Reproducibility of the assessment of myocardial function through tissue Doppler imaging in very low birth weight infants

OBJECTIVES: This study aims to analyze the variability between two trained neonatologists when performing consecutive echocardiograms using tissue Doppler imaging (TDI) and conventional methods in very low birth weight infant (VLBWI). METHODS: Two serial echocardiograms were performed in 30 VLBWI infants. The echocardiographic parameters analyzed were tricuspid annular plane systolic excursion (TAPSE), A', E', and S' waves, and myocardial performance index acquired by TDI (MPI-TDI) of both ventricles and shortening fraction (SF). The intra-observer and inter-observer agreements and the intra-operator agreement were analyzed using quantitative and qualitative statistical methods. RESULTS: The intra-observer agreement was very good, TAPSE, and TDI-derived parameters had an intra-class correlation (ICC) > 0.8. TDI-derived velocities had a coefficient of variation (COV) < 11%, while MPI-TDI had a COV between 20%-28%. The inter-observer agreement was excellent. There was greater variability when analyzing intra-operator agreement, with the least variable parameter being TAPSE. According to PABAK, the variability presented moderately substantial agreement. CONCLUSIONS: Tricuspid annular plane systolic excursion is very reproducible between observers and operators. Measurements of TDI wave velocities are more reproducible than MPI-TDI. TDI is sufficiently reproducible in the VLBWI if adequate training is performed, and guidelines are followed to obtain standardized echocardiographic images.