Autopulse Advertisement in Critical Care Medicine

We’ve all seen folks come in via EMS with mechanical devices performing automated chest compressions.  These probably do a lovely job of freeing up paramedics from performing uninterrupted CPR, but their relationship to outcomes has been typically uncertain.

This meta-analysis and systematic review, however, reports these devices are superior to manual chest compression – with an OR of 1.6 towards increased return of spontaneous circulation.  Considering the copious evidence towards improved outcomes by minimizing interruptions during CPR, this would be an important finding, and tailors nicely with the expected advantage of mechanical compression devices.

However, this COI statement covering each of the four authors might also be in some fashion related to the positive results reported here:
“Dr. Westfall has received modest research grant support from ZOLL Medical Corporation. Mr. Krantz has received significant research grant support from ZOLL Medical Corporation. Mr. Mullin has served as a consultant for ZOLL Medical Corporation. Dr. Kaufman is an employee of ZOLL Medical Corporation.”

Unsurprisingly, these authors also demonstrate one of the overlooked evils of meta-analyses – the obfuscation of source COIs.  This JAMA article from 2011 does a lovely job describing this critical problem, and, as expected, these conflicted authors ignore the pervasive sponsorship bias present in their selected review.  Additionally, half the articles are only conference abstracts, suffering from results and methods not subject to the same level of rigorous peer review.

It really ought to be rather embarrassing for the editors of this journal to be approving such a clearly flawed vehicle – essentially blatant advertising for their $15,000 medical device – for publication.  No better, Journal Watch Emergency Medicine gives this article a bland and un-insightful thumbs-up.

“Mechanical Versus Manual Chest Compressions in Out-of-Hospital Cardiac Arrest: A Meta-Analysis”
www.ncbi.nlm.nih.gov/pubmed/23660728‎