Background: Endocannabinoids (ECs) are rapidly acting immune-modulatory lipid-signaling molecules that are important for adaptation to stressful and aversive situations. They are known to interact with glucocorticoids and other stress-responsive systems. Maladaptation to acute or chronic stress represents a major risk factor for the development of psychiatric disorders. In the present study, we administered stress doses of hydrocortisone in a prospective, randomized, placebo-controlled double-blind study in patients undergoing cardiac surgery (CS) to examine the relationship between the use of glucocorticoids, plasma EC levels, and the occurrence of early postoperative cognitive dysfunction (delirium) and of later development of depression. Methods: We determined plasma levels of the ECs anandamide and 2-arachidonoylglycerol (2-AG) in CS patients of the hydrocortisone (n=56) and the placebo group (n=55) preoperatively, at postoperative day (POD) 1, at intensive care unit discharge, and at 6 months after CS (n=68). Postoperative delirium was diagnosed according to Diagnostic and Statistical Manual of the American Psychiatric Association IVth Edition (DSM-IV) criteria, and depression was determined by validated questionnaires and a standardized psychological interview (Structured Clinical Interview for DSM-IV). Results: Stress doses of hydrocortisone did not affect plasma EC levels and the occurrence of delirium or depression. However, patients who developed delirium on POD 1 had significantly lower preoperative 2-AG levels of the neuroprotective EC 2-AG (median values, 3.8 vs. 11.3 ng/ml; p=0.03). Preoperative 2-AG concentrations were predictive of postoperative delirium (sensitivity=0.70; specificity=0.69; cutoff value=4.9 ng/ml; receiver operating characteristic curve area=0.70; 95% confidence interval=0.54-0.85). Patients with depression at 6 months after CS (n=16) had significantly lower anandamide and 2-AG levels during the perioperative period. Conclusions: A low perioperative EC response may indicate an increased risk for early cognitive dysfunction and long-term depression in patients after CS. Glucocorticoids do not seem to influence this relationship.