Here are the three algorithms that can predict heart failure

Predicting in advance who is at risk of developing heart failure and intervening before the disease manifests itself. This is the aim of three new predictive models developed by an international collaboration of researchers also coordinated by Human Technopole in Milan. The results of the studies have been published in the European Heart Journal. The research offers novel tools to estimate the risk of heart failure in people with different clinical profiles: from individuals without previous heart problems, to those who have already had heart attacks or strokes, to patients with advanced forms of the disease.

Heart failure is one of the most widespread and growing cardiovascular diseases today. According to the most recent estimates, it affects over 60 million people worldwide and about 800,000 in Italia, with 80,000 new cases every year. The ageing of the population, together with the spread of diabetes and obesity and increased survival after heart attacks and strokes, contributes to increasing its incidence. In this context, the ability to predict individual risk becomes a crucial tool for prevention and better treatment planning. The three new models are based on advanced statistical algorithms that analyse large amounts of clinical data collected in international studies and use information readily available in everyday medical practice, such as age, blood pressure, presence of diabetes or kidney function.

"Heart failure can start long before symptoms and present itself in different ways depending on each individual's medical history," explains Emanuele Di Angelantonio, director of Human Technopole's Health Data Science Centre and one of the scientific coordinators of the research. With these three models, we provide tools based on solid data that can be applied throughout the entire course of the disease, from prevention to the most advanced forms'. The research is part of Human Technopole's Flagship Research Programme dedicated to cardiovascular and metabolic diseases, which aims to understand the genetic, environmental and lifestyle factors underlying these diseases. 'These studies demonstrate how advanced clinical data analysis can transform large collections of information into concrete tools for medicine and public health,' observes Marino Zerial, Director of Human Technopole. The aim is to develop new approaches to prevention and personalised medicine'.

The three algorithms are designed to respond to different stages of the clinical pathway. The first, called Score2-HF, concerns people over 40 who have never had cardiovascular disease. Based on data from more than 600,000 individuals from 14 European countries and validated on more than 1.3 million people, it estimates the probability of developing heart failure in the next 10 or 30 years. The model complements the tools already used to assess the risk of heart attack and stroke in the general population. The second model, Smart2-HF, is aimed instead at patients who have already had cardiovascular events such as heart attack or stroke but have not yet developed heart failure. Built on almost 8,000 patients and validated on more than 240,000 cases from different international cohorts, it allows estimation of both ten-year and lifelong risk. The third model, Life-Preserved, is dedicated to patients suffering from heart failure with preserved ejection fraction, a form of the disease in which the heart contracts normally but is too stiff to fill adequately with blood. Based on data from more than 20,000 patients and validated on more than 28,000 additional cases, it helps predict cardiovascular hospitalisations and mortality, providing physicians with support in treatment decisions.

The ultimate goal is to make these tools increasingly integrated into daily clinical practice, in order to identify the most vulnerable patients at an early stage and intervene with more targeted therapeutic strategies.