Performance of the ROX index to predict intubation in immunocompromised patients receiving high-flow nasal cannula for acute respiratory failure

Annals of Intensive Care

Table 3 Adjusted ROX index 6 h after HFNC onset according other risk factors of intubation in all cohorts (multivariate analysis)

Variables	OR (95% CI)	p
Oxygen flow at randomization > 9 L/min	1.10 (0.64–1.88)	0.27
IVNICTUS study	1.03 (0.58–1.83)	0.29
ARF related to pneumonia	1.06 (0.63–1.76)	0.26
SOFA Day 1 (per point increase)	1.17 (1.08–1.27)	0.04
ROX H6 (per point increase)	0.89 (0.82–0.96)	0.04

Oxygen flow at ICU admission was dichotomized below or over 9 l/min
Patients form IVNICTUS study were compared to patients from HIGH study
ARF aetiology was dichotomized between ARF related to bacterial or viral pneumonia versus all other reason of ARF
SOFA Day 1 was sequential organ failure assessment
ROX index was used as continuous variable
Characteristics of the multivariate model
Hosmer and Lemeshow goodness of fit test p = 0.65
The AUC of the ROC curve of the predictive model = 0.66
The relationship between ROX index and lod(odds) of intubation was linear
ARF acute respiratory failure, NIV non-invasive ventilation, OR odds ratio, CI confidence interval