Therapeutic Effects of Sodium–Glucose Cotransporter 2 Inhibitors in Patients with Heart Failure with Preserved Ejection Fraction: From Outcome Improvement to Potentially Favourable Influences on Atrial Fibrillation Burden, Atrial Fibrillation Progression and Atrial Cardiomyopathy

This article refers to ‘Empagliflozin in heart failure with preserved ejection fraction with and without atrial fibrillation’ by G. Filippatos et al., published in this issue on pages 970–977. Atrial fibrillation (AF) and heart failure (HF) frequently coexist, but the link is particularly tight in HF with preserved ejection fraction (HFpEF). Taking into account the whole spectrum of HF patients, the prevalence of AF appears to be higher when HF is diagnosed in association with higher values of left ventricular ejection fraction (EF).1 Indeed, there is an important interdependence between HFpEF and AF, related to common pathophysiological pathways involving inflammation and fibrosis, leading to a frequent convergence of the two clinical manifestations.2 Sodium–glucose cotransporter 2 (SGLT2) inhibitors are a cornerstone of pharmacological treatment of both diabetes and HF,3, 4 and in consideration of the mutual interrelationships2, 3 there is growing interest in evaluating the clinical effects on the risk of AF and AF-related outcomes. In this issue of the Journal, Filippatos et al.5 present an interesting pre-defined secondary analysis of the EMPEROR-Preserved trial, comparing the effects of empagliflozin versus placebo on primary and secondary endpoints, as well as on safety outcomes, stratified according to the presence of AF at baseline, defined as AF reported in any electrocardiogram before empagliflozin initiation or in medical history. The occurrence of new-onset AF was also evaluated. Overall, 52% of the HF patients with EF >40% had AF, confirming the frequent association between AF and HFpEF/HF with mildly reduced EF.1 Patients with AF had worse New York Heart Association (NYHA) class, higher heart rate, higher N-terminal pro-B-type natriuretic peptide concentrations, and lower estimated glomerular filtration rate (eGFR) levels, as compared to patients without AF. In the placebo group, the primary endpoint of cardiovascular death or HF hospitalization occurred in 18.7% of the patients with baseline AF, as compared to 15.3% of patients without AF (hazard ratio [HR] 1.26, 95% confidence interval [CI] 1.04–1.51). Similarly, also the risk of HF hospitalizations was higher in patients with AF (HR 1.38, 95% CI 1.06–1.79). Of note, the risk of the primary endpoint or of HF hospitalizations did not differ by type of AF, comparing persistent/permanent versus paroxysmal AF. Treatment with empagliflozin reduced cardiovascular death or HF hospitalization as compared to placebo to a similar extent in patients with and without AF. Moreover, the beneficial effects on total HF hospitalizations and annual eGFR decline were similar in patients with and without AF. This is a very reassuring finding, since AF is frequently a marker of increasing clinical complexity and high burden of comorbidities6 and a lack of benefit of some effective drugs, specifically beta-blockers, was previously highlighted by pooled analysis of randomized controlled trials (RCTs) when patients were in AF rather than in sinus rhythm.7 The analysis of the EMPEROR-Preserved trial found no significant differences in the incidence of stroke between treatment arms in patients either with or without a history of AF. Stroke was not an adjudicated endpoint in the EMPEROR-Preserved trial and therefore this observation is far from being conclusive, even if it is noteworthy that a recent meta-analysis focused on SGLT2 inhibitors in the specific setting of cardiovascular prevention in diabetic patients found a neutral effect on the risk of stroke (for all types of strokes, as well as for fatal stroke, non-fatal stroke, ischaemic stroke or transient ischaemic attacks), even if it highlighted a marked reduction of haemorrhagic stroke.8 Further studies are needed to assess if the beneficial effects of SGLT2 inhibitors on HF hospitalizations and cardiovascular mortality in HFpEF are also associated with a positive impact on stroke, taking into account both AF-associated and non-AF-associated stroke. Atrial fibrillation is a progressive disease9, 10 and it is known that there is a convergence between AF and HFpEF,2 and that the risks of stroke and death increase when AF follows HF, in a temporal sequence.11 From this perspective, it will be important to assess if SGLT2 inhibitors may reduce the incidence of new-onset AF and the progression of AF to a higher burden, usually expressed by the transition to permanent AF. In the analysis presented by Filippatos et al.,5 no significant differences were found in the incidence of new-onset AF between the treatment arms, both in patients with or without diabetes, and this is an observation that deserves some comments. Indeed, in interpreting these results it is important to consider that new-onset AF was not a pre-defined endpoint in studies on SGLT2 inhibitors, and all the analyses on AF incidence were done post-hoc, evaluating trial data reporting AF as a potential adverse effect, thus with an inevitable risk of under-reporting. A meta-analysis based on RCTs performed for evaluating SGLT2 inhibitors both in diabetic and non-diabetic patients found a reduction in AF incidence with SGTL2 inhibitors versus placebo when administered to diabetic patients (odds ratio [OR] 0.80, 95% CI 0.65–0.99), but without significant effects when the drugs were administered to the whole group of HF patients (OR 0.70, 95% CI 0.43–1.13).12 The finding that AF was not reduced in the setting of HF trials (considering both HF with reduced EF and HFpEF) when comparing SGLT2 inhibitors versus controls is confirmed by another recent meta-analysis (OR 0.83, 95% CI 0.68–1.01).13 We think that there is a need for further evaluations on the effects of SGLT2 inhibitors on AF in HF, and specifically in HFpEF. Furthermore, all the data collected on AF in RCTs performed so far were related to clinical AF, therefore with an underestimation of the overall burden of AF, since the latter should also consider asymptomatic AF.9 Moreover, in the perspective of a comprehensive overview of the potential therapeutic effects of SGLT2 inhibitors on AF in HFpEF, it will be important to investigate the impact of these agents on progression of AF burden, and on related clinical events, also in terms of HF hospitalizations, stroke and death. This type of very accurate assessments can nowadays be performed through continuous rhythm monitoring by means of cardiac implantable electronic devices.14 Taking into account the convergence of AF and HFpEF, on the basis of common pathophysiological pathways involving inflammation,2 there is a need to analyse and characterize the extent of atrial cardiomyopathy, whose evolution may modulate the clinical course of AF and HF, with important implications for patient outcomes. Indeed, there is growing interest in the role of atrial cardiomyopathy not only with regard to AF and HFpEF clinical convergence but also with regard to the risk of stroke and clinical outcome, independently of AF.2, 9 The SGLT2 inhibitors may favourably interact with the pathophysiological pathways of the myopathic process characterizing HFpEF at the atrial and ventricular level15, 16 (Figure 1), with a potential for being beneficial also for AF, in terms of reduced incidence, slowed progression and facilitated response to rhythm-control interventions, but dedicated studies are required for clinical confirmation, translating basic science into clinical practice. In summary, the clinical data reported by Filippatos et al.5 are reassuring with regard to the possibility of a favourable clinical response to empagliflozin in HFpEF independently of AF, but open the way to more focused studies, specifically targeted to assess the effects of SGLT2 inhibitors on AF burden, AF progression, and atrial cardiomyopathy, also taking into account the different phenotypes that can be identified among patients affected by HFpEF, with important differences in the prevalence of AF and in the burden of associated comorbidities.17, 18 Conflict of interest: G.B. reports small speaker fees from Bayer, Boehringer Ingelheim, Boston, Daiichi Sankyo, Janssen, and Sanofi, outside of the submitted work. All other authors have nothing to disclose.