Multivariate logistic regression analysis of anti-PF4/heparin antibody seroconversion
| Predictors . | Univariate analysis . | Multivariate analysis . | ||||
|---|---|---|---|---|---|---|
| OR . | 95% CI . | P . | OR . | 95% CI . | P . | |
| Female gender | 3.09 | 1.86-5.13 | <.001 | 3.19 | 1.91-5.34 | <.001 |
| Rheumatoid arthritis | 0.64 | 0.40-1.02 | .058 | 0.52 | 0.32-0.84 | .008 |
| History of venous thrombosis | 3.23 | 1.22-8.59 | .025 | 2.61 | 0.93-7.29 | .067 |
| TKA surgery | 1.77 | 1.33-2.35 | <.001 | 1.76 | 1.31-2.38 | <.001 |
| Spinal anesthesia | 1.58 | 1.17-2.13 | .003 | 1.43 | 1.03-1.97 | .031 |
| Pharmacologic thromboprophylaxis | ||||||
| UFH | 1.41 | 0.82-2.41 | .212 | 1.76 | 0.98-3.15 | .059 |
| LMWH | 0.84 | 0.59-1.20 | .330 | 1.00 | 0.67-1.48 | .999 |
| Fondaparinux | 1.60 | 1.22-2.11 | .001 | 1.48 | 1.08-2.02 | .014 |
| Mechanical thromboprophylaxis | ||||||
| Use of GCSs | 0.37 | 0.27-0.50 | <.001 | 0.48 | 0.34-0.68 | <.001 |
| Use of DMT (foot pump or IPCD) | 2.42 | 1.65-3.55 | <.001 | 2.01 | 1.34-3.02 | .001 |
| Predictors . | Univariate analysis . | Multivariate analysis . | ||||
|---|---|---|---|---|---|---|
| OR . | 95% CI . | P . | OR . | 95% CI . | P . | |
| Female gender | 3.09 | 1.86-5.13 | <.001 | 3.19 | 1.91-5.34 | <.001 |
| Rheumatoid arthritis | 0.64 | 0.40-1.02 | .058 | 0.52 | 0.32-0.84 | .008 |
| History of venous thrombosis | 3.23 | 1.22-8.59 | .025 | 2.61 | 0.93-7.29 | .067 |
| TKA surgery | 1.77 | 1.33-2.35 | <.001 | 1.76 | 1.31-2.38 | <.001 |
| Spinal anesthesia | 1.58 | 1.17-2.13 | .003 | 1.43 | 1.03-1.97 | .031 |
| Pharmacologic thromboprophylaxis | ||||||
| UFH | 1.41 | 0.82-2.41 | .212 | 1.76 | 0.98-3.15 | .059 |
| LMWH | 0.84 | 0.59-1.20 | .330 | 1.00 | 0.67-1.48 | .999 |
| Fondaparinux | 1.60 | 1.22-2.11 | .001 | 1.48 | 1.08-2.02 | .014 |
| Mechanical thromboprophylaxis | ||||||
| Use of GCSs | 0.37 | 0.27-0.50 | <.001 | 0.48 | 0.34-0.68 | <.001 |
| Use of DMT (foot pump or IPCD) | 2.42 | 1.65-3.55 | <.001 | 2.01 | 1.34-3.02 | .001 |
Multivariate logistic regression was performed to identify risk factors independently associated with anti-PF4/heparin antibody seroconversion after controlling simultaneously for potential confounders. Variables, which were selected by univariate logistic regression analysis with a P value <.2 using the χ2 test or Fisher’s exact test (see Table 1), were included into a multivariate logistic regression model with stepwise forward selection method with forced entry of the variables gender, surgical type, and each pharmacologic prophylaxis, which were identified as risk factors for anti-PF4/heparin antibody formation in the previous studies. A 2-tailed P value of <.05 was considered significant.