The ratio of early transmitral blood flow velocity over tissue Doppler early diastolic mitral annulus velocity (E/e′) was found unreliable for estimating pulmonary capillary wedge pressure (PCWP) in patients with decompensated systolic heart failure (HF). The objective of this study was to test its reliability in stable HF. Therefore, 130 consecutive patients with a left ventricular (LV) ejection fraction of <35% and stable HF underwent right-sided cardiac catheterization and transthoracic echocardiography with measurement of transmitral flow velocities (E, A) and mitral annulus velocities during systole (s′) and diastole (e′). Mean age was 56 ± 11 years and mean LV ejection fraction was 28 ± 8%; 48% had PCWP of >15 mm Hg. E/e′ septal correlated more strongly with PCWP (r = 0.53) compared with E/e′ lateral (r = 0.41) and E/e′ mean (r = 0.50; all p values <0.001). The area under the receiver operating characteristic curve (AUC) of E/e′ ratios for PCWP estimation was 0.79 (95% confidence interval [CI] 0.70 to 0.87) for E/e′ septal , 0.72 (95% CI 0.63 to 0.82) for E/e lateral , and 0.79 (95% CI 0.70 to 0.87) for E/e mean (all p values <0.0001). AUCs of E/e septal and E/e mean did not vary with s′ septal , QRS width, or resynchronization. Using a cutoff of 8, negative predictive value of E/e′ septal was 89% and negative likelihood ratio of 0.15. E/e′ lateral showed good diagnostic performance only in patients with s′ lateral of >4.5 cm/s (n = 77, 59%; AUC = 0.82; 95% CI 0.71 to 0.92; s′ lateral of ≤4.5 cm/s: AUC = 0.54; 95% CI 0.38 to 0.70; p = 0.005). In conclusion, e′ is useful for estimating LV filling pressure in stable severe systolic HF. E/e′ septal showed good diagnostic performance for detecting normal filling pressures.
Elevated left ventricular (LV) filling pressure (LVFP) is associated with mortality in patients with heart failure (HF). The reference standard for measuring LVFP is an invasive measurement of pulmonary capillary wedge pressure (PCWP). Echocardiographic estimation of LVFP might avoid invasive measurement. Parameters most strongly correlated to PCWP are the ratios of early transmitral blood flow velocity (E) over early diastolic velocity of the septal and lateral mitral annulus (e′). E/e′ cutoffs indicating elevated LVFP in current guidelines were determined from studies that included few or no patients with low LV ejection fraction (LVEF). Whether the E/e′ ratio is effective in detecting elevated LVFP in patients with low LVEF is unclear. In decompensated HF, Mullens et al found none of the 3 E/e′ ratios (E/e′ lateral , E/e′ septal , and E/e′ mean ) useful, whereas Nagueh et al found relatively preserved reliability of the E/e′ mean ratio in a small cohort of patients. Severe impairment of LV contractility is associated with decreased tissue Doppler imaging (TDI) velocities. We hypothesized that this contractility-related decrease might impair the E/e′ ratio independently of PCWP. The objective of the present study was, therefore, to evaluate the diagnostic performance of the 3 E/e′ ratios for detecting PCWP elevation in patients with stable severe systolic HF.
Methods
This cross-sectional study was approved by the Henri Mondor Assistance Publique-Hôpitaux de Paris Ethics Committee as part of a genetic study supported by a publicly funded French research agency (Délégation Régionale à la Recherche Clinique, www.drc.aphp.fr , registration number, P020902). Patients gave their written consent before study inclusion including for elective right-sided cardiac catheterization, which was a part of the genetic study (phosphodiesterase 5 gene polymorphism).
Consecutive patients with symptomatic HF and LVEF of ≤35% were prospectively screened on admission in the heart failure unit of the cardiology department of Mondor Hospital from 2005 to 2008. Patients were preselected if they were receiving appropriate medication and had been clinically stable for at least 1 month. In these patients, the reason for admission was routine programed follow-up for reevaluation and optimization of drug treatment. Patients with congenital heart defects, significant heart valve disease, or with severe pulmonary disease were excluded. Postcapillary pulmonary hypertension was not an exclusion criterion. We also excluded patients with coronary artery disease requiring revascularization.
Within a maximum of 48 hours after study inclusion, each patient underwent a clinical examination; measurements of brain natriuretic peptide, serum creatinine, and serum sodium levels; echocardiography; and right-sided cardiac catheterization. In most patients echocardiography and right-sided cardiac catheterization were performed within 24 hours. The medical treatment was left unchanged until completion of all investigations.
Right-sided cardiac catheterization was performed in the fasting state without sedation using a Swan-Ganz catheter (Edwards Lifesciences, Irvine, California). The procedure was performed in the morning to limit hemodynamic changes linked to fasting. All medications were given as usual on the day of the procedure. Systemic blood pressure, central venous pressure, pulmonary artery pressure, and PCWP were determined during an average of 5 end-expiratory cardiac cycles, at steady state, with the patient in the supine position. The 0-mm Hg reference level was the midaxillary line with the patient in the supine position. The pressure transducer (Merit Medical Systems, South Jordan, Utah) was carefully maintained at the same level throughout the procedure. PCWP was assessed in all patients in West’s zone 3, as verified by fluoroscopy. The cardiac index was determined using the thermodilution method, as the mean of the results obtained with 3 rapid injections of 10 mL of ice-cold saline solution. All hemodynamic measurements were performed by the same physician (P-FL), who was blinded to the echocardiographic data and PCWP. A cut-off value of ≥15 mm Hg was applied to define patients with significantly elevated PCWP.
Two-dimensional and Doppler (pulsed, continuous, and TDI) echocardiographies were performed by a single physician (TD) in all study patients using a VIVID 7 machine with a 3.5-MHz transducer (GE Vingmed, Horten, Norway). All measurements were performed at rest in the left lateral recumbent position. Data were stored using EchoPAC (GE Vingmed) and analyzed by an echocardiographer who was blinded to all clinical and outcome data. Two-dimensional, M-mode, and Doppler echocardiography measurements and quantifications were performed as recommended by the American Society of Echocardiography, as the mean of 3 to 5 end-expiratory cycles. LVEF was calculated according to the biplane method of discs (Simpson’s rule). Peak transmitral early diastolic velocity (E) was measured using pulse-wave Doppler echocardiography. Peak systolic and early diastolic velocities of the lateral (s′ lateral and e′ lateral ) and septal (s′ septal and e′ septal ) mitral annulus were measured using pulsed TDI and their mean was calculated (s′ mean and e′ mean ). The E/e′ lateral , E/e′ septal , and E/e′ mean ratios were calculated to estimate LVFP. Pulmonary artery systolic pressure (sPAP) was assessed by measuring the gradient between the right ventricle and the right atrium using the peak velocity (V max ) of tricuspid regurgitation ( sPAP = 4 × V max 2 + right atrial pressure ) . Right atrial pressure was estimated to be 10 mm Hg in all patients. The echocardiographer was blinded to the results of right-sided cardiac catheterization.
Patient characteristics were described as numbers (percentages) for qualitative variables, means (standard deviation) for normally distributed quantitative variables, and median (interquartile range) for non–normally distributed quantitative variables. Patients were classified into 2 groups according to whether PCWP was <15 mm Hg or ≥15 mm Hg. The 2 groups were compared using the chi-square test or Fisher’s exact test for qualitative variables and the nonparametric Wilcoxon test for quantitative variables. Correlations between E/e′ and PCWP were assessed using Spearman correlation coefficients. The diagnostic performance of the 3 E/e′ ratios for estimating PCWP was assessed by computing the area under the receiver operating characteristic curve (AUC), with the 95% confidence interval (95% CI). AUCs were compared using the Delong test for nonindependent samples. Sensitivity, specificity, positive predictive value (PPV), negative predictive value, positive likelihood ratio (LR + ), and negative likelihood ratio were calculated using previously reported cutoffs for E/e′ septal (8 to 15), E/e′ mean (8 to 13), and E/e′ lateral (8 to 12). Stratified analyses were done according to the cause of HF (ischemic, yes/no), cardiac resynchronization therapy (CRT; yes/no), QRS width (<120 ms/≥120 ms), and s′ velocities. The s′ velocity cutoffs were determined to obtain the best AUC for E/e′ when appropriate. Two-tailed p values <0.05 were considered significant. Analyses were performed using STATA 11.0 (StataCorp 2009, College Station, Texas).
Results
Of the 130 included patients, 111 (85.4%) had complete echocardiographic data. The first 19 patients did not have septal mitral annulus TDI recording. Table 1 lists the baseline demographic, clinical, and laboratory data in the overall population and in the groups with normal (<15 mm Hg) or increased (≥15 mm Hg) PCWP. No patient was receiving inotropic support. Patients with PCWP elevation had functional impairments with higher New York Heart Association classes, lower peak VO 2 , lower serum sodium level, and higher serum brain natriuretic peptide level, compared with the other patients ( Table 1 ).
| Variable | All Patients, n = 130 | PCWP (mm Hg) | p ∗ | |
|---|---|---|---|---|
| <15, n = 68 | ≥15, n = 62 | |||
| Age (yrs) | 56 (12) | 55 (10) | 57 (13) | 0.27 |
| Men | 106 (82) | 58 (85) | 48 (77) | 0.25 |
| Body mass index (kg/m²) | 26.5 (4.5) | 26.3 (4.7) | 26.7 (4.4) | 0.58 |
| Coronary heart disease | 53 (42) | 25 (38) | 28 (47) | 0.32 |
| New York Heart Association class | ||||
| II | 35 (27) | 26 (38) | 9 (14) | <0.001 |
| III | 75 (58) | 38 (56) | 37 (60) | |
| IV | 20 (15) | 4 (6) | 16 (26) | |
| Atrial fibrillation | 9 (7) | 3 (5) | 6 (10) | 0.31 |
| Hypertension | 30 (23) | 15 (22) | 15 (24) | 0.77 |
| Hyperlipidemia | 12 (9) | 4 (6) | 8 (13) | 0.17 |
| Diabetes mellitus | 27 (21) | 12 (18) | 15 (24) | 0.36 |
| Systolic blood pressure (mm Hg) | 106 (20) | 107 (18) | 105 (22) | 0.49 |
| Diastolic blood pressure (mm Hg) | 65 (13) | 66 (11) | 64 (15) | 0.38 |
| QRS width (ms) | 125 (32) | 124 (29) | 126 (35) | 0.83 |
| 6-minute walking distance (m) | 449 (98) | 454 (100) | 443 (98) | 0.65 |
| Peak VO 2 (ml/kg/min) | 16.1 (5.1) | 17.6 (4.8) | 14.3 (4.9) | 0.001 |
| Sodium (mmol/L), median (IQR) | 139 (137–140) | 139 (138–141) | 138 (136–140) | 0.03 |
| Creatinine (μmol/L), median (IQR) | 118 (98–137) | 107 (94–136) | 121 (104–140) | 0.08 |
| Brain natriuretic peptide (ng/L), median (IQR) | 344 (107–680) | 123 (68–357) | 577 (338–1,114) | <0.0001 |
| Treatment | ||||
| Angiotensin-converting enzyme inhibitor | 102 (80) | 55 (83) | 47 (77) | 0.37 |
| Angiotensin receptor blocker | 26 (21) | 13 (20) | 13 (22) | 0.74 |
| β blocker | 117 (92) | 64 (97) | 53 (87) | 0.05 |
| Aldosterone antagonist | 86 (68) | 48 (73) | 38 (62) | 0.21 |
| Loop diuretic | 108 (85) | 53 (80) | 55 (90) | 0.12 |
| CRT | 43 (37) | 17 (27) | 26 (49) | 0.01 |
∗ p Value by chi-square test, Fisher’s exact test, or Wilcoxon rank sum test, as appropriate.
Adequate hemodynamic measurements, mitral inflow, and mitral annular TDI signals were obtained in all 130 patients ( Table 2 ). The prevalence of PCWP elevation was 48%. Patients with elevated PCWP had significantly more severe LV and left atrial dilation, lower LVEF, and a more restrictive LV filling pattern as reflected by a higher E/A ratio (2.7 ± 1.5) and shorter E-wave deceleration time (138.6 ± 58 ms). Table 2 lists the TDI mitral annulus velocities (s′, e′, and a′) and the values of the 3 E/e′ ratios (E/e′ lateral , E/e′ septal , and E/e′ mean ). Lateral s′ was significantly higher than septal s′ (5.1 ± 2.1 vs 4.3 ± 1.7 cm/s; p <0.001). E′ velocities correlated with s′ values (lateral r = 0.47, p <0.001; septal r = 0.51, p <0.001). TDI velocities of the septal mitral annulus were lower than those of the lateral mitral annulus. Patients with higher PCWP values had lower TDI velocities compared with those with normal PCWP, resulting in higher E/e′ septal than E/e′ lateral values in patients with PCWP elevation compared with those with normal PCWP.
| Variable | All Patients, n = 130 | PCWP (mm Hg) | p ∗ | |
|---|---|---|---|---|
| <15, n = 68 | ≥15, n = 62 | |||
| Echocardiography | ||||
| LVEDD index (mm/m²) | 37 (5) | 36 (5) | 38 (5) | 0.006 |
| LVEF (%) | 28 (8) | 31 (8) | 24 (7) | <0.0001 |
| Left atrial surface (cm²) | 26 (9) | 23 (9) | 29 (8) | 0.0002 |
| E (cm/s) | 80 (28) | 67 (22) | 95 (26) | <0.0001 |
| A (cm/s) | 56 (27) | 64 (26) | 47 (25) | 0.0004 |
| E/A | 1.9 (1.4) | 1.3 (0.9) | 2.7 (1.5) | <0.0001 |
| E deceleration time (ms) | 162 (76) | 183 (84) | 139 (58) | 0.0007 |
| s lateral (cm/s) | 5.1 (2.1) | 5.8 (2.3) | 4.3 (1.5) | <0.0001 |
| s septal (cm/s) | 4.3 (1.7) | 4.8 (1.7) | 3.9 (1.6) | 0.003 |
| s mean (cm/s) | 4.8 (1.6) | 5.3 (1.7) | 4.2 (1.4) | 0.0001 |
| e′ lateral (cm/s) | 6.3 (3.0) | 6.6 (2.9) | 6.0 (3.0) | 0.23 |
| e′ septal (cm/s) | 5.1 (2.5) | 5.3 (2.6) | 4.7 (2.4) | 0.17 |
| e′ mean (cm/s) | 5.8 (2.3) | 6.1 (2.3) | 5.4 (2.3) | 0.08 |
| a′ lateral (cm/s) | 5.6 (2.6) | 6.4 (2.5) | 5.4 (2.3) | 0.0004 |
| a′ septal (cm/s) | 5.4 (2.5) | 6.5 (2.4) | 4.1 (2.0) | <0.0001 |
| E/e′ lateral | 15.5 (9.3) | 12.1 (6.5) | 19.2 (10.4) | <0.0001 |
| E/e′ septal | 18.5 (10.7) | 13.9 (7.1) | 24.0 (11.8) | <0.0001 |
| E/e′ mean | 15.9 (8.4) | 12.3 (5.5) | 20.1 (9.3) | <0.0001 |
| Right-sided cardiac catheterization | ||||
| PCWP (mm Hg) | 16.3 (8.6) | 9.9 (3.4) | 23.4 (6.7) | <0.0001 |
| Systolic PAP (mm Hg) | 39 (15) | 28 (7) | 51 (13) | <0.0001 |
| Diastolic PAP (mm Hg) | 18 (9) | 12 (5) | 25 (8) | <0.0001 |
| Mean PAP (mm Hg) | 26 (11) | 18 (5) | 35 (8) | <0.0001 |
| Right atrial pressure (mm Hg) | 9 (5) | 6 (3) | 12 (5) | <0.0001 |
| Cardiac index (L/min/m²) | 2.2 (0.6) | 2.3 (0.5) | 2.1 (0.6) | 0.06 |
∗ p Value of the chi-square test. Fisher’s exact test or Wilcoxon rank sum test as appropriate.
Spearman correlation coefficients for E/e′ and PCWP were 0.53 for E/e′ septal , 0.41 for E/e′ lateral , and 0.50 for E/e′ mean , with all p values <0.0001. AUCs of E/e′ septal , E/e′ lateral , and E/e′ mean were 0.79 (95% CI 0.70 to 0.87), 0.72 (95% CI 0.63 to 0.82), and 0.79 (95% CI 0.70 to 0.87), respectively ( Figure 1 ). Differences between septal and lateral AUCs were not significant (p = 0.16). For all 3 E/e′ ratios, stratification on origin of HF (ischemic, yes/no), CRT (yes/no), and QRS width (<120 ms/≥120 ms) did not improve the diagnostic performance of the E/e′ ratios for estimating PCWP (data not shown), and neither did stratification of E/e′ septal and E/e′ mean on s′ velocity. However, the diagnostic performance of E/e′ lateral was improved by stratification on s′ lateral velocity ( Figure 2 ). An s′ lateral velocity cutoff of 4.5 cm/s yielded the highest AUC (0.82; 95% CI 0.71 to 0.92; n = 77, 59%). Conversely, in patients with s′ lateral ≤4.5 cm/s, the AUC of E/e′ lateral was only 0.54 (95% CI 0.38 to 0.70). The difference between the 2 AUCs was statistically significant (p <0.01).
