CPR: What’s better mechanical or manual compressions?

Meta-analysis of the available research looks at primary and secondary patient outcomes of applying mechanical chest compression devices following out-of-hospital cardiac arrest.

By Kenny Navarro

Medic 58 responds with Engine 19 on a reported unconscious person. Engine 19 arrives to find a 60-year-old male who collapsed while watching a football game on television. Family members performed chest compressions with the assistance of the emergency medical dispatcher.

The crew from Engine 19 takes over chest compressions and immediately attaches the AED. The AED detects a shockable rhythm and recommends a shock. The crew momentarily interrupts CPR to deliver the shock and immediately resumes chest compressions.

Medic 58 arrives on scene before the AED completes a second analysis of the patient’s cardiac rhythm. Medic 58 is equipped with a mechanical CPR device that delivers chest compressions via an integrated suction cup.

During the set up of this device, the AED performs a second analysis and recommends another shock. After delivering the shock, the mechanical CPR device becomes operational and manual chest compressions cease.

One of the medics inserts a supraglottic airway and confirms proper placement. The initial end-tidal carbon dioxide (ETCO2) value is 18 mm Hg. The ECG monitor displays an organized rhythm at a rate of 30 complexes per minute.

After administration of epinephrine, the rhythm rate begins to increase and the medics notice the ETCO2 value slowly begins to rise.

Eight minutes after arriving on scene, the ETCO2 value is 58 mm Hg. The patient has regained a pulse but remains unconscious and is not breathing. Medic Davis turns the mechanical chest compressor off but leaves the device in place. Before leaving the scene, Medic Gutierrez performs a 12-lead ECG, which reveals the presence of a large inferior wall ST-segment myocardial infarction (STEMI).

Transport to the STEMI center is uneventful and the patient is quickly transferred to the cath lab. The cardiologist congratulates the medics on a successful resuscitation, but the humble medics reply the switch from manual to mechanical chest compressions is primarily responsible for the save.

Study review
Some have argued that the failure to show improved outcomes associated with mechanical chest compression may be more reflective of the limitations and execution of the study designs than an indictment of the machines [1]. In an attempt to address this issue, a research team conducted a meta-analysis of all evidence, both randomized and observational, published on the subject since the year 2000 [2].

A meta-analysis is a rigorous review method that allows researchers to integrate the research findings of several small trials into one analysis [3]. By pooling these results in one analysis, researchers can establish a more accurate estimate of a treatment effect than can be determined by any single small study alone.

The ultimate goal of any resuscitation attempt is to return the patient to as normal a life as possible — a discharge from the hospital with favorable neurologic function. However, this outcome measure is a reflection of both in- and out-of-hospital care, along with a number of other factors.

In some cases, care delivered in the hospital could make prehospital interventions appear more or less effective than they really are. With this in mind, the research team chose survival to hospital admission as the primary outcome for the present investigation rather than discharge from the hospital.

The team also measured and reported on a number of secondary outcomes, such as return of spontaneous circulation (ROSC) in the field, survival to hospital discharge and favorable neurological outcome at discharge. These secondary outcomes provide useful information about CPR machines, but one must be careful when interpreting effectiveness based on secondary outcomes.

Study methodologies are determined by the research question and the primary outcome variable used to answer the question. Attempting to draw firm conclusions based on a secondary variable may require a completely different methodological approach.

In conducting this meta-analysis, the research team chose an inclusion period that used the European Resuscitation Council/American Heart Association guidelines from 2000, 2005 and 2010. Changes in assessment sequences, ventilation recommendations, compression rates and depths, and a host of other recommendations during this period, could have affected the outcome.

To address that concern, the research team conducted a sensitivity analysis, which allows the researchers to see the effects caused by changing any one of those variables. A sensitivity analysis allows the researchers to determine which variables have the most influence on the outcome [4].

Results of the meta-analysis
Twenty studies met the inclusion criteria for this meta-analysis. Of those, five had a randomized design and the remaining studies were observational in nature.

The authors of the present study categorized the data from the five randomized trial as high quality with little risk of bias. The data from the observational studies was categorized as good to moderate quality.

Combining the study populations of each of the included investigations yielded a meta-analysis population of 21,363 patients, of which 9,391 received mechanical chest compressions and 11,972 received manual compression. EMS personnel used the Autopulse in 11 studies, the LUCAS device in eight studies and both devices in one study.

For the primary endpoint of survival to hospital admission, the research team calculated odds ratios separately for the randomized (high-quality data) and the observational studies (good to moderate data). Pooling data from the randomized studies did not alter the odds of survival to hospital admission between the two groups.

However, after pooling data from the observational studies, patients who received mechanical chest compressions were 42 percent more likely to survive to hospital admission when compared with those patients receiving manual chest compression.

When using secondary outcome variables, the research team found very similar results. Pooling the high-quality data from the randomized studies found no differences between the two groups in the likelihoods of achieving ROSC, surviving to hospital discharge or having favorable neurological status at the time of discharge.

However, pooling data from the observational studies demonstrated a 74 percent increase in the likelihood of achieving ROSC for patients who received mechanical chest compressions. Unfortunately, survival to hospital discharge and favorable neurological status did not appear to be influenced by the use of mechanical chest compression devices.

Regardless of whether the patient received mechanical or manual chest compressions, there did not appear to be any differences in survival rates associated with the specific guidelines (year 2000, 2005, and 2010) used by the rescue teams.

What this means for you
This meta-analysis represents the most comprehensive review of the use of mechanical chest compression devices in the out-of-hospital environment to date. The routine application of mechanical chest compression devices does not appear to result in improved clinical outcomes for patients who suffer OOH cardiac arrest.

This meta-analysis confirmed the results of a previous meta-analysis of randomized trials of OOH cardiac arrest, which found the odds of achieving ROSC were not significantly different between mechanical CPR devices and manual CPR [5]. In the previous meta-analysis, patients who received CPR from a load-distributing band instead of manual CPR were 62 percent more likely to achieve ROSC.

However, the variability of reporting longer-term outcome data in the studies used for the meta-analysis prohibited the authors from drawing any firm conclusions about whether mechanical CPR devices improve long-term survival.

In each patient encounter in the observational trials, EMS personnel had a choice of whether to provide manual CPR only or to apply a mechanical CPR device. This type of study protocol can lead to selection bias.

If, for example, rescuers believe mechanical CPR is superior to manual CPR, the rescue team could choose to apply the mechanical device only when they feel the patient has a good chance of survival — younger, presenting in a shockable ECG rhythm, witnessed cardiac arrest receiving bystander BLS, but perform manual CPR on those judged to be beyond help — older, unknown down time, presenting with asystole.

Selection bias could make the mechanical CPR machine look more effective than it really is, which may have occurred in these observational trials.

Study limitations
Randomization helps to reduce the effects of selection bias. A meta-analysis that includes only randomized trial results provides a clearer picture of the effects of the intervention. One might ask then, why researchers do not design more randomized trials instead of observational trials. The simple answer is that randomized trials are significantly more complex and expensive to execute.

The current meta-analysis included five randomized trials with high-quality data. Even after pooling the data, limitations associated with the individual studies could explain why the meta-analysis did not find superiority of one CPR method to another.

For example, an independent data safety monitoring board prematurely halted the AutoPulse Assisted Prehospital International Resuscitation trial when an interim progress analysis discovered worsened neurological outcomes for patients who received mechanical chest compressions [6]. The ASPIRE authors noted a significant difference between the two groups in the amount of time it took to deliver the first defibrillation shock, with the mechanical compression group waiting an average of 2.1 minutes longer.

This delay could reflect the unintended consequences of on-scene prioritization of CPR machine deployment rather than early defibrillation. Likewise, enthusiasm for mechanical chest compressions may have resulted in resuscitation attempts for patients who otherwise might have been pronounced dead on the scene. This is supported by the fact that medics enrolled more asystolic patients in the mechanical chest compression group than in the manual compression group.

The multi-center LUCAS in Cardiac Arrest (LINC) trial had a similar limitation [7]. The study protocol required application and use of the CPR machine before delivering the first shock to the experimental group. This created a significant difference between how quickly patients in the two groups received the first shock after paramedics arrived on scene.

Since the American Heart Association acknowledges the most important interval directly affecting survival rates is the interval from arrival to first shock [8], this delay could have masked any survival benefits provided by mechanical CPR.

Researchers in the LINC trial also did not monitor study protocol adherence however, an analysis of the defibrillation data suggests the medics did not follow the study protocol for the mechanical chest compression group in about one-fourth of the cases.

Similarly, only 60 percent of the patients randomized to receive mechanical chest compression actually did in the Pre-hospital Randomised Assessment of a Mechanical Compression Device in Cardiac Arrest trial [9]. Medics documented a number of reasons why patients in the remaining 40 percent did not receive mechanical chest compressions, including crew error, patients being too large or inadequate crew training in the use of the device.

In contrast, researchers in the Circulation Improving Resuscitation Care Trial did monitor protocol compliance in both the control and the experimental groups, although this trial used a load-distributing band instead of a device with an integrated suction cup [10]. In this trial, the medics appeared to follow the study protocol but still could not demonstrate superiority of one chest compression method over the other.

Although the CIRC researchers did measure chest compression fractions in both groups (which were similar), they were not able to consistently measure chest compression depth. Therefore, any unknown dissimilarities in CPR depth could be responsible for the outcome.

In each of the randomized trials used for the meta-analysis, medics provided manual chest compressions while applying the mechanical compression devices. This essentially created study groups of manual and mechanical compressions versus manual compressions alone.

In one study, medics performed manual chest compressions for almost three minutes before having the mechanical chest compression device activated [11]. Poor quality manual compressions for this amount of time before beginning mechanical compressions could make the machines appear less effective.

Despite the perceived short term benefits associated with mechanical chest compression devices, this meta-analysis of the available research does not support a routine strategy of applying mechanical chest compression devices in an effort to improve clinical outcomes following out-of-hospital cardiac arrest.

Although the latest American Heart Association Guidelines establish manual chest compressions as the preferred method of initial circulatory support following cardiac arrest, they do acknowledge situations when mechanical chest compressions are an acceptable alternative [12].

These situations include conditions that make manual chest compressions impossible or unsafe to deliver, in situations where there are a limited number of rescuers, and in a moving ambulance. EMS personnel with access to these devices must be adequately trained in their efficient use and develop an application strategy that limits interruptions in chest compressions.


  1. Paradis, N. A., Young, G., Lemeshow, S., Brewer, J. E., & Halperin, H. R. (2010). Inhomogeneity and temporal effects in AutoPulse Assisted Prehospital International Resuscitation—an exception from consent trial terminated early. Journal of Emergency Medicine, 28(4), 391-398. doi:10.1016/j.ajem.2010.02.002.
  2. Bonnes, J. L., Brouwer, M. A., Navarese, E. P., Verhaert, D. V., Verheugt, F. W., Smeets, J. L., & de Boer, M. J. (2016). Manual cardiopulmonary resuscitation versus CPR including a mechanical chest compression device in out-of-hospital cardiac arrest: A comprehensive meta-analysis from randomized and observational studies. Annals of Emergency Medicine, 67(3), 349-360. doi:10.1016/j.annemergmed.2015.09.023
  3. Haidich, A. B. (2010). Meta-analysis in medical research. Hippokratia, 14(Suppl 1), 29-37.
  4. Taylor, M. (2009). What is a sensitivity analysis? Retrieved from http://www.medicine.ox.ac.uk/bandolier/painres/download/whatis/What_is_sens_analy.pdf
  5. Westfall, M., Krantz, S., Mullin, C., & Kaufman, C. (2013). Mechanical versus manual chest compressions in out-of-hospital cardiac arrest: A meta-analysis. Critical Care Medicine, 41(7), 1782–1789. doi:10.1097/CCM.0b013e31828a24e3
  6. Hallstrom, A., Rea, T. D., Sayre, M. R., Christenson, J., Anton, A. R., Mosesso, V. N. Jr., Ottingham, L. V., Olsufka, M., Pennington, S., White, L. J., Yahn, S., Husar, J., Morris M. F., & Cobb, L. A. (2006). Manual chest compression vs. use of an automated chest compression device during resuscitation following out-of-hospital cardiac arrest. Journal of the American Medical Association, 295(22), 2620–2628. doi:10.1001/jama.295.22.2620
  7. Rubertsson, S., Lindgren, E., Smekal, D., Östlund, O., Silfverstolpe, J., Lichtveld, R. A., Boomars, R., Ahlstedt, B., Skoog, G., Kastberg, R., Halliwell, D., Box, M., Herlitz, J., & Karlsten, R. (2014). Mechanical chest compressions and simultaneous defibrillation vs conventional cardiopulmonary resuscitation in out-of hospital cardiac arrest: The LINC randomized trial. Journal of the American Medical Association, 311(1), 53-61. doi:10.1001/jama.2013.282538
  8. Sinz, E., Navarro, K, & Soderberg, E. (Eds.). (2013). ACLS for experienced providers. Dallas, TX: American Heart Association.
  9. Perkins, G. D., Lall, R., Quinn, T., Deakin, C. D., Cooke, M. W., Horton, J., Lamb, S. E., Slowther, A. M., Woollard, M., Carson, A., Smyth, M., Whitfield, R., Williams, A., Pocock, H., Black, J. J., Wright, J., Han, K., Gates, S.; & the PARAMEDIC Trial Collaborators. (2015). Mechanical versus manual chest compression for out-of-hospital cardiac arrest (PARAMEDIC): A pragmatic, cluster randomised controlled trial. Lancet, 385(9972), 947-955. doi:10.1016/S0140-6736(14)61886-9
  10. Wik, L., Olsen, J. A., Persse, D., Sterz, F., Lozano, M. Jr., Brouwer, M. A., Westfall, M., Souders, C. M., Malzer, R., van Grunsven, P. M., Travis, D. T., Whitehead, A., Herken, U. R., & Lerner, E. B. (2014). Manual vs. integrated automatic load-distributing band CPR with equal survival after out of hospital cardiac arrest. The randomized CIRCtrial. Resuscitation, 85(6), 741-748. doi:10.1016/j.resuscitation.2014.03.005
  11. Smekal, D., Johansson, J., Huzevka, T., & Rubertsson, S. (2011). A pilot study of mechanical chest compressions with the LUCASTM device in cardiopulmonary resuscitation. Resuscitation. 82(6), 702-706. doi:10.1016/j.resuscitation.2011.01.032
  12. Brooks, S. C., Anderson, M. L., Bruder, E., Daya, M. R., Gaffney, A., Otto, C. W., Singer, A. J., Thiagarajan, R. R., & Travers, A. H. (2015). Part 6: Alternative techniques and ancillary devices for cardiopulmonary resuscitation. 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation, 132(suppl 2), S436–S443. doi:10.1161/CIR.0000000000000260

About the author

Kenny Navarro is an Assistant Professor in the Department of Emergency Medicine at the University of Texas Southwestern Medical School at Dallas. He also serves as the AHA Training Center Coordinator for Tarrant County College. Mr. Navarro serves as an Emergency Cardiovascular Care Content Consultant for the American Heart Association, served on two education subcommittees for NIH-funded research projects, as the Coordinator for the National EMS Education Standards Project, and as an expert writer for the National EMS Education Standards Implementation Team. Send correspondence concerning any articles in this section to Kenneth W. Navarro, The University of Texas Southwestern Medical School at Dallas, 5323 Harry Hines Blvd MC 8890, Dallas, Texas 75390-8890, or e-mail kenny.navarro@ems1.com.

Comments - Add Yours

Leave a Reply

Your email address will not be published. Required fields are marked *