Respiratory Trials·org

Topic

Sepsis & septic shock

Fluids, vasopressors, steroids, and the rise and fall of protocolised care.

Protocolised resuscitation is dead. A single-centre trial of early goal-directed therapy appeared to cut mortality dramatically (Rivers EGDT) and reshaped sepsis care worldwide, then failed to replicate three times over (ProCESS, ARISE, ProMISe). The likeliest explanation is that usual care caught up: early recognition, prompt antibiotics and fluids became standard *because* of the original trial, leaving nothing for the protocol to add.

Fluid strategy matters less than the argument suggests. Neither restriction in established septic shock (CLASSIC) nor at the point of sepsis-induced hypotension (CLOVERS) changed mortality — and the CLOVERS authors were careful to say 'or higher'. More promising is what to resuscitate *towards*: targeting capillary refill time, which needs nothing but a finger and a stopwatch, outperformed usual care (ANDROMEDA-SHOCK-2).

Steroids remain genuinely unsettled. Two large trials published in the same issue disagreed: hydrocortisone alone did not reduce mortality (ADRENAL) while hydrocortisone with fludrocortisone did (APROCCHSS). Fludrocortisone, bolus versus infusion, and a sicker population are the candidate explanations, and none has been tested head to head.

Vasopressors are chosen on physiology rather than evidence of survival benefit: neither vasopressin added to catecholamines (VASST) nor angiotensin II in refractory shock (ATHOS-3, whose endpoint was blood pressure) has shown a mortality effect.

Antibiotics should be early and short. Every hour of delay after hypotension was associated with higher mortality (Kumar 2006) — retrospective, and no one will randomise delay. Seven days matched fourteen for bloodstream infection in 3608 patients (BALANCE), and daily procalcitonin guidance shortens courses relative to usual care (ADAPT-Sepsis) — though whether it adds anything to simply stopping at seven days is untested.

Protocolised early resuscitation

Trial Year Therapy Effect Finding
ProMISe 2015 Early goal-directed therapy Strict EGDT gave identical 90-day mortality at higher cost
ARISE 2014 Early goal-directed therapy Early goal-directed therapy did not reduce 90-day mortality
ProCESS 2014 Protocol-based resuscitation Protocolised resuscitation did not beat usual care in septic shock
Rivers EGDT 2001 Early goal-directed therapy A protocolised resuscitation bundle appeared to cut mortality dramatically in early septic shock

Choice of vasopressor

Trial Year Therapy Effect Finding
ATHOS-3 2017 Angiotensin II Angiotensin II raised blood pressure in catecholamine-refractory shock
VASST 2008 Low-dose vasopressin Adding low-dose vasopressin to catecholamines did not improve survival

Duration of antibiotic therapy

Trial Year Therapy Effect Finding
ADAPT-Sepsis 2025 Procalcitonin-guided antibiotic duration Procalcitonin-guided stopping shortened antibiotic courses without increasing deaths
BALANCE 2025 Seven-day antibiotic course A week of antibiotics matched a fortnight for bloodstream infection

How much fluid

Trial Year Therapy Effect Finding
CLOVERS 2023 Restrictive fluid strategy with early vasopressors Neither restrictive nor liberal early fluid strategy was better in sepsis-induced hypotension
CLASSIC 2022 Restrictive fluid strategy Restricting fluid in established septic shock did not reduce mortality

Steroids in septic shock

Trial Year Therapy Effect Finding
ADRENAL 2018 Hydrocortisone Hydrocortisone resolved shock faster without reducing mortality
APROCCHSS 2018 Hydrocortisone plus fludrocortisone Hydrocortisone with fludrocortisone reduced 90-day mortality by 6 percentage points

Invasive monitoring in shock

Trial Year Therapy Effect Finding
Deferred arterial catheter 2025 Deferred arterial catheterisation Deferring the arterial line in shock was as safe as inserting one early

Timing of antibiotics

Trial Year Therapy Effect Finding
Kumar 2006 retrospective cohort 2006 Early effective antimicrobial therapy Every hour of delay to effective antibiotics in septic shock was associated with higher mortality

What to resuscitate towards

Trial Year Therapy Effect Finding
ANDROMEDA-SHOCK-2 2025 Capillary refill time-targeted resuscitation Resuscitating towards capillary refill time beat usual care on a hierarchical composite

Key references

Guidelines, standards and reviews for this topic.