Table 2. Adjustment for the effect of antipyretic therapy for patients with sepsis by treating temperature as continuous variable (maximum body temperature during ICU stay).
| Odds Ratio | Odds ratio | Lower limit of 95% CI | Upper limit of 95% CI | P |
|---|---|---|---|---|
| Age | 1.02 | 1.01 | 1.02 | <0.001 |
| SOFA | 1.22 | 1.20 | 1.24 | <0.001 |
| Sex | 1.12 | 1.003 | 1.251 | 0.044 |
| Care unit (MICU as the reference) | ||||
| SICU | 0.70 | 0.54 | 0.90 | 0.006 |
| CCU | 0.96 | 0.83 | 1.10 | 0.524 |
| CSRU | 0.46 | 0.40 | 0.53 | <0.001 |
| Lactate (normal as the reference) | ||||
| Hyperlactamia (>2 mmol/l) | 1.30 | 1.15 | 1.48 | <0.001 |
| Missing values | 0.47 | 0.40 | 0.56 | <0.001 |
| Maximum temperature during ICU stay (with each 1°C increase) ¶ | 1.15 | 1.07 | 1.23 | <0.001 |
| Antipyretic therapy | .0000164 | 1.98e-08 | .013631 | <0.001 |
| Antipyretic therapy×Tmax ‡ | 1.33 | 1.12 | 1.58 | <0.001 |
¶ The main effect of Tmax (OR = 1.15) is exponentiation of the slope in the regression model when no antipyretic therapy was given (antipyretic therapy = 1).
‡ The coefficient of the interaction term was the difference of the slope in groups with and without antipyretic therapy. An OR>1 indicates a positive coefficient and steeper slope of the antipyretic therapy group than non-antipyretic group.