Net Ultrafiltration Rate and Mortality in Critically Ill Patients: A Multicenter Cohort Study With Analysis of Fluid...

To investigate the association between net ultrafiltration (NUF) rate and fluid balance changes with 28-day mortality in critically ill patients receiving continuous renal replacement therapy (CRRT).

Multicenter retrospective observational study.

Data from two European mixed medical-surgical ICUs: University Hospital Bern, Switzerland, and Amsterdam University Medical Center, the Netherlands.

Adult critically ill patients receiving CRRT for at least 24 hours.

None.

We studied 973 critically ill adults undergoing a total of 6914 days of CRRT. NUF rates and fluid balance changes were calculated from hourly data and categorized into four predefined groups. NUF rates were classified as no NUF, low (< 1.01 mL/kg/h), moderate (1.01-1.75 mL/kg/h), or high (> 1.75 mL/kg/h). Fluid balance change was stratified as negative (< -1.5 L/day), moderate negative (-1.5 to -0.5 L/day), neutral (-0.5 to 0.5 L/day), or positive (> 0.5 L/day). Using multivariable regression, random forest modeling, and mediation analysis, we observed no independent association between NUF rate and 28-day mortality after accounting for fluid balance. Mediation analysis suggested that the apparent survival advantage observed among patients with higher NUF rates was largely mediated through achievement of a more negative fluid balance, rather than a direct effect of NUF rate itself. In contrast, more positive fluid balance changes were linked to adverse clinical outcomes including 28-day mortality.

In this two-center cohort of CRRT patients, NUF rate was not independently associated with mortality. Instead, the key determinant of outcome was fluid balance change. These findings suggest clinicians should not make NUF decisions in isolation, but rather should be using NUF as a means to achieve patient-specific fluid balance targets.