Septic shock is a frequent complication associated with high mortality in patients with malignancies. The best transfusion strategy (restrictive or liberal) for the resuscitation of septic shock remains a controversial issue, in relation with potentially discrepant goals of tissue oxygenation and transfusion sparing.

In this study, the investigators propose to address the efficacy of two RBC transfusion strategies (liberal or restrictive) in restoring appropriate tissue oxygenation as well as their tolerance.

The investigators designed a prospective randomized multicenter trial aimed at comparing liberal and restrictive RBC transfusion strategies applied during the first 48 hours of resuscitation in cancer patients with septic shock and anemia.

Septic shock is a frequent complication in patients with malignancies and remains affected with a mortality rate higher than 50%.

Red blood cell (RBC) transfusion remains a major issue for critically ill cancer patients who frequently display anemia as a result of malignant bone marrow involvement or imposed by cytotoxic treatments. However, our current practice of RBC transfusion in the intensive care unit (ICU) is drawn from general populations. Several case-control studies suggested that RBC transfusion was associated with higher mortality and increased incidence of ICU-acquired complications in critically ill patients. In 1999, a restrictive strategy of non-leucodepleted RBC transfusion to maintain hemoglobin above 7 g/dL was shown to be as effective as a liberal transfusion strategy aimed to maintain haemoglobin > 10 g/dL in critically ill patients. As of today, the current recommendations for RBC transfusion remain largely based on this study which excluded patients with a history of anemia. Leucodepletion that is now routinely implemented in France might be associated with fewer transfusion-related events. Indeed, some recent studies challenged the restrictive strategy and suggested that a higher transfusion threshold might be beneficial in septic patients for whom oxygen delivery is of paramount importance. Most importantly, a hemodynamic support algorithm for severe sepsis also known as early goal-directed therapy (EGDT) included a hematocrit target of 30%. The majority of EGDT-treated patients received RBC transfusion within the early 72 hours of resuscitation, thereby representing a major difference compared to standard treatment, but the prognostic value of RBC transfusion was not specifically assessed. A recent case-control study also suggested that RBC transfusion was an independent predictor of survival in patients with septic shock.

Cancer patients with septic shock and hemoglobin level < 9 g/dL initiation will be randomized to the interventional arm (liberal transfusion strategy to maintain hemoglobin > 9 g/dL) or to the control arm (restrictive transfusion strategy to maintain hemoglobin > 7 g/dL) in a 1:1 ratio.

Patients from the intervention arm will have their hemoglobin level maintained above 9 g/dL for the whole time under vasopressors, for a maximum of 28 days. After weaning from vasopressor, the transfusion threshold will be lowered to 7 g/dL as recommended by the SSC guidelines. In case of shock relapse requiring reintroduction of vasopressors, the transfusion threshold will be upgraded back to 9 g/dL until next catecholamine weaning.

In the control arm, the transfusion threshold will be 7 g/dL until ICU discharge regardless of catecholamine administration.

The primary objective of the study will be the restoration of tissue oxygenation as assessed by lactate clearance at 12 hours following randomization. The secondary endpoints will be related to restoration of tissue oxygenation at alternative time points (6h, 24h, 36h, 48h) as assessed as above, the 7-day, 28-day, in-ICU and in-hospital mortality rate, changes in organ failures over the first 48 hours and 7 days, duration for organ failure supports, the development of acute ischemic and thrombotic events (myocardial infarction, mesenteric ischemia, ischemic stroke, limb ischemia, deep venous thrombosis) over the first 7 days.

An interim analysis on the primary endpoint has been pre-planned at the end of the follow up of half of patients included.

Inclusion Criteria:

• Age > 18 years
• Evolutive (complete remission < 2 years) malignant solid tumor or chronic or acute hematological malignancy
• Septic shock defined as:
• Presumed or documented infection
• Acute circulatory failure defined as hypotension requirement of vasoactive drugs for more than one hour (norepinephrine or epinephrine ≥ 0.1 µg/kg/min)
• Tissue hypoxia defined by arterial lactate level > 2 mmol/L within 3 hours prior to inclusion
• Hemoglobin level < 9 g/dL
• Informed consent from patients or surrogates

Exclusion Criteria:

• Acute life-threatening bleeding
• Ongoing acute coronary syndrome or any other acute ischemic condition
• End-of-life decisions at the time of ICU admission
• Refusal of transfusions for personal beliefs
• Lack of social security coverage
• Pregnancy or breastfeeding