The prognostic impact of net ultrafiltration intensity in critically ill patients receiving continuous renal replacement therapy: a multivariable and propensity-matched analysis

Jin Lü, Peiyun Li, Yongjiang Tang, Wanhong Yin, Fang Wang, Ling ZhangRenal 2025 · DOI

TL;DR

This study investigates the optimal rate of fluid removal (net ultrafiltration intensity, UF<sup>net</sup>) in critically ill patients undergoing continuous renal replacement therapy (CRRT). They found that a moderate UF<sup>net</sup> intensity (1.01-1.75 mL/kg/h) during the initial 72 hours of CRRT is associated with lower 28-day mortality, suggesting a "sweet spot" for fluid management in these vulnerable patients.

What Problem Does This Paper Address?

Critically ill patients often develop acute kidney injury (AKI) and fluid overload (FO), which are significant contributors to mortality. Continuous renal replacement therapy (CRRT) is a common treatment to manage fluid balance in these patients. However, the precise intensity of fluid removal (net ultrafiltration, UF<sup>net</sup>) that optimizes patient outcomes, particularly survival, remains a subject of debate. Current guidelines offer limited specific guidance on this crucial parameter. Previous studies have yielded conflicting results, with some suggesting higher UF<sup>net</sup> is detrimental and others indicating a benefit from more aggressive fluid removal. This paper aims to clarify the relationship between UF<sup>net</sup> intensity and mortality in a contemporary cohort of critically ill patients receiving CRRT, specifically investigating whether a moderate UF<sup>net</sup> intensity offers a survival advantage.

Methods

This was a retrospective observational study conducted at a single center in China. The researchers analyzed data from 683 critically ill patients who received CRRT for more than 72 hours. Patients were stratified into three groups based on their UF<sup>net</sup> intensity during the initial 72 hours of CRRT: low (<1.01 mL/kg/h), moderate (1.01-1.75 mL/kg/h), and high (>1.75 mL/kg/h).

Key analytical techniques included:

  • Kaplan-Meier survival curves and log-rank tests: Used to compare survival probabilities between the UF<sup>net</sup> intensity groups.
  • Cox proportional hazards models: Employed to assess the association between UF<sup>net</sup> intensity and 28-day mortality, both unadjusted and adjusted for various confounding variables (age, sex, weight, baseline creatinine, APACHE II score, SOFA score, fluid overload, ICU admission diagnosis, and time from ICU admission to CRRT).
  • Propensity Score Matching (PSM): A 1:1:1 matching strategy was used to balance baseline characteristics between the groups and further control for confounding factors, aiming to establish a more robust association.
  • Multivariable generalized additive model: Used to explore the non-linear association between UF<sup>net</sup> intensity and mortality.

Sensitivity analyses were performed for 60-day and 90-day mortality, and subgroup analyses were conducted based on admission diagnosis.

Datasets

The study utilized data from electronic CRRT machine records and electronic medical records of patients admitted to the West China Hospital of Sichuan University between January 2021 and September 2023. This represents a single-center dataset of critically ill patients undergoing CRRT. No new benchmarks or datasets were introduced.

Key Insights and Ideas

  • The "Sweet Spot" for Fluid Removal: The most striking finding is the identification of a moderate UF<sup>net</sup> intensity (1.01-1.75 mL/kg/h) as being associated with the lowest 28-day mortality. This suggests that both too little and too much fluid removal can be detrimental.
  • J-Shaped Relationship Confirmed: The results align with previous research suggesting a J-shaped relationship between UF<sup>net</sup> intensity and mortality, where extreme values on either end of the spectrum are associated with worse outcomes.
  • Intensity Matters Beyond the First 72 Hours: While the study focused on the initial 72 hours, the findings suggest this early period is critical for setting the trajectory of fluid management and its impact on survival.
  • UF<sup>net</sup> Intensity and Other Outcomes: Higher UF<sup>net</sup> intensity (>1.75 mL/kg/h) was linked to longer CRRT duration, while the moderate intensity group experienced longer ICU and hospital stays, likely reflecting improved survival.
  • Regional Applicability: The study aimed to assess if international findings on UF<sup>net</sup> thresholds were applicable to an Asian population, suggesting that these moderate intensity ranges may have broader relevance.

Key Results

  • Compared to the moderate UF<sup>net</sup> intensity group (1.01-1.75 mL/kg/h), both the low (<1.01 mL/kg/h) and high (>1.75 mL/kg/h) UF<sup>net</sup> intensity groups were associated with significantly higher 28-day mortality, even after multivariable adjustment and propensity score matching.
  • Specifically, the adjusted hazard ratio (HR) for 28-day mortality was 1.54 (95% CI: 1.24-1.91, p < .001) for the low UF<sup>net</sup> group and 1.27 (95% CI: 1.03–1.57, p = .024) for the high UF<sup>net</sup> group, relative to the moderate group.
  • Sensitivity analyses showed similar trends for 60-day and 90-day mortality.
  • Subgroup analyses by admission diagnosis did not reveal significant differences in the impact of UF<sup>net</sup> intensity on mortality risk.
  • Patients in the high UF<sup>net</sup> intensity group had a longer CRRT duration.
  • Patients in the moderate UF<sup>net</sup> intensity group had increased length of hospital and ICU stay.

Limitations and Future Work

The authors acknowledge several limitations:

  • Single-Center, Retrospective Design: This limits the ability to establish a causal relationship and generalizability to other ICU populations.
  • Potential for Residual Confounding: Despite extensive adjustments, unmeasured confounders may still influence the observed associations.
  • Missing Data: 23 patients were excluded due to missing critical variables, though sensitivity analyses suggested this did not materially affect outcomes.
  • Incomplete Diuretic Documentation: The use and dosing of diuretics prior to CRRT were not fully documented, which could confound results.
  • Limited Sample Size: The sample size may have affected the robustness of subgroup analyses.
  • Fluid Overload Measurement: The study found FO >5% was not a significant prognostic factor, potentially due to difficulties in accurately assessing true fluid balance in transferred patients.
  • Fixed Timeframes: The analysis focused on fixed early timeframes (24, 48, 72 hours) and did not capture the dynamic trajectory of fluid removal over time.

Future work suggested includes:

  • Prospective Validation: Conducting prospective studies to validate the optimal UF<sup>net</sup> thresholds.
  • Multicenter Studies: Expanding research to multicenter ICU cohorts to improve generalizability.
  • Time-Series Analysis: Employing time-series methods or trajectory-based clustering to better understand the dynamic evolution of UF<sup>net</sup> adjustments in response to a patient's clinical status.

Summary updated

Abstract

UF<sup>net</sup> intensity between 1.01 and 1.75 mL/kg/h during the first 72 h of CRRT was associated with lower 28-day mortality compared to both lower and higher UF<sup>net</sup> intensities. However, future studies are needed to better define optimal UF<sup>net</sup> thresholds in multicenter ICU cohorts.
BibTeX
@inproceedings{lu2025prognostic,
  author = {Jin Lü and Peiyun Li and Yongjiang Tang and Wanhong Yin and Fang Wang and Ling Zhang},
  title = {{The prognostic impact of net ultrafiltration intensity in critically ill patients receiving continuous renal replacement therapy: a multivariable and propensity-matched analysis}},
  booktitle = {Renal},
  year = {2025},
  doi = {10.1080/0886022x.2025.2537810},
}
Figure