Emergency Department Imaging of Pediatric Patients with Blunt Thoracic Trauma: A Systematic Review, Meta-Analysis, and Practice Management Guideline

Abstract

Pediatric blunt torso trauma is a common reason for emergency department visits. Computed tomography (CT) is a highly sensitive tool for diagnosis of thoracic cavity injuries. However, the prevalence of thoracic injuries requiring intervention is lower in children compared with adults, and ionizing radiation from CT is associated with higher potential harms in this population. The role of screening chest x-ray and point-of-care ultrasound in identifying thoracic injuries requiring operative and procedural intervention has not been systematically reviewed previously. Four relevant questions related to imaging after blunt thoracic trauma in pediatric patients were developed using clearly defined Population (P), Intervention (I), Comparison (C), and appropriately selected Outcomes (O) (PICO). A systematic review and meta-analysis were conducted using the Grading of Recommendations Assessment, Development and Evaluation methodology. The multispecialty working group reached consensus on the final evidence-based recommendation. A total of 10,447 articles were screened, and data from 3 studies were included in the final meta-analysis. All the studies were retrospective in design, and the quality of the available evidence was determined to be very low. Based on the limited data, we were only able to make a recommendation on one of the 4 drafted PICO questions. In pediatric (<18 years) patients with blunt trauma presenting to the emergency department who have no or minor traumatic findings on initial screening chest imaging, we conditionally recommend that chest CT not be routinely used to identify thoracic injuries requiring procedural or operative intervention, or to identify thoracic injuries leading to mortality.

1. Introduction

Trauma and unintentional injury are the leading causes of mortality in the pediatric age group.¹ Each year, more than 10 million children in the United States are treated in the emergency department (ED) for injury-related complaints.² The thoracic cavity is injured in up to 12% of children admitted to a hospital with blunt trauma.³ The most common thoracic injuries, in order of prevalence, are pulmonary contusion, rib fracture, pneumothorax, and hemothorax, accounting for approximately 53%, 50%, 37%, and 13% of reported injuries, respectively.4, 5, 6, 7 The incidence of thoracic aortic injuries in pediatric blunt thoracic trauma is extremely uncommon (about 0.1%).⁸

In many cases, the management of pediatric thoracic trauma injuries is conservative and symptom-based, with lower mortality and morbidity for these injuries than seen in the adult population.^9,10 Occult pneumothoraces and hemothoraces generally are managed with observation alone,¹¹ and the Western Trauma Association suggests observation for pneumothoraces of <20% of chest volume on chest x-ray or 35 mm from the chest wall on CT scan in the stable patient.¹² Rib fractures do increase mortality,¹³ likely due to associated injuries, but typically isolated closed rib fractures are treated in the outpatient setting with supportive care, such as analgesia and incentive spirometry. The complication rate from pulmonary contusions is low, with one study reporting no intubations in children with CT-demonstrated pulmonary contusions over a 2-y period.¹⁴ The hazard ratio (HR) of pneumonia following minor chest trauma is 1.23 in children within 1 y, far less than the HR of 1.77 from an upper respiratory tract illness.¹⁵

Computed tomography (CT) is often regarded as the gold standard in diagnosing thoracic injuries in trauma. However, it also exposes children to potentially harmful radiation, requiring consideration of risks and benefits of advanced imaging. Thoracic CT typically delivers a radiation dose between 2 and 8 mSv, compared with a dose of 0.02 mSv for a plain film of the chest.¹⁶ This translates into a solid malignancy lifetime risk between 6.1 per 10,000 CTs for preteen boys to 30.5 per 10,000 CTs for girls aged 5 to 9 years.¹⁷ Another study estimates a malignancy rate of 1 cancer for every 37 actionable injuries identified.¹⁸ The relatively low rate of intervention required in pediatric blunt thoracic trauma coupled with the radiation risk of CT creates a dilemma in the management of children presenting to the ED with blunt chest trauma. Sufficient advanced imaging must be done to identify fatal or actionable injuries without exposing children with minor or no injuries to unnecessary radiation. In addition to low radiation protocols, choosing the right study for the right indication has been the focus of a number of organizations, including the American Academy of Pediatrics Choosing Wisely Campaign.¹⁹

The objective of our study was to use the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology to perform a systematic review and meta-analysis to generate evidence-based recommendations regarding the utility of different imaging modalities to evaluate for injuries leading to death or procedural intervention in pediatric patients presenting to the ED following blunt chest trauma.

2. Objectives

The guideline was developed by assembling a working group consisting of pediatric emergency physicians, pediatric and trauma/acute care surgeons, and pediatric radiologists. Outcomes were selected in accordance with GRADE methodology.²⁰ Four relevant questions related to imaging after thoracic trauma in pediatric patients were developed using Population (P), Intervention (I), Comparison (C), and Outcome (O) (PICO) format.

The following PICO questions were formulated prior to the literature search:

PICO 1: In pediatric (<18 years) patients with blunt trauma presenting to the ED with a minor injury burden (P), should screening chest imaging (I) versus no screening chest imaging (C) be used to identify thoracic injuries requiring procedural or operative intervention (O1) and to identify thoracic injuries leading to mortality (O2)?

PICO 2: In pediatric patients (<18 years) with blunt trauma (P) presenting to the ED, should initial evaluation with chest x-ray (I) versus point-of-care ultrasound performed by bedside clinicians (C) be used to identify thoracic injuries requiring procedural or operative intervention (O1) and to identify thoracic injuries leading to mortality (O2)?

PICO 3: In pediatric patients (<18 years) with blunt trauma presenting to the ED, who have no or minor traumatic findings on initial screening chest imaging (P), should CT chest (I) versus no chest CT (C) be used to identify thoracic injuries requiring procedural or operative intervention (O1) and to identify thoracic injuries leading to mortality (O2)?

PICO 4: In intubated pediatric patients (<18 years) with blunt trauma who have no acute traumatic findings on initial screening chest imaging (P), should CT chest (I) versus no chest CT (C) be used to identify thoracic injuries requiring procedural or operative intervention (O1) and to identify thoracic injuries leading to mortality (O2) and minimize delay in procedural or operative intervention for injury (O3)?

3. Methods

3.1. Identification of References

A health sciences librarian (CG) assisted in developing a comprehensive search strategy to identify relevant studies in PubMed, Scopus, and CINAHL. The search strategy incorporated both controlled vocabulary (eg, MeSH terms and other database-specific subject headings) and relevant keywords to capture studies addressing 3 primary concepts: (1) blunt chest or thoracic trauma, (2) diagnostic imaging—including sonography and tomography, and (3) physical examination or diagnostic assessment. Boolean operators and database-specific syntax were used to combine terms across the 3 concepts.

The initial search was conducted on May 5, 2022, in all 3 databases, covering records from database inception through that date, with no date restrictions applied. To ensure inclusion of newly published studies, the search was rerun on October 25, 2024.

Eligible studies included published observational or randomized controlled trials involving patients who received both the intervention and comparator. Studies were excluded if they were case series, case reports, commentaries, animal studies, operative technique articles, or not published in English. Although review articles were excluded from formal inclusion, their reference lists were screened to identify any additional relevant primary studies not captured by the database searches.

All search results were imported into EndNote (Clarivate) to screen for duplication, then uploaded to Covidence (www.covidence.org), where any remaining duplicate citations were identified and removed. Titles and abstracts were screened independently by 2 reviewers, with disagreements resolved by a third reviewer. Full-text articles selected for inclusion were also reviewed independently by 2 reviewers, with conflicts adjudicated by a third reviewer. The screening and selection process adhered to PRISMA guidelines. The review protocol was registered with PROSPERO (CRD42022326867).

3.2. Data Extraction and Management

Data were extracted in duplicate using Excel (Microsoft, Redmond, WA) and meta-analysis was conducted using RevMan Online (revman.cochrane.org) and MetaAnalysis Online (MetaAnalysisOnline.com).²¹ As one of the articles (Azari) included 2 different patient periods, we included those separately within the meta-analysis. Hazard ratios (and 95% CIs) for dichotomous outcomes and differences in means (and 95% CIs) were calculated for the intervention groups and comparison groups using OpenEpi.²² Statistical significance was considered P<.05. Heterogeneity was calculated and quantified with I². A low degree of heterogeneity was considered I² values less than 50%, those with moderate heterogeneity had I² values between 50% and 74%, and high heterogeneity was considered I² values greater than 75%.²³

3.3. Methodological Quality Assessment

This review adhered to the validated GRADE methodology.^24,25 The quality of the evidence was assessed with GRADEPro (www.gradepro.org). Risks of bias, inconsistency, indirectness, imprecision, and publication bias were all considered. Based on the results of the meta-analyses and the evidence quality, all members of the working group then voted on recommendations for each PICO question, taking into consideration the quality of the evidence, the relationship of benefits and harms, patient values and preferences, cost, and resource utilization.

4. Results

A total of 10,447 articles met the initial search definitions. After abstract and title screening, 534 manuscripts were identified for full-text review. Ultimately, data from 3 studies were included in the final meta-analysis (Fig 1).

Figure 1.

Flow diagram for study selection for analysis (PICO 3).

PICO 1

We did not find any studies that directly compared chest imaging versus no chest imaging in our population of interest. No further analysis was performed.

PICO 2

We found 2 studies that examined the use of ultrasound in thoracic trauma. One of them focused on identification of occult pneumothoraces, but not specifically on thoracic injuries leading to mortality or requiring procedural/operative intervention.²⁶ The other was a retrospective study of lung ultrasound examinations performed by ultrasound technicians and interpreted by a radiology resident or attending physician.²⁷ Neither of them directly addressed our PICO questions and was further analyzed.

PICO 3: Quantitative Synthesis

We found 3 studies including 4 patient data sets with available data for meta-analysis.28, 29, 30 All were retrospective studies. Year of publication ranged from 2016 to 2020. Total number of subjects included was 2859. Tests for heterogeneity for both were high, suggesting considerable variability within the data sets. In assessing missed injury, pooled results showed that CT scan reduced missed injury by a HR of 30.0 (95% CI 3.49 to 258.40, P<.001) (Fig 2A). In addition, CT scan reduced missed injury requiring procedural or operative intervention by a HR of 6.23 (95% CI 0.54 to 72.00, P=.14) (Fig 2B). Individual study risk of bias is shown in Table S1. A total of 12 patients had injuries identified by CT requiring procedural or operative intervention (Table). Four of them had a negative chest x-ray. Two cases of aortic injuries were identified by chest CT, both in the same case series.²⁹ These 2 patients had abnormalities found on chest x-ray (rib fractures, pneumothorax), although the aortic injury was identified only on CT. There was no mortality that occurred as a result of missed injury on screening chest imaging.

Figure 2.

A, Forest plot of the meta-analysis results for PICO question 3 (missed injury). B, Forest plot of the meta-analysis results for PICO question 3 (injuries requiring procedural or operative intervention).

Table.

List of patients whose chest computed tomography (CT) identified injuries requiring procedural or operative intervention (each row represents a single patient).

	Study	Age	Chest x-ray findings	Additional chest CT findings	Intervention as a result of additional chest CT findings	Death from chest injury?
1	Golden30	N/A	Negative	Small pneumothorax	Chest tube placed before exploratory laparotomy and positive pressure ventilation	No
2	Stephens29	15	Negative	Pneumothorax	Chest tube	No
3	Stephens	2	Negative	Pneumothorax	Chest tube placed while in operating room	No
4	Stephens	15	Negative	Hemothorax, rib fractures	Chest tube placed before anesthesia/ positive pressure ventilation	No
5	Stephens	17	Clavicle fracture	Sternal fracture, pneumothorax	Chest tube	No
6	Stephens	7	Pneumomediastinum	Pneumothorax, rib fractures	Chest tubes, negative bronchoscopy and esophagram	No
7	Stephens	18	Rib fracture	Pneumothorax	Chest tube	No
8	Stephens	15	Rib fractures	Hemopneumothorax	Chest tube	No
9	Stephens	16	Sternoclavicular joint dislocation	Pneumomediastinum, compression subclavian vein	Open reduction internal fixation of clavicle, negative bronchoscopy	No
10	Stephens	15	Pneumothorax, Pneumomediastinum	Rib fractures, clavicle fracture	Open reduction internal fixation of ribs for flail chest, repair of lung laceration, negative echocardiogram, negative bronchoscopy, chest tubes	No
11	Stephens	1	Rib Fractures, Pneumothorax	Dissecting aortic aneurysm, hemomediastinum	Beta-blockers for aortic dissection, chest tube	No
12	Stephens	13	Pneumothorax	Thoracic aortic injury, mediastinal hematoma, rib fracture	Stent graft aorta, chest tubes	No

PICO 3: Grading the Evidence

Certainty of outcome effect was assessed using the GRADE framework. The risk of bias was serious, as most studies included were retrospective. The risk of inconsistency was serious, given moderate heterogeneity. The risk of indirectness was not serious due to compatible outcomes across studies. The risk of imprecision was serious due to wide CI. The overall certainty of evidence was determined to be very low (Fig 3).

Figure 3.

GRADE evaluation of evidence for PICO 3.

PICO 3: Recommendation

In pediatric patients (<18 years) with blunt trauma presenting to the ED who have no or minor traumatic findings on initial screening chest imaging, we conditionally recommend against routinely using chest CT to identify thoracic injuries requiring procedural or operative intervention, or to identify thoracic injuries leading to mortality.

PICO 4

We did not find any studies that examined our outcomes of interest among intubated patients as a separate group. No further analysis was performed.

5. Discussion

In this systematic review and meta-analysis using GRADE methodology, we found that chest CT is much better than chest x-ray in identifying injuries in the thoracic cavity, yielding an HR of 30.0. Nevertheless, only a small number of these “missed” injuries (12 out of 2859, or 0.4%) led to procedural or operative interventions, and none of them resulted in mortality. The prevalence of aortic injury was 2/2859, or less than 0.1%, in our included patients. After carefully weighing the potential harm from radiation exposure and potential benefits from detection of significant injuries, the working group decided to conditionally recommend against routine chest CT imaging in nonintubated pediatric blunt trauma patients presenting to the ED if screening chest imaging shows no or only minor traumatic findings.

Studies have shown that in pediatric blunt trauma patients, a selective CT imaging strategy is associated with similar rates of mortality, intensive care unit length of stay, and complications compared to a liberal imaging strategy.^31,32 In the case of thoracic injuries, this is likely driven by CT findings that mostly do not lead to escalation of care or acute procedural intervention^3,18,33, 34, 35, 36, 37 and a very low incidence of thoracic vascular injuries.³⁸ Yet overutilization of CT often happens, even in low acuity patients, and considerably more at non-pediatric trauma centers.^39,40 We hope that our findings and recommendation would help promote thoughtful and responsible thoracic imaging in stable pediatric blunt trauma patients. This can be used in conjunction with existing guidelines to risk stratify patients for selective CT ordering.^41,42

Point-of-care ultrasound (POCUS) is frequently used to evaluate for pneumothorax, hemothorax, pericardial effusion, and tamponade in pediatric blunt torso trauma patients as part of the E-FAST examination protocol.⁴³ In a recently published Cochrane review, POCUS was found to be 96% sensitive and 99% specific for detecting chest injuries in adult patients.⁴⁴ However, our group was unable to find any studies comparing the test characteristics of POCUS to chest x-ray in identifying thoracic injuries requiring procedural or operative intervention or leading to mortality in pediatric blunt trauma patients. Further research is needed in this area to elaborate the utility of POCUS in this population.

We were unable to find any studies directly comparing screening versus no screening chest imaging in pediatric blunt trauma patients presenting to the ED with a minor injury burden. This is likely due to the low prevalence of significant injuries and perceived low radiation risk from a single screening chest x-ray in this group of patients. We were also unable to find any studies examining the role of chest CT in identifying thoracic injuries leading to mortality or requiring procedural or operative intervention in intubated pediatric blunt trauma patients. Possible reasons may include low prevalence of intubated pediatric trauma patients and rarity of isolated thoracic injuries in intubated patients. In both cases, potential large sample size might have been a significant barrier for aspiring researchers.

6. Limitations

Many of the included studies were retrospective in design. Potential limitations include selection bias, possible missing/incomplete data or undetected injures, and lack of control for confounders. The total number of patients with outcomes of interest (death, procedural or operative intervention) was also low. Hence, our findings would need to be confirmed by prospective studies with a large sample size. The majority of the studies were conducted in academic, Level 1 pediatric trauma referral center settings, which may affect their applicability in community hospitals and lower-resourced settings. Study heterogeneity was moderate in most of the reviewed topics, and the number of studies included was small for PICO 3. Looking at the forest plots for PICO 3, one of the included studies (Azari post 2020) seemed to show a different HR trend compared with the rest of the group. This might be due to a spurious effect of small magnitude, Hawthorne effect from the quasi-experimental study design, advancement in technology (in this case resolution of plain x-ray) over time, or change in practice in management of certain thoracic injuries over time. Our group did not include a patient representative, so patient values and preferences were not directly taken into consideration in our recommendations.

7. Conclusion

In this systematic review and meta-analysis using GRADE methodology, we conditionally recommend against routine chest CT imaging in pediatric patients presenting with blunt torso trauma to the ED, who have no or minor traumatic findings on initial screening chest imaging based on limited evidence.

Funding and Support

By JACEP Open policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist.