Lung ultrasound boosts accuracy of pneumonia diagnosis in ED

Lung ultrasound proves to be a useful tool for screening community-acquired pneumonia (CAP) in the emergency department (ED), with a recent study showing that its use brings down diagnostic uncertainty from 73 percent to 14 percent.

“We also observed that lung ultrasound findings resulted in changes in antibiotic prescriptions in 30 percent of the patients,” according to a team of researchers from Centre Hospitalier Universitaire de Nantes in France.

“In our study, we applied the lung ultrasound criteria for pneumonia … [which involved] direct visualization of a condensation or signs of a localized pulmonary oedema (B lines >3 by spot),” they pointed out. [Chest 2009;136:1014-1020; Intensive Care Med 2012;38:577-591]

A total of 148 adult patients (average age 72 years, 47.3 percent female) with a clinical suspicion of CAP underwent lung ultrasound. A clinical CAP probability according to four-level Likert scale (definite, probable, possible, and excluded) was established on two separate occasions: once, after routine diagnostic procedure (clinical, radiological and laboratory tests); second, based on the ultrasound result. An adjudication committee composed of three independent experts established the final CAP probability at hospital discharge.

Routine diagnosis procedure involved laboratory tests in 135 patients and chest X-ray (CXR) in 138 patients. The resulting CAP probabilities were definite in 34 (23 percent) patients, probable in 52 (35 percent), possible in 56 (38 percent), and excluded in six (4 percent). Antibiotic treatment was prescribed for 106 patients (72 percent). [BMJ Open 2021;11:e046849]

Lung ultrasound altered CAP probability classification in 106 patients (72 percent). CAP was classified as definite in 97 patients (66 percent), probable in 13 (9 percent), possible in eight (5 percent), and excluded in 30 (20 percent). Other diagnoses were chronic obstructive pulmonary disease exacerbation (n=11, 7 percent), acute heart failure (n=11, 7 percent), other or no diagnosis (n=14, 9 percent).

Results of the lung ultrasound also led to the proposal of changes in antibiotic prescription in 47 patients (32 percent), with 21 antibiotic treatments initiated whereas 26 were discontinued.

Meanwhile, adjudication committee CAP probability was definite in 81 patients (55 percent), probable in 16 (11 percent), possible in 12 (8 percent), and excluded in 39 (26 percent). Using these classifications as reference, 39 out of 148 routine diagnosis procedure CAP probability classifications were correct, while 109 lung ultrasound CAP probability were correct (27 percent vs 77 percent), indicating greater concordance with the ultrasound classifications (p<0.001). The corresponding Cohen’s kappa coefficients between routine diagnosis procedure and lung ultrasound, according to adjudication committee, were 0.07 (95 percent confidence interval [CI], 0.04–0.11) and 0.61 (95 percent CI, 0.55–0.66).

“Indeed, lung ultrasound has been recognized for several years to have a very good diagnostic performance in CAP with a positive likelihood ratio (LR) of 16.8 and a negative LR of 0.07 (meta-analysis, including 10 studies dating from 1996 to 2013),” the researchers noted. “Our results are in line with these previous studies, which show very good LRs with a much higher diagnostic certainty of CAP with lung ultrasound.” [Respir Res 2014;15:50]

Despite the presence of several limitations, including the interpretation of CXR results by the patient’s treating emergency physician and not by a radiologist and the diagnostic testing being conducted at the discretion of the treating physician, the study clearly demonstrates that “adding lung ultrasound to the routine diagnosis procedure could improve CAP diagnosis accuracy and could help to reduce diagnosis uncertainty and unnecessary antibiotic prescriptions,” according to the researchers.