Introduction:
Echo and cardiac magnetic resonance imaging (CMR) are often used in evaluation of patients with heart failure (HF) whether associated with systolic dysfunction or preserved ejection fraction (EF). We evaluated the relative merits of EF (by CMR and 2D echo) versus echo strain (by 2D and 3D speckle tracking echo STE) for imaged-based recognition of patients with HF.
Methods:
117 subjects (81 normals and 36 HF) were evaluated. There were no significant differences in age or gender between normals (59+/-14 yrs, 39 M) and HF (54+/-14 yrs, 29 M). 2D biplane echo and CMR provided EF while 2D and 3DSTE yielded global longitudinal (GLS %) and circumferential (GCS%) strains. Since 3DSTE was optimized for strain rather than EF, 3DEF was not evaluated. HF diagnosis was based on dyspnea, plasma N-terminal pro-B type brain natriuretic peptide and EF.
Results:
As expected, mean EF and strains differed between normals and HF (table). Using a backward elimination regression model, only 2DEF, CMREF, 2DGCS% and 3DGLS% were retained for further analysis. ROC-guided criterion values (for HF recognition) for 2DEF, CMREF, 2DGCS% and 3DGLS% were >52%, >50%, >-22% and >-16%, respectively. Using these values, the sensitivity, specificity, and odd-ratios for HF recognition with each parameter were computed (table).
Conclusions:
CMREF provided the highest sensitivity and specificity for HF recognition. The most sensitive echo measure was 3DGLS% while the most specific echo measure was 2DEF. However, in a mixed population of normals and HF, whether associated with systolic dysfunction or preserved EF, 2DGCS% and 3DGLS% had the highest odds ratios.
Imaging in heart failure with preserved ejection fraction: insights into echocardiography and cardiac magnetic resonance imaging
