From a recent evidence summary:
Recent findings: The efficacy of NIV is variable depending on the cause of the episode of ARF. In community-acquired pneumonia, NIV is often associated with poor response, with better response in patients with preexisting cardiac or respiratory disease. In patients with pandemic influenza H1N1 and severe ARF, NIV has been associated with high failure rates but relatively favorable mortality. In acute respiratory distress syndrome, NIV should be used very cautiously and restricted to patients with mild–moderate acute respiratory distress syndrome without shock or metabolic acidosis due to the high failure rate observed in several reports. Despite limited evidence, NIV may improve the outcomes of patients with chest trauma and severe ARF. In postoperative ARF, both continuous positive airway pressure and NIV are effective to improve clinical outcomes, particularly in those with abdominal, cardiac, and thoracic surgery.
Summary: Although patients with severe hypoxemic ARF are, in general, less likely to be intubated when NIV is used, the efficacy is different among these heterogeneous populations. Therefore, NIV is not routinely recommended in all patients with severe hypoxemic ARF.