Mechanical cardiopulmonary resuscitation in adult cardiac arrest: setting-specific associations and mediation through resuscitation process quality

Mechanical cardiopulmonary resuscitation (mCPR) devices aim to improve outcomes after cardiac arrest by delivering consistent, high-quality chest compressions. However, the mechanisms underlying their clinical effects remain incompletely understood. This study aimed to evaluate the association of mCPR with return of spontaneous circulation (ROSC) and survival to hospital discharge, and to explore the potential mediating role of CPR quality indicators. In this multicenter retrospective cohort (2022–2024), we included adults with cardiac arrest who received mechanical or manual CPR. The prespecified primary analysis was stratified by arrest location (OHCA vs. IHCA). Within each stratum, confounding was addressed using stabilized inverse-probability-of-treatment weights truncated at the 1st/99th percentiles; covariate balance (target SMD < 0.10) and common support were verified. Weighted logistic regression with Huber–White robust standard errors then estimated adjusted odds ratios (aORs) for ROSC and survival to hospital discharge. Exploratory mediation analysis evaluated indirect effects through CPR quality metrics, including compression rate, compression depth, and no-flow time. Among 723 adults (mechanical = 193; manual = 530), the prespecified stratified IPTW analysis showed that mechanical cardiopulmonary resuscitation was associated with higher odds of ROSC and survival to discharge in out-of-hospital cardiac arrest (OHCA)—ROSC aOR 3.65 (95% CI 1.94–6.87); survival aOR 5.01 (2.41–10.43)—but not in in-hospital cardiac arrest (IHCA)—ROSC 1.24 (0.56–2.76); survival 2.71 (0.92–8.02). The treatment-by-location interaction was significant for ROSC (p = 0.038). Findings were directionally consistent in overlap-weighting and propensity-score matching sensitivity analyses (IHCA estimates remained imprecise). Exploratory mediation suggested indirect effects via improved compression rate/depth and reduced no-flow time. mCPR was associated with better outcomes in OHCA but not IHCA; exploratory mediation suggests a pathway via CPR-process metrics. These observational findings support context-specific implementation and further prospective study. Standardized neurologic outcomes (e.g., CPC or mRS) were not uniformly available, which limits interpretation of survival endpoints.”