Background—This observational study sought to identify the off-label use pattern of recombinant activated factor VII (rFVIIa) in cardiac surgery and to identify predictors of its effectiveness and risk.

Methods and Results—At 18 Canadian centers, 522 nonhemophiliac cardiac surgical patients received rFVIIa during the period 2003 through 2006; data were available, and retrospectively collected, on 503 patients. The median (quartile 1, quartile 3) units of red blood cells transfused from surgery to therapy and in the 24 hours after therapy were 8 (5, 12) and 2 (1, 5), respectively (P<0.0001). Mortality rate was 32%, and mortality or major morbidity rate was 44%. These rates were within expected ranges (mortality, 27% to 35%; mortality or morbidity, 39% to 48%), which were calculated with a separate cohort of cardiac surgical patients who did not receive rFVIIa used as reference. Independent predictors of complications included instability before therapy (multiple inotropes or intra-aortic balloon pump) and increasing red blood cell units transfused before and after therapy. Variables independently associated with nonresponse included abnormal coagulation parameters and >15 red blood cell units transfused before therapy.

Conclusions—In Canada, rFVIIa is used primarily when standard interventions have failed to control bleeding. In this setting, rFVIIa is associated with reduced blood product transfusions and, after risk adjustment, does not appear to be associated with increased or decreased complication rates. The effectiveness of the drug may be enhanced if it is given early in the course of refractory blood loss in the setting of adequate amounts of circulating coagulation factors.