ROX index in acute hypoxic respiratory failure predicts high-flow nasal cannula outcomes

Key takeaways:

• Capturing ROX index during the first 12 hours of high-flow nasal cannula therapy yielded favorable sensitivity and specificity values.
• Varying acute hypoxic respiratory failure causes did not alter specificity.

BOSTON — When capturing ROX index in acute hypoxic respiratory failure, time from the start of high-flow nasal cannula therapy impacted the index’s sensitivity and specificity, according to data presented at the CHEST Annual Meeting.

“I believe there has to be still a lot of research with the ROX index, but once we achieve a good point, or rather a point where we can start using ROX index repeatedly, then I believe [high-flow nasal cannula] therapy, as well as the ROX index together, can make managing patients with respiratory failure a lot easier,” Inban Pugazhendi, MD, internal medicine resident at NYMC, Saint Claire’s Hospital and St. Mary’s General Hospital, said during his presentation.

Searching the databases of PubMed, Cochrane Central, Embase and Scopus, Pugazhendi and colleagues reviewed and evaluated 15 studies including 3,888 adults with acute hypoxic respiratory failure receiving high-flow nasal cannula (HFNC) therapy to establish if the ROX index — SpO2/FiO2 to respiratory rate ratio — can predict HFNC outcomes.

The risk of bias was minimal in most of the studies (11 out of 15), with fewer studies having high risk (1 out of 15) or moderate risk (3 out of 15), Pugazhendi said during the presentation.

Additionally, Pugazhendi reported that 4.96 was the average ROX cutoff value within the included studies.

Researchers reported “considerable heterogeneity in sensitivity [I² = 84.37%] and specificity [I² = 91.55%]” in the studies and noted the threshold effect as the probable culprit behind this disparity via the hierarchical summary receiver-operating characteristic (HSROC) curve.

Both the sensitivity and specificity values were lower than the area under HSROC curve value (0.66 and 0.73 vs. 0.83), but Pugazhendi noted “a good result.”

“For [an] index which can be calculated in seconds at the bedside, such a sensitivity and specificity index can be a game changer going forward,” Pugazhendi said.

Additionally, the diagnostic odds ratio was 5.4, according to the abstract.

Time elapsed since the start of HFNC therapy at the moment of capturing ROX appeared to impact the index’s sensitivity and specificity. Researchers found that measuring ROX during the first 6 hours of HFNC yielded comparable sensitivity (0.66) and specificity (0.74) values, and this was also the case if ROX was measured between the first 6 and 12 hours of HFNC (sensitivity, 0.64; specificity, 0.71).

In terms of disease etiology, varying causes of acute hypoxic respiratory failure (namely COVID-19) did not alter the “good specificity” of ROX to predict HFNC outcomes, according to researchers.

“If at all there is a limitation, what I would say is we were not able to go too much into the subgroup analysis,” Pugazhendi said during the presentation.