Baseline cortisol levels, total proteins, and eosinophil count as predictors of hemodynamic response to steroid treatment in septic shock.

Background: Our aim was to study if baseline serum cortisol is related to the hemodynamic response to steroid treatment in septic shock patients and if the measurement of total proteins and eosinophil count improves its accuracy.

Methods: A retrospective analysis was performed in 66 consecutive surgical septic shock patients receiving steroid treatment. Four criteria were chosen to define hemodynamic improvement based on the combination of noradrenaline (NA) withdrawal (at 24 and 48 hour) and an increase of the hemodynamic index (HI = mean arterial pressure/NA dose) of 150% at 24 hour and of 350% at 48 hour. The accuracy of the serum cortisol to predict the hemodynamic response to steroid treatment following the four criteria was determined by receiver operating characteristic curve analysis.

Results: The largest area under curve was found for the NA withdrawal or an increase of the hemodynamic index >350% at 48 hour after starting the steroid treatment (area under curve, 0.686; 95% CI, 0.553-0.819; p = 0.01). This criteria was met by 35 patients (53%) and was associated with a lower mortality (25.7% vs. 67.7%, p = 0.001). However, no clear serum cortisol cutoff value for the diagnosis of adrenal insufficiency based on the hemodynamic response could be found. Neither the baseline proteins nor the eosinophils improved the accuracy of cortisol to predict hemodynamic improvement.

Conclusion: The measurement of serum cortisol in surgical septic shock patients does not accurately predict the hemodynamic response to steroids. No clear cutoff value for cortisol, alone or in combination with total protein and/or eosinophil count, can be defined to indicate steroid treatment.