Highlights
– In a randomized, cross‑over trial of 20 critically ill patients on CRRT, a 250‑ml net ultrafiltration (UFNET) challenge identified those who converted from preload‑independent to preload‑dependent during therapy (overall post‑challenge preload‑dependence 33%).
– The relative change in calibrated cardiac index (∆CIUFC) between before and after the 250‑ml UFNET challenge showed acceptable diagnostic accuracy (mITT AUROC 0.83; ITT AUROC 0.74).
– No statistically significant difference was found between a fast (15‑min) and slow (30‑min) removal of 250 ml, and a calibrated cardiac index (CCI) variation ≥5% during a pre‑challenge postural maneuver predicted de novo preload‑dependence (AUROC 0.82).
Background
Fluid management is a central challenge in critically ill patients, especially those receiving continuous renal replacement therapy (CRRT). While net ultrafiltration (UFNET) can remove excess fluid, excessive or poorly tolerated UFNET may precipitate hemodynamic instability by reducing cardiac preload. Distinguishing patients who are preload‑dependent (i.e., whose cardiac output increases meaningfully with preload augmentation) from those who are preload‑independent helps clinicians tailor fluid removal safely. Conventional dynamic tests (passive leg raising or mini fluid challenges) assess fluid responsiveness, but their role in predicting tolerance to UFNET during CRRT has been less explored.
Study design
Biscarrat et al. conducted a single‑center, randomized, cross‑over diagnostic trial (NCT05214729) in adult ICU patients receiving CRRT who were preload‑independent at enrollment and monitored with a calibrated continuous cardiac index (CCI) system. The diagnostic maneuver under evaluation was a 250‑ml net ultrafiltration challenge (UFNET challenge) delivered either rapidly over 15 minutes (fast challenge) or more slowly over 30 minutes (slow challenge). Each patient underwent both challenge rates in randomized order, separated by a 24‑hour washout period.
Before and after each UFNET challenge, investigators performed a postural maneuver (PM) that evaluated preload‑dependence using relative CCI changes; the PM performed after the challenge served as the reference standard to dichotomize outcomes (responder = positive PM, non‑responder = negative PM). The primary diagnostic measure was the relative change in CCI between before and after the UFNET challenge (∆CIUFC); performance was assessed by area under the receiver operating characteristic curve (AUROC). Analyses included intention‑to‑treat (ITT) and a modified intention‑to‑treat (mITT) set excluding challenges found to be preload‑dependent at the start.
Key findings
Twenty patients were included, yielding 36 UFNET challenges: 19 fast and 17 slow. Main results were:
- Rate of preload‑dependence after UFNET (ITT): 33% (12/36; 95% CI 19–51%).
- Diagnostic performance of ∆CIUFC (ITT): AUROC 0.74 (95% CI 0.58–0.88), indicating fair diagnostic accuracy.
- After excluding five challenges retrospectively identified as already preload‑dependent at challenge start (mITT, N = 31), AUROC improved to 0.83 (95% CI 0.66–0.99), indicating good discriminatory performance.
- There was no statistically significant difference in ∆CIUFC performance between fast (15‑min) and slow (30‑min) challenges in either ITT or mITT analyses.
- Pre‑challenge CCI variation during the PM (i.e., the postural maneuver performed before the UFNET challenge) predicted de novo preload‑dependence: AUROC 0.82 (95% CI 0.65–0.98) with an optimal threshold of +5% change in CCI.
In practical terms, about one in three preload‑independent patients became preload‑dependent after removing 250 ml during CRRT. The ∆CIUFC after a small, controlled UFNET served as a useful indicator of this conversion, and a modest rise in CCI during a pre‑challenge PM (≥5%) identified patients at higher risk.
Interpretation and clinical implications
These findings suggest that a small, standardized UFNET challenge (250 ml) can function as a bedside diagnostic tool to uncover latent preload dependence in patients who appear preload‑independent at baseline. Using calibrated pulse‑contour cardiac index monitoring, clinicians can measure ∆CIUFC and interpret a significant fall in cardiac index after 250 ml removal as indicating that further UFNET may induce hemodynamic compromise.
Key implications:
- Risk stratification before larger or prolonged net ultrafiltration: a positive UFNET challenge could prompt reduced UFNET targets, closer hemodynamic monitoring, or temporary suspension of UFNET while addressing potential reversible contributors to intolerance.
- Integration with pre‑challenge PM: a CCI rise ≥5% during a PM before UFNET may identify patients at heightened risk of becoming preload‑dependent should UFNET be applied, enabling preemptive adjustments.
- Choice of removal rate: the study did not show a clear hemodynamic advantage to slower removal of 250 ml (30 min) versus faster removal (15 min), though it was underpowered to conclusively compare rates for clinical endpoints; clinicians should still individualize UFNET rate based on overall hemodynamics and tolerance.
Strengths
The trial employed a randomized cross‑over design, which reduces between‑patient confounding by letting each participant serve as their own control. The use of calibrated continuous cardiac index monitoring provided high‑resolution hemodynamic data, and the study focused on a pragmatic, small UFNET volume that is simple to implement at the bedside.
Limitations
Important limitations temper the conclusions and their generalizability:
- Small sample size and single‑center conduct limit statistical power and external validity. The subgroup analyses comparing fast and slow challenges were likely underpowered.
- The diagnostic reference standard—the PM performed after the UFNET challenge—relies on CCI measurement and the assumption that PM‑induced CCI changes reflect true preload responsiveness; different monitoring technologies or PM methods may alter thresholds and performance.
- Five challenges were excluded post‑hoc because patients were retrospectively considered preload‑dependent at baseline, an approach that may introduce bias and limit the purely prospective nature of the diagnostic assessment.
- The study evaluated a short‑term physiologic endpoint (conversion to preload‑dependence) rather than patient‑centered outcomes such as hemodynamic instability requiring intervention, organ dysfunction, duration of CRRT, ICU length of stay, or mortality.
- CCI monitoring requires an arterial line and specific calibration—this technology is not universally available, and pulse contour systems differ in performance.
Expert commentary and mechanistic insight
From a physiological standpoint, removal of intravascular volume during UFNET reduces venous return and right‑sided preload. In patients whose cardiac function lies on the ascending portion of the Frank‑Starling curve (preload‑dependent), even modest reductions in preload can reduce stroke volume and cardiac output. A controlled, small UFNET dose acts like a targeted mini‑challenge in the opposite direction of a fluid bolus: rather than testing whether CO increases when preload is augmented, it tests whether CO falls when preload is decreased. This makes the UFNET challenge a conceptually coherent and direct test of tolerance to volume removal.
Clinically, a pragmatic pre‑UFNET workflow could combine a brief PM and a 250‑ml UFNET challenge to inform UFNET strategy. A pre‑UFNET PM showing a small, reproducible rise in CCI (≥5%) should prompt caution, and an immediate decline in CCI after 250 ml removal would suggest limiting or delaying planned fluid removal.
Practical recommendations
Until larger confirmatory studies are available, clinicians may consider the following cautious approach for patients on CRRT who appear preload‑independent but for whom substantial UFNET is planned:
- Perform a brief postural maneuver (if the patient’s condition permits) and observe CCI (or equivalent cardiac output measure); a CCI rise ≥5% may indicate latent preload dependence.
- If available, perform a controlled 250‑ml UFNET challenge and observe ∆CIUFC. A meaningful reduction in CCI after removal suggests intolerance to further UFNET and should prompt revision of fluid removal goals and closer hemodynamic support.
- Individualize UFNET rate and cumulative targets based on the challenge response, underlying cardiac function, vasopressor requirements, and global goals of care.
Future research directions
Key unanswered questions that warrant larger, multicenter investigations include:
- Whether a UFNET challenge‑guided strategy reduces clinically relevant outcomes (symptomatic hypotension, days on CRRT, ICU length of stay, or mortality) compared with standard care.
- Optimal UFNET challenge volume and rate, and whether thresholds differ across patient subgroups (e.g., heart failure, septic shock, ARDS).
- Comparative performance of different hemodynamic monitoring platforms and noninvasive surrogates for centers without calibrated pulse‑contour devices.
- Integration of UFNET challenge into protocols for fluid removal in post‑resuscitation or volume‑overloaded patients on CRRT.
Conclusions
Biscarrat and colleagues provide proof‑of‑concept that a simple 250‑ml UFNET challenge can identify preload‑independent ICU patients who will become preload‑dependent during CRRT. The diagnostic performance was acceptable (mITT AUROC 0.83), and a modest CCI response to a pre‑challenge postural maneuver (≥5%) identified patients at risk. While promising, these results require validation in larger, multicenter trials that link UFNET challenge results to clinical outcomes. Meanwhile, the UFNET challenge can be considered as an adjunctive bedside test to individualize fluid removal during CRRT in settings with appropriate hemodynamic monitoring capability.
Funding and trial registration
Clinical trial registration: NCT05214729. Funding sources and detailed disclosures are reported in the original publication (Biscarrat et al., Crit Care 2025).
References
1) Biscarrat C, Deniel G, Chivot M, Yonis H, Chauvelot L, Mezidi M, Richard JC, Bitker L. Diagnostic performance of a 250‑ml net ultrafiltration challenge to identify risk of preload‑dependence in critically ill patients undergoing continuous renal replacement therapy: a randomized, cross‑over trial. Crit Care. 2025 Oct 21;29(1):446. doi: 10.1186/s13054-025-05674-3. PMID: 41121412; PMCID: PMC12541968.
2) Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group. KDIGO Clinical Practice Guideline for Acute Kidney Injury. Kidney Int Suppl. 2012;2:1–138.
