What is the ROX index (ratio of SpO₂/FiO₂ to respiratory rate) in oxygen therapy?

What is the ROX Index in Oxygen Therapy?

The ROX index is a simple bedside calculation—SpO₂/FiO₂ divided by respiratory rate—that predicts whether high-flow nasal cannula (HFNC) therapy will succeed or fail in patients with acute hypoxemic respiratory failure, helping clinicians decide when to continue HFNC versus escalating to intubation. 1

Definition and Calculation

The ROX index is calculated as:

ROX = (SpO₂/FiO₂) / Respiratory Rate

Where:

• SpO₂ is oxygen saturation measured by pulse oximetry (as a percentage, not decimal)
• FiO₂ is the fraction of inspired oxygen (expressed as a decimal, e.g., 0.60 for 60%)
• Respiratory rate is breaths per minute 1, 2

For example, a patient with SpO₂ 95%, FiO₂ 0.50, and respiratory rate 25 breaths/min would have a ROX index of (95/0.50)/25 = 7.6 1

Clinical Application and Timing

The ROX index should be calculated at 2,6, and 12 hours after initiating HFNC therapy, with prediction accuracy improving over time—the 12-hour measurement provides the most reliable prediction of treatment outcome. 2

Interpretation Thresholds for Pneumonia/Hypoxemic Respiratory Failure:

• ROX ≥ 4.88 at 12 hours: Strongly predicts HFNC success; continue therapy 1, 2
• ROX < 3.85 at 12 hours: High risk of HFNC failure; strongly consider intubation 2
• ROX < 2.85 at 2 hours: Very high risk of failure; prepare for intubation 2
• ROX < 3.47 at 6 hours: High risk of failure; escalate monitoring 2

The hazard ratio for intubation with ROX ≥ 4.88 at 12 hours is 0.291 (95% CI 0.161-0.524), meaning patients meeting this threshold have approximately 70% lower risk of requiring mechanical ventilation 2

Disease-Specific Considerations

COPD Exacerbations:

For COPD patients on HFNC or NIV, higher ROX thresholds are required—a ROX > 6.88 provides optimal sensitivity (62%) and specificity (57%) for predicting treatment success, compared to the 4.88 threshold used in pneumonia. 3

This critical difference exists because COPD patients have different respiratory mechanics and often baseline hypercapnia, making the standard pneumonia-derived cutoffs less accurate 3

Modified ROX Indices:

Two variations have been studied:

• ROX-HR index: (ROX index / heart rate) × 100, which may provide earlier prediction of failure, particularly in post-extubation patients 4

  • ROX-HR > 6.80 at 10 hours predicts success in acute respiratory failure 4
  • ROX-HR > 8.00 at 10 hours predicts success post-extubation 4

• mROX index: Uses PaO₂ instead of SpO₂ in the calculation (PaO₂/FiO₂ / respiratory rate), which may have higher predictive accuracy at 2 hours (AUROC 0.73 vs 0.68 for standard ROX) 5

  • Baseline mROX > 7.1 showed 100% specificity for HFNC success 5

Clinical Decision Algorithm

When initiating HFNC for acute hypoxemic respiratory failure:

1. Calculate baseline ROX index before starting HFNC 2
2. Initiate HFNC per standard protocols (typically 40-60 L/min flow, FiO₂ titrated to SpO₂ ≥ 92%) 6
3. Recalculate ROX at 2,6, and 12 hours 2
4. If ROX < 2.85 at 2 hours: Prepare for intubation; failure is highly likely 2
5. If ROX < 3.47 at 6 hours: Intensify monitoring; consider ICU transfer if not already there 2
6. If ROX < 3.85 at 12 hours: Strongly consider intubation; continued HFNC carries high failure risk 2
7. If ROX ≥ 4.88 at any time point (2,6, or 12 hours): Continue HFNC; success is likely 2
8. Monitor trend: Patients who fail typically show minimal improvement or declining ROX values over the first 12 hours 2

Common Pitfalls and Caveats

Do not delay intubation in patients with ROX < 3.85 at 12 hours simply because they appear subjectively comfortable—the index identifies patients at high risk for sudden decompensation. 2

The ROX index was validated primarily in pneumonia patients; apply different thresholds for COPD (ROX > 6.88 for success) to avoid premature intubation. 3

Ensure accurate FiO₂ measurement—many clinicians mistakenly use the HFNC device setting rather than measured delivered FiO₂, which can vary based on patient inspiratory flow and mouth breathing. 1

The ROX index does not apply to patients requiring HFNC for reasons other than acute hypoxemic respiratory failure (e.g., post-operative prophylaxis, comfort care). 1, 2

Among the index components, SpO₂/FiO₂ carries greater weight than respiratory rate in predicting outcome, but both are necessary for optimal prediction. 2

Evidence Quality

The ROX index was derived and validated in multicenter prospective cohort studies with area under the receiver operating characteristic curves ranging from 0.68-0.76, representing moderate to good discriminatory ability 1, 2. The index has been externally validated across multiple populations, though primarily in pneumonia patients 1, 2. The COPD-specific threshold requires further validation but represents the best available evidence for this population 3.