Heterogeneity in the Effect of Early Goal-Directed Therapy for Septic Shock: A Secondary Analysis of Two Multicenter International Trials
Authors' conclusion
Treatment response to EGDT varied across patients in two multicenter RCTs with large benefits for some patients while others were harmed. Patient characteristics, including albumin, were most important in identifying HTE.
Study context
- Population
- Adults presenting to the ED with severe sepsis or septic shock; 81 sites in Australia, New Zealand, Hong Kong, Finland, Republic of Ireland and the US81 sites
- Comparison
- Early goal-directed therapy (EGDT) vs usual care
Methodological detailsmachine learning (local-linear random forest), derivation and validation cohorts by random split of sites, multicentre, international
- Machine learning (local-linear random forest)
- Derivation and validation cohorts by random split of sites
- Multicentre
- International
Primary result
As reported in the abstract
HTE in validation cohort, by quintile of predicted iARD · 90 days
EGDT vs usual careEstimate (95% CI)
Suggesting harm from EGDT
absolute risk difference (quintile 1) −8%(−19% to 4%)
Suggesting benefit
absolute risk difference (quintile 5) 12%(1–23%)
EGDT vs usual careEstimate (95% CI)
Suggesting harm from EGDT
relative risk reduction (quintile 1) 1.34(0.89–2.01)
Suggesting benefit
relative risk reduction (quintile 5) 0.64(0.42–0.96)
- Heterogeneity of treatment effect (HTE) confirmed: interaction between iARD prediction and treatment
- P < 0.001
Further findings
Sensitivity analyses (cohorts swapped)
- Similar findings
Contributors to HTE
- Pre-intervention albumin contributed the most to HTE
- Analyses of individual EGDT components inconclusive
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Heterogeneity in the Effect of Early Goal-Directed Therapy for Septic Shock: A Secondary Analysis of Two Multicenter International Trials. Crit Care Med. Published 23 October 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →