Interaction of vasopressin infusion, corticosteroid treatment, and mortality of septic shock*

OBJECTIVE: Vasopressin and corticosteroids are often added to support cardiovascular dysfunction in patients who have septic shock that is nonresponsive to fluid resuscitation and norepinephrine infusion. However, it is unknown whether vasopressin treatment interacts with corticosteroid treatment. DESIGN: Post hoc substudy of a multicenter randomized blinded controlled trial of vasopressin vs. norepinephrine in septic shock. SETTING: Twenty-seven Intensive Care Units in Canada, Australia, and the United States. PATIENTS: : Seven hundred and seventy-nine patients who had septic shock and were ongoing hypotension requiring at least 5 microg/min of norepinephrine infusion for 6 hours. INTERVENTIONS: Patients were randomized to blinded vasopressin (0.01-0.03 units/min) or norepinephrine (5-15 microg/min) infusion added to open-label vasopressors. Corticosteroids were given according to clinical judgment at any time in the 28-day postrandomization period. MEASUREMENTS: The primary end point was 28-day mortality. We tested for interaction between vasopressin treatment and corticosteroid treatment using logistic regression. Secondary end points were organ dysfunction, use of open-label vasopressors and vasopressin levels. MAIN RESULTS: There was a statistically significant interaction between vasopressin infusion and corticosteroid treatment (p = 0.008). In patients who had septic shock and were also treated with corticosteroids, vasopressin, compared to norepinephrine, was associated with significantly decreased mortality (35.9% vs. 44.7%, respectively, p = 0.03). In contrast, in patients who did not receive corticosteroids, vasopressin was associated with increased mortality compared with norepinephrine (33.7% vs. 21.3%, respectively, p = 0.06). In patients who received vasopressin infusion, use of corticosteroids significantly increased plasma vasopressin levels by 33% at 6 hours (p = 0.006) to 67% at 24 hours (p = 0.025) compared with patients who did not receive corticosteroids. CONCLUSIONS: There is a statistically significant interaction between vasopressin and corticosteroids. The combination of low-dose vasopressin and corticosteroids was associated with decreased mortality and organ dysfunction compared with norepinephrine and corticosteroids.