Table 3.

Multivariate analysis of variables associated with outcomes in the overall population

	Adjusted odds ratio (95 % CI)	p value
Variables independently associated with intubation a
Simplified Acute Physiology Score II, per point	1.04 (1.00–1.08)	0.04
Noninvasive ventilation as a first-line therapy	3.25 (1.39–7.60)	0.007
Use of vasopressors within 24 h after ICU admission	4.12 (1.32–12.84)	0.02
Variables independently associated with mortality at day 28 b
Age (per year)	1.03 (1.00–1.07)	0.04
Use of vasopressors within 24 h after ICU admission	2.83 (1.02–7.91)	0.047
Noninvasive ventilation as a first-line therapy	3.70 (1.49–9.19)	0.005

^aNon-collinear variables included in the logistical regression model were Simplified Acute Physiology Score II, Noninvasive ventilation as a first-line therapy, use of vasopressors within 24 h after ICU admission, SpO₂ at ICU admission, cause of respiratory failure and PaCO₂ as a continuous variable. The year of ICU admission was forced in the model

^bNon-collinear variables included in the logistical regression model were age, PaO₂-to-FiO₂ ratio at ICU admission, use of noninvasive ventilation as a first-line therapy, type of immunosuppression, use of vasopressors in the 24 h after ICU admission, cause of respiratory failure and PaCO₂ > 45 mmHg. The year of ICU admission was forced in the model