Incidence, predictability, and outcomes of systemic venous congestion following a fluid challenge in initially fluid-tolerant preload-responders after cardiac surgery: a pilot trial

Background Fluid administration has traditionally focused on preload responsiveness (PR). However, preventing fluid intolerance, particularly due to systemic venous congestion (VC), is equally important. This study evaluated the incidence and predictability of VC following a 7 ml/kg crystalloid infusion in fluid-tolerant preload-responders and its association with adverse outcomes. Methods This single-center, prospective, observational study (May 2023-July 2024) included 40 consecutive patients who were mechanically ventilated within 6 h of intensive care unit (ICU) admission after elective open-heart surgery and had acute circulatory failure. Patients were eligible if they were both fluid-tolerant and preload-responsive. PR was defined as a >= 12% increase in left-ventricular outflow tract velocity time integral (LVOT-VTI) 1 min after a passive leg raising (PLR) test. VC was defined by a portal vein pulsatility index (PVPI) >= 50%. Patients received a 7 ml/kg Ringer's Lactate infusion over 10 min. The primary outcome was the incidence of VC 2 min post-infusion (early-VC). Secondary outcomes included VC at 20 min, the incidence of acute kidney injury (AKI) and severe AKI at 7 days, and ICU length of stay (LOS). Results 45% of patients developed early-VC, with VC persisting in only 5% at 20 min. One-third of patients developed AKI, with 17.5% progressing to severe AKI. The median ICU LOS was 4 days. Patients with early-VC had significantly higher central venous pressure, lower mean perfusion pressure, worse baseline right ventricular function, and a higher incidence of severe AKI. While LVOT-VTI returned to baseline by 20 min in both groups, PVPI remained elevated in early-VC patients (p < 0.001). The LVOT-VTI versus PVPI regression line showed similar slopes (p = 0.755) but different intercepts (p < 0.001), indicating that, despite fluid tolerance and PR at baseline, early-VC patients had reduced right ventricular diastolic reserve (RVDR). Post-PLR PVPI predicted early-VC with an area under the curve of 0.998, using a threshold of 44.3% (p < 0.001). ConclusionPost-PLR PVPI effectively predicts fluid-induced early-VC in fluid-tolerant preload-responders, identifying those with poor RVDR. Its use can guide fluid management in cardiac surgery patients, helping to prevent unnecessary fluid administration and associated complications.