Table 3.
Multivariable analyses, MetS vs. non-MetS
| Outcomes | MetS (n = 188) | Non-MetS (n = 99) | Risk difference, % (95% CI) | Adjusted ORa (95% CI) | P value |
|---|---|---|---|---|---|
| Hospital mortality, n (%) | 48 (25.5) | 10 (10.1) | 15.4 (6.8–24.0) | 3.42 (1.52–7.69) | 0.0030 |
| ICU requirement, n (%) | 106 (56.4) | 24 (24.2) | 32.1 (21.1, 43.2) | 4.59 (2.53–8.32) | <0.0001 |
| ARDSb, n (%) | 69 (36.7) | 11 (11.1) | 25.6 (16.3–34.9) | 4.70 (2.25–9.82) | <0.0001 |
| IMV, n (%) | 90 (47.9) | 18 (18.2) | 29.7 (19.3–40.1) | 4.71 (2.50–8.87) | <0.0001 |
| LOS, mean ± SD, days | 14.1 ± 10.6 | 10.7 ± 10.1 | n/a | n/a | 0.0097 |
| LOS, median (IQR), days | 11 (7.0–20.5) | 7 (4.0–14.0) | n/a | n/a | 0.0062 |
| Hospital readmission,c n (%) | 16/140 (11.4) | 6/89 (6.7) | 4.7 (−1.7 to 12.1) | 1.19 (0.40–3.61) | 0.7533 |
n/a, not available.
a
Multivariable logistic regression model adjusted for age, sex, race, hospital site, and Charlson Comorbidity Index.
b
ARDS defined according to Berlin criteria.
c
Hospital readmission data were available only for case subjects who survived to discharge.