Data on the optimal period for RRT weaning in critically ill patient are scarce. The current practice for RRT weaning is based on urine output, the threshold of which is debatable. Two recent observational studies have shown that an increase in urinary creatinine or urea concentrations is a better predictive marker of RRT weaning than urine output.

An unjustified delay in RRT weaning leads to numerous complications such as catheter-related infections, delay of the patient's functional recovery, severe ionic disorder, bleeding, and induced hemodynamic instability. It also induces an increase workload for careers and in cost without any additional benefit for the patient. Conversely, too early weaning inevitably limits the prevention on fluid accumulation that is independently associated with an increased risk of mortality and inevitably leads to resumption of RRT requiring reinsertion of dialysis catheter resulting in potential complications.

A multicentre randomized controlled trial will be then necessary and only able to identify the optimal RRT weaning strategy.

The main objective is to compare two RRT weaning strategies on RRT duration in critically ill patients with acute kidney injury: a strategy based on combined criteria (urine output + urinary parameters) as compared to a single strategy based only on urine output.

The study protocol will be an open-label, two parallel group, multicenter, randomized, controlled clinical trial, in which enrolled ICU adult patients will have RRT weaning based either on urine output alone (single strategy) or on urine output and urinary parameters (combined strategy).

When the urine output is greater than 500ml/24h, the enrollment must be performed within 24hours in 2 groups:.

" Single strategy ": In the single strategy, RRT weaning will be achieved when urine output exceeds 500ml/24h without diuretics or 2000ml/24h with diuretics use.

" Combined strategy": In the combined strategy, when urine output exceeds 500ml/day with or without diuretic use, RRT will be stopped during 48h to assess urinary indices (urinary creatinine and urea). Soon as urinary indices are higher than thresholds values (urinary creatinine > 5.2mmol/day and urinary urea > 1.35mmol/kg/day, RRT will be weaned. If they are lower, a RRT session will be perform after which the weaning process will be resume.

The primary endpoint is the number of RRT-free days at D30 with at least 7 consecutive days alive and without RRT.

RRT is an invasive high-cost treatment with potentially severe complications. However, there is a major dearth of data on RRT weaning in the ICU. The weaning strategy using only urine output is based on scarce and debatable data. The two observational studies concluded independently that the urinary concentration of creatinine or urea would be able to predict RRT weaning. These new data propose a novel combined strategy based on urine output and urinary parameters which are more specific to renal recovery and potentially better suited to determine the optimal period of RRT weaning.

This trial aims to determine the best RRT weaning strategy and so indirectly to determine the optimal duration of RRT in the ICU. Our trial methodology is able to determine clearly if one of the two weaning strategies is more effective and will lead to more health benefits and fewer risks for ICU patients.It is an open-label, two parallel group, multicenter, randomized, controlled clinical trial, in which enrolled ICU adult patients will have RRT weaning based either on urine output alone (single strategy) or on urine output and urinary parameters (combined strategy).

Critically ill patients aged 18 and over, admitted to the ICU, receiving or having received invasive mechanical ventilation and/or catecholamine infusion at least 48h, with acute kidney injury, at KDIGO 3 stage associated with oliguria at least < 200ml/ 24h before RRT initiation, treated with intermittent or continuous renal replacement therapy and resumption of urine output > 300ml/24h with or without diuretic use; will be included in the study.

Enrolled patients will be randomized into two groups, the single strategy or the combined strategy. When the urine output is greater than 500ml/24h, the enrollment must be performed within 24hours:

"Single strategy": RRT weaning will be achieved when urine output exceeds 500ml/24h without diuretics or 2000ml/24h with diuretics use.

"Combined strategy": When urine output exceeds 500ml/day with or without diuretic use, RRT will be stopped during 48h to assess urinary indices (urinary creatinine and urea). Soon as urinary indices are higher than thresholds values (urinary creatinine > 5.2mmol/day and urinary urea > 1.35mmol/kg/day), RRT will be weaned. If they are lower, a RRT session will be perform after which the weaning process will be resume.

After the RRT weaning, RRT will be resumed both groups if there is at least one of the following criteria:

• oliguria (urine outpout <300ml/24h with or without diuretic use) or anuria >72h;
• serum urea concentration >40mmol/L,
• serum potassium concentration of more than 5.5mmol/L despite medical treatment
• pH <7.15 in a context of pure metabolic acidosis (PaCO2 <35 mmHg) or in a context of mixed acidosis with PaCO2 ≥50mmHg without possibility of increasing alveolar ventilation and despite medical treatment;
• Acute pulmonary edema due to fluid overload responsible for severe hypoxemia requiring oxygen flow rate >5 l/min to maintain a SpO2 ≥ 95% or requiring an FiO2 >50% in patients on high-flow cannula oxygen therapy or invasive or non-invasive mechanical ventilation and despite diuretic therapy (equivalent to furosemide dose of 1mg/kg at least).

If several weaning are performed because RRT is resumed many times during the ICU stay, the patient will keep the assigned weaning strategy at randomization to D30.

The days elapsed between two RRT sessions are not considered as RRT-weaned days if they are less than 7 consecutive days, and are not taken into account.

The primary endpoint is the number of RRT-free days at D30 with at least 7 consecutive days alive and without RRT.

To highlight a minimal clinically difference of 2 RRT-weaned days between groups for a two-sided type I error at 5% and a statistical power greater than 90% [55], we have estimated that 600 patients (300 by group) will be necessary, with 1) variability (standard-deviation and interquartile range) of RRT-free days ranged between (median [interquartile range]) 17 [2-26] vs. 19 [5-29] , and 12 [1; 25] vs. 16 [2; 28] and 2) 30-day mortality around 35 to 45%.

The choice of the difference of 2 RRT-weaned days was determined according to clinical relevance corresponding to a reduction of at least one RRT session.

To compare two RRT weaning strategies on its duration in patients: a strategy based on combined criteria (urine output + urinary parameters) as compared to a single strategy based only on urine output, the primary analysis will be performed by Student t-test or the nonparametric Mann-Whitney if the assumptions of the t-test are not met. The homoscedasticity will be analyzed using Fisher-Snedecor test. Results will be expressed as effect-size and 95% confidence interval.

Inclusion Criteria:

• Adults (18 years of age or older)
• Admitted to the ICU
• Receiving or having received invasive mechanical ventilation and/or catecholamine infusion (epinephrine or norepinephrine or dobutamine) at least 48h
• With acute kidney injury, at KDIGO 3 (See Appendix 1) stage and associated with oliguria at least < 200ml/ 24h before RRT initiation
• Treated with intermittent or continuous renal replacement therapy
• Resumption of urine output > 300ml/24h with or without diuretic use

Exclusion Criteria:

• Preexisting Chronic kidney disease with estimated glomerular filtration rate (eGFR) < 30ml/min
• AKI caused by vascular occlusion, glomerulonephritis, vasculitis, post-renal obstruction, thrombotic microangiopathy, tumor lysis syndrome
• RRT for another cause than AKI (eg: drug intoxication,...)
• Decision to forgo life-sustaining treatment including RRT
• Cirrhosis with Child-Pugh score of C or hepatorenal syndrome
• Kidney transplantation
• Patient already enrolled in the study
• Participation in another clinical trial assessing the impact or duration of RRT
• Pregnancy in progress or planned during the study period or breastfeeding women
• Patients protected by law (Art. L1121-6 and L1121-8 of the Code de la Santé Publique) : Adult protected by law or patient under guardianship or curatorship
• Subjects not covered by public health insurance
• Absence of written informed consent from the patient or his or her proxy (if present) before inclusion or when possible when the patient has been included in an emergency setting