Agreement between central venous pressure and right atrial pressure in critically ill patients: In search of the optimal external reference point – A prospective clinical study

Background Right atrial pressure (RAP) is a key determinant of venous return in critically ill patients. Central venous pressure (CVP) is commonly used as a surrogate, but accuracy depends on correct transducer leveling. Evidence on the optimal external reference point, particularly across body positions, remains limited. Methods In this prospective study, 16 sedated, mechanically ventilated ICU patients underwent direct RAP measurement using a solid-state catheter positioned in the mid-right atrium. CVP was recorded using three external reference points: one-third and one-half of the anteroposterior (AP) chest diameter at the fourth intercostal space, and 5 cm below the sternal angle. Measurements were obtained in supine, semi-recumbent (30°), and Trendelenburg (−10°) positions. Agreement between CVP and RAP was assessed using Bland–Altman analysis, intraclass correlation coefficients (ICC), and coverage within ±2 mmHg. Results CVP referenced to one-third of the AP chest diameter showed the closest agreement with RAP across body positions (bias −0.9 to −0.1 mmHg; ICC 0.79–0.90; 94–100% of values within ±2 mmHg). CVP referenced to one-half of the AP chest diameter consistently overestimated RAP by approximately 1–2 mmHg, with variable agreement across positions. The sternal-angle reference showed greater position-dependent error, with wide limits of agreement in the semi-recumbent position (−7.8 to +2.2 mmHg) and only 31–44% of measurements within ±2 mmHg in the semi-recumbent and Trendelenburg positions. RAP decreased in the semi-recumbent position and increased in Trendelenburg (both p < 0.01). Conclusions In sedated, mechanically ventilated ICU patients, referencing CVP to one-third of the AP chest diameter showed the closest agreement with directly measured RAP across body positions, whereas fixed sternal angle referencing introduced greater variability and position-dependent error.