Limitations
This study has some limitations to acknowledge. The lower sample size than initially planned may have caused a beta error (absence of significance in the intubation rate due to a low number of patients). Failure of therapy with a PEEP of 5 was extremely high (78%) and it is possible that further
continuation of the study might have required an interim analysis to discuss eventual amendments. Furthermore, even if not pre-designed by our protocol, we were not able to identify any specific subgroup of patients who might have benefit by different PEEP regimes. However, it must be remarked that patients with respiratory insufficiency due to bronchiolitis have frequently very similar clinical characteristics and those with bacterial coinfection be well represent a subgroup of more severe patients, refractory to CPAP approach. This study is also lacking a third “NIV arm” (applying a different interface, i.e., face mask) that may be a potentially interesting comparator, especially in most severe cases. However, the initial intention of this study was to apply helmet in all bronchiolitis. HCPAP support requires a learning curve for optimization of the effectiveness. In this study, we recruited centers with proven and skilled experience in HCPAP treatment for bronchiolitis, thus limiting the external validity of the study. It is also likely that this contributed to the early interruption of the trial, given the unexpectedly low enrollment rate with respect to the planned schedule. Furthermore, the intubation rate was heterogeneous across study sites, likely representing the challenge of the non-invasive respiratory management of these patients and the potential differences in HCPAP setup. However, in all centers, the primary outcome was not associated with initial PEEP level. Another potential weakness of the study was the absence of control regarding the timing of admission to PICU for bronchiolitis; this may vary significantly among different centers due to organizational aspects and availability of PICU beds. However, the ideal timing for PICU transfer is currently unknown, this did not appear to be different in the two groups and even if proactive admission to a PICU may affect the outcomes, this has not yet been proven. Furthermore, among the inclusion criteria, the maximum duration of 72 hours of HFNC therapy partially limited this weakness. Finally, bacterial coinfection was diagnosed through a nasal swab, which do not allow a definitive diagnosis of actual lung infection.