By Samuel Nadler, MD, PhD

Critical Care, Pulmonary Medicine, The Polyclinic Madison Center, Seattle; Clinical Instructor, University of Washington, Seattle

Dr. Nadler reports no financial relationships relevant to this field of study.

Septic shock carries a significant risk of mortality despite increasing knowledge of its pathophysiology and clinical management. Studies dating back to the 1960s suggested steroid treatment may alter the course of septic shock and led to the concept of critical illness-related corticosteroid insufficiency.1

In 2002, Annane et al demonstrated a mortality benefit for patients with septic shock given the combination of hydrocortisone and fludrocortisone, generating interest in this therapy and changing guidelines.2 However, the 2008 CORTICUS trial demonstrated no mortality benefit and a re-evaluation of this guideline.3

Two recent trials, ADRENAL and APROCCHSS, have provided more data regarding steroid therapy for septic shock.4,5 Comparing these seminal studies provides context for the decision about whether to treat septic shock with steroid therapy.


Annane et al published results from a placebo-controlled, randomized, double-blind trial of 299 patients with septic shock.2 The authors enrolled adult patients with documented or strong suspicion of infection, alterations in body temperature, tachycardia, systolic blood pressure < 90 mmHg for one hour despite fluid administration, and vasopressors with organ dysfunction defined by low urine output, elevated arterial lactate, or need for mechanical ventilation. Patients were excluded if they presented with acute myocardial infarction, pulmonary embolism, advanced cancer, AIDS, or contraindications to or pre-existing indications for steroid therapy. Participants were stratified further by a 250 mcg tetracosactrin stimulation test as responders or non-responders. The intervention group received hydrocortisone 50 mg intravenously every six hours and enteral fludrocortisone 50 mcg daily for seven days. Regarding the primary endpoint, 28-day mortality in non-responders was significantly lower in the steroid group (53%) than in the placebo group (63%), with an odds ratio (OR) of 0.54 (95% confidence interval [CI] 0.31-0.97; P = 0.04).

In all patients, the OR for 28-day survival was 0.65 (95% CI, 0.39-1.01; P = 0.09) in the steroid vs. placebo groups. The non-responders also demonstrated both decreased ICU mortality and hospital mortality when given steroids. Time to vasopressor therapy withdrawal with steroid therapy was shorter in both the non-responder group (7 days vs. 10 days; P = 0.001) and the overall patient population (7 days vs. 9 days; P = 0.01). No adverse events were ascribed to steroid replacement therapy. Specifically, there were no increased rates of infections, bleeding, or delirium.

The authors of the CORTICUS study, published in 2008, enrolled 499 patients with sepsis and examined the response to 50 mg of hydrocortisone intravenously every six hours.3 The inclusion criteria included evidence of infection, the systemic response to infection, and onset of shock within 72 hours. Exclusion criteria included poor prognosis and recent steroid use. All patients underwent a cosyntropin stimulation trial. The primary endpoint was 28-day mortality in non-responders. Notably, power calculation of this trial cited a sample size of 800 patients to achieve a statistical power of 80% to detect a 10% decrease in absolute mortality, assuming 50% mortality.

This trial demonstrated no difference in mortality with hydrocortisone compared with placebo in non-responders (39.2% vs. 36.1%; P = 0.69) or all patients (34.3% vs. 31.5%; P = 0.51). There was no difference in ICU mortality, death during hospitalization, death at one year, or length of stay. Oddly, time to reversal of shock with steroids was statistically lower in those patients who responded to cosyntropin (2.8 days vs. 5.8 days; P < 0.001) and all patients (3.3 days vs. 5.8 days; P < 0.001), but not in non-responders (3.9 days vs. 6.0 days; P = 0.06).

The authors of the HYPRESS trial, published in 2016, further evaluated the effect of hydrocortisone alone in patients with severe sepsis.6 The inclusion criteria for this trial included evidence of infection with a systemic response, organ dysfunction not present for more than 48 hours, but excluded patients with septic shock, defined by persistent hypotension despite fluids and vasopressors. The study population received hydrocortisone 50 mg bolus followed by a continuous infusion of 200 mg per day for five days with tapering over the next six days. The primary endpoint was the occurrence of septic shock within 14 days. Power calculation in this study planned an 80% chance to detect a 15% difference, assuming a rate of septic shock in the study population of 40% with 380 total patients.

Of the 353 patients who were included in the intention-to-treat analysis, the rates of septic shock at 14 days in the placebo and study group were very similar (22.9% and 21.2%, respectively; P = 0.70). No significant differences were noted in many secondary endpoints, although the rates of delirium were lower in the hydrocortisone group. The rate of hyperglycemia was significantly higher with steroid administration. Another notable difference was the relatively higher proportion of patients with pneumonia and respiratory tract infections in the placebo group compared with the study group.

Most recently, the authors of two additional studies (ADRENAL and APROCCHSS) examined the response of patients with sepsis to steroids.4,5 ADRENAL investigators randomized adult patients on mechanical ventilation with suspicion of infection, two or more systemic inflammatory response syndrome (SIRS) criteria, and the need for vasopressors or inotropes for at least four hours. Patients who received etomidate, exhibited other indications for steroids, or were expected to die within 90 days were excluded. The study compared a continuous infusion of hydrocortisone 200 mg daily for seven days vs. placebo. The primary outcome was all-cause mortality at 90 days. A population of 3,800 patients provided the trial a 90% power to detect a 5% absolute difference in the primary outcome, with an estimated baseline mortality of 33%. However, there was no difference in 90-day mortality between the treatment and control groups (27.9% and 28.8%, respectively; P = 0.50). There were improvements in the secondary outcomes of median time to shock resolution, median time to discharge, and median time to cessation of mechanical ventilation. These authors also noted more adverse events in the hydrocortisone group, including hyperglycemia, hypernatremia, hypertension, and myopathy.

In contrast, APROCCHSS originally was designed to be a 2 × 2 factorial study examining activated protein C (APC) and steroids in septic shock. During the trial, APC was removed from the market, but the trial continued, focusing on the steroid effects. Inclusion criteria included indisputable or probable septic shock for < 24 hours as defined by a sequential organ failure assessment score of 3-4 in at least two organ systems and vasopressor therapy. Exclusion criteria included septic shock for > 24 hours, pregnancy or lactation, or underlying conditions that could affect short-term survival. The study group received 50 mg hydrocortisone intravenously every six hours and 50 mcg enteral fludrocortisone daily for seven days. Again, cosyntropin stimulation trials were performed. The primary outcome was 90-day all-cause mortality.

Here, the baseline mortality was assumed to be 45% and a total of 1,280 patients would be required to detect an absolute 10% difference. The primary outcome was realized in 43% of the steroid group and 49.1% of the placebo group, with a relative risk of death of 0.88 (P = 0.03). Mortality at ICU discharge, hospital discharge, and day 180 were all statistically less in the steroid-treated group. Furthermore, the secondary outcomes of vasopressor-free, ventilator-free, and organ failure-free days were fewer in the treatment arm. Again, more hyperglycemia was noted with steroid administration, but there was no difference in infection, bleeding, or myopathy rates between the two groups.


With these trials demonstrating conflicting outcomes with differing interventions, how can providers make decisions regarding steroid treatment for sepsis? By comparing each study design and specific patient factors within each study, some conclusions can be drawn (See Table 1).

In terms of study design, the most important differences were in the intervention arms. Those studies that demonstrated mortality benefits included both hydrocortisone and fludrocortisone. It might be that both medications are required for benefit. However, the authors of the 2010 COIITS study compared hydrocortisone to hydrocortisone plus fludrocortisone, which showed no difference in mortality, although this study did not include a placebo arm without steroid therapy and was not powered specifically to detect differences in mortality.7

The power of each study to evaluate mortality also differed. After the 2002 study generated excitement for steroid therapy, the CORTICUS study in 2008 failed to show an improvement in mortality. However, this study was underpowered to detect changes in mortality. Both 2018 studies were adequately powered for their primary outcomes, although both the overall mortality and number of patients enrolled in the ADRENAL study were lower than the figures used in the power calculation. An additional difference was the length of steroid treatment. Both positive studies used steroids for seven days, then stopped. ADRENAL specified treatment up to seven days, while the other negative trials tapered steroids up to 11 days. It is possible that the adverse effects of steroid treatments increasingly outweighed the benefits with longer treatments. Patient factors also differed considerably between these studies. Studies with positive results demonstrated higher overall mortality (See Table 1). This a result of differing inclusion criteria. The inclusion criteria in the two positive studies specified significant organ dysfunction with few exclusion criteria.

CORTICUS included patients with organ dysfunction, but excluded patients with poor prognosis, immunosuppression, or life expectancy of < 24 hours, not an uncommon occurrence with severe sepsis. HYPRESS specifically excluded patients with septic shock and demonstrated the lowest overall mortality of the studies. ADRENAL included SIRS criteria, but did not specify organ dysfunction, and excluded patients who were expected to die within 90 days from comorbidities. Also notable was the much higher proportion of patients with pulmonary infections in the positive studies.

While in ADRENAL, 33.8% and 36.5% of the steroid and placebo groups, respectively, had pulmonary infections; in the APROCCHSS study, 58% and 60.7%, respectively, had pulmonary infections. This is notable, as a growing body of literature suggests steroid therapy can alter the outcomes of severe pneumonia.8

What the outcomes of each study had in common also are informative. The 2002 Annane et al study demonstrated that steroids led to a decreased time to vasopressor therapy withdrawal in both non-responders and all patients. CORTICUS reported a decreased time to reversal of shock for all patients as well as decreased time on mechanical ventilation. ADRENAL demonstrated that hydrocortisone administration was associated with decreased median time to shock reversal, ICU length of stay, and median time on mechanical ventilation.

APROCCHSS demonstrated statistically significant improvements in vasopressor-free days and organ failure-free days, with a nonsignificant trend toward increased ventilator-free days. Except for the original Annane et al study, all reported hyperglycemia as an adverse effect of steroid treatment.

However, the COIITSS study did not demonstrate that intensive control of hyperglycemia with insulin in steroid-treated patients improved outcomes, raising the question of the clinical significance of transient hyperglycemia in this population.7


Much uncertainty remains regarding the benefits of steroid therapy for sepsis. Furthermore, it seems unlikely that additional, well-powered studies will be undertaken to address the questions of patient selection, treatment intervention, duration of therapy, or infusions vs. bolus therapy.9 Current studies seem to indicate potential benefits for patients with higher levels of sepsis severity, especially in those presenting with pneumonia as the causative factor. As the only positive trials included fludrocortisone, the best evidence would be to use both intravenous hydrocortisone and enteral fludrocortisone if treating severe sepsis with steroids. There seems to be some agreement that steroid therapy in sepsis leads to shorter time to vasopressor withdrawal and the duration of need for mechanical ventilation. Predicting which patients will require longer duration of vasopressors and mechanical ventilation is difficult. Ultimately, in those patients presenting with severe sepsis with rising vasopressor needs who no longer appear to be fluid responsive, the addition of hydrocortisone and fludrocortisone may improve outcomes.


  1. Marik PE. Critical illness-related corticosteroid insufficiency. Chest 2009;135:181-193.
  2. Annane D, Sébille V, Charpentier C, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA 2002;288:862-871.
  3. Sprung CL, Annane D, Keh D, et al. Hydrocortisone therapy for patients with septic shock. N Engl J Med 2008;358:111-124.
  4. Venkatesh B, Finfer S, Cohen J, et al. Adjunctive glucocorticoid therapy in patients with septic shock. N Engl J Med 2018;378:797-808.
  5. Annane D, Renault A, Brun-Buisson C, et al. Hydrocortisone plus fludrocortisone for adults with septic shock. N Engl J Med 2018;378:809-818.
  6. Keh D, Trips E, Marx G, et al. Effect of hydrocortisone on development of shock among patients with severe sepsis: The HYPRESS randomized clinical trial. JAMA 2016;316:1775-1785.
  7. COIITSS Study Investigators; Annane D, Cariou A, Maxime V, et al. Corticosteroid treatment and intensive insulin therapy for septic shock in adults: A randomized controlled trial. JAMA 2010;303:341-348.
  8. Stern A, Skalsky K, Avni T, et al. Corticosteroids for pneumonia. Cochrane Database Syst Rev 2017;12:CD007720.
  9. Suffredini AF. A role for hydrocortisone therapy in septic shock? N Engl J Med 2018;378:860-861.


Table 1. Comparison of Major Studies of Steroids for Sepsis


of Patients


Length of Treatment (Days)

Study Mortality

in Mortality

Power to Detect Change?

Annane et al2


Hydrocortisone and fludrocortisone boluses


28-day mortality of 58%





Hydrocortisone bolus


28-day mortality of 33%

Not significant




Hydrocortisone infusion


28-day mortality of 8.5%

Not significant




Hydrocortisone infusion

Up to 7

90-day mortality of 28%

Not significant




Hydrocortisone and fludrocortisone boluses


90-day mortality of 46%