically relevant and non-relevant WRF in patients with acute HF.Methods and resultsWe retrospectively analysed data from a single-centre registry on acute HF at our university hospital between January 2015 and June 2021. This study included patients with acute HF aged ≥20 years who were immediately hospitalized and had brain natriuretic peptide levels ≥100 pg/mL at admission. WRF was defined as increases of ≥0.3 mg/dL and >25% in serum creatinine level from admission to discharge. FIB-4 scores were calculated before discharge. The primary endpoint was all-cause mortality within 1 year of discharge. Based on the presence of WRF and whether FIB-4 scores were above the median, patients were stratified into four groups: no WRF and lower FIB-4 scores, no WRF and higher FIB-4 scores, WRF and lower FIB-4 scores, and WRF and higher FIB-4 scores. The patients were followed up via clinical visits or telephone interviews. Clinical outcomes were collected from the electronic medical records.ResultsOf the 969 patients hospitalized for acute HF (76 ± 11 years, 59% men), 118 patients (12%) had WRF at discharge. The median (interquartile range) FIB-4 score at discharge was 2.36 (1.55-3.25). The primary endpoint occurred in 136 patients (14.0%). The 1 year mortality rates were 10.5% in the no WRF and lower FIB-4 scores (≤2.36) group (n = 428), 16.1% in the no WRF and higher FIB-4 scores (>2.36) group (n = 423), 12.5% in the WRF and lower FIB-4 scores group (n = 56), and 25.8% in the WRF and higher FIB-4 scores group (n = 62) (P = 0.005). Kaplan-Meier analysis demonstrated higher all-cause mortality in the WRF and higher FIB-4 group (log-rank P = 0.003). In the Cox regression analysis, only the WRF and higher FIB-4 scores group was associated with an increased risk of mortality compared with the no WRF and lower FIB-4 scores group (hazard ratio = 2.11, 95% confidence interval: 1.07-4.18, P = 0.032), despite adjusting for other confounding factors.ConclusionsFIB-4 is a valuable risk stratification marker for WRF in patients with acute HF. The underlying mechanism and potential clinical importance of these observations require further investigation.

Project description:AimsAcute decompensated heart failure (ADHF) is a frequent cause of hospitalization for patients with heart disease, and ADHF patients are at high risk of heart failure (HF) re-hospitalization. Residual congestion at discharge is also a strong predictor of poor outcomes and re-hospitalization for ADHF patients. However, the impact of residual congestion at discharge on worsening renal function (WRF) in both high-aged and older patients remains uncertain because previous studies of WRF in ADHF patients were conducted for older patients. We therefore designed and conducted a retrospective, population-based study using the Kobe University Heart Failure Registry in Awaji Medical Center (KUNIUMI) Registry to investigate the association of residual congestion at discharge with WRF in ADHF patients according to age.Methods and resultsWe studied 966 hospitalized ADHF patients with a mean age of 80.2 ± 11.4 years from among 1971 listed in the KUNIUMI Registry. WRF was defined as an increase of ≥0.3 mg/dL in the serum creatinine level during the hospital stay compared with the value on admission. The primary endpoint was defined as cardiovascular death or HF re-hospitalization after discharge over a mean follow-up period of 2.0 ± 0.1 years. The primary endpoint was recorded for 369 patients (38.2%). As expected, patients with both WRF and residual congestion at discharge had significantly less favourable outcomes compared with those without one of them, and patients without either of these two characteristics had the most favourable outcomes, whereas those with residual congestion and with WRF had the least favourable outcomes. Moreover, WRF was significantly associated with worse outcomes for high-aged patients ≥80 years old, but not for those <80 years old if decongested. Multivariable Cox regression analysis showed that both residual congestion at discharge and WRF were the independent predictors of outcomes for high-aged patients, but residual congestion at discharge, not WRF, was the independent predictor of outcomes for older patients.ConclusionsAssociation of residual congestion at discharge with WRF for hospitalized ADHF patients can differ according to age. Our findings showed the importance of WRF and residual congestion at discharge for high-aged ADHF patients and of aggressive diuresis to alleviate congestion for older ADHF patients for better management of such patients in a rapidly ageing society.

Project description:AimsThis study aims to develop and validate an optimal model for predicting worsening heart failure (WHF). Multiple machine learning (ML) algorithms were compared, and the results were interpreted using SHapley Additive exPlanations (SHAP). A clinical risk calculation tool was subsequently developed based on these findings.Methods and resultsThis nested case-control study included 200 patients with chronic heart failure (CHF) from the China-Japan Friendship Hospital (September 2019 to December 2022). Sixty-five variables were collected, including basic information, physical and chemical examinations, and quality of life assessments. WHF occurrence within a 3-month follow-up was the outcome event. Variables were screened using LASSO regression, univariate analysis, and comparison of key variables in multiple ML models. Eighty per cent of the data was used for training and 20% for testing. The best models were identified by integrating nine ML algorithms and interpreted using SHAP, and to develop a final risk calculation tool. Among participants, 68 (34.0%) were female, with a mean age (standard deviation, SD) of 68.57 (12.80) years. During the follow-up, 60 participants (30%) developed WHF. N-terminal pro-brain natriuretic peptide (NT-proBNP), creatinine (Cr), uric acid (UA), haemoglobin (Hb), and emotional area score on the Minnesota Heart Failure Quality of Life Questionnaire were critical predictors of WHF occurrence. The random forest (RF) model was the best model to predict WHF with an area under the curve (AUC) (95% confidence interval, CI) of 0.842 (0.675-1.000), accuracy of 0.775, sensitivity of 0.900, specificity of 0.833, negative predictive value of 0.800, and positive predictive value of 0.600 for the test set. SHAP analysis highlighted NT-proBNP, UA, and Cr as significant predictors. An online risk predictor based on the RF model was developed for personalized WHF risk assessment.ConclusionsThis study identifies NT-proBNP, Cr, UA, Hb, and emotional area scores as crucial predictors of WHF in CHF patients. Among the nine ML algorithms assessed, the RF model showed the highest predictive accuracy. SHAP analysis further emphasized NT-proBNP, UA, and Cr as the most significant predictors. An online risk prediction tool based on the RF model was subsequently developed to enhance early and personalized WHF risk assessment in clinical settings.

Dataset Information

Machine Learning Approach to Understand Worsening Renal Function in Acute Heart Failure.

Publications

Machine Learning Approach to Understand Worsening Renal Function in Acute Heart Failure.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets