Diastolic function assessment based on a semi-automated computing of strain–volume loops

This commentary refers to ‘New expectations for diastolic function assessment based on a semi-automated computing of strain volume loops’ by A. Hubert et al. pp. 1366--71. Accurate assessment of diastolic function among heart failure patients with preserved ejection fraction (HFpEF), remains one of the most demanding fields of echocardiography.1 Therefore, exploration of new diastolic indices, such as strain–volume loops (SVLs), recently discussed in the paper by Hubert et al.,2 sounds quite promising. However, despite the authors’ fair intentions, the results presented in their article do not comply with its initial aims. It is evidence-based that validation of newly introduced parameters requires two important conditions: firstly, comparison of the ‘new-comers’ against ‘ground truth’ indices, and secondly, enrolment of an appropriate population with a unique characteristic to be detected by the suggested index. In the case of Hubert’s study, neither of these two prerequisites is met. Isolated echocardiographic parameters, such as E/A or E/E′, chosen by authors to compare the diagnostic utility of SVLs, are far from being considered ‘golden standards’ in the estimation of filling pressures, especially in the HFpEF population recruited.1 Additionally, comparison of receiver operator characteristic curves is not intended to highlight the best diastolic parameter—as it was stated in the aims of the study (that would happen e.g. if wedge pressure was selected as a classifier)—but rather highlights the best discriminator of one group from another (e.g. amyloidosis vs. HFpEF). The authors’ notion that specific groups reflect certain diastolic dysfunction stages, and therefore could be used as classification models, may be considered an oversimplification. This is because many other confounding variables (such as wall thickness magnitude or underlying histology in HFpEF group; parameters never reported in the study) may enter the formula,3 distorting the conclusions reached. Application (but not citation) of ejection fraction global longitudinal strain ratio (EFSR) further supports our argument that the examined variables, including SVLs, are tested as HFpEF discriminators rather than diastolic function parameters. EFSR, introduced by our group, is solely a systolic phase indicator of the dissociation between ejection fraction and subendocardially originated longitudinal strain, caused by the gradient of amyloid infiltration in the myocardium.4 Finally, it is quite unfortunate that only the loop area is examined in this study and not other parameters of SVLs (Figure 1).5 A typical example of a strain–volume loop with its components in a patient with amyloidosis. EDV, end-diastolic volume; ESV, end-systolic volume. In conclusion, SVLs constitute a very interesting approach to the study of patients with HFpEF. Despite that, the methodological issues previously mentioned reduce the true impact of the study, necessitating careful planning of future projects so as to meet readers’ expectations. Conflict of interest: none declared.