FAST Ultrasound Examination as a Predictor of Outcomes After Resuscitative Throacotomy: A Prospective Evaluation
I read Inaba et al1 with interest and much appreciation of the authors’ contribution to the field of trauma. However, I found the study's finding contradicted the title and the language of the study. Furthermore, because this was the study of the utility of a test (a cardiac ultrasound), I wondered whether the authors chose the correct endpoint (a gold standard) for their study sensitivity–specificity analysis.
To most of us practicing trauma surgeons, we often use cardiac ultrasound to confirm a “no cardiac motion” after the patient has been arrested for a certain amount of downtime, so that we do not have to perform a resuscitative thoracotomy (RT). Inaba's study's main finding supported this practice (n = 126 with “no cardiac motion,” none survived). No one uses cardiac ultrasound to predict survival, though sometimes we use it to diagnose cardiac tamponade which leads us to perform RT. Hence, the point that 9/54 (17%) of patients with “cardiac motion” went on to survive or become organ donors seemed irrelevant, but that seemed to be the main emphasis of the study. This notion was well pointed out by the manuscript discussant, Dr David Spain. This leads to our next question.
So, if the utility of the cardiac ultrasound (the test) was to confirm “no cardiac motion” (which should be a “positive” test), the study endpoint (the gold standard) would be death. This would give the utility of this test a sensitivity of 73%, specificity of 100%, positive predictive value of 100%, negative predictive value of 17%, and accuracy of 75%. We realized Inaba et al used “cardiac motion” for his “positive test” and “survival” as an endpoint to calculate sensitivity–specificity, but why? Wasn’t that counterintuitive to the role of cardiac ultrasound?
Finally, because Inaba et al have suggested that one can use cardiac ultrasound to avoid performing unnecessary RT, will the author and the group at the University of Southern California that is known for being very aggressive in performing RT,1,2 averaging 5 to 6 RTs per month, use such tools to triage their decision and not to perform RT? Inaba et al1 pointed out that 59.4% [it should be 70% (126/180)] of their RTs could have been avoided. It would be beneficial if the University of Southern California group can perform a follow-up study that demonstrates to the trauma community that using cardiac ultrasound can help decrease an unnecessary RT.
In conclusion, I thank Inaba et al for their works. The findings and conclusion, though contradictory, still stimulate interest and intellectual discussion, which I believe is the goal of the journal publication.