Sepsis is a dysregulated host response to infection that leads to life-threatening organ dysfunction and represents a major healthcare problem. Septic shock is the most severe form, characterized by increased capillary permeability and vasodilation, resulting in hypotension and tissue hypoxia. Early identification and treatment of tissue hypoperfusion are pivotal components of initial resuscitation to limit progression to multiple organ dysfunction and death. The 2021 Surviving Sepsis Guidelines recommend guiding initial resuscitation by targeting decreases in serum lactate levels in patients with elevated lactate. However, although elevated lactate levels may reflect tissue hypoxia, serum lactate is not a direct marker of tissue perfusion. Hyperlactatemia may be attributable to mechanisms other than tissue hypoperfusion, such as accelerated aerobic glycolysis driven by excessive β-adrenergic stimulation or impaired clearance (e.g., in liver failure).

The venous-to-arterial carbon dioxide partial pressure difference (CO₂ gap), which is inversely related to cardiac output, has been shown to reflect the adequacy of venous blood flow to remove CO₂ from tissues. The CO₂ gap is closely linked to microcirculatory blood flow during the early resuscitation phase of septic shock and may effectively identify persistent tissue hypoperfusion in shock states. A persistently high CO₂ gap during early resuscitation has been associated with significantly higher 28-day mortality and increased Sequential Organ Failure Assessment (SOFA) scores. Moreover, the CO₂ gap has been shown to respond to changes in cardiac output during inotrope infusion in patients with low blood flow, suggesting that its assessment could be useful for therapeutic adjustments. Therefore, there are compelling arguments to evaluate the usefulness of the CO₂ gap in guiding early resuscitation in patients with septic shock.

The investigators postulated that CO₂ gap-guided early resuscitation may be more effective in improving outcomes than lactate-guided resuscitation.

Main objective: The aim of the CARBON trial is to compare a veno-arterial CO2 difference-guided resuscitation strategy (CO2gap-guided strategy) with a lactate level-guided resuscitation on mortality in adults intensive care unit (ICU) patients fulfilling the SEPSIS-3 criteria consensus definition.

HYPOTHESIS: The investigators hypothesized that a CO2gap-guided resuscitation strategy during early septic shock would reduce mortality compared with a lactate level-guided resuscitation.

Inclusion Criteria:

• Patients aged 18 years or older AND

• Acutely admitted to a study ICU AND

• Primary diagnosis of septic shock according to the Sepsis-3 criteria and defined as:

  • A suspected or documented site of infection or positive blood culture AND
  • Acute increase of at least 2 points in the Sequential Organ Failure Assessment (SOFA) score consequent to the infection AND
  • Having a serum lactate level >2 mmol/l AND
  • Requirement of vasopressors (any dose of norepinephrine) to maintain mean arterial pressure (MAP) ≥65 mmHg despite adequate fluid resuscitation (at least 1L of IV fluid in the last 24 hours prior to screening)

Exclusion Criteria:

• Septic shock for more than 12 hours at the time of screening
• Primary cause of hypotension not due to sepsis (e.g., acute bleeding)
• Decision not to resuscitate (or to limit full care) or not to intubate taken before obtaining consent
• Death is deemed to be imminent or inevitable or patients with an underlying disease process with a life expectancy of less than 3 months
• Anticipated surgery during the first 24 hours after randomization
• Patient or their relatives' refusal to participate
• Patients participating in another RCT with interventions possibly compromising the primary outcome
• Prior enrollment in the CARBON trial
• Known to be pregnant.
• Legal protection (i.e., incompetence to provide consent and no guardian or incarceration)
• No affiliation with the French health care system

Data will be collected and entered into the electronic web-based case report form by trial or clinical trained personal blinded to the allocation group.

Statistical analyses will be performed with the statistician blinded to the allocation group.