Predictors of prolonged vasopressin infusion for the treatment of septic shock.

PURPOSE: Prolonged catecholamine use has been linked with poor clinical outcomes, including higher mortality. The objective was to identify characteristics that may be predictive of prolonged arginine vasopressin (AVP) use for 7 days or more in patients with septic shock. MATERIALS AND METHODS: This was a retrospective nested cohort analysis of adult patients receiving AVP as initial hemodynamic support for septic shock, either alone or in combination with norepinephrine, between 2008 and 2010. RESULTS: Univariate factors predictive of patients requiring extended AVP support were peripheral vascular disease (PVD) (48% vs 18%, P = .001), congestive heart failure (30% vs 12%, P = .024), and acute kidney injury (AKI) (83% vs 49%, P = .003). Patients requiring extended AVP support more frequently experienced a new intensive care unit (ICU) arrhythmia, typically atrial fibrillation (39% vs 7%, P < .001), and had higher 28-day mortality (74% vs 20%, P < .001). Multivariate analysis revealed that the strongest independent predictors of prolonged AVP dependence were new ICU arrhythmia (odds ratio [OR], 5.3; 95% confidence interval [CI], 1.6-17.8), PVD (OR, 4.3; 95% CI, 1.4-13.1), and AKI (OR, 3.9; 95% CI, 1.1-14.5). CONCLUSIONS: Patients with preexisting PVD and AKI and those experiencing a new ICU arrhythmia on AVP may be more likely to remain on AVP for 7 or more days.