High-Flow Nasal Cannula in Pediatric Patients: A Survey of Clinical Practice

Discussion

In this survey of RTs, there was no consensus for the definition of HFNC in pediatric patients, or how to set or adjust flow. Most respondents used CPAP or NIV when patients failed HFNC. Most respondents indicated that RTs assessed patients every 4 h. More than three quarters of respondents delivered aerosol therapy via HFNC, and most used a vibrating mesh nebulizer placed on the dry side of the humidifier.

HFNC was defined as any heated gas delivered by nasal cannula for 49% of respondents, while 21% indicated they set the flow at or above the patient's inspiratory flow. Importantly, no respondents indicated that they used a scientifically validated method to ensure set flow was greater than or equal to the patient's inspiratory flow. Walsh and Smallwood¹³ have put forth a method to approximate the initial flow setting for HFNC based on the patient's age and weight; to our knowledge, however, this method has not been validated in clinical practice.

Clearly defining HFNC is important because a standard definition for HFNC would allow clinicians to compare patient outcomes across centers, to provide standard definitions for clinical trials, and to provide program benchmarking. The use of a clear definition of HFNC has important patient safety implications as well, because many centers require patients on HFNC to be in the ICU environment. Determining which patient populations, set flows, and F_IO2 requirements are safe to be outside the ICU are important areas where additional data are needed. Therapeutic goals also need to be taken into consideration, as patients placed on HFNC to receive aerosol therapy or therapeutic gases, or decrease discomfort associated with regular nasal cannula use, may not require ICU-level care. The frequency of monitoring and assessments by RTs could be determined on the basis of therapeutic goals to optimize therapy and to avoid unnecessary assessments. These parameters will likely need to be determined according to the needs of individual institutions and their comfort level utilizing HFNC in areas with less frequent monitoring than the ICU or the emergency department.

Optimal methods to determine initial flow and to adjust the flow for HFNC are unknown. Initial flows in clinical trials have used 2 L/kg/min in infants post cardiac surgery and in infants with moderate/severe bronchiolitis or pneumonia, 1 L/kg/min for infants with mild/moderate bronchiolitis, and 1–3 L/kg/min for subjects in the pediatric ICU with respiratory distress due to bronchiolitis and pneumonia.^{10–12,14,15} Weiler et al¹⁶ recently evaluated HFNC titration in 21 children (median age 6 mo, median weight 6.5 kg) and found the optimal flow to decrease work of breathing (as measured by the pressure rate product) was 1.5–2 L/kg/min. Importantly, the authors did not increase flow above 2 L/kg/min due to concerns about air leak and subject tolerance.¹⁶ Only 19% of our respondents indicated that a weight-based method was used to set initial inspiratory flow. The most common response was per provider orders; we did not ask specifically what methods were used by the providers to determine initial flow. Conceptually, using patient weight seems reasonable, but older children may not tolerate flows of 2 L/kg/min. In addition, some children may be significantly under or over their predicted body weight, thus making flow based on weight suboptimal. Using age instead of weight is another strategy, but many patients are underweight or overweight based on their age, especially children with chronic respiratory illnesses who tend to be below their predicted body weight.¹⁷ Direct comparisons of different methods for setting and adjusting flow are needed. Individual inspiratory flows also likely vary substantially between disease states, even for patients who are the same age, height, and weight. Device limitations prevent delivering > 2 L/kg/min for larger children because most devices have a maximum flow of 60 L/min.

Most of our respondents indicated that NIV or CPAP was the next intervention for patients who did not respond well to HFNC. We did not ask respondents what criteria were used to determine HFNC failure; however, it is important to recognize when HFNC is failing and escalate to appropriate level of support. When starting HFNC, clear clinical goals and criteria for escalation should be defined by the team, as well as whether the next step will be NIV, CPAP, or intubation. A delay in escalation to NIV may be associated with worse outcomes in children with asthma.¹⁸

Aerosol delivery via HFNC is somewhat controversial.^19,20 Bench studies to date indicate that aerosol is delivered to the airway; however in vivo data from humans are lacking. Aerosol delivery is also affected by flow, type of system used, cannula size, and type of nebulizer used.²¹ The majority of our respondents used vibrating mesh nebulizers to deliver aerosol therapy to patients, a method that has several advantages compared to jet nebulizers as it does not add flow to the system and does not require removal from the system for therapy. The optimal delivery method is also affected by the type of commercially available HFNC set-up used by individual centers. Differences between the CHA respondents and AARConnect respondents may have been due to the system(s) used at each center; we did not ask respondents what system(s) they used. Future research should focus on in vivo studies of not only aerosol deposition, but patient-oriented outcomes such as length of stay for asthmatic patients or hospital admission rate for those in the emergency department.

This study has several limitations. The relatively small sample size and demographic differences between the AARConnect respondents and CHA may have biased the results. Respondents may have had special interest in HFNC practice. There were no statistically significant differences between the groups for clinical practice; however, the survey was underpowered to detect differences. In an effort to keep the survey brief, subtleties in clinical practice could not be detected. Only members of the CHA and AARC members with access to the AARConnect Neonatal/Pediatric, Management, and Help Line sections were surveyed, and these respondents may not be representative of RTs as a whole. Questions may not have been worded clearly, and by their nature surveys cannot evaluate nuances in clinical practice. Finally, to keep the survey focused, we did not ask any questions regarding HFNC weaning methods or strategies. These topics are also debated and are deserving of further study.