Correlation of Random Cortisol Levels, Corticosteroid Initiation, and Vasopressor Use

Introduction: Corticosteroids use in critically ill patients with shock remains debated due to inconsistent mortality outcomes. Current guidelines recommend initiation of steroids based on fluid and vasopressor response, but cite low quality of evidence. This study aimed to evaluate the benefit of steroid initiation on vasopressor use and resolution of shock in relation to baseline random cortisol levels. We hypothesized that lower random cortisol levels would correlate with improved outcomes in patients treated with steroids.


Methods: This retrospective chart review evaluated 84 patients who received corticosteroids in addition to vasopressors in the intensive care units (ICU) at UCHealth Memorial Hospital Central from January 2023 to December 2023. The primary outcome was time to resolution of shock following steroid administration in each cortisol level group ( < 10 mcg/dL vs 10-20 mcg/dL vs >20 mcg/dL). Secondary outcomes included: ICU length of stay (LOS), 28-day mortality, incidence of new blood glucose readings >180 mg/dL, and steroid regimens used.


Results: There was a significant difference between the cortisol level groups ( < 10 mcg/dL vs 10-20 mcg/dL vs >20 mcg/dL) with respect to median time from initiation of steroids to resolution of shock (37.2 hours vs 48.8 hours vs 65.6 hours, respectively; p = 0.013). A pairwise comparison between groups showed a significant difference in the primary outcome between the cortisol < 10 mcg/dL and >20 mcg/dL groups only (p = 0.005). Secondary outcomes showed no significant difference between groups in ICU LOS (5 days vs 7 days vs 9 days, p = 0.37), 28-day mortality (12.5% of patients vs 2% of patients vs 19% of patients; p = 0.25), time to initiation of steroids (6.9 hours vs 18.3 hours vs 7.1 hours; p = 0.06), new hyperglycemia (26% of patients vs 51% of patients vs 29% of patients; p = 0.08), or hydrocortisone equivalents used in each group (300 vs 300 vs 213; p = 0.99). The most frequent steroid regimens were hydrocortisone 100mg IV three times daily, hydrocortisone 50mg IV four times daily, and methylprednisolone 20mg IV twice daily.


Conclusions: This study suggests that in patients with shock treated with steroids, baseline random cortisol levels < 20 mcg/dL are associated with a shorter time to resolution of shock and no difference in 28-day mortality.