Tissue Doppler Imaging as a Prognostic Marker for Cardiovascular Events in Heart Failure with Preserved Ejection Fraction and Atrial Fibrillation


Although atrial fibrillation (AF) is associated with adverse cardiovascular (CV) outcomes, the prognosis of heart failure (HF) with preserved ejection fraction (HFPEF) with AF is still uncertain. This study was designed to evaluate whether the occurrence of CV events in patients with HFPEF and AF could be predicted by tissue Doppler imaging (TDI) of mitral annular velocity.


Clinical and echocardiographic data from January 2004 to December 2005 in patients with HFPEF and AF were investigated in this retrospective study. The development of CV events was defined as the composite of CV death, recurrent HF, and ischemic stroke.


Of 148 patients (mean age 68 ± 10 years; 64% were men), 35 clinical events, including 2 cardiac deaths, 22 recurrent HFs, and 11 ischemic strokes, were identified during a median follow-up of 27 months. In univariate analyses, age, ejection fraction, left atrial dimension, systolic (s’), and early diastolic (e’) mitral annular velocities were correlated with clinical events. Multivariate analyses revealed that old age, enlarged left atrial dimension, and lower s’ and e’ remained independent predictors of outcomes. Furthermore, patients with both s’ < 5 cm/s and e’ < 7 cm/s experienced more frequent clinical events (hazard ratio 12.2; 95% confidence interval, 1.62-92.5; P = .015).


s’ and e’, particularly their combination, seem to be useful predictors of CV events in patients with HFPEF with AF.

Atrial fibrillation (AF) is one of the most common causes of heart failure (HF), particularly with the deterioration of systolic function over the long term, and it is easily manifested among elderly patients with HF, irrespective of ischemic or nonischemic causes. AF poses an additional risk to low ejection fraction (EF) or cardiac output, leading to increased morbidity, and it is believed to be a surrogate marker of left ventricular (LV) dysfunction and poor prognosis in patients with systolic HF.

The subtle systolic dysfunction reported in previous studies through extensive workup of HF with preserved EF (HFPEF) seems to underlie the clinical course of HF. Thus, the systolic function of HFPEF with AF is also intuitively expected to have a downhill course. However, although HFPEF with AF is frequently encountered in clinical practice, its clinical course has not received much attention. Epidemiologic studies demonstrated that patients with HFPEF had poor prognosis or mortality, as in those with systolic HF, and AF was the important background factor related to diastolic dysfunction. However, few publications have focused on the outcome of HFPEF in patients with AF and the detailed echocardiographic findings in such patients. Furthermore, little is known about the added risk of AF in patients with HFPEF, whereas few studies have revealed conflicting results regarding the impact of AF.

Tissue Doppler imaging (TDI) recently has been used to assess systolic LV function, and mitral annular velocity measured with TDI has emerged as a useful parameter because it generates additional information to LV diastolic pressure and function. Accordingly, studies that have investigated HFPEF using TDI demonstrate the existence of systolic dysfunction even with normal EF.

Thus, we hypothesized that patients with HF and AF showing systolic dysfunction would have more frequent clinical events than those without systolic dysfunction, although they all have normal or preserved EF. We aimed to assess the relationship between clinical events and echocardiographic parameters in terms of TDI for the prediction of the occurrence of cardiac adverse events.

Materials and Methods

Study Population

This was a retrospective study of patients with HFPEF and AF that was conducted at a single cardiovascular center of tertiary university hospital from January 1, 2004, to December 31, 2005. Electronic medical records (EMRs) of all consecutive patients with HFPEF and AF aged more than 18 years who had been evaluated for dyspnea as their first manifestation of HF during an outpatient workup were reviewed. Criteria for study eligibility were a) the presence of signs or symptoms of clinically stable New York Heart Association class I, II, or III; b) no medical history of a diagnosis of HF or hospital admission for HF; and c) left ventricular ejection fraction (LVEF) > 50% and evidence of diastolic LV dysfunction by echocardiography, as proposed in the guidelines of the European Society of Cardiology. Criteria for study ineligibility were a) no prior detailed study of the cause of HF; b) significant valvular diseases (moderate or severe grade); and c) the presence of moderate or advanced lung, renal, or liver diseases. The institutional review board approved the study protocol. Informed consent form was waived because of the retrospective nature of this study.

Echocardiographic Examination

Echocardiographic data, which were obtained from the hospital computerized database, included the measurements of cardiac dimensions and LVEF, which had been determined according to the recommendations of the American Society of Echocardiography using a Vivid 7 ultrasound system (GE, Vingmed Ultrasound, Horten, Norway) equipped with TDI capabilities. LV end-diastolic and end-systolic dimensions, end-diastolic thickness of the interventricular septum and LV posterior wall, and left atrial (LA) dimension were assessed from parasternal long-axis views by 2-dimensional guided M-mode. LVEF was calculated from apical views, using the modified biplane Simpson’s method. LV diastolic filling patterns were assessed by the mitral inflow pulsed-wave Doppler velocity, and the following parameters were obtained: peak early diastolic transmitral velocity (E), its deceleration time, and the isovolumic relaxation time. Pulsed TDI was performed adjusting the controls to obtain the best-quality images, and mitral annular peak systolic (s’) and early diastolic (e’) velocities were also obtained with the sample volume positioned at the septal annulus on the apical 4-chamber view. Furthermore, the E/e’ ratio was derived from dividing the early mitral inflow velocity by the early diastolic mitral annular velocity. For the AF study, each of the conventional and TDI parameters were obtained over 5 consecutive beats and then averaged.

Clinical Outcomes

Adverse cardiac events were defined as the composite of cardiovascular death, rehospitalization for management of cardiac decompensation, and ischemic stroke, which were determined from hospital electronic medical record or by telephone contact. The diagnosis of ischemic stroke was made when patients presented with symptoms of stroke, and hemorrhage was excluded by relevant findings on computed tomography or magnetic resonance imaging. We also excluded patients with transient ischemic attack.


Results are presented as mean ± standard deviation for continuous variables and as percentage for categoric variables. Differences between the 2 groups were determined using the chi-square test or the unpaired t test. Cox proportional hazards regression analysis was used to explore the association between mitral annular velocities and the occurrence of cardiac events. Receiver operating characteristic (ROC) curves were used to evaluate the ability of echo parameters to predict the occurrence of cardiac events and to determine the optimal cutoff values for sensitivity and specificity. A P value < .05 was considered significant. All statistical analyses were performed using the SPSS 11.0 for Windows (SPSS Inc, Chicago, IL).


Among a total of 168 patients with HFPEF and AF whose EMRs were reviewed for the study, 20 were excluded because values of TDI for mitral annular velocities were not feasible as the result of technical problems (poor echo window with lower quality and higher noise velocity recordings) at baseline screening. Thus, the study population consisted of 148 patients who had complete data. During a median follow-up of 27 months, there were 35 clinical events (2 cardiac deaths, 22 recurrent, decompensated HFs, and 11 ischemic strokes). Although patients who experienced clinical events were significantly older, there were no significant differences in blood pressure, New York Heart Association class, diabetes, hypertension, and medical history between those who experienced clinical events and those who did not experience cardiac events ( Table 1 ).

Table 1

Baseline clinical characteristics of study patients

Variables Clinical events (-)
(N = 113)
Clinical events (+)
(N = 35)
P value
Age, yrs 67.4 ± 9.9 74.3 ± 10.2 <.001
Male, n (%) 76 (67.3) 18 (51.4) .109
Height, cm 163.5 ± 9.9 160.4 ± 9.6 .113
Weight, kg 64.4 ± 12.1 60.5 ± 12.4 .107
BSA, m 2 1.71 ± 0.20 1.63 ± 0.21 .059
Systolic BP, mm Hg 119.8 ± 16.3 114.8 ± 16.1 .116
Diastolic BP, mm Hg 72.4 ± 11.9 68.9 ± 11.6 .133
Heart rate, bpm 80.9 ± 18.8 82.6 ± 22.8 .656
NYHA class
I (%) 20 (17.7) 7 (20.0) .804
II (%) 84 (74.3) 24 (68.6) .519
III (%) 9 (8.0) 4 (11.4) .507
Diabetes, n (%) 20 (17.7) 8 (22.9) .471
Hypertension, n (%) 42 (37.2) 13 (37.1) 1.000
CAD, n (%) 4 (3.5) 3 (8.6) .357
Stroke, n (%) 28 (24.8) 9 (25.7) 1.000
Beta-blockers, n (%) 49 (44.5) 20 (52.6) .452
ACEi or ARBs, n (%) 47 (42.7) 14 (36.8) .571
Anticoagulants, n (%) 40 (36.4) 11 (28.9) .436
Antiplatelet agents, n (%) 49 (44.5) 22 (57.9) .189
Diuretics, n (%) 54 (47.8) 22 (62.9) .127

ACEi, Angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; BP, blood pressure; BSA, body surface area; CAD, coronary artery disease; NYHA, New York Heart Association.

Data are expressed as mean ± standard deviation or as number (percentage).

Echocardiographic analysis revealed a trend toward a greater LA dimension and a significantly lower EF in patients with events ( Table 2 ). However, the mean EF was sufficiently above the normal range (>55%) in the events group. In TDI, s’ and e’ were significantly lower in the events group ( P < . 001, respectively) ( Figure 1 ), whereas early mitral inflow was similar in both groups. Thus, the E/e’ ratio was significantly higher in the events group ( P = . 001). The area under the curve (AUC) in ROC analysis for the prediction of adverse cardiac events is shown in Figure 2 . The 2 greatest AUCs for clinical outcomes were s’ and e’ (AUC = 0.74 and 0.73, respectively), and the AUC of s’ approached that of e’, the E/e’ ratio, and EF when each of the AUC values was compared. Cutoff levels of all parameters for the prediction of clinical events were sought using ROC analysis ( Table 3 ). ROC analysis revealed that all parameters (s’, e’, E/e’ ratio and EF) except LA dimension had significantly discriminative ability (all P < .05), and an optimal cutoff value of each parameter could be obtained. The Kaplan–Meier plot of event-free survival during the follow-up time period revealed cardiac events differences according to risk stratification using the s’ and e’ cutoff values derived from the ROC analysis and the combination of the s’ and e’ levels ( Figure 3 ). Patients with both s’ < 5 cm/s and e’ < 7 cm/s had significantly more frequent clinical events (log-rank χ 2 = 16.65, P < . 001).

Table 2

Baseline echocardiographic characteristics of study patients

Variables Clinical events (-)
(N = 113)
Clinical events (+)
(N = 35)
P value
LVEDD, cm 5.00 ± 0.55 5.15 ± 0.63 .161
LVESD, cm 3.28 ± 0.54 3.45 ± 0.59 .118
LAD, cm 4.72 ± 0.80 5.02 ± 0.82 .055
PWT, cm 0.94 ± 0.15 0.95 ± 0.14 .706
IVST, cm 0.95 ± 0.16 0.92 ± 0.20 .385
EF, % 61.6 ± 7.26 58.1 ± 7.76 .016
PASP, mm Hg 33.4 ± 9.34 36.2 ± 13.0 .217
DT, ms 181.0 ± 51.0 197.9 ± 60.7 .121
IVRT, ms 98.1 ± 21.6 104.9 ± 24.6 .131
E, m/s 0.92 ± 0.24 0.97 ± 0.39 .359
s’, cm/s 5.77 ± 1.24 4.71 ± 1.17 <.001
e’, cm/s 7.70 ± 2.17 6.00 ± 2.06 <.001
E/e’ ratio 13.2 ± 6.12 18.1 ± 10.6 .001

E, Early diastolic mitral inflow; e’, peak early diastolic mitral annular velocity; DT, deceleration time; EF, ejection fraction; IVRT, isovolumic relaxation time; IVST, interventricular septal wall thickness; LAD, left atrial dimension; LVEDD, left ventricular end-diastolic dimension; LVESD, left ventricular end-systolic dimension; PASP, pulmonary artery peak systolic pressure; PWT, left ventricular posterior wall thickness; s’, peak systolic mitral annular velocity.

Data are expressed as mean ± standard deviation.

Figure 1

Measurement of tissue Doppler velocity at the septal mitral annulus in patients with HFPEF and AF. A patient with event ( right ) had lower peak systolic and early diastolic velocities than a patient without event ( left ).

Figure 2

ROC analysis for the prediction of adverse cardiac events. e’, Peak early diastolic mitral annular velocity; LAd, left atrial dimension; LVEF, left ventricular ejection fraction; s’, peak systolic mitral annular velocity.

Table 3

Predictive characteristics of echocardiographic variables for the occurrence of cardiovascular adverse events

Variables Cutoff value Sensitivity (%) Specificity (%) AUC (95% CI) P value AUC comparison,
P value
s’ (cm/s) 5.0 52.3 82.4 0.74 (0.639∼0.834) <.001
e’ (cm/s) 7.0 64.9 71.6 0.73 (0.630∼0.832) <.001 .919
E/e’ ratio 15.2 50.0 77.5 0.65 (0.539∼0.761) .008 .188
EF (%) 55.5 78.4 50.0 0.64 (0.527∼0.752) .014 .100
LAD (cm) 4.99 50.0 66.7 0.60 (0.493∼0.707) .079 .056

AUC, Area under the curve; e’, peak early diastolic mitral annular velocity; EF, ejection fraction; LAD, left atrial dimension; s’, peak systolic mitral annular velocity.

P values are displayed for the differences in AUC values compared with the AUC value for s’.

Figure 3

Kaplan–Meier plot of event-free survival during the follow-up period in study patients according to cutoff values of peak systolic (s’) and early diastolic (e’) mitral annular velocities.

In univariate analyses for the prediction of adverse cardiac events, in addition to confounders such as age and EF, TDI parameters (s’ and e’) and LA dimension showed a significant hazard ratio to clinical events ( Table 4 ). Moreover, multivariate regression analyses showed that LA dimension and the combined s’ and e’ still remained significant predictors after adjustment for other parameters, including age and EF, compared with those without events; patients with both s’ < 5 cm/s and e’ < 7 cm/s had a significantly high risk for clinical events (hazard ratio 12.22; 95% confidence interval [CI], 1.62-92.5; P = . 015).

Table 4

Univariate and multivariate analyses for prediction of the occurrence of cardiovascular adverse events

Univariate Multivariate
Variables HR 95% CI P value HR 95% CI P value
Age 1.07 1.03∼1.11 <.001 1.06 1.02∼1.10 .003
Diabetes 1.53 0.69∼3.38 .293 1.09 0.45∼2.65 .854
Hypertension 1.07 0.54∼2.12 .848 0.85 0.41∼1.75 .660
Ischemia 1.72 0.52∼5.63 .372 1.51 0.83∼6.11 .110
Stroke 1.00 0.47∼2.13 .993 1.54 0.66∼3.58 .315
LA dimension 1.57 1.04∼2.38 .034 1.74 1.09∼2.77 .021
Ejection fraction 0.95 0.90∼0.99 .023 0.98 0.94∼1.02 .299
s’ and e’ 2.93 1.67∼5.16 <.001
s’ ≥ 5 cm/s and e’ ≥ 7 cm/s 0.22 0.07∼0.71 .012
s’ < 5 cm/s or e’ < 7 cm/s 1.12 0.57∼2.20 .736 7.44 0.93∼59.79 .059
s’ < 5 cm/s and e’ < 7 cm/s 2.23 1.15∼4.34 .018 12.22 1.62∼92.50 .015

Only gold members can continue reading. Log In or Register to continue

Jun 16, 2018 | Posted by in CARDIOLOGY | Comments Off on Tissue Doppler Imaging as a Prognostic Marker for Cardiovascular Events in Heart Failure with Preserved Ejection Fraction and Atrial Fibrillation

Full access? Get Clinical Tree

Get Clinical Tree app for offline access