A new method to estimate left ventricular circumferential midwall systolic function by standard echocardiography: Concordance between models and validation by speckle tracking. (15th January 2016)

• Record Type: Journal Article
• Authors: Ballo, Piercarlo; Nistri, Stefano; Bocelli, Arianna; Mele, Donato; Dini, Frank L.; Galderisi, Maurizio; Zuppiroli, Alfredo; Mondillo, Sergio
• Is Part Of: International journal of cardiology. Volume 203(2016)
• Page Start: 947
• Page End: 958
• Publication Date: 2016-01-15
• Subjects: Echocardiography -- Systole -- Left ventricular function -- Speckle tracking; Cardiology -- Periodicals; Electronic journals; 616.12
• DOI: 10.1016/j.ijcard.2015.11.051
• Languages: English
• ISSNs: 0167-5273

Abstract

Abstract: Background: Assessment of left ventricular circumferential (LVcirc ) systolic function by standard echocardiography can be performed by estimating midwall fractional shortening (mFS) and stress-corrected mFS (ScmFS). Their determination is based on spherical or cylindrical LV geometric models, which often yield discrepant values. We developed a new model based on a more realistic truncated ellipsoid (TE) LV shape, and explored the concordance between models among hypertensive patients. We also compared the relationships of different mFS and ScmFS estimates with indexes of LVcirc systolic strain. Methods: In 364 hypertensive subjects, mFS was determined using the spherical (mFSspher ), cylindrical (mFScyl ), and TE model (mFSTE ). Corresponding values of ScmFSspher, ScmFScyl, and ScmFSTE were obtained. Global circumferential strain (GCS) and systolic strain rate (GCSR) were also measured by speckle tracking. Results: The three models showed poor concordance for the estimation of mFS, with average differences ranging between 11% and 30% and wide limits of agreement. Similar results were found for ScmFS, where reclassification rates for the identification of abnormal LVcirc systolic function ranged between 18% and 29%. When tested against strain indexes, mFSTE and ScmFSTE showed the best correlations (R = 0.81 and R = 0.51, p < 0.0001 for both) with GCS and GCSR. Multivariable analysis confirmed that mFSTE and ScmFSTE showed the strongest independent associations with LVcirc strain measures. Conclusions: Substantial discrepancies in LVcirc midwall systolic indexes exist between different models, supporting the need of model-specific normative data. The use of the TE model might provide indexes that show the best associations with established strain measures of LVcirc systolic function.