Global Longitudinal Strain as a Predictor of Risk in Patients with Heart Failure with Preserved Ejection Fraction: A Deserved Role?

This article refers to ‘Clinical implications of subclinical left ventricular dysfunction in heart failure with preserved ejection fraction: The PARAGON-HF study’ by M. Minamisawa et al., published in this issue on pages 871–881. Diagnosis of heart failure with preserved ejection fraction (HFpEF) is challenging as indicated by the various definitions utilized across studies. According to the European Society of Cardiology (ESC) guidelines, patients diagnosed with HFpEF are those who present symptoms and signs of heart failure (HF), along with evidence of structural and/or functional cardiac abnormalities and/or elevated natriuretic peptides (NPs), and with a left ventricular ejection fraction (LVEF) of ≥50%.1 This specific subtype, which differs in terms of its epidemiology, diagnosis, and treatment from HF with reduced ejection fraction (HFrEF), represents a unique challenge in clinical practice. Comparatively, HFpEF patients tend to be female, elderly, and commonly have associated comorbidities such as hypertension, diabetes, obesity, chronic renal failure, and atrial fibrillation, in contrast to those with HFrEF.2 Despite the increasing prevalence and poor prognosis associated with HFpEF, identifying patients who will experience unfavourable outcomes within this heterogeneous group remains exceedingly difficult. Beyond LVEF, a prompt and accurate diagnosis is therefore a key challenge for improving the prognosis of these patients. Some authors suggest that global longitudinal strain (GLS) could be a useful criterion for predicting the prognosis of patients. However, studies supporting this idea are often biased, of small sample size, and yield contradictory results.3-5 In this issue of the Journal, Minamisawa and co-authors assessed the prognostic value of GLS in a sample of HFpEF patients derived from the specific echocardiography sub-study of PARAGON-HF.6 The specific echocardiography sub-study of PARAGON-HF included 790 patients with adequate imaging for GLS who were thus included in this study.7 The outcomes assessed were the PARAGON-HF primary composite outcome.8 This editorial serves us as an opportunity to provide additional insights regarding this publication and the challenges represented by HFpEF. Regarding the results, among the 790 patients included in the study, the mean absolute GLS was 16.1%, and 47% of the population had evidence of abnormal GLS (absolute GLS <16%). A total of 604 patients (77%) had normal LVEF and of these, 37% had abnormal GLS. Worse GLS was associated with reduced LVEF, impaired right ventricular function, greater left ventricular size, wall thickness, mass and relative wall thickness. GLS was significantly associated with NP levels such that lower values of GLS were associated with higher levels of NPs. Lower GLS, even after multivariable adjustment, was associated with a higher risk of total hospitalization for HF or cardiovascular death and total HF hospitalization alone. There were no significant associations between GLS and biomarkers of extracellular matrix homeostasis including N-terminal propeptide of collagen I and III, tissue inhibitor of matrix metalloproteinase, carboxyl-terminal telopeptide of collagen type I, and soluble ST2. While LVEF is recognized as a crucial prognostic factor in HF patients, it is important to acknowledge its limitations which include for example its tendency to detect systolic dysfunction at a relatively advanced stage and its imperfect reproducibility.9 Although numerous therapies have improved the prognosis of patients with HFrEF, these same treatments have not demonstrated efficacy in the HFpEF population.1 This highlights the limitations in relying only on LVEF for prognostic evaluation of patients with HFpEF. Minamisawa's study convincingly demonstrates the utility of GLS in risk stratification for individuals with HFpEF. It is particularly useful given that these results are consistent with other studies, however, with smaller sample sizes and several bias, making their applicability in clinical practice more challenging. GLS serves as a predictive indicator for adverse outcomes such as HF hospitalization, cardiovascular mortality, and demonstrates correlation with NP levels. Although a suggested cut-off point of 16% in absolute values from existing literature10 was not confirmed through multivariable analysis in this study, it remains possible to identify a subgroup of individuals at higher risk of unfavourable progression. The use of vendor-independent software to assess strain in this analysis and the calculation of GLS performed only in the apical four-chamber and two-chamber views contrary to recommendations could limit these results. A vendor-independent software indeed limits the inter-vendor variability of measurements and we congratulate the authors for performing the strain as such. Finally, the inclusion criteria, particularly with a LVEF >45%, do not allow for generalization to the entire population of HFpEF, especially considering that even the ESC definition requires a LVEF >50%. Given the established relationship between low LVEF and impaired GLS, the inclusion of patients with LVEF <50% holds significant clinical relevance. While closely related from a prognosis perspective, HFpEF and HF with mildly reduced ejection fraction differ in several aspects, particularly in terms of epidemiology and therapeutic approaches. Considering them as a single entity would overlook the heterogeneity of these diseases.11 The lack of correlation observed in this study between markers of cardiac fibrosis and HF underscores the complexity of HFpEF pathophysiology. The pathophysiology of HFpEF is highly complex, involving diastolic left ventricular dysfunction, systolic dysfunction, ventricular dyssynchrony, chronotropic incompetence, atrial dysfunction and fibrillation, cardiovascular reserve dysfunction, right ventricular dysfunction and pulmonary hypertension.12 We now understand that HFpEF is not solely a heart condition but rather a systemic process leading to left ventricular hypertrophy, remodelling, fibrosis, and stiffness. The aetiology of this pathophysiology remains unclear, multifactorial and related to systemic inflammation leading to endothelial dysfunction, microvascular dysfunction and general oxidative stress.2 Microvascular dysfunction contributes significantly to structural and functional alterations.13, 14 We can no longer consider HFpEF without taking into account the underlying microvascular dysfunction. Numerous studies have shown that HFpEF should not be regarded as a single entity but rather as a heterogeneous spectrum comprising various epidemiological, clinical, biomarker, and prognostic profiles within the population. This recognition emphasizes the necessity for prognostic scores integrating a comprehensive array of parameters that evaluate these multifaceted elements to effectively identify high-risk patients. GLS should be viewed as one component within a broader array of clinical, biological, electrocardiographic, and multimodal imaging parameters. By integrating GLS alongside other relevant clinical data, a more nuanced understanding of HFpEF pathophysiology and prognosis can be achieved, facilitating tailored management strategies and improved patient outcomes (Figure 1). Several researchers have used machine learning techniques to identify phenotypes to help them to identify patients with poor prognosis.15 In conclusion, the study conducted by Minamisawa and colleagues provides further evidence supporting the significance of GLS in risk assessment for individuals with HFpEF. Future studies focusing exclusively on patients with LVEF >50% and integrating GLS with additional clinical, biological, genetic and imaging parameters are mandatory for enhancing prognostic accuracy in this heterogeneous population. Conflict of interest: none declared.