Date: February 14th, 2018

Reference: Venkatesh S et al. Adjunctive Glucocorticoid Therapy in Patients with Septic Shock. NEJM January 2018.

Guest Skeptic: Dr. Rory Spiegel (@EMNerd_) is a clinical instructor at University of Maryland, a recent graduate of Stony Brook’s Resuscitation Fellowship, and a current Critical Care fellow at University of Maryland. He writes an excellent blog called EM Nerd, which he describes as nihilistic ramblings.

Case: 64-year-old male presents to your emergency department with worsening abdominal pain, nausea, vomiting and anorexia for the past week. On presentation he is lethargic and hypotensive. He requires control of his airway and is given a 30 cc/kg fluid bolus and started on norepinephrine. His urine analysis is consistent with a urinary tract infection. Over the course of his emergency department stay he has escalating vasopressor requirements. After starting vasopression, you ask yourself if you should be adding steroids for potential relative adrenal insufficiency?

Background: Corticoid steroids have had a tumultuous history when it comes to their use for the treatment of septic shock. The underlying physiological reasoning is for the most part simple.

There is a certain subset of patients in septic shock whose adrenal axis functions well enough to support them in a state of health but are unable to support them in a state of unwell. It is this state of relative adrenal insufficiency that we hope to combat when administering steroids in patients with septic shock

But when it comes to the clinical evidence supporting the use of glucocorticoid therapy for the treatment of septic shock, we have existed in a state of ambiguity. There have been two contradictory randomized control trials.

The first, the Annane et al trial published in JAMA in 2002, suggested a mortality benefit in favor of the corticosteroid group in the subset of patients who were found to have relative adrenal insufficiency.

The second, the larger CORTICUS study, published in the NEJM in 2008 found no benefit of the use of corticosteroids compared to placebo.

Now these contradictory results could have come about for a number of reasons. First, the patients in the Annane et al trial represented a much sicker population, isolating the subset of patients who truly benefit from steroid supplementation.

While the negative results observed in the CORTICUS trial were the consequence of enrolling a much healthier population and washing out any chance of identifying the underlying signal.

On the other hand, the results of the Annane et al trial may have been due to statistical noise and a small sample size, and CORTICUS represents nothing more than the expected regression to the mean.

There was also the HYPRESS trial that we reviewed on SGEM#168 with Salim Rezaie from REBEL EM. This looked at hydrocortisone to prevent patients with severe sepsis in developing septic shock. While the primary outcome showed no difference between glucocorticoid treatment and placebo, there was also no secondary mortality benefit observed.

The SGEM bottom line from that episode was that the use of hydrocortisone in adult patients with severe sepsis to prevent septic shock cannot be recommended at this time.

Clinical Question: Does the administration of a continuous infusion of hydrocortisone to a general population of patients in septic shock on mechanical ventilation improve survival?

Reference: Venkatesh S et al. Adjunctive Glucocorticoid Therapy in Patients with Septic Shock. NEJM January 2018.

  • Population: Adults, age 18 or older who required mechanical ventilation, with septic shock (two or more SIRS criteria on vasoactive agents for at least 4-hours)
    • Excluded: Patients likely to receive glucocorticoids for an indication other than septic shock, had received etomidate during the hospitalization, were considered to be likely to die from a pre-existing disease within 90 days or had treatment limitations in place, or had met all the inclusion criteria for >24 hours.
  • Intervention: IV infusion of hydrocortisone at a dose of 200 mg per day for seven days or until death or discharge from the ICU
  • Comparison: Placebo infusion
  • Outcome:
    • Primary: 90-day mortality from any cause
    • Secondary: 28-day mortality from any cause, time to resolution of shock, recurrence of shock, ICU length of stay, hospital length of stay, days of mechanical ventilation, renal replacement therapy, incidence of new-onset bacteremia or fungemia and blood transfusions in the ICU.

Authors’ Conclusions: Among patients with septic shock undergoing mechanical ventilation, a continuous infusion of hydrocortisone did not result in lower 90-day mortality than placebo.”

checklistQuality Checklist for Randomized Clinical Trials:

  1. The study population included or focused on those in the emergency department. No. This was primarily an ICU population
  2. The patients were adequately randomized. Yes
  3. The randomization process was concealed. Yes
  4. The patients were analyzed in the groups to which they were randomized. Yes
  5. The study patients were recruited consecutively (i.e. no selection bias). Yes
  6. The patients in both groups were similar with respect to prognostic factors. Yes
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. Yes
  8. All groups were treated equally except for the intervention. Yes
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes
  10. All patient-important outcomes were considered. Yes
  11. The treatment effect was large enough and precise enough to be clinically significant. No

Key Results: They enrolled 3,658 patients into the study. The mean age of patients was early 60’s and approximately 60% were male.

No difference in all-cause mortality at 90 days

  •  Primary Outcome (90-day Mortality):
    • 27.9% in the intervention arm vs 28.8% in the control arm
    • Odd ratio 0.95 (95% CI, 0.82 to 1.10; P=0.50)
  • Secondary Outcomes (treatment vs. placebo):
    • 28-day mortality (22% vs. 24%)
    • Time to resolution of shock (3 vs. 4 days)
    • Recurrence of shock (20% vs, 18%)
    • ICU length of stay (10 vs. 12 days)
    • Hospital length of stay (39 vs. 43 days)
    • Mechanical ventilation (6 vs. 7 days)
    • Renal replacement therapy (43% vs. 40%)
    • Incidence of new-onset bacteremia or fungemia (14% vs 14%)
    • Blood transfusions in the ICU (37% vs. 42%)
  • Adverse Events: 27 vs. 6 adverse events with hyperglycemia being the most common (6) in the hydrocortisone group.

Screen Shot 2015-04-25 at 3.11.12 PM

This was a very well-done, randomized, blinded, multi-center trial with rigorous methods.

1) Power: It was a little under powered because the prevalence of mortality was a little less than their estimated 33% for mortality and they were a little under on their 3,800-enrollment target. I do not think this fundamentally changes their conclusions or invalidates their results.

2) Secondary Outcomes: One of the biggest issues in the trial is how exactly do we interpret secondary outcomes? The authors noted that patients in the hydrocortisone group had a faster time to resolution of shock, a shorter time to discharge from the ICU, and a shorter duration of initial mechanical ventilation. But we all know the dangers of interpreting secondary outcomes as truth. The study was not powered for these outcomes and should be considered hypothesis generating.

Moreover, although time to ICU discharge was statistically shorter in patients randomized to receive hydrocortisone, the authors found no difference in number of days alive and outside the ICU, or days alive and outside the hospital. And while duration of initial mechanical ventilation was shorter in the hydrocortisone group, there was no difference in days alive and free of mechanical ventilation nor a difference in the rate of recurrent mechanical ventilation. These inconsistencies make one question the true clinical meaning of these outcomes

3) Statistical vs. Clinical: The difference between statistical and clinical significance is a point we often make on the SGEM. Yes, the authors found a shorter time to resolution of shock in the hydrocortisone group, but the mean difference in MAP between the two groups was only 5.39 mm Hg, and there was no difference in the daily dose of norepinephrine. Many findings, especially those measuring continuous variables like differences in blood pressure can be statistically different. It’s always important to judge whether these findings have any effect on clinical outcomes.

4) Regression to the Mean: This is an important concept when appraising a study. The lead author of this trial presented its results at this year’s Critical Care Reviews conference in Ireland, put on by our friend Rob McSweeny, and during the presentation he showed the results of the various interim analyses that were conducted during the study

During the first analysis, which was over 900 patients, twice that of the CORTICUS trial, there was a statistically significant improvement in mortality favoring the patients who received hydrocortisone. But over the course of the trial as the sample size grew larger and larger, this difference disappeared. This is the perfect example of how unstable trials are at small sample sizes. Now if this study was stopped early after this first interim analysis, we all would have thought steroids were helpful in septic shock, which of course would have been incorrect.

It is not atypical that trials with small sample sizes to have wide variations in point estimates around the true mean that meet our standard for statistical significance difference, and it is only after a larger validation study is conducted that the point estimate settles around the true effect size.

5) Adverse Events: Like many studies, this trial was not as rigorous in assessing adverse events. It was up to the individual clinician to decide if any of the adverse events were related to the trial. There were four times as many adverse events in the hydrocortisone group, but these events were not independently evaluated. Regardless, if there is no mortality benefit with treatment then why should patients accept any level of harm?

Comment on Authors’ Conclusion Compared to SGEM Conclusion: The authors conclusion that glucocorticoids do not improve mortality in septic shock is a reasonable assessment. What conclusion we can draw from the secondary outcomes is less clear.

SGEM Bottom Line: The administration of a continuous infusion of hydrocortisone to a general population of patients in septic shock on mechanical ventilation cannot be recommended at this time.

It is hard to view these results as anything but a negative study. While there are some potential signals of benefit in the patients who received hydrocortisone, they were fleeting and ultimately had minimal influence on patient centered outcomes. If a true benefit does exist, it is small and would require an impossibly large trial to empirically demonstrate.

Despite this I think the results of the ADRENAL trial will paradoxically increase the use of glucocorticoids in patients with septic shock. The desire to act in the face of critical illness is strong. And the influential power of a statistically significant p-value is strong.

People should read the article called Don’t just do something, stand there! The value and art of deliberate clinical inertia (Emerg Med Austrialas 2018).The senior author on it is Dr. Daniel Fatovich and it is an excellent explanation on why it can be hard to avoid unnecessary tests and treatments.

When faced with a patient in refractory shock, many will continue to give corticosteroids in the hopes they may help the individual patient laying before them. It will just happen with an added degree of hand-wringing, having read the literature refuting this practice and simply want nothing to do with its conclusions

Dr. Rory Spiegel

Dr. Rory Spiegel

Case Resolution: Hydrocortisone was held and the patient’s shock worsened over the next few hours. A CT abdomen/pelvis was ordered because of a concern of lack of source control. The imaging identified an obstructive kidney stone with an associated pyelonephritis. Urology was consulted, and a drain was placed. Over the next 24-hours the patient’s shock resolved, and he came off his vasopressors

Clinical Application: In an undifferentiated cohort of patients with septic shock the addition of glucocorticoid therapy does not improve mortality.

What Do I Tell My Patient’s Family?  Your husband/father is fairly sick. We have resuscitated him with IV fluids and medications that will bring his blood pressure up to a safe level. Recent literature suggests that the addition of glucocorticoids does not improve outcomes when compared to standard care. Importantly a source was identified, and we know think we have control of his infection.

Keener Kontest: Last weeks’ winner was Shawn Murphy a PA from Parry Sound, Ontario, Canada. He knew the fantastic acronym they gave to the study on IV antibiotics in the pre-hospital setting for sepsis was called the PHANTASi Trial (Prehospital ANTibiotics Against Sepsis) trial . Shawn has won before but this time he will be getting one of the cool NEW skeptical prizes.

Listen to the SGEM podcast on iTunes to hear this weeks’ question. If you know the answer send an email to TheSGEM@gmail.com with “keener” in the subject line. The first correct answer will receive one of our cool new skeptical prizes.

FOAM logoOther FOAMed:

  • EM Nerd: The Case of the Relative Insufficiency
  • REBEL EM: The ADRENAL Trial – Steroids in Septic Shock
  • St. Emlyn’s: The End of the ‘Roid?’
  • The Bottom Line: Adjunctive Corticosteroid Treatment in Critically Ill Patients with Septic Shock
  • EM Literature of Note: The Definitive Word on Steroids in Septic Shock
  • EMCrit: Metabolic Sepsis Resuscitation – Strike Hard, Strike Fast, No Remorse

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.

Screen Shot 2017-12-15 at 10.53.29 AM