TABLE 2.
Characteristics of Patients With Sepsis Acquired in the Community, in Hospital, and in the ICU
| Source of Sepsis | Community | Hospital | ICUa | p for Comparison |
|---|---|---|---|---|
| Number of admissions | 7,940 | 4,079 | 2,040 | |
| Women, n (%) | 3,253 (41.0) | 1,718 (42.1) | 719 (35.2) | < 0.0001 |
| Age, median (IQR) | 60.1 (45.4–73.1) | 65.1 (51.6–75.5) | 61.3 (47.5–72.0) | < 0.0001 |
| Severity of sepsis | ||||
| Septic shock, n (%) | 2,493 (31.4) | 1,036 (25.4) | 415 (20.3) | < 0.0001 |
| SOFA score, median (IQR) | 8 (5–11) | 7 (5–11) | 9 (6–11) | < 0.0001 |
| Change in component SOFA score on the day that criteria for sepsis were met, mean (sd) | ||||
| Cardiovascular | 2.28 (1.58) | 2.07 (1.58) | 1.04 (1.42) | < 0.0001 |
| Respiratory | 2.48 (1.11) | 2.55 (1.10) | 0.87 (1.10) | < 0.0001 |
| Renal | 0.84 (1.22) | 0.88 (1.21) | 0.21 (0.62) | < 0.0001 |
| Coagulation | 0.74 (1.01) | 0.81 (1.24) | 0.14 (0.59) | < 0.0001 |
| CNS | 2.05 (1.64) | 1.80 (1.56) | 1.18 (1.44) | < 0.0001 |
| Liver | 0.53 (0.90) | 0.55 (0.96) | 0.17 (0.54) | < 0.0001 |
| Relative contribution of organ system to overall delta SOFA on the day that sepsis criteria were met (%) | ||||
| Cardiovascular | 24.7 | 22.8 | 26.3 | < 0.0001 |
| Respiratory | 32.0 | 34.5 | 26.6 | < 0.0001 |
| Renal | 8.8 | 9.6 | 6.2 | < 0.0001 |
| Coagulation | 7.7 | 8.2 | 3.6 | < 0.0001 |
| CNS | 21.6 | 19.5 | 32.8 | < 0.0001 |
| Liver | 5.3 | 5.4 | 4.4 | < 0.0001 |
| Outcomes | ||||
| ICU length of stay, d, median (IQR) | 3.8 (1.8–9.0) | 4.2 (2.0–9.0) | 18.2 (11.1–30.1) | < 0.0001 |
| ICU mortality, n (%) | 1,022 (12.9) | 760 (18.6) | 484 (23.7) | < 0.0001 |
| ICU mortality for septic shock, n (%) | 622 (24.9) | 349 (33.7) | 170 (41.0) | < 0.0001 |
IQR = interquartile range, SOFA = Sequential Organ Failure Assessment.
aData are shown for the first episode per admission for ICU-acquired sepsis. Hospital- or community-acquired sepsis could only occur once per ICU admission, by definition. P values compare the relevant estimate from community-, hospital-, and ICU-acquired sepsis (proportion tests for categorical variables, Kruskal-Wallis tests for variables with median and IQR quoted, and analysis of variance tests for other variables).