The SONIC CENTRAL study - Central venous pressure in hospitalized patients: accuracy of physical examination and point-of-care ultrasonography of the jugular vein and inferior vena cava.

BackgroundEstimating central venous pressure (CVP) is essential in the diagnostic evaluation and treatment guidance of most hospitalized patients. It is unknown how different noninvasive, bedside methods to estimate CVP correlate with each other and which method has the best accuracy.Research QuestionWhich noninvasive bedside method to estimate CVP has the best accuracy to detect elevated CVP?Study Design and MethodsDuring this prospective, single-center, observational study, we included patients admitted to the ward or ICU who already had an indwelling central venous catheter and who did not undergo positive pressure ventilation. We measured height of the fluid column in the external jugular vein (EJV) according to Lewis's and Borst's method (EJV height), maximum and minimum diameters and height of the fluid column of the internal jugular vein (IJV; IJV height) using ultrasound, and diameters of the inferior vena cava (IVC) throughout a respiratory cycle and sniffing. We then compared these measurements with intravenously measured CVP.ResultsNinety patients were included. Twenty-seven patients (30%) showed CVP of ≥ 10 mm Hg. All measurements had a significant correlation with CVP, except for the diameter of the IJV. Areas under the receiver operating characteristic curve for IJV height, EJV height, maximum diameter, and collapsibility on inspiration of the IVC were 0.85, 0.80, 0.78, and 0.76 respectively. The interobserver agreement was good to excellent. We estimated continuous likelihood ratios for the measurements to aid clinical decision-making.InterpretationEJV height, IJV height, IVC diameter, and IVC collapsibility can be used to identify an elevated CVP in hospitalized patients. Among these, ultrasonographic estimation of the height of the fluid column in the IJV is quick and easy and allows identification of an elevated CVP with the best reproducibility and accuracy.Trial RegistryNational Trial Register; No.: NL9111