Interrater reproducibility of the 2016 American society of echocardiography left ventricular diastolic function guidelines

Background Little data exist regarding interreader variability of diastolic measurements and their application by the 2016 American Society of Echocardiography left ventricular (LV) diastolic function guidelines. Methods Volunteers (n = 49) were recruited from an outpatient cardiology practice. The presence and grade of diastology dysfunction (DD) was determined by the 2016 LV diastology guideline algorithm. We determined the mean, standard deviation, coefficient of variation, and intraclass correlation coefficient (ICC) for each measurement and Fleiss K-statistic to define differences in grading DD. We determined predictors associated with disagreement of DD grade using odds ratios. Results The mean LVEF was 56%, LAVI 32 ml/m(2), and peak TR velocity was 2.3 m/s. The ICC for mitral inflow and tissue Doppler velocities were >.90, for LV volumes were .80-.86, and for LA volume was .56. The Fleiss K-value for the agreement of the presence of DD was .68 and for DD grade was .59. Variables with increased odds of disagreement were (1) at least one reader considering a TR signal uninterpretable (OR 12.0; 95% CI 1.3-109.6), (2) at least one reader assessing both LVEF 50%-55% and LAVI 29-39 ml/m(2) (OR 9.3; 95% CI 1.0-87), and (3) at least one reader assessing LVEF 50-55% (OR 3.8; 95% CI 1.1-13.4). Conclusions Using the 2016 ASE/EACVI diastology guidelines, we found excellent interrater reliability of Doppler measurements, moderate-good interrater reliability of volumetric measurements, and moderate-good but not excellent agreement for diastology grade.