Derivation and validation of a machine learning-based risk prediction model in patients with acute heart failure.

Risk stratification is important in patients with acute heart failure (AHF), and a simple risk score that accurately predicts mortality is needed. The aim of this study is to develop a user-friendly risk-prediction model using a machine-learning method. A machine-learning-based risk model using least absolute shrinkage and selection operator (LASSO) regression was developed by identifying predictors of in-hospital mortality in the derivation cohort (REALITY-AHF), and its performance was externally validated in the validation cohort (NARA-HF) and compared with two pre-existing risk models: the Get With The Guidelines risk score incorporating brain natriuretic peptide and hypochloremia (GWTG-BNP-Cl-RS) and the acute decompensated heart failure national registry risk (ADHERE). In-hospital deaths in the derivation and validation cohorts were 76 (5.1 %) and 61 (4.9 %), respectively. The risk score comprised four variables (systolic blood pressure, blood urea nitrogen, serum chloride, and C-reactive protein) and was developed according to the results of the LASSO regression weighting the coefficient for selected variables using a logistic regression model (4 V-RS). Even though 4 V-RS comprised fewer variables, in the validation cohort, it showed a higher area under the receiver operating characteristic curve (AUC) than the ADHERE risk model (AUC, 0.783 vs. 0.740; p = 0.059) and a significant improvement in net reclassification (0.359; 95 % CI, 0.10-0.67; p = 0.006). 4 V-RS performed similarly to GWTG-BNP-Cl-RS in terms of discrimination (AUC, 0.783 vs. 0.759; p = 0.426) and net reclassification (0.176; 95 % CI, -0.08-0.43; p = 0.178). The 4 V-RS model comprising only four readily available data points at the time of admission performed similarly to the more complex pre-existing risk model in patients with AHF.

Misumi K ，Matsue Y ，Nogi K ，Fujimoto Y ，Kagiyama N ，Kasai T ，Kitai T ，Oishi S ，Akiyama E ，Suzuki S ，Yamamoto M ，Kida K ，Okumura T ，Nogi M ，Ishihara S ，Ueda T ，Kawakami R ，Saito Y ，Minamino T ... -  《-》

Usefulness of Incorporating Hypochloremia into the Get With The Guidelines-Heart Failure Risk Model in Patients With Acute Heart Failure.

Although hypochloremia is strongly associated with adverse prognosis in acute heart failure (AHF), it is unknown whether incorporating hypochloremia into the preexisting risk model improves the model performance. We calculated the Get With The Guidelines-Heart Failure (GWTG-HF) risk score in 1,428 patients with AHF (derivation cohort) and developed 2 risk scores incorporating brain natriuretic peptide (BNP) into the GWTG-HF risk score (GWTG-BNP risk score) and incorporating both BNP and hypochloremia (GWTG-BNP-Cl risk score). Hypochloremia was defined as <98 mmol/L. The external validation and comparison of model performance were performed in an independent group of 1,256 patients with AHF (validation cohort). All models were tested for in-hospital mortality. Hypochloremia was observed in 9.4% and 12.2% of the derivation and validation cohorts, respectively. Hypochloremia was an independent predictor of in-hospital mortality in the derivation cohort (odds ratio 2.02; p = 0.028). In the validation cohort, the GWTG-HF, GWTG-BNP, and GWTG-BNP-Cl risk scores demonstrated good discrimination (area under the curve: 0.742, 0.749, and 0.763, respectively). However, the GWTG-BNP-Cl risk score was more reliable than the GWTG-HF and GWTG-BNP risk scores in risk reclassification (net reclassification improvement: 0.491 and 0.408, respectively; p <0.01 for both). Moreover, this score demonstrated a good calibration of the GWTG-BNP-Cl model (Hosmer-Lemeshow test: p = 0.479). In conclusion, incorporating hypochloremia into the preexisting risk model improves the model performance.

Misumi K ，Matsue Y ，Nogi K ，Sunayama T ，Dotare T ，Maeda D ，Kagiyama N ，Kasai T ，Kitai T ，Oishi S ，Akiyama E ，Suzuki S ，Yamamoto M ，Kida K ，Okumura T ，Nogi M ，Ishihara S ，Ueda T ，Kawakami R ，Saito Y ，Minamino T ... -  《-》

Characterization of hemodynamically stable acute heart failure patients requiring a critical care unit admission: Derivation, validation, and refinement of a risk score.

Most patients with acute heart failure (AHF) admitted to critical care units (CCUs) are low acuity and do not require CCU-specific therapies, suggesting that they could be managed in a lower-cost ward environment. This study identified the predictors of clinical events and the need for CCU-specific therapies in patients with AHF. Model derivation was performed using data from patients in the ASCEND-HF trial cohort (n=7,141), and the Acute Heart Failure Emergency Management community-based registry (n=666) was used to externally validate the model and to test the incremental prognostic utility of 4 variables (heart failure etiology, troponin, B-type natriuretic peptide [BNP], ejection fraction) using net reclassification index and integrated discrimination improvement. The primary outcome was an in-hospital composite of the requirement for CCU-specific therapies or clinical events. The primary composite outcome occurred in 545 (11.4%) derivation cohort participants (n=4,767) and 7 variables were predictors of the primary composite outcome: body mass index, chronic respiratory disease, respiratory rate, resting dyspnea, hemoglobin, sodium, and blood urea nitrogen (c index=0.633, Hosmer-Lemeshow P=.823). In the validation cohort (n=666), 87 (13.1%) events occurred (c index=0.629, Hosmer-Lemeshow P=.386) and adding ischemic heart failure, troponin, and B-type natriuretic peptide improved model performance (net reclassification index 0.79, 95% CI 0.046-0.512; integrated discrimination improvement 0.014, 95% CI 0.005-0.0238). The final 10-variable clinical prediction model demonstrated modest discrimination (c index=0.702) and good calibration (Hosmer-Lemeshow P=.547). We derived, validated, and improved upon a clinical prediction model in an international trial and a community-based cohort of AHF. The model has modest discrimination; however, these findings deserve further exploration because they may provide a more accurate means of triaging level of care for patients with AHF who need admission.

Raslan IR ，Brown P ，Westerhout CM ，Ezekowitz JA ，Hernandez AF ，Starling RC ，O'Connor C ，McAlister FA ，Rowe BH ，Armstrong PW ，van Diepen S ... -  《-》

Validation of the Get With The Guideline-Heart Failure risk score in Japanese patients and the potential improvement of its discrimination ability by the inclusion of B-type natriuretic peptide level.

Detailed characteristics of patients with acute heart failure (AHF) in Japan have not been elucidated. Furthermore, international application of risk models obtained in the United States has not been validated. We evaluated the Get With The Guidelines-Heart Failure (GWTG-HF) risk score performance in AHF patients enrolled in the West Tokyo Heart Failure registry, a large, ongoing, prospective, multicenter cohort registry. Variables required for the GWTG-HF risk score were race, age, systolic blood pressure, heart rate, blood urea nitrogen level, sodium concentration, and presence of chronic obstructive pulmonary disease. Score discrimination and calibration were evaluated by the c statistic, Hosmer-Lemeshow statistic, and visual plotting. We conducted additional analyses to determine whether other variables improved the performance of the score. The primary outcome was in-hospital mortality. In total, 1,876 patients were admitted for AHF between April 2006 and August 2014. The patients were predominantly men (60.6%), with a mean age of 73.3 ± 13.6 years. Sixty-eight (3.6%) patients died during hospitalization. The GWTG-HF risk score showed acceptable discrimination; the c statistic for in-hospital mortality in this cohort was 0.763 (95% CI, 0.700-0.826). The calibration plot showed good conformance between the predicted and observed in-hospital mortality. Notably, addition of B-type natriuretic peptide level to the conventional GWTG-HF score significantly improved the discrimination (c statistic, 0.818; 95% CI, 0.771-0.865). The GWTG-HF risk score can be applied in Japanese AHF patients with good discrimination and calibration. Furthermore, addition of B-type natriuretic peptide level improves discrimination and could be considered in future risk models.

Shiraishi Y ，Kohsaka S ，Abe T ，Mizuno A ，Goda A ，Izumi Y ，Yagawa M ，Akita K ，Sawano M ，Inohara T ，Takei M ，Kohno T ，Higuchi S ，Yamazoe M ，Mahara K ，Fukuda K ，Yoshikawa T ，West Tokyo Heart Failure Registry Investigators ... -  《-》

Machine Learning-Based Models Incorporating Social Determinants of Health vs Traditional Models for Predicting In-Hospital Mortality in Patients With Heart Failure.

Segar MW ，Hall JL ，Jhund PS ，Powell-Wiley TM ，Morris AA ，Kao D ，Fonarow GC ，Hernandez R ，Ibrahim NE ，Rutan C ，Navar AM ，Stevens LM ，Pandey A ... -  《-》