Table 2.
Comparison of ROC and MPM thresholds using logistic regression analysis
| Models and predictors | OR (95 % CI) | p | AUC | AIC |
|---|---|---|---|---|
| ROC threshold | ||||
| BNP ≥ 116 pg/ml or NTproBNP ≥ 277.5 pg/ml; | 8.4 (4.98,14.11) | <0.001 | 0.768 | 431.63 |
| Ref = BNP < 116 or NTproBNP < 277.5 pg/ml | ||||
| Surgery type | ||||
| Aortoiliac; Ref = Infrainguinal surgery | 2.4 (1.33, 4.37) | 0.004 | ||
| Diabetes | ||||
| Yes; Ref = No | 2.0 (1.10, 3.54) | 0.023 | ||
| MPM threshold | ||||
| BNP ≥ 115.57 pg/ml or NTproBNP ≥ 246.7 pg/ml; | 8.5 (5.03, 14.41) | <0.001 | 0.777 | 430.16 |
| Ref = BNP < 115.57 pg/ml or NTproBNP < 246.7 pg/ml | ||||
| Surgery type | ||||
| Aortoiliac; Ref = Infrainguinal surgery | 2.5 (1.40, 4.60) | 0.002 | ||
| Diabetes | ||||
| Yes; Ref = No | 2.1 (1.15, 3.71) | 0.015 | ||
ROC threshold is an indicator variable of BNP and NTproBNP thresholds determined by the ROC curve method (Allison 2012); MPM threshold is an indicator variable of BNP and NTproBNP thresholds determined by the minimum p value method
Ref reference level, OR odds ratio, CI confidence interval, AUC area under the ROC curve