Skip to main content

Foundational Trials

ADRENAL

Image: Shutterstock / luchschenF

Paper

Publication

  • Title: Adjunctive Glucocorticoid Therapy in Patients with Septic Shock
  • Acronym: ADRENAL
  • Year: 2018
  • Journal published in: New England Journal of Medicine
  • Citation: Venkatesh B, Finfer S, Cohen J, et al. Adjunctive glucocorticoid therapy in patients with septic shock. N Engl J Med. 2018;378(9):797-808.

Context & Rationale

  • Background
    • Low-dose corticosteroids have long been used in septic shock to treat catecholamine-refractory vasoplegia and to address concerns about critical-illness-related corticosteroid insufficiency.
    • Earlier randomised trials and meta-analyses suggested more rapid shock reversal with corticosteroids, but mortality effects were inconsistent and safety concerns persisted (e.g., hyperglycaemia, neuromuscular weakness, infection).
    • Consequently, practice varied internationally, and guideline recommendations were conditional and dependent on shock refractoriness.
  • Research Question/Hypothesis
    • In adults with septic shock receiving invasive mechanical ventilation and vasopressors, does adjunctive intravenous hydrocortisone (200 mg/day by continuous infusion) reduce 90-day all-cause mortality compared with placebo?
  • Why This Matters
    • Corticosteroids are inexpensive, widely available, and commonly administered in septic shock; even small benefits (or harms) would have major population impact.
    • A large, blinded trial was required to define whether the primary value proposition is survival benefit, faster shock resolution, or both.
    • Clarifying effect size and safety was essential for harmonising international guidance and reducing unwarranted practice variation.

Design & Methods

  • Research Question: Does adjunctive hydrocortisone (200 mg/day continuous infusion) reduce 90-day mortality in mechanically ventilated adults with septic shock compared with placebo?
  • Study Type: Multicentre, international, randomised, double-blind, placebo-controlled, parallel-group trial conducted in 69 medical-surgical ICUs (Australia, New Zealand, United Kingdom, Saudi Arabia, Denmark).
  • Population:
    • Adults (≥18 years) in ICU with septic shock (suspected/proven infection, ≥2 SIRS criteria, vasopressor/inotrope-dependent hypotension despite fluids) and receiving invasive mechanical ventilation.
    • Required continuous vasopressor/inotrope infusion for ≥4 hours at randomisation; randomisation within 24 hours of meeting shock criteria.
    • Key exclusions: no clinician equipoise regarding corticosteroids; systemic glucocorticoids already given or clearly indicated; etomidate during the current admission; pregnancy; prior enrolment in the trial.
  • Intervention:
    • Hydrocortisone sodium succinate 200 mg/day by continuous intravenous infusion (delivered as two consecutive 12-hour infusions).
    • Continued until death, ICU discharge, or day 7 (whichever occurred first).
  • Comparison:
    • Matched placebo infusion delivered using the same schedule, preparation, and masking.
    • All other care (antimicrobials, haemodynamic targets, ventilation strategy, glycaemic management) per treating team and local standards.
  • Blinding: Double-blind (patients, clinicians, investigators, and outcome assessors); identical study drug kits prepared/dispensed by pharmacy.
  • Statistics: 3800 patients were required to detect a 5 percentage-point absolute reduction in 90-day mortality (from 33% to 28%) with 90% power at a two-sided 5% significance level (allowing 1% withdrawal); primary analysis was modified intention-to-treat with prespecified adjusted regression models. 12
  • Follow-Up Period: 90 days for the primary outcome (all-cause mortality).

Key Results

This trial was not stopped early. Recruitment reached the prespecified sample size; interim analyses did not trigger stopping for benefit or harm.

Outcome Hydrocortisone Placebo Effect p value / 95% CI Notes
90-day all-cause mortality (primary) 511/1832 (27.9%) 526/1826 (28.8%) OR 0.95 95% CI 0.82 to 1.10; P=0.50 Prespecified adjusted logistic regression.
28-day all-cause mortality 410/1841 (22.3%) 448/1840 (24.3%) OR 0.89 95% CI 0.76 to 1.03; P=0.13 Not statistically significant.
Time to resolution of shock Median 3 days (IQR 2–5) Median 4 days (IQR 2–9) HR 1.32 95% CI 1.23 to 1.41; P<0.001 Faster vasopressor independence; meets Bonferroni threshold (P≤0.005).
Time to ICU discharge Median 10 days (IQR 5–30) Median 12 days (IQR 6–42) HR 1.14 95% CI 1.06 to 1.23; P<0.001 Earlier ICU discharge; meets Bonferroni threshold (P≤0.005).
Time to cessation of initial mechanical ventilation Median 6 days (IQR 3–18) Median 7 days (IQR 3–24) HR 1.13 95% CI 1.05 to 1.22; P<0.001 Meets Bonferroni threshold (P≤0.005).
Blood transfusion 683/1848 (37.0%) 773/1855 (41.7%) OR 0.82 95% CI 0.72 to 0.94; P=0.004 Meets Bonferroni threshold (P≤0.005).
New-onset bacteraemia or fungaemia 262/1853 (14.1%) 262/1860 (14.1%) OR 1.00 95% CI 0.86 to 1.16; P=0.96 No signal of increased bloodstream infection.
Any adverse event related to trial regimen 21 (1.1%) 6 (0.3%) Not reported P=0.001 Myopathy 2 vs 0; GI bleeding 3 vs 0; “other” 16 vs 5; anaphylaxis 0 vs 0.
Any serious adverse event related to trial regimen 4 (0.2%) 2 (0.1%) Not reported P=0.41 Myopathy 2 vs 0; ischaemic bowel 1 vs 0; serious infection 0 vs 1; bleeding/abdominal-wound dehiscence 0 vs 1 each; anaphylaxis 0 vs 0.
  • Hydrocortisone did not reduce 90-day mortality (OR 0.95; 95% CI 0.82 to 1.10; P=0.50), but confidence intervals are compatible with modest benefit or harm.
  • Hydrocortisone accelerated shock resolution (median 3 vs 4 days; HR 1.32) and ICU discharge (median 10 vs 12 days; HR 1.14), supporting a “faster recovery” rather than “survival” signal.
  • Adverse events attributed to the regimen were uncommon but more frequent with hydrocortisone (1.1% vs 0.3%); serious adverse events were rare.

Internal Validity

  • Randomisation and allocation:
    • Central web-based randomisation (1:1), stratified by admission type (medical vs surgical) and trial site.
    • Allocation concealment maintained through identical masked study-drug kits dispensed by pharmacy; double blinding reduced selection and performance bias.
  • Dropout/exclusions after randomisation:
    • 3800 patients were randomised; 90-day vital status was ascertained in 3658 patients (1832 hydrocortisone vs 1826 placebo).
    • Primary analyses therefore relied on an “available data” modified intention-to-treat population; missingness was small and numerically balanced.
  • Performance/detection bias:
    • Blinding was feasible (identical infusion preparations) and outcomes were largely objective (mortality, time to shock resolution, ICU discharge), limiting detection bias.
  • Protocol adherence and separation of the variable of interest:
    • Trial drug delivery: hydrocortisone 200 mg/day vs placebo 0 mg/day (both as continuous infusion).
    • Mean duration of study infusion: 5.1 ± 2.0 days (hydrocortisone) vs 5.6 ± 2.0 days (placebo); compliance 95.2% vs 94.9%.
    • Non-trial systemic corticosteroids (days 1–14): 137/1853 (7.4%) vs 163/1860 (8.8%), supporting good treatment separation with limited contamination.
  • Baseline characteristics and illness severity:
    • Groups were well balanced: APACHE II median 25 (IQR 19–31) in both groups; SOFA median 10 (IQR 7–13) in both.
    • Haemodynamics at randomisation were similar: mean arterial pressure 66.8 ± 12.9 mmHg vs 67.0 ± 12.3 mmHg; noradrenaline dose 0.35 ± 0.50 vs 0.33 ± 0.46 μg/kg/min.
  • Heterogeneity:
    • 69 ICUs across five countries improves precision and reduces centre effects; the primary outcome effect was reported as similar across six prespecified subgroups.
  • Timing and dose:
    • Time from shock onset to randomisation: 20.9 ± 91.9 hours (hydrocortisone) vs 21.2 ± 83.4 hours (placebo), indicating variable enrolment timing with substantial dispersion.
    • Fixed-dose hydrocortisone (200 mg/day) reflects common practice; the continuous-infusion strategy may reduce peak-trough variability but may not replicate intermittent bolus regimens.
  • Outcome assessment and statistical rigour:
    • Primary analysis used prespecified adjusted regression models; protocol and statistical analysis plan were published before trial completion. 12
    • Secondary outcomes were numerous; nominal statistical significance should be interpreted alongside multiplicity considerations (reported separately in the trial supplement).

Conclusion on Internal Validity: Overall, internal validity appears strong given concealed randomisation, robust blinding, objective outcomes, and high protocol adherence, with only modest limitations from incomplete ascertainment of 90-day vital status in a small proportion of randomised patients.

External Validity

  • Population representativeness:
    • Represents a typical high-acuity ICU septic shock cohort: mean age ~62 years, APACHE II 25, vasopressor-dependent shock with mandatory invasive ventilation.
    • Important exclusions (e.g., etomidate exposure, clear steroid indication, lack of equipoise) may reduce applicability to some real-world practice contexts.
  • Applicability to practice and health systems:
    • The intervention is simple and scalable (200 mg/day hydrocortisone infusion) and aligns with regimens commonly used in resource-rich ICUs.
    • Generalisability is strongest for ventilated, vasopressor-dependent septic shock; applicability is less direct for non-ventilated shock, earlier/lower-severity sepsis, or settings where hydrocortisone is routinely paired with fludrocortisone.

Conclusion on External Validity: External validity is moderate-to-high for mechanically ventilated adults with vasopressor-dependent septic shock managed in high-resource ICUs, but narrower for broader sepsis populations and alternative steroid strategies.

Strengths & Limitations

  • Strengths:
    • Large sample size (n=3800), multicentre and international, enhancing precision and reducing single-centre bias.
    • Double-blind, placebo-controlled design with clinically relevant, objective endpoints.
    • Pragmatic delivery within routine ICU care, improving clinical interpretability.
    • Prospectively published protocol and statistical analysis plan. 12
  • Limitations:
    • Eligibility required invasive mechanical ventilation and excluded etomidate exposure, narrowing applicability.
    • Primary outcome ascertainment was incomplete in a minority of randomised patients due to consent/availability of follow-up data.
    • Hydrocortisone was tested without routine fludrocortisone; results may not generalise to combined glucocorticoid–mineralocorticoid strategies.
    • Multiple secondary endpoints raise multiplicity concerns for “recovery” outcomes outside prespecified adjustments.

Interpretation & Why It Matters

  • Clinical interpretation
    • Hydrocortisone (200 mg/day infusion) should be viewed as an adjunct that shortens vasopressor-dependent shock and accelerates ICU discharge, rather than as a therapy that improves survival.
    • In practice, the main expected benefit is earlier haemodynamic stabilisation (median shock resolution 3 vs 4 days) with potential downstream effects on ICU resource use and liberation from organ support.
    • Given the low absolute frequency of regimen-attributed serious adverse events (0.2% vs 0.1%), the risk–benefit profile is most favourable in patients with persistent vasopressor requirement despite adequate resuscitation.

Controversies & Other Evidence

  • Contemporary discordant trial results (ADRENAL vs APROCCHSS):
    • APROCCHSS (hydrocortisone plus fludrocortisone) reported lower 90-day mortality than placebo (43.0% vs 49.1%; RR 0.88; 95% CI 0.78 to 0.99), whereas ADRENAL (hydrocortisone alone) did not show a mortality reduction. 6
    • Key explanatory differences include illness severity (higher baseline mortality in APROCCHSS), steroid regimen (bolus dosing and mandatory fludrocortisone vs hydrocortisone infusion alone), and eligibility criteria (e.g., ADRENAL excluded etomidate). 3
  • Mineralocorticoid supplementation (fludrocortisone) as a potential effect modifier:
    • Correspondence emphasised mechanistic plausibility for added mineralocorticoid activity and the uncertainty about whether fludrocortisone meaningfully changes outcomes beyond hydrocortisone alone. 4
    • Direct head-to-head randomised evidence comparing hydrocortisone-alone versus hydrocortisone-plus-fludrocortisone strategies remains limited, so regimen choice is partly inference-based.
  • Longer-term outcomes:
    • At 6 months, mortality was similar: 35.9% (hydrocortisone) vs 35.4% (placebo); OR 0.99; 95% CI 0.86 to 1.13; P=0.83. 5
  • How the totality of evidence has been synthesised since 2018:
    • A patient-level meta-analysis (including contemporary large trials) reported consistent improvements in shock reversal and ICU course, without a clear overall survival signal for low-dose hydrocortisone strategies. 11
    • An updated 2024 systematic review and meta-analysis of corticosteroids in sepsis/septic shock concluded that corticosteroids likely improve shock-related outcomes and may modestly reduce mortality, with recognised metabolic adverse effects; certainty varied by endpoint and population. 12
  • Guideline impact:
    • Surviving Sepsis Campaign (2021) suggests intravenous corticosteroids for adults with septic shock and ongoing vasopressor requirement, reflecting prioritisation of faster shock resolution and shorter ICU stay over mortality benefit. 9
    • A 2024 focused guideline update on corticosteroids in sepsis similarly supports a conditional recommendation, emphasising patient selection (persistent vasopressor dependence) and balancing benefits against metabolic harms. 10
  • Foundational explanatory context from earlier RCTs:
    • Earlier studies generated the central “mortality vs shock reversal” tension: a hydrocortisone–fludrocortisone strategy suggested survival benefit in septic shock in 2002, whereas CORTICUS showed faster shock reversal without mortality benefit. 78

Summary

  • In ventilated adults with septic shock requiring vasopressors, hydrocortisone 200 mg/day by continuous infusion did not reduce 90-day mortality versus placebo (27.9% vs 28.8%; OR 0.95; 95% CI 0.82 to 1.10).
  • Hydrocortisone accelerated shock resolution (median 3 vs 4 days; HR 1.32) and ICU discharge (median 10 vs 12 days; HR 1.14).
  • There was no increase in new-onset bloodstream infection (14.1% vs 14.1%); blood transfusion was less frequent with hydrocortisone (37.0% vs 41.7%; OR 0.82).
  • Regimen-attributed adverse events were uncommon but more frequent with hydrocortisone (1.1% vs 0.3%); serious adverse events were rare (0.2% vs 0.1%).
  • ADRENAL reframed corticosteroids as a “recovery-accelerating” adjunct in refractory septic shock, shaping modern conditional guideline recommendations.

Overall Takeaway

ADRENAL is a landmark pragmatic ICU trial because it provided high-precision evidence that hydrocortisone (200 mg/day infusion) does not confer a mortality benefit in ventilated septic shock, while consistently shortening vasopressor-dependent shock and improving ICU throughput. Its clinical importance lies in reframing steroid use toward carefully selected patients with persistent vasopressor requirements, where faster shock reversal is the dominant and reproducible benefit.

Overall Summary

  • Hydrocortisone 200 mg/day infusion: no 90-day survival benefit, but faster shock resolution and earlier ICU discharge.
  • Current guidance supports conditional use in ongoing vasopressor-dependent septic shock, with attention to regimen choice (with vs without fludrocortisone) and metabolic harms.

Bibliography