Are Two Shocks Better Than One?: March 2023 Annals of Emergency Medicine Journal Club.

Cheskes S, Verbeek PR, Drennan IR, et al. Defibrillation strategies for refractory ventricular fibrillation. N Engl J Med. 2022;387:1947-1956. In patients experiencing out-of-hospital cardiac arrest (OHCA) with refractory ventricular fibrillation, do alternative defibrillation strategies, including vector-change (VC) defibrillation and double sequential external defibrillation (DSED) improve survival to hospital discharge? Design: Cluster randomized controlled trial. Setting: Six Canadian paramedic services. Population: Adult patients with OHCA due to refractory ventricular fibrillation. Intervention: Two alternative defibrillation strategies, VC defibrillation and DSED. Primary Outcome: Survival to hospital discharge. Secondary outcomes: Termination of ventricular fibrillation, return of spontaneous circulation, and survival with a good neurologic outcome. Sponsors: Funded by the Heart and Stroke Foundation of Canada ClinicalTrials.gov Identifier: NCT04080986 Among the 405 patients enrolled, a total of 136 patients (33.6%) were assigned to standard defibrillation, 144 (35.6%) to VC defibrillation, and 125 (30.9%) to DSED. The authors found that both DSED and VC defibrillation were associated with a higher chance of survival to hospital discharge compared with standard defibrillation (30.4% vs 13.3%; relative risk, 2.21; 95% confidence interval [CI], 1.33 to 3.67 for the DSED analysis and 21.7% vs 13.3%; relative risk, 1.71; 95% CI, 1.01 to 2.88 for the VC defibrillation analysis). Among the secondary outcomes, DSED but not VC defibrillation was associated with a higher chance of survival with a good neurologic outcome compared with standard defibrillation (27.4% vs 11.2%; relative risk, 2.21; 95% CI, 1.26 to 3.88). Overall, the trial conducted by Cheskes et al contributes meaningfully to our understanding of the use of alternative defibrillation strategies for refractory ventricular fibrillation before widespread incorporation of DSED and VC. These promising results should be confirmed in future randomized trials with larger sample sizes. Greater statistical power will improve the reliability of the observed effect size, while likewise reducing natural imbalances between baseline characteristics and prognostic factors between study cohorts. 1.Previous studies examining alternative defibrillation strategies have been confounded by resuscitation time bias. Describe resuscitation time bias, and how the design by Cheskes et al potentially limited its effects on their results. Resuscitation time bias occurs when an experimental intervention is employed late in the clinical course, after initial resuscitative efforts. As such, the patients captured for study enrollment are, by definition, nonresponders who differ in underlying physiology and prognosis from the general population of OHCA and VC. This difference in prognosis also includes the length of resuscitative efforts before randomization, an independent predictor of poor outcomes. The trial published by Cheskes et al limited the effects of resuscitation time bias through the process of randomization. In this case the authors used a cluster randomized trial design in which the paramedic service was randomized to defibrillation strategy. In this way the authors ensured that all patients received the defibrillation strategy which they were assigned at similar points in the resuscitation, avoiding the association with alternative defibrillation strategies and ultra-refractory arrest.2.This trial was stopped early because of COVID-19 pandemic-related operational challenges. As a result, the trial did not reach the planned sample size. How might this affect its results? Trials stopped early often tend to overestimate the effect size of an intervention in question.1Guyatt G.H. Briel M. Glasziou P. Bassler D. et al.Problems of stopping trials early.BMJ. 2012; 344e3863Crossref Scopus (85) Google Scholar Early in a trial’s enrollment it is not uncommon to observe large variations in the effect size, as the small sample size heightens the influence of random error. As the sample size grows these variations will diminish, regressing toward the true effect size. It ought also to be noted trials stopped early may also underestimate an intervention’s true effect size, but such results are more likely have a longer lead time toward publication, and to be published in less-prominent journals. The average effect on published outcomes, therefore, remains to overestimate the effect size of the intervention in question.2Pocock S.J. Hughes M.D. Practical problems in interim analyses, with particular regard to estimation.Control Clin Trials. 1989; 10: 209-221SAbstract Full Text PDF PubMed Scopus (113) Google Scholar The underlying principle on which trials are stopped early remains biased toward exaggerating effect sizes, regardless of whether they are stopped because of prespecified stopping rule or an external event such as the COVID-19 pandemic.3Montori V.M. Devereaux P.J. Adhikari N.K. et al.Randomized trials stopped early for benefit: a systematic review.JAMA. 2005; 294: 2203-2209Crossref PubMed Scopus (571) Google Scholar In the case of Cheskes et al, although the premature stoppage appears to be beyond the control of the investigators, it still limits the certainty with which we can view their results.