High-Velocity Nasal Insufflation for Respiratory Failure

Dr. Jerome Martin


Dyspnea and acute respiratory failure are among the top 5 reasons for patients to present to the Emergency Department. High-velocity nasal insufflation (HVNI), a type of high-flow nasal cannula (HFNC), decreases dead space to augment ventilation in addition to providing increased oxygenation at flow rates of 25-35 L/min. Typically, HFNC is used in hypoxemic respiratory failure and noninvasive positive pressure ventilation (NIPPV) is used for hypercapnic or mixed. HVNI could be helpful in supporting both hypoxic and hypercapnic respiratory failure presenting to the ED as it does provide some ventilatory support. When compared to NIPPV it not only improves patient comfort, but has increased ease of use.


Is there data to support the use of HVNI in undifferentiated respiratory failure?


Doshi, P; Whittle, J; Bublewicz, M; et al. High-velocity Nasal Insufflation in the treatment of respiratory failure: A randomized clinical trial. Annals of Emergency Medicine. July 2018; 7 (1): 73-83. [paper]

Study Design

Prospective, multicenter (5 southeastern United States centers), parallel group, randomized controlled trial of two noninvasive ventilator support modalities: HVNI and NIPPV. All respiratory interventions were tracked for 72 hours after intervention. The primary outcome was treatment failure defined as endotracheal intubation and mechanical ventilation. For secondary outcomes they also looked at crossover to another study arm as a failure, as well as other biochemical, physiological, and subjective markers of respiratory status. The study used a noninferiority model to determine how HFNI compares to traditional NIPPV. 204 patients were enrolled in the trial, 104 were randomized into the HVNI group and 100 were randomized into the NIPPV group. Analysis was done in an intention to treat model.


  • Intubation rate of the HVNI group was 7% (95% CI 2-12%).

  • Intubation rate of the NIPPV group was 13% (95% CI 6-20%).

  • Risk difference=-6%; 95% confidence interval -14% to 2% meets noninferiority.

  • Failure rate was 26% in the HVNI group (95% CI 17-24%) with the majority of those crossing over to NIPPV.

  • Failure rate was 17% in the NPPV group (95% CI 2-20%).

  • Risk difference 9%; confidence interval –2% to 20% meets noninferiority.


HVNI is noninferior to NIPPV for initial treatment of undifferentiated respiratory failure.


  • Wouldn’t it be great if we had an easier to use, more comfortable modality for patients who initially present to the ED with respiratory failure? And this paper tried to address this, but there were many limitations.

  • One obvious limitation is the inability to blind the treatment team by nature of the devices contributing to a possible early bias. The determination of treatment failure is sometimes subjective and based on the physician’s clinical judgement. The opinion of the treating physician in regards to HVNI as an effective treatment or not can skew the results.

  • Crossover from HVNI to NIPPV was huge! Almost a quarter of patients in the HVNI arm ended up on NIPPV. In an intention to treat model this causes a big problem when determining noninferiority. The study was also not powered for subanalysis specific to different types of respiratory failure. When looking at the secondary outcomes there was no consensus on arterial versus venous blood gas samples, which limited the interpretation of blood gas parameters such as PaO2.

  • Also, the setup used is not necessarily what people may have in their EDs for HFNC. Some places just have a large bore NC connected to wall oxygen turned up to 15L/min. This trial was looking at a system with a blender to provide pressurized air and oxygen at 25-30L/min. So we definitely cannot extrapolate this data to other HFNC system.

  • The results of the article showed that there was not a statistical difference in the failure rate of HFNI vs NPPV in the initial treatment of undifferentiated respiratory failure. In my opinion that makes this an important article because many patients find greater comfort in the use of HFNI over NIPPV. For me it makes me feel more justified in utilizing HVNI at least initially in the ED to help respiratory distress. I can see it being of benefit in the patient who arrives anxious and doesn’t want to try the restrictive mask. HVNI also allows for another noninvasive modality attempt prior to the definitive intubation.

  • This paper also makes a strong argument for the use of HVNI prior to any ABG/VBG results. The full adoption of such a practice does require some more study. Further investigation into the efficaciousness of both of these respiratory support modalities depending on the specific cause of respiratory failure may be beneficial. Since the study wasn’t powered for subgroup analysis we are still left with the older trials which show that HFNC is useful in hypoxemic respiratory failure, but the utilization in hypercapnic failure compared to NIPPV needs further review.