|Home | About | Journals | Submit | Contact Us | Français|
Thank you for the opportunity to respond to Dr. Paul Chan’s insightful comments regarding our manuscript “The Utility of Therapeutic Hypothermia for Post-Cardiac Arrest Syndrome Patients With an Initial Non-Shockable Rhythm.”1 In this manuscript, we utilized propensity score matching in a quasi-experimental design to mimic a randomized controlled trial (RCT). Propensity score analyses have been both promoted and criticized in the literature, with a known limitation being the impact of unmeasured variables; however, there have been no published randomized trials of targeted temperature management versus no temperature management for patients with initial non-shockable rhythms due to various barriers. Our analysis, which included 519 patients with an initial non-shockable rhythm from the Penn Alliance for Therapeutic Hypothermia (PATH) Registry between 2000–2013, was an achievable, alternative approach, offering some insights on this important question.
Dr. Chan raises concern that the variables included in our derivation of a propensity score did not reflect patients’ severity of illness and did not account for date (year) of the arrest and hospital location, as cardiac arrest outcomes have improved over time and variability in hospital performance has been observed.2,3 He also suggested that a much larger cohort would have been helpful. In deriving our propensity score we applied the findings of Lindner et al., who determined that age, sex, initial rhythm, witnessed arrest, duration of arrest and location of arrest all contribute to an individual’s propensity to receive therapeutic hypothermia.4 Additionally, in order to perform this quasi-experimental analysis, our untreated cohort included a larger number of patients from the early years of data collection, when hypothermia was under utilized, and patients who were potential candidates for therapy did not receive it. In order to address Dr. Chan’s concerns, we re-calculated our propensity score to include both year of arrest and treating hospital. We determined the odd’s ratios (OR) for good neurologic outcome and for survival to hospital discharge to be 2.36 (95% CI: 1.31–4.27) and 2.23 (95% CI: 1.35–3.71), respectively. In order to further explore the timing phenomenon, we applied the previously described propensity score derivation to the subset of patients who were treated after 2007 and found similar results for good neurologic outcome and survival to hospital discharge (OR 2.66 (95% CI: 1.44–4.91) and 2.40 (95% CI: 1.41–4.06), respectively). Incorporating the two variables suggested by Dr. Chan into a more robust propensity score confirms our published findings and strengthens the results.
In response to Dr. Chan’s comment calling for a repeat study in a larger cohort utilizing a more robust propensity score, we agree and hope that the recommendations from the recent IOM report calling for a national registry on cardiac arrest might allow for such an analysis. In addition, Lascarrou and colleagues recently published the study protocol for HYPERION, a multicenter trial in which 543 patients successfully resuscitated from non-shockable cardiac arrest will be randomized to therapeutic hypothermia versus therapeutic normothermia.5 We eagerly await the results of this ambitious multicenter trial.