Background: Dyspnoea is one of the most common reasons for emergency department (ED) visits, and admission due to dyspnoea is associated with a poor prognosis. Point-of-care ultrasound (PoCUS) improves ED diagnosis of dyspnoeic patients, but it remains uncertain whether serial PoCUS in this setting impacts prognosis.
The objectives were to investigate temporal changes in serial PoCUS findings in dyspnoeic ED patients and to examine if PoCUS findings were associated with short-term mortality and hospitalisation.

Methods: This is a substudy of a randomised, controlled, blinded-outcome multicentre trial in Denmark between October 9, 2019, and May 26, 2021, with a follow-up period of 30 days. In a random sample, patients aged 18 or older admitted with a primary complaint of dyspnoea and no trauma were included and allocated 1:1 to a serial ultrasound or control group. Patients in both groups received standard care within one hour of arrival, including a single cardiopulmonary PoCUS. In the serial ultrasound group, two additional cardiopulmonary PoCUS investigations were performed at two-hour intervals. The main study showed that treatment guided by serial PoCUS significantly reduced the degree of dyspnoea compared to usual care. In this secondary analysis, we examined if the PoCUS findings changed within the first 5 hours in the ED by comparing median values by visualisation of box plots. Associations between the sum of B-lines, inferior vena cava collapsibility index (IVC-CI), tricuspid annular plane systolic excursion (TAPSE), right ventricular dilatation, and ejection fraction (EF) and 30-day all-cause mortality and 30-day hospitalisation after discharge were examined with a multivariable logistic regression model with the Firth method applied because of low event rates, adjusted for age, sex, and baseline vital signs. The sample size was powered to the primary outcome in the main study (change in the degree of dyspnoea).

Results: Overall, 206 patients were included. The median age was 76 years, sex was evenly distributed, and the most common comorbidities were chronic obstructive pulmonary disease and arterial hypertension. Besides decreasing B-lines, PoCUS findings did not fluctuate during the initial ED stay. Thirty-day all-cause mortality was 9 (4.3%), and 47 patients (22.8%) were hospitalised. PoCUS findings were not associated with mortality or hospitalisation. Age > 80 years was most strongly associated with prognosis.

Discussion and conclusion: PoCUS is essential for the initial diagnostic workup. In the short term, only B-lines showed dynamic changes. However, PoCUS parameters do not provide prognostic information, but a larger sample size is necessary to draw definitive conclusions. Nevertheless, the potential use of serial focused lung ultrasound and B-line trajectories by ED physicians as a tool to tailor treatment, along with usual care, is a promising approach.