Efficacy and safety of mechanical versus manual compression in cardiac arrest - A Bayesian network meta-analysis

Aim: To compare relative efficacy and safety of mechanical compression devices (AutoPulse and LUCAS) with manual compression in patients with cardiac arrest undergoing cardiopulmonary resuscitation (CPR). Methods: For this Bayesian network meta-analysis, seven randomized controlled trials (RCTs) were selected using PubMed/Medline, EMBASE, and CENTRAL (Inception-31 October 2017). For all the outcomes, median estimate of odds ratio (OR) from the posterior distribution with corresponding 95% credible interval (Cr I) was calculated. Markov chain Monte Carlo (MCMC) modeling was used to estimate the relative ranking probability of each intervention based on surface under the cumulative ranking curve (SUCRA). Results: In analysis of 12, 908 patients with cardiac arrest [AutoPulse (2, 608 patients); LUCAS (3, 308 patients) and manual compression (6, 992 patients)], manual compression improved survival at 30 days or hospital discharge (OR, 1.40, 95% Cr I, 1.09-1.94), and neurological recovery (OR, 1.51, 95% Cr I, 1.06-2.39) compared to AutoPulse. There were no differences between LUCAS and AutoPulse with regards to survival to hospital admission, neurological recovery or return of spontaneous circulation (ROSC). Manual compression reduced the risk of pneumothorax (OR, 0.56, 95% Cr I, 0.33-0.97); while, both manual compression (OR, 0.15, 95% Cr I, 0.01-0.73) and LUCAS (OR, 0.07, 95% Cr I, 0.00-0.43) reduced the risk of hematoma formation compared to AutoPulse. Probability analysis ranked manual compression as the most effective treatment for improving survival at 30 days or hospital discharge (SUCRA, 84%). Conclusions: Manual compression is more effective than AutoPulse and comparable to LUCAS in improving survival at 30 days or hospital discharge and neurological recovery. Manual compression had lesser risk of pneumothorax or hematoma formation compared to AutoPulse.