Author + information
- Published online December 4, 2017.
- John L. Colquitt, MD∗ (, )
- Aamir Jeewa, MD,
- Shaine A. Morris, MD, MPH,
- Kristen Sexson Tejtel, MD, PhD, MPH,
- William J. Dreyer, MD,
- Susan W. Denfield, MD and
- Ricardo H. Pignatelli, MD
- ↵∗Division of Cardiology, Department of Pediatrics, Baylor College of Medicine and Texas Children’s Hospital, 6621 Fannin Street, MC-19345C, Houston, Texas 77030
Noninvasive measures to assess allograft status and prognosis in children post–heart transplantation (HT) have long been an area of investigation (1), yet robust markers of allograft dysfunction remain elusive. Recent adult guidelines recommend the serial evaluation of global longitudinal strain (GLS) in HT recipients to detect subclinical left allograft dysfunction (2). We hypothesized that abnormal GLS obtained by speckle-tracking echocardiography would predict late allograft failure in pediatric HT recipients.
We prospectively-recruited 104 pediatric HT patients to undergo speckle-tracking echocardiography prior to surveillance cardiac catheterization and endomyocardial biopsy. Standard 2-dimensional echocardiograms were performed using a GE Vivid E9 or E7 ultrasound system (GE Healthcare, Waukesha, Wisconsin). Longitudinal strain of the left ventricle was assessed using apical 4-, 3-, and 2-chamber images, and offline analysis of digitally stored images was performed using EchoPAC software version 110.0.2 (GE Healthcare). The primary study endpoint was allograft failure, defined as cardiac death, hospice enrollment, or listing for repeat HT. Patients with evidence of rejection (greater than or equal to grade 2R) on histologic review of the enrollment biopsy specimen were excluded, as GLS can be reduced during episodes of acute rejection (3), and due to the potential of acute rejection to introduce error as both a cause of death and as a mediator of allograft failure. Cox multivariable regression models were created using: 1) continuous echocardiographic measures; and 2) data-driven threshold values to potentially provide improved clinical utility to the measures. Interobserver variability of GLS was excellent (intraclass correlation coefficient 0.94, n = 20 studies).
Median age at enrollment was 10.5 years (interquartile range [IQR]: 6.0 to 16.2 years) and median duration of follow-up was 3 years (IQR: 2.5 to 3.4 years). Median age at transplant was 3.2 years (IQR: 0.8 to 11.1 years), and median allograft age at the time of the echocardiogram was 3.3 years (IQR: 0.9 to 7.1 years). Twenty-one (20%) patients developed allograft failure at a median follow-up of 2.6 years (IQR: 1.7 to 3.2 years) post-echocardiogram. Factors associated with allograft failure by univariable analysis included older age at transplant, higher mean pulmonary capillary wedge pressure, cardiac allograft vasculopathy, lower left ventricular ejection fraction (LVEF), higher mitral early diastolic velocity (E)/late diastolic velocity (A) ratio, larger left atrial volume, and reduced GLS and GLS rate. By multivariable analysis using continuous echocardiographic measures, reduced GLS (hazard ratio [HR]: 1.2 per 1% increase; 95% confidence interval [CI]: 1.02 to 1.40; p = 0.02), larger left atrial volume (HR: 1.6 per 10 ml/m2 increase; 95% CI: 1.1 to 2.3; p = 0.02), and cardiac allograft vasculopathy (HR: 3.6; 95% CI: 1.4 to 9.4; p < 0.01) remained independent predictors of allograft failure.
Using accepted normal values, LVEF was transformed into a categorical variable. To identify a clinically useful threshold value for GLS, analysis by quartiles supplemented by a receiver operating curve analysis were performed. A GLS of –18% or higher (less negative) had a sensitivity of 91% and specificity of 55% for allograft failure, with an area under the curve of 0.710 (p < 0.01). The categorical variables LVEF and GLS were then combined to create an additional nested variable, consisting of 3 subgroups: group 1 (reference group) had normal LVEF and better strain, group 2 had either lower LVEF or worse strain, and group 3 had both lower LVEF and worse strain (Figure 1). Factors independently associated with allograft failure were EF <55% or GLS >–18% (HR: 4.0; 95% CI: 1.1 to 14.4; p = 0.03), or presence of both EF <55% and GLS >–18% (HR: 6.7; 95% CI: 1.6 to 27.3; p < 0.01). Among 44 patients with at least 2 echocardiograms (median elapsed time between studies 249 days [IQR: 165 to 347 days]), 10 of whom developed allograft failure at a median of 1.9 years (IQR: 1.6 to 2.6 years), worsening GLS between examinations predicted allograft failure (HR: 1.1 per 1% increase; 95% CI: 1.02 to 1.30; p = 0.03, secondary univariable analysis (Figure 1).
This study shows that reduced baseline GLS and worsening GLS at follow-up provides important prognostic information in pediatric HT patients and may enhance the identification of high-risk HT patients, independent of and incremental to LVEF. The small number of outcomes limits the strength of the analysis, but our results support the integration of GLS into surveillance protocols following pediatric HT.
Please note: The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation
- Badano L.P.,
- Miglioranza M.H.,
- Edvardsen T.,
- et al.