Elevated Plasma-Free Cortisol Concentrations and Ratios Are Associated With Increased Mortality Even in the Presence of Statin Therapy in Patients With Severe Sepsis*

Background: Dissociation between plasma-free cortisol and total cortisol profiles exists in critical illness. Data on plasma-free cortisol are based on either calculated values or immunoassay-based measurements. Both have significant limitations. Statins have been advocated as a therapy in sepsis. Whether they impact on plasma cortisol through inhibition of cholesterol synthesis is unclear. Objectives: In patients enrolled into a randomized trial of statins in sepsis (n = 250), we examined the association of mass spectrometry measured plasma-free cortisol, plasma-free cortisol/plasma total cortisol ratios, and outcome and the impact of concomitant statin therapy on cortisol profiles in 80 steroid naïve patients. Patients: Two hundred twenty serial measurements of plasma-free cortisol, plasma total cortisol, and interleukin-6 were collected from 80 patients (43 placebo and 37 statins). Data from 10 volunteers were used as controls. Measurements and Main Results: Data are presented as median and interquartile range. Compared with controls, in severe sepsis, baseline plasma total cortisol was elevated two-fold (463 nmol/L [284–742 nmol/L] vs 245 nmol/L [200–299 nmol/L], p < 0.001), plasma-free cortisol 20-fold (75 nmol/L [20–151 nmol/L] vs 5 nmol/L [5–7 nmol/L], p < 0.001), and plasma-free cortisol/plasma total cortisol ratio six-fold (0.15 vs 0.02, p = 0.058). Baseline interleukin-6 was elevated at 121 pg/mL (65–611 pg/mL). In severe sepsis, there were no differences in plasma total cortisol (p = 0.66), plasma-free cortisol (p = 0.77), and interleukin-6 (p = 0.29) between statins and placebo groups. Plasma-free cortisol, plasma total cortisol, and plasma-free cortisol/plasma total cortisol were positively correlated with interleukin-6 (p = 0.0001, p < 0.0004, and p < 0.001, respectively) and day 90 mortality (p = 0.03, p = 0.03, and p = 0.058, respectively). Elevated plasma-free cortisol/plasma total cortisol ratios were associated with increased length of stay (p = 0.04). Baseline plasma-free cortisol, plasma total cortisol, and plasma-free cortisol/plasma total cortisol ratios were higher in nonsurvivors as compared with survivors (174 nmol/L [77–329 nmol/L] vs 57 nmol/L [17–122 nmol/L], p = 0.016; 890 nmol/L [333–1,430 nmol/L] vs 408 nmol/L [269–681 nmol/L], p = 0.035; and 0.19 [0.13–0.29] vs 0.14 [0.07–0.20]; p = 0.054, respectively). Conclusions: In severe sepsis, plasma-free cortisol increase is 10-fold greater than that of plasma total cortisol. Both are similarly associated with inflammatory response and mortality. Elevated plasma-free cortisol/plasma total cortisol ratios were associated with increased length of stay. Statin therapy does not influence the plasma cortisol profiles in patients with severe sepsis.