- Design
- Multicohort retrospective study, US health system 2012–2024
- Population
- 15,732 adults with septic shock (Sepsis-3 and CDC criteria)
- Primary outcome
- Incidence of refractory septic shock and hospital mortality
- Effect
- 21.8% refractory; mortality 64.4%; adjusted OR 4.87 (95% CI 4.46–5.31)
This retrospective multicohort study, published in June 2026, applied new consensus criteria for refractory septic shock to 15,732 adults with septic shock in a large US health system (2012 to 2024). Refractory shock was defined as a noradrenaline-equivalent dose above 0.5 µg/kg/min together with lactate above 2 mmol/L.
Refractory septic shock occurred in 21.8% (218 per 1,000 with septic shock; 95% CI 211 to 224). Hospital mortality in this group was 64.4%, and risk-adjusted odds of death were 4.87 times higher than in septic shock without these features (95% CI 4.46 to 5.31).
The criteria are simple and bedside-measurable, which makes them useful for escalation decisions, research enrolment and conversations with families. They identify a group in whom standard care is failing; adjuncts such as vasopressin and hydrocortisone, if not already started, and source control need urgent review.
- Track noradrenaline-equivalent dose in µg/kg/min and lactate together; both are needed for the criteria.
- Do not wait for noradrenaline above 0.5 µg/kg/min to consider vasopressin and hydrocortisone; current sepsis guidance suggests considering them at lower doses.
- Re-examine source control, including drains, lines and surgical review.
- Use the high mortality to open early, honest conversations with families about goals of care.
Why it matters
It turns 'refractory' from a feeling into a defined, measurable threshold that flags high mortality.
Don't overread it
Retrospective US data; the criteria mark high risk but do not show which treatment changes it.
The statistics, in plain English
An odds ratio of 4.87 means the odds of death were nearly five times higher, after adjusting for age, comorbidity and organ failure. This is association from routine records; the criteria identify risk but do not show which treatments change it.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for emergency & critical care, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free