Anaemia and postoperative blood loss are frequent in cardiac surgery. The efficacy of allogenic blood transfusion as an oxygen carrier, diminishes the longer it is storaged and some inmunomodulation effects have been described. Several recent studies show an increased mortality risk associated with transfusions. Accurate anticipation of blood needs is the first step toward a saving protocol and diverse independent predictive variables associated with transfusion needs have been described, but no universal model exists due to institutional facility differences, lack of agreement in thresholds, and heterogeneicity of studied variables. Based on evidence from recent studies, individualized predictive models are advised. Objectives of study: (1) to describe transfusional incidence; (2) to develop and validate a predictive rule to quantify the transfusional risk of off-pump cardiac surgery patients and (3) to obtain a prognostic index, easily applicable in the pre-operatory setting. Method: Logistic regression methology was applied on a retrospective cohort (n=310). Seven variables formed Sp_SinCEC predictive model: age, sex, weight, haemoglobin, creatinine, unstable angina and number of predicted bypasses. Sp_SinCEC was externally validated (ICC=0,94) on a prospective cohort (n=78). Cindex was 0,89. The model was then compared with others that have already been published: Karkouti et al (c-index= 0,85), TRUST (c-index=0,79) and TRACK (c-index=0,84) This tool (1) provides objective information about patients likelihood of needing blood transfusion; (2) should allow a more cost-efficient multimodal blood-saving resource assignation (3) and could be useful as a classifiying tool to assign patients on experimental clinical trials.
|Date of Award||2 Feb 2011|
|Supervisor||Francisco Javier Rius Cornado (Director), Ignacio Jose Gich Saladich (Director) & Eduardo Muñiz Diaz (Director)|