National Early Warning Score 2 provides independent prognostic information in patients hospitalized with heart failure with preserved ejection fraction

Background

Heart failure (HF) hospitalisation is a strong predictor of long-term mortality among HF patients. Current European Society of Cardiology HF guidelines do not include recommendations for post-discharge long-term risk prediction for acute HF, despite several cardiac biomarkers and clinical scoring systems being able to predict long-term prognosis. In Europe, the clinical scoring system National Early Warning Score 2 (NEWS2) is increasingly being used for in-hospital risk prediction, and several recent studies have demonstrated a potential to utilize NEWS2 for long-term risk prediction.

Purpose

To assess the long-term prognostic accuracy of NEWS2 in patients hospitalised with acute HF in comparison to N-terminal pro-B-type natriuretic peptide (NT-proBNP) and high-sensitivity cardiac troponin T (hs-cTnT), and their differences in prognostic accuracy for HF subgroups with reduced (HFrEF) and preserved (HFpEF) ejection fraction.

Methods

In the Akershus Cardiac Examination 2 Study, we measured hs-cTnT and NT-proBNP and calculated NEWS2 within 24 hours of admission in patients hospitalised due to acute dyspnea (n=314). Patients with acute HF as adjudicated index diagnosis were included in the current sub-study (n=143). We assessed post-discharge long-term risk of all-cause mortality by receiver operating characteristic (ROC) analysis and Cox regression models.

Results

During median follow-up of 26 (quartile 1–3 8–33) months, 66 patients (46%) died. Patients with poor prognosis had higher NEWS2 (mean 5.9 vs 4.4 points, p=0.004), hs-cTnT concentrations (median 49 vs 27 ng/L, p=0.002), and NT-proBNP concentrations (median 5776 vs 2593 ng/L, p<0.001) compared to patients with favourable prognosis. Stratified by HF subgroups, the area under the ROC curve (ROC AUC) for NEWS2 to predict mortality was superior for HFpEF (ROC AUC 0.79, 95% CI 0.66–0.92) compared to HFrEF (ROC AUC 0.57, 95% CI 0.45–0.69; p=0.012) (Figure). In contrast, prognostic accuracy in HFrEF and HFpEF were comparable for hs-cTnT (ROC AUC 0.64 vs 0.67, p=0.76) and NT-proBNP (ROC AUC 0.72 vs 0.71, p=0.94) (Table). After adjustment for age, sex, body mass index, systolic blood pressure, history of coronary artery disease and estimated glomerular filtration rate, NEWS2 was associated with mortality for HFpEF patients (HR 1.37, 95% CI 1.13–1.66), while no association was found for NEWS2 to predict mortality in HFrEF patients (HR 1.03, 95% CI 0.93–1.15; p for interaction=0.016).

Conclusion(s)

NEWS2 is a superior prognostic marker among patients hospitalised with HFpEF compared to HFrEF and might be a valuable and readily available tool to prognosticate patients at hospital discharge.

Funding Acknowledgement

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Research grant from the Norwegian Research Council and internal grants from Akershus University Hospital to Torbjørn Omland and Helge Røsjø.

Markers of subclinical cardiac disease associate with thresholds for pre-diabetes and diabetes in the general population: data from the ACE 1950 Study

Background

Diabetes mellitus (DM) is associated with increased risk of left ventricular (LV) remodeling and incident heart failure. However, the associations between dysglycemia and subclinical cardiac disease in middle-aged subjects recruited from the general population are not established.

Purpose

To assess the associations of dysglycemia and diagnostic DM thresholds with indices of subclinical cardiac injury and dysfunction in the general population.

Methods

We included participants born in 1950 from the Akershus Cardiac Examination 1950 Study with available biomarker measurements (n=3,688). We used regression models and restricted cubic splines (knots selected from lowest Akaike Information Criterion) to assess the association between glycated hemoglobin A1c (HbA1c) and cardiac troponin T (cTnT), N-terminal pro-B-type natriuretic peptide (NT-proBNP), C-reactive protein (CRP), and echocardiographic parameters. We classified participants with self-reported diagnosis of DM or HbA1c ≥6.5% (48 mmol/L) as DM, participants with HbA1c 5.7–6.5% as pre-DM, and participants with HbA1c <5.7% (39 mmol/mol) as no-DM.

Results

Mean age was 63.9±0.7 years, mean body mass index (BMI) 27.2±4.4 kg/m2, and 1,795 participants (49%) were women. DM was classified in 380 participants (10%), pre-DM in 1,630 participants (44%) and no-DM in 1,678 participants (46%). Increasing HbA1c concentrations were associated with younger age, male sex, obesity, hypercholesterolemia, hypertension, and established coronary artery disease in adjusted analyses. In models adjusted for age, sex, BMI, smoking, hypertension, atrial fibrillation, coronary artery disease and renal function, greater HbA1c was associated with increasing logcTnT and logCRP concentrations, decreasing logNT-proBNP concentrations and worse global longitudinal strain and E/e' (p<0.001 for all). LV mass index was not associated with HbA1c in adjusted models (p=0.23). All five associations were non-linear in the total study population (p<0.001 for non-linearity for all) with robust, linear associations in the pre-DM range of HbA1c, also in adjusted models, and attenuated associations in the no-DM and DM range (Figure 1).

Conclusion

We found robust, linear associations between HbA1c and indices of subclinical cardiac injury and dysfunction among participants classified as pre-DM, while associations were more attenuated among participants with DM. Preventive measures for cardiovascular disease should be considered also in patients with dysglycemia and HbA1c below the established cutoff for DM.

Figure 1. P-values for overall trend

Funding Acknowledgement

Type of funding source: Public hospital(s). Main funding source(s): Akserhus University Hospital