ROX index performance to predict high-flow nasal oxygen outcome in Covid-19 related hypoxemic acute respiratory failure

Annals of Intensive Care

Table 2 ROX index accuracy to predict HFNO failure at different times after HFNO initiation

Time after HFNO initiation	Maximum of Youden’s index	Rox index cut-off	Sensitivity*	95% CI limits for sensitivity		Specificity	95% CI limits for specificity		AUROC
Time after HFNO initiation	Maximum of Youden’s index	Rox index cut-off	Sensitivity*	Lower	Upper	Specificity	Lower	Upper	AUROC
H0	0.393	6.20	63.3%	48.3%	76.6%	76.0%	61.8%	86.9%	0.70
H2-H4	0.413	6.88	78.0%	62.4%	89.4%	63.3%	48.3%	76.6%	0.71
H6	0.361	7.00	75.7%	58.8%	88.2%	60.4%	45.3%	74.2%	0.68
H12	0.320	8.73	87.1%	70.2%	96.4%	44.9%	30.7%	59.8%	0.66
H18	0.398	6.96	69.0%	49.2%	84.7%	70.8%	55.9%	83.0%	0.70
H24	0.490	6.96	77.3%	54.6%	92.2%	71.7%	56.5%	84.0%	0.75
H48	0.343	5.00	43.8%	19.8%	70.1%	90.5%	77.4%	97.3%	0.67
H72	0.543	5.78	80.0%	44.4%	97.5%	74.3%	56.7%	87.5%	0.77

HFNO high-flow nasal oxygen therapy, AUROC area under the receiver operating characteristic curve, CI confidence interval; * = proportion of patients with ROX index ≤ cut-off among those who failed HFNO within 28 days after ICU admission