Nesecc.org

NeSECC editie 2007 nr. 2; p.22
Introduction
In 2006, a blood conservation study was used. Binary variables were analyses with initiated in the Amphia hospital in Breda, logistic regression analysis; for these vari- The Netherlands, with the primary aim of ables, the odds ratio and 95% confidence achieving an increase of transfusion free interval are reported in addition to p-val- patients with 30-50% and a reduction of ue.
transfusion of al ogeneic blood products Surgical process variables
verse clinical outcomes, such as postop- More off pumps and higher operative tem- erative infarction, renal failure, infection, perature. Similar aorta occlusion times prolonged length of stay, and death.
A blood conservation protocol was intro- within non off pumps. Large increase of duced at May 17th 2006 as a guideline for use of cell saver, no increase of cell saved blood conservation and transfusion prac- volume given use of cell saver. Increase of tice in the care process of cardiac surgery. patients receiving autologous blood trans- Patient data
In this study, 517 patients who underwent fusion volume given this transfusion. a cardiac surgical intervention in the pe- Medication: decrease in patients receiv- riods of May 17th to September 30th 2006 ing aprotinine, large increase in patients form the blood conservation group. We receiving tranexamic acid or cyklolapron. compared the data of this patient group No change in given heparin amount, de- retrospectively with data of 580 patients crease in heparin/protamine ratio, so an who underwent in the period of May 17th increase in protamine amount given the – September 30th 2005 (control group). Data analysis
Transfusion related outcome
The preoperative characteristics of both Decrease in amount of blood loss; de- patient groups were compared. To ana- crease of patients with major blood loss. lyze the differences in categorical data, Large increase of allogeneic blood trans- we used the Pearson Chi-squared test, or fusion free patients (54%), decrease on Fisher’s exact test where any expected cell amount of allogeneic blood transfusion count was less than 5. Continuous vari- volume (37% in median), decrease of ables were analyzed using the unpaired patients who received major allogeneic Student t-test where a Gaussian distribu- blood transfusion volume. Decrease of al- tion could be demonstrated; otherwise logeneic blood transfusion volume given no major blood loss; no decrease of alloge- Subsequently, we compared the surgical neic blood transfusion volume in patients process and transfusion related variables with major blood loss. No decrease in he- and clinical outcomes for both patients moglobin and hematocrit values at several groups. Differences in the distribution of time points in the process. No difference continuous variables were analyzed using in inotropics and vasopressin use.
the unpaired Student t-test where a Gaus- sian distribution could be demonstrated; NeSECC editie 2007 nr. 2; p.23
PRE (n=580)
POST (n=517)
General
Age (mean±sd)
Co-morbidity
EuroSCORE (median, interquartile range)
Preoperative treatment
Aspirin and clopidogrel < 5 days before surgery (%)
Table 1: Comparison of baseline characteristics between the control group (PRE) and the blood conservation Figure 1. Amphia transfusion tree model (n=3803) NeSECC editie 2007 nr. 2; p.24
Clinical outcomes
risk of 0.77 to male patients with a simi- No differences in outcomes, with one ma- lar hemoglobin value and an increased jor exception: large increase of occurrence EuroSCORE (> 3). Lowest transfusion of infections. Figure 1 shows the Amphia risk (0.21) was found for relatively young transfusion risk tree model. The tree male patients (< 70) undergoing isolated model includes preoperative hemoglobin CABG surgery with a good preopera- as primary split. In addition, the variables tive hemoglobin value (> 8.4), a relatively EuroSCORE, female, body mass index, high body mass index (> 25), and without age, and isolated CABG or valve surgery a high EuroSCORE (< 8). The indicator were included as predictive features. The variable of the use of the blood conserva- tree model assigns a transfusion risk of tion protocol was not included in the tree 0.90 to female patients with a low hemo- model, though. globin value (< 8.4), and a transfusion

Source: http://www.nesecc.org/fjc_documents/0702_a4.pdf