Evaluation of Prognostic Factors for High-Flow Nasal Cannula Therapy Failure in Infants with Bronchiolitis: A Prospective Study

doi:10.21203/rs.3.rs-4714148/v1

Download PDF

Research Article

Evaluation of Prognostic Factors for High-Flow Nasal Cannula Therapy Failure in Infants with Bronchiolitis: A Prospective Study

https://doi.org/10.21203/rs.3.rs-4714148/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Objective

This study aims to examine factors associated with High-flow nasal cannula (HFNC) therapy failure in infants with moderate to severe bronchiolitis, thereby optimizing and personalizing management by identifying patients at risk of treatment failure.

Methods

A prospective cohort study was conducted over two years (September 2021 - March 2023) in the pediatric intensive care unit of a tertiary-level university hospital. Infants aged 0–12 months with moderate to severe bronchiolitis requiring HFNC were included. Exclusion criteria were previous noninvasive or invasive mechanical ventilation, tracheostomy, uncorrected cyanotic congenital heart disease, or chronic respiratory or neuromuscular pathology. Clinical and biological characteristics were assessed at initiation and at H2 of HFNC treatment. Logistic regression was used to identify risk factors for HFNC therapy failure.

Results

Of the 154 patients included, 100 (64.9%) responded to HFNC, while 54 (35.1%) required escalation of care. Univariate analysis showed significant differences in age and weight at admission between responders and non-responders. Multivariate analysis identified weight below the 10th percentile, FiO2 requirement ≥ 42.5%, and Wang score ≥ 10 at HFNC initiation as predictors of failure. Persistence of tachycardia, tachypnea, hypercapnia, and FiO2 ≥ 48.5% after 2 hours of HFNC were also predictors of failure. The most frequent causes of HFNC failure were persistent increased respiratory effort (64.8%) and severe apnea (18.5%).

Conclusions

Key predictors of HFNC therapy failure include low weight, high initial FiO2 requirements, and severe clinical scores at initiation, as well as persistent clinical instability during the first hours of therapy. These findings can guide clinical decisions and improve outcomes for infants with bronchiolitis.

Bronchiolitis

infant

high flow nasal cannula

prognosis

Bronchiolitis remains one of the leading causes of hospital admission in infants, particularly during the early months of life. Despite numerous advancements in the understanding of this respiratory disease, its treatment continues to be a significant challenge for healthcare professionals. Among the increasingly utilized therapeutic interventions in bronchiolitis management, high-flow nasal cannula therapy (HFNC) has garnered particular interest due to its potential effectiveness in improving oxygenation and reducing the need for invasive ventilation [1, 2].

However, despite its growing use, it is increasingly evident that HFNC therapy is not uniformly effective in all infants with bronchiolitis. Some patients do not respond favorably to this therapeutic modality and require an escalation of care [3]. The underlying factors that influence this variable response to HFNC therapy remain poorly understood.

The present two-year prospective cohort study aimed at a comprehensive examination of factors associated with HFNC therapy failure in infants with moderate to severe bronchiolitis. By investigating these factors, we seek to optimize and personalize bronchiolitis management by identifying patients at risk of HFNC therapy failure.

This prospective cohort study was conducted in the pediatric intensive care unit (PICU) of a tertiary-level university hospital between September 2021 and March 2023. The study protocol was approved by the local ethics committee of the Fattouma Bourguiba University Hospital in Monastir, Tunisia, in accordance with the Declaration of Helsinki and good clinical practice. Informed consent was obtained from the parents of all participants.

Inclusion criteria were patients aged 0–12 months with a clinical diagnosis of moderate to severe bronchiolitis requiring HFNC. Patients were admitted from the emergency department, the pediatric ward, or transferred from a community hospital. Exclusion criteria included previous noninvasive ventilation (NIV) or invasive mechanical ventilation, tracheostomy, uncorrected cyanotic congenital heart disease, or chronic respiratory or neuromuscular pathology.

Oxygen requirements were defined according to the World Health Organization's recommendation if the transcutaneous oxygen saturation of hemoglobin (SpO2) was < 94% at room air. The Wang score was used to assess the severity of bronchiolitis in these infants [4]. SpO2 was measured using a pulse oximeter at room air. Two observers independently assessed all children on the bronchiolitis severity score. The first observer was a resident in pediatrics, and the second observer was a senior pediatric physician.

Protocol and Evaluation

The variables evaluated at admission included: age, sex, gestational age at birth, birth weight and weight at admission, signs of respiratory distress, SpO2 at room air, lung auscultation, respiratory rate, heart rate, Wang clinical severity score, blood gas analysis and FiO2 were assessed at initiation and at H2 of HFNC treatment. Hypercapnia was defined as a PaCO2 is greater than 45 mm Hg. Respiratory rate and heart rate were classified as above or below the 95th percentile based on population-based reference values [5].

HFNC device

The HFNC devices used were Optiflow™ Airvo2 (Fisher & Paykel Healthcare, Auckland, New Zealand). Optiflow™ Junior 2 nasal cannulas (Fisher & Paykel Healthcare, Auckland, New Zealand) were used in three sizes, adapted to each patient, namely: OPT314 Neonatal, OPT316 Infant, and OPT318 Pediatric.

All children included in our study received a flow rate of 2L/kg/min of a heated and humidified gas mixture with titration of FiO₂ according to the target SpO₂.

Outcome

Patients were classified according to a binary outcome as either responding to treatment or failing to respond to treatment. Treatment failure was objectively defined as the need to switch from HFNC to NIV or invasive mechanical ventilation. The decision to start HFNC and to escalate respiratory support was based on physician discretion.

Statistical analysis

Demographic characteristics are presented as numbers and percentages for categorical variables, means and standard deviations (SD) for normally distributed continuous variables, and median and quartiles (Q1, Q3) for non-normally distributed continuous variables. To compare baseline characteristics with the HFNC responder and failure groups, chi-square tests or Fisher's exact test were used for categorical variables, and t-tests or Mann-Whitney U tests were used for continuous variables.

To identify risk factors associated with HFNC failure, a binary logistic regression was carried out. Crude and adjusted odds ratios with 95% confidence intervals (CI) were reported. All variables with p < 0.25 in the univariate analysis were included in the multivariable model. A p value of 0.05 was considered statistically significant

During the study period 623 infants under the age of one year who were hospitalized in our department for lower airway respiratory distress. A total of 154 patients met the inclusion criteria for our study. Of these, 100 (64.9%) were HFNC responders and 54 (35.1%) failed HFNC treatment (Fig. 1).

Univariate analysis

The baseline characteristics of the HFNC responder and failure groups are presented in Table 1. The groups differed in terms of age at admission (p < 0.001) and weight at admission (p < 0.001). There were no differences in other baseline demographic data, including gender, gestational age, or medical history.

Table 1

Patient’s characteristics
Patient characteristics	Success group	Failure group	Total	p
Median age (days) (Q1 ; Q3)	73.5 (43.5 ; 143)	43 (28.75 ; 67)	62.5 (37 ; 120)	< 10^− 3
Median weight (grams) (Q1 ; Q3)	6500 (5500 ; 8000)	4400 (3675 ; 5000)	5500 (4575 ; 7000)	< 10^− 3
Weight < 10th percentile N (%)	1 (1%)	6 (11,11%)	5 (3,24%)	0.004
Gestational age ≥ 37 weeks N (%) < 37 weeks N (%)	87 (87%)	43 (79.6%)	130 (84.4%)	0.229
Gestational age ≥ 37 weeks N (%) < 37 weeks N (%)	13 (13%)	11 (20.4%)	24 (15.6%)
Median weight at birth (grams) (Q1 ; Q3)	3200 (2812 ; 3695)	3200 (2775 ; 3542)	3200 (2800 ; 3600)	0.482
Non-cyanotic CHD* N (%)	8 (8%)	2 (3.7%)	10 (6.5%)	0.302
History of bronchiolitis N (%)	12 (12%)	6 (11%)	18 (11.68%)	0.870
History of MV** N (%)	8 (8%)	6 (11%)	14 (9%)	0.522
Family history of atopy N (%)	27 (27%)	13 (24.1%)	40 (26%)	0.693
Median duration of symptom progression (days) (Q1 ; Q3)	2 (1 ; 4)	3 (2 ; 4)	2.5 (2 ; 4)	0.539
CHD: Congenital heart disease, MV*: Mechanical ventilation

Most of base line clinical severity symptoms were comparable between the two study groups, except for oxygen saturation, which was lower in the failure group. Additionally, the presence of respiratory pauses, apnea, or abdominal breathing, were significantly more frequent in the failure group. Patient’s clinical and blood gas monitoring parameters during the first 2 hours of HFNC treatment are presented in Table 2. For FiO2 at H0, the area under the ROC curve is 0.709 with a 95% confidence interval (0.621, 0.798). The minimum threshold is 42.5% with a sensitivity of 61.1% and a specificity of 76%. For FiO2 at H2, the area under the ROC curve is 0.711 with a 95% confidence interval (0.620, 0.801). The minimum threshold is 48.5% with a sensitivity of 43.3% and a specificity of 87.9% (Fig. 2).

Table 2

Patient’s clinical and blood gas monitoring parameters during the first 2 hours of HFNC treatment RR: respiration rate; HR: heart rate; CRP: C reactive protein; WBC: white blood cells
Patient Factors	Success group	Failure group	Total	p-value
At initiation of HFNC
RR at H0 ≥ 95% percentile N (%)	77 (77%)	37 (68.5%)	114 (74%)	0.252
Retractions
Subcostal recession N (%)	98 (98%)	53 (98.1%)	151 (98%)	0.949
Intercostal recession N (%)	91 (91%)	50 (92.6%)	141 (91.55%)	0.743
Paradoxal breathing N (%)	3 (3%)	8 (14.8%)	11 (7.14%)	0.04
Apnea N (%)	1 (1%)	10 (18.5%)	11 (7.14%)	< 10^− 3
Median SpO₂ at H0 (Q1 ; Q3)	91% (89% ; 95%)	90% (86.7% ; 92%)	90% (88% ; 94%)	0.024
Median FiO₂ at H0 (Q1 ; Q3)	38% (30% ; 40%)	50% (40% ; 55%)	40% (30% ; 50%)	< 10^− 3
Auscultation Crackles N (%) Wheezing N (%)	34 (34%)	22 (40,7%)	56 (36,4%)	0.407
Auscultation Crackles N (%) Wheezing N (%)	79 (79%)	37 (68.5%)	116 (75.2%)	0.150
HR at H0 ≥ 95% percentile N (%)	23 (23%)	13 (24.1%)	36 (23.4%)	0.881
Median of Wang Score (Q1 ; Q3)	7 (5.2 ; 9)	9.5 (8 ; 10.25)	8 (6 ; 10)	< 10^− 3
Wang score ≥ 10 N (%)	41 (41%)	42 (77.8%)	83 (53.9%)	< 10^− 3
Median pH (Q1 ; Q3)	7.42 (7.38 ; 7,54)	7.43 (7.38 ; 7,47)	7.42 (7.38 ; 7.45)	0.220
Initial hypercapnia N (%)	17 (17%)	9 (16.66%)	26 (16.88%)	0.958
Median CRP (Q1 ; Q3)	10.25 (3.5 ; 26.75)	11.6 (1.76 ; 33,3)	11 (3 ; 27.78)	0.814
High or low WBC count N (%)	33 (34%)	13 (25%)	46 (29.87%)	0.256
At H2 of HFNC
RR at H2 ≥ 95% percentile N (%)	28 (28%)	37 (68.5%)	65 (42.2%)	< 10^− 3
Median FiO₂ at H2 (Q1 ; Q3)	38% (30% ; 40%)	50% (40% ; 57.5%)	40% (30% ; 50%)	< 10^− 3
HR at H2 ≥ 95% percentile N (%)	5 (5%)	20 (37%)	25 (16.2%)	< 10^− 3
Median pH at H2 (Q1 ; Q3)	7.42 (7.37 ; 7.47)	7.45 (7.41 ; 7.51)	7.44 (7.40 ; 7.48)	0.009
Hypercapnia at H2 N (%)	4 (4%)	15 (27.8%)	19 (12.3)	< 10^− 3

The most frequent cause of HFNC treatment failure was persistent increased respiratory effort (64.8%), followed by the occurrence of severe apnea (18.5%). Two patients required mechanical ventilation due to pneumothorax (3.7%) and one patient was intubated for severe sepsis (1.9%).

Multivariate analysis

To determine early factors independently associated with HFNC failure, a binary logistic regression analysis was performed at H0 and H2 of HFNC initiation

The multivariate analysis presented in Table 3, demonstrated that at initiation of HFNC in bronchiolitis patients, a weight less than 10th percentile, an FiO2 requirement ≥ 42.5% and a Wang score ≥ 10 were predictors factors for HFNC failure, while persistence of tachycardia, tachypnea, hypercapnia and an FiO2 ≥ 48.5% after 2 hours of HFNC treatment were also predictors of failure.

Table 3

Early factors independently associated with HFNC failure in bronchiolitis.
Factor		Odds ratio	CI at 95%	p
H0	Wang score ≥ 10	1.5	(1.23 ; 1.86)	0.004
	FiO2 ≥ 42,5%	4.8	(2.18 ; 10.75)	0.013
	Hypotrophy (weight < 10th percentile)	13.2	(1.17 ; 18.61)	0.004
H2	HR ≥ 95th percentile	3.1	(1.005–10.994)	0.049
	RR ≥ 95th percentile	3.1	(1.29–7.39)	0.011
	Hypercapnia	4.7	(1.27–17.33)	0.02
	FiO₂ ≥ 48.5%	4.9	(1.96–12.02)	0.001
HR: Heart rate; RR: Respiratory rate; CI: confidence interval

Our study is one of the few prospective studies identifying predictors of HFNC failure in bronchiolitis patients. The results suggest that at the initiation of HFNC in bronchiolitis patients, a Wang score ≥ 10, an FiO2 > 42.5%, and failure to thrive could be predictive factors of HFNC failure. Additionally, tachycardia, tachypnea ≥ 95th percentile for age, hypercapnia, and an FiO2 ≥ 48.5% observed at H2 of HFNC treatment could also be predictive factors of failure.

High-flow nasal cannula (HFNC) has been confirmed to be safe and beneficial in moderate to severe bronchiolitis [1]. Its superiority to standard oxygen treatment has been proven in terms of therapy failure. The physiological benefits, such as the supply of heated and humidified air, decrease the work of breathing and the respiratory rate, improve gas exchange, and decrease the need for intubation and mechanical ventilation [3]. Despite these benefits, a significant portion of these patients required escalation of respiratory support and ultimately intubation. Literature review demonstrates a failure rate of 6–29.3% [6–8]. Our study revealed a failure rate of 35%, significantly higher than the majority of studies examining HFNC failure factors in the treatment of acute bronchiolitis. For instance, Aydin et al. reported a failure rate of 11.9% [7], and Betters et al. found an even lower failure rate of about 6% [6]. However, D'Alessandro et al. observed in a recent study published in 2021 that the failure rate of HFNC was 29.33% [8].

Several hypotheses could explain this variability in results. Firstly, the initial severity of bronchiolitis appears to be a key factor. Infants with moderate to severe bronchiolitis have an increased risk of treatment failure with HFNC due to the severity of their lung pathology [7]. In our study, we included only infants with a moderate to severe Wang score. This may explain the higher failure rate in our population compared to other studies that did not specify the severity of respiratory distress before initiating HFNC [6, 7]. Secondly, factors such as age, comorbidities, vaccination status, and the presence of concurrent infections can also influence the response to HFNC treatment [9]. Another relevant hypothesis is the variability in defining "HFNC treatment failure." In some studies, failure was defined as the need for intubation or the occurrence of cardiorespiratory arrest [6]. Additionally, considerations related to airway management, healthcare professionals' clinical experience, as well as the duration and consistency of HFNC use, may also play a crucial role in the observed failure rate [9].

Assessing HFNC treatment failure is crucial for understanding the challenges encountered in managing infants with bronchiolitis. There is no clear consensus about which patients are the best candidates for HFNC therapy and which factors can predict HFNC failure.

Risk factors for HFNC failure in bronchiolitis have been studied by Abboud et al., who concluded that patients with lower RR and higher pCO2 had a higher risk of HFNC failure [9]. Guillot et al. studied 61 patients aged 0 to 24 months with bronchiolitis and reported that higher pCO2 was the only predictive factor of HFNC failure [10]. This result was consistent with our study, which concluded that hypercapnia at the second hour of HFNC management increases the risk of failure by 4.69.

A particularly notable observation in our study was that 6 out of 7 patients who were underweight at admission progressed to HFNC treatment failure (p = 0.004). This finding is consistent with previous studies showing that underweight infants often have more severe respiratory difficulties and may be less responsive to non-invasive interventions, including HFNC [11, 12].

The use of the Wang score has proven to be an effective tool in assessing the severity of the illness. This score, based on objective clinical criteria, provides a systematic and standardized evaluation of bronchiolitis severity. By integrating this score into our analysis, we were able to stratify infants into different severity groups and identify a significant correlation with HFNC treatment failure. Infants with higher scores, indicating increased bronchiolitis severity, were more likely to not respond adequately to HFNC treatment. Several studies have utilized this score to assess acute bronchiolitis severity. Pinto et al [13]. used the Wang score in their prospective study to evaluate its relevance, comparing it to other scores. They concluded it had sufficient internal validity and good reliability.

Several studies have attempted, retrospectively, to find a correlation between the fraction of inspired oxygen and the risk of HFNC treatment failure. Betters et al. [6] conducted a retrospective study on infants with bronchiolitis placed on HFNC. Using a logistic regression model, they showed that patients who did not respond favorably to HFNC required higher FiO2 (OR: 38.3; 95% CI p = 0.002). Sunkonkit et al. [14] demonstrated that maximum FiO2 requirements > 60% increased the risk of failure by 4.23 (p = 0.021). Our study appears to have better accuracy as we sequentially studied FiO2 requirements at multiple treatment intervals.

Close monitoring of patients on HFNC is essential as these patients are at risk of deteriorating at any time, and therapeutic escalation is possible. Aydin et al.[7] concluded that elevated heart rate and respiratory rate observed during the second hour of the monitoring period could be predictive factors for HFNC failure. This result confirms the findings of another prospective pilot study of 61 cases of bronchiolitis conducted by Mayfield et al. [15], which showed that infants who responded favorably to HFNC therapy had a significant decrease in heart rate as early as the first hour and a significant decrease in respiratory rate by the third hour. Similarly, the lack of decrease in respiratory rate was considered a predictive factor for HFNC failure in the study by Abboud et al. [9]. Our study corroborates these findings, as we found that the persistence of tachypnea and/or tachycardia increases the risk of failure by 3.1.

This study has limitations, including being conducted in a single center. With no guidelines for HFNC use in bronchiolitis, the decision to initiate and/or escalate care was based on clinical discretion. A pragmatic approach was adopted, assuming that patients requiring oxygen supplementation would have an SpO2 of less than 94% in ambient air. Secondly, from a methodological standpoint, the investigating physicians were involved in the decision-making regarding HFNC therapy.

The results of this study identify patient and clinical factors that should be considered when initiating HFNC therapy for bronchiolitis and may help identify patients at greater risk for respiratory deterioration who may require close monitoring and preparedness for possible escalation to NIMV or invasive MV.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Author Contribution

S.Z. and A.M. initiated the study. S.C., F.T. and C.C. supervised the study and participated in the interpretation of results and manuscript preparation. S.Z. and M.B. conducted statistical analysis and contributed to result interpretation. All authors read and approved the final manuscript.

Dafydd C, Saunders BJ, Kotecha SJ et al (2021) Efficacy and safety of high flow nasal oxygen for children with bronchiolitis: systematic review and meta- analysis. BMJ Open Resp Res 8:e000844. 10.1136/bmjresp-2020-000844
Weichler K, El Assal O, Michael F, Pollauf L (2019) The Clinical impact of High Flow Nasal Cannula Utilization in the Emergency Department on the Endotracheal Intubation Rate for Infants and Young Children with Severe Bronchiolitis. Pediatrics 2_MeetingAbstract144. 416. https://doi.org/10.1542/peds.144.2MA5.416
Matthew J, Lipshaw AA, Vukovic P, Dean O, Semenova Y, Zhang M, Eckerle (2021) Eileen Murtagh Kurowski. High-Flow Nasal Cannula in Bronchiolitis at a Pediatric Emergency Department: Trends and Outcomes. Hosp Pediatr 11(2):119–125. https://10.1542/hpeds.2020-002774
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. https://10.1164/ajrccm/145.1.106
Bonafide CP, Brady PW, Keren R, Conwaay PH, Marsolo K, Daymont C (2013) Development of heart and respiratory rate percentile curves for hospitalized children. Pediatrics 131(4):e1150–e1157
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. https://10.1002/ppul.23626
Aydın O, Aydin EA, Birbilen AZ, Tekşam Ö (2021) Predictive factors of high-flow nasal cannula oxygen therapy failure in children with respiratory distress treated in a Pediatric Emergency Department. Turk J Pediatr 63(6):1012–1019. https://10.24953/turkjped.2021.06.009
D'Alessandro M, Vanniyasingam T, Patel A, Gupta R, Giglia L, Federici G, Wahi G (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–e235. https://10.1093/pch/pxaa087
Abboud AP, Roth JP, Skiles LC, Stolfi A, Rowin EM (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–e349. https://10.1097/PCC.0b013e31825b546f
Guillot C, Le Reun C, Behal H, Labreuche J, Recher M, Duhamel A, Leteurtre S (2018) First-line treatment using high-flow nasal cannula for children with severe bronchiolitis: Applicability and risk factors for failure. Arch Pediatr 25(3):213–218. https://10.1016/j.arcped.2018.01.003
Chevret L, Mbieleu B, Essouri S, Durand P, Chevret S, Devictor D (2005) [Bronchiolitis treated with mechanical ventilation: prognosis factors and outcome in a series of 135 children]. Arch Pediatr 12(4):385–390. https://10.1016/j.arcped.2005.01.008
Sinha IP, McBride AKS, Smith R, Fernandes RM (2015) CPAP and High-Flow Nasal Cannula Oxygen in Bronchiolitis. Chest 148(3):810–823. https://10.1378/chest.14-1589
Pinto FR, Correia-Costa L, Azevedo I (2020) Comparison of Kristjansson Respiratory Score and Wang Respiratory Score in infants with bronchiolitis in a hospital emergency department. Hong Kong Physiother J 40(2):145–153. https://10.1142/S1013702520500146
Sunkonkit K, Kungsuwan S, Seetaboot S, Reungrongrat S (2022) Factors associated with failure of using high flow nasal cannula in children. Clin Respir J 16(11):732–739. https://10.1111/crj.13533
Mayfield S, Bogossian F, O'Malley L, Schibler A (2014) High-flow nasal cannula oxygen therapy for infants with bronchiolitis: pilot study. J Paediatr Child Health 50(5):373–378. https://10.1111/jpc.12509

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Evaluation of Prognostic Factors for High-Flow Nasal Cannula Therapy Failure in Infants with Bronchiolitis: A Prospective Study

Status:

Version 1

Abstract

Objective

Methods

Results

Conclusions

Figures

Introduction

Methods

Statistical analysis

Results

Discussion

Conclusion

Declarations

Funding

Author Contribution

References

Additional Declarations

Status:

Version 1