Date: October 28th, 2014

Guest Skeptics: Dr. Matt Dawson. Director of Point of Care Ultrasound at the University of Kentucky. Co-creator of Ultrasound Podcast, Introduction to Bedside Ultrasound Volumes 1 and 2 digital textbooks, One Minute Ultrasound smartphone app, Sonocloud, and other random ventures. He has received numerous awards teaching awards.

Dr. Mike Mallin. Director of Emergency Ultrasound and the Emergency Ultrasound Fellowship at the University of Utah. He is particularly interested in echocardiography and has sat for and passed the Echo Boards. He is published in multiple journals that can be found in your trash can.

Case: 66-year-old man develops sudden onset back/flank pain. He has a history of hypertension and smokes cigarettes. You are concerned about a potential acute abdominal aneurysm. A CT abdomen has been ordered and is pending.

Questions: How reliable is a bedside emergency department ultrasound for detecting acute abdominal aneurysm?

Screen Shot 2014-11-02 at 7.22.25 AMBackground: The prevalence of triple A is 1.3-15% in emergency department adults and increases with age. It’s greatest in males over 65 with a history of smoking and hypertension, in fact the Cochrane Review actually recommends screening ultrasounds in these patients.

The patients population seen the in emergency department and critical care, the symptomatic one, the prevalence is as high as 23%.

These are patients that are the absolute definition of “can’t miss” The mortality with rupture approximately 90 percent.

There is no combination of history or physical findings that can reliably exclude acute abdominal aneurysm.

Reference: Rubano et al. Systematic Review: Emergency Department Bedside Ultrasonography for Diagnosing Suspected Abdominal Aortic Aneurysm. Acad Emerg Med 2013

This paper is part of evidence based diagnostic series in Academic Emergency Medicine. The decision editor for this is the same guy who wrote the book on Evidence Based Emergency Care, Dr. Chris Carpenter.

  • Population: Adult patients with suspected AAA (7 studies, n=655 patients
  • Intervention: ED US by emergency physicians
  • Comparison: Confirmation gold standards (Table 1 – radiologist interpretation of ED US, radiologist US, CT abdomen, angiography and laparotomy results)
  • Outcome: Diagnostic performance of ED US by emergency physicians to detect triple A

Authors’ Conclusions: “Seven high-quality studies of the operating characteristics of ED bedside US in diagnosing AAA were identified. All showed excellent diagnostic performance for emergency bedside US to detect the presence of AAA in symptomatic patients.”

checklist-cartoonQuality Checklist for Systematic Review:

  1. The diagnostic question is clinically relevant with an established criterion standard. Yes
 Comment: See Table 1: various gold standards including radiologist interpretation of ED US, radiologist US, CT abdomen, angiography and laparotomy results.
  2. The search for studies was detailed and exhaustive. Unsure
  3. The methodological quality of primary studies were assessed for common forms of diagnostic research bias. Yes
  4. The assessments of studies were reproducible. Yes
  5. There was low heterogeneity for estimates of sensitivity or specificity. No
  6. The summary diagnostic accuracy is sufficiently precise to improve upon existing clinical decision making models. Yes

Key Results: 

  • Sensitivity = 0.99 (0.95-1.00); heterogeneity (I2) = 13.2%
  • Specificity = 0.99 (0.97-0.99); heterogeneity (I2) = 46.8%
  • LR+: 10.8-infinity
  • LR-: 0.00-0.025

Screen Shot 2015-04-25 at 3.11.12 PM

They had well described methods in this study. They used the MOOSE (Meta-Analyses and Systematic Reviews of Observational Studies) reporting guidelines. They also used QUADAS (Quality Assessment of Diagnostic Accuracy Studies) tool to evaluate the quality of the studies included.

Overall prevalence of triple A in various studies: 4.8-60.6%. This could potentially influence diagnostic test performance for negative (NPV) and positive predictive value (PPV) in studies with variable prevalence of triple A.

While disease prevalence impacts PPV and NPV it does not affect positive and negative likelihood ratios.

Screen Shot 2014-11-02 at 7.15.38 AMThe search was limited to published English language manuscripts. We do like to see a more exhaustive search. There may be great papers in other languages. But if we never search those other languages we will never find them.

This can also be one of the leaks in the knowledge translation pipe. That is why the SGEM is starting to podcast in French as well as English to reach a greater audience and cut the KT window down to <1yr.

There was a moderate heterogeneity detected between studies (up to 50%), necessitating use of random-effects 
analyses; attributed to operator training and experience.

Random effect model we cannot assume that the true effect size is identical between the studies. We suspect that there are other reasons influencing the effect size besides sampling error. Therefore, when heterogeneity is high we must use the random effect model to analyze the data.

There was little commentary on inter-rater reliability in any of the included studies.

Studies were at risk of verification bias. This is when the treating physician aware of test result, influences ordering of reference standard test. This can lead to a risk of overestimating sensitivity.

The studies included were also at risk of test review bias. This is where the interpreter of reference standard result influences the interpretation of the emergency department ultrasound result.

There are many forms of bias unique to diagnostic research. One of the best papers is by Kohn et al called Understanding the direction of bias in studies of diagnostic test accuracy.

Another issue was “indeterminate scans” were coded as false positives as they triggered a need for further imaging/investigations to avoid missed symptomatic triple A’s. This would maximize specificity at the expense of sensitivity.

The authors suggest that it may be more conservative to code indeterminate scans as false negative. This would optimize sensitivity and recalculate new Sensitivity/Specificity and Likelihood Ratios which they did not do. However, with this coding system, both Sensitivity and Specificity were very high, suggesting that coding treating indeterminate scans as potential positives does not lead to increased patient harm.

Another thing is the issue of emergency department ultrasound training. There seems to be little evidence-based consensus on what constitutes adequate training for competence in emergency department ultrasound for various diagnostic entities.

All Canadian Emergency Medicine training programs now incorporate emergency department training.

American College of Graduate Medical Education requires emergency ultrasound training for all Emergency Medicine residents. Abdominal aortic aneurysm is one of those modalities that is required.

Comment on Authors’ Conclusion Compared to Our Conclusion: Overall agreement with Authors’ conclusion, with caveat of training requirements, and limitations in search strategy.

SGEM Bottom Line: Emergency department US, when applied by “trained” emergency physicians, is an excellent accurate diagnostic modality to detect triple A’s in symptomatic adult patients. When in doubt, go on to more definitive imaging.

Case Resolution: You do a bedside ultrasound of the 66 year-old man with a history of hypertension and smoking and discover a 6cm acute abdominal aneurysm and contact vascular surgery.

Clinical Application:  Adult emergency department patients with any presentation consistent with potential triple A should be scanned.

What Do I Tell My Patient:  Emergency department ultrasound is a safe, noninvasive, painless and accurate way to check to see if you have a serious life threatening condition called triple A (acute abdominal aneurysm).

Keener Kontest: Last weeks winner was Nadia Awad an Emergency Medicine Pharmacist from New Jersey. She knew the Jadad score was named after Alejandro R. Jadad Bechara, a Colombian Canadian physician who developed the scoring system while he was a research fellow at the University of Oxford.

Listen to the podcast to hear this weeks keener question. If you know the answer send an email to TheSGEM@gmail.com with “keener” in the subject line. The first person to correctly answer the question will receive a cool skeptical prize.

Upcoming Conferences:

Remember to be skeptical of anything you learn,even if you heard it on the Skeptics’ Guide to Emergency Medicine.