Description

Fluid overload is a frequent and clinically relevant complication in critically ill patients with acute kidney injury (AKI). In this setting, continuous renal replacement therapy (CRRT) is frequently used not only for solute control but also as a strategy for controlled fluid removal through net ultrafiltration (UFNET). Although UFNET is central to de-resuscitation, the individual tolerance to fluid removal is highly variable and is not fully captured by blood pressure monitoring alone.

The concept of ultrafiltration intolerance remains poorly standardized. In clinical practice, intolerance is often recognized only after overt hemodynamic instability occurs, such as hypotension, escalation of vasoactive support, or interruption of fluid removal. However, reductions in cardiac output, impaired tissue perfusion, or worsening venous congestion may precede overt hypotension. Therefore, a multiparametric assessment may allow earlier identification of patients at risk.

This is a prospective, observational, analytical study in critically ill adult patients with AKI receiving CRRT with prescribed UFNET. The study does not assign or modify therapeutic interventions. CRRT modality, anticoagulation strategy, UFNET prescription, vasopressor management, fluid administration, and all other clinical decisions will remain under the responsibility of the treating clinical team according to routine care.

The study will characterize the physiological response to UFNET using a structured multiparametric monitoring approach. Recorded domains will include conventional macrohemodynamic variables, vasoactive support, selected advanced hemodynamic variables when available, focused cardiac ultrasound, venous congestion assessment, peripheral perfusion parameters, fluid balance variables, and selected biochemical markers. Functional hemodynamic maneuvers may be performed when feasible and clinically safe.

Data will be collected prospectively using a standardized case report form. Variables will be recorded at predefined time points before and during the early phase of UFNET, with additional off-schedule recordings if clinical signs compatible with intolerance occur. Source data will be obtained from the electronic or paper medical record, bedside monitoring systems, CRRT prescription and treatment records, laboratory results, and ultrasound assessments performed as part of clinical evaluation.

A data dictionary will define each variable, including its source, units, coding, and expected physiological range when applicable. Data quality procedures will include review of completeness, range checks, consistency checks between related variables, and verification of clinically implausible values against source records. The research team will periodically review entered data to identify missing, inconsistent, or out-of-range values.

Data will be anonymized before analysis. No directly identifiable patient information will be stored in the final analytical database. Access to the study database will be restricted to authorized study investigators. Data will be stored using password-protected institutional or investigator-controlled systems according to local confidentiality and ethical requirements.

The planned sample size is 128 participants, including an estimated analytical sample of 116 participants and an approximate 10% over-recruitment to account for incomplete data, missing assessments, or inability to definitively adjudicate the outcome. The sample size was based on an analytical case-control approach aimed at identifying factors associated with ultrafiltration intolerance.

Missing data will be evaluated before statistical analysis. Variables with substantial missingness may be excluded from inferential analyses. For variables with acceptable levels of missingness, available-case analysis will be performed. The extent and pattern of missing data will be reported.

Statistical analysis will include descriptive statistics, comparison between patients who develop ultrafiltration intolerance and those who do not, and exploratory modeling to identify factors independently associated with intolerance. Continuous variables will be summarized using median and interquartile range or mean and standard deviation, as appropriate. Categorical variables will be summarized as frequencies and percentages. Group comparisons will be performed using appropriate parametric or non-parametric tests according to data distribution. Multivariable logistic regression may be used to explore independent predictors, with covariate selection based on clinical relevance and number of events.

The overall objective of this study is to improve the characterization of ultrafiltration intolerance during CRRT and to generate evidence that may support individualized, physiology-guided fluid removal strategies in critically ill patients with AKI.