Morpho-mechanistic screening criteria for the echocardiographic detection of rheumatic heart disease

Introduction: Screening echocardiography, guided by the current World Heart Federation (WHF) criteria, has important limitations that impede the establishment of large-scale rheumatic heart disease (RHD) control programmes in endemic regions. The criteria misclassify a significant number of normal cases as borderline RHD. Prior attempts to simplify them are limited by incorporation bias due to the lack of an externally validated, accurate diagnostic test for RHD. We set out to assess novel screening criteria designed to avoid incorporation bias and to compare this against the performance of the current WHF criteria.

Methods: The performance of the WHF and the morpho-mechanistic (MM) RHD screening criteria (a novel set of screening criteria that evaluate leaflet morphology, motion and mechanism of regurgitation) as well as a simplified RHD MM 'rule-out' test (based on identifying a predefined sign of anterior mitral valve leaflet restriction for the mitral valve and any aortic regurgitation for the aortic valve) were assessed in two contrasting cohorts: first, a low-risk RHD cohort consisting of children with a very low-risk RHD profile. and second, a composite reference standard (CRS) RHD-positive cohort that was created using a composite of two criteria to ensure a cohort with the highest possible likelihood of RHD. Subjects included in this group required (1) proven, prior acute rheumatic fever and (2) current evidence of predefined valvular regurgitation on echocardiography.

Results: In the low-risk RHD cohort (n=364), the screening specificities for detecting RHD of the MM and WHF criteria were 99.7% and 95.9%, respectively (p=0.0002). The MM rule-out test excluded 359/364 cases (98.6%). In the CRS RHD-positive cohort (n=65), the screening sensitivities for the detection of definite RHD by MM and WHF criteria were 92.4% and 89.2%, respectively (p=0.2231). The MM RHD rule-out test did not exclude any cases from the CRS RHD-positive cohort.

Conclusion: Our proposed MM approach showed an equal sensitivity to the WHF criteria but with significantly improved specificity. The MM RHD rule-out test excluded RHD-negative cases while identifying all cases within the CRS RHD-positive cohort. This holds promise for the development of a two-step RHD screening algorithm to enable task shifting in RHD endemic regions.