Abstract
Background
Several studies have demonstrated the importance of left ventricular (LV) global longitudinal strain (GLS) as a reliable prognostic indicator in patients with heart failure (HF). These studies have included few African American (AA) patients, despite the growing prevalence and severity of HF in this patient population.
Hypothesis
LV GLS predicts long‐term HF admission and all‐cause mortality in AA patients with chronic HF on optimal guideline‐directed medical therapy (GDMT).
Methods
We enrolled 207 AA adults, age 56 ± 14.5 years, with New York Heart Association (NYHA) class I through III HF on optimal GDMT from the University of Illinois HF clinic between November 2001 and February 2014. LV GLS was assessed by velocity vector imaging using 2‐, 3‐, and 4‐chamber views. Patients were followed for HF admissions and death for 3 ± 3.0 years. LV GLS value of −7.95 was used as the optimal cutoff point that maximizes sensitivity and specificity
Results
LV GLS < −7.95% was significantly associated with higher all‐cause mortality and HF admissions in Kaplan‐Meier survival curves (log‐rank P < 0.001). After incorporation in multivariate Cox proportional hazard models, GLS < −7.95% was found to be an independent predictor of all‐cause mortality (hazard ratio [HR] = 4.04; 95% confidence interval [CI]: 1.07‐15.32; P = 0.04] and HF admissions (HR = 3.86; 95% CI: 1.38‐10.77; P = 0.010).
Conclusions
In AA patients with chronic stable HF on GDMT, more impaired LV GLS (< −7.95%) is a strong and independent predictor of long‐term all‐cause mortality and HF admissions.
Keywords: Heart failure/cardiac transplantation/cardiomyopathy/myocarditis, African Americans, Mortality, Readmission, Speckle‐tracking Strain
1. INTRODUCTION
Despite advances in the diagnostic techniques and treatment for heart failure (HF), the disease burden continues to grow, affecting millions of people and accounting for well over a million hospitalizations annually.1 Several prognostic indicators have been investigated in an attempt to identify those at highest risk of hospitalizations and morbidity.2 Reduction in left ventricular (LV) function has long been established as a strong predictor of mortality among patients with HF.2, 3 However, several more objective and possibly more sensitive methods for assessing LV function have been emerging, 1 of which includes LV strain.4, 5, 6, 7
LV mechanics are complex, coordinated actions involving multidimensional movement including longitudinal and circumferential shortening and radial thickening. Myocardial strain is a method to measure these multidimensional movements, and may lead to a better understanding of myocardial deformation in LV dysfunction. 6 Strain can be determined by 2‐dimensional speckle tracking, which is an echocardiographic method based on tracking of characteristic speckle patterns created by constructive and destructive interference patterns of ultrasound beams in the myocardium.8 Several studies have demonstrated the importance of global longitudinal strain (GLS) and strain rate (change in myocardial strain over time) as reliable indicators of prognosis in patients with acute and chronic systolic HF, post–myocardial infarction, and ischemic cardiomyopathy. 2, 5, 9, 10 However, these studies have included few African American patients.
African American patients have a 50% higher relative incidence of heart failure, more advanced disease severity, and increased morbidity compared with the general population.11 The mechanism leading to the development of heart failure in this patient population is not yet clear. Socioeconomic status, environmental exposures, and genetic differences likely play a role.11, 12 The majority of African American patients have nonischemic cardiomyopathy including hypertensive heart disease, dilated cardiomyopathy, and alcoholic cardiomyopathy.12 Despite the differences in etiology, and possibly genetics and underlying pathophysiology of heart failure in African American patients, studies assessing prognostic indicators in heart failure have included few African Americans. One study, which included equal proportions of African Americans and Caucasians, found that African Americans (compared with whites) had disproportionately worse GLS and comorbidity burden, regardless of left ventricular ejection fraction (LVEF) as determined by echocardiography.13 Given the differences in the underlying etiology, outcomes, and possible mechanism of the heart in African Americans, we sought to evaluate the prognostic value of LV GLS as a predictor of HF admissions and all‐cause mortality in the African American patient population.
2. METHODS
2.1. Study population
Inclusion criteria consisted of patients with a diagnosis of stage C congestive heart failure with preserved or reduced LVEF, African American race, age ≥18 years, and New York Heart Association (NYHA) class I through III symptoms. A total of 207 African Americans were enrolled from the heart failure clinic at the University of Illinois Hospital and Health Sciences System (UI‐Health) between November 2001 and February 2014. Patients with heart failure with reduced ejection fraction (EF) were on target or maximally tolerated doses of guideline‐directed medical therapy including angiotensin‐converting enzyme inhibitors or angiotensin receptor blockers and β‐blocker therapy. Patients with active malignancy or acute coronary syndrome within 2 months were excluded.
The study protocol was approved by the UI‐Health Institutional Review Board. All patients provided written, informed consent for study participation. On the day of enrollment, a cardiologist performed a complete medical history and physical examination, and a clinical pharmacist obtained a medication history. Complete transthoracic echocardiograms were performed and stored for analysis of echocardiographic parameters and LV strain. Demographic and additional clinical data were obtained from the electronic medical record. Heart failure admissions and all‐cause mortality were prospectively tracked through the UI‐Health electronic medical record throughout the follow‐up period, which was from the time of study enrollment through 2015. Heart failure admission was defined as a hospitalization with a primary diagnosis of acute decompensated heart failure. The Social Security Death Index served as an additional source for patient survival status.
2.2. Transthoracic echocardiography and GLS
Standard 4‐chamber, 3‐chamber, and 2‐chamber apical views and parasternal short‐axis views of the left ventricle were obtained using commercially available ultrasound systems (Siemens SC2000 and Siemens Sequoia 512; Siemens Medical Systems, Erlangen, Germany). All images were stored digitally and analyzed with offline software (Syngo Dynamics 9.0; Siemens Medical Systems) by an independent investigator who was blinded to the clinical data. Standard echocardiographic data were recorded, including LV dimensions and EFs, LV mass index, septal and posterior wall thickness, and left atrial volume according to published guidelines.14 Additionally, pulsed‐wave Doppler mitral inflow peak E‐ and A‐wave velocities, E‐wave deceleration time, and pulsed‐wave tissue Doppler of the lateral and septal mitral annular diastolic velocities (e'), and right ventricular systolic velocity in the apical 4‐chamber view were obtained.
Speckle tracking for myocardial strain was performed using Velocity Vector Imaging software (Siemens Medical Systems). Briefly, 1 cardiac cycle was acquired from apical 4‐chamber and 2‐chamber views. The endocardial border was traced manually in the end diastolic frame. The software detected motion along a user‐defined endocardial‐tracing border by applying successive tracking steps. From this tracking, longitudinal strain was determined from the endocardial trace. The software calculates the average strain values for 6 LV segments for each of apical 2‐, 3‐, and 4‐chamber views (18 total segments). Peak GLS was defined as the peak negative value on the strain curve during the entire cardiac cycle averaged from the 18 segments. Assessment of LV was regarded as suboptimal when speckle tracking could not be obtained in <5 of the 6 myocardial segments in any 1 of the apical views. In the cases where there was suboptimal tracking in an apical view, GLS values were averaged from the analyzable views, as long as there was a minimum of 2 optimal apical views. A single researcher who was blinded to other clinical data performed all primary measurements.
2.3. Statistical analysis
The number of patients and percentages were computed for categorical variables. Means, standard deviations, and medians were computed for continuous variables. The receiver operating characteristic (ROC) curve, which is a measure of the discriminatory power, was drawn to assess the ability of GLS to predict all‐cause mortality and HF admissions. The optimal GLS threshold value was determined as the value that maximized the average of sensitivity and specificity. Clinical characteristics were compared between patients with a GLS level above and below the threshold value (−7.95%) as determined by the ROC curve, by χ2 test for categorical variables, and by t test for continuous variables. Strain values > −7.95 denote improved strain (more negative), whereas strain values < −7.95denote worsening strain (less negative), as defined by the European Association of Cardiovascular Imaging/ American Society of Echocardiography/Industry Task Force consensus document.8 Univariate and stepwise multivariable Cox regression analyses of variables thought to be clinically predictive of outcome were used to assess the association of GLS with time to all‐cause mortality and HF admissions. Variance inflation factors (VIF) were used to test for multicollinearity. VIF was found to be <2 for all variables included in the multivariable analyses, indicating a very low likelihood of collinearity to be present among the included variables. Accordingly, hazard ratio (HR) and 95% confidence interval (CI) were calculated. The Kaplan‐Meier curves and log‐rank tests were used to compare the times to all‐cause mortality and the first heart failure admission across 2 GLS groups. Statistical analyses were conducted using SPSS 21(IBM, Armonk, NY). The significance level was set at P ≤ 0.05.
3. RESULTS
3.1. Study population
Of the 207 patients enrolled, 184 (89%) patients had LV GLS that could be measured. The mean age of the study population was 56.5 ± 14.5 years, 80 (44%) were men, and 39 (22%) had ischemic heart failure. The mean GLS in the study population was −8.8% ± 4.4%.
The area under the ROC curve for the prediction of all‐cause mortality using LV GLS was 0.665 (P = 0.002), and 0.593 for prediction of HF admission (P = 0.054). The optimal GLS threshold value for prediction of all‐cause mortality was −7.95%, corresponding to a sensitivity of 0.75, specificity of 0.60, positive predictive value of 0.31, and negative predictive value of 0.90. For the prediction of HF admissions, the specific and sensitivity for the same GLS threshold was 0.63 and 0.60, respectively, with a positive predictive value of 0.36 and negative predictive value of 0.82. GLS was lower than the threshold value of −7.95% in 85 (46%) patients. Baseline characteristics of patients with a GLS above and below the threshold value are summarized in Table 1. Patients in the more impaired GLS (< −7.95%) group had a higher percentage of male population (P = 0.003), higher prevalence of atrial fibrillation/flutter (P < 0.001), lower LVEF (P < 0.001), and higher LV mass index (P = 0.001). Other characteristics, including heart failure therapy, are summarized in Table 1.
Table 1.
Clinical characteristics by LV GLS
| Characteristic | LV GLS | P | |
|---|---|---|---|
| ≥ −7.95%, n = 99 | < −7.95%, n = 85 | ||
| Age, y | 55.2 ± 15.2 | 58.0 ± 13.6 | 0.188 |
| Male sex | 33 (33%) | 47 (55%) | 0.003 |
| BMI, kg/m2 | 34.5 ± 9.5 | 32.9 ± 10.5 | 0.285 |
| BSA, m2 | 2.1 ± 0.3 | 2.1 ± 0.38 | 0.841 |
| Hypertension | 80 (81%) | 67 (79%) | 0.854 |
| Atrial fibrillation/flutter | 10 (10%) | 32 (38%) | <0.001 |
| Ischemic heart disease | 17 (18%) | 22 (27%) | 0.205 |
| Tobacco use | 51 (52%) | 47 (55%) | 0.766 |
| Cocaine abuse | 4 (4%) | 7 (8%) | 0.351 |
| Medication use | |||
| ACE inhibitor or ARB | 93 (95%) | 80 (94%) | 1.000 |
| β‐Blocker | 96 (98%) | 84 (99%) | 1.000 |
| Hydralazine plus nitrate | 8 (8%) | 20 (24%) | 0.007 |
| Ca channel blockers | 28 (29%) | 11 (13%) | 0.012 |
| Loop diuretics | 78 (80%) | 76 (91%) | 0.063 |
| Digoxin | 23 (24%) | 32 (38%) | 0.052 |
| Statin | 50 (62%) | 50 (68%) | 0.503 |
| NYHA functional class | 0.150 | ||
| Class I | 30 (31%) | 17 (20%) | |
| Class II | 30 (31%) | 24 (28%) | |
| Class III | 38 (39%) | 44 (52%) 1 | |
| Blood pressure, mm Hg | |||
| Systolic | 128.2 ± 24.9 | 122.3 ± 23.8 | 0.104 |
| Diastolic | 72.7 ± 13.1 | 72 ± 15 | 0.746 |
| Heart rate, bpm | 72 ± 12 | 74.5 ± 15 | 0.192 |
| S3, % | 0 (0%) | 4 (5%) | 0.044 |
| JVD, above 8 cm | 4 (4%) | 8 (9%) | 0.230 |
| Rales | 1 (1%) | 3 (4%) | 0.337 |
| GFR | 72.7 ± 27.7 | 64.3 ± 25.7 | 0.037 |
| LDL | 94.5 ± 36.4 | 95.8 ± 32.9 | 0.809 |
| Mean LVEF, % | 44.0 ± 9.0 | 24.3 ± 8.0 | <0.001 |
| LVEF <40% | 35 (35%) | 81 (95%) | <0.001 |
| LV relative wall thickness, cm | 0.4 ± 0.2 | 0.2 ± 0.1 | 0.001 |
| LV mass index, g/m2 | 101.6 ± 54.8 | 132.3 ± 63.5 | 0.001 |
| LA volume, mm3 | 69.1 ± 26.8 | 92.4 ± 31.6 | <0.001 |
| Mitral E/A ratio | 1.3 ± 1.0 | 1.9 ± 1.3 | 0.001 |
| Mitral annulus TDI | |||
| e′ septal cm/s | 6.2 ± 2.2 | 5.8 ± 2.54 | 0.351 |
| e′ lateral cm/s | 9.3 ± 4.4 | 8.2 ± 3.7 | 0.101 |
| E/e′ average ratio | 11.1 ± 7.7 | 12.3 ± 6.1 | 0.284 |
Abbreviations: ACE, angiotensin‐converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; BSA, body surface area; e', early diastolic mitral annular velocity; E/e', ratio of mitral peak velocity of early filling to early diastolic mitral annular velocity; GFR, glomerular filtration rate; JVD, jugular venous distension; LVEF, left ventricular ejection fraction; LV GLS, left ventricular global longitudinal strain; NYHA, New York Heart Association; S3, the third heart sound; SD, standard deviation; TDI, tissue Doppler imaging.
Data are presented as mean ± SD or no. (%).
This was the only subgroup that was significantly different; however, there was no overall significant difference.
3.2. LV GLS and all‐cause mortality
Overall mean time to death or censoring was 3.6 ± 2.9 years, with a mean of 3.0 ± 2.6 years in the more‐impaired GLS group and 3.9 ± 3.6 years in the less‐impaired group. Thirty‐six (20%) of the 184 patients died over the course of the follow‐up period: 26 (31%) died from the more‐impaired GLS group and 10 (10%) from the less‐ impaired GLS group (P < 0.001). The mean baseline GLS was −6.9% ± 4.7% among patients who died and −9.2% ± 4.3% among the survivors (P = 0.006) at the end of study period.
Kaplan‐Meier survival curves (Figure, A) demonstrated significantly higher all‐cause mortality across time in the more impaired GLS group (log‐rank P < 0.001). On univariate Cox proportional hazard analyses, more‐impaired GLS was significantly associated with all‐cause mortality either as a continuous variable (HR = 1.14; 95% CI: 1.04‐1.24; P = 0.004) or dichotomous variable using the threshold of −7.95% (HR = 3.74; 95% CI: 1.79‐7.80; P < 0.001) (Table 2). Additional variables that were significantly associated with all‐cause mortality in univariate analysis were age (P = 0.001), ischemic heart disease (P < 0.001), history of atrial fibrillation or flutter (P < 0.001), glomerular filtration rate (GFR) (P < 0.001), and LV mass index (P = 0.003).
Table 2.
Univariate Cox regression hazard analysis for prediction of all‐cause mortality and HF admission
| All‐Cause Mortality | HF Admission | |||||
|---|---|---|---|---|---|---|
| Variable | HR | 95% CI | P Value | HR | 95% CI | P Value |
| LV GLS < −7.95% | 3.74 | 1.79–7.80 | <0.001 | 2.62 | 1.46–4.72 | 0.001 |
| LV GLS as a continuous variable | 1.14 | 1.04–1.24 | 0.004 | 1.08 | 1.01–1.16 | 0.021 |
| LVEF | 0.97 | 0.95–1.00 | 0.051 | 0.98 | 0.96–1.00 | 0.099 |
| Age | 1.04 | 1.02–1.06 | 0.001 | 1.02 | 1.00–1.04 | 0.017 |
| Male | 1.37 | 0.71–2.27 | 0.351 | 0.99 | 0.56–1.73 | 0.965 |
| NYHA functional class II | 1.90 | 0.73–4.91 | 0.186 | 2.34 | 0.95–5.75 | 0.064 |
| NYHA functional class III | 2.14 | 0.89–5.13 | 0.089 | 3.06 | 1.33–7.05 | 0.008 |
| Ischemic heart disease | 3.37 | 1.71–6.63 | <0. 001 | 1.79 | 0.97–3.29 | 0.063 |
| Atrial fibrillation/flutter | 3.93 | 1.95–7.91 | <0.001 | 1.67 | 0.87–3.18 | 0.122 |
| GFR | 0.975 | 0.96–0.99 | 0.001 | 0.98 | 0.97–0.99 | <0.001 |
| Loop Diuretics | 2.29 | 0.55–9.60 | 0.258 | 1.25 | 0.53–2.94 | 0.616 |
| LV relative wall thickness (cm) | 1.24 | 0.13–11.71 | 0.849 | 1.16 | 0.21–6.53 | 0.866 |
| LV mass index (g/m2) | 1.01 | 1.00–1.02 | 0.003 | 1.00 | 1.00–1.01 | 0.456 |
| Mitral E/A ratio | 1.33 | 1.00–1.77 | 0.053 | 1.21 | 0.94–1.56 | 0.131 |
Abbreviations: CI, confidence interval; GFR, glomerular filtration rate; HF, heart failure; HR, hazard ratio; LV, left ventricular; LV GLS, left ventricular global longitudinal strain; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association.
Figure 1.
(A) Kaplan‐Meier curves of all‐cause mortality for patients stratified by left ventricular global longitudinal strain (LV GLS) > −7.95% versus ≤ −7.95% demonstrating significantly higher all‐cause mortality across time in the more impaired GLS group (log‐rank P < 0.001). (B) Kaplan‐Meier curves of heart failure (HF) hospitalization for patients stratified by LV GLS > −7.95% vs ≤ −7.95% demonstrating significantly higher HF admission rate across time in the more‐impaired GLS group (log‐rank P = 0.001).
After incorporating age, sex, NYHA class, history of ischemic heart disease, GFR, and LVEF in a multivariate model, GLS < −7.95% was found to be a significant predictor of all‐cause mortality (HR = 4.04; 95% CI: 1.07‐15.32; P = 0.040). GLS as a continuous variable was also found to be an independent predictor after incorporating age, sex, NYHA class, history of ischemic heart disease, and GFR in the model (HR = 1.11 for each 1% worsening GLS from a GLS of −7.95%; 95% CI: 1.01‐1.22; P = 0.028).
3.3. LV GLS and heart failure admissions
Overall mean time to HF admission or censoring was 3.2 ± 3.0 years, with a mean of 2.5 ± 2.6 years for the more‐impaired GLS group and 3.8 ± 3.3 years for the less‐impaired GLS group. Forty‐nine (27%) patients were admitted for HF over the course of follow‐up period: 31 (37%) patients from the more‐impaired GLS group and 18 (18%) from the less‐impaired GLS group (P = 0.007). The mean baseline LV GLS was −7.7% ± 4.2% among patients who were admitted and −9.1% ± 4.5% among the ones who were not (P = 0.058).
Kaplan‐Meier survival curves (Figure, B) demonstrated significantly higher HF admission rate in the more‐impaired GLS group (log‐rank P = 0.001). On univariate Cox proportional hazard analyses, more‐impaired GLS was significantly associated with HF admissions either as a continuous variable (HR = 1.08; CI: 1.01‐1.16; P = 0.021) or dichotomous variable using the threshold of −7.95% (HR = 2.62; 95% CI: 1.46‐4.72; P = 0.001) (Table 2). Additional variables that were significantly associated with HF admissions in univariate analysis were age (P = 0.017), NYHA class III (P = 0.008), and GFR (P = 0.001). Controlling for age, sex, NYHA class, history of ischemic heart disease, atrial fibrillation/flutter, GFR, LVEF, and use of loop diuretics, GLS < −7.95% was found to be a significant predictor of HF admissions (HR = 3.86; 95% CI: 1.38‐10.77; P = 0.010) (Table 3). The continuous GLS variable was not significantly associated with HF admissions after incorporating age, sex, NYHA class, history of ischemic heart disease, atrial fibrillation/flutter, GFR, and use of loop of diuretics (HR = 1.05; 95% CI: 0.98‐1.13; P = 0.193).
Table 3.
Multivariate Cox regression hazard analysis for prediction of HF admission
| Heart Failure Admission | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| All Patients (N = 207) | LVEF <40% (N = 116) | No Atrial Fibrillation/Flutter (N = 142) | |||||||
| Variable | HR | 95% CI | P | HR | 95% CI | P | HR | 95% CI | P |
| LV GLS < −7.95% | 3.77 | 1.34–10.57 | 0.012 | 4.33 | 1.09–17.16 | 0.037 | 4.22 | 1.21–14.66 | 0.024 |
| LV GLS as a continuous variable | 1.03 | 0.87–1.2 | 0.736 | 1.06 | 0.83–1.35 | 0.625 | 1.02 | 0.83–1.24 | 0.875 |
| LVEF | 1.03 | 0.99–1.06 | 0.194 | 1.10 | 0.97–1.11 | 0.254 | 1.03 | 0.99–1.08 | 0.187 |
| Age | 1.01 | 0.98–1.04 | 0.558 | 1.00 | 0.97–1.04 | 0.892 | 1.01 | 0.99–1.04 | 0.481 |
| Male | 1.11 | 0.59–2.08 | 0.745 | 1.14 | 0.54–2.41 | 0.724 | 1.40 | 065–3.00 | 0.386 |
| NYHA class II | 2.16 | 0.85–5.50 | 0.107 | 2.60 | 0.74–9.15 | 0.136 | 1.89 | 0.67–5.33 | 0.230 |
| NYHA class III | 2.53 | 1.03–6.24 | 0.043 | 3.26 | 0.99–10.63 | 0.050 | 2.34 | 0.85–6.40 | 0.099 |
| Ischemic heart disease | 1.21 | 0.61–2.40 | 0.579 | 1.30 | 0.55–3.06 | 0.554 | 1.51 | 0.68–3.31 | 0.309 |
| Atrial fibrillation/flutter | 1.01 | 0.49–2.10 | 0.971 | 1.01 | 0.45–2.28 | 0.988 | — | — | — |
| GFR | 0.99 | 0.97–1.00 | 0.070 | 0.98 | 0.97–0.99 | 0.033 | 0.99 | 0.98–1.01 | 0.252 |
| Use of loop diuretics | 1.03 | 0.41–2.59 | 0.957 | 1.59 | 0.34–7.58 | 0.558 | 0.84 | 0.84–2.21 | 0.722 |
Abbreviations: CI, confidence interval; GFR, glomerular filtration rate; HR, hazard ratio; LV GLS, left ventricular global longitudinal strain; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association.
3.4. Patients without atrial fibrillation
3.4.1. LV GLS and all‐cause mortality
On univariate Cox proportional hazard analyses, more impaired GLS was significantly associated with all‐cause mortality as a dichotomous variable using the threshold of −7.95% (HR = 4.06; 95% CI: 1.64‐9.99; P = 0.002) but not as a continuous variable (HR = 1.10; 95% CI: 0.99‐1.22; P = 0.069). After incorporating age, sex, NYHA class, history of ischemic heart disease, GFR, and LVEF in a multivariate model, GLS < −7.95% demonstrated a trend toward predicting all‐cause mortality (HR = 6.47; 95% CI: 0.92‐45.36; P = 0.060) (Table 4).
Table 4.
Multivariate Cox regression hazard analysis for prediction of all‐cause mortality
| All‐Cause Mortality | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| All Patients (N = 207) | LVEF <40% (N = 116) | No Atrial Fibrillation/Flutter (N = 142) | |||||||
| Variable | HR | 95% CI | P | HR | 95% CI | P | HR | 95% CI | P |
| LV GLS < −7.95% | 4.04 | 1.07–15.32 | 0.040 | 1.75 | 0.35–8.77 | 0.498 | 6.47 | 0.92–45.36 | 0.060 |
| LV GLS as a continuous variable | 1.11 | 1.01–1.22 | 0.028 | 1.32 | 0.99–1.76 | 0.062 | 1.09 | 0.74–1.41 | 0.914 |
| LVEF | 1.02 | 0.97– 1.08 | 0.425 | 0.97 | 0.91–1.05 | 0.501 | 1.04 | 0.96–1.12 | 0.334 |
| Age | 1.01 | 0.98–1.05 | 0.416 | 1.01 | 0.96–1.05 | 0.788 | 1.02 | 0.98–1.06 | 0.416 |
| Male | 1.47 | 0.66–3.29 | 0.340 | 2.03 | 0.79–5.23 | 0.144 | 1.51 | 0.46–4.98 | 0.500 |
| NYHA class II | 1.52 | 0.52–4.36 | 0.441 | 1.17 | 0.35–3.92 | 0.797 | 1.15 | 0.31–4.30 | 0.840 |
| NYHA class III | 2.26 | 0.78–6.52 | 0.133 | 1.51 | 0.44–5.23 | 0.517 | 1.73 | 0.45–6.63 | 0.426 |
| Ischemic heart disease | 2.12 | 0.98–4.58 | 0.056 | 2.95 | 1.13–7.69 | 0.027 | 3.39 | 1.13–10.12 | 0.029 |
| GFR | 0.99 | 0.97–1.01 | 0.141 | 0.98 | 0.96–0.99 | 0.041 | 0.99 | 0.98–1.02 | 0.859 |
Abbreviations: CI, confidence interval; GFR, glomerular filtration rate; HR, hazard ratio; LV GLS, left ventricular global longitudinal strain; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association.
3.4.2. LV GLS and heart failure admissions
On univariate Cox proportional hazard analyses, more impaired GLS was significantly associated with HF admissions as a dichotomous variable using the threshold of −7.95% (HR = 2.64; 95% CI: 1.36‐5.12; P = 0.004), and trended toward significance as a continuous variable (HR = 1.07; 95% CI: 0.99‐1.16; P = 0.066). Controlling for age, sex, NYHA class, history of ischemic heart disease, GFR, LVEF, and use of loop diuretics, GLS < −7.95% was found to be significant predictor of HF admissions (HR = 4.21; 95% CI: 1.21‐14.66; P = 0.024) (Table 3). The continuous GLS variable was not significant in the same model.
3.5. Patients with LVEF < 40%
3.5.1. LV GLS and all‐cause mortality
On univariate Cox proportional hazard analyses, more‐ impaired GLS was significantly associated with all‐cause mortality either as a continuous variable (HR = 1.24; 95% CI: 1.07‐1.42; P = 0.004) or dichotomous variable using the threshold of −7.95% (HR = 4.38; 95% CI: 1.31‐14.60; P = 0.016). After incorporating age, sex, NYHA class, history of ischemic heart disease, GFR, and LVEF in a multivariate model, GLS as a continuous variable (HR = 1.32; 95% CI: 0.99‐1.76; P = 0.062) was found to have tendency toward predicting all‐cause mortality, but not as a dichotomous variable < −7.95% (HR = 1.75; 95% CI: 0.35‐8.77; P = 0.498) (Table 4).
3.5.2. LV GLS and heart failure admissions
On univariate Cox proportional hazard analyses, more‐impaired GLS was associated with a trend toward HF admissions as a continuous variable (HR = 1.12; 95% CI: 0.99‐1.27; P = 0.070) and was significant as a dichotomous variable using the threshold of −7.95% (HR = 3.14; 95% CI: 1.21‐8.16; P = 0.019). Controlling for age, sex, NYHA class, history of ischemic heart disease, GFR, LVEF, and use of loop diuretics, GLS < −7.95% was found to be significant predictor of HF admissions (HR = 4.33; 95% CI: 1.09‐17.16; P = 0.037) (Table 3). The continuous GLS variable was not significant in the same model.
4. DISCUSSION
Our data demonstrate that more‐impaired LV GLS with a cutoff value of < −7.95% is a strong and an independent predictor of long‐term all‐cause mortality and HF admissions in African American patients with chronic HF. This is the first study to our knowledge that has specifically assessed LV GLS in a chronic HF population of only African American patients. Prior studies have included largely Caucasian patients and few African Americans. Interestingly, in prior studies in largely Caucasian populations, similar findings have been demonstrated. A study by Motoki et al conducted on patients with chronic HF found that GLS correlated with LV cardiac structure and LVEF, and that GLS of < −6.95% predicted adverse events (death, HF hospitalizations, and cardiac transplant) after adjustment for age, E/septal e', and N‐terminal pro‐brain natriuretic peptide and LVEF. 15 As in our study, severe LV strain abnormality (GLS < −6.95 in the study by Motoki et al) was demonstrated in the overall heart failure population but only included patients with reduced EF (LVEF <35%). On multivariate analysis in our study, HF patients with both preserved and reduced EF were included, and still showed the predictive power of LV GLS independent of LVEF. In a substudy analyses of HF patients with LVEF <40% and in patients without atrial fibrillation, LV GLS < −7.95% remained an independent predictor of HF admission but not mortality, likely due to the reduced statistical power.
In another study by Zhang et al, LV GLS was associated with worse cardiac outcomes, and when combined with EF added incremental value to EF in the prediction of adverse outcomes.10 These studies evaluated composite outcomes; however, our study assessed all‐cause mortality and HF hospitalizations separately. Assessing these separately strengthens the findings that strain predicts HF readmissions, as HF hospitalizations would decline if death rates were higher in this subset. As noted earlier, this is the first study, to our knowledge, to demonstrate the prognostic value of the LV GLS in African Americans with chronic HF with both preserved and reduced EFs on the most current medical treatment as recommended by the American College of Cardiology/American Heart Association guidelines.16 We chose to study patients who were managed in the HF clinic to determine the prognostic value of LV GLS while controlling for differences that may occur secondary to changes and titration of HF therapeutics. Also, LV strain was analyzed as a categorical value, with a cutoff value as well as a continuous variable. Though strain as a continuous variable was also significant for all‐cause mortality in the overall study population in the multivariable model, the use of a cutoff value for prognosticating outcome is likely more relevant in the clinical setting. Also, as seen in this study, the hazard ratios were larger with the strain as a cutoff value.
Our study has a few additional differences as compared to prior studies. Our study population included a higher proportion of females and higher prevalence of hypertension, and did not include any patients with class IV NYHA HF. The higher prevalence of hypertension in African Americans compared to European Americans is expected, as it has been shown in previous studies. The lack of patients with class IV NYHA HF may explain the obtained LV GLS threshold of −7.95%, which corresponds to less‐impaired LV mechanics compared to the previous studies that used LV GLS of −7% and tertiles divided at −6.5% and −9.6%.10
Our results suggest the usefulness of the LV GLS to risk African American patients with HF. Whether patients identified as high risk would benefit from more aggressive medical therapy or closer clinical follow‐up is an area for future research.
4.1. Study limitations
Echocardiograms were not performed at the time of enrollment as patients were enrolled from the cardiology clinic and not from the echocardiography lab. We decided to use the date of echocardiogram as day 0 in the follow‐up process to reflect the true follow‐up period from the day LV GLS was measured. Also, there are multiple strain software programs commercially available with known small intervendor differences in obtained GLS values. The results obtained using the strain software in this study may differ with values performed by other strain software. We also collected data from UI‐Health medical system. Whether patients had outcomes in other institutions that were not reported by the patient is not known.
5. CONCLUSION
In African American patients with chronic stable HF, more‐impaired LV GLS (< −7.95%) is a strong and independent predictor of long‐term all‐cause mortality and HF admissions, superior to LVEF and other known clinical risk predictors.
Conflict of interests
The authors declare no potential conflict of interests.
Kansal MM, Mansour IN, Ismail S, Bress A, Wu G, Mirza O, Marpadga R, Gheith H, Kim Y, Li Y, Cavallari L and Stamos TD. Left ventricular global longitudinal strain predicts mortality and heart failure admissions in African American patients. Clin Cardiol. 2017;40:314–321. 10.1002/clc.22662
Larisa Cavallari is currently at the Department of Pharmacotherapy and Translational Research, University of Florida, Gainesville, Florida.
REFERENCES
- 1. Mozaffarian D, Benjamin EJ, Go AS, et al. Heart disease and stroke statistics—2016 update: a report from the American Heart Association. Circulation. 2016;133:e38–e360. [DOI] [PubMed] [Google Scholar]
- 2. Levy WC, Mozaffarian D, Linker DT, et al. The Seattle Heart Failure Model: prediction of survival in heart failure. Circulation. 2006;113:1424–1433. [DOI] [PubMed] [Google Scholar]
- 3. Sweitzer NK, Lopatin M, Yancy CW, Mills RM, Stevenson LW. Comparison of clinical features and outcomes of patients hospitalized with heart failure and normal ejection fraction (> or = 55%) versus those with mildly reduced (40% to 55%) and moderately to severely reduced (<40%) fractions. Am J Cardiol. 2008;101:1151–1156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Mignot A, Donal E, Zaroui A, et al. Global longitudinal strain as a major predictor of cardiac events in patients with depressed left ventricular function: a multicenter study. J Am Soc Echocardiog. 2010;23:1019–1024. [DOI] [PubMed] [Google Scholar]
- 5. Ersboll M, Valeur N, Mogensen UM, et al. Relationship between left ventricular longitudinal deformation and clinical heart failure during admission for acute myocardial infarction: a two‐dimensional speckle‐tracking study. J Am Soc Echocardiog. 2012;25:1280–1289. [DOI] [PubMed] [Google Scholar]
- 6. Hashimoto I, Li XK, Bhat AH, Jones M, Zetts AD, Sahn DJ. Myocardial strain rate is a superior method for evaluation of left ventricular subendocardial function compared with tissue doppler imaging. J Am Coll Cardiol. 2003;42:1574–1583. [DOI] [PubMed] [Google Scholar]
- 7. Li X, Jones M, Wang HF, et al. Strain rate acceleration yields a better index for evaluating left ventricular contractile function as compared with tissue velocity acceleration during isovolumic contraction time: an in vivo study. J Am Soc Echocardiogr. 2003;16:1211–1216. [DOI] [PubMed] [Google Scholar]
- 8. Voigt JU, Pedrizzetti G, Lysyansky P, et al. Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging. J Am Soc Echocardiog. 2015;28:183–193. [DOI] [PubMed] [Google Scholar]
- 9. Bertini M, Ng ACT, Antoni ML, et al. Global longitudinal strain predicts long‐term survival in patients with chronic ischemic cardiomyopathy. Circ Cardiovasc Imaging. 2012;5:383–391. [DOI] [PubMed] [Google Scholar]
- 10. Zhang KW, French B, Khan AM, et al. Strain improves risk prediction beyond ejection fraction in chronic systolic heart failure. J Am Heart Assoc. 2014;3:e000550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Franciosa JA, Ferdinand KC, Yancy CW; Consensus Statement on Heart Failure in African Americans Writing Group. Treatment of heart failure in African Americans: a consensus statement. Congest Heart Fail . 2010;16:27–38. [DOI] [PubMed] [Google Scholar]
- 12. Yancy CW, Strong M. The natural history, epidemiology, and prognosis of heart failure in African Americans. Congest Heart Fail. 2004;10:15–18. [DOI] [PubMed] [Google Scholar]
- 13. Selvaraj S, Aguilar FG, Martinez EE, et al. Association of comorbidity burden with abnormal cardiac mechanics: findings from the HyperGEN study. J Am Heart Assoc. 2014;3:e000631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of, Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging . 2016;17:412. [DOI] [PubMed] [Google Scholar]
- 15. Motoki H, Borowski AG, Shrestha K, et al. Incremental prognostic value of assessing left ventricular myocardial mechanics in patients with chronic systolic heart failure. J Am Coll Cardiol. 2012;60:2074–2081. [DOI] [PubMed] [Google Scholar]
- 16. Committee Members Writing, Yancy CW, Jessup M, et al; American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines. Circulation . 2013;128:e240–e327. [DOI] [PubMed] [Google Scholar]