Multimorbidity in Heart Failure: Leveraging Cluster Analysis to Guide Tailored Treatment Strategies

Current Heart Failure Reports https://doi.org/10.1007/s11897-023-00626-w Multimorbidity in Heart Failure: Leveraging Cluster Analysis to Guide Tailored Treatment Strategies Mariëlle C. van de Veerdonk1,2 · Gianluigi Savarese3 · M. Louis Handoko2 · Joline W.J. Beulens4,5,6 · Folkert Asselbergs1,7 · Alicia Uijl1,3 Accepted: 16 August 2023 © The Author(s) 2023 Abstract Review Purpose This review summarises key findings on treatment effects within phenotypical clusters of patients with heart failure (HF), making a distinction between patients with preserved ejection fraction (HFpEF) and reduced ejection fraction (HFrEF). Findings Treatment response differed among clusters; ACE inhibitors were beneficial in all HFrEF phenotypes, while only some studies show similar beneficial prognostic effects in HFpEF patients. Beta-blockers had favourable effects in all HFrEF patients but not in HFpEF phenotypes and tended to worsen prognosis in older, cardiorenal patients. Mineralocorticoid receptor antagonists had more favourable prognostic effects in young, obese males and metabolic HFpEF patients. While a phenotype-guided approach is a promising solution for individualised treatment strategies, there are several aspects that still require improvements before such an approach could be implemented in clinical practice. Summary Stronger evidence from clinical trials and real-world data may assist in establishing a phenotype-guided treatment approach for patient with HF in the future. Keywords Heart failure · Machine learning · Clustering · Phenotyping · Precision medicine · Treatment response Introduction * Alicia Uijl 1 Department of Cardiology, Amsterdam University Medical Centers, Amsterdam Cardiovascular Sciences, University of Amsterdam, Meibergdreef 9, 1105, AZ, Amsterdam, The Netherlands 2 Department of Cardiology, Amsterdam University Medical Centers, Amsterdam Cardiovascular Sciences, Vrije Universiteit Amsterdam, Meibergdreef 9, 1105, AZ, Amsterdam, The Netherlands 3 Division of Cardiology, Department of Medicine, Karolinska Institutet, Stockholm, Sweden 4 Department of Epidemiology and Data Science, Amsterdam University Medical Centers, Vrije Universiteit Amsterdam, Amsterdam Public Health Institute, Amsterdam, The Netherlands 5 Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands 6 Amsterdam Public Health Research Institute, Amsterdam, The Netherlands 7 Health Data Research UK London, Institute for Health Informatics, University College London, London, UK Heart failure (HF) is a complex clinical syndrome that has been characterized as a global pandemic. According to the Global Burden of Diseases in 2017, there were an estimated 64.3 million prevalent HF patients worldwide [1]. Despite advances in evidence-based treatment of patients with HF, the disease is still paired with a substantial morbidity and mortality, with a 5-year mortality rate of 60% [2, 3]. In addition, 60% of patients are readmitted within 1 year after their initial diagnosis of HF, of which almost one-third have HF as primary cause of hospitalization [3, 4] The goals of medical therapies for HF are to reduce symptoms, improve quality of life, prevent recurrent hospitalisations for HF, halt or reverse deterioration of cardiac function, and improve survival [5]. Patients are treated with a “one-size-fits-all” approach in which all therapies are considered for all patients following the guidelines with a selection based on ejection fraction (EF) and comorbidities. To date, this approach has worked well in patients with HF and reduced EF (HFrEF; EF ≤ 40%). However, treatment implementation in daily clinical practice has been 13 Vol.:(0123456789) Current Heart Failure Reports suboptimal [6, 7]. It is suggested that there could be a benefit from personalisation of treatment sequencing for patients with HFrEF to accomplish more effective treatment [8•], which could potentially be achieved via patient phenotyping. The “one-size-fits-all” approach seems less fruitful in patients with HF and preserved ejection fraction (HFpEF; EF ≥ 50%), where to date, only sodium-glucose co-transporter 2 inhibitors (SGLT2i) have shown benefit [9, 10]. Perhaps, it is not the drugs that are ineffective, but rather it is the enormous heterogeneity of the patient population that predisposes the clinical trials to disappointing results in HFpEF [11, 12]. Patient phenotyping to personalise therapy has therefore frequently been suggested to disentangle the heterogeneity of the patient population. To personalise therapies, several studies have investigated cluster analyses to discover distinct subgroups of HF patients based on their characteristics. This has led to a proliferation of clustering studies, different phenogroups based on comorbidity profiles, and different hypotheses on the origin of these clusters. Thus far, there have been no implications for daily clinical practice and how patients are treated based on clustering studies. This review therefore summarises key findings on treatment effects within phenotypical clusters of HF patients, making a distinction between patients with HFpEF and HFrEF. In addition, future directions with regard to a “phenotype-guided” treatment approach will be discussed. Hypothesis for a Phenotype‑Guided Approach The current foundations for HFrEF treatment consist of modulation of the renin-angiotensin-aldosterone system and the sympathetic nervous system by angiotensin-converting enzyme (ACE)-inhibitors, angiotensin receptor neprilysin inhibitors (ARNI), beta-blockers, mineralocorticoid receptor antagonists (MRA), and SGLT2i [5]. All treatments have shown to improve symptoms and survival and reduce the number of hospitalisations [5]. SGLT2i have most recently been included as they have shown to improve cardiovascular mortality, reduce HF symptoms, and improve quality of life [13, 14]. The benefit of applying comprehensive combination therapy (including ARNI/MRA/beta-blocker/ SGLT2i) instead of single-agent or dual agent therapies of the most commonly used agents (i.e. ACE-inhibitors and/ or beta-blockers) has been demonstrated in meta-analyses [15, 16] and was suggested in analyses from three clinical trials (PARADIGM-HF, EMPHASIS-HF, and DAPA-HF) [13, 17, 18]. Despite the overwhelming evidence from clinical trials, real-world data suggest that the implementation in daily practice is falling behind. Patients do not meet target doses, 13 there is clinical inertia, or there are concerns with tolerability in those with impaired renal function, anaemia, atrial fibrillation (AF), lung and liver disease, or hyperkalaemia [6, 7]. Although the prevalence of comorbidities in HF clinical trials has increased over time, inclusion of patients with these comorbidities remain limited, complicating the application of evidence to individual patients [19]. Adjusting a priority or sequence in the available guideline directed medical th (...truncated)