ABO phenotype splits bleeding risk from long-term survival after CABG
In a retrospective multicenter cohort of 2,836 patients undergoing isolated coronary artery bypass grafting, ABO blood-group phenotype tracked with two different postoperative trajectories: early bleeding and long-term mortality. Blood group O carried the greater 30-day bleeding burden, while blood group A was linked to higher adjusted all-cause mortality over follow-up. The contrast is clinically uncomfortable because the same inherited marker that may favor one perioperative outcome appears to disfavor another.
Follow-up was substantial. The reverse Kaplan–Meier median was 7.8 years, with 277 deaths and 229 bleeding events captured across the cohort. Those numbers give the analysis enough duration to move beyond the immediate postoperative window and into the longer arc of survival after surgery, where platelet biology, vascular risk, and background cardiovascular disease all continue to exert pressure on outcomes. The study examined blood groups A, AB, B, and O, using O as the reference category.
The mortality signal was concentrated in blood group A. During follow-up, deaths occurred in 130 of 1,104 patients with blood group A, 25 of 246 with AB, 50 of 533 with B, and 72 of 953 with O. After multivariable adjustment, blood group A was associated with higher long-term all-cause mortality than blood group O, with an adjusted hazard ratio of 1.50 and a 95% CI of 1.12 to 2.00. The global ABO comparison reached statistical significance as well. In practical terms, a common inherited phenotype tracked with a measurable difference in survival after bypass surgery, even after accounting for other factors in the model.
One sentence stands out sharply: blood group O was the bleeding-heavy group, not the mortality-heavy one.
Hemostatic biology offers a plausible explanation for the early postoperative pattern
The 30-day bleeding data moved in the opposite direction from the mortality data. Bleeding occurred in 103 of 953 patients with blood group O, compared with 77 of 1,104 in group A, 10 of 246 in AB, and 39 of 533 in group B. Adjusted hazard ratios for 30-day bleeding were 0.64 for A, 0.35 for AB, and 0.67 for B, each relative to O, with a global ABO _P_ value below 0.001. O blood type has long been associated with lower levels of certain von Willebrand factor-related hemostatic proteins and a different balance of coagulation activity, which offers a biologically coherent route to more postoperative bleeding after a procedure as hemostatically demanding as CABG. In a surgery where chest tube output, transfusion decisions, re-exploration thresholds, and anticoagulation management already occupy narrow margins, inherited differences in clot formation can become visible quickly.
The early bleeding pattern deserves attention because it is not merely a laboratory curiosity. After CABG, bleeding can amplify transfusion exposure, prolong intensive care stays, increase reintervention risk, and complicate recovery in ways that are difficult to fully erase later. The analysis does not quantify those downstream effects here, but the observed distribution of 30-day bleeding events suggests that ABO-related coagulation differences may be relevant at the bedside long before any late survival signal becomes apparent.
No convincing heterogeneity emerged in the exploratory subgroup analyses by age, sex, surgical urgency, or diabetes status, although the number of bleeding events limited the power of those comparisons. The absence of obvious effect modification leaves the core pattern intact across the examined strata, but it does not convert the association into a decision rule. The study was observational, so residual confounding remains possible, and the authors were explicit that the findings are hypothesis-generating rather than proof of clinical utility.
What the study can and cannot support in current CABG practice
For surgeons, anesthesiologists, and perioperative teams, the immediate question is whether ABO phenotype should influence planning. On the evidence presented here, the answer is no. The analysis does not establish incremental predictive value, and it does not show that knowing blood group changes outcomes when added to standard risk assessment. Blood type may be biologically informative, but that is not the same as clinically actionable.
Still, the study adds weight to a growing view that inherited hematologic traits can shape different phases of cardiac surgical risk. The early postoperative signal aligns with coagulation biology; the later mortality signal may reflect broader vascular or thrombotic mechanisms, background comorbidity burden, or an interaction between phenotype and long-term cardiovascular risk. Because the study included isolated CABG only, the findings should not be generalized beyond that population without caution. The cohort also spanned 2010 to 2025, a period during which surgical practice, transfusion thresholds, and perioperative management may have changed.
A final limitation sits in plain sight: the analysis identifies association, not causation. Without prospective validation, external replication, and a demonstration that ABO-informed risk stratification improves care, the result remains an interesting signal rather than a bedside tool. Even so, the pattern is hard to ignore—blood group O tracked with more early bleeding, and blood group A tracked with worse long-term survival after isolated CABG.