Table 4.
Independent predictors of transfusion. Multivariable logistic regression analysis identified several parameters. The strongest predictor was severe anemia with hematocrit level below 30% followed by congestive heart failure, dependent function status, preoperative coagulopathy and creatinine elevation > 1.2 mg/dl. Intriguingly, BMI > 30 kg/m2 was protective. These data were the basis for the development of a risk prediction model with a c-statistic for discrimination of 0.81. Model score points are shown for each predictor.
| Odds Ratio | 95% Confidence Interval | Model Score | |
|---|---|---|---|
| BMI > 30 kg/m2 | 0.46 | 0.29 – 0.70 | −2 |
| Dependent Functional Status | 2.7 | 1.5 – 5.0 | 2 |
| Female | 1.6 | 1.1 – 2.4 | 1 |
| Hematocrit Level | |||
| < 30 % | 57.1 | 29.5 – 110.7 | 8 |
| 30–40 % | 5.7 | 3.1 – 10.3 | 4 |
| 40+ % | reference | – | 0 |
| History of Congestive Heart Failure | 2.8 | 1.1 – 7.1 | 2 |
| Preoperative Coagulopathy | 2.5 | 1.7 – 3.6 | 2 |
| Preoperative Creatinine > 1.2 mg/dl | 2.3 | 1.6 – 3.3 | 2 |
| Preoperative Dyspnea | 2 | 1.3 – 3.0 | 1 |