Diagnostic Accuracy of Tissue Doppler Index E/e' for Evaluating Left Ventricular Filling Pressure and Diastolic Dysfunction/Heart Failure With Preserved Ejection Fraction: A Systematic Review and Meta-Analysis

Abstract

Background: Tissue Doppler index E/e' is used clinically and in multidisciplinary research for estimation of left ventricular filling pressure (LVFP) and diastolic dysfunction (DD)/heart failure with preserved ejection fraction (HFpEF). Its diagnostic accuracy is not well studied.

Methods and results: From the PubMed, Scopus, Embase, and Cochrane databases, we identified 24 studies reporting E/e' and invasive LVFP in preserved EF (≥50%). In random-effects models, E/e' had poor to mediocre linear correlation with LVFP. Summary sensitivity and specificity (with 95% CIs) for the American Society of Echocardiography-recommended E/e' cutoffs (lateral, mean, and septal, respectively) to identify elevated LVFP was estimated by using hierarchical summary receiver operating characteristic analysis. Summary sensitivity was 30% (9-48%), 37% (13-61%), and 24% (6-46%), and summary specificity was 92% (82-100%), 91% (80-99%), and 98% (92-100%). Positive likelihood ratio (LR+) was <5 for lateral and mean E/e'. LR+ was slightly >10 for septal E/e' obtained from 4 studies (cumulative sample size <220). For excluding elevated LVFP, summary sensitivity for E/e' (lateral, mean, and septal, respectively) was 64% (38-86%), 36% (3-74%), and 50% (14-81%), while summary specificity was 73% (54-89%), 83% (49-100%), and 89% (66-100%). Because of data set limitations, meaningful inference for identifying HFpEF by using E/e' could not be drawn. With the use of quality assessment tool for diagnostic accuracy studies (Quality Assessment of Diagnostic Accuracy Studies questionnaire), we found substantial risks of bias and/or applicability.

Conclusions: There is insufficient evidence to support that E/e' can reliably estimate LVFP in preserved EF. The diagnostic accuracy of E/e' to identify/exclude elevated LVFP and DD/HFpEF is limited and requires further validation in a well-designed prospective clinical trial.

Keywords: E/e'; diagnostic accuracy; diastolic dysfunction; heart failure with preserved ejection fraction; left ventricular filling pressure; tissue Doppler imaging.

© 2016 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley Blackwell.

Figures

Figure 1

Summary of the literature search. Primary analysis studies include data for patients with LVEF ≥50%. Supplemental analysis studies include data either for patients with LVEF ≥40% and/or with preselected echocardiographic indices (eg, the ratio of the early (E) to late (A) ventricular filling velocities (E/A) <0.9). DD/HFpEF represents invasively proven DD/HFpEF (i.e. clinical diagnosis of DD/HFpEF based on clinical sign and symptoms with evidence of elevated LVFP or impaired LV relaxation/chamber stiffness with or without additional biochemical markers and/or other ancillary tests). DD indicates diastolic dysfunction; HF, heart failure; LVEF, left ventricular ejection fraction; LVFP, left ventricular filling pressure; pEF, preserved ejection fraction.

Figure 2

Summary of quality assessment analysis (Quality Assessment of Diagnostic Accuracy Studies [QUADAS 2]). A, QUADAS 2 bar charts for primary analysis studies (n=24). B, QUADAS 2 bar charts for all selected studies (n=39).

Figure 3

Diagnostic accuracy of E/è recommended by the American Society of Echocardiography (ASE) to identify elevated left ventricular filling pressure (LVFP). A through C, Analysis for E/èlateral (>12). A, Individual studies (reference number as listed in the main text is indicated in brackets) with corresponding LVFP measurements, sample size, elevated LVFP prevalence, diagnostic 2×2 data (true positive [TP], false positive [FP], false negative [FN], true negative [TN]), and corresponding values of sensitivity (Sens.) and specificity (Spec.) with 95% CI are described. Summary heterogeneity is described by I2 statistic. B, The Rutter and Gatsonis11 hierarchical summary receiver operating characteristic (HSROC) analysis for recommended E/è cutoff to identify elevated LVFP is depicted. Summary sensitivity, summary specificity with 95% CI, and corresponding positive likelihood ratio (LR+) are depicted. C, Positive predictive value (PPV)–prevalence relationship for E/è to identify elevated LVFP using summary sensitivity and specificity HSROC points. Prevalence of elevated LVFP corresponding to 50% PPV for E/è is highlighted. *TP, FP, FN, TN values were extracted from the graphical data representation of LVFP vs E/è in study results; for such study, column presenting patient number (N) include 2 numbers: first number is actual counted patients in the plot, and second number is total patients in the study group. D through F, Analysis for E/èmean (>13). G through I, Analysis for E/èseptal (>15). Other description is same as for A through C.

Figure 4

Diagnostic accuracy of E/è recommended by the American Society of Echocardiography to identify normal left ventricular filling pressure (LVFP). A through C, Analysis for E/èlateral (<8). A, Individual studies (reference number as listed in the main text is indicated in brackets) with corresponding LVFP measurements, sample size, normal LVFP prevalence, diagnostic 2×2 data, and corresponding values of sensitivity and specificity with 95% CIs are described. Summary heterogeneity is described by I2 statistic. B, Hierarchical summary receiver operating characteristic (HSROC) analysis for recommended E/è cutoff to identify normal LVFP is depicted. Summary sensitivity, summary specificity with 95% CI, and corresponding positive likelihood ratio (LR+) are depicted. C, Positive predictive value (PPV)–prevalence relationship for E/è to identify normal LVFP using summary sensitivity and specificity HSROC points. Prevalence of normal LVFP corresponding to 50% PPV for E/è is highlighted. *Same as in Figure 3. D through F, Analysis for E/èmean (<8). G through I, Analysis for E/èseptal (<8). Other description is same as for A through C.

Figure 5

Estimates for use of American Society of Echocardiography–recommended E/èseptal cutoffs in patient group with varying prevalences of elevated left ventricular filling pressure (LVFP). A through C, Summary outline of application of E/èseptal for evaluating elevated and normal LVFP in representative examples with disease prevalence of elevated LVFP set at 10%, 50%, and 90% for n=100. More than half of the patients are in the indeterminate zone regardless of the disease prevalence. In the low‐prevalence (A, 10%) scenario, abnormal E/è is noted in a few patients and can be misleading in a substantial number of these patients. Normal E/è value is suggestive of normal LVFP in the majority of patients. For intermediate prevalence (B, 50%), abnormal E/è is suggestive of elevated LVFP, while normal E/è value is suggestive of normal LVFP in the majority of patients. In the high‐prevalence (C, 90%) scenario, even more patients are found in the indeterminate zone. Abnormal E/è is suggestive of elevated LVFP in most instances. Normal E/è is misleading in most cases here and cannot be used to rule out elevated LVFP. D and E, Estimated 2×2 distributions of patients after application of E/èseptal cutoffs with hierarchical summary receiver operating characteristic–calculated summary sensitivity and specificity for evaluating elevated (D) or normal (E) LVFP (Figures 3 and 4).