Table 3.
The hazard for 28-day mortality as a function of time to antibiotics in ED patients with suspected infection in three categories of illness severity
| Crude model (HR (95% CI) | Corrected model (HR (95% CI) | P (Corrected model) | |
|---|---|---|---|
| PIRO group 1 to 7 (n = 413) | |||
| Antibiotics <1 hour (reference category) | 1 | 1 | 0.422a |
| Antibiotics 1–3 hours | 0.95 (0.17-5.18) | 2.55 (0.36-18.25) | 0.352 |
| Antibiotics >3 hours | 1.98 (0.36-10.78) | 5.31 (0.43-68.16) | 0.191 |
| Type of hospital (academic versus urban) | 0.06 (0.007-0.48) | 0.008 | |
| P score | 2.53 (1.41-4.56) | 0.002 | |
| Appropriateness of antibiotics | 0.33 (0.06-1.66) | 0.180 | |
| Amount of oxygen (L/min) | 1.17 (1.02-1.35) | 0.028 | |
| Amount of fluids (L/ ED stay) | 1.65 (0.76-3.59) | 0.205 | |
| PIRO group 8 to 14 (n = 532) | |||
| Antibiotics <1 hour (reference category) | 1 | 1 | 0.676a |
| Antibiotics 1–3 hours | 1.11 (0.62-1.99) | 1.25 (0.62-2.31) | 0.488 |
| Antibiotics >3 hours | 0.65 (0.22-1.90) | 0.86 (0.28-2.63) | 0.786 |
| Type of hospital (academic versus urban) | 0.35 (0.19-0.68) | 0.002 | |
| P score | 1.29 (1.14-1.46) | <0.001 | |
| R score | 0.74 (0.62-0.89) | 0.001 | |
| O score | 1.28 (1.10-1.50) | 0.002 | |
| Amount of oxygen (L/min) | 1.04 (1.00-1.09) | 0.067 | |
| PIRO group >14 (n = 223) | |||
| Antibiotics <1 hour (reference category) | 1 | 1 | 0.978a |
| Antibiotics 1–3 hours | 1.10 (0.62-1.97) | 0.99 (0.53-1.87) | 0.983 |
| Antibiotics >3 hours | 0.93 (0.36-2.43) | 1.11 (0.40-3.08) | 0.849 |
| Type of hospital (academic versus urban) | 1.57 (0.83-3.00) | 0.166 | |
| O score | 1.15 (1.00-1.34) | 0.056 | |
| β-blocker use | 0.96 (0.51-1.79) | 0.892 | |
| Appropriateness of antibiotics | 0.35 (0.19-0.62) | <0.001 | |
| Amount of oxygen (L/min) | 1.01 (0.98-1.04) | 0.414 | |
| Amount of fluids (L/ED stay) | 0.88 (0.65-1.20) | 0.425 | |
Cox regression analysis was performed with time to antibiotics divided into three categories. In the corrected model regression coefficients were corrected for possible predefined confounders: PIRO score; β-blocker, statin, and antibiotic use prior to ED presentation; appropriateness of initial antibiotics in the ED; amount of fluids (L); amount of oxygen (L/min) in the ED; and hospital (academic versus urban hospital). aOverall P value for categories of time to antibiotics. The P values were not used to construct the model.
CI, Confidence interval; ED, Emergency department; HR, Hazard ratio; ICU, Intensive care unit; PIRO, Predisposition, Infection, Response, and Organ failure score (as a measure of illness severity).