Feasibility and Accuracy of Cardiac Right-to-Left-Shunt Detection in Children by New Transpulmonary Ultrasound Dilution Method.

Abstract

Transpulmonary ultrasound dilution (TPUD) method, a novel indicator dilution (ID) technique for cardiac output measurement, detects and quantifies shunts, both in children and adults. However, its accuracy and reproducibility in cardiac right-to-left-shunt (RLS) detection have not been investigated. In a prospective observational study, we assessed the validity of TPUD algorithm for RLS detection in children with congenital heart disease (CHD) and proven RLS in comparison with controls without shunts between February 2010 and October 2011. As TPUD algorithm was unknown, we tested ID curve morphology, appearance time (AT) and central blood volume index (CBVI) as diagnostic criteria. TPUD identified RLS correctly in all 16 RLS subjects [median age (range): 18months (1month-15years 6months)] and excluded RLS in all 26 controls [74months (8months to 17years 4months)]. AT was significantly shorter in RLS (P<0.05). Applying only AT (93.8% sensitivity, 92.3% specificity), RLS can be detected by shortening of ≥1.69s of normally expected AT. RLS ID curves were subdivided into four morphological categories: (I) hump-on-upslope (n=5); (II) double-hump (n=3); (III) pseudonormal (n=3); (IV) abnormal width (n=5). No correlation was found between specific type of CHD and RLS categories. CBVI measurements were significantly smaller in RLS categories I-III than in controls (P<0.05). TPUD appears to be a valid method for cardiac RLS detection. Shortened AT and low CBVI are reliable parameters for RLS identification. RLS categories have specific implications for cardiac output, blood volume and RLS fraction measurements. TPUD is valuable to monitor shunt direction and magnitude to optimise haemodynamic and respiratory therapy.