Table 2.
Multivariable Analyses: Predictors of Hospital Mortality in Patients with a Primary Diagnosis of TTP (only subjects who underwent TPE are included in this analysis.)
| Predictor variables in the multivariable model | Odds Ratio Estimates | 95% Wald Confidence Limits | P-value | |
|---|---|---|---|---|
| Platelet Transfusion | 2.19 | 1.19 | 4.05 | <0.05 |
| Intracranial Hemorrhage | 6.05 | 1.58 | 23.24 | <0.01 |
| Arterial Thrombosis | 6.73 | 1.11 | 40.91 | <0.05 |
| Myocardial Infarction | 2.32 | 1.16 | 4.63 | <0.05 |
| Ischemic Stroke | 2.42 | 1.17 | 5.01 | <0.05 |
| Age (>=60 years) | 3.47 | 2.14 | 5.63 | <.0001 |
| Renal Failure | 2.56 | 1.46 | 4.47 | <0.01 |
| Non Central Nervous System Bleeding** | 1.21 | 0.62 | 2.38 | 0.57 |
| Gender** | 1.10 | 0.67 | 1.79 | 0.71 |
| Race (Caucasians versus African Americans)** | 1.47 | 0.88 | 2.45 | 0.14 |
Area under receiver operating characteristic curve: 78.6%
Significant factors are in boldface type. Model is also adjusted for Gender, Non Central Nervous System Bleeding and Race (Caucasian versus African American)
Good model calibration and fit per the hosmer-lemeshaw goodness-of-fit test