- Design
- Post hoc biomarker analysis of a double-blind factorial randomised trial
- Population
- 173 adults with septic shock in UK ICUs (VANISH)
- Primary outcome
- Interaction of proenkephalin with treatment on kidney-failure-free days and 28-day mortality
- Effect
- High PENK: kidney-failure-free days 11 (hydrocortisone) vs 26 (placebo); interaction p = 0.01
VANISH randomised patients with septic shock in UK ICUs to vasopressin or noradrenaline and, in a factorial design, hydrocortisone or placebo. This post hoc analysis measured plasma proenkephalin A, a marker of glomerular filtration, in 173 participants and split them at 89 pmol/L.
High proenkephalin was linked to 28-day mortality (OR 3.24, 95% CI 1.62–6.74). It did not modify the effect of vasopressin. But among the 111 patients who reached the maximum vasopressor dose, it did appear to modify the effect of hydrocortisone on kidney-failure-free days: with high proenkephalin, median days were 11 with hydrocortisone against 26 with placebo, while with low proenkephalin both groups had 28 (interaction p = 0.01). Mortality followed the same pattern — 50% with hydrocortisone and 32% with placebo in the high group — but that interaction was not significant (p = 0.14).
Corticosteroid trials in septic shock have disagreed, and heterogeneous treatment effects are a plausible reason. This analysis is small, exploratory and post hoc, so it should change nothing at the bedside today. It does suggest kidney function at baseline may be worth testing as a predictor of steroid response in future trials.
- Continue to use hydrocortisone in septic shock according to current guidelines; this analysis does not change them.
- The signal of harm with hydrocortisone came from a small post hoc subgroup of 111 patients.
- Proenkephalin is not a routine test and should not guide steroid decisions outside research.
- Expect baseline kidney function to feature in future trials of steroids in sepsis.
Why it matters
It offers one testable reason why steroid trials in septic shock keep disagreeing.
Don't overread it
This is an exploratory post hoc subgroup; it does not show that hydrocortisone harms patients with impaired kidney function.
The statistics, in plain English
An interaction p value of 0.01 in one of several post hoc comparisons, in a subgroup of 111 patients, is easily a chance finding. The mortality pattern looks dramatic, but the interaction was not significant, so it cannot be relied on.
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