Table 3.

A regression model with NIV failure as the dependent variable

	ACPE		COPD		OHS
	P value*	Final model ExpB (CI)	P value*	Final model ExpB (CI)	P value*	Final model ExpB (CI)
Age, yearsa	0.03	1.096 (1.035–1.161)	0.029	NS	0.104	NS
Maleb	0.534		0.558		1
BMIa	0.699		0.715		0.269
Smokerb	1		0.552		1
Drinkerb	0.580		0.450		1
Respiratory ratea	0.068	1.094 (1.002–1.194)	0.555		0.301
SBP, mmHga	0.325		0.972		0.001	1.040 (1.014–1.067)
DBP, mmHga	0.724		0.385		0.337
Charlson indexa	0.880		0.824		0.585
Glasgow scalea	0.267		0.334		0.159	NS
APACHE II scorea	0.143		0.749		0.180
PaO2, mmHga	0.859		0.109	1.019 (1.000–1.019)	0.819
PaO2/FiO2 a	0.861		0.213		0.540
PaCO2, mmHga	0.781		0.119	NS	0.101	NS
pH ≤ 7.25b	0.359		0.865		0.083	NS
FVC %a			0.774
FEV1 %a			0.472
RICU type 1b	0.696		0.864		0.778
PS modeb	1		1		0.316
CPAP/NIVb	1		1		1
DNI orderb	0.562		0.033	NS	0.388
pH ≥ 7.35 at 1 hb	0.027	NS	0.042	0.341 (0.117–0.999)	0.058	NS

Abbreviations: ExpB exponent B, CI confidence interval, NS not significant, BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, RICU respiratory intermediate care unit, PS pressure support, CPAP/NIV previous continuous positive airway pressure or noninvasive ventilation, DNI do not intubate

^acontinuous variables in the model, ^bcategorical variables in the model, and ^* P values from univariate analysis. Variables with a P ≤ 0.2 were included in regression analysis