Premium
A point score system for predicting the likelihood of blood transfusion after hip or knee arthroplasty
Author(s) -
Larocque B. J.,
Gilbert K.,
Brien W. F.
Publication year - 1997
Publication title -
transfusion
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.045
H-Index - 132
eISSN - 1537-2995
pISSN - 0041-1132
DOI - 10.1046/j.1537-2995.1997.37597293874.x
Subject(s) - medicine , logistic regression , blood transfusion , arthroplasty , receiver operating characteristic , surgery , univariate analysis , multivariate analysis
BACKGROUND : Given the high cost of autologous blood donation for elective surgery, it would be desirable to predict which patients are most likely to benefit from the procedure. The purpose of this study was to develop a point score system for predicting the likelihood of blood transfusion in hip and knee arthroplasty. STUDY DESIGN AND METHODS : A database of 599 patients undergoing elective surgery at a teaching hospital was used for the analysis. Variables were analyzed to determine their univariate association with postoperative blood transfusion. Significant factors were entered into a multiple logistic regression model, and a point score system was developed on the basis of the regression coefficients. Four strata of transfusion risk were constructed. RESULTS : Factors independently associated with blood transfusion included preoperative hemoglobin, type of arthroplasty, primary versus revision surgery, autologous donor status, and patient weight. Four factors were used to create a point score system with four strata. The likelihood of blood transfusion for patients in the four risk strata was 1.7, 11.0, 40.0, and 78.3 percent. The calculated area under the receiver operating characteristic curve was 0.86. CONCLUSION : The likelihood of a postoperative blood transfusion can be predicted by using this simple point score system. Autologous blood donation can subsequently be targeted to the high‐risk patients.