Abstract
PURPOSE: It is not clear whether vasopressin versus norepinephrine changed mortality in clinical practice in the Vasopressin and Septic Shock Trial (VASST) coordinating center hospital after VASST was published. We tested the hypothesis that vasopressin changed mortality compared to norepinephrine using propensity matching of vasopressin to norepinephrine-treated patients in the VASST coordinating center hospital before (SPH1) and after (SPH2) VASST was published. METHODS: Vasopressin-treated patients were propensity score matched to norepinephrine-treated patients based on age, APACHE II, respiratory, renal, and hematologic dysfunction, mechanical ventilation status, medical/surgical status, infection site, and norepinephrine dose. The propensity score estimated the probability that a patient would have received vasopressin given baseline characteristics. For sensitivity analysis, we then excluded patients who had underlying severe congestive heart failure. The primary outcome was 28-day mortality. RESULTS: Vasopressin- and norepinephrine-treated patients were similar after matching in SPH1 (pre-VASST); vasopressin-treated patients (n = 158) had a significantly higher mortality than norepinephrine-treated patients (n = 158) (60.8 vs. 46.2%, p = 0.009). In SPH2 after matching, the 28-day mortality rates were not significantly different; 31.2% and 26.9% in the vasopressin (n = 93) and norepinephrine (n = 93) groups, respectively (p = 0.518). The day 1 vasopressin dose in SPH1 vs. SPH2 was 0.036 units/min (SD 0.009) vs. 0.032 units/min (SD 0.005), p = 0.001, significantly lower in SPH2 after VASST. CONCLUSIONS: Before VASST, vasopressin use was associated with increased mortality compared to norepinephrine in the VASST coordinating center hospital. After VASST, there was no difference in mortality between vasopressin- and norepinephrine-treated patients. This may be the first retrospective propensity-matched cohort study of a sepsis treatment in a center that had previously coordinated a large pivotal randomized controlled trial of that treatment and could be a useful approach for other sepsis therapies. TRIAL REGISTRATION: Registration: ISRCTN94845869.