Starting vasopressin within three hours of high-dose noradrenaline was associated with lower mortality

Design

Target trial emulation using two ICU electronic health record databases

Population

3,810 adults with septic shock whose noradrenaline reached ≥0.25 µg/kg/min before vasopressin

Primary outcome

28-day mortality

Effect

48.1% vs 53.0%; risk difference −5.0 points (95% CI −6.8 to −3.2); HR 0.82 (0.78–0.85)

This target trial emulation used two large US intensive care databases to compare starting vasopressin within 3 hours against 3 to 6 hours after noradrenaline reached 0.25 µg/kg/min in adults with septic shock. 3,810 patients were eligible, and a clone-censor-weight method was used to mimic randomisation.

Weighted 28-day mortality was 48.1% with ultra-early initiation and 53.0% with early initiation, a difference of 5 percentage points (HR 0.82). Ultra-early vasopressin was also associated with less renal replacement therapy (HR 0.68), though not with less acute kidney injury.

The finding supports adding vasopressin promptly once noradrenaline requirement crosses a moderate dose, consistent with how many guidelines already frame it. But this is observational data, however carefully analysed, and unmeasured differences between patients started early and late — including how sick clinicians judged them to be — could explain part of the result.

• Consider adding vasopressin when noradrenaline reaches about 0.25 µg/kg/min rather than escalating noradrenaline alone.
• Avoid letting the decision drift for hours once the threshold is reached.
• Reassess volume status, source control and steroids alongside vasopressor escalation.
• Monitor for digital and mesenteric ischaemia after starting vasopressin.

Why it matters

It suggests the timing of the second vasopressor, not only the choice, may matter.

Don't overread it

This is an emulated trial from observational data and cannot exclude confounding by illness severity.

The statistics, in plain English

A target trial emulation analyses routine data as if it were a randomised trial, defining eligibility, timing and comparison groups in advance. It reduces some biases but cannot balance factors that were not recorded, which is what randomisation does.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.