Correlation of the ROX index with outcomes.

Background
There is no existing reliable and practical method for predicting the prognosis of acute respiratory distress syndrome (ARDS).
Objective
We aimed to clarify the association between the ROX index, which is calculated as the ratio of peripheral oxygen saturation divided by the fraction of inspired oxygen to the respiratory rate, and the prognosis of patients with ARDS under ventilator support.
Methods
In this single-center retrospective cohort study from prospectively collected database, eligible patients were categorized into three groups based on ROX tertiles. The primary outcome was the 28-day survival, and the secondary outcome was 28-day liberation from ventilator support. We performed multivariable analysis using the Cox proportional hazards model.
Results
Among 93 eligible patients, 24 (26%) patients died. The patients were divided into three groups according to the ROX index (< 7.4, 7.4–11, ≥ 11), with 13, 7, and 4 patients dying in the groups, respectively. A higher ROX index was associated with lower mortality; adjusted hazard ratios [95% CIs] for increasing tertiles of ROX index: 1[reference], 0.54[0.21–1.41], 0.23[0.074–0.72] (P = 0.011 for trend) and a higher rate of successful 28-day liberation from ventilator support; adjusted hazard ratios [95% CIs] for increasing tertiles of ROX index: 1[reference], 1.41[0.68–2.94], 2.80[1.42–5.52] (P = 0.001 for trend).
Conclusions
The ROX index at 24 h after initiating ventilator support is a predictor of outcomes in patients with ARDS and might inform initiation of more advanced treatments.