Intravenous fluids are one of the keystones in the initial management of patients with septic shock, but they inevitably lead to a fluid overload, which is associated with poor outcome.

So far no studies have evaluated the interest of a restrictive strategy for managing fluid intake targeting all non-resuscitative fluids (fluids for maintenance and drug dilution as well as nutrition) and especially the impact of this restrictive strategy on fluid overload.

The hypothesis of this research is that an optimised restrictive strategy targeting all non-resuscitative fluids in patients hospitalised in the intensive care unit for septic shock, will have an impact on fluid balance in these patients.

Sepsis and its most severe form, septic shock, are the cause of nearly 30% of admission in intensive care unit (ICU). The related mortality rate varies from 20 to 60% according to the studied population and the type of the study. With the concomitant and rapid treatment of the source of infection, hemodynamic optimization with fluid infusion (fluid resuscitation) is the cornerstone of management of patients in septic shock. Nevertheless, fluid overload has been shown to be associated with poor outcome. The negative effects of a positive fluid balance on ICU outcomes for septic patients has been well established.

In patients with acute respiratory distress syndrome, a restrictive fluid therapy has been associated with shorter use of mechanical ventilation. In patients with acute renal failure, fluid overload is associated with longer need of renal replacement therapy. In patients with septic shock numerous non-randomised studies reported that a positive fluid balance at Day 3, Day 5 was associated with an increased risk of mortality.

Actually, in patients with septic shock the part of resuscitation fluid is only the third of the total fluid input that can attain 25 liters at day 3. Some studies reported more specifically the role of the different types of fluid in the fluid balance at D3 and D5. In a retrospective single centre study including 14654 ICU patients over 10 years, the authors distinguished resuscitation, replacement and maintenance fluids, blood products, nutrition and fluid creep (to concentrate electrolytes, keep venous access and to infuse continuous medications). They reported that maintenance fluids and fluid creep represent more than 50% of fluid input during the first 5 ICU days with significant sodium and chloride loads. In a retrospective multi-centre study involving 400 patients mechanically ventilated > 24 h in 10 ICUs, the authors reported that maintenance and fluid creep represented more than 50% of the fluids infused during the first 3 ICU days. In this study, fluid balance at Day3 was an independent risk factor for 30-day mortality.

All these findings could lead to hypothesize that a restrictive strategy in fluid input for the maintenance of fluid and creep fluid could be associated with a better outcome without impairing the resuscitation phase.

To our knowledge, to date, no study has focused on a restrictive strategy targeting non-resuscitative fluid input to optimize fluid balance in patients with septic shock. OPTIFLUID will be the first randomised controlled trial with the objective of reducing fluid balance through a fluid restriction targeted at all fluids other than resuscitative fluids.

The only existing interventional studies on targeted restriction were limited to resuscitative intravenous fluids, and they failed showing a benefit.

By optimizing fluid therapy, our approach is likely to improve patient-centred outcomes, but in this pilot study we focus only on the fluid balance and the safety.

Inclusion Criteria:

• Patient suffering from an infection already documented or suspected
• Patient with organ dysfunction defined as an acute change in SOFA score ≥ 2 points
• Patient with refractory hypotension requiring vasopressors to maintain a mean arterial pressure ≥ 65 mm Hg in spite of resuscitation with an adequate volume of intravenous fluids (according to the recommendations of the Surviving Sepsis campaign). The treatment by vasopressor must have been started less than 24 hours before inclusion and randomization.
• Patient with hyperlactatemia (blood lactate > 2 mmol / L). This criterion must be fulfilled at least once for the purpose of the diagnosis within the 24 hours preceding inclusion and randomization.
• Patient or close relative / legal representative / family member / curator / tutor must have given written informed consent and signed the consent form for the patient included in an emergency situation
• Patient must be covered by a health insurance scheme.
• Adult patient (≥18 years) under 85 years old (< 85).

Exclusion Criteria:

• Patient who has had a previous episode of septic shock requiring vasopressor administration or mechanical ventilation or renal replacement therapy during the current ICU stay.
• A patient whose life expectancy is shortened due to his/her initial health condition or a moribund patient whose life expectancy is less than 48 hours or whose decision to limit comprehensive care was made before including the patient.
• Patient who has had a recent cardiac arrest (during current hospital stay)
• Patient requiring emergency renal-replacement therapy (hyperkalemia [potassium > 6.5 mmol/l] refractory to medical treatment and/or metabolic acidosis [pH <7.15 and partial pressure of carbon dioxide (pCO2) <45 mmHg] refractory to medical treatment and/or pulmonary edema in anuric patients who do not respond to diuretic therapy).
• Patient with KDIGO 3 acute kidney injury and likely to require renal-replacement therapy within the next 24 hours, as determined by the patient's clinician in charge
• Patients with end-stage chronic renal failure or patients already undergoing chronic dialysis
• Severely malnourished patient in whom nutritional support is an emergency procedure (body mass index <18 kg / m2)
• Patient participating in or having participated in an interventional study with similar patient outcome in the previous 3 months.
• Patient in an exclusion period determined by another study.
• Patient under legal protection.
• Pregnant patient (positive serum pregnancy test).