The β3-Adrenoreceptor Agonist Mirabegron for the Treatment of Combined Pre- and Post-Capillary Pulmonary Hypertension: New Pathway or Dead End?

This article refers to ‘β3 adrenergic agonist treatment in chronic pulmonary hypertension associated with heart failure (SPHERE-HF): a double blind, placebo-controlled, randomized clinical trial’ by A. García-Álvarez et al., published in this issue on pages 373–385. In patients with heart failure (HF), pulmonary hypertension (PH) is common and a marker of an advanced disease stage with poor exercise tolerance and a serious prognosis. Independent of the underlying cardiac disease, PH is the result of the backward transmission of an elevated left atrial (LA) pressure into the pulmonary veins leading to an increase in the mean pulmonary artery wedge pressure (mPAWP) and thereby also the mean pulmonary artery pressure (mPAP). Initially, this is a purely passive phenomenon with a normal transpulmonary gradient and pulmonary vascular resistance (PVR), respectively (isolated post-capillary PH). In more advanced disease stages, pulmonary vascular remodelling with a rise in PVR may occur, and the mPAP is the result of both an elevated mPAWP and an elevated transpulmonary gradient (i.e. the product of PVR and cardiac output), a constellation referred to as combined pre- and post-capillary PH (CpcPH). Presence of CpcPH is particularly relevant for patients with HF with preserved ejection fraction (HFpEF), because in contrast to HF with reduced ejection fraction (HFrEF) there are still very limited options to treat HFpEF and thereby to lower mPAWP and potentially also PVR.1, 2 Several trials evaluating the impact of drugs specifically targeting the pulmonary vasculature (drugs approved for the treatment of pulmonary arterial hypertension) have been conducted in patients with HFpEF and PH with overall mixed results.3-7 On this background, researchers have made efforts to develop innovative treatments for CpcPH. In the present issue of the Journal, Garcia-Alvarez et al.8 report the results of the randomized controlled double-blind multicentre SPHERE-HF study evaluating the effects of the oral β3-adrenoreceptor (β3AR) agonist mirabegron in CpcPH patients. Previous pre-clinical research had revealed favourable effects of β3AR agonism on ventricular and vascular function in HF. In endothelial cells, β3AR stimulation leads to vasorelaxation by enhanced nitric oxide generation and downstream soluble guanylate cyclase signalling.9 Mirabegron was previously found to reduce PVR and to improve right ventricular (RV) function in a large animal model of post-capillary PH.10 Mirabegron has already been approved for use in humans for the treatment of overactive bladder. For the present study, CpcPH was defined based on the 2015 European Society of Cardiology/European Respiratory Society (ESC/ERS) definition [(1) mPAP ≥25 mmHg, (2) mPAWP ≥15 mmHg, and (3) PVR ≥3 WU and/or diastolic pressure gradient ≥7 mmHg]11 but the previous 2011 ESC/ERS definition was also allowed (transpulmonary gradient ≥12 mmHg). For the final analysis 66 patients were available. The majority of patients were elderly women with HFpEF and atrial fibrillation. The mean mPAP, mPAWP, and PVR were 42 mmHg, 22 mmHg, and 4.3 WU, respectively. Patients were treated with oral mirabegron 50–200 mg/day (median dose 150 mg/day) or placebo for 16 weeks. The trial was negative for the primary endpoint, that is the change in PVR. Mirabegron treatment had no clear effect on mPAP and mPAWP either. The secondary endpoints 6-min walking distance, N-terminal-pro-B-type natriuretic peptide and quality of life were not affected. However, the authors reported a positive result for one of the secondary endpoints: an increase in RV ejection fraction as assessed by cardiac magnetic resonance imaging (MRI) or computed tomography in the mirabegron compared to the placebo group.8 The authors are to be commended for conducting this important trial evaluating an innovative treatment approach using state-of-the-art techniques. However, how should we interpret these results? At the first sight the lack of effect on PVR is in contrast to a recent study among patients with advanced HFrEF (New York Heart Association class III/IV), where oral mirabegron treatment (300 mg/day) during 1 week led to a reduction in PVR.12 That study was very small (n = 20), and the data on changes in PVR in absolute terms are somewhat difficult to interpret because overall the baseline PVR was relatively low, and the mirabegron group had a numerically clearly higher baseline PVR than the placebo group (3.5 vs. 1.7 WU). Still, there was a signal for a reduction in PVR and an increase in stroke volume index in the mirabegron arm.12 There are relevant differences in PH pathophysiology between HFpEF and HFrEF including a higher LA stiffness for a given LA size and a higher PVR for a given mPAWP in HFpEF compared to HFrEF. Among patients with HFpEF, those with atrial fibrillation typically have higher mPAWP (despite similar left ventricular end-diastolic pressure [LVEDP]) than those without. Left ventricular (LV) remodelling in HFrEF is substantially different from HFpEF, and there are several drugs with proven impact on morbidity and mortality in HFrEF, and many of these lead to a reverse LV remodelling process. There is also some albeit limited evidence that miragebron may have favourable effects on LV structure and function in HFrEF.13 In contrast, there is only one class of drugs with proven clinical benefit in HFpEF, that is the inhibitors of the sodium–glucose cotransporter 2 (SGLT2i) but the effect of SGLT2i on LV structure in HFpEF is less clear. In the present study, there was a statistically significant reduction in LA size in the mirabegron group, which however, was not accompanied by any evidence of LV remodelling (mass, dimensions, tissue Doppler measurements). Unfortunately, LA dimensions by MRI were not reported. Thus, it seems that mirabegron treatment did not lead to a relevant LV–LA remodeling in this very difficult to treat population, which is consistent with the lack of effect on mPAWP.8 The authors pointed out a significant positive result for one important secondary endpoint, that is RV ejection fraction by MRI. Unfortunately, findings were not absolutely consistent. There was no reduction in mPAP and PVR, and there was no improvement of the tricuspid annular plane systolic excursion (TAPSE)/systolic pulmonary artery pressure (sPAP) ratio, an accepted non-invasive marker of RV–pulmonary artery coupling. Although cardiac MRI is a well established technique to quantify RV function in PH (also included in the risk stratification algorithm for pulmonary arterial hypertension in the 2022 ESC/ERS guidelines14), we would expect at least some correlation with echo measures. This was not the case however, although comprehensive echocardiography including RV strain was performed. When looking at the detailed MRI data (online supplementary Table S3) it becomes obvious that the increase in RV ejection fraction following mirabegron therapy was the consequence of an increase in RV end-diastolic volume while end-systolic volume was unchanged. This would correspond to an uncommon type of RV reverse remodelling. The results are disappointing but are they surprising? A critical reappraisal of an animal study10 underlying the design of SPHERE-HF reveals that the pig model of post-capillary PH (pulmonary vein banding) was not fully representative for CpcPH in humans, and that the results are difficult to interpret for the following reasons. This was a model of pulmonary vein stenosis or veno-occlusive disease rather than CpcPH because the LV and LA, and thus the LVEDP were normal. Importantly, mPAWP was not measured in this study, and the transpulmonary gradient was calculated as the difference between mPAP and LVEDP.10 However, the banding mimics a rise in LA pressure and mPAWP (a haemodynamic key component of CpcPH), and using the LVEDP as a substitute for mPAWP is likely to lead to an overestimation of PVR in this setting. The mPAP and LVEDP did not significantly change following oral mirabegron treatment for 14 days but the (overestimated) PVR was reduced because of an increase in cardiac output. Thus, we must consider the results of SPHERE-HF neutral. This does not exclude the possibility that β3AR stimulation may be a useful therapeutic approach in HFrEF. Still, SPHERE-HF is a very important trial because it contributes new data on the role of drugs specifically targeting pulmonary vascular disease in CpcPH in the context of HFpEF. Figure 1 summarizes the few studies in this field. In the study with the highest baseline PVR, the endothelin antagonist macitentan had no effect on haemodynamics but increased the risk of fluid retention.3 The SPHERE-HF population is that with the second highest PVR among these studies. The studies evaluating the phosphodiesterase-5 inhibitor sildenafil or the soluble guanylate cyclase stimulator riociguat included patients with lower PVR (Figure 1). The 2022 ESC/ERS guidelines advise against the use of pulmonary arterial hypertension-specific drugs in PH in the context of HF (class III recommendation) but they also state that an individualized treatment approach is needed in patients with very advanced CpcPH (PVR >5 WU).14 What is next? There is now an effective treatment for HFpEF, that is SGLT2i dapagliflozin and empagliflozin. The clinical value of these substances was not established yet when SPHERE-HF was designed. Interestingly, empagliflozin was recently found to reduce the diastolic pulmonary artery pressure in patients with HF (HFrEF and HFpEF) and overall mild PH with a previously implanted pulmonary artery pressure monitor, and this effect occurred independent of LV ejection fraction.15 Thus, an SGLT2i is indicated in every HFpEF patient and must belong to the background therapy of all HFpEF patients in future studies. SPHERE-HF was a negative (or actually neutral) trial in terms of results but a very positive example of a true CpcPH trial. Similar but larger trials evaluating innovative treatment strategies also looking at clinical endpoints will be required to move the field forward. Conflict of interest: none declared.