“Our finding raises awareness about the importance of implementing and not withdrawing medical treatments in daily clinical practice even if patients with heart failure experience an improvement in symptoms and pump function,” says the study’s last author Gianluigi Savarese, docent and senior lecturer at the Department of Clinical Science and Education, Södersjukhuset, Karolinska Institutet, and senior cardiologist at Södersjukhuset. 

Using data from the Swedish Heart Failure Registry, RiksSvikt, the researchers have analysed more than 8,700 patients with heart failure whose heart pump function, the ejection fraction, was initially impaired (below 40 per cent) but later improved to 40 per cent or more.  

By linking RiksSvikt with other national registers, the researchers from Karolinska Institutet and Linköping University, Sweden, and the University of Naples, Italy, among others, were able to track hospitalisations and deaths in patients who stopped or continued various heart failure treatments. 

Increased morbidity and mortality 

The observational study shows that patients who stopped taking medication of the type RASi (renin-angiotensin system inhibitors), ARNi (angiotensin receptor-neprilysin inhibitors) or MRA (mineralocorticoid receptor antagonists) had between 36 and 38 per cent higher risk of heart-related death or hospitalisation within one year of medication discontinuation.  

However, stopping beta-blocker medication was only associated with a higher risk in patients whose heart function had only moderately improved. 

“Our results show that heart failure medications continue to provide important benefits even when heart function has improved,” says Gianluigi Savarese. “This supports the current recommendation to continue with RASi/ARNi and MRA treatment, but also opens up the possibility of reconsidering whether beta-blockers can be discontinued in certain patients whose heart function has recovered well.” 

More tailored treatment strategies 

Since it was an observational study, no firm conclusions can be drawn about causality. It cannot be ruled out that the results were influenced by residual confounding. The researchers are now planning further studies to confirm the results.