End-stage renal disease (ESRD) presents a significant health burden and is associated with high cardiovascular morbidity and mortality. This is particularly true in African Americans who generally have higher rates of cardiovascular mortality. Outcomes in ESRD are related to extent of cardiovascular disease, but markers for outcome are not clearly established. Global longitudinal strain (GLS) has emerged as an important measure of left ventricular systolic function that is additive to traditional ejection fraction (EF). It can be measured on routine digital echocardiography and is reproducible. This study tested the hypothesis that GLS is associated with mortality in black Americans with ESRD and preserved EF. Forty-eight outpatients undergoing hemodialysis, 59.4 ± 13.3 years, with EF ≥50% were prospectively enrolled. GLS, measured by an offline speckle tracking algorithm, ranged from −8.6% to −22.0% with a mean of −13.4%, substantially below normal (−16% or more negative). The prevalence of left ventricular systolic dysfunction, as determined by GLS, was 89%. Patients were followed for an average of 1.9 years; all-cause mortality was 19% (9 deaths). GLS was significantly associated with mortality (hazard ratio 1.15, 95% confidence interval 1.02 to 1.30, p = 0.02), whereas EF was not. After adjustment for multiple potential confounders (age, gender, race, smoking, hypertension, diabetes, hyperlipidemia, coronary disease, heart failure, and EF), GLS remained strongly associated with mortality (hazard ratio 1.30, 95% confidence interval 1.10 to 1.56, p = 0.002). In conclusion, GLS is an important index in patients with ESRD, which is additive to EF as a marker for mortality in this high-risk group.
Patients with end-stage renal disease (ESRD) have limited life expectancy with most dying of cardiovascular disease. African Americans have higher cardiovascular mortality rate than whites and have been reported to have worse mortality among those with ESRD. Echocardiography is a powerful diagnostic tool that conveys important prognostic information, in particular, by determining left ventricular (LV) ejection fraction (EF). Newer echocardiographic measurements have the potential to improve risk prediction in general, and in ESRD in particular, even in patients with preserved EF. Global longitudinal strain (GLS) is a more sensitive measure of LV systolic function than EF. By measuring displacement between 2 points in a myocardial segment strain provides a means to determine myocardial deformation, which relates to fibrosis and contractile function. Averaging strain in the various myocardial segments, as imaged in the apical views, yields GLS. This measure is reproducible and predictive of cardiovascular events generally and in a number of specific settings (e.g., heart failure, aortic stenosis, and acute myocardial infarction). We hypothesized that GLS, as measured on standard 2-dimensional echocardiography, would be a marker of mortality risk among African American patients with ESRD and a preserved EF.
Methods
The study sample was obtained from a group of 75 patients with ESRD who were prospectively enrolled from outpatient dialysis centers. The study was approved by the Einstein Healthcare Network’s Institutional Review Board, and each subject signed informed consent. All were clinically stable, and all had been receiving hemodialysis for at least 6 months (mean 3.8 ± 2.5 years). None of the subjects had greater than mild valvular heart disease. This study focuses on the 65 subjects who were African Americans. From this group, we excluded 6 subjects with an EF of <50% leaving 59 subjects. A further 6 subjects were excluded because of wall motion abnormalities (2 subjects), severe pulmonary hypertension (2 subjects), aortic stenosis (1 subject), and moderate mitral regurgitation (1 subject). Of the remaining 53 subjects, 48 had images suitable for speckle strain analysis and form the study sample. All echocardiograms were obtained on a nondialysis day to avoid the hemodynamic and volume shifts that can occur in relation to hemodialysis.
Echocardiographic images of the left ventricle were acquired in apical 4- and 2-chamber views. Longitudinal strains were then measured using a DICOM-based speckle tracking system (2D Cardiac Performance Analysis, version 1.1; TomTec Imaging System, Munich, Germany). This software can analyze 2-dimensional echocardiographic data from various ultrasound machines. Each apical view was divided into 6 segments, and peak systolic longitudinal strain was calculated for each of these 12 segments; GLS was determined as the average of all regional strains.
Baseline demographic and clinical characteristics were recorded, including age, gender, race, and presence or absence of various clinical risk factors such as active smoking, hypertension, diabetes, hyperlipidemia, coronary artery disease, and heart failure. Vital status was determined by phone contact, search of hospital records, and query of the Social Security Death Index. Total length of follow-up was determined from the date of the index echocardiogram to study end (June 1, 2013) or the date of death.
Variables are displayed as mean and standard deviation for continuous variables and as numbers and percentages for categorical variables. Proportional hazards ratios were calculated to test the independent predictive value of GLS, both as a dichotomized variable and as a continuous variable. Unadjusted results are presented and results after controlling for various confounders: model 1 adjusted for age and gender; model 2 adjusted for the age and gender plus clinical factors (smoking, hypertension, diabetes, hyperlipidemia, coronary artery disease, heart failure, and EF). A 2-tailed p value <0.05 was considered significant. Statistical analyses were performed using JMP 9 software (SAS Institute Inc., Cary, North Carolina).
Results
Baseline clinical and echocardiographic characteristics are listed in Table 1 . Twenty-six subjects (54%) were men. Mean age was 59.4 ± 13.3 years. Follow-up ranged from 87 days to 1,047 days with an average of 713 days (1.9 years; median follow-up of 741 days = 2.0 years). During this time, there were 9 deaths recorded (19%). GLS was found to be normally distributed, ranging from −8.6% to −22.0%. Average value for GLS was −13.4%. This is considerably below normal, generally taken to be −16% or more negative. Using this metric, 89% of our sample had abnormal LV systolic function.
| Variable | |
|---|---|
| Age (years) | 59.4 ± 13.3 |
| Sex (male) | 26 (54%) |
| Diabetes Mellitus | 28 (58%) |
| Hypertension | 45 (94%) |
| Smoker | 10 (21%) |
| Dyslipidemia ∗ | 23 (48%) |
| Coronary Artery Disease | 14 (29%) |
| History of Heart Failure | 12 (25%) |
| Years on dialysis | 4.1 ± 2.3 |
| Septal Wall Thickness (cm) | 1.2 ± 0.2 |
| Posterior Wall Thickness (cm) | 1.3 ± 0.2 |
| Left Atrial Dimension (cm) | 4.5 ± 0.6 |
| Left Ventricular Dimension (cm) | 4.7 ± 0.7 |
| Ejection Fraction (%) | 59 ± 8 |
∗ Dyslipidemia defined as total cholesterol >200 mg/dl and/or triglycerides >150 mg/dl.
Results of the various proportional hazards models are presented in Table 2 . In unadjusted analysis, GLS was significantly associated with mortality with a hazard ratio (HR) of 1.15 (i.e., 15% increase in mortality for each 1% less negative GLS), 95% confidence interval (CI) 1.02 to 1.30, p = 0.02. Adjusting for age and gender (model 1), GLS remained predictive (HR 1.21, 95% CI 1.06 to 1.39, p = 0.005). In the fully adjusted model (including clinical variables, model 2), the HR was 1.30, 95% CI 1.10 to 1.56, p = 0.002. Smoking (p = 0.02) and diabetes (p = 0.04) were also independently associated with mortality. A GLS cutoff of −16% has been previously reported to have prognostic value in patients with advanced chronic kidney disease. Applying this value to our sample, those with GLS less negative than this (worse function) had a risk ratio for mortality of 4.1 (p = 0.02). After adjustment for age, gender, and clinical variables, GLS less negative than −16% was associated with an 8.2 times greater risk of mortality (p = 0.01, Figure 1 ).
| Model | Hazard Ratio | 95% CI | P value |
|---|---|---|---|
| Unadjusted | 1.15 | 1.02 – 1.30 | 0.02 |
| Model 1 Adjusted for Age and Sex | 1.21 | 1.06 – 1.39 | 0.005 |
| Model 2 Fully Adjusted (Model 1 + clinical factors ∗ ) | 1.30 | 1.10 – 1.56 | 0.002 |
∗ Clinical Factors = diabetes, hypertension, hyperlipidemia, smoking, diagnosed coronary artery disease, history of heart failure, and ejection fraction.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
