Ultrasound lowers diagnostic imaging referral in suspected PE, but has high failure rate

Key takeaways:

• Use of an ultrasound protocol in patients with suspected PE lowered referral to diagnostic imaging by almost 50%.
• The failure rate found with this tool was determined to be “too high.”

Referral to diagnostic imaging decreased with use of an ultrasound protocol in patients with suspected pulmonary embolism, according to a presentation at the European Respiratory Society International Congress.

However, the failure rate associated with ultrasound use was deemed unacceptable, according to researchers.

“In a recent publication by Charlotte Dronkers, MD, PhD, and colleagues, it was described that the prevalence of pulmonary embolism in patients referred for diagnostic imaging is declining on a European scale, and the authors called for additional approaches of improving selection of patients with suspected pulmonary embolism for diagnostic imaging,” Casper Falster, MD, PhD fellow in the Odense Respiratory Research Unit in Denmark, said during the presentation. “We thought to ourselves could ultrasound be an aid in this respect?”

Following promising results from a meta-analysis and small, descriptive study of 75 patients referred for diagnostic imaging, Falster and colleagues proceeded to conduct a multicenter, randomized controlled trial comprised of 150 adults with suspected PE to determine if ultrasound can reduce the use of diagnostic imaging referral in this patient population. All adults had either a Wells score higher than 6 or, heightened age-adjusted D-dimer if the Wells score was between 0 to 6.

A focused lung, cardiac and deep venous ultrasound was performed in 73 patients, whereas 77 patients went straight to diagnostic imaging and acted as the control group.

Patients who underwent an ultrasound had the possibility for one of three outcomes based on a bespoke protocol developed with findings from the previous studies conducted by Falster and colleagues. A diagnosis of PE could be dismissed (no signs of PE, low clinical suspicion or differential diagnosis), confirmed (visible venous thrombus, two or more subpleural infarctions, McConnell’s sign or D-sign) or deemed unclear based on the outlined criteria, resulting in referral to diagnostic imaging. Notably, those with confirmed PE requiring hospitalization also received a referral to imaging.

“We decided to go ahead with the same protocol but we made a slight adjustment to increase the sensitivity,” Falster said. “If you would like to use ultrasound to dismiss pulmonary embolism, you could not have pulmonary embolism as the equally or most likely diagnosis when you calculated the Wells score. So if you had a high clinical suspicion, ultrasound could not rule out pulmonary embolism and you had to be referred to diagnostic imaging.”

Researchers reported confirmed PE in 26 patients from the control group.

Among those who underwent an ultrasound, 11 had a confirmed PE diagnosis, including eight patients who had to be hospitalized. Diagnostic imaging confirmed the diagnosis in all eight patients, Falster said. Ultrasound was unable to make a clear judgement of confirmation or dismissal of PE in 32 patients.

Out of the 32 patients referred to diagnostic imaging, researchers confirmed PE in 12 of them. This result demonstrates a 45.2% (95% CI, 33.4%-56.6%) decrease in referral to diagnostic imaging, according to Falster.

Further, ultrasound classified 30 patients (41%) with a dismissed diagnosis of PE; however, when assessed again at 3 months, two of these patients had PE for a failure rate of 6.6% (95% CI, 1.8%-21%).

“The International Society of Thrombosis and Hemostasis has this formula for calculating an acceptable failure rate, and in our instance, the acceptable failure rate would be 2%,” Falster said during the presentation. “So naturally this is too high, but I think it’s important to consider that this study was not powered to finally assess this, and as you can appreciate, the 2% are included in the confidence intervals.”

References:

• Dronkers CEA, et al. J Thromb Haemost. 2017;doi:10.1111/jth.13654.
• Falster C, et al. Thorax. 2021;doi:10.1136/thoraxjnl-2021-216838.
• Falster C, et al. Ultrasound Int Open. 2022;doi:10.1055/a-1971-7454.