E-FAST for Pneumothorax – Not So FAST?
March 7, 2022
Written by Megan Hilbert
This study suggested ultrasonographic evaluation of pneumothorax was not as sensitive as previously thought, but there are some issues we need to discuss.
Why does this matter?
Previous review articles and meta-analyses have shown that sensitivity of ultrasound supersedes that of supine chest X-ray in the trauma patient. This study turns all of this on its head…or does it?
Not so fast with your “Not so FAST”
This was a retrospective study completed at a Level 1 Trauma Center where a registered sonographer completed a CUST (complete ultrasonography of trauma) exam on trauma patients. They found that supine chest x-ray was more sensitive (43%) at detecting pneumothorax as compared to ultrasound (35%), which was the opposite of their proposed hypothesis.
These results give me pause, and by digging further into the paper I have a few concerns. While these exams were completed on trauma patients, only those with pneumothorax (as diagnosed on CT) were included in the analysis. Another concern is that the authors themselves noted, “the pleural views were not universally performed with the same rigor as in a prospective study designed specifically to detect pneumothorax.” And while I agree that a research focus can increase the rigor of investigation, I would posit that ultrasonographic evaluation of the unstable trauma patient necessitates a rigorous investigation at baseline. We are asking a very focused clinical question – Can I identify why my patient is unstable? If you aren’t intentional in your evaluation sonographically, then it is better to not pick up the probe at all.
These authors should be applauded for publishing a “negative” study (where the results actually contradict their hypothesis). There are, however, many methodological issues that call into question the validity and generalizability of this papers’ claims (i.e. retrospective nature of the study, CUST being considered a false negative if a chest tube was placed within 8 hours of presentation). Perhaps most importantly, these studies were completed by sonographers who traditionally are not taught lung ultrasound and which does not mimic real life scenarios (completion of POCUS by Emergency Physician or Trauma Surgeon). I encourage readers to ask – Was this the best study design to determine diagnostic accuracy (or lack thereof)?
Conflict of Interest: While I am the first to admit that I subscribe to the ultrasound way of life, it is a tool that I use to help guide my clinical decision making; not one to replace it. I am open to hearing any and all criticism of its use and identifying its inherent limitations in order to provide my patients with the best possible care.
Not so FAST – Chest ultrasound underdiagnoses traumatic pneumothorax. J Trauma Acute Care Surg. 2022 Jan 1;92(1):44-48. doi: 10.1097/TA.0000000000003429.
For additional review of methodological issues, check out Dr. Dan Kim’s twitter review.
Editor’s Note: Again, 100% of included patients had pneumothorax on CT. Could this have introduced bias that confounded the accuracy of the CUST exam? For example, could these patients have been more severely ill, necessitating a quicker, less thorough ultrasound exam? Patients without CUST were excluded. Were there systematic differences in patients who were excluded? We don’t know. A better study design to determine diagnostic accuracy would have been to use the test (E-FAST in this case) as a screening tool prospectively in undifferentiated patients with torso trauma and compare US with subsequent CT. Authors noted they followed STROBE guidelines for observational studies, but this was a diagnostic accuracy study; STARD guidelines would have been more appropriate. Bottom line: Use caution in applying this study to practice. ~Clay Smith
Reviewed by Jeremy Boyd and Clay Smith