Continuous kidney replacement therapy (CKRT) serves as the preferred extracorporeal kidney support therapy for critically ill patients with hemodynamic instability. Although Kidney Disease: Improving Global Outcomes (KDIGO) recommends a delivered CKRT dose of 20–25 mL/kg/h, its effect on mortality remains unclear. This study aims to evaluate the association between CKRT intensity and mortality.

We conducted a retrospective analysis on 2,265 patients from the Medical Information Mart for Intensive Care IV version 3.1 database. The delivered CKRT dose (mL/kg/h) was classified into low (<20), moderate (20–25), and high (>25). The outcome was 90-day mortality after CKRT initiation. Hazard ratios (HRs) and 95% confidence intervals (CIs) were determined by Cox regression models, supported by multinomial propensity score weighting. The dose–response relationship was assessed using restricted cubic spline models and an additional dose categorization.

Low, moderate, and high intensities were administered to 275 (12.1%), 478 (21.1%), and 1,512 (66.8%) patients, respectively. Low-intensity (HR = 0.62, 95% CI: 0.51–0.76, p <0.001) and moderate-intensity CKRT (HR = 0.79, 95% CI: 0.68–0.91, p = 0.001) were associated with reduced mortality compared with high-intensity. Sensitivity analyses showed a nonlinear dose–response relationship at doses <35, with the lowest mortality observed at 15–19.9.

Low and moderate intensities were associated with reduced mortality, with the greatest survival benefit observed at doses slightly below the KDIGO-recommended range. These findings suggest that this range should be further examined in randomized clinical trials.