Studies in cardiac surgery and orthopedic surgery suggest that a lower transfusion trigger (Hgb 8.0 g/dL) can improve outcomes in patients. In surgical oncology, transfusion has been suggested to increase subsequent metastasis, either through immune modulation and/or microvascular sludging. The literature linking transfusion to metastasis is confounded by greater blood loss which may reflect greater tumor burden.
Researchers in Sao Paolo Brazil randomized ~200 patients undergoing major abdominal cancer surgery who were subsequently randomized to transfusion triggers of Hgb = 7.0 (restrictive) vs 9.0 g/dL(liberal).
Their results differed from the studies mentioned above. They found that the restrictive group had a significantly higher composite endpoint of major morbidity and mortality (19.6% in liberal group vs 35.6% in restrictive group). 30 day mortality was 8.2% (liberal) vs 22.8% (restrictive). The vast majority of deaths were from sepsis and multisystem organ failure.
I think the study was extremely well done. However, I question the high rates of mortality. While the authors quote European and South American series with similar death rates, the rates are higher than in some North American series. One big question: were these patients routinely admitted to ICU based on the planned operation, or was ICU admission a dynamic decision based upon intraoperative events (blood loss, transfusion and volume resuscitation, hypotension, length of surgery, etc). If it was routine admission, then I particularly believe the mortalities to be high.
The authors point out that their restrictive trigger of Hgb 7.0 g/dL might have been too low. As always, further studies (and attempts to replicate in other settings) will be useful.