The timing of renal replacement therapy (RRT) in the context of severe acute kidney injury (AKI) is one the most debated issues in critical care medicine. The Artificial Kidney Initiation in Kidney Injury (AKIKI) was the first large prospective multicenter randomized trial published on this topic. This study conducted by our team (published in the New England Journal of Medicine, July 2017) showed no significant difference between an early and delayed RRT initiation strategy in term of mortality. Nearly 50% of patients escaped RRT in the delayed strategy and this strategy was associated with less catheter-related infections and faster renal function recovery. Two (serum urea concentration \>40 mmol/l and oliguria/anuria for more than 72 hours) of the 5 criteria which mandated RRT in the delayed strategy are still open to debate since they have never been shown to put patient at danger. To go further into our investigation of RRT criteria,the investigators designed a study that would compare the "delayed strategy" used in AKIKI that can now be considered as "standard" with another in which RRT is delayed for a longer period in the absence of a potentially severe complication (such as hyperkalemia or severe overload pulmonary edema).
The AKIKI 2 study will be a prospective, multicenter, open-label, two-arm randomized trial.
The study will involve 2 stages (an observational stage and a randomization stage) At any time during these two stages, the occurrence of a potentially severe condition defined below will mandate strong consideration for immediate RRT initiation unless a medical treatment can very rapidly resolve the situation.
Criteria that mandate strong consideration for RRT initiation at any time during the 2 stages of the study
1. Serum potassium concentration \>6 mmol/l
2. Serum potassium concentration \>5.5 mmol/l persisting despite medical treatment
3. Arterial blood pH \<7.15 in a context of pure metabolic acidosis (PaCO2 \<35 mmHg) or in a context of mixed acidosis with a PaCO2 \>50 mmHg without possibility of increasing alveolar ventilation
4. Acute pulmonary edema due to fluid overload despite diuretic therapy leading to severe hypoxemia requiring oxygen flow rate \>5 l/min to maintain SpO2\>95% (oxygen saturation) or FiO2\>50% (inspired oxygen fraction) under invasive or non-invasive mechanical ventilation
* Observational Stage: All patients receiving (or who have received for the present episode) intravenous catecholamines and/or invasive mechanical ventilation and presenting with AKI classification stage 3 of KDIGO classification \[defined by at least one of the following criteria: serum creatinine concentration of more than 4 mg/dl (354 µmol/liter) or greater than 3 times the baseline creatinine level, anuria (urine output of 100 ml/day or less) for more than 12 hours, oliguria (urine output below 0.3 ml/kg/h or below 500 ml/day) for more than 24 hours\] and who have no potentially severe condition mandating strong consideration for immediate RRT initiation as described above will be included in an observational study. Clinical and metabolic conditions will be closely monitored and RRT consideration will be mandatory if one or more of these conditions (see above) occurs.
* Randomization stage: Patients presenting, either immediately after inclusion or during their follow-up in the observational study, one or both of following criteria: a serum urea concentration \>40 mmol/l and/or an oliguria/anuria for more than 72 hours even in the absence of an above-mentioned potentially severe condition will be randomly allocated to one of the two study treatment arms. One arm is termed "standard strategy" (which was the delayed arm of AKIKI 1) and the other a "delayed RRT strategy".
"Standard strategy": RRT is initiated within 12 hours after documentation of serum urea concentration \>40 mmol/l and/or an oliguria/anuria for more than 72 hours. The timing of initiation will be recorded and RRT sessions will be performed until criteria for cessation are observed (see below).
"Delayed strategy": RRT will be initiated only if one or more of above-mentioned potentially severe situations (see "Criteria that mandate strong consideration for RRT initiation at any time") occur (identical to those used during the observational study) or if serum urea concentration reaches 50 mmol/L (this level being chosen in order to avoid extreme elevations of serum urea levels as mentioned above). As already explained, the physician in charge will be allowed to try a medical treatment of hyperkalemia, acidosis or pulmonary edema and the decision to start RRT or not will be taken by him, according to his usual practice. The duration of anuria will not constitute a criterion per se for RRT initiation. When required, RRT will be performed with the same modalities and stopped according to the same criteria as in the standard strategy, with special care to avoid dialysis disequilibrium syndrome.
The choice of RRT modality (intermittent or continuous technique) will be left at the study site discretion. Several RRT modalities can be used in the same patient, according to attending physician's indication. Duration of and interval between sessions, and device settings as well as modality of anticoagulation are left at the investigator's discretion. However, RRT will be prescribed and monitored according to national guidelines in order to ensure optimal efficacy of RRT.
In case of RRT initiation in a context of high serum urea concentration (\> 40 mmol/l), prevention of dialysis disequilibrium syndrome will be recommended. The management will be left at clinician's discretion and will include one or several of the following measures :
* Slow, gentle initial hemodialysis (dialysis time \<2 hours and low blood flow rate)
* Increasing dialysate sodium levels
* Administration of osmotically active substance by using a high-glucose-concentration dialysate or administering hypertonic glucose in the venous line of the dialyser during dialysis.
All study centers have extensive experience in both acute kidney injury management and RRT techniques.
Renal replacement therapy discontinuation will be contemplated when spontaneous diuresis \>500 ml/24h, and highly recommended if diuresis is \>1000 ml/24h without diuretic administration or \>2000 ml/24h, in patients receiving diuretics.
Renal replacement therapy cessation will be mandatory if diuresis (as defined above) is present and serum creatinine level decreases spontaneously.
If improvement of renal function is insufficient to achieve a spontaneous decrease in creatinine level and/or if diuresis becomes lower than 1000 ml/24h without diuretics (or lower than 2000 ml/24h under diuretics), RRT will be resumed.
Renal function recovery will be defined according to urine output as mentioned above and RRT cessation without need for resumption in the following 7 days.
RRT will be initiated only if one or more of above-mentioned potentially severe situations occur (identical to those used during the observational study. The physician in charge will be allowed to try a medical treatment.) or if serum urea concentration reaches 50 mmol/L (this level being chosen in order to avoid extreme elevations of serum urea levels as mentioned above). Patients will not receive RRT whatever the duration of anuria/oliguria if none of the above-mentioned indications for RRT is present. When required, RRT will be performed with the same modalities and stopped according to the same criteria as in the "standard" RRT strategy, with special care to avoid dialysis disequilibrium syndrome.
The decision to initiate RRT in the "delayed strategy" arm of the trial will have to be approved by the attending physician(s) involved in patient's care in order to make sure that it corresponds to her/his usual practice.
