Infants with bronchiolitis with high-flow nasal cannula in the paediatric ward. Is there a role for the ROXi (Respiratory rate–Oxygenation index) to predict transfer requirement to the PICU?

doi:10.21203/rs.3.rs-4992200/v1

Introduction: The respiratory rate–oxygenation index (ROXi) has been shown to be a reliable tool to predict the risk for HFNC failure in adult patients with lower respiratory tract infections. In paediatrics, the reliability of this index to predict HFNC failure appeared mitigated. In specific population, such as infants with moderate to severe bronchiolitis hospitalized in paediatric ward with HFNC, its ability to predict transfer requirement to the PICU needs to be further evaluated. The main objective of this study was to evaluate the ability of the ROXI collected at initiation of HFNC in the paediatric ward to predict the need for PICU transfer and/or HNFC failure in cases of moderate to severe bronchiolitis.

Methods: A retrospective review of patients aged 0 to 6 months with bronchiolitis who received HFNC within seven tertiary paediatric hospital over the last 5 epidemic seasons from 2018 to 2023 was conducted. Demographic, clinical, and biochemical variables were collected at admission and at the beginning of HFNC therapy support. Initial management and its evolution were described. Patients were compared depending on their transfer to the PICU during hospital stay. HFNC failure was defined as the need for Non-Invasive Ventilation. Multivariable regression analysis was used to determine parameters associated with transfer to the PICU and HFNC failure.

Results: We included 383 infants in this multicentric study (median age 63 days [7; 192]). 76 patients (20%) requiring HNFC were finally transferred to the PICU with a median of 2 days after the hospitalization. Only 40 children (10%) benefited from ventilatory escalation and for children transferred to a PICU, only one patient was intubated. In our population, the optimal ROX index for prediction of PICU requirement and HNFC failure of 6.9 (sensibility 53.1% and specificity 79.8%) and 7.6 (sensibility 62.5% and specificity 66.8%), respectively. In the multivariate analysis, explanatory variables for both transfer to the PICU and HFNC failure were preterm birth, younger age (under 3-month-old), and mWCAS ≥ 3. Besides, SpO2 ≤ 92% at admission and hypotonia were considered as risks factor for transfer and HFNC failure, respectively

Conclusion: HFNC appeared to be a safe tool for the management of moderate to severe bronchiolitis in the paediatric ward. However, it is necessary to identify patients for whom management remains safe in the paediatric ward, and in this context ROXi appears to be an interesting marker.

bronchiolitis

high-flow nasal cannula

paediatric ward

non-invasive ventilation

Pediatric intensive care unit

• Over the past decade, there has been a growing interest in the use of high flow nasal canula (HFNC) for moderate to severe bronchiolitis hospitalised outside the paediatric intensive care unit (PICU)

• Patients that could be safely managed with HFNC in in this setting remain to be determined and to this extend the ability of the respiratory rate–oxygenation index (ROXi) needs to be further evaluated.

What is New:

• The ROXi appears to be an interesting identifier of patient with moderate to severe bronchiolitis who could be safely managed in the paediatric ward

• The length of HFNC support was shown to be longer for patients who were not referred to the PICU, highlighting the importance for PICU teams to provide remote support to the paediatric teams in charge of those patients.

In high-income countries, bronchiolitis is the main cause of hospitalization in infants [1]. Among hospitalized patients, two to 6% required to be transferred to the paediatric intensive care units (PICU). Determining which patients need to be transferred to the PICU could be challenging. In 2023, the Groupe Francophone de Réanimation et Urgence Pédiatriques (French‑speaking group of paediatric intensive and emergency care; GFRUP) issued expert recommendations on the management of acute severe bronchiolitis in infants under 12 months of age admitted to the PICU [2]. In these guidelines, non-invasive ventilatory support (NIV), continuous positive airway pressure (CPAP), is recommended as first-line treatment of severe bronchiolitis. However, for less severe patients, high-flow nasal cannula (HFNC) could be considered as the first step of the respiratory support [2]. Over the past decade, there has been a growing interest in the use of HFNC outside the PICU due to its relative safety, ease of use and comfort [3–6]. Since then, the use of HFNC has spread to conventional paediatric wards and paediatric emergency departments [3–6]. However, there is still no consensus on the optimal use of HFNC outside the PICU. Besides, the patients that could be safely managed with HFNC in in this setting remain to be determined, as HFNC failure occurs in 10–15% [3–6]. In this scenario, the worsening of respiratory distress in these infants requires an increase in non-invasive or invasive respiratory support, which should probably be managed within a PICU [7, 8].

Several studies have identified demographic and physiologic parameters associated with HFNC failure, but none appeared to be totally reliable to be used as a triage tool [8–13]. In adult population, the respiratory rate–oxygenation index (ROXi) has been shown to be a reliable tool to predict the risk for HFNC failure in patients with lower respiratory tract infections [14, 15]. In the PICU, the reliability of this index to predict early response to HFNC appeared mitigated [16–18]. Nonetheless, its reliability to predict transfer requirement from the paediatric ward to the PICU of infants with moderate to severe bronchiolitis with HFNC remains promising and needs to be further evaluated [19, 20].

The primary objective of this study was to evaluate the ability of the ROXI collected at initiation of HFNC in the paediatric ward to predict the need for PICU transfer and/or HNFC failure. The secondary objective was to determine other the predictive factors of HFNC failure and transfer requirements. The tertiary objective was to describe the characteristics and management of infants under 6 months old with bronchiolitis managed with HFNC in the paediatric ward.

Design and settings.

A multicentre retrospective cohort study was performed over the epidemic seasons (October to February) from 2018 to 2023 in seven paediatric and/or neonatal conventional or intermediate care units in Basse Normandie, France, representing 80% of the units in this region. The seven participating units were all in a general or a university hospital and were all affiliated to the same PICU. Patients aged 0 to 6 months with a clinical diagnosis of bronchiolitis supported with HFNC and hospitalized in one of the participating centres were included in the study. Patients were excluded if they received non-invasive ventilation (NIV) prior to HFNC, HFNC ordered by the reference PICU team awaiting transfer to PICU, HFNC for less than two hours or if their parents opposed the use of the data for any retrospective study. This study was approved by the Local Research Ethics Committee for health research of the University Hospital of Caen on June 19th, 2023 (ID 4375). The Institutional Review Board waived the need for informed consent.

General management

All patients were treated according to the unit's standards of practice, in terms of monitoring, nutritional support or adjuvant therapy. Patient management, especially the introduction of HFNC, was left to the discretion of the medical team in charge of the patient and guided by the paediatrician’s clinical and paraclinical assessment. On a regular basis, in our region, HFNC was indicated when patient presented a respiratory rate superior at 60–70 per minute, or hypercapnia with partial pressure of carbon dioxide (PCO2) between 50 and 60 mmHg, or acidosis with pH between 7.25 and 7.34, or need standard oxygen therapy superior at 2 litters per minute to maintain oxygen saturation (SpO2) superior at 94%, or modified Wood's Clinical Asthma Score (m-WCAS) [21] superior at 3. No protocol was specifically designed for this study. The devices used were Airvo2 or Optiflow (Fisher & Paykel Healthcare). The HFNC settings (inspired fraction of oxygen (FiO2) and flow rate were defined and adjusted by the paediatrician based on its personal expertise.

Study protocol

Patients were retrospectively selected from the medical information department of each participating hospital. The selection of the files was made based on the coded diagnosis given at the end of the hospitalization, following the International Statistical Classification of Diseases and Related Health Problems, Tenth revision (ICD-10): 1) J21.0 – Acute Bronchiolitis due to respiratory syncytial virus; 2) J21.1 – Acute Bronchiolitis due to Human metapneumovirus; 3) J21.8 – Acute Bronchiolitis due to other specified organisms; 4) J21.9 – Acute Bronchiolitis, unspecified [22]. Patients were retrospectively allocated into two groups (PICU transfer group and paediatric group) depending on the transfer of the patient to the PICU during its hospital stay. Transfer requirement was decided by the intensivist paediatricians working in the referral PICU based on the description of the patient clinical status. On a regular basis, in our region, patients with a respiratory rate superior at 70 or inferior at 30 per minute, or more than three significant apnoea per hour, or acidosis with pH inferior at 7.25 and hypercapnia with PCO2 superior at 60 mmHg, or hypoxia with SatO2 inferior at 92% despite FiO2 superior at 60% or neurologic failure usually referred to the PICU to received NIV.

Endpoint

HFNC failure was defined as the need for mechanic ventilation, either invasive or non-invasive, including CPAP and bilevel positive airway pressure (BiPAP). Decision to upgrade the ventilatory support was left to the discretion of the medical team in charge of the patient based on regional practices.

Data extraction

Patient data was extracted from the final medical report and input on Excel spreadsheet. Demographic data (age, sex, past medical history, birth term, birth weight, admission weight, pre-hospital symptoms, time of hospital admission and viral cause) were collected upon admission. Past medical history was categorized as twin pregnancy, first bronchiolitis, previous ICU admission for respiratory support, bronchopulmonary dysplasia, passive smoking, family history of asthma or atopy in a first-degree relative. We reported clinical, biochemical and imaging data at admission and at the beginning of HFNC therapy, including the Wang Bronchiolitis Severity Score (WBSS) [23] and m-WCAS [21] both retrospectively evaluated. We also collected chest X-ray interpretation when it was carried out. Initial management and its evolution were described, including type and volume of hydration and/or feeding support, and adjuvants therapies. ROXI at HFNC initiation was retrospectively calculated. It was defined as the ratio of oxygen saturation as measured by pulse oximetry/oxygen fraction (FiO2) to respiratory rate [14]. We retrieved time to HFNC initiation, need for escalation to another ventilation support treatment, time of transfer and reason, total duration of HFNC and length of stay.

Statistical analysis

This study was a descriptive cohort study in which all patients meeting the criteria in the given period were to be included. No sample size was calculated. Variables distributions was assessed by the Shapiro-Wilk comparison test. Continuous variables were expressed as mean (standard deviation) or median [min-max] according to their distribution and were compared with Student T test or Mann-Whitney U test as appropriate. Qualitative variables were described as absolute numbers (percentage) and were compared by Khi-2-test, with Monte Carlo simulation with 2000 replicates when appropriate (small sample size < 5). The univariate and multivariate logistic regressions were performed to determine risk factors for PICU transfer and HFNC failure. The variables significantly associated with PICU transfer on univariate analyses (p < 0.05), clinically relevant according to the literature, with limited missing values (< 10) and no colinear with other covariates were considered for inclusion in the multivariate logistic regression analysis. We explored collinearity among these potential risk factors by calculating the variance inflation factor (VIF), with a value above 5 indicating collinearity. In the multivariate model, we used a backward procedure among qualified variables for selecting independent risk factors associated with PICU transfer, with the threshold of p-value < 0.05 to remain in the model. In the multivariate model, we include the ROXi as a dichotomic variable, we chose the ROXi thresholds of 6.67 and 5.39 based on Kannikeswaran et al [20]. To assess the accuracy of ROXi for correctly classifying patients who would require PICU and succeed or failed HFNC, the receiver operating characteristic (ROC) curve was performed, and the area under the ROC curve (AUROC) was calculated. According to the ROC analysis, the best cut-off points were calculated using Youden’s index. Length of HFNC in the whole cohort and in HFNC responders and total length of non-invasive support (HFNC + NIV) in both groups were depicted using Kaplan–Meier curves and compared with the Log-rank test. The level of statistical significance was set at a p-value < 0.05. Statistical analyses were performed using open-access R software, Version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria).

Flowchart and patients’ demographic

During the study period, 729 infants under 6-month-olds were hospitalized for bronchiolitis with oxygen supply, among them 346 were not included because they only received standard oxygen therapy. We therefore included and analysed 383 (53%) infants. Among them 76 patients (20%) were transferred to the PICU (Figure 1). HFNC was administered at admission for 214 infants (56%), and later during the paediatric stay for the remaining 169 (44 %). Patients transferred in the PICU received less often HFNC at admission (29 (38%) vs 185 (60%), p<0.001), but no difference in the delay between admission and initiation of HFNC was observed between groups. ICU transfer was due to hypercapnic acidosis (34/76 (45%)), aggravation of respiratory distress (25/76 (33%)), persistent hypoxia with FiO2 > 50% (11/76 (15%)), neurological failure (3/76 (4%)) or logistical issues (3/76 (4%)). All ICU transfer were completed before the sixth day with a median of 2 days [1.0 – 6.0]. Most of the patient in the PICU transfer group issued HFNC failure (61/76 (80.2%)). Only one patient was intubated following septic shock. In the no transfer group, HFNC failure was observed in 12 (4%) patients, all of them were hospitalized in the intermediate care unit of one centre. Overall, HFNC failure was observed in 73 (19%) patients.

Demographic characteristics at admission to the paediatric unit are detailed in the Table 1. Patients transferred to the PICU weigh less, were younger (p<0.001) and were more often born preterm (p=0.004). The most common agent was respiratory syncytial virus (78%), especially in the ICU transfer group (233 patients, 86%, p=0.019) (Table 1).

Clinical and paraclinical features at admission and HFNC initiation

Table 2 depicted the clinical and paraclinical characteristics of the patients at admission and at HFNC initiation. At admission, the transfer to the PICU was associated with a higher heart rate, a lower SpO2, an history of hypotonia, a lower pH, and an abnormal chest X-ray, especially showing associated pneumonia or atelectasis. At HFNC initiation, the transfer to the PICU was associated with a higher respiratory rate, a higher respiratory distress with use of accessory muscles and higher severity scores (Wang and m-WCAS), neurologic failure, deterioration of general status and apnoea, as well as a lower pH, and an abnormal chest X-ray (Table 2). The ROXI was lower in the transfer group (6.75 [2.14; 25.40] vs 8.66 [2.94; 29.10]; p<0.001.).

Risk factors of transfer and HFNC failure

ROXi as a continuous variable or as a dichotomic variable with a threshold of 5.39 was not considered in the multivariate model as an explanatory variable for both transfer to the PICU and HFNC failure (Table 3 and supplemental digital content 1). However, it became discriminant at a threshold of 6.67 for both transfer to the PICU (OR [95%CI]: 3.0 [1.4 - 6.2], p = 0.004) and HFNC failure (OR [95%CI]: 2.5 [1.2 - 5.3], p = 0.02) (supplemental digital content 1). The ROC curves determined, in our population, the optimal ROX index for prediction of need for transfer to the PICU and HNFC failure of 6.86 and 7.6, respectively (figure 2).

Among the clinically and statistically significant variables included in the multivariate logistic regression, apart from ROXi, 7 parameters remained in the model after the backward procedure. The independent explanatory variables for both transfer to the PICU and HFNC failure were preterm, younger age (between 3 and 1 month and under 1 month), and mWCAS ≥ 3. SpO2 ≤ 92% at admission in room air and hypotonia were considered as risks factor for transfer and HFNC failure, respectively (Table 3 and supplemental digital content 1).

Ventilatory support management

The mean length of HFNC was significatively lower in the transfer group (12 hours [0.5 - 120] vs 90 hours [10 - 240], p<0.001), even when comparison was restricted to the HFNC responders (49 [6 - 120] vs 96 [10 - 240]; p<0,01). Likewise, there was a shorter total length of ventilatory support (HFNC+NIV) in the transfer group with a median of 78 hours [2 - 210] vs 96 hours [10 - 268] (p=0.042, log rank=) (Figure 3). The strategy for weaning patients from ventilatory support was significantly different between the groups. Patients in the PICU were more often switched from NIV or HFNC to oxygen (51 (70%)) whereas they were maintained on HFNC until complete resolution of oxygen requirements (282 (92%)) in the paediatric ward (p < 0.001).

In this multicentric study, we showed that 53% of the patients hospitalized outside the PICU received HFNC. Among them, 20% were finally transferred to the PICU within 6 days. In France, 53.3% of hospital used HFNC in acute bronchiolitis outside the PICU [5]. A similar rate of use is reported in other countries with variation related to indications, flow rate, feeding, and protocol use [24, 25]. Over the last two decades, PICU admissions for bronchiolitis multiplied by 10 [26]. The burden of bronchiolitis for the PICU is concerning and even more aggravated by the difficulties faced by the public hospital system due to a lack of hospital beds and a shortage of human and material resources. From this point of view, the results of the clinical trial of Franklin et al. showing a lower rate of escalation of care due to treatment when HFNC was used early during the hospital admission compared to standard oxygen therapy in general paediatric ward were promising [4]. However, the cost-effectiveness of this approach remains challenging, as well as safety, especially regarding the appropriate monitoring of patient requiring HFNC in the paediatric ward. Even though, several studies have shown that the use of HFNCs in a general ward was safe, with very few adverse events described [27–28], most of these studies were conducted with a favourable high nurse to patient ratio, which questioned the external applicability of their results. However, although reassuring about procedural safety and promising about the need for NIV, most studies have shown that HFNC, compared with standard oxygen, does not prevent patients from being admitted to the PICU, even when used early in the emergency department [3, 4].

To this regard identifying the patients who will require transfer to the PICU because of a higher risk of HFNC failure or other complications is mandatory. The question of identifying non-responders to HFNC has already been raised several times [3–6, 9–13]. In this context, this multicentre study in conventional paediatric wards in France, confirmed that prematurity, young age (< 3 months old), high respiratory score, and hypoxemia were risk factors of HFNC failure and PICU requirement, in adequation with data from the literature [3–6, 9–13].

Although these results are useful in guiding practitioners in their decision to transfer patients to the PICU, we note that a reliable indicator is lacking. ROXI has been identify as a good predictor of HFNC success in adult patients with lower respiratory tract infections [14] and recently some teams have assessed the feasibility of extending this to the paediatric population, in the PICU and in the paediatric ward as-well [16–18]. Regarding patients with bronchiolitis, Kannikeswaran et al. observed an association between ROXI and the need of positive pressure ventilation in children < 2 years old on HFNC in the paediatric emergency department [20]. Lately, in the PICU, Milesi et al. concluded that ROXI, calculated at H0 and H1 on 286 patients, had no early discriminatory capacity to predict HFNC failure [16]. They concluded that ROX-HR (= 100x (ROX/Heart Rate)) performances were better but poorly discriminant with a value at H1 4.07 (3.17; 5.41), and their result conforms with Webb et Al. [18]. In contrast, Nascimento et al., in a prospective and multicentric cohort of 102 infants in the PICU, observed that the ROXI at 12h from HFNC initiation was a discriminant tool to difference HFNC failure from success (5.8 (95% CI 4.7– 7.1) vs 7.7 (95% CI 7.2–8.2), p = 0.005), with an area under the curve of 0.716 (p = 0.016) [17]. In our cohort, we found that ROXI was significantly higher in the group of patients that were not transferred to the PICU. We identified that by applying to our population, the thresholds proposed by Kannikeswaran et al, patients with a ROXi > 6.67 have up to a 6-fold and 5-fold greater odds of transfer to the PICU or HFNC failure, respectively. Unfortunately, we were not able to record ROXI variation at different time points, nor to evaluate the ROX-HR. Nevertheless, further validation on a larger scale is mandatory. Multimodal evaluation of these patients seems mandatory and the development of a clinical score combining ROXI and m-WACS as well as other factors (Age, term of birth) is to be considered [29].

For patients who were not referred to the PICU, our study shows that their median duration of HFNC support was 2 times longer than those admitted to the PICU, and 4 times longer when considering all types of ventilatory support. These results are even more problematic given that these patients are hospitalized during the epidemic season, when paediatric care capacity is exceeded. In this context, it is important for PICU teams to provide remote support to the paediatric teams in charge of those patients, using modern day communication tools and telemedicine [30].

This study has several limits. First, this is an observational and retrospective study with no control group, so we could not assess outcomes in children not treated with HFNC and we may face information and confusion bias. Even if multicentric character, it was not a national scale, so our population may not reflect general population. Thus, given the limited sample size, further evidence is required to extrapolate these results and conclusions. Besides, in the absence of protocol for the use of HFNC and for PICU transfers, these decisions were left at the discretion of the medical team and consequently not entirely uniform. Unfortunately, we were not able to perform dynamic evaluation of ROXi and were limited regarding other risk factors. This dynamic assessment at different time points is important in daily practice and should be considered in future studies [11]. Finally, in comparison with Milesi et al. study [6], our patients appeared to be less ill according to Wang's severity score, with a third classified as "mild", a lower mean PCO2, and ventilatory escalation for only 19% of our cohort. This result could indicate that HFNC is mainly an alternative tool in less severe bronchiolitis.

Our study investigated the demographic, clinical and paraclinical factors associated with HFNC-treated bronchiolitis in general paediatric wards in children under 6 months of age and its management. In our population, 20% of these children were transferred to a PICU, mostly during the first 48 hours. Hence, identifying patients for whom management remains safe in the paediatric ward is essential and, in this context, ROXi appears to be an interesting marker that needed further investigations. Besides, as the length of HFNC support was shown to be longer for patients who were not referred to the PICU, PICU teams should be inspired to provide remote support to the paediatric teams in charge of those patients, to limit the service overload during epidemic season.

BiPAP Bi-level Positive Airway Pressure

CPAP Continuous Positive Airway Pressure

FiO2 Inspired Fraction of Oxygen

GFRUP Groupe Francophone de Réanimation et Urgence Pédiatriques

HFNC High-flow Nasal Cannula

HR Heart Rate

ICU Intensive Care Unit

LOS Length of Stay

m-WCAS modified-Wood's Clinical Asthma Score

NAVA Neurally Adjusted Ventilatory Assist

NIV Non-Invasive Ventilation

PCO2 Partial Pressure of Carbon Dioxide

PEEP Positive End Expiratory Pressure

PICU Pediatric Intensive Care Unit

RR Respiratory Rate

SpO2 Oxygen Saturation

WBSS Wang Bronchiolitis Severity Score

Acknowledgements: We acknowledge with gratitude those respected colleagues who were involved in the management of the patients and who provided the patients’ data.

Authors’ Contributions: CLP, CM, NS and DWB developed the study protocol; CLP performed data extraction; DC and DWB performed statistical analysis; CLP, JCC, CA and DB conceptualized the article. CLP, DC, CM, NS and DWB drafted the first version of the manuscript. All authors reviewed, edited, and approved the manuscript.

Availability of data and material: All data that support the findings of this study will be available from the corresponding author upon reasonable request.

Ethics approval: Our protocol was analysed within the Research Ethics Committee (CLERS) and was approved on on June 19th, 2023 (ID 4375). This study protocol was deemed to comply with the legal standard applicable in France and with the European regulation on the protection of individuals.

Informed consent: not applicable

Ethical clearance: not applicable

Funding: No funding was received for the conduct of the review

Competing interest: none was declared.

Dalziel SR, Haskell L, O’Brien S, Borland ML, Plint AC, Babl FE et al (2022) Bronchiolitis Lancet 400(10349):392–406
Milési C, Baudin F, Durand P, Emeriaud G, Essouri S, Pouyau R et al (2023) Clinical practice guidelines: management of severe bronchiolitis in infants under 12 months old admitted to a pediatric critical care unit. Intensive Care Med 49(1):5–25
Durand P, Guiddir T, Kyheng C, Blanc F, Vignaud O, Epaud R et al (2020) A randomised trial of high-flow nasal cannula in infants with moderate bronchiolitis. Eur Respir J 56(1):1901926
Franklin D, Babl FE, Schlapbach LJ, Oakley E, Craig S, Neutze J et al (2018) A Randomized Trial of High-Flow Oxygen Therapy in Infants with Bronchiolitis. N Engl J Med 378(12):1121–1131
Panciatici M, Fabre C, Tardieu S, Sauvaget E, Dequin M, Stremler-Le Bel N et al (2019) Use of high-flow nasal cannula in infants with viral bronchiolitis outside pediatric intensive care units. Eur J Pediatr 178(10):1479–1484
Kepreotes E, Whitehead B, Attia J, Oldmeadow C, Collison A, Searles A et al (2017) High-flow warm humidified oxygen versus standard low-flow nasal cannula oxygen for moderate bronchiolitis (HFWHO RCT): an open, phase 4, randomised controlled trial. Lancet 389(10072):930–939
Vahlkvist S, Jürgensen L, la Cour A, Markoew S, Petersen TH, Kofoed PE (2020) High flow nasal cannula and continuous positive airway pressure therapy in treatment of viral bronchiolitis: a randomized clinical trial. Eur J Pediatr 179(3):513–518
Groupe Francophone de Réanimation et d’Urgences Pédiatriques (GFRUP), Milési C, Essouri S, Pouyau R, Liet JM, Afanetti M et al (2017) High flow nasal cannula (HFNC) versus nasal continuous positive airway pressure (nCPAP) for the initial respiratory management of acute viral bronchiolitis in young infants: a multicenter randomized controlled trial (TRAMONTANE study). Intensive Care Med 43(2):209–216
D’Alessandro M, Vanniyasingam T, Patel A, Gupta R, Giglia L, Federici G et al (2021) Factors associated with treatment failure of high-flow nasal cannula among children with bronchiolitis: a single-centre retrospective study. Paediatr Child Health 26(5):e229–235
Kelly GS, Simon HK, Sturm JJ (2013) High-flow nasal cannula use in children with respiratory distress in the emergency department: predicting the need for subsequent intubation. Pediatr Emerg Care 29(8):888–892
Abboud PA, Roth PJ, Skiles CL, Stolfi A, Rowin ME (2012) Predictors of failure in infants with viral bronchiolitis treated with high-flow, high-humidity nasal cannula therapy. Pediatr Crit Care Med 13(6):e343–349
Baquedano-Lobera I, Bardella-Gil C, García-Iñiguez JP (2022) Analysis of predictors of response to high-flow oxygen nasal cannula therapy in a pediatric intensive care unit. Bol Med Hosp Infant Mex 79(4):222–227
Er A, Çağlar A, Akgül F, Ulusoy E, Çitlenbik H, Yılmaz D et al (2018) Early predictors of unresponsiveness to high-flow nasal cannula therapy in a pediatric emergency department. Pediatr Pulmonol 53(6):809–815
Roca O, Caralt B, Messika J, Samper M, Sztrymf B, Hernández G et al (2019) An index combining respiratory rate and oxygenation to predict outcome of nasal high-flow therapy. Am J Respir Crit Care Med 199(11):1368–1376
Junhai Z, Jing Y, Beibei C, Li L (2022) The value of ROX index in predicting the outcome of high flow nasal cannula: a systematic review and meta-analysis. Respir Res 23(1):33
Milesi C, Nogue E, Baleine J, Moulis L, Pouyau R, Gavotto A et al (2024) ROX (Respiratory rate- OXygenation) index to predict early response to high-flow nasal cannula therapy in infants with viral bronchiolitis. Pediatr Pulmonol [In press]
Nascimento MS, Zólio BA, Vale LAPA, Silva PA, de Souza L, Gonçalves TS LHR, et al (2024) ROX index as a predictor of failure of high-flow nasal cannula in infants with bronchiolitis. Sci Rep 14(1):389
Webb LV, Chahine R, Aban I, Prabhakaran P, Loberger JM (2022) Predicting High-Flow Nasal Cannula Therapy Outcomes Using the ROX-HR Index in the Pediatric ICU. Respir Care respcare.09765.
Yildizdas D, Yontem A, Iplik G, Horoz OO, Ekinci F (2021) Predicting nasal high-flow therapy failure by pediatric respiratory rate‐oxygenation index and pediatric respiratory rate‐oxygenation index variation in children. Eur J Pediatr 180(4):1099–1106
Kannikeswaran N, Whittaker P, Sethuraman U (2022) Association between respiratory rate oxygenation index and need for positive pressure ventilation in children on high flow nasal cannula for bronchiolitis. Eur J Pediatr 181(11):3977–3983
Duarte-Dorado DM, Madero-Orostegui DS, Rodriguez-Martinez CE, Nino G (2013) Validation of a scale to assess the severity of bronchiolitis in a population of hospitalized infants. J Asthma 50:1056–1061
International statistical classification of diseases and related health problems 10th revision. [Online].2019. [Cited 2024 Jan 4]. https://icd.who.int/browse10/2019/en
Wang EE, Milner RA, Navas L, Maj H (1992) Observer agreement for respiratory signs and oximetry in infants hospitalized with lower respiratory infections. Am Rev Respir Dis 145(1):106–109
Novak C, Vomiero G, de Caen A, Cooke S (2021) Current practices and policies regarding the use of high-flow nasal cannula on general pediatric inpatient wards in Canada. Paediatr Child Health 26(7):414–420
Turnham H, Agbeko RS, Furness J, Pappachan J, Sutcliffe AG, Ramnarayan P (2017) Non- invasive respiratory support for infants with bronchiolitis: a national survey of practice. BMC Pediatr 17(1):20
Linssen RS, Teirlinck AC, van Boven M, Biarent D, Stona L, Amigoni A et al (2022) Increasing burden of viral bronchiolitis in the pediatric intensive care unit; an observational study. J Crit Care 68:165–168
Betters KA, Gillespie SE, Miller J, Kotzbauer D, Hebbar KB (2017) High flow nasal cannula use outside of the ICU; factors associated with failure. Pediatr Pulmonol 52(6):806–812
Demir Ş, Anıl M, Yalçın G, Bardak Ş, Demir G, Berksoy E et al (2022) Early predictors of high- flow nasal cannula oxygen treatment failure in patients with respiratory distress admitted to the pediatric emergency department. Turk J Pediatr 64(4):648–657
Freire G, Kuppermann N, Zemek R, Plint AC, Babl FE, Dalziel SR et al (2018) Predicting Escalated Care in Infants With Bronchiolitis. Pediatrics 142(3):e20174253
Jacovas VC, Chagas MEV, Constant HMRM, Alves SS, Krauzer JRM, Guerra LR et al (2021) Telemedicine in Pediatric Intensive Care Units: Perspectives From a Brazilian Experience. Curr Pediatr Rep 9(3):65–71

Table 1. Patients’ characteristics prior to admission
Characteristics	All (n=383)			PEDIATRIC (n=307)		PICU (n=76)	*p-value*
Age of admission (days) — §	63 [7 ; 192]			70 [7 ; 192]		36 [10 ; 181]	<0.001*
< 1 month old	80 (21)			48 (16)		32 (42)	<0.001*
1 to 3 months old	175 (46)			141 (46)		34 (45)
> 3 months old	128 (33)			118 (38)		10 (13)
Weight of admission (kg) — §	5.135 [2.385 ; 10.680]			5.210 [2.680 ; 10.680]		4.010 [2.385 ; 9.560]	<0.001*
Sex — n (%) Male / Female	200 (52) 183 (48)			163 (53) 144 (47)		37 (49) 39 (51)	0.511
Birth weight (kg) — n (%)							0.39
< 3rd percentile	19 (5)			14 (4)		5 (7)
3rd - 10th percentile	24 (6)			17 (6)		7 (9)
> 10th percentile	340 (89)			276 (90)		64 (84)
Gestational age at birth (weeks of amenorrhea) — n (%)							0.004*
< 28	8 (2)			7 (2)		1 (1)
28 - 32	3 (1)			3 (1)		0
32 - 37	38 (10)			22 (7)		16 (21)
> 37	334 (87)			275 (90)		59 (78)
History and comorbidity — n (%) †
Twin pregnancy	18 (5)	14 (5)			4 (5)		0.76
First bronchiolitis	341 (89)	268 (87)			73 (10)		0.035*
ICU admission for respiratory support	28 (7)	19 (6)			9 (12)		0.14
Bronchopulmonary dysplasia	4 (1)	3 (1)			1 (1)		1
Passive smoking	20 (5)	14 (5)			6 (8)		0.26
Family history of asthma and/or atopia in a first-degree relative	70 (18)	61 (2)			9 (12)		0.137
Congenital heart disease	5 (1)	5 (2)			0		0.6
Pre-hospital symptom
Duration of symptoms (days) — §	3 [0 ; 13]	3 [0 ; 13]			3 [0 ; 7]		0.001*
Fever — n (%)	129 (34)	111 (36)			18 (24)		0.043*
Apnea — n (%)	11 (3)		6 (2)			5 (7)	0.044*
Cyanosis — n (%)	8 (2)		6 (2)			2 (3)	0.66
Hypotonia — n (%)	11 (3)		8 (3)			3 (4)	0.696
Viral cause — n (%)							0.019*
Respiratory syncytial virus	298 (78)		233 (76)			65 (86)
Rhinovirus/enterovirus	22 (6)		15 (5)			7 (9)
Other virus	12 (3)		12 (4)			0
Undetermined	51 (13)		47 (15)			4 (5)
Viral coinfection	23 (6)		15 (5)			8 (11)	0.099

§Median [min ; max].

*p-value < 0.05 was considered statistically significant.

†Individuals may have more than one type of previous medical history; column percentages will not add to 100%.

Table 2: Clinical and paraclinical data at admission and HFNC initiation

Characteristics	At admission in pediatric department				At HNFC initiation
Characteristics	All n=383	Pediatric n=307	PICU n=76	*p-value*	All n=383	Pediatric n=307	PICU n=76	*p-value*
*Heart rate (beats/min) — n (%)*				0.025*				1
< 80	0	0	0		4 (1)	3 (1)	1 (1)
80 - 160	204 (53)	174 (57)	30 (39)		270 (71)	217 (71)	53 (70)
160 - 180	139 (36)	103 (33)	36 (47)		86 (22)	69 (22)	17 (22)
> 180	40 (11)	30 (10)	10 (13)		23 (6)	18 (6)	5 (7)
*Respiratory* *rate* *(breaths/min) —* n *(%)*				0.558				*0.0015*
< 20	6 (2)	5 (2)	1 (1)		8 (2)	3 (1)	5 (6)
20 - 40	92 (24)	79 (26)	13 (17)		114 (30)	99 (32)	15 (20)
40 - 60	187 (49)	148 (48)	39 (51)		147 (38)	125 (41)	22 (29)
60 - 70	70 (18)	54 (17)	16 (21)		78 (20)	56 (18)	22 (29)
> 70	28 (7)	21 (7)	7 (10)		36 (10)	24 (8)	12 (16)
*Transcutaneous oxygen saturation (%) without oxygen supply — n (%)*				<0.001*
< 88	11 (3)	3 (1)	8 (10)
88 - 92	32 (8)	20 (7)	12 (16)
> 92	340 (89)	284 (92)	56 (74)
*Use of* *accessory* *muscles —* n *(%)*				0.483				0.005*
No retraction	54 (14)	44 (15)	10 (13)		29 (9)	24 (8)	5 (7)
Mild	110 (29)	93 (30)	17 (22)		61 (16)	56 (19)	5 (7)
Moderate	145 (38)	111 (36)	34 (45)		126 (33)	106 (35)	20 (26)
Severe	74 (19)	59 (19)	15 (20)		161 (42)	115 (38)	46 (60)
Missing data					6	6	0
*Apnea —* n *(%)*	10 (3)	7 (2)	3 (4)	0.407	36 (9)	22 (7)	14 (18)	0.007*
*Hypotonia —* n *(%)*	20 (5)	12 (4)	8 (11)	0.035*	35 (9)	18 (6)	17 (22)	<0.001*
*Retrospective evaluation of Wang score — n (%)*				0.428				<0.001*
Mild	175 (46)	143 (47)	32 (42)		114 (30)	101 (33)	13 (17)
Moderate	185 (48)	148 (48)	37 (49)		228 (60)	182 (60)	46 (61)
Severe	23 (6)	16 (5)	7 (9)		38 (10)	21 (7)	17 (22)
Missing data					3	3	0
*Retrospective evaluation of Woodm score — n (%)*				0.166				<0.001*
0 - 2,5	300 (78)	245 (80)	55 (72)		247 (65)	213 (70)	34 (45)
≥ 3	83 (22)	62 (20)	21 (27)		133 (35)	91 (30)	42 (55)
Missing data					3	3	0
*Retrospective calculation of ROX index after initiation of HFNC*					8.4 [2.1;29.1]	8.7 [2.9;29.1]	6.8 [2.1;25.4]	<0.001*
pH	7.37 [7.15;7.48]	7.38 [7.24;7.48]	7.36 [7.15;7.46]	0.009*	7.36 [7.20 ; 7.52]	7.37 [7.13 ; 7.45]	7.35 [7.20 ; 7.52]	0.031*
*PCO2* *(mmHg)*	47.8 ± 9.5	46.4 ± 10.4	50.1 ± 8.8	0.673	50.0 [25.0 ; 87.0]	50.0 [25.0 ; 87.0]	50.0 [35.9 ; 78.0]	0.639
*Lactate (mmol/L)*	1.6 [0.9;10.0]	1.5 [0.9 ; 3.0]	1.9 [0.9 ; 10.0]	0.894	27.8 ± 3.8	27.9 ± 3.7	27.8 ± 3.4	0.434
*Bicarbonate (mmol/L)*	26.3 ± 3.5	25.8 ± 3.7	27.2 ± 3.4	0.954	1.65 [0.7 ; 3.7]	1.6 [0.7 ; 2.5]	1.7 [0.9 ; 3.7]	0.705
*Chest X-ray interpretation — n (%)*				0.018*				0.028*
Normal	52 (28)	44 (27)	8 (23)		16 (20)	10 (20)	6 (20)
Bronchitis	86 (46)	76 (47)	10 (29)		23 (29)	18 (37)	5 (17)
Pneumonia	30 (16)	22 (14)	8 (23)		24 (30)	15 (31)	9 (30)
Chest distension	13 (7)	11 (7)	2 (6)		2 (3)	2 (4)	0
Atelectasis	16 (9)	9 (6)	7 (20)		14 (18)	4 (8)	10 (33)

§Median [min ; max].

*p-value < 0.05 was considered statistically significant.

Table 3: Multivariate analysis. Predictive factors for transfer to the PICU and NIV requirement.

	*Transfer to the PICU*			*NIV requirement*
	Events n = 76	OR [95%CI]	*p-value*	Events n = 73	OR [95%CI]	*p-value*
Gestational age at birth
> 37	59	Reference		59	Reference
≤ 37	17	3.7 [1.6 - 8.9]	0.003	14	2.5 [1.0 - 6.3]	0.05
Age
> 3 months	10	Reference		5	Reference
[1-3] months	34	3.8 [1.5 - 10.3]	0.006	37	11.0 [3.4 - 52.1]	<0.001
< 1 month	32	13.5 [4.8 - 42.5]	<0.001	31	31.7 [8.7 - 162]	<0.001
Apnea§
No	60	Reference		59	Reference
Yes	16	1.3 [0.5 - 3.4]	0.5	14	1.0 [0.4 - 2.6]	>0.9
Hypotonia§
No	58	Reference		55	Reference
Yes	18	1.8 [0.7 - 4.4]	0.2	18	2.8 [1.1 - 7.3]	0.03
SpO2 at admission
> 92%	56	Reference		58	Reference
≤ 92%	20	2.4 [1.0 - 5.7]	0.05	15	1.2 [0.4 - 3.3]	0.8
Heart rate at HFNC initiation
< 180 rpm	66	Reference		65	Reference
≥ 180 rpm	10	1.3 [0.5 - 3.5]	0.6	8	1.5 [0.6 - 3.8]	0.4
m-WCAS at HFNC initiation
0-2.5	34	Reference		38	Reference
≥ 3	42	3.0 [1.5 - 6.3]	0.002	35	2.1 [1.0 - 4.3]	0.05
ROXI†
≥ 5.39	53	Reference		46	Reference
< 5.39	11	2.1 [0.7 - 6.5]	0.2	10	2.8 [0.9 - 8.8]	0.08

§At any moment before initiation of high flow nasal canula. † missing data.

OR [95%CI]: Odds ratio with 95% confidence interval.

PICU: Pediatric Intensive Care Unit; NIV: Non-Invasive Ventilation; m-WCAS: modified Wood's Clinical Asthma Score; ROXI: Respiratory rate–Oxygenation index

No competing interests reported.

BronchioHFNCSDC1.docx
Supplementary content:
- Supplemental digital content 1: Predictive factors for transfer to the PICU and NIV requirement, multivariate analysis.

Infants with bronchiolitis with high-flow nasal cannula in the paediatric ward. Is there a role for the ROXi (Respiratory rate–Oxygenation index) to predict transfer requirement to the PICU?

Status:

Version 1

Abstract

Figures

What is Known

Introduction

Methods

General management

Study protocol

Endpoint

Data extraction

Statistical analysis

Results

Discussion

Conclusion

Abbreviations

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1