Oct 282016

The idea that augmenting cortisol levels to normal or supranormal levels must be somehow beneficial in septic shock has a compelling basis in physiology and intuition. For physicians, injecting powerful synthetic hormones to restore homeostasis to save a dying patient is a seductive fulfillment of the scientist-savior fantasy.

So intensivists were primed to believe the results of a single 2002 randomized trial showing corticosteroids improved survival in adrenally-impaired patients with septic shock -- despite its 70% ICU mortality and confusing statistical methods. And why not, when it was followed in 2005 by a study showing corticosteroids delayed septic shock development in 46 patients with community-acquired pneumonia, and numerous small studies seemed to support the physiologic benefits of corticosteroids in severe infections?

But that was a more innocent era, when critical care docs could feel their sepsis dragon-slaying powers growing almost daily, thanks to an ever-expanding armory including intensive insulin therapy, Xigris, and the magic incantations known as sepsis protocols. There were so many treatments with accepted efficacy, we had to put them in bundles to make them easier to carry. Systemic corticosteroids were quickly added to the panel of standard treatments for patients with severe sepsis and septic shock.

Then CORTICUS (2008) showed no mortality benefits from corticosteroids among 500 slightly less-ill patients with septic shock, regardless of adrenal function. Steroids did bring faster resolution of shock, but with a suspicion of higher infection rates in the steroid-treated. Corticosteroids became less of a reflex treatment, and were reserved for patients with the worst septic shock (variously called "refractory" or "vasopressor-dependent"; the definition varied by whom you asked).

Eight years later, we can add to the discussion the results of 2016's HYPRESS, a randomized trial of about 350 adults with severe sepsis (enrolled back when that was a thing), who did not have shock at the time of presentation.

What They Did

Patients with sepsis without shock at presentation got either placebo or hydrocortisone 200 mg daily for 5 days, then tapered over 6 more days. Adrenal function (ACTH stimulation) was not checked consistently and was not included in the analysis.

What They Found

Corticosteroids did not improve any important measures. Patients who received steroids developed septic shock at the same rate as those who did not (~20%). Mortality in the ICU, hospital, or at 28, 90, or 180 days were all equivalent statistically. Numerically, more patients who had received steroids tended to die over time (27% vs 22% at 180 days) - not statistically significant.

Patients getting steroids developed significantly more hyperglycemia. They had numerically more secondary infections (22% vs 17%), also not statistically significant.

What It Means

Trials conducted on sepsis are notoriously hard to compare, because enrolled patient populations are so different, and sepsis mortality has been steadily declining over the past 20 years. Almost no single trial should be considered definitive, and HYPRESS is no exception.

For those reasons, and for its relatively survival-prone patient population, HYPRESS does not provide enough information to advise intensivists to stop prescribing corticosteroids for the very sickest patients with severe septic shock on vasopressors. But HYPRESS should help dispel any magical thinking leading physicians to prescribe corticosteroids to patients with sepsis, who are not in shock.

Read more: Effect of Hydrocortisone on Development of Shock Among Patients With Severe Sepsis: The HYPRESS Randomized Clinical Trial. JAMA. 2016 Oct 3 online.

Get our weekly email update, and explore our library of practice updates and review articles.

PulmCCM is an independent publication not affiliated with or endorsed by any organization, society or journal referenced on the website. (Terms of Use | Privacy Policy)


Corticosteroids for sepsis didn’t prevent septic shock (HYPRESS trial)