Correspondence

Transcutaneous Pacing in Patients with Asystolic Cardiac Arrest

N Engl J Med 1993; 329:1277October 21, 1993

Article

To the Editor:

Cummins et al. (May 13 issue)1 are to be complimented on their community-based interventional study to assess the efficacy of out-of-hospital transcutaneous pacing in asystole. They demonstrated a twofold increase in survival in the intervention group but overall judged the treatment to be ineffective, because their observed odds ratio of 2.05 “did not achieve statistical significance.” I am reluctant to accept their interpretation.

Only 40 percent of the patients in the intervention group were treated. The authors list the reasons for this low figure, among which were a staggered training schedule and the steady accumulation of negative outcomes, which caused the paramedics to omit treatment. These shortcomings of design or sources of bias make the use of the term “intention-to-treat analysis” inappropriate. I submit that under these circumstances it is incumbent on the authors to test the hypothesis that failure to deliver the treatment may have accounted for what they interpret as treatment failure. This may be done by analyzing their data according to “treatment delivered.” The data in the published article are insufficient to allow readers to do so themselves.

The authors do not formally address the low power of the study, as surely they should have in a study that they interpret as negative2. Formal calculation of power allows readers to determine to what extent the presumed negative findings may be due to chance alone. Readers can then decide for themselves whether to accept the authors' conclusions.

Carmel Armon, M.D.
Loma Linda University Medical School and Medical Center, Loma Linda, CA 92354

2 References

1
Cummins RO, Graves JR, Larsen MP, et al. Out-of-hospital transcutaneous pacing by emergency medical technicians in patients with asystolic cardiac arrest. N Engl J Med 1993;328:1377-1382
Full Text | Web of Science | Medline

2
Pocock SJ. Clinical trials: a practical approach. Chichester, England: John Wiley, 1983.

Author/Editor Response

The authors reply:

To the Editor: We respect Dr. Armon's attempt to discover some positive themes in our study of transcutaneous pacing for out-of-hospital asystolic cardiac arrest. No amount of statistical analysis, however, can change our observations that patients in asystole did not respond to transcutaneous pacing when it was delivered before hospital admission. In our study we observed a 4 percent survival rate in the study group and a 2 percent survival rate in the controls (P not significant). This tempts us to emphasize the positive: one group had twice as high a survival rate as the other. In absolute terms, however, we must still confront the fact that 96 percent of the people in one group died, and 98 percent in the other.

We analyzed our data using the intention-to-treat method. Statistical propriety requires such an analysis. This analytic approach does not cause bias in the allocation of treatment, as Dr. Armon implies, but instead reduces the possibility of bias. We do present data in Table 2 of our article that permit an analysis of the patients as actually treated. Our results are the same: 1.8 percent (2 of 112 patients) survived among the patients who received pacing, and 1.9 percent (5 of 259 patients) survived in the control group (again, P not significant). Pacing produced no improvement in survival for people with out-of-hospital asystolic cardiac arrest.

We agree that the size of our study conferred low power, with only an 80 percent chance of detecting a difference as large as 100 percent in survival between the two groups (5 percent alpha). Although a larger sample would be desirable, at some point common sense and practicality must prevail. The core question our study addressed was whether transcutaneous pacing by the first emergency personnel to respond, with large expenditures for equipment and training, should be implemented nationwide. Our answer is negative. We could find no evidence that out-of-hospital transcutaneous pacing for the asystolic heart is effective.

Richard O. Cummins, M.D.
University of Washington, Seattle, WA 98195

Judith Reid Graves, R.N., E.M.T.-P.
Center for Evaluation of Emergency Medical Services, Seattle, WA 98164

Citing Articles (1)