Plasma fibrinogen level and severe bleeding in cardiac surgery: an observational post hoc study of the ALBICS trial

Fibrinogen is widely used in cardiac surgery. Still, reported associations between plasma fibrinogen level and chest tube drainage are weak. The results of trials on fibrinogen supplementation are contradictory. We aimed to investigate how the plasma fibrinogen level relates to severe bleeding and resternotomy for bleeding in patients undergoing on-pump cardiac surgery.

We conducted an observational post hoc study of 1,386 patients undergoing on-pump cardiac surgery enrolled in the Albumin in Cardiac Surgery (ALBICS) trial. We assessed severe bleeding with the Universal Definition of Perioperative Bleeding classification (UDPB), categorized as "UDPB-low" (classes 0-2) and "UDPB-high" (classes 3-4) and as resternotomy. We measured plasma fibrinogen levels preoperatively and 30 min after protamine administration ("post-cardiopulmonary bypass [CPB]").

The incidences of UDPB-high and resternotomy were 8.1% (112/1,386) and 3.6% (50/1,386). No patient with preoperative a fibrinogen level > 4.7 g·L^-1 (90/1,386; 6.5%) had UDPB-high or resternotomy. After adjustment for hemostatic laboratory values, preoperative fibrinogen predicted UDPB-high (odds ratio [OR], 0.77; 95% confidence interval [CI], 0.60 to 0.98; per standard deviation [SD] [0.9 g·L^-1]) and resternotomy (OR, 0.65; 95% CI, 0.43 to 0.98; per SD [0.9 g·L^-1]). No patient with a post-CPB fibrinogen level > 3.1 g·L^-1 (73/1,386; 5.3%) had UDPB-high or required resternotomy. Post-CPB fibrinogen predicted UDPB-high (OR, 0.51; 95% CI, 0.33 to 0.77; per preoperative SD [0.9 g·L^-1]) and resternotomy (OR, 0.31; 95% CI, 0.16 to 0.62; per preoperative SD [0.9 g·L^-1]).

The preoperative fibrinogen level had borderline and the post-CPB fibrinogen level had strong associations with severe bleeding and resternotomy in patients undergoing on-pump cardiac surgery. Further research is required to delineate whether the observed association represents a cause-and-effect relationship.