Substituting Lung Ultrasound for Chest Xray in Pediatric Pneumonia
Feasibility and Safety of Substituting Lung Ultrasound for Chest Radiography When Diagnosing Pneumonia in Children: A Randomized Controlled Trial
Background
Pneumonia is a leading cause of death in children worldwide. Physical exam and symptoms are not reliable for diagnosis, and chest xray is not available in all parts of the world. Lung ultrasound is known to be accurate for pneumonia and has more portability, less radiation, and less cost.
Questions
Is substituting a lung ultrasound for a chest xray in suspected pneumonia feasible and safe?
Can lung ultrasound reduce chest xrays needed?
Population
Urban pediatric emergency department
Convenience sample enrolled 24 hours a day, 7 days a week
Inclusion:
-age: birth to 21 yo
-suspicion of pneumonia requiring CXR for evaluation (fever, cough, tachycardia, abnormal findings on auscultation)
Exclusion:
-previously performed CXR
-hemodynamically unstable
Who did the ultrasounds?
15 pediatric emergency medicine attendings and fellows, varying levels of experience
Training consisted of 1 hour lung ultrasound training session prior to start of study
-30 minutes lecture on recognizing pathology and pitfalls
-30 minutes hands on scanning on normal models
Design
Randomized to intervention and control in blocks of variable lengths. Intention to treat analysis.
Not blinded
Primary Outcome: difference in frequency of chest xray between the two groups
Stratified by level of experience: novice ≤25 LUS, experienced >25 LUS
Delineated sub-centimeter consolidations = < 1 cm and not seen on Chest xray
Secondary Outcomes Measured:
-Unscheduled heathcare visits 1-2 weeks after ED visit
-Antibiotic use
-ED length of stay (LOS)
-Hospital admission rates
Missed pneumonia defined as diagnosed by healthcare provider on repeat ED or healthcare visit (clinical or radiographic evidence) AND with the initiation of antibiotics
Power analysis: powered to detect 15% or more absolute reduction in CXR use
Intervention
Those randomized to intervention received:
1) Lung ultrasound
2) CXR only in the following circumstances:
a) discretion of the enrolling clinician
b) clinical uncertainty of findings
c) request of referring or admitting physician
d) request of patient guardian.
Control: CXR first, then all received lung ultrasound
Treatment decisions left to treating clinician's discretion
2 View CXRs interpreted by radiologists blinded to results
Scan
Linear high frequency transducer
6 zone scanning protocol (3 per side), perpendicular planes at each spot
Midclavicular Anterior
Midclavicular Posterior
Midaxillary - from axilla to diaphragm
Positive findings = lung consolidation with air bronchograms
Negative findings = B-lines, subpleural consolidations without air bronchograms (considered likely viral etiology)
Lung Ultrasound Video Referenced in Study
Another Lung Ultrasound Example Referenced in Study
Another instructional Lung Ultrasound Video Referenced
More great links to help you understand lung ultrasound
5 Minute Sono - Air bronchograms
Results
191 enrolled and randomized, groups similar
Pneumonia in 14/103 (13.6%) investigational arm, 12/88 (13.6%) in control
Primary Findings
Reduction in cxr by 38.8% in intervention group
Number needed to scan (NNS) = 2.5 to save 1 child an xray
If xray had not been requested by others (n =29), there would have been a total of 67% reduction in cxr. (NNS = 1.5)
Stratified by experience:
Novice → 30% reduction in cxr
Experienced → 60.6% reduction in cxr
Stratified by age:
Less than or equal to 2 yo → 47.9% reduction
Greater than 2 yo → 30.9% reduction
Other Findings
No difference in missed pneumonia: Investigation group: 0%, Control 0%
No difference between groups for unscheduled heathcare visits, antibiotics use, radiographic pneumonia, ED length of stay, or hospital admission
However, in the LUS group that did not have chest xrays ordered (n= 40) there was reduction in LOS by 48 minutes
Mean lung ultrasound examination 7 minutes
Cost calculation - reduction in cxr saved approximately $9,200
Kappa between sonologist and expert doing quality assurance was 0.81
Limitations
Single center trial, Pediatric ED
Convenience sample
Not blinded to CXR when performing ultrasound
Excluded unstable patients
Could not calculate accuracy because no gold standard used between both groups (not everyone had cxr)
Unclear how many pneumonias were missed on ultrasound but caught on cxr when sonologist thought cxr was necessary. In the ultrasound first arm, there was 63/103 that got cxr. 29 of those 63 were ordered at the request of admitting physician, referring physician, or guardian. Therefore, there were 34/103 (33%) cases in which xray was ordered by the sonologist in the ultrasound first arm, but we don't know the specific reasons why.
Take Home Points
1. A lung ultrasound first approach in suspected pediatric pneumonia resulted in reduction in chest xray by 38.8%.
2. No pneumonias were missed in the lung ultrasound first arm (which included the use of chest xray when results unclear).
Our score
