The Timing of Initiating Hydrocortisone and Long-term Mortality in Septic Shock

Abstract

Background: Previous studies on the association between the timing of corticosteroid administration and mortality in septic shock focused only on short-term mortality and produced conflicting results. We performed a retrospective review of a large administrative database of intensive care unit (ICU) patients to evaluate the association between the timing of hydrocortisone initiation and short- and long-term mortality in septic shock. We hypothesized that a longer duration between the first vasopressor use for sepsis and steroid initiation was associated with increased mortality.

Methods: Data were extracted from the Medical Information Mart in the Intensive Care-IV database. We included adults who met Sepsis-3 definition for septic shock and received hydrocortisone. The exposure of interest was the time in hours from vasopressor use to hydrocortisone initiation (>12 as late and ≤12 as early). The primary outcome was 1-year mortality. Secondary outcomes included 28-day mortality, 90-day mortality, in-hospital mortality, and length of hospital stay. Cox proportional hazard models were used to estimate the association between exposure and mortality. Competing risk regression models were used to evaluate the association between exposure and length of hospital stay.

Results: A total of 844 patients were included in this cohort: 553 in the early group and 291 in the late group. The median time to hydrocortisone initiation was 7 hours (interquartile range, 2.0-19.0 hours). After multivariable Cox proportional hazard analysis, we found that hydrocortisone initiation >12 hours after vasopressor use was associated with increased 1-year mortality when compared with initiation <12 hours (adjusted hazard ratio, 1.39; 95% confidence interval, 1.13-1.71; P = .002, E-value = 2.13). Hydrocortisone initiation >12 hours was also associated with increased 28-day, 90-day, and in-hospital mortality and prolonged length of hospital stay.

Conclusions: In patients with septic shock, initiating hydrocortisone >12 hours after vasopressor use was associated with an increased risk of both short-term and long-term mortality, and a prolonged length of hospital stay.