Mechanical versus manual chest compressions for cardiac arrest

Background Mechanical chest compression devices have been proposed to improve the effectiveness of cardiopulmonary resuscitation (CPR). Objectives To assess the effectiveness of mechanical chest compressions versus standard manual chest compressions with respect to neurologically intact survival in patients who suffer cardiac arrest. Search strategy We searched the Cochrane Central Register of Controlled Studies (CENTRAL) on The Cochrane Library, MEDLINE, EMBASE, Science Citation abstracts, Biotechnology and Bioengineering abstracts and Clinicaltrials.gov in November 2009. No language restrictions were applied. Experts in the field of mechanical chest compression devices and manufacturers were contacted. Selection criteria We included randomised controlled trials (RCTs), cluster RCTs and quasi-randomised studies comparing mechanical chest compressions to manual chest compressions during CPR for patients with atraumatic cardiac arrest. Data collection and analysis Two authors (SCB and LJM) abstracted data independently. Disagreement between reviewers was resolved by consensus and a third author (BB) if consensus could not be reached. The methodologies of selected studies were evaluated for risk of bias by a single author (SCB). The primary outcome was survival to hospital discharge with good neurologic outcome. We used the DerSimonian & Laird method (random-effects model) to provide a pooled estimate for relative risk with 95% confidence intervals. Main results Four trials, including data from 868 patients, were included in the review. The overall quality of included studies was poor and significant clinical heterogeneity was observed. Only one study (N = 767) reported survival to hospital discharge with good neurologic function (as defined as a Cerebral Performance Category score of 1 or 2), demonstrating reduced survival with mechanical chest compressions when compared with manual chest compressions (RR 0.41 (95% CI 0.21-0.79). Data from other studies included in this review were used to calculate relative risks for having a return of spontaneous circulation (2 studies, N = 51, pooled RR 2.81, 95% CI 0.96 to 8.22) and survival to hospital admission (1 study, N = 17, RR 4.13, 95% CI 0.19 to 88.71) in patients who received mechanical chest compressions versus those who received manual chest compressions. Authors' conclusions There is insufficient evidence from human RCTs to conclude that mechanical chest compressions during cardiopulmonary resuscitation for cardiac arrest is associated with benefit or harm. Widespread use of mechanical devices for chest compressions during cardiac is not supported by this review. More RCTs that measure and account for CPR process in both arms are needed to clarify the potential benefit from this intervention.