Table 2.

Multivariate analysis for explaining early intubation

	Odds ratio [95% confidence interval]	P-value	% of r2 contribution	Covariate rank in r2 contribution
Use of accessory respiratory musclesa	5.63 [2.83–11.82]	< 0.001	12.8	3
pH (for each 0.1 decrease)a	1.51 [1.09–2.97]	0.02	3.8	7
Glasgow score (reference ≥ 14)a			39.5	1
10–13 vs ≥ 14	3.03 [1.20–7.63]	0.01
<10 vs ≥ 14	39.95 [10.13–134.85]	< 0.001
Inability to clear tracheal secretionsa	2.64 [1.38–3.96]	0.02	6.02	6
Vasopressor (for each 1 μg/kg/min increase)a	2.34 [1.38–3.96]	0.001	5.7	5
Lactate (for each 1 mmol/L increase)a	1.11 [1.01–1.21]	0.02	8.6	4

	Variance estimate [95% confidence interval]
Variable effect: center	0.50 [0.12–1.81]	< 0.01	16.3	2

Overall r² of the model was 0.6

Parameters entered in the model not presented in the table, with a P-value > 0.1: age, sex, weight, NYHA status, chronic respiratory failure, chronic renal failure, immunosuppression, pulmonary and urinary site of infection, pathogen, respiratory rate, PaCO₂, PaO₂/FiO₂ ratio, fluid loading from first hypotension to H0

^aWorst value recorded between H0 and H8, or between H0 and immediately before intubation if performed before H8