From Pages to Practice

By Bhavna Seth, M.D.

Published February 28, 2018


You are on call overnight in the intensive care unit (ICU) and admit a 70-year-old woman, Mrs. Roma, who is already intubated due to respiratory distress and sepsis. You institute your ICU’s sepsis protocol, which includes providing crystalloids, initiating antibiotics as soon as possible, and checking the patient’s lactate (which turns out to be normal). You anticipate that she will defervesce soon, but 1 hour after starting the sepsis protocol, her mean arterial pressure drops in the 50s and her lactate rises. You give more fluids, obtain central access, initiate norepinephrine and note some improvement. However, her norepinephrine requirements escalate during the next 3 hours and you add a second vasopressor. You also consider the role of adding glucocorticoids and wonder how effective they are in patients with sepsis.

Over the last 50 years, the role of glucocorticoids in the management of critical illness and infections has continued to be debated. On the one hand, an “adrenal response” is known to be crucial for survival in physiological stress, and benefits of administering exogenous glucocorticoids may improve adrenergic responsiveness and mitigate possible critical illness-related corticosteroid insufficiency (CIRCI). On the other hand, glucocorticoids can antagonize the endogenous immune defense mechanisms involved in clearing infections. To date, we have learned that high-dose glucocorticoid regimens are associated with worse outcomes, but reports of low-dose glucocorticoids in sepsis have been hamstrung by conflicting results.

This week, NEJM published two much-anticipated, well-powered, randomized, blinded, multicenter, controlled trials of low-dose glucocorticoids in critically ill medical and surgical patients: The Adjunctive Corticosteroid Treatment in Critically Ill Patients with Septic Shock (ADRENAL) trial and the Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHSS) trial. Both trials included patients with clearly defined vasopressor-dependent shock and respiratory failure leading to the use of mechanical ventilation. Both trials also included details of antimicrobial therapy, assessment of 90-day all-cause mortality as the primary outcome, and robust secondary outcomes and adverse events. The combined sample size of more than 5000 patients (3658 and 1241, respectively) dwarfs those of previous studies.

The ADRENAL trial compared hydrocortisone (200 mg/day) to placebo for 7 days or until death or discharge (whichever came first). The APROCCHSS study compared patients who received 7 days of hydrocortisone (200 mg/day) plus fludrocortisone (50 µg tablet/day) to those who did not (placebo group). However, the primary outcome for the two trials differed greatly: In ADRENAL, 90-day all-cause mortality was 27.9% with hydrocortisone and 28.8% with placebo (P=0.50); in APROCCHSS, it was 43.0% with hydrocortisone plus fludrocortisone vs. 49.1% with placebo (P = 0.03). For secondary outcomes, both trials indicated that hydrocortisone was associated with an increase in the number of days free of shock or organ failure and more rapid cessation of mechanical ventilation. Rates of serious adverse events were low, with higher risk of hyperglycemia with bolus glucocorticoid doses. No differences were noted in rates of renal-replacement therapy or incidence of new-onset bacteremia or fungemia (ADRENAL trial).

Why did the primary outcomes differ so greatly? First, the patient populations were not completely comparable. Patients in the ADRENAL trial had a higher rate of surgical admissions (31.5% vs. 18.3%). Prior studies indicate that surgery induces a Th2 immunosuppressive state, and glucocorticoids compound this, increasing the risk of secondary infections. Patients in the ADRENAL trial also had lower rates of renal-replacement therapy (12.7% vs 27.6%), blood infection (17.3% vs. 36.6%), pulmonary infection (35.2% vs. 59.4%), and urinary tract infection (7.5% vs. 17.7%), and a higher rate of abdominal infection (25.5% vs. 11.5%). Second, the two studies used different tools to assess illness severity at recruitment: the ADRENAL trial used the Acute Physiology and Chronic Health Evaluation II (APACHE II) score and the APROCCHSS trial used the Sequential Organ Failure Assessment (SOFA) score and the Simplified Acute Physiology Score II (SAPSE II). Finally, previous studies have failed to demonstrate a benefit from the addition of fludrocortisone in septic shock (as in the APROCCHSS trial).

Unfortunately, for Mrs. Roma’s management, there is no clear indication whether glucocorticoids will be beneficial to her survival. However, considering the safety profile and the likelihood of short-term clinical improvement, for now, using glucocorticoids in septic and critically ill patients remains a clinical alternative until stronger evidence to the contrary emerges.

Browse more From Pages to Practice

Bhavna Seth, Resident at Boston Medical Center
Bhavna Seth, Resident at Boston Medical Center