The first time I saw one of these machines I thought we were in a dystopian novel. The robot-like machine is bolted to the patient, and it provides steady compressions without tiring. The patient I saw it used on was an older women. I had to look away.
In this video, the earnest clinician calls it “truly a lifesaver device.”
The idea is sound: survival of out of hospital cardiac arrest is dismal. It’s around 10% in most places in the US. Higher in Scandinavia. Proper CPR with the right placement, depth and rate of compressions is important. But humans vary; we tire and switch out to other people. The mechanical CPR devices purport to mitigate that variability.
A recent observational study in Circulation cleverly looked at rates of survival after emergency medicine services adopted use of the devices. Let’s look at that paper, then I will show you the trial data. In the end, I will ask why these devices are so popular.
Circulation Paper
The team used a cardiac arrest registry to assess agency-level rates of survival from out-of-hospital-cardiac arrest (OHCA) in the 2 years before and after their first documented use of a mechanical CPR device.
They looked at survival in 49 agencies that adopted devices vs 73 “control” agencies that never did. To be fair, there was a lot of selection in centers and patients. They excluded patients thought not to have a cardiac arrest, and those agencies without 2 full years of data.
But in the end they had 14k patients with data from before mechanical CPR and nearly 18k after mechanical CPR. They first looked at survival over time in the control arm to be sure that time wasn’t the cause of better survival.
The results were all negative.
Control agencies showed no survival trend over time (~9.6–10.6% favorable neurological survival annually) — confirming the design’s core assumption held.
At agencies that introduced mechanical CPR devices, survival was essentially flat before and after adoption: favorable neurological survival went from 8.9% to 8.3%; survival to discharge went from 11.0% to 10.0%. Neither change was statistically significant.
This held regardless of how heavily an agency used the device (<20% vs. >20% of cases), and held whether or not 9-1-1-witnessed arrests were included.
While there were limitations of this data, including the lack of granular data on both patients and agencies, there was not any signal of benefit in the outcome of interest.
Note here that the outcome of interest was not the quality of CPR, which may have been better, but survival and survival with neurologic recovery. Here, there was no signal of benefit.
You’d think if these things were so great there would be an uptick in outcomes in agencies that bought and used these devices.
Randomized Trial Data
The RCT data is similar.
The PARAMADIC trial (LANCET) randomized approximately 6000 patients to the device vs standard CPR. Survival at 30 days was similar (6% vs 7%) with an odds ratio of 0.86 (95% CI 0.64-1.15).
The CIRC trial (RESUSCITATION) randomized 4700 patients to mechanical CPR (Auto-Pulse device) vs standard CPR. The adjusted odds ratio of survival to hospital discharge for mechanical CPR vs manual CPR, was 1.06 (95% CI 0.83–1.37). The trial was declared positive because it had a non-inferiority design, but you can see there is no significant difference in outcome.
The LINC trial (JAMA) randomized 2600 patient to mechanical CPR to manual CPR. There were also no significant differences in 4-hour survival, survival to discharge or survival with good neurologic outcome.
A Cochrane Systematic review of 11 trials also found no evidence of benefit and a small signal of harm.
Comments
I’ve looked at this data previously and asked an emergency medicine specialist why these devices are so popular. He could not give an obvious answer. Maybe it’s the music that goes along with the short videos.
Also, maybe buried in this negative data are patients who could benefit. But I doubt it.
Why? Well, let’s say the device is marginally better at CPR than humans. Good CPR is one thing on the causal chain to survival with good function. But there are many others: the underlying disease, the time to starting CPR, the time to defibrillation, and the quality of care in the hospital. In other words: a lot of things besides CPR.
Given the aggressiveness of these devices and their costs (approximately $20,000 plus disposables x many ambulances), I would think we would abandon them. That is, if the treatment of patients with out of hospital cardiac arrest is to remain an evidence-based intervention.
I am interested in any experts who can counter this slew of negative data and convince me the jackhammer-like devices are worth using.
As a resident, when I worked in the hospital for shifts that lasted up to 24 hours, I participated in probably a dozen codes, all using manual compression, multiple medications and electric shocks when appropriate. Every single one of them was successful in that the patient's pulse was restored. But none of them was successful in that the patient left the hospital alive - or with any semblance of a mental or functional life. In every case, the patient was elderly and had died due to a medical condition that had eroded the organs of their bodies by degrees over years, and so reviving them merely delayed an inevitable ending before they died a second time. Generally, every one of the dozen people in the room know this from the start: the familiar shared glance that communicates a shared knowledge of the futility of what we are doing, the adrenaline rush that is vaguely unsatisfying as the aides change out the bespattered sheets and gowns while the sanitation crew sweeps up the debris of syringe caps and bloody gauze pads and stepped-on tubing. Being DNR/DNI would have been a mercy, not just to the medical staff, but to the patients themselves.
The real question is not whether mechanical CPR is better than manual. Eventually, John Henry will get beaten by the machine and his heroism and humanity and sweat will become the stuff of legend and folklore. The real question is whether we should be doing most of these resuscitations at all. The price of the machines is tiny compared to the cost of ICU care over the subsequent days and weeks, or the longterm care in the few cases where the patient leaves the hospital to "convalesce," all of them now, finally, DNR.
As an aside - as healthcare costs continue to rise, we need to ask ourselves how health systems, physicians, and other clinicians contribute, because ultimately someone is paying for this device. Our natural inclination is to blame others - pharma, the insurance companies, others, etc - rather than looking at how we contribute. Most insurance companies are happy to expand coverage for drugs or devices - they just pass the cost on to who is paying for it. However, for Medicaid and Medicare it has a huge impact on state and federal budgets.