Background: Point-of-care ultrasonography (US) is routinely used during thoracentesis to identify a safe fluid pocket and assess for complications such as pneumothorax. Although US has higher diagnostic accuracy than chest radiography (CXR)—meta-analysis estimates pooled sensitivity and specificity of 0.87 vs 0.46—post-procedure CXRs remain common. Most post-thoracentesis pneumothoraces are small, clinically insignificant, and rarely require intervention. US is rapid, cost-effective, radiation-free, and immediately available to the proceduralist. Professional guidelines and large cohort studies show that routine post-thoracentesis CXR provides minimal clinical benefit and rarely changes management. Choosing Wisely also recommends against routine chest radiography after uncomplicated thoracentesis in asymptomatic patients, citing low clinical yield and potential cost savings. The campaign estimates that avoiding 100 unnecessary CXRs annually could save approximately $7,000 in direct imaging costs. Broader implementation of bedside lung US has been associated with decreased imaging utilization and lower radiology costs in inpatient workflows. Data on resident-performed thoracenteses are limited. This study evaluates the diagnostic value of post-procedure lung US within a structured teaching-hospital procedural service.
Methods: We performed a retrospective analysis of thoracenteses conducted by our procedural team from January 2020 to October 2024, identified via CPT codes. Adults ≥18 years with documented post-procedure lung-sliding US and a follow-up CXR were included. Collected data included demographics, indications, comorbidities, timing of US and CXR, lung-sliding results, CXR findings, CT chest (if obtained), complications, and whether a resident performed the procedure. CT served as the gold standard when US was abnormal, but CXR was negative.
Results: Of 326 charts reviewed, 262 met inclusion criteria. Mean age was 63.1 years; 54.2% were men. The average delay between lung US and CXR interpretation was 8 hours 46 minutes. Pneumothorax occurred in 21 patients (8.0%); 76.2% were small, one required chest-tube placement, and three were already intubated. Nine pneumothoraces (42.9%) were found in supervised resident-performed procedures. Most affected patients had significant multisystem disease, including heart failure, cirrhosis, and renal disease. US sensitivity was 4.8%, specificity 99.1%, positive predictive value 33.3%, and negative predictive value 91.8%.
Conclusions: Lung US reliably identifies normal post-procedure lung sliding, and routine CXR after uncomplicated thoracentesis rarely changes management. Many pneumothoraces found on CXR were trivial and may represent atmospheric air introduced during catheter removal rather than true complications, which is a factor to consider in the resident procedures. Operational challenges during the COVID-19 pandemic, including delays in imaging interpretation and missing follow-up CXRs, likely affected documentation and data completeness. Our findings also highlight the need to standardize procedural documentation, especially reporting of lung US findings. Overall, these results support using post-procedure lung US as the primary evaluation tool after uncomplicated thoracenteses in a teaching-hospital setting. Selective CXR use for symptomatic or high-risk patients may optimize safety, efficiency, and cost-effectiveness while enhancing procedural training for residents.