Sources of variation in assessing left atrial functions by 2D speckle-tracking echocardiography

Left atrial (LA) strain and strain rate, determined by speckle-tracking echocardiography (STE), are reproducible indices to assess LA function. Different normal ranges for LA phasic functions have been reported. We investigated the role of the reference point (P- and R-wave), gain, and region of interest (ROI), as the major sources of variation when assessing LA function. 52 subjects were evaluated for LA conventional and STE analysis. 45 of them (46 ± 14 years, 26 men) were feasible for concomitant LA deformation, and LA phasic volumes and ejection fractions (LAEF) evaluation. First, we compared the P- and R-wave methods, for the evaluation of the LA functions. We used diastolic mitral profile to clearly delineate the time intervals for each LA function. For the P-wave method, active function was assessed from negative global strain as a difference between the strain at pre-atrial contraction and strain just before mitral valve closure (GSA-), and late diastolic strain rate (GSRL); passive function from positive strain at MVO (GSA+), and from early negative diastolic strain rate (GSRE); reservoir function from the sum of GSA- and GSA+ (TGSA), and positive strain rate at the beginning of LV systole (GSR+). For the R-wave method we used the same SR parameters. The active function was evaluated by late positive global strain (GSAC), the reservoir by positive peak before the opening of the mitral valve (TGSA), and conduit function by the difference between TGSA and GSAC (GSA+). Then, by using P-wave method, we measured all previously described parameters for different gains-minimum (G0), medium (G12), and maximum (G24), and for different ROIs-minimum (ROI0), step 1 (ROI1), and 2 (ROI2). Feasibility of the LA strain measurements was 87 %. Active LA function was similar in the absolute value (GSAC and GSA-), whereas passive and reservoir functions were significantly higher (GSA+, TGSA) with the R-wave method. Active LAEF correlated with GSA- measured by the P-wave (r = -0.44, p = 0.002), but not with the GSAC measured by the R-wave method. Similar correlations were found for passive and reservoir LAEF with correspondent strain parameters, only with P-wave method. There were no differences between methods regarding SR indices and their correlations with correspondent LAEFs. Increase of gain from minimum to maximum overestimated all measured LA functions (all p < 0.05). Intermediary changes did not have a significant impact on the measurement of active and conduit function, but they do have on the measurement of the reservoir function. Increase of ROI from minimum to ROI2 was associated with an overestimation of all measurements of atrial functions (all p < 0.05). For all parameters, except GSR+, a decrease of atrial S and SR values from minimum ROI to step 1 was recorded. For GSA+, TGSA, GSRE a decrease of S and SR values with each ROI step was recorded. The two methods used to assess LA functions by STE do not provide similar results. The R-wave method essentially ignores negative peak, creating a positive strain for atrial contraction, and also provides higher values for the reservoir and conduit functions, by comparison with the P-wave method. Increase of gain overestimates, whereas increase of ROI underestimates all parameters of LA functions. Therefore, we suggest that P-wave as a reference point, a medium gain, and a minimum ROI should be used as the best choice for a correct assessment.