Real-World Failure Patterns of Mechanical CPR Devices

Mechanical cardiopulmonary resuscitation (mCPR) devices are widely used in emergency department and prehospital cardiac arrest care, yet systematic characterization of device failure patterns remains limited. This work is particularly important given current American Heart Association guidelines, which no longer recommend routine mechanical CPR use but endorse selective deployment in scenarios where manual CPR quality or provider safety cannot be maintained, making device reliability central to patient safety.

We analyzed 6,509 FDA’s MAUDE database reports of mCPR device failure events from 2015–2025 using a trained artificial intelligence classification system to categorize failures by mechanism. Events were examined overall, by manufacturer, and longitudinally. Mechanical failure narratives were further reviewed to identify recurring root-cause themes.

Mechanical and battery failures were the predominant failure mode (3,359 reports or 51.6%) across the study period, remaining the leading category annually (45–58%). Load sensor and compression delivery failures (20.7%) and software failures (18.1%) comprised most remaining events. Root-cause analysis of mechanical failures revealed common themes of manufacturing or component defects, device aging or exceeding service life, brake gap or tolerance drift, load-cell failures, and impact-related damage. The largest manufacturer, Manufacturer A, reported failure patterns that mirrored the overall dataset. In contrast, Manufacturer B devices demonstrated a higher proportion of ineffective compression-related failures (42.6%) but matching mechanical failures (47.5%). Over time, load sensor failures increased in relative frequency after 2018, peaking at approximately 28% of annual reports.

More than half of reported mCPR device failures are mechanical and battery-related and persist as the dominant failure mode over a decade of real-world use. Many failures cluster around predictable mechanisms such as aging, component wear, and tolerance drift, suggesting opportunities for improved preventive maintenance, clearer device end-of-life criteria, and targeted design modifications.