K.K. is supported in part by the Canadian Institutes of Health Research and the Canadian Blood Services. T.M.Y. is supported in part by the Canadian Institutes of Health Research and the Heart and Stroke Foundation of Ontario.
The independent association of massive blood loss with mortality in cardiac surgery
Version of Record online: 22 SEP 2004
Volume 44, Issue 10, pages 1453–1462, October 2004
How to Cite
Karkouti, K., Wijeysundera, D. N., Yau, T. M., Beattie, W. S., Abdelnaem, E., McCluskey, S. A., Ghannam, M., Yeo, E., Djaiani, G. and Karski, J. (2004), The independent association of massive blood loss with mortality in cardiac surgery. Transfusion, 44: 1453–1462. doi: 10.1111/j.1537-2995.2004.04144.x
- Issue online: 22 SEP 2004
- Version of Record online: 22 SEP 2004
- Received for publication April 22, 2004; revision received May 31, 2004, and accepted June 2, 2004.
BACKGROUND: Although the association between massive perioperative blood loss (MBL) and adverse outcomes is well recognized, it is unclear whether MBL is an independent risk factor or, instead, simply a marker for other adverse events or severity of illness. The objective of this cohort study was to quantify the independent association of MBL in cardiac surgery with all-cause in-hospital mortality.
STUDY DESIGN AND METHODS: Data were prospectively collected on consecutive patients who underwent cardiac surgery with cardiopulmonary bypass at a quaternary-care academic center from 1999 to 2003. The number of red blood cell (RBC) units transfused within 1 day of surgery was used as a surrogate measure of perioperative blood loss. Receiver-operating characteristic curve analyses were employed to identify the most appropriate cutoff for defining MBL. The independent association of MBL with mortality was determined with multivariable logistic regression analyses. Bootstrapping and sensitivity analyses were used to confirm the validity of the results.
RESULTS: MBL was defined as receiving at least 5 units of RBCs within 1 day of surgery. Of 9215 patients analyzed, 1.8 percent (n = 169) died and 9.7 percent (n = 890) had MBL. After adjusting for multiple potential confounders (including perioperative adverse events), MBL was associated with an 8.1-fold (95% confidence interval, 3.9-17.0) increase in the odds of death. This risk estimate was stable across different modeling conditions as well as in bootstrap sampling.
CONCLUSION: MBL after cardiac surgery has a strong, independent association with in-hospital mortality.