TOPLINE:

Researchers developed and externally validated a risk model (SMART2-HF) that predicted the risk for heart failure (HF) in patients with established atherosclerotic cardiovascular disease (ASCVD) and flagged many deemed low risk for recurrent cardiovascular events as having a higher risk for HF.

METHODOLOGY:

• Researchers developed and externally validated the SMART2-HF model to predict the risk for incident HF in patients with established ASCVD but no prior diagnosis of HF.
• They used data from a prospective Dutch cohort linked to national hospital and mortality registries.
• The development cohort included 7698 patients comprising 2065 women and 5633 men with a history of ASCVD and no prior HF, with median ages of 61 and 62 years, respectively.
• The model was externally validated in 240,741 patients up to 80 years of age from six international cardiovascular databases across Europe and globally. The median follow‑up duration across datasets ranged from roughly 1.8 to 12 years.
• The primary outcome was incident HF, defined as the composite outcome of the first hospitalisation with HF or HF-related death. Model performance was assessed by discrimination using Harrell's C statistic.

TAKEAWAY:

• During a median follow-up of 10.6 years, incident HF occurred in 1031 patients (13%) in the development cohort; 96% of events were captured by hospitalisations and 15% were preceded by another cardiovascular event (most often myocardial infarction).
• In external validation, 24,885 incident HF events (10.3%) occurred. The model achieved a pooled C statistic of 0.696, with consistent performance observed between sexes and across risk regions and different ASCVD subtypes.
• In one external validation dataset, approximately 25% of patients judged as low risk for recurrent cardiovascular events were classified by the SMART2‑HF model as having a higher risk for HF.

IN PRACTICE:

"The SMART2-HF model could identify a high-risk population with the largest absolute benefit from preventive treatment and therefore optimize the deployment of (expensive) lifelong therapies. In addition, the SMART2-HF model could aid in deciding upon referral to specialty clinics for management for patients currently managed in primary care," the researchers of the study wrote.

SOURCE:

This study was led by Tessa H. Reitsma, University Medical Centre Utrecht, Utrecht, Netherlands. It was published online on March 11, 2026, in European Heart Journal.

LIMITATIONS:

The SMART2-HF model was derived from a single Western European cohort with limited sample size, particularly among women. HF events were identified through hospital diagnoses, potentially underestimating the absolute risk for HF in cases managed exclusively in primary care. Validation data sources predominantly consisted of individuals of White European ancestry, which may have limited the generalisability of the findings.

DISCLOSURES:

No disclosures or conflicts of interest statements were explicitly provided for the study.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.