Burden of disease and change in practice in critically ill infants with bronchiolitis

Discussion

This large binational study, including over 9000 critically ill infants with bronchiolitis, demonstrates that severe bronchiolitis is responsible for a huge burden of disease, resulting in over USD30 million direct costs each year in Australia and New Zealand. Bronchiolitis has a major impact on ICU resource consumption, accounting for >25% of non-elective PICU admissions in infants. In the absence of high-grade evidence for treatment of bronchiolitis in ICU, we observed a dramatic increase in the number of children admitted to intensive care in Australia and New Zealand with bronchiolitis, concomitant with the increased use of HFNC therapy. In view of close to zero mortality in infants with severe bronchiolitis, these findings indicate an urgent need for future studies investigating whether a proportion of these patients may be safely managed outside the ICU, hence reducing associated healthcare costs.

By estimating population-based ICU admission rates, we were able to document a major change in practice over the past decade, with a dramatic increase in ICU admissions in infants with bronchiolitis that do not require invasive ventilation. The increase observed in general ICUs may have been enhanced by reporting bias, as more general ICUs have contributed data to the study database over recent years. However, the findings were reproduced in subgroup analyses restricted to paediatric ICUs, and restricted to those centres that had provided data for the entire study duration. Infants with bronchiolitis admitted in recent years to ICU were less likely to have underlying comorbidities, were older, and their predicted mortality was lower, which suggests changing thresholds for ICU admission. At the same time, we observed a considerable variability in practice between units, with over six-fold difference in risk-adjusted intubation rates that was not explained by ICU type, size, nor major patient factors. Our findings support a recent North American study on bronchiolitis that observed >3.5-fold variation in the risk of intubation between paediatric ICUs, and a high variability in the proportion of patients exposed to other non-evidence-based interventions [15]. While bronchiolitis hospital admission rates and length of stay have been proposed as benchmarks in the NHS Atlas of Variation in Healthcare (www.chimat.org.uk/variation), ICU resource use has not been investigated. Notably, the study period witnessed the implementation of rapid response teams and early warning tools (EWTs) in the major paediatric hospitals, followed by regional hospitals. While EWTs have been shown to predict the need for ICU admission, and reduce cardiopulmonary arrests in the ward setting [16], to the best of our knowledge they have not been validated specifically for bronchiolitis. Indeed, infants with severe bronchiolitis commonly manifest tachycardia, tachypnoea, and increased work of breathing, and may trigger EWTs relatively rapidly, with an unknown impact on health outcomes and resource utilisation.

Our data show that this increase in ICU admission rates was associated with a rapid expansion in costs, despite a reduction in LOS in intensive care and a shorter duration of ventilation. Our costs are comparable to those in a recent European study, which reported a four times higher cost for children requiring ICU compared to those treated on the ward (EUR8061 versus EUR1834) [17].

Our study demonstrates that, despite lack of larger trials on HFNC use in infants, HFNC therapy has become the most common support mode for critically ill infants with bronchiolitis in Australia and New Zealand recently, with over 70% of admissions in 2014 supported with HFNC. At the same time, while the proportion of admissions requiring intubation and invasive ventilation dropped, we did not observe a significant reduction in the absolute number of children requiring invasive ventilation. Our observation parallels those of several single-centre studies that reported a drop in intubation rates in patients with severe bronchiolitis following the increased use of NIV and HFNC [10, 12, 13, 18–22], and suggests that these have led to a major change of practice in the absence of higher grade evidence. Recent large randomised trials in adults have demonstrated efficacy of HFNC therapy for acute hypoxemic respiratory failure [23, 24]. In the neonatal population, HFNC therapy was shown to be non-inferior to nasal CPAP [25, 26]. While the physiological effects of HFNC therapy, including provision of 4–5 cmH₂O CPAP, reduction in inspiratory work of breathing through provision of flows matching or greater than peak inspiratory flows, and washout of nasopharyngeal dead space have been independently confirmed [27, 28], there is a lack of randomised controlled trials to support either outcome benefit or cost benefit of HFNC therapy in the paediatric age group. A recent adult single-centre study reported higher mortality in patients treated with HFNC that required delayed emergency intubation [29]. In our cohort, the increasing use of HFNC was not associated with a delay in intubations, and in fact the proportions of late intubations, mortality, length of mechanical ventilation, and length of ICU stay decreased.

Advantages of HFNC therapy include an excellent safety profile, low equipment costs, easy application, and increased patient comfort compared to other forms of respiratory support. As a result, the application of HFNC therapy outside ICU settings has been increasingly tested, and single-centre studies suggest safety, feasibility, and efficiency of such an approach [18], although the translation of this approach into better outcomes remains to be proven. Bronchiolitis is a relatively uniform disease with close to zero mortality, patients mostly demonstrate gradual rather than precipitous deterioration, and predictors can be used to stratify severity [30]. As a result, this disease may in fact be optimally suited to design interventional trials aiming to reduce the number of intensive care admissions [31].

We believe the results of this study have several implications for the design of further research on bronchiolitis. First, they highlight an urgent need to validate early warning tools for bronchiolitis, as these may have a direct impact on intensive care admission rates. Secondly, improved markers of severity, including viral genomic load [32], may assist in optimising risk stratification to the target groups most likely to develop severe respiratory failure. Thirdly, the variation in respiratory support practice suggests a lack of standardised protocols [5, 8]. Fourthly, the dramatic change of practice seen in respiratory support modes warrants a paediatric ICU trial of HFNC therapy versus NIV versus low-flow oxygen for infants and children with respiratory failure [24]. Finally, our data suggest that it is likely that a considerable proportion of intensive care patients may be good candidates to be considered for treatment outside ICU; whether HFNC therapy provides a cost benefit in such a setting has to be tested by large trials [31]. Importantly, trials on optimal respiratory support outside ICU may inform on the development and implementation of cost-effective interventions relevant at a global scale, with the potential to translate into mortality benefits in resource-poor countries [33].

Several limitations of this study need to be considered. While we were not able to assess overall hospital admissions due to bronchiolitis, we captured every patient admitted to a general or paediatric ICU in Australia and New Zealand providing data to the registry. Although bronchiolitis is well-defined as a clinical entity [8], overlap with reactive airway diseases may lead to diagnostic challenges, potentially resulting in overdiagnosis of the disease. The use of HFNC was not prospectively captured prior to 2010, and we therefore cannot comment on how frequently HFNC was used as a respiratory support mode prior to 2010. Furthermore, we did not have access to detailed microbiological data, and extended nasopharyngeal virus polymerase chain reaction tests were not consistently performed in centres throughout the study. Because the large majority of infants with bronchiolitis admitted to ICU did not receive intra-arterial monitoring on admission, we were unable to stratify by the degree of hypoxaemia. Finally, inspired oxygen fraction (F_IO2) to arterial oxygen saturation measured by pulse oximetry ratios could not be used as the true F_IO2 is unknown in HFNC or low flow oxygen therapy. In addition, we did not have access to carbon dioxide tension values measured on admission to ICU; however, elevated carbon dioxide tension values represent a common finding in bronchiolitis and may not necessarily indicate a need for intubation [34].

In conclusion, severe bronchiolitis remains responsible for a huge burden of disease. We observed a major change in practice in the management of severe bronchiolitis with dramatically increased use of HFNC therapy despite the lack of high-grade evidence for treatment benefit in this age group. Our data suggest that thresholds to admit bronchiolitis patients to intensive care have reduced over the past decade, with a major impact on healthcare-related costs and resource utilisation in ICUs in Australia and New Zealand. International trials addressing the risk stratification and safe management of bronchiolitis outside intensive care settings are urgently warranted.