Platelet Transfusion in Cardiac Surgery: A Systematic Review and Meta-Analysis

Background

Blood transfusion is a well-established independent risk factor for mortality in patients undergoing cardiac surgery but the impact of platelet transfusion is less clear. We performed a systematic review and meta-analysis of observational studies comparing outcomes of patients who received platelet transfusion after cardiac surgery.

Methods

We searched MEDLINE and EMBASE databases to January 2019 for studies comparing perioperative outcomes in patients undergoing cardiac surgery with and without platelet transfusion.

Results

There were nine observational studies reporting 101,511 patients: 12% with and 88% without platelet transfusion. In unmatched/unadjusted studies, patients who received platelet transfusion were older, with greater incidence of renal, peripheral, and cerebrovascular disease, myocardial infarction, left ventricular dysfunction, and anemia. They were more likely to have nonelective, combined surgery; preoperative hemodynamic instability and endocarditis; and more likely to be receiving clopidogrel preoperatively. Perioperative complications were significantly increased without adjusting for these baseline differences. After pooling only matched/adjusted data, differences were not found between patients who did receive platelets and patients who did not in operative mortality (risk ratio [RR] 1.26; 95% confidence interval [CI], 0.69 to 2.32, P = .46, five studies), stroke (RR 0.94; 95% CI, 0.62 to 1.45; P = .79; five studies), myocardial infarction (RR 1.29; 95% CI, 0.95 to 1.77; P = .11; three studies), reoperation for bleeding (RR 1.20; 95% CI, 0.46 to 3.18; P = .71; three studies), infection (RR 1.02; 95% CI, 0.86 to 1.20; P = .85; six studies); and perioperative dialysis (RR 0.91; 95% CI, 0.63 to 1.32; P = .62; three studies).

Conclusions

After accounting for baseline differences, platelet transfusion was not linked with perioperative complications in cardiac surgery patients. Given the small number of observational studies, these findings should be considered hypothesis generating.