|
Correction
to
Plaisance et al.,
|
| |||||||||||||||||||||||||||||||||||
Additional opportunities for error arise when, as in the study by Plaisance et al., unblinded investigators both supervise the use of the device and directly manage the resuscitation that determines its success. If the chance of survival increases, the quality of neurologic outcome becomes crucial, but scoring is reliable only when blinded investigators with proved interrater reliability directly examine patients with use of validated instruments.3 In the study by Plaisance et al., subjective outcomes were obtained from third parties (proved unreliable in previous research) by unblinded investigators, and the reproducibility of ratings was not tested.
The sophisticated, intensive resuscitation resources of the Service d'Aide Médicale Urgente are the diametric opposite of most of the world's emergency medical services. In the current study, 25 percent of one-year survivors had an arrest that was witnessed by advanced-life-support personnel (and apparently another 25 percent had an arrest that was witnessed by basic-life-support personnel). These are small, atypical subpopulations, different from those with the citizen-witnessed arrests studied in most CPR research, and should be analyzed separately in the Utstein model.
I admire the investigators' hard work, but it is well known that we researchers (like all humans) are strongly motivated to find better treatments. The less rigorous a study's methods, the greater the likelihood that it will yield positive results.4 Even when investigators are blinded, failure to implement blinding as rigorously as possible increases the odds of a positive result by 40 percent.5 There were so many opportunities for positive-outcome bias in this investigation (and previous investigations) of active compression–decompression CPR that it would be surprising statistically if some studies did not report efficacy.
Michael Callaham, M.D.
University of California San Francisco
San Francisco, CA 94143-0208
References
Michael R. Bubb, M.D.
University of Florida
Gainesville, FL 32610
Moreover, and more important, their conclusion is based on an increase in long-term survival by a factor of less than three (5 percent, vs. 2 percent with standard CPR) in a population with cardiac arrest with a remarkably high incidence of asystole as the first monitored rhythm. Five years ago, a doubling of the rate of good outcomes as a result of external pacing in a smaller but similar population was not considered a result substantial enough to alter current guidelines.1 The obvious question — whether emergency medical systems that have already achieved far better outcomes without the active compression–decompression device than the French emergency medical system have achieved with it, should nevertheless adopt this technique — remains unanswered.
Patrick Martens, M.D.
Sint Jan Hospital
8000 Brugge, Belgium
References
To the Editor: Dr. Callaham is aware of the hemodynamic data from studies in animals and in humans in support of active compression–decompression CPR. Unlike his prehospital field experience, in which use of the gauge was not emphasized and the device was not widely used before its efficacy was evaluated, we had the benefit in Paris of knowing with a high degree of certainty that active compression–decompression was being performed correctly, since it had been introduced into the Paris emergency medical system two years before the study began. Although the limitations of performing a new technique in an unblinded manner are important, issues related to proper training, use of the gauge, and widespread implementation of a technique before it is compared with the widely accepted standard are most likely more important. There are probably many opportunities for positive-outcome bias in research, but there are also opportunities for negative-outcome bias, as stated by Nolan et al.1 Finally, the survivors in the standard-CPR group, not the active compression–decompression group, had a very high percentage of arrests witnessed by advanced-life-support personnel.
As noted by Dr. Bubb, few patients in Paris receive CPR from a bystander. This is an unfortunate situation but may slowly be improving, as noted in the conclusion of our article. Nonetheless, the standard-CPR group and the active compression–decompression group were similar with regard to the number of patients who received bystander CPR, but the outcomes were clearly better in the active compression–decompression group.
In response to Dr. Martens: we did not randomize by patient number because of the logistic problem that would be involved in trying to coordinate all the rescue units at any point in time. Our low randomization errors support our approach.
In view of a meaningful improvement in outcome when this technique is performed by well-trained personnel, we continue to support, as do other investigators,2,3,4 more widespread use of this technique.
Patrick Plaisance, M.D.
Lariboisière University Hospital
75475 Paris, France
Editor's note: The following disclosure statement was omitted from the article entitled "A Comparison of Standard Cardiopulmonary Resuscitation and Active Compression–Decompression Resuscitation for Out-of-Hospital Cardiac Arrest," by Plaisance et al. "Dr. Lurie is a co-inventor of the CardioPump, the device used to perform active compression–decompression cardiopulmonary resuscitation, and is entitled to potential royalties from the sale of the device." We regret the omission.
References
| |||||||||||||||||||||||||||||||||||
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
