The Left Atrium: A Window Into Cardiopulmonary Haemodynamics in Heart Failure with Preserved Ejection Fraction

This article refers to ‘Left atrial enlargement is associated with pulmonary vascular disease in heart failure with preserved ejection fraction’ by E.K. Gard et al., published in this issue on pages 806–814. Left atrial (LA) myopathy, defined as ‘any complex of structural, architectural, contractile or electrophysiological changes affecting the atria with the potential to produce clinically-relevant manifestations’ is present in certain individuals with heart failure with preserved ejection fraction (HFpEF).1 Indeed, in patients with HFpEF, myocardial remodelling may occur in the left atrium independent of the left ventricle, and is associated with a unique proteomic signature reflective of extracellular matrix remodelling and immune dysregulation.2, 3 The pathophysiology of LA myopathy in HFpEF is complex, with mechanical, biochemical, and electrical mechanisms all contributing. Specifically, presence, compared to absence, of HFpEF is associated with higher degrees of LA fibrosis, abnormal ion channel expression,4 impaired contractile function,5 and increased markers of inflammation including tumour necrosis factor-α, NADPH oxidase, and interleukin-6.6, 7 In addition, patients with HFpEF have relatively high LA pressure that may be due to left ventricular (LV) diastolic dysfunction, intrinsic LA stiffness, or ‘atrial’ mitral regurgitation.2 Increased LA pressure is subsequently associated with higher wall stress, further LA fibrosis, and increases in LA volume (LAV).2 In this issue of the Journal, Gard and colleagues investigated the relationship between indexed LAV (LAVI) and invasive haemodynamics at rest and exertion in a selection of patients from the REDUCE LAP-HF trial and an internal cohort of patients with exertional dyspnoea.8 The primary findings of the study are as follows: (i) higher LAVI was associated with blunted increase in cardiac output during exercise, higher mean pulmonary arterial (PA) pressure, and higher pulmonary vascular resistance (PVR); (ii) there was no association between LAVI and resting pulmonary capillary wedge pressure (PCWP) after adjustment for multiple comparisons; and (iii) higher LAVI was associated with decreased LA strain. The authors therefore conclude that in patients with HFpEF, presence of increased LAV is associated with a more advanced HFpEF phenotype as defined by PA haemodynamics, which may occur independent of increases in LA pressure. This study adds to the literature by suggesting an alternate mechanism of LA remodelling in HFpEF, independent of pressure overload. However, the reader should be mindful of important limitations when interpreting the results of this retrospective study. First, multivariate adjustment for potential confounders, namely atrial fibrillation (AF), was not performed to evaluate for potentially independent associations of LAVI and haemodynamic measures. It is well known that AF and HFpEF commonly coexist, and the presence of AF in HFpEF is associated with poor prognosis.9 While the authors performed a subgroup analysis to exclude patients who were in AF during the time of the studies obtained, there was no specific adjustment for history of AF. Indeed, even when a patient with HFpEF has paroxysmal AF and is in sinus rhythm during the time of an echocardiogram, a history of AF can significantly impact LA measurements and haemodynamics.10 This is especially relevant in the present study, as 33% of patients in the current study had a history of AF. Second, it is important to recognize that using LAV at end-systole to calculate LAVI may have impacted results. LAV can be measured at end-systole (LAVmax) or end-diastole (LAVmin). The left atrium enlarges during the course of LV systole as a result of longitudinal motion of the left ventricle from base to the apex of the heart. Thus, LAVmax is partially confounded by LV longitudinal systolic function. LAV derived at end-diastole (LAVmin) reflects LA volume independent of LV systolic function, and may more accurately reflect the effect of chronically elevated pressure (i.e. PCWP) upon the left atrium. Indeed, prior studies have suggested that LAVmin is a better predictor of HFpEF-associated morbidity than LAVmax.11 Given that HFpEF is associated with reduced LV longitudinal systolic function as measured by LV global longitudinal strain, LAVmax may be an imperfect measures of true LA remodelling.12 As LAVmax and LAVmin are therefore distinct, their associations with measures such as PCWP may not be identical. In addition, because only a minority of patients included in analysis had data on LA strain, the conclusions drawn from the strain analyses may not be comprehensive. Prior studies have shown that LA strain is a sensitive indicator of LA myopathy in HFpEF that may detect subtle changes in myocardial function prior to observation of anatomic changes (including increased LAV).13 Due to these strong associations, future studies with larger samples are needed to draw conclusions regarding associations between LA strain and invasive haemodynamics in HFpEF. These associations would be particularly interesting to investigate early in the HFpEF disease course, before macroscopic LA remodelling occurs. Importantly, the absent association between LA volume and PCWP in the current study must be interpreted carefully. First, LAV was evaluated as a categorical variable as opposed to a continuous variable, which may decrease power to detect associations. Second, HFpEF is a disease of exertional intolerance, and many patients with HFpEF have normal PCWP at rest. Over time, patients with HFpEF may have increases in PCWP with exertion that is associated with remodelling of the left atrium and, ultimately, the pulmonary vasculature. While adjusted results in the present study demonstrate that higher LAVI was associated with increased PVR, but not PCWP, we contend that this does not necessarily mean that LA remodelling occurs due to reasons independent of LA pressure. Due to rises in LA pressure and pulmonary pressures with exertion, remodelling may subsequently occur in the pulmonary circulation leading to combined pre- and post-capillary pulmonary hypertension. This process may occur among individuals with normal resting PCWP. In the recently conducted REDUCE LAP-HF II trial, which investigated the utility of an inter-atrial shunt device on outcomes in patients with HFpEF, exercise-induced LA hypertension in the absence of resting LA hypertension was common in HFpEF.14 We therefore believe that the association between LAVI and increased PVR at rest in the current study may be attributed to exercise-induced increases in LA pressure that have an impact on LA structure and function, and, ultimately, PA pressure, over time. Despite these limitations, the authors should be congratulated for completing a retrospective study that combines echocardiographic and haemodynamic data to further understand our knowledge of HFpEF phenotypes. As the authors suggest, integration of echocardiographic, clinical, and haemodynamic data has the potential to inform future treatment strategies for patients with HFpEF. In particular, because patients with HFpEF have both increased LAV and PA pressures, it would be reasonable to assume that decreasing LA pressure could lead to improvement in overall haemodynamics and symptoms. However, the recently published REDUCE LAP-HF II trial demonstrated no difference in cardiovascular death, stroke, or heart failure events between intervention (inter-atrial shunt device) and sham control groups.13 Subgroup analysis, however, revealed that placement of the shunt was associated with improvement in outcomes in a cohort of patients without latent pulmonary vascular disease (PVD). Latent PVD can only be detected upon invasive haemodynamic measurement with exercise, which can be cumbersome. Thus, there is a need to identify patients who may ultimately benefit from such device therapies for HFpEF using non-invasive measures. While the current study cannot provide this detail of information, this investigation suggests that patients with very large left atria may have significant pulmonary vascular remodelling that may lead to elevation in PVR and less benefit from an inter-atrial shunt device. Recent data from the REDUCE LAP-HF II cohort, however, suggest that there are several patients who may have LA remodelling without profound elevation in resting or exercise PVR. It is therefore important to evaluate the ability of other non-invasive modalities, such as LA reservoir strain imaging on echocardiography, cardiac magnetic resonance imaging, or circulating proteomic or metabolomic biomarkers, that may identify individuals with LA myopathy but without latent PVD. Such detailed characterization of the left atrium in HFpEF offers promise to identify populations within the overarching HFpEF syndrome that may derive benefit from targeted therapies (Figure 1).15, 16 In aggregate, these results suggest that patients with impaired LA strain, who do not yet have increased LAV, may represent a vulnerable cohort who may derive particular benefit from targeted HFpEF interventions. As the heart failure community continues to investigate phenotypes and treatments for HFpEF, continuing to integrate echocardiographic and haemodynamic data, as Gard and colleagues have done, will likely produce informative results. R.B.P. is supported by KL2TR001424 from the National Center for Advancing Translational Sciences, U.S. National Institutes of Health (NIH). Conflict of interest: G.P.: none declared. R.B.P. received consultant fees from International Consulting Associates, Inc.