Retrospective evaluation of the respiratory rate–oxygenation index to predict the outcome of high-flow nasal cannula oxygen therapy in dogs (2018–2021): 81 cases

Objective

To evaluate the respiratory rate–oxygenation index (ROX), modified ROX index (ROX-HR), and the ratio of pulse oximetry saturation (Spo₂) to Fio₂ (SF) to determine if these indices over time are predictive of outcome in dogs treated with high-flow nasal cannula oxygen therapy (HFNC).

Design

Retrospective study.

Setting

Two university teaching hospitals.

Animals

Eighty-one client-owned dogs treated with HFNC for hypoxemic respiratory failure.

Interventions

None.

Measurements and Main Results

The ROX was defined as the SF divided by the respiratory rate (RR), and the ROX-HR was defined as the ROX divided by the heart rate multiplied by 100. The overall success rate of HFNC was 44% (n = 36/81). Dogs weaned from HFNC had a significantly higher ROX (P < 0.0001) at 1–3, 5–10, 12, and 15 hours than dogs that failed HFNC. Both the ROX and SF showed excellent discriminatory power in predicting HFNC failure at 6 hours, with an area under receiver operating curve of 0.85 (95% confidence interval: 0.72–0.99; P < 0.002) and 0.86 (95% confidence interval: 0.73–0.99; P < 0.001), respectively. The optimal cutoff values for predicting HFNC failure at 6 hours were a ROX ≤3.68 (sensitivity 72%, specificity 92%) and an SF ≤143 (sensitivity 79%, specificity 93%).

Conclusions

These results suggest that similar to people, the ROX and SF are useful predictors of HFNC failure. These indices are easy to measure at the bedside and may have clinical use. Future prospective studies are warranted to confirm the findings and to optimize cutoff values in a larger population of dogs undergoing HFNC.