Table 2.
Logistic regression analysis of risk factors for corticosteroid dependence before start of corticosteroid
|
P value1 |
OR (95%CI) | ||
| Univariate | Multivariate | ||
| Tenesmus | 0.012 | 0.013 | 0.336 (0.147-0.768) |
| Anemia | 0.041 | 0.338 | 0.576 (0.233-1.422) |
| Hypokalemic (< 3.0 mmol/L) | 0.065 | 0.060 | 0.121 (0.014-1.028) |
| Diarrhea (> 6/d) | 0.135 | ||
| ESR (> 15 mm/h) | 0.193 | ||
| Weight loss | 0.260 | ||
| Tachycardia (≥ 90/min) | 0.280 | ||
| Hypoproteinemia (< 3.5 mg/dL) | 0.373 | ||
1
Data with univariate P < 0.5 are not displayed. All the variables were statistically analyzed. In univariate analysis, tenesmus and anemia before start of corticosteroid treatment were related with corticosteroid dependence (P = 0.012, 0.041). All the variables with P < 0.100 were further analyzed by logistic regression. After multivariate analysis, tenesmus at baseline was found to be associated with corticosteroid dependence (OR = 0.336, 95%CI: 0.147-0.768; P = 0.013). ESR: Erythrocyte sediment rate.