The ROX index is a simple bedside calculation that helps clinicians predict whether a patient with breathing failure will need to be placed on a mechanical ventilator. It stands for Respiratory rate–OXygenation index, and it combines three measurements that are already being monitored in any hospital setting: blood oxygen saturation, the concentration of oxygen being delivered, and how fast the patient is breathing. The formula is SpO2/FiO2 divided by respiratory rate.
How the ROX Index Is Calculated
The formula has three components. SpO2 is the oxygen saturation reading from a standard pulse oximeter, expressed as a percentage. FiO2 is the fraction of inspired oxygen, meaning how much supplemental oxygen the patient is receiving (room air is 0.21, or 21%). Respiratory rate is the number of breaths per minute. You divide SpO2 by FiO2 first, then divide that result by the respiratory rate.
For example, a patient with an oxygen saturation of 96%, receiving 50% oxygen, breathing 25 times per minute would have a ROX index of (96/50)/25 = 7.68. A higher number is better because it means the patient is maintaining good oxygen levels without breathing excessively fast or requiring very high oxygen concentrations.
Why It Was Developed
The ROX index was originally introduced to solve a specific problem in intensive care: deciding when high-flow nasal cannula therapy (HFNC) is working and when it’s failing. HFNC delivers heated, humidified oxygen at high flow rates through nasal prongs, and it has become a common first-line treatment for patients with severe breathing difficulty from pneumonia and other causes of low oxygen levels.
The challenge is that HFNC can sometimes mask worsening respiratory failure. A patient may look stable on the surface while their breathing muscles are working dangerously hard. If intubation (placing a breathing tube) is delayed too long, outcomes get significantly worse. The ROX index gives clinicians an objective number to track over time, rather than relying solely on clinical impression.
Key Threshold Values
The most widely referenced thresholds come from a validation study published in the American Journal of Respiratory and Critical Care Medicine. These values are assessed at specific time points after starting high-flow nasal cannula therapy:
- ROX ≥ 4.88 at 2, 6, or 12 hours after starting HFNC was consistently associated with a lower risk of needing intubation. Patients above this threshold had a high chance of HFNC success.
- ROX < 3.85 at 12 hours indicated a high risk of HFNC failure, and intubation should be seriously considered.
- ROX < 2.85 at 2 hours and < 3.47 at 6 hours were early warning signals that the therapy was not working.
The recommended approach is to monitor the ROX index repeatedly over time, with particular attention from the 12-hour mark onward. A single measurement is less useful than the trend. In one study, any decrease in the ROX index after starting HFNC was a strong predictor of eventual intubation, with patients whose values dropped being roughly 15 times more likely to need a breathing tube.
A separate study found that an initial ROX value below 5 at the time HFNC was started was itself suggestive of progression to mechanical ventilation.
Performance Compared to Other Measures
The ROX index generally performs better than looking at respiratory rate alone or the oxygen saturation-to-FiO2 ratio alone, though the advantage is not always statistically significant. Its strength is that it captures two different dimensions of respiratory distress in a single number. A patient can have acceptable oxygen levels while breathing dangerously fast, or a normal respiratory rate while requiring very high oxygen. Either situation alone might not trigger alarm, but the ROX index flags both.
In COVID-19 patients specifically, the ROX index proved useful but with some caveats. COVID pneumonia affects breathing patterns differently than typical bacterial pneumonia. Some patients maintained deceptively normal respiratory rates despite significant lung involvement, which could lead to a falsely reassuring ROX value. Researchers debated whether higher thresholds might be needed for COVID patients, though no consensus cutoff was established.
The Modified ROX-HR Index
One recognized limitation of the standard ROX index is that it ignores heart rate, which is another important sign of physiological stress. A modified version called ROX-HR incorporates heart rate into the formula: (SpO2/FiO2) divided by (respiratory rate × heart rate), then multiplied by 100.
In studies comparing the two, ROX-HR showed slightly better predictive accuracy than the standard ROX index, with the area under the curve improving from 0.822 to 0.834 in one study of patients after extubation. That improvement was not statistically significant, but the logic is sound: a rising heart rate often accompanies worsening respiratory failure, and capturing it adds another dimension to the assessment. A ROX-HR value above 12.13 at 2 hours was associated with 72% sensitivity and 86% specificity for predicting HFNC success in that study.
Limitations to Keep in Mind
The ROX index is a snapshot. It captures one moment in time and can be influenced by factors that have nothing to do with how well the lungs are working. Fever, pain, anxiety, moving around in bed, and even acidosis from other causes can all increase respiratory rate and temporarily push the number down. Conversely, sedating medications or fatigue can slow breathing and artificially inflate the value.
The index also does not account for the flow rate of oxygen being delivered. Two patients can have identical ROX values while one receives 30 liters per minute and the other receives 60 liters per minute. The patient on higher flow is clearly in a more precarious situation, but the ROX index treats them the same. Higher flow rates create positive pressure in the airway, help flush out carbon dioxide, and reduce the work of breathing, all of which can make a patient’s numbers look better than their underlying condition warrants.
For these reasons, the ROX index works best as one input among many rather than a standalone decision tool. Clinicians use it alongside their physical examination, imaging, blood gas results, and the overall trajectory of the patient’s illness. Its greatest value is in providing an easily calculated, reproducible number that can be tracked over hours to reveal trends that might otherwise be missed.