New studies inform ongoing controversies about steroids for patients with septic shock and about crystalloid solutions for fluid resuscitation in the intensive care unit.
On March 1, 2018, three studies that generated much discussion in the critical care community were published in the New England Journal of Medicine. Two of these studies focused on use of corticosteroids in treating patients with septic shock; in the third study, researchers examined whether crystalloid choice in intensive care unit (ICU) patients influenced outcomes. Are these trials practice-changing?
Should steroids be given to septic shock patients?
The controversy regarding corticosteroids for treating patients with septic shock has been ongoing for nearly 2 decades. These two new trials add to the debate but probably won’t end it, because they generated partially conflicting results. In one trial, APROCCHSS, 90-day mortality was significantly lower in patients who were treated with both the glucocorticoid hydrocortisone (50 mg every 6 hours for 1 week) and the mineralocorticoid fludrocortisone than in placebo recipients (43% vs. 49%; NEJM JW Gen Med Mar 1 2018 and N Engl J Med 2018; 378:797). In contrast, the other trial (ADRENAL) was a comparison of hydrocortisone alone versus placebo, and mortality was virtually the same in both groups – about 28% (NEJM JW Gen Med Apr 15 2018 and N Engl J Med 2018; 378:809). Key differences between the trials were use of a mineralocorticoid and higher overall mortality (suggesting a sicker patient population) in APROCCHSS. Notably, in both studies, the mean duration of septic shock was shorter in the steroid groups; in ADRENAL, this translated into less time in the ICU. Corticosteroid side effects were minimal in both trials.
Multiple trials now have shown that steroids shorten the duration of septic shock. If this effect shortens the length of ICU stay, as it did in ADRENAL, steroid use might result in cost savings and less arduous hospitalizations for some patients and families. In other words, even if the mortality benefit is marginal, these secondary effects might be worthwhile, given the low cost and apparent absence of harm from relatively brief courses of moderate-dose steroids.
After talking to several colleagues in Seattle and across the country, my sense is that these studies will reinforce previous practice preferences, whatever they might have been. Those who previously were steroid skeptics will not necessarily change their practice, whereas clinicians who had low thresholds for giving steroids will continue to do so and will note that APROCCHSS supports their practice. Like many of my peers, I will continue using glucocorticoids for patients with refractory septic shock who are on escalating doses of vasopressors or who require multiple vasopressors. In my discussions, reactions to adding fludrocortisone were mixed. My take is that fludrocortisone is inexpensive and low risk, so I probably will add it when I start glucocorticoids.
Is a balanced crystalloid better than normal saline for ICU patients?
The third trial (SMART) was conducted because of concern about potential adverse renal effects of the high chloride content of normal saline. Investigators compared normal saline with “balanced” crystalloid solutions (either lactated Ringer’s solution or Plasma-Lyte A) in more than 15,000 patients in ICUs at Vanderbilt University. The primary outcome was major adverse kidney events — a composite outcome that included death, renal-replacement therapy, or doubling of creatinine at discharge. Patients in the normal saline group had more primary outcome events than those in the balanced solution group (15% vs. 14%); this small difference was statistically significant for the composite outcome, but no significant difference was found for any individual component (NEJM JW Gen Med Apr 15 2018 and N Engl J Med 2018; 378:819).
In my discussions with other intensivists, most told me that their practices already were changing to preferential use of lactated Ringer’s instead of normal saline, except in unique patient populations (e.g., those with traumatic brain injury). So, although debate continues on how to interpret the results of SMART, and experts express caution about using a single-center trial to drive practice, the results reinforce the practice of reaching for lactated Ringer’s first, for most critically ill patients who require fluid resuscitation.
ARYAN'S BLOG 