The use of corticosteroids in septic shock has been extensively studied. Early investigations determined that high-dose corticosteroids in septic shock are not beneficial and may be harmful [2
]. Interest was renewed with the observation of hypothalamic-pituitary-adrenal axis dysfunction in patients with septic shock [4
]. When defined as an increase in plasma cortisol of ≤9 mcg/dl sixty minutes after administration of 250 mcg corticotropin, relative adrenal insufficiency (RAI) occurs in approximately 41–63% of patients with sepsis, is predictive of death [4
] and is associated with a blunted response to vasopressors that can be reversed by hydrocortisone [9
]. Under this premise, initial studies of stress-dose corticosteroids (200–300 mg hydrocortisone per day) in septic shock were conducted, demonstrating rapid shock reversal [10
]. Subsequently, in a multi-center trial in 300 patients with septic shock refractory to volume resuscitation and vasopressors, Annane and colleagues found that the administration of hydrocortisone 50 mg intravenously every 6 hours and fludrocortisone 50 mcg per day reduced 28-day mortality by 10% in patients with RAI [8
]. At the time of publication, this was the most definitive trial of stress-dose steroids in septic shock, greatly influencing intensivists and rapidly became the standard of care [14
The Corticosteroid Therapy of Septic Shock (CORTICUS) study evaluated the efficacy and safety of low-dose hydrocortisone therapy in a broad population of patients with septic shock, including patients who responded to a corticotropin test, in whom a benefit was unproven. Patients were enrolled if they had clinical evidence septic shock with onset within 72 hours of enrollment. Shock was defined by a systolic blood pressure (SBP) <90 mmHg despite fluid resuscitation or a vasopressor requirement for at least one hour. All patients underwent a corticotropin stimulation test. Somewhat surprisingly, the use of low-dose hydrocortisone had no significant effect on 28-day mortality, regardless of the patients' adrenal responsiveness to corticotropin. The proportion of patients in whom reversal of shock was achieved was similar in the two groups, though this goal was achieved earlier in patients who received hydrocortisone. New infection, hypernatremia and hyperglycemia occurred more frequently in the hydrocortisone group compared to placebo.
CORTICUS is the largest study to date to address the role of corticosteroids in septic shock. Yet, the study has limitations, the most important of which is inadequate power. The study was stopped prematurely due to slow recruitment, termination of funding, and time expiry of the trial drug. As such, only 500 of the intended 800 patients were enrolled. This, coupled with a lower control death rate than anticipated, resulted in the trial having less than 35% power to detect a relative risk reduction of 20% for the primary outcome [15
]. Selection bias is another potential limitation. Physicians who were convinced of the benefit of steroids may have been reluctant to withhold this therapy from their sickest patients, thereby excluding the group of patients that theoretically had the most to gain. The lower than expected mortality rate in the control group supports this notion. To better understand the potential influence of this limitation, it would have been useful for the authors to have provided information about the patients who were screened but not included in the study, such as those who were excluded because they were already receiving corticosteroids.
In comparing CORTICUS and the study by Annane and colleagues, there are important methodological differences, which may in part explain their differing findings. In the Annane study, patients were enrolled within eight hours of onset of shock and were still hypotensive (SBP <90 mmHg for at least one hour) despite fluid resuscitation and vasopressor therapy. In contrast, CORTICUS enrolled patients with evidence of shock within the previous 72 hours, as manifest by either hypotension after fluid resuscitation or a vasopressor requirement for at least 1 hour. This led to a disparity in severity of illness between the trials, with Annane and colleagues enrolling a sicker group of patients as measured by SAPS II scores and control group mortality (table). These observations bring into question not only the issue of timing, but also whether sicker patients might be more likely to benefit, as was seen with recombinant human activated protein C [16
]. CORTICUS patients more commonly had post-surgical, hospital-acquired, and abdominal infections. Patients with these characteristics may respond differently to steroid therapy than the primarily medical sample studied by Annane and colleagues. Finally, the trials also employed different steroid treatment protocols. The Annane trial used a fixed dose of hydrocortisone along with fludrocortisone for a total of 7 days; whereas in CORTICUS, a tapering dose of hydrocortisone (without fludrocortisone) for a total of 11 days was used. Whether the use of mineralocorticoids is important or a shorter, fixed dose regimen could have made a difference remain important and unanswered questions.
Comparison of patient characteristics in CORTICUS and the study of Annane and colleagues.
There are two additional studies addressing the use of corticosteroids in septic shock that should be mentioned. The Combination of Corticotherapy and Intensive Insulin Therapy for Septic Shock (COIITSS) study is completed, but not yet published [17
]. This study used a factorial design in 508 adults with septic shock to simultaneously compare hydrocortisone alone versus hydrocortisone plus fludrocortisone and intensive insulin therapy versus conventional glucose control. The other study, Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHS), is ongoing [18
]. APROCCHS aims to compare the efficacy and safety of recombinant human activated protein C to that of low dose of corticosteroids and to investigate the interaction between these drugs in 1280 adults with septic shock.