Screening for rheumatic heart disease: The reliability of anterior mitral valve leaflet thickness measurement.

Abstract

Studies determining the reliability of the World Heart Federation (WHF) anterior mitral valve leaflet (AMVL) measurement are limited by the introduction of bias in their test-retest analyses. This study sought to determine the reliability of the current AMVL measurement while controlling for systematic bias. Retrospective analysis of echocardiographic data from 16 patients with previous acute rheumatic fever was performed. Included in this study was an optimized cine loop of the mitral valve (MV) [reader-optimized measurement (ROM]) in the parasternal long-axis view and an optimized still image of the MV obtained from the same cine loop [specialist-optimized image (SOI)]. Each still image and associated cine loop was quadruplicated and randomized to determine intra- and inter-rater agreement and quantify the impact of zoom on AMVL measurement. Specialist-optimized image without zoom reflected the highest degree of agreement in both cohorts with an ICC of 0.29 and 0.46. The agreement in ROM images without zoom was ICC of 0.23 and 0.45. The addition of zoom to SOI decreased agreement further to an ICC of 0.20 and 0.36. The setting associated with the poorest agreement profile was ROI with zoom with an ICC of 0.13 and 0.34, respectively. The intra-rater agreement between readers in both cohorts was moderate across all settings with an ICC ranging between 0.64 and 0.86. The WHF AMVL measurement is only moderately repeatable within readers and demonstrates poor reproducibility that was not improved by the addition of a zoom-optimized protocol. Given our study findings, we cannot advocate the current WHF AMVL measurement as a reliable assessment for RHD.