Coronary Microvascular Dysfunction in Heart Failure with Preserved Ejection Fraction – Adding New Pieces to the Jigsaw Puzzle

This article refers to ‘Endothelium-dependent and independent coronary microvascular dysfunction in patients with heart failure with preserved ejection fraction’ by J.H. Yang et al., published in this issue on pages 432–441. According to recent epidemiologic data, approximately 50% of patients with heart failure (HF) have HF with preserved ejection fraction (HFpEF). Hypertension, obesity, and coronary artery disease (CAD) are risk factors for HFpEF and the prevalence of this condition increases with age. There is experimental and clinical evidence that coronary endothelial inflammation and imbalance in nitric oxide (NO) production/utilisation, induced by co-morbidities, are crucial in the pathophysiology of HFpEF.1, 2 These abnormalities are in turn responsible for coronary microvascular dysfunction (CMD) and ischaemia, cardiomyocyte stiffening and hypertrophy, and/or myofibroblast activation, leading to left ventricular (LV) loss of lusitropic function.3 Pre-clinical studies in rodents confirm the role of CMD in the pathogenesis of HFpEF4 and microvascular rarefaction was described in autoptic specimens from HFpEF patients.5 However, clinical evidence of CMD in HFpEF has, so far, mainly been indirect and based on retrospective studies using non-invasive imaging to exclude obstructive CAD. The recent PROMIS-HFpEF trial was the first prospective, multinational, multicentre, observational study, which enrolled patients with HFpEF without obstructive CAD. It demonstrated that the coronary flow reserve (CFR), which in the absence of obstructive CAD is an index of coronary microcirculatory function, was reduced in 75% of the patients.6 A CFR in the lower range, measured with transthoracic Doppler echocardiography, was associated with several indices of HFpEF including higher N-terminal pro-B-type natriuretic peptide, evidence of systemic endothelial dysfunction, lower tricuspid annular plane systolic excursion and impaired right ventricular free wall strain. The relationship between CFR, myocardial injury and diastolic dysfunction was further studied by Taqueti et al.7 by means of positron emission tomography. They found that patients with impaired CFR had worse diastolic function and that detectable levels of troponin were associated with diastolic dysfunction only in the presence of impaired CFR. Despite the HFpEF paradigm, which postulates coronary endothelial dysfunction as a pathogenetic mechanism underlying HFpEF, no study has yet specifically evaluated coronary endothelium-dependent microvascular function in HFpEF, whereas a number of studies have revealed that endothelium-independent CMD is often present in patients with HFpEF. In the current issue of the Journal, Yang et al.8 provide further insights into the mechanisms of CMD in patients with HFpEF. The study is unique in its kind since it is the first study assessing both coronary endothelium-dependent and -independent vascular function using gold standard invasive coronary physiology assessment in HFpEF patients. A cohort of 162 consecutive patients with HFpEF underwent invasive coronary physiologic testing and echocardiography to assess endothelium-dependent and -independent coronary microvascular function in parallel with LV diastolic relaxation and filling pressure. Overall, CMD was present in 117/162 patients (72%). In 47 patients (29%), there was evidence of endothelium-dependent CMD, in 53 (33%) CMD was endothelium-independent and combined in 17 cases (10%). Compared with patients with normal coronary microvascular function, those with evidence of CMD had decreased diastolic relaxation velocity and higher filling pressure. Furthermore, endothelium-independent microvascular dysfunction was associated with increased mortality. The results of this study confirm that CMD is highly prevalent in patients with HFpEF and shed some light on the mechanisms of CMD in these patients. In keeping with the HFpEF paradigm, which postulates that CMD, due to inflammation-induced impaired NO bioavailability, triggers HFpEF-related cardiac remodelling and dysfunction,3 endothelium-dependent microvascular dysfunction was present in 54% of HFpEF patients with CMD. Specifically, 40% patients had isolated endothelium-dependent CMD and 14% had evidence of both endothelium-dependent and -independent CMD. The percentage of HFpEF patients with isolated endothelium-dependent CMD might be a reflection of this selected study cohort. Hence, these findings need cautious interpretation. A key point of the study is that CMD, similarly to HFpEF, can be manifold and the existence of different CMD forms in HFpEF should be identified in order to unravel its role and weight in the pathophysiology of HFpEF. Accordingly, Dryer et al.9 recently demonstrated the presence of four distinct coronary physiology groups in HFpEF patients combining invasive measurements of CFR and index of microvascular resistance. Both structural and functional alterations of the coronary microcirculation can be at the origin of CMD that can be endothelial-dependent, -independent, or both. Endothelial-dependent CMD arises from an imbalance in endothelium-derived relaxing factors like NO end endothelium-derived vasoconstrictors like endothelin,10 whereas endothelial-independent CMD might be ascribed both to undue vasoconstriction (unbalanced increase of sympathetic drive and increased adenosine turnover) and to hypertrophic inward remodelling of coronary arterioles.11 The latter mechanisms might have caused both impaired endothelium-independent CMD and impaired LV diastolic function in approximately 60% of the studied CMD patients. In agreement with the findings of Obokata et al.,12 the authors suggest that endothelium-independent CMD can initiate a vicious circle of myocardial ischaemia and injury particularly during physical activity, which, in the long term, may lead to interstitial fibrosis affecting LV function. The finding that in obesity-related HFpEF, the relationships between CFR and LV relaxation velocities were the most robust, lends some support to this hypothesis. The microvascular rarefaction and consequently the inadequate perfusion of myocytes in obese patients, together with the greater cardiac metabolic demand, may facilitate the development of demand ischaemia and LV dysfunction. Taken all together, these detrimental factors could provide an explanation to the worse outcome observed in patients with endothelium-independent CMD. The findings of Yang et al.8 highlight the diversity of CMD in HFpEF, and the need to categorise endothelial-dependent and -independent CMD with specific non-invasive and invasive coronary physiology testing to tailor therapeutic regimens. The findings of Yang et al.8 may partly explain the disappointing results of NO donors and phosphodiesterase 5A inhibitors13, 14 and hint at novel therapeutic strategies for patients with HFpEF. Adenosine kinase (ADK), an enzyme, which metabolises adenosine to AMP, has been shown to be up-regulated in arterioles isolated from the right atrial appendages of HFpEF patients and in obese ZSF1 rats (an animal model of HFpEF).15 The same authors demonstrated that adenosine or ADK inhibitor ABT-702 restored arteriolar vasodilatation both in isolated rat and human arterioles and improved conducted vasodilatation. Further mechanistic studies are needed to corroborate this hypothesis. Nevertheless, it is tempting to believe that pharmacological inhibition of ADK, which amplifies a key metabolic vasoregulatory mechanism by endogenous adenosine production and couples metabolic demand with myocardial perfusion, could be more effective than nitrates and phosphodiesterase 5A inhibitors, which induce a non-selective coronary blood flow.15 In conclusion, our understanding about the role of CMD in HFpEF is still in its childhood and albeit clinical studies have confirmed its prevalence in HFpEF,6 deeper insights are still lacking to define accurately the causal role, pathophysiology, assessment and treatment of CMD in patients with HFpEF. Prospective clinical trials aimed at assessing both endothelial-dependent and -independent CMD and (both cardiac and all-cause) mortality are highly required. The findings of Yang et al.8 point out that beyond NO-related CMD and subsequent fibrosis and LV stiffness, other mechanisms may underlie CMD in HFpEF and aggravate LV diastolic dysfunction and blunting of myocardial flow reserve. Finally, given the relevance of myocardial ischaemia in HFpEF, the impact of an impaired regulation of vasomotor tone in CMD and HFpEF should not be neglected (Figure 1). Conflict of interest: none declared.