Cutting the Gordian knot of left ventricular diastolic dysfunction: Role of opportunistic screening models

Left ventricular diastolic dysfunction (LVDD) is the result of impaired left ventricular (LV) relaxation and increased LV chamber stiffness, which results in elevated filling pressures. ¹ LVDD is characterised by impaired functional parameters (LV relaxation, increased LV stiffness) and is frequently accompanied by structural abnormalities, such as increased interstitial collagen deposition and modified extracellular matrix proteins. ²

Early detection of LVDD is one of the most important approaches in preventing heart failure (HF). The body of evidence demonstrates that LVDD is a progressive abnormality leading to myocardial remodelling and finally to overt HF. Patients with symptomatic LVDD and a LV ejection fraction of 50% or greater are described as heart failure with preserved ejection fraction (HFpEF). ³

The prevalence of LVDD is approximately 30% in the general population, as estimated by echocardiography, ² and depends of the age and associated comorbidities. ⁴ Aging is strongly associated with LVDD, even following adjustment for various risk factors, including ischaemic heart diseases (IHD), female gender, type 2 diabetes mellitus (T2DM), obesity, arterial hypertension and LV hypertrophy.^5–7 In routine clinical practice asymptomatic individuals may frequently have multiple and interrelated risk factors associated with LVDD. Therefore, a timely recognition of individuals at highest risk of developing LVDD in the community is important to prevent progression to overt HFpEF. The optimal screening model is hard to define, but one of the basic requirements will be to use easily applicable clinical variables strongly associated with LVDD. Given the progressive nature of LVDD and a paucity of data regarding the relation with established risk factors, every effort is welcome.

In a previous issue of the journal, Gohar et al. ⁸ present a research paper on the sex-specific screening models for the early identification of patients at high risk of LVDD, with or without symptoms of HF. Following the exclusion of patients with LV systolic dysfunction/heart failure with reduced ejection fraction (HFrEF), they analysed 1371 elderly patients (≥60 years) from the four primary care HF screening studies (STRETCH, TREE, UHFO-COPD and UHFO-DM).^9–12 All included patients had HFpEF and echocardiographic evidence of LVDD. For the development of prediction models, Gohar et al. ⁸ selected 11 candidate predictors (age, history of IHD, atrial fibrillation (AF), hypertension, T2DM, angina pectoris, shortness of breath at least when walking at a normal pace (MRC breathlessness scale ≥3), ankle oedema, pulse pressure, body mass index (BMI) and the use of β-blocker therapy). These variables were entered into logistic regression analyses and risk scores were separately developed for men and women. Their predictive validity to differentiate LVDD correctly was analysed by C-statistics. ⁸ Men and women were analysed separately, because previous research has shown gender heterogeneity in LVDD predictors.^13–15 Finally, two prediction models were constructed for either sex. The first model considered clinical predictors and a high-risk category due to multimorbidity and polypharmacy. However, this model excluded N-terminal pro B-type natriuretic peptide (NT-proBNP). The second model comprised the same clinical predictors but included NT-proBNP greater than 125 pg/mL. Stratification into low, moderate and high-risk categories required higher score values for accurate prediction, compared with score values that included NT-proBNP. The internal/external validation principle was applied by combining individual patient data from multiple studies. In addition, after applying the male model in women and vice versa, it was evident that LVDD prediction was most accurately achieved by using age, β-blocker therapy and NT-proBNP, in both men and women. ⁸ The male model also included BMI, exercise intolerance, increased pulse pressure and a history of IHD in addition to age and β-blocker therapy.

The constructed screening models represent simple, easy applicable tools which can be used in everyday clinical practice. The inclusion of NT-proBNP, although not necessary, increased their predictive validity in both sexes. One of the important limitations is that screening models have not been truly externally validated. Furthermore, the prediction of HFpEF in women, based on only two clinical characteristics (age and beta-blocker therapy), may be oversimplified, even in multi-comorbid, elderly individuals. Some components (e.g. MRC breathlessness scale) are not routinely assessed. The definition of LVDD was not in accordance with the last update of the recommendation from the American Society of Echocardiography (ASE) and the European Association of Cardiovascular Imaging (EACVI), albeit this issue has probably not significantly influenced patient classification. Also, the degree of LVDD was not assessed.

The prevalence of HFpEF is rapidly rising.^16,17 It currently affects up to one in 10 elderly individuals and is becoming the most common phenotype of HF. ¹⁶ HFpEF confers high morbidity and mortality, which underscores the importance of early detection and timely intervention in patients at risk.^16,18 Similar to the present study by Gohar et al., ⁸ a prediction model for HFpEF (H₂FPEF score) based on easy-to-assess clinical and echocardiographic characteristics was recently introduced for the differential diagnosis of dyspneic patients. ¹⁹ However, unlike screening models developed by Gohar et al., ⁸ the prognostic validity of the H₂FPEF score was further rigorously assessed using invasive haemodynamic exercise testing. ¹⁹ The screening models described in the present study and the diagnostic H₂FPEF score ¹⁰ had three variables in common, namely increasing age, BMI and antihypertensive treatment (including beta-blocker use and polypharmacy), which are all strong predictors of LVDD. However, unlike the screening model of Gohar et al., ⁸ the H₂FPEF score did not include NT-proBNP. The reasons why NT-proBNP had no added value in the H₂FPEF score may be a considerable proportion (24%) of missing data, ¹⁹ and an inverse relationship between high BMI and NT-proBNP levels.^20,21 The present study included a mostly overweight population, which might not have affected the predictive value of NT-proBNP. Also, unlike the diagnostic H₂FPEF score, echocardiography was not a part of the models of Gohar et al., ⁸ because these models were developed to screen patients in the community who should be referred for further echocardiographic examination to confirm or exclude LVDD.

The use of echocardiography in diagnosing LVDD may be challenging. ² Differentiation between normal and abnormal findings can be hampered by an overlap between Doppler indices in healthy individuals and those with cardiovascular diseases or comorbidities (e.g. IHD, valve disease, atrial fibrillation, hypertension, T2DM). ¹ As healthy elderly people have greater LV stiffness and slower myocardial relaxation compared with younger individuals, Doppler indices need to be interpreted in accordance with age.^22,23 The latest recommendation by the ASE and EACVI has standardised and simplified the echocardiographic evaluation of LVDD in daily practice and recommended the use of age-appropriate cut-off values.^1,2 In addition, ‘exercise diastology’ can be considered in selected patients suspected of having LVDD, not detected by resting echocardiography.^1,3,24 To refine LVDD assessment further, left ventricle and atrial speckle tracking of mechanical deformation (strain and strain rate) and cardiac magnetic resonance have incremental diagnostic and prognostic value in HFpEF.^25–30 However, these sophisticated techniques are not commonly available and could not be a part of a screening tool.

The most important aspect of screening for LVDD is to enable a timely intervention which could prevent progression into HFpEF.^3,7,31 At present, available therapies cannot restore normal LV diastolic function. ³² However, in the randomised STOP-HF trial, greater use of renin–angiotensin–aldosterone system blockade in patients with LV systolic/diastolic dysfunction was associated with less overt HF at follow-up. ³¹ While several therapeutic options for HFpEF are under clinical assessment (e.g. sacubitril/valsartan, sodium-glucose cotransporter type II inhibitors), ³² the focus of clinical management for HFpEF should be on identification of at-risk patients and treatment of comorbidities frequently accompanying LVDD (e.g. atrial fibrillation, hypertension, T2DM).^33–35 Therefore, community-based opportunistic screening of patients at high risk of having LVDD, who do not currently receive medical therapy, could be beneficial and life-saving.

In conclusion, the study of Gohar et al. ⁸ provides practical screening models easily applicable in the community, which could fulfil the currently unmet requirement for a viable strategy of opportunistic screening for LVDD. These models could facilitate the implementation of effective pharmacological and lifestyle management to reduce the growing burden of HFpEF.