Introduction
Acute bronchiolitis is the first episode of respiratory distress with wheezing or decreased alveolar air entry caused by inflammation of the lower airway in infants under two years of age. It is usually preceded by a catarrhal presentation of the upper airways with rhinitis, cough, or fever1-4. Its diagnosis is essentially clinical2,5,6. Bronchiolitis is a pathology affecting the lower respiratory tract, usually of viral etiology2,7. Although most cases are self-limited and can be managed at home, bronchiolitis is the leading cause of hospital admission in infants during the winter months1,2,8. Between 1-5% of patients require hospital admission, and 5-15% need respiratory support in a Pediatric Intensive Care Unit (PICU). Acute bronchiolitis is the most frequent infectious cause of hospitalization in infants with no underlying chronic disease1,7,9,10.
Treatment options are limited as the available evidence does not support the routine use of bronchodilators, anticholinergics, inhaled or systemic corticosteroids, antiviral drugs, or antibiotics. Respiratory support remains the mainstay of treatment due to the lack of effective medications to treat this pathology2,10-13.
In recent years, high-flow oxygen (HFO) therapy with nasal cannula has been described as a valuable and safe alternative to conventional oxygen therapy to treat these patients2,8,14,15. It has been observed that its use reduces the need for invasive and non-invasive respiratory support to treat respiratory distress12,16-20. In addition, it could provide adequate comfort with fewer side effects21-24. However, as it has not been established in which patients its success may be more likely, it is not known who would be suitable candidates for its application2,10,24,25.
On this basis, the main objective of this study was to characterize the patients with bronchiolitis in whom HFO is used in our PICU. As secondary objectives, we sought to evaluate the effectiveness of HFO in these patients and the response rate and to analyze possible variables associated with a greater probability of success or failure that could serve as predictors for the management of these patients.
Methods
We conducted a cross-sectional, retrospective, observational study on infants < 24 months (including patients < 28 days) admitted to the Pediatric Intensive Care Unit of the Hospital Universitario Miguel Servet de Zaragoza for respiratory distress between January 2015 and March 2019. We included patients admitted for respiratory distress in the presence of bronchiolitis. We excluded those patients who were already receiving intensive respiratory support (invasive or non-invasive mechanical ventilation) at the time of admission to the unit.
Microsoft Excel and SPSS (Statistical Package for the Social Sciences) for Windows were used to create the database and analyze the data.
The descriptive results were expressed as the arithmetic mean, median, and standard deviation. The Kolmogorov-Smirnov test was used to analyze the distribution of quantitative variables since the sample size was > 30 patients. The variables were considered to follow a normal distribution when a p-value > 0.05 was obtained.
HFO success was established when patients did not require intensification of respiratory support and HFO failure in cases where non-invasive or invasive mechanical ventilation was needed. The validated SCORE Wood-Downes-Ferrés bronchiolitis severity scale was applied to all patients to standardize the severity assessment.
For the contrast of hypotheses between two qualitative variables, we applied the c2 test. To compare quantitative variables related to the success or failure of the HFO, we used the Student’s t-test for variables with a normal distribution or the Mann-Whitney’s U test for those variables that did not follow a normal distribution.
To compare two or more quantitative variables, we applied ANOVA if the quantitative variable followed a normal distribution or the Kruskal-Wallis test in the opposite case.
For linear correlation between quantitative variables with normal distribution, Pearson’s correlation was applied, and for correlation between qualitative variables with no normal distribution, Spearman’s correlation was used.
The limit of statistical significance accepted for the analysis was 95%. A statistically significant difference was considered when the p-value < 0.05.
For the present study, we followed the protocols established by the hospital for access to medical record data, publication of patient data, and divulgation among the scientific community, always respecting patient privacy.
Results
We included 149 patients who received HFO as the first respiratory support therapy on admission (or continued with its application if they had started such treatment in their hospital of origin). The mean age at admission was 4.58 months ± 6.41 (median 2 months), and 60.4% were male and 39.6% female. Clinical variables before the onset of HFO are shown in Table 1.
Mean | Median | Standard deviation | Min | Max | |
---|---|---|---|---|---|
pCO2 pre-HFO | 58.19 mmHg | 56.50 mmHg | 17.82 | 29 | 123 |
HR pre-HFO | 165 bpm | 165 bpm | 27 | 94 | 257 |
RR pre-HFO | 60 Bpm | 60 Bpm | 14 | 22 | 98 |
SatO2 pre-HFO | 95.85% | 97% | 5.47 | 82 | 100 |
FiO2 pre-HFO | 0.39 | 0.35 | 0.173 | 0.21 | 1 |
SCORE | 7 | 7 | 2 | 2 | 11 |
HR: heart rate (beats per minute); RR: respiratory rate (breaths per minute); pCO2: partial pressure of carbon dioxide (mmHg); SatO2: oxygen saturation; FiO2: fraction of inspired oxygen; SCORE: validated Wood-Downes-Ferrés bronchiolitis severity scale.
In 112 patients (75.2%), an adequate response to HFO was observed and did not require escalation to other respiratory support modalities. In 37 patients (24.8%), HFO failed, and escalation or change to another respiratory support modality was necessary.
No statistically significant differences were found in sex, age (in months), age group, and weight between patients who responded adequately to HFO and those in whom this measure failed. There were also no differences in terms of the history of prematurity, respiratory pathology, bronchopulmonary dysplasia, neurological pathology, or cardiac pathology between the responder and non-responder groups. Similarly, no statistically significant differences were found between the two groups in the nasopharyngeal aspirate virus results (Table 2).
HFO success >(n = 112) | HFO failure (n = 37) | p-value | |||
---|---|---|---|---|---|
Sex | |||||
Male | 72 | 64.3% | 18 | 48.6% | 0.092 |
Female | 40 | 35.7% | 19 | 51.4% | |
Age (months) | 4.40 | 4.46 | 0.88 | ||
Age (months) | |||||
< 28 days | 24 | 21.4% | 8 | 21.6 | 0.395 |
1 – 6 | 62 | 55.4% | 23 | 62.2% | |
6 – 12 | 14 | 12.5% | 2 | 5.4% | |
12 – 18 | 8 | 7.1% | 3 | 8.1% | |
18 – 24 | 4 | 3.6% | 1 | 2.7% | |
Weight (kg) | 5.29 | 5.16 | 0.58 | ||
History of prematurity | |||||
No | 78 | 70% | 24 | 64.5% | 0.588 |
Yes | 34 | 30% | 13 | 35.1% | |
History of bronchopulmonary dysplasia | |||||
No | 103 | 91.7% | 34 | 92% | 0.98 |
Yes | 9 | 8% | 3 | 8.1% | |
History of respiratory disease | |||||
No | 90 | 80.4% | 30 | 81% | 0.923 |
Yes | 22 | 19.6% | 7 | 18.9% | |
History of cardiac disease | |||||
No | 87 | 77.7% | 30 | 81% | 0.662 |
Yes | 25 | 22.3% | 7 | 19% | |
History of neurological disease | |||||
No | 98 | 87.5% | 30 | 81% | 0.267 |
Yes | 14 | 12.5% | 7 | 19% | |
NPA | |||||
Negative | 26 | 23.3% | 9 | 28.1% | 0.89 |
Positive | 86 | 76.7% | 28 | 71.8% |
HFO: high-flow oxygen therapy; NPA: nasopharyngeal aspirate.
Regarding clinical variables before initiating oxygen therapy (respiratory rate and heart rate), it was observed that the group of HFO responders showed lower heart and respiratory rates at admission than the group of non-responders. However, these differences were not statistically significant (Table 3). Regarding gasometry variables, higher pCO2 levels were found in HFO non-responders than responders (Table 3). However, these differences were not significant (p = 0.083). Given this tendency, we analyzed whether there was a pCO2 level at which HFO failure could be predicted. We found that a baseline pCO2 level ≥ 75 mmHg was a predictor of HFO therapy failure (p = 0.043). The SCORE for bronchiolitis severity was significantly higher in the group in which HFO failed than the group in which it was effective (p = 0.032). The SCORE for bronchiolitis severity could also be a predictor of HFO therapy failure (Table 3).
HFO responders | HFO non-responders | p-value* | |
---|---|---|---|
HR (bpm) | 160 | 170 | 0.521 |
RR (BPM) | 58 ± 14 | 63 ± 16 | 0.13 |
pCO2 (mmHg) | 56.59 ± 15.31 | 62.54 ± 23.04 | 0.083 |
SCORE | 6 | 7 | 0.032 |
Hours of evolution | 36 | 48 | 0.39 |
SatO2 (%) | 97 | 96 | 0.425 |
FiO2 | 0.35 | 0.35 | 0.655 |
SatO2/FiO2 | 279.3 ± 83.2 | 270.9 ± 82 | 0.597 |
FiO2: fraction of inspired oxygen; HFO: high-flow oxygen therapy; HR: heart rate (beats per minute, bpm); RR: respiratory rate (breaths per minute, bpm); pCO2: partial pressure of carbon dioxide; SatO2: oxygen saturation; SCORE: validated Wood-Downes-Ferrés bronchiolitis severity scale.
The table shows the means and standard deviations for the variables that follow a normal distribution and the medians for those with a non-normal distribution.
Mann-Whitney’s U test (nonparametric test), Student’s t test (parametric test). p- value < 0.05 was considered statistically significant.
Discussion
Acute bronchiolitis in infants is the leading cause of hospital admission in the winter months, and up to 5-15% of them require admission to a PICU1,2,7,8.
Recently, HFO with nasal cannula has been described as a valuable and safe alternative to conventional oxygen therapy for treating patients with acute respiratory distress14. Multiple studies have concluded the efficacy and adequate clinical response of patients after initiating this respiratory support system2,11,21-23. In addition, other reports support this system in an inpatient ward or pediatric emergency room due to the good results obtained and the reduced need for admission to an Intensive Care Unit8,11,26.
Due to the increased use of HFO in infants with bronchiolitis, research has been conducted to determine which demographic, clinical, or analytical characteristics of the patients could select or predict a greater probability of success when using this respiratory support. It has been described that pCO2 levels, pH, respiratory rate, and heart rate14,22,27-29 before initiating HFO therapy may be related to its success or failure.
Here, we analyzed a sample of 149 patients. Most of them were infants < 6 months of age who were admitted to initiating HFO for respiratory failure. In this population, the success rate of HFO was 75%, with 15 patients requiring invasive mechanical ventilation for nonresponse. Other studies have reported a similar response rate (between 60-90%, depending on the source)3,12,17.
When analyzing patient characteristics, no statistically significant differences were observed between the responder and non-responder groups regarding sex, age (in months), weight, history of prematurity, bronchopulmonary dysplasia, respiratory disease, neurological disease, or cardiac disease. Other studies that also analyzed these differences between responders and non-responders to HFO have reported similar results14,22,27-29.
Lower heart rate and respiratory rate were observed in patients who responded adequately to HFO compared to non-responders. However, these differences were not significant. In contrast, statistically significant differences in respiratory rate between responders and non-responders were found in other studies22,27,28 where the respiratory frequency was higher in the group in which oxygen therapy failed. Consistent with other studies14,22,27, no statistically significant differences were found in heart rate between the HFO responder and non-responder groups.
Differences in pCO2 levels obtained by measuring blood gases before initiating HFO were also observed in this study between responder and non-responder groups. The pCO2 levels were higher in the nonresponder group than in the responder group. In addition, moderate-severe hypercapnia with a pCO2 level ≥ 75 mmHg was identified as a predictor of HFO failure (p = 0.043). These results are consistent with other studies14,22,27, in which elevated pCO2 levels were also associated with HFO failure.
When analyzing the differences between responder and non-responder groups, the SCORE for bronchiolitis severity was also significantly higher in the group of non-responders (p = 0.032).
In summary, we identified moderate-severe hypercapnia with a pCO2 level ≥ 75 mmHg and a higher level of respiratory symptoms severity (according to the SCORE Wood-Downes-Ferrés bronchiolitis severity scale) were predictors of HFO failure in a PICU. The rest of the variables analyzed were not related to a higher probability of success or failure of HFO.
Based on these results, initial moderate-severe hypercapnia (pCO2 level ≥ 75 mmHg) and SCORE of bronchiolitis severity according to the Wood-Downes-Ferrés rating scale are proposed as predictors of HFO failure. The presence of these factors represents a possible risk in the initial evaluation of the patient and increases the probability of requiring intensive respiratory support with non-invasive or invasive mechanical ventilation. However, these data come from a single center. Thus, given the limited sample size, further evidence is required to extrapolate these results and conclusions.