Dataset Information

Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone.

ABSTRACT: BACKGROUND:Cardiovascular diseases kill approximately 17 million people globally every year, and they mainly exhibit as myocardial infarctions and heart failures. Heart failure (HF) occurs when the heart cannot pump enough blood to meet the needs of the body.Available electronic medical records of patients quantify symptoms, body features, and clinical laboratory test values, which can be used to perform biostatistics analysis aimed at highlighting patterns and correlations otherwise undetectable by medical doctors. Machine learning, in particular, can predict patients' survival from their data and can individuate the most important features among those included in their medical records. METHODS:In this paper, we analyze a dataset of 299 patients with heart failure collected in 2015. We apply several machine learning classifiers to both predict the patients survival, and rank the features corresponding to the most important risk factors. We also perform an alternative feature ranking analysis by employing traditional biostatistics tests, and compare these results with those provided by the machine learning algorithms. Since both feature ranking approaches clearly identify serum creatinine and ejection fraction as the two most relevant features, we then build the machine learning survival prediction models on these two factors alone. RESULTS:Our results of these two-feature models show not only that serum creatinine and ejection fraction are sufficient to predict survival of heart failure patients from medical records, but also that using these two features alone can lead to more accurate predictions than using the original dataset features in its entirety. We also carry out an analysis including the follow-up month of each patient: even in this case, serum creatinine and ejection fraction are the most predictive clinical features of the dataset, and are sufficient to predict patients' survival. CONCLUSIONS:This discovery has the potential to impact on clinical practice, becoming a new supporting tool for physicians when predicting if a heart failure patient will survive or not. Indeed, medical doctors aiming at understanding if a patient will survive after heart failure may focus mainly on serum creatinine and ejection fraction.

SUBMITTER: Chicco D

PROVIDER: S-EPMC6998201 | biostudies-literature | 2020 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone.

Chicco Davide D Jurman Giuseppe G

BMC medical informatics and decision making 20200203 1

<h4>Background</h4>Cardiovascular diseases kill approximately 17 million people globally every year, and they mainly exhibit as myocardial infarctions and heart failures. Heart failure (HF) occurs when the heart cannot pump enough blood to meet the needs of the body.Available electronic medical records of patients quantify symptoms, body features, and clinical laboratory test values, which can be used to perform biostatistics analysis aimed at highlighting patterns and correlations otherwise und ...[more]

PMID: 32013925

Similar Datasets

Project description:Background and aimsPatients with heart failure (HF) with reduced left ventricular (LV) ejection fraction (EF) have a heterogeneous prognosis, and assessment of coronary physiology with coronary flow velocity (CFV) and coronary flow velocity reserve (CFVR) may complement established predictors based on wall motion and EF.Methods and resultsIn a prospective multicenter study design, we enrolled 1,408 HF patients (age 66 ± 12 years, 1,035 men), with EF <50%, 743 (53%) with coronary artery disease, and 665 (47%) with normal coronary arteries. Recruitment (years 2004-2022) involved 8 accredited laboratories, with inter-observer variability <10% for CFV measurement. Baseline CFV (abnormal value >31 cm/s) was obtained by pulsed-wave Doppler in mid-distal LAD. CFVR (abnormal value ≤2.0) was assessed with exercise (n = 99), dobutamine (n = 100), and vasodilator stress (dipyridamole in 1,149, adenosine in 60). Inducible myocardial ischemia was identified with wall motion score index (WMSI) stress > rest (cut-off Δ ≥ 0.12). LV contractile reserve (CR) was identified with WMSI stress < rest (cutoff Δ ≥ 0.25). Test response ranged from score 0 (EF > 30%, CFV ≥ 32 cm/s, CFVR > 2.0, LVCR present, ischemia absent) to score 5 (all steps abnormal). All-cause death was the only endpoint. Results. During a median follow-up of 990 days, 253 patients died. Independent predictors of death were EF (HR: 0.956, 95% CI: 0.943-0.968, p < 0.0001), CFV (HR: 2.407, 95% CI: 1.871-3.096, p < 0.001), CFVR (HR: 3.908, 95% CI: 2.903-5.260, p < 0.001), stress-induced ischemia (HR: 2.223, 95% CI: 1.642-3.009, p < 0.001), and LVCR (HR: 0.524, 95% CI: 0.324-.647, p = 0.008). The annual mortality rate was lowest (1.2%) in patients with a score of 0 (n = 61) and highest (31.9%) in patients with a score of 5 (n = 15, p < 0.001).ConclusionHigh resting CFV is associated with worse survival in ischemic and nonischemic HF with reduced EF. The value is independent and additive to resting EF, CFVR, LVCR, and inducible ischemia.

Project description:AimsThis study aimed to investigate the prognostic value of dynamic changes in left ventricular ejection fraction (EF) for cardiovascular (CV) outcomes in an all-comer heart failure (HF) population with reduced EF (HFrEF, EF < 40%). We sought to identify independent factors related to improvement in EF and to identify risk factors for increased risk of CV events in the subgroups of improved EF (iEF) and non-improved EF (niEF), respecively.Methods and resultsThis is a retrospective sub-analysis from the REDEAL HF trial, which included consecutive patients with chronic HF who were hospitalized from July 2009 to December 2017. Baseline and follow-up echocardiography data (interval ≥12 months) of 573 consecutive patients with HFrEF were analysed. iEF was defined as absolute improvement in EF ≥ 10% and follow-up EF over 40%. The primary endpoint was defined as a composite endpoint of cardiovascular (CV) death, CV hospitalization, or appropriate implantable cardioverter-defibrillator (ICD) therapy for ventricular arrhythmia. EF improved in 37.2% of patients with HFrEF during follow-up (median period of 17 months). iEF was independently associated with shorter HF duration (>4 vs. ≤4 years, odd ratio [OR] = 0.477, 95% CI 0.305-0.745), no coronary artery disease (CAD vs. no CAD, OR = 0.583, 95% CI 0.396-0.858), and no ICD implantation (ICD vs. no ICD, OR = 0.341, 95% CI 0.228-0.511). Compared with niEF, iEF was significantly and independently associated with lower all-cause mortality (22.1% vs. 31.1%, P = 0.019; hazard ratio [HR] = 0.674, 95% CI 0.469-0.968), lower CV mortality (8.9% vs. 16.1%, P = 0.015; HR = 0.539, 95% CI 0.317-0.916), and lower CV events risk (27.2% vs. 49.2%, P < 0.001; HR 0.519, 95% CI 0.381-0.708), after adjustment for age, sex, duration of HF, and other clinical risk factors. Hypertension (HR = 2.452, P = 0.032) and elevated N-terminal prohormone of brain natriuretic peptide (NT-proBNP >1153 pg/mL, HR = 4.372, P < 0.001) were identified as independent risk factors for CV events in the iEF subgroup. ICD implantation (HR = 1.533, P = 0.011), elevated NT-proBNP (HR = 1.626, P = 0.018), increased left atrial volume index (HR = 1.461, P = 0.021), reduced lateral mitral annular plane systolic excursion (HR = 1.478, P = 0.025), and reduced tricuspid plane systolic excursion (HR = 1.491, P = 0.039) were identified as risk factors for CV events in the niEF subgroup.ConclusionsImprovement in EF is independently related to the longer survival and lower CV related mortality and hospitalization rate of HFrEF. Elevated baseline NT-proBNP is identified as the strongest prognostic factor associated with increased CV events risk in HFrEF patients both with and without improved EF, regardless of age, sex, duration of HF, and other clinical risk factors.

Dataset Information

Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone.

Publications

Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets