| Prednisone | Control | Relative risk (95% CI) | p-value | |
| Primary efficacy end-point | ||||
| ICU mortality | 17/111 (15.3) | 15/106 (14.2) | 1.08 (0.6–2.05) | 0.81 |
| ICU mortality in patients ventilated with NIV | 8/76 (10.5) | 8/71 (11.3) | 0.93 (0.37–2.35) | 0.88 |
| ICU mortality in patients ventilated conventionally | 9/35 (25.7) | 7/35 (20.0) | 1.28 (0.54–3) | 0.57 |
| Secondary end-points | ||||
| NIV failure | 12/76 (15.7) | 9/71 (12.7) | 1.25 (0.56–2.8) | 0.59 |
| Mechanical ventilation duration days | 6 (4–12) | 6 (3.8–12) | 0.87 | |
| ICU length of stay days | 9 (6–14) | 8 (6–14) | 0.88 | |
| Safety end-point | ||||
| Hyperglycaemic episodes requiring initiation or alteration of insulin therapy | 55/111 (49.5) | 35/106 (33.0) | 1.5 (1.08–2.08) | 0.015 |
Data are presented as n/N (%) or median (interquartile range), unless otherwise stated. ICU: intensive care unit, NIV: noninvasive ventilation.