Regional Pulmonary Involvement in Bronchiolitis: Insights From Lung Ultrasound

ABSTRACT

Background

Acute bronchiolitis is a leading cause of pediatric hospitalization, with severe cases necessitating ventilatory support. Lung ultrasound (LUS) is emerging as a valuable tool for assessing respiratory conditions, yet its utility in evaluating regional heterogeneity in bronchiolitis remains underexplored.

Objectives

This study aimed to assess the regional distribution of pulmonary lesions in infants with bronchiolitis using LUS and explore their association with the need for ventilatory support.

Methods

A prospective study of 160 infants with bronchiolitis was conducted at a tertiary care center. LUS was performed within the first 12 h of admission, with pulmonary regions scored based on the Brat scoring system. Patients were categorized into a favorable outcome group and a ventilatory support group, and the severity of regional lung lesions was analyzed.

Results

Median age was 65.5 days (IQR 38–118.5; range 11–314). Infants requiring ventilatory support exhibited higher regional LUS scores—particularly in lateral‐superior, lateral‐inferior, posterior‐superior, and posterior‐inferior zones (p = 0.001); posterior regions showed the highest prevalence of severe lesions. In multivariable analysis, involvement of specific zones independently predicted ventilatory support, notably right lateral‐superior (OR 4.6, 95% CI 2.12–9.86), left lateral‐superior (OR 3.7, 95% CI 1.78–7.86), left posterior‐superior (OR 2.0, 95% CI 1.23–3.51), and left posterior‐inferior (OR 2.1, 95% CI 1.20–3.71).

Conclusions

Our findings highlight the heterogeneous distribution of pulmonary involvement in bronchiolitis and underscore the potential role of LUS in severity stratification. However, the study′s single‐center design necessitates cautious interpretation, with further research needed to validate these results and expand the clinical application of LUS in bronchiolitis management.

1. Introduction

Acute bronchiolitis (AB) is a common viral infection in infants under 12 months and a leading cause of pediatric hospitalizations worldwide [1]. For severe cases, it can require ventilatory support, ranging from high‐flow nasal cannula (HFNC) therapy to mechanical ventilation, often necessitating admission to the intensive care unit (ICU) [2]. This places a significant burden on healthcare systems and highlights the need for accurate and timely assessment of disease severity.

While clinical parameters such as respiratory distress, oxygen saturation, and feeding difficulties provide essential insights into disease progression, they are often insufficient for objective, and reliable monitoring. This underscores the need for alternative tools that are simple, noninvasive, and repeatable, to help classify the severity of bronchiolitis in hospitalized infants.

Lung ultrasound (LUS) has emerged over the past decade as a valuable imaging modality for assessing pediatric respiratory diseases [3, 4, 5]. Its bedside applicability, non‐invasiveness, and real‐time results make it particularly advantageous for use in infants [6]. Numerous LUS scoring systems have been proposed [7, 8, 9, 10], varying based on whether they focus on a global or detailed view of the lungs, and whether they incorporate or not, assessments of posterior regions. However, the regional distribution of pulmonary involvement in bronchiolitis, as assessed by LUS, has been relatively underexplored.

This study aims to evaluate the regional distribution of pulmonary lesions in infants with bronchiolitis using lung ultrasound. Specifically, it seeks to determine whether the pulmonary involvement is homogeneous or heterogeneous and to explore its correlation with the need for ventilatory support.

2. Patients and Methods

2.1. Study Population and Setting

This prospective, observational, and analytical study was conducted from January to March 2024 to evaluate the regional distribution of pulmonary lesions detected via LUS in infants hospitalized for bronchiolitis. The study was approved by the local ethics committee of Fattouma Bourguiba University Hospital in Monastir, Tunisia, in accordance with the Declaration of Helsinki and principles of good clinical practice (ethical approval number: FB.011/2023).

Bronchiolitis was defined as a viral episode of respiratory distress, accompanied by coryza, cough, crepitations, and/or wheezing, occurring in infants under the age of 1 year [11]. No routine viral testing (PCR or antigen detection) was performed. The diagnosis of bronchiolitis was based on clinical criteria in line with established guidelines. We used convenience sampling of consecutive eligible admissions, enrolling all infants hospitalized with bronchiolitis between January and March 2024. No a priori sample size calculation was performed. To prevent overfitting, the exclusion criteria were based on established literature [3]. Children were excluded if they presented with immunosuppression, complex congenital heart disease, pneumonia, neuromuscular disorders, cystic fibrosis, bronchopulmonary dysplasia, a history of inhaled foreign body aspiration, infantile asthma, or unstable critical conditions requiring immediate life‐saving interventions. Parental consent was a mandatory prerequisite for inclusion, and children whose LUS could not be performed due to time constraints were also excluded. After obtaining parental consent, all participants underwent a clinical assessment, the severity of bronchiolitis was evaluated using the Wang score [12]. The indications of ventilatory support are defined according to the good practice recommendations for the management of bronchiolitis [2, 13].

2.2. Study Groups

Patients were monitored throughout their hospitalization following the standard department protocol. Clinical surveillance included regular assessment of respiratory rate, oxygen saturation, respiratory effort, and Wang scores. Based on their clinical course, patients were categorized into two groups:

• ‐Favorable outcome group: Patients with progressive improvement of respiratory symptoms who did not require ventilatory support.
• ‐Ventilatory support group: Patients who showed insufficient improvement despite standard therapeutic interventions and required ventilatory support (HFNC, noninvasive or invasive mechanical ventilation).

Primary outcome. The primary endpoint was escalation to ventilatory support (HFNC, NIV, or IMV). Infants managed on room air or simple supplemental oxygen were categorized as no ventilatory support.

2.3. Lung Ultrasound

LUS was performed using a Samsung HM70A machine equipped with a 10 MHz linear probe. The standard abdominal preset was used with a mechanical index set below 0.7, intermediate gain to ensure clear visualization of the pleural line, a single focus on the pleural line, and a depth of 3–5 cm. Each LUS examination was conducted within the first 12 h of hospitalization.

Three pediatricians, each with more than 1 year of experience in pediatric LUS, performed the examinations. The pediatricians performing LUS were independent of clinical decision‐making; treating physicians were blinded to LUS results. Because scans were performed at the bedside, examiners were not strictly blinded to contemporaneous clinical signs.

Each lung was divided into 12 regions following a modified three‐zone per hemithorax protocol inspired by the Bedside Lung Ultrasound in Emergency (BLUE) protocol described by Lichtenstein [14]. The hemithorax was divided into anterior, lateral, and posterior zones, further classified into upper and lower regions (divided by the internipple line). The anterior axillary line defined the anterior chest, while the posterior axillary line and the spine defined the posterior chest. Posterior scanning was optimized by positioning patients in lateral decubitus or rolling them onto their side (Figure 1). The regions were classified as follows: right anterior superior (RAS) and right anterior inferior (RAI), left anterior superior (LAS) and left anterior inferior (LAI), right lateral superior (RLS) and right lateral inferior (RLI), left lateral superior (LLS) and left lateral inferior (LLI), right posterior superior (RPS) and right posterior inferior (RPI), left posterior superior (LPS) and left posterior inferior (LPI).

Figure 1.

Schematic representation of the lung areas investigated by lung ultrasound. Chest regions were identified by parasternal, anterior axillary, internipple lines and spine. 1 = Right anterior superior; 2 = Right Anterior Inferior; 3 = Left Anterior Superior; 4 = Left Anterior Inferior; 5 = Left Posterior Superior; 6 = Left Posterior Inferior; 7 = Right Posterior Superior; 8 = Right Posterior Inferior; 9 = Left Lateral Superior; 10 = Left Lateral Inferior.

Lung ultrasound findings were scored using the Brat et al. scoring system [15] as follows:

• ‐Score 0: Normal lung sliding, predominantly A‐lines, and/or < 3 B‐lines per lung segment.
• ‐Score 1: ≥ 3 B‐lines per lung segment without consolidation or “white‐out.”
• ‐Score 2: Consolidated B‐lines or “white‐out” without subpleural consolidation or pleural effusion.
• ‐Score 3: Subpleural consolidation accompanied by findings of score 1 or 2.

Each lung field was scored, and a total score (0–36) was calculated. This bilateral summation aligns with the original Brat framework, which aggregates per‐zone scores across both hemithoraces to reflect overall aeration impairment; this is appropriate in bronchiolitis, where involvement is typically bilateral and heterogeneous.

Scores were recorded on a standardized form for subsequent analysis.

2.4. Statistical Analysis

Statistical analysis was performed using the SPSS software (IBM Statistical Package for the Social Science Statistics, version 25.0). Quantitative data were expressed as mean ± standard deviation (SD) or median with interquartile range (IQR), depending on the data distribution. Qualitative variables were presented as counts and percentages. A univariate analysis was conducted to compare LUS scores and their regional distribution between the two study groups. For qualitative variables, the Chi‐square test or Fisher's exact test was performed, as appropriate. For quantitative variables, the Mann‐Whitney U test was used for non‐normally distributed data. To identify region‐specific ultrasound signatures of severity, we carried a multivariable logistic regression with regional LUS scores as covariates, estimating their mutually adjusted associations with the need for ventilatory support. We did not include contemporaneous clinical severity variables (Wang score, respiratory rate, SpO₂) because they are downstream of LUS abnormalities and inform escalation decisions.

To assess interobserver reliability, the Intraclass Correlation Coefficient (ICC) was calculated using a two‐way random‐effects model with absolute agreement. LUS examinations from 30 randomly selected infants were independently analyzed by three operators, each blinded to the evaluations of the others. ICC values were interpreted as follows: > 0.90, excellent agreement; 0.76–0.90, good agreement; 0.50–0.75, moderate agreement; and < 0.50, poor agreement.

In this manuscript, we followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines to ensure comprehensive and transparent reporting of our study. The STROBE checklist was used to guide the reporting of our research methods and results.

3. Results

A total of 160 consecutive infants diagnosed with acute bronchiolitis were included in the study. The median age was 65.5 days (IQR: 38–118.5), and the median weight was 5400 g (IQR: 4350–6700 g). The median length of hospital stay was 4 days (IQR: 3–6). Based on the need for ventilatory support during hospitalization, the cohort was divided into two groups: 73 infants (45.6%) exhibited a favorable clinical course and did not require ventilatory support, while 87 infants (54.4%) required some form of ventilatory support during their hospital stay. Within the no ventilatory support group, patients included both those on room air and those receiving simple supplemental oxygen; primary analyses contrasted ventilatory support versus no ventilatory support as prespecified.

Illness day at admission was 3 days (IQR: 2–4; range 1–30) overall. It was 3 (2–5; range: 1–30) in the no ventilatory support group and 3 (2–3; range: 1–15) in the ventilatory support group, with a significant difference (p = 0.032).

There were no statistically significant differences between the two groups regarding weight or medical history. While overall age distribution did not differ significantly between groups, neonates < 28 days were more frequent among infants requiring ventilatory support (19.5% vs. 6.8%, p = 0.021). Compared with infants managed on room air or simple oxygen, those requiring ventilatory support had higher LUS scores (median 17 IQR: 15–19 vs. 11 IQR: 8–13, p = 0.001), higher clinical severity (Wang 7 IQR: 5–9 vs. 4 IQR: 3–5, p = 0.001), lower SpO₂ (93% IQR: 90–96 vs. 95% IQR: 93–97, p = 0.001), higher respiratory rate (60 IQR: 54–67 vs. 58 IQR: 50–64, p = 0.013), and a longer length of stay (5 IQR: 4–8 vs. 3 IQR: 2–5 days, p = 0.032). CRP did not differ significantly between groups (p = 0.22). (Table 1)

Table 1.

Demographic, clinical, biological and comorbidity characteristics between the two study groups.

	Group 1: room air or simple oxygen therapy	Group 2: Ventilatory support	p value
n	73	87	—
Gender (male)*	38 (52%)	51 (58.6%)	0.40
Age (days)**	73 (39–129.5)	61 (36–111)	0.13
Age < 28 days*	5 (6.85%)	17 (19.5%)	0.021
History of prematurity*	8 (11%)	14 (16.1%)	0.65
Birth weight (kg)**	3.2 (3–3.6)	3.2 (2.8–3.6)	0.74
IUGR*	10 (13.7%)	13 (14.9%)	0.95
History of mechanical ventilation*	2 (2.7%)	4 (4.6%)	0.55
Illness day at admission (d)**	3 (2–5)	3 (2–3)	0.032
Weight (kg)**	5.5 (4.3–6.8)	5.2 (4.5–6.5)	0.51
Hypotrophy*	12 (16.4%)	8 (9.2%)	0.17
Wang score**	4 (3–5)	7 (5–9)	0.001
SpO₂ (%)**	95 (93–97)	93 (90–96)	0.001
Respiratory rate**	58 (50–64)	60 (54–67)	0.013
Heart rate**	133 (124–147)	140 (126–150)	0.14
LOS (days)**	3 (2–5)	5 (4–8)	0.032
pH**	7.46 (7.38–7.50)	7.41 (7.37–7.46)	0.05
PCO2**(mmHg)	31.9 (30–36.8)	36 (30–42.3)	0.23
CRP**(mg/L)	15 (5.7–45)	11.9 (3–30.1)	0.22
LUS score**	11 (8–13)	17 (15–19)	0.001
Death or sequel*	0	0	—

Note: Bold values are statistically significant results (p < 0.05).

^*Data presented in frequency and percentage.

^**Data presented in median and interquartile range. HFNC (High flow nasal cannula); LOS (length of stay); Illness Day at admission was defined as the number of calendar days from symptom onset to hospital admission.

The distribution of ultrasound scores revealed significant differences between the two study groups. For each explored pulmonary region (anterior, lateral, and posterior), severe lesions (scores of 2 or 3) were more frequent in the upper regions compared to the corresponding lower regions. Notably, the posterior‐superior regions of both the right and left lungs exhibited the highest proportions of severe lesions, with 69.4% and 66.3% of patients affected, respectively. In the right lung, severe lesions were observed in 33.2% of patients in the anterior regions and 45% in the posterior regions. Similarly, in the left lung, 34.4% of patients demonstrated severe lesions in the anterior regions, compared to 46.9% in the posterior regions (Table 2). Interobserver agreement for lung ultrasound (LUS) scores was assessed on a subset of 30 randomly selected examinations, independently evaluated by three operators. The intraclass correlation coefficient (ICC) was 0.86 (95% CI: 0.76–0.92), indicating good to excellent agreement.

Table 2.

Regional distribution of pulmonary ultrasound scores in bronchiolitis.

		0	1	2	3
Right lung	Anterior superior	10 (6.3%)	61 (38.1%)	61 (38.1%)	28 (17.5%)
	Anterior inferior	85 (53.1%)	54 (33.8%)	17 (10.6%)	4 (2.5%)
	Lateral superior	31 (19.4%)	84 (52.5%)	33 (20.6%)	12 (7.5%)
	Lateral inferior	51 (31.9%)	86 (53.8%)	21 (13.1%)	2 (1.3%)
	Posterior superior	16 (10%)	33 (20.6%)	57 (35.6%)	54 (33.8%)
	Posterior inferior	70 (43.8%)	51 (31.9%)	31 (19.3%)	8 (5%)
Left lung	Anterior superior	13 (8.1%)	57 (35.6%)	71 (44.4%)	19 (11.9%)
	Anterior inferior	77 (48.1%)	67 (41.9%)	14 (8.7%)	2 (1.3%)
	Lateral superior	19 (11.9%)	84 (52.5%)	47 (29.3%)	10 (6.3%)
	Lateral inferior	42 (26.3%)	86 (53.7%)	30 (18.7%)	2 (1.3%)
	Posterior superior	21 (13.1%)	33 (20.6%)	55 (34.4%)	51 (31.9%)
	Posterior inferior	66 (41.3%)	56 (35%)	32 (20%)	6 (3.7%)

The regional distribution analysis of ultrasound scores revealed significant differences between the two groups in several pulmonary regions. In the right lung, patients requiring ventilatory support demonstrated significantly higher scores in the lateral‐superior, lateral‐inferior, posterior‐superior, and posterior‐inferior regions compared to infants who improved on ambient air or simple nasal cannula oxygen therapy (p = 0.001 for each region). Similar trends were observed in the left lung, where scores were also significantly higher in the lateral‐superior, lateral‐inferior, posterior‐superior, and posterior‐inferior regions of patients requiring ventilatory support. Conversely, some regions, such as the anterior‐superior and anterior‐inferior, did not show significant differences in scores between the two groups (Table 3).

Table 3.

Comparison of the regional distribution of ultrasound scores between the two groups.

Region		Total (Median and IQR)	Room air or simple nasal cannula group (Median and IQR)	Ventilatory support group (Median and IQR)	p value
Right lung	Anterior superior	2 (1–2)	2 (1–2)	2 (1–3)	0.050
	Anterior inferior	0 (0–1)	0 (0–1)	1 (1–2)	0.293
	Lateral superior	1 (1–2)	1 (0–1)	1 (1–2)	0.001
	Lateral inferior	1 (0–1)	0 (0–1)	1 (1–1)	0.001
	Posterior superior	2 (1–3)	2 (1–2)	2 (2–3)	0.001
	Posterior inferior	1 (0–1)	0 (0–1)	1 (0–2)	0.001
Left lung	Anterior superior	2 (1–2)	1 (1–2)	2 (1–2)	0.069
	Anterior inferior	1 (0–1)	0 (0–1)	1 (0–1)	0.244
	Lateral superior	1 (1–2)	1 (1–1)	1 (1–2)	0.001
	Lateral inferior	1 (0–1)	1 (0–1)	1 (1–2)	0.001
	Posterior superior	2 (1–3)	1 (1–2)	3 (2–3)	0.001
	Posterior inferior	1 (0–1)	0 (0–1)	1 (0–2)	0.001

Note: Bold values are statistically significant results (p < 0.05).

A multivariate logistic regression model was utilized to evaluate the association between specific pulmonary regions (e.g., anterior‐superior, lateral‐inferior, posterior‐superior) and the likelihood of requiring ventilatory support. By design, no contemporaneous clinical severity indices (Wang, respiratory rate, SpO₂) were included in the multivariable model. The analysis demonstrated that the lateral‐superior and posterior regions of the left lung, along with the lateral‐superior and anterior‐superior regions of the right lung, were the most significant predictors of ventilatory support requirement (Table 4).

Table 4.

Pulmonary regions predictive of ventilatory support requirement.

Lung	Region	Odds ratio	95% CI	p value
Left	Lateral superior	3.7	(1.78–7.86)	0.001
	Posterior superior	2	(1.23–3.51)	0.006
	Posterior inferior	2.1	(1.20–3.71)	0.009
Right	Anterior superior	2.2	(1.24–3.82)	0.007
	Lateral superior	4.6	(2.12–9.86)	< 0.001

4. Discussion

Our study aimed to evaluate the regional heterogeneity of pulmonary lesions in infants hospitalized for bronchiolitis using LUS. Our findings revealed that the distribution of pulmonary involvement in bronchiolitis is not uniform, with specific regions, particularly the posterior and lateral zones, being more frequently and severely affected. These results provide valuable insights into the pathophysiology of bronchiolitis and underscore the potential of LUS as a tool for stratifying disease severity and guiding management decisions. Our assessment of interobserver agreement (ICC = 0.86) indicates that LUS scoring in bronchiolitis can achieve good reproducibility among trained operators.

These region‐specific LUS findings are physiologically plausible. In supine infants with viral bronchiolitis, dependent posterior regions receive relatively less ventilation compared with perfusion, which promotes mucus pooling, atelectasis, and the appearance of coalescent B‐lines or subpleural consolidations on LUS. Conversely, anterior regions benefit from greater intercostal traction and better aeration [16, 17, 18, 19]. The left posterior predominance we observed is consistent with the compressive effect of the cardiac mass, which can reduce local compliance and ventilation on the left side in neonates and young infants [20]. Finally, relatively greater involvement in upper regions may result from airway geometry (more acute branching angles that increase resistance) and the pyramidal thoracic shape in early infancy, which narrows the superior transverse diameter and further raises resistance in upper‐lobe bronchi [21, 22, 23]. Together, these mechanisms provide a coherent physiological basis for the posterior–lateral dominance and for the independent association of specific regions with ventilatory support observed in our study.

Previous studies have consistently demonstrated that higher global LUS scores are associated with more severe bronchiolitis and worse outcomes. The LUSBRO study reported excellent accuracy of a global ultrasound score for predicting PICU admission (AUC 0.93) and correlation with hospital stay [24]. Similarly, Hernández‐Villarroel et al. [25] found that higher scores at emergency department (ED) presentation predicted the need for ventilatory support within 24 h (AUC 0.85, NPV 96%), while a recent U.S. study showed that higher LUS scores correlated with escalating support and longer length of stay [9]. In our study because all participants were already hospitalized and LUS was performed within 12 h of admission, risk of hospitalization or ventilatory support need nat emergency department (ED) triage was not assessable by design. A systematic review confirmed that LUS predicts PICU admission, support escalation, and hospitalization duration across multiple cohorts [26]. More recent works have refined this prognostic role: Camporesi et al. [4] developed a combined clinical‐ultrasound score incorporating posterior zones, improving prediction of PICU admission, and another multicenter Italian cohort identified greater involvement of superior lobes as a specific marker of severity [27]. Our findings are consistent with this growing body of evidence confirming the prognostic value of LUS in bronchiolitis. By evaluating the regional distribution of lesions across 12 lung zones, our study adds additional detail to prior work, indicating that posterior and lateral regions—and in particular upper and left posterior areas—carry prognostic weight. This suggests that not only the extent but also the distribution of abnormalities may be clinically relevant for risk stratification.

These findings implications should be interpreted cautiously within the context of this study′s limitations. The higher prevalence of severe lesions in posterior regions emphasizes the need for thorough LUS evaluations, particularly in critically ill infants. Incorporating regional assessments into routine LUS protocols could enhance the accuracy of severity stratification and inform targeted interventions. For example, infants with pronounced posterior or lateral involvement may require closer monitoring and earlier consideration of ventilatory support. This region‐specific approach could also optimize resource allocation in busy pediatric units.

Several limitations warrant consideration in our study. First, this was a single‐center study, which may limit generalizability to other settings and case‐mixes. Second, we did not collect longitudinal lung ultrasound data: we captured days from symptom onset, LUS was obtained once early after admission, so time‐aligned trajectories between ultrasound and clinical course could not be established. As a result, we could not determine temporal precedence between changes in LUS burden and clinical worsening or recovery, or evaluate the response to therapy over time; thus, causal or time‐ordered inferences should be made with caution. Third, decisions to initiate ventilatory support were made by treating clinicians according to departmental recommendations and the contemporaneous clinical picture. Although we separated roles (LUS operators were independent from management) and treating physicians were blinded to LUS results, LUS examiners were not strictly blinded to contemporaneous clinical signs at the bedside. This may introduce observer bias; we mitigated this risk through standardized acquisition/scoring and by demonstrating good interobserver agreement (ICC = 0.86). Inter‐physician variability in thresholds, patient‐level factors not captured by our models, and contextual elements (e.g., timing, workload) may have introduced non‐differential noise and residual confounding by indication, potentially biasing the observed associations between regional LUS scores and ventilatory support toward or away from the null. Fourth, we did not adjust our multivariable models for contemporaneous bedside clinical severity measures (Wang score, respiratory rate, and SpO₂) to avoid over‐adjustment for mediators on the causal pathway between LUS abnormalities and clinical escalation; nevertheless, this analytic choice may allow residual confounding, and findings should be interpreted accordingly.

5. Conclusion

Our study highlights the significant regional heterogeneity of pulmonary lesions in infants with bronchiolitis, particularly emphasizing the predominance of severe involvement in the posterior and lateral lung zones. These findings underline the potential utility of LUS in stratifying disease severity and optimizing management strategies. However, the single‐center design and reliance on clinical judgment for ventilatory support decisions necessitate cautious interpretation of our results. Further multicenter studies are warranted to validate these observations and refine the clinical application of lung ultrasound in bronchiolitis.

Author Contributions

Seyfeddine Zayani: investigation, software, writing – original draft. Farah Thabet: supervision, validation, writing – review and editing, formal analysis. Abir Daya: investigation. Olfa Betbout: conceptualization, data curation, resources. Chokri Chouchane: methodology, project administration. Slaheddine Chouchane: funding acquisition, visualization, supervision.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data available on request due to privacy/ethical restrictions.