The New England Journal of Medicine January 10, 2008

Hydrocortisone Therapy for Patients with Septic Shock

Charles L. Sprung, Djillali Annane, Didier Keh, Rui Moreno, Mervyn Singer, Klaus Freivogel, Yoram G. Weiss, et al.

Bottom Line

In adult patients with septic shock, low-dose hydrocortisone did not improve 28-day mortality, regardless of corticotropin test response, but it did hasten the reversal of shock while increasing the risk of superinfections.

Key Findings

1. No significant difference in overall 28-day mortality: 34.3% (86 of 251) in the hydrocortisone group versus 31.5% (78 of 248) in the placebo group (P=0.51) [1.1.1].

2. No mortality benefit was observed among corticotropin non-responders: 28-day mortality was 39.2% with hydrocortisone versus 36.1% with placebo (P=0.69).

3. No mortality benefit was observed among corticotropin responders: 28-day mortality was 28.8% with hydrocortisone versus 28.7% with placebo (P=1.00).

4. Hydrocortisone significantly shortened the median time to reversal of shock (approximately 3.3 days vs. 5.8 days, P<0.001).

5. Hydrocortisone was associated with higher rates of superinfections, including new sepsis and septic shock (33% vs. 26%; RR 1.27), and an increased incidence of hyperglycemia (85% vs. 72%).

Study Design

Design

Randomized Controlled Trial

Double-Blind

Sample

499

Patients

Duration

28 days

Median

Setting

Multicenter, international

Population Adult intensive care unit patients showing clinical evidence of infection, systemic response, organ dysfunction attributable to sepsis, and the onset of shock (defined as systolic blood pressure <90 mm Hg despite adequate fluid or vasopressor therapy) within the preceding 72 hours.

Intervention Hydrocortisone 50 mg administered intravenously every 6 hours for 5 days, followed by a gradual dose taper over 6 days (total duration of 11 days).

Comparator Matching intravenous placebo administered on the identical dosing schedule.

Outcome Rate of death from any cause at 28 days among patients who did not have an adequate response to a corticotropin (ACTH) stimulation test (non-responders).

Study Limitations

• The trial was underpowered for its primary endpoint; enrollment was halted at 499 patients instead of the projected 800 due to slow recruitment and exhausted funding [1.2.6].

• The 72-hour enrollment window following the onset of shock was unusually long, which may have resulted in missing the optimal early therapeutic window for corticosteroids.

• The overall mortality rate in the placebo group (31.5%) was significantly lower than in previous positive steroid trials, such as the 63% mortality seen in the 2002 Annane trial, indicating a less severely ill patient cohort that might not derive as much benefit from corticosteroids.

• A notable proportion of patients received etomidate for intubation prior to enrollment; etomidate profoundly suppresses adrenal function and may have confounded corticotropin test results and clinical outcomes.

Clinical Significance

The CORTICUS trial fundamentally challenged the routine use of corticosteroids in all patients with septic shock. By disproving a mortality benefit and questioning the clinical utility of the corticotropin (ACTH) stimulation test to guide therapy, it shifted the global paradigm. Although hydrocortisone successfully reversed shock more rapidly, the absence of survival benefit combined with the increased risk of adverse events (hyperglycemia and superinfections) led to major revisions in the Surviving Sepsis Campaign guidelines. Consequently, corticosteroids are now typically reserved as a second-line therapy for patients whose septic shock remains refractory to adequate fluid resuscitation and vasopressor administration.

Historical Context

Following the landmark 2002 trial by Annane et al., which reported a substantial survival benefit from hydrocortisone and fludrocortisone in corticotropin 'non-responders', the concept of 'relative adrenal insufficiency' became widely accepted. Routine ACTH testing and empirical steroid administration became standard practices in septic shock. The CORTICUS trial (2008) contradicted these practices by showing no mortality benefit and increased infection risks, which largely eliminated routine ACTH testing from standard sepsis protocols and tempered enthusiasm for universal steroid use. This intense debate set the stage for subsequent mega-trials, such as ADRENAL and APROCCHSS in 2018, which further delineated the distinct effects of corticosteroids on shock resolution versus overall survival.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

How does hydrocortisone physiologically contribute to the reversal of shock, and why might it simultaneously increase the risk of superinfections in patients with septic shock?

Key Response

This explores the dual pharmacological role of glucocorticoids in sepsis: they upregulate alpha-1 adrenergic receptors and improve vascular responsiveness to catecholamines (reversing shock), while their immunosuppressive effects, such as the inhibition of NF-kB and decreased cytokine production, impair the host's ability to fight off secondary pathogens (increasing superinfections).

Resident

Based on the CORTICUS trial, in which specific clinical scenario is the initiation of low-dose hydrocortisone still considered appropriate for a patient in septic shock, and why is the corticotropin stimulation test no longer routinely used?

Key Response

Residents must know that steroids are now reserved only for vasopressor-refractory shock to hasten shock reversal. CORTICUS demonstrated that the corticotropin (ACTH) test did not predict which patients would experience a mortality benefit, making it clinically obsolete for guiding steroid initiation in septic shock.

Fellow

How do the differences in patient populations, baseline illness severity, and time to intervention between the Annane 2002 trial and the CORTICUS trial explain the conflicting mortality outcomes regarding hydrocortisone in septic shock?

Key Response

Critical care fellows must synthesize conflicting landmark trials. Annane enrolled sicker patients with higher baseline mortality much later in their course, whereas CORTICUS enrolled a broader, slightly less severe cohort earlier. This suggests that steroids may only offer survival benefits in a specific phenotype of profound, refractory shock rather than a general septic shock population.

Attending

Given that hydrocortisone hastens shock reversal but does not improve 28-day mortality and increases the risk of superinfections, how should we balance these competing outcomes when discussing prognosis, therapeutic goals, and resource allocation in the ICU?

Key Response

Attendings must weigh the operational benefit of faster liberation from the ICU (off pressors) against the real risk of nosocomial harm (superinfections) without a survival advantage. This requires nuanced shared decision-making, acknowledging that physiological improvement does not always translate to improved patient-centered outcomes.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

The CORTICUS trial struggled with slow enrollment and was ultimately underpowered for its original mortality endpoint. How would the use of a continuous or ordinal outcome measure, such as 'vasopressor-free days', alter the statistical efficiency and interpretability of such critical care trials compared to a binary 28-day mortality endpoint?

Key Response

This question probes study design tradeoffs. Using continuous variables like vasopressor-free days increases statistical power and might better capture the drug's true biological effect (shock reversal), especially when 28-day mortality is heavily confounded by competing risks in the ICU.

Journal Editor

Etomidate, a known adrenal suppressant, was frequently used for intubation prior to randomization in the CORTICUS trial. How does the post-hoc handling of these patients complicate the interpretation of the corticotropin response and the internal validity of the study?

Key Response

A peer reviewer would heavily scrutinize the etomidate confounding. Etomidate causes transient adrenal insufficiency, which artificially inflates the 'non-responder' ACTH group. This muddies the biological plausibility of the intervention and threatens the validity of the primary subgroup analysis regarding relative adrenal insufficiency.

Guideline Committee

The Surviving Sepsis Campaign guidelines suggest against routine IV corticosteroids, reserving them for cases where adequate fluid resuscitation and vasopressor therapy cannot restore hemodynamic stability. How does the CORTICUS trial directly inform the strength of this recommendation and the exclusion of routine ACTH testing?

Key Response

SSC guidelines moved away from routine ACTH testing and universal steroids directly because of CORTICUS. The committee weighed the trade-off between faster shock reversal (benefit) and superinfection risk without mortality benefit (harm), correctly down-grading steroids to a weak recommendation as a 'rescue' therapy only.

Clinical Landscape

Noteworthy Related Trials

2002

Annane et al.

n = 299 · JAMA

Tested

Hydrocortisone 50mg q6h plus fludrocortisone 50mcg daily

Population

Patients with septic shock requiring vasopressors

Comparator

Placebo

Endpoint

28-day mortality

Key result: Showed a significant reduction in 28-day mortality for patients with relative adrenal insufficiency who received corticosteroids.

2018

ADRENAL Trial

n = 3,658 · NEJM

Tested

Hydrocortisone 200mg per day as a continuous infusion

Population

Patients with septic shock undergoing mechanical ventilation

Comparator

Placebo

Endpoint

90-day mortality

Key result: Found no significant difference in 90-day mortality, though the hydrocortisone group had faster resolution of shock.

2018

APROCCHSS Trial

n = 1,241 · NEJM

Tested

Hydrocortisone 50mg q6h plus fludrocortisone 50mcg daily

Population

Adults with severe septic shock

Comparator

Placebo

Endpoint

90-day mortality

Key result: Demonstrated lower 90-day mortality in the hydrocortisone-plus-fludrocortisone group compared to placebo.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis