Predictive factors of prolonged mechanical ventilation, overall survival, and quality of life in patients with post-thymectomy myasthenic crisis

Table 2 Univariate and multivariate analyses of risk factors that may prolong mechanical ventilation

Variables	Cox model	p value	Hazard ratio (95.0% CI for HR)
Alcohol status^b	Univariate analysis	0.040	0.524 (0.283–0.970)
	Multivariate analysis^a	0.049	0.517 (0.267–0.998)
MGFA classification	Univariate analysis	0.001	0.789 (0.685–0.908)
	Multivariate analysis^a	0.007	0.811 (0.697–0.944)
Clavien-Dindo classification	Univariate analysis	0.003	0.748 (0.617–0.907)
	Multivariate analysis^a	0.013	0.783 (0.645–0.950)
Osserman’s classification	Univariate analysis	0.006	0.680 (0.515–0.897)
QMG score	Univariate analysis	0.010	0.948 (0.910–0.987)
Preoperative crisis^b	Univariate analysis	0.036	0.550 (0.314–0.963)
Preoperative max pyridostigmine	Univariate analysis	0.025	0.998 (0.996–1.000)
Lung function	Univariate analysis	0.005	0.738 (0.597–0.912)

We defined “status = 1” as tracheal extubation and “time” as the duration of mechanical ventilation in the analysis of Cox proportional hazard regression analyse
^aMultivariate Cox analysis by forward step
^bAlcohol status is defined as (1) never, (2) occasional, (3) excessive, and (4) dependence. Preoperative crisis is defined as the times of myasthenic crisis before surgery

ISSN: 1477-7819