Skip to main content

Table 4 Univariate analyses and multivariate logistic regression model for factors potentially associated with bad outcome

From: Early prediction of hospital outcomes in patients tracheostomized for complex mechanical ventilation weaning

  Univariate regression Multivariate model
  OR (CI 95%) p value OR (CI 95%) VIF p value
BMI 1.181 (1.07–1.32) 0.0009 1.205 (1.09–1.36) 1.003 0.0008
Age 1.038 (1.01–1.08) 0.0253 1.044 (1.00–1.09) 1.037 0.0463
Sex 2.061 (0.73–6.42) 0.2967    
Number of comorbidities 1.248 (0.73–2.15) 0.4471    
Clinical Frailty Score 0.986 (0.76–1.27) 0.8269    
NRS score at ICU admission 1.129 (0.88–1.48) 0.385    
SAPS II at ICU admission 1.011 (0.99–1.04) 0.3227    
SOFA score at ICU admission 1.010 (0.87–1.17) 0.8776    
Type of ICU admission (medical/surgical) 0.560 (0.22–1.44) 0.3203    
Neurological cause for intubation 0.849 (0.30–2.30) 0.8038    
VT/PBW 1.160 (0.81–1.68) 0.5643    
PEEP 1.026 (0.79–1.34) 0.996    
Dynamic plateau pressure 0.898 (0.78–1.02) 0.1097    
Percentage of days with sedation use 0.213 (0.03–1.26) 0.1011 0.208 (0.02–1.57) 1.035 0.14
Percentage of days with opioids use 0.293 (0.02–5.01) 0.7923    
Percentage of days with NMBA use 0.592 (0.07–4.19) 0.401    
Control ventilation before tracheostomy 0.383 (0.07–1.97) 0.1597    
1st separation attempt 1.062 (0.95–1.19) 0.2298    
Any separation attempt 0.564 (0.20–1.57) 0.3983    
Sedation use (day before tracheostomy) 1.556 (0.54–4.91) 0.551    
Opioids use (day before tracheostomy) 3.949 (0.96–26.82) 0.2531    
Tracheostomy technique (percutaneous vs surgical) 1.427 (0.42–5.70) 0.728    
Time from intubation to tracheostomy 1.006 (0.95–1.07) 0.921    
  1. BMI body mass index, NRS nutrition risk screening, ICU intensive care unit, SAPS II Simplified Acute Physiology Score II, SOFA score Sequential Organ Failure Assessment score, VT/PBW tidal volume divided by predicted body weight, PEEP positive end-expiratory pressure, NMBA neuromuscular blocking agents
  2. Left p values calculated using univariate logistic regression for each variable. Right p values calculated with multiple logistic regression model, which included BMI, age and sedation use