Randomized Clinical Trial (RCT) evaluating the effects of low-dose prolonged hydrocortisone infusion on resolution of MODS in severe sepsis.
Patients: severe sepsis < 48h ICU entry - stratified by group (A without shock and MODS < 2; B with shock or MODS > 3). Treatment: Hydrocortisone infusion (10 mg/hr) for 7 days.Primary end point: group A improvement in MODS; group B resolution of shock by day 7.
Eighty patients: 48 treated vs. 32 placebo (Group A: 17 vs. 13; Group B: 31 vs. 19). Baseline differences (treated vs. control) included (i) higher APACHE III scores (75± 25 vs. 66 ± 11, p = 0.03) and (ii) a trend to higher rates of baseline mechanical ventilation (79% versus 59%, p = 0.06). There was no difference in baseline adrenal insufficiency (35% versus 31%, p = 0.74). Treated group had a significant reduction in median day 7 CRP levels (1.9 (23) versus 13.6 (12), p < 0.0001), duration of mechanical ventilation (5 (5) versus 10 (12), p=0.006), ICU duration (6 (6) versus 12 (18), 0=0.006, and hospital stay (9.5 (8) versus 19 (14), p=0.0005). Despite these highly favorable surrogate endpoints, the ITT analysis demonstrated that treatment with hydrocortisone resulted in worse ICU survival (62.5% versus 81.3%, p=0.07) and hospital survival (52.1% versus 81.3%, p=0.008). In per-protocol analysis, causes of deaths included: 8 progression of shock or MODS, 6 sudden cardiac death (SCD) in patients with ASCVD, and 7 progression of underlying disease.
These findings are limited by a small sample sizes, significant baseline imbalances, unusually excellent outcome in the placebo group, and the significant co-morbidities in the treated group. The potential effect of rebound inflammation in this group of patients was previously reported (Nawab et al. Am J Respir Crit Care Med 2007; 175: A594).
In sepsis trials, mortality can be affected by factors not related to the acute infection per se. Mortality outcome trials should incorporate an expanded description of mortality causes.
Gianfranco Meduri, No Financial Disclosure Information; No Product/Research Disclosure Information