Project Details

Sensitivity of lung ultrasound in Ugandan children with radiographic pneumonia

Nathaniel Meuser-Herr, MD

Summary

BACKGROUND: Prompt recognition and treatment of pneumonia is crucial, as the case-fatality rate in untreated children is high, often exceeding 20%. However, the complications of integrated management of pneumonia with often co-existing malaria, malnutrition and diarrheal illnesses impact the sensitivity of the diagnostic criteria. Lung ultrasound (L-US) has been shown to be comparable or even more sensitive than CXR in the diagnosis of pediatric pneumonia in multiple studies completed in developed countries. GAP: No studies of L-US in pediatric pneumonia have been conducted in low-middle income countries where children have risk factors for more severe disease. HYPOTHESIS: We hypothesize that L-US will perform with at least 80% sensitivity and that with a standardized ultrasound education course and technique, L-US operators will demonstrate almost perfect agreement. METHODS: A L-US and CXR performed on all children, between the ages of 2-months and 5-years, presenting with a diagnosis of pneumonia by WHO criteria. Operators and radiologists will be blinded to the alternate imaging findings. Interobserver agreement will be tested in comparison of L-US findings between the bedside operator and a secondary expert review of the ultrasound. RESULTS: 300 children were enrolled over a 12-month period. 4 were excluded due to incomplete acquisition of the L-US and/or CXR. 2 independent radiologists agreed on the interpretation of 293 of the 296 CXRs. Bedside L-US interpretation was 94.7% sensitive [74.0-99.9] in detecting the 19/293 primary endpoint pneumonia consolidations on CXR. Specificity was 58.0% [51.9-63.9] in ruling out consolidations. The bedside L-US and secondary expert review of the L-US clips were 70% agreeable with a kappa of 0.37. IMPACT: Lung ultrasound has the potential to become a validated diagnostic study in pediatric pneumonia in the setting of low-middle income countries. It can be taught and implemented in rural areas where CXR is not available. This would build capacity in the diagnosis, referral, and management of severe pneumonia and serve as a reliable diagnostic test to rule out disease.