Akintoye · Circulation. Cardiovascular quality and outcomes 2018 · multicenter randomized placebo-controlled trial · n=1516

Fish Oil and Perioperative Bleeding.

Cited 73 times in the scientific literature.

Level 2 - randomized trial

Individual randomized controlled trial

PubMed 30571332 · doi:10.1161/CIRCOUTCOMES.118.004584 · record verified 2026-08-30

What was done

A multinational, placebo-controlled randomized trial evaluated 1,516 patients (mean age 63 years) undergoing planned cardiac surgery (52% coronary artery bypass graft, 50% valve surgery). Participants were randomized to perioperative fish oil (eicosapentaenoic acid plus docosahexaenoic acid; 8–10 g/day for 2–5 days preoperatively, followed by 2 g/day postoperatively) or placebo. The primary outcome was major perioperative bleeding defined by the Bleeding Academic Research Consortium (BARC). Secondary outcomes included bleeding by TIMI and ISTH definitions, chest tube output, and total units of blood transfused. Achieved plasma phospholipid omega-3 fatty acid levels were assessed on the morning of surgery.

What was found

Major BARC bleeding occurred in 92 patients (6.1%). Compared with placebo, fish oil did not increase BARC bleeding risk (odds ratio 0.81, 95% CI 0.53–1.24; absolute risk difference -1.1%, 95% CI -3.0% to 1.8%). Findings were similar for secondary bleeding definitions. Total blood units transfused were significantly lower in the fish oil group than the placebo group (mean 1.61 vs 1.92 units, P < 0.001). Higher achieved morning-of-surgery omega-3 levels were associated with lower BARC bleeding risk compared with the lowest quartile (third quartile OR 0.30, 95% CI 0.11–0.78; fourth quartile OR 0.36, 95% CI 0.15–0.87).

Why it matters

These findings counter common clinical recommendations to discontinue fish oil prior to surgery, indicating high-dose perioperative fish oil does not raise bleeding risk in cardiac surgery and may reduce transfusion needs.

Limits

The trial only evaluated patients undergoing cardiac surgery, limiting generalizability to non-cardiac or emergency operations. Preoperative dosing was short-term (2–5 days), which may not capture the effects of chronic use. Numeric results for chest tube output and secondary bleeding definitions were not reported in the abstract.

Cited by