The clinical influence of hand-held ultrasound examinations by general practitioners in patients with suspected heart failure supported by tools for automatic quantification and telemedicine

Abstract

Abstract Funding Acknowledgements Type of funding sources: Public Institution(s). Main funding source(s): Research Council of Norway, Norwegian University of Science and Technology Nord-Trøndelag Hospital Trust Background Heart failure (HF) affects large populations. Echocardiography is the diagnostic cornerstone related to identification of cardiac pathology and classification of the disease. The availability of echocardiography may be a bottleneck in HF diagnostics. Hand-held ultrasound devices (HUDs) are easy to use and have a high availability. So far, the clinical limitations of HUDs have been related to user experience and lack of quantification of cardiac function. Purpose We aimed to evaluate the clinical influence of HUDs by general practitioners (GPs), supported by automatic tools for quantification of ejection fraction (autoEF) and mitral annular plane systolic excursion (autoMAPSE), as well as telemedical support in HF diagnostics. Methods Five GPs with limited ultrasound experience participated in the study. They were equipped with a HUD with automatic tools for LV quantification (autoEF and autoMAPSE). All GPs underwent lectures and 6 days of practical training. In total, 167 patients referred to an outpatient cardiac clinic with suspected HF were included. The GPs examined the patients in chronological stages. First by a clinical examination. Secondly, by adding HUD, then the automatic quantification tools, and finally all recordings were transferred for telemedical support by an external cardiologist. At all stages, the GPs considered whether the patients had HF. A cardiologist’s clinical and echocardiographic examination served as a reference. Results Age was mean ± SD 70 ± 13 years, mean BMI was 29 ± 19 kg/m2 and mean NT-pro-BNP 702 ± 1216 ng/L. Sinus rhythm was present in 76%. Ejection fraction was mean 53 ± 10 %. Of the 167 patients, 28 were diagnosed with HF. The clinical importance of the HUD examinations is shown in the Figure. The GPs correctly classified 55% of the patients (15 with and 77 without HF) after the clinical evaluation. After adding the HUD examination, the proportion increased to 71% (19 with and 99 without HF). The proportion was highest after telemedical support (74%; 20 with and 103 without). The use of HUDs with and without telemedicine improved the clinical benefit compared to the clinical evaluation, p <0.001. There was no significant benefit after adding the automatic tools (autoEF 55%; 15 with and 77 without, autoMAPSE 54%; 17 with and 73 without). The proportion of uncertain assessment were lowest for HUD and telemedicine (clinical 43 (26%), HUD only 20 (12%), autoMAPSE 37 (25%), autoEF 41 (25%), telemedicine 25 (15%)). Conclusion By adding HUD examinations to the clinical evaluation of suspected HF, the GPs diagnostic precision improved both in ruling in and ruling out HF. The use of HUDs combined with telemedical support had the highest proportion of correct reclassification. Tools for automatic quantification of LV function were not of clinical importance. Future studies are needed to evaluate how automatic quantification of cardiac function can support inexperienced users. Abstract Figure.