Documentation:FIB book/Biomechanics of Cardiopulmonary Resuscitation CPR-Related Injuries: Comparative Analysis between Manual and Mechanical Chest Compression Systems

Introduction

Anatomical diagram of the human sternum and rib cage, highlighting the manubrium (green), the body of the sternum (blue), and the xiphoid process (purple).

Cardiopulmonary resuscitation (CPR) is a critical lifesaving intervention used when the heart stops beating, aiming to maintain blood circulation to vital organs until normal heart function can be restored . CPR can triple the chance of survival following cardiac arrest by keeping oxygenated blood flowing to the brain and other vital organs . This process involves applying rhythmic chest compressions, which can be delivered manually or mechanically via the use of automated devices such as the LUCAS 2 system . This device administers consistent and controlled chest compressions using a piston system driven by either air or battery power.

While CPR is essential as a lifesaving intervention, the force required to circulate blood also places substantial stress on the thorax. The most common CPR-related thoracic injuries are rib fractures and sternal fractures , highlighting a need for balancing effective compressions and the risk of injury. Large clinical datasets show that rib fractures occur in roughly 60-70% of CPR cases and sternal fractures occur in 30-40% . Much of the existing evidence comes from cadaver-based loading studies, real-time measurements collected during live CPR, and research using mechanical compression devices. Understanding the biomechanics behind these injuries is essential because it allows healthcare professionals and engineers to identify how the thorax responds to these forces to improve CPR techniques, device design, and patient outcomes that balance effective blood flow with reduced injury risk. Injury biomechanics research in CPR helps understand safe loading tolerances for the thorax and guides prevention strategies which may include refining recommended compression depths and improving the design and function of mechanical CPR devices.

In contrast to automotive crash testing that involves dynamic impacts at relatively much higher speed and lasting shorter in duration, CPR is less regulated and presents a unique loading condition. The repeated chest compressions necessary are closest to an intermediate regime between quasi-static and dynamic impact loading. These compressions generate substantial thoracic deflections at moderate rates. The recommended guideline is to provide compressions to a depth of at least 2 inches (5 cm) to a maximum of 2.4 inches (6 cm) at a rate of 100 to 120 compressions per minute , with minimal muscle tension. Because of this, as well as the limited presence of CPR biomechanics in literature as most studies are clinical, insights from both quasi-static and dynamic loading literature were used to supplement this review to help contextualize thoracic behavior under CPR-relevant loading conditions.

This review first examines the biomechanics of the thorax under loading, then evaluates factors contributing to injury risk in both manual and mechanical CPR. Strengths, limitations, and controversies within the existing literature will be discussed, and gaps to guide future research will be addressed.

To gather literature review findings, peer-reviewed articles were identified using search terms such as CPR biomechanics, thoracic injury, rib fracture, and mechanical chest compression. The team selected studies based on relevance to thoracic mechanics, injury outcomes, and force characteristics.

Overview of Thoracic Biomechanics Under Loading

Literature directly addressing the biomechanics of CPR-induced injuries is limited. Most of the CPR literature focuses on clinical and physiological aspects, rather than injury biomechanics. CPR is unique in the sense that it falls within an intermediate loading regime between quasi-static and dynamic impact loading, such as frontal crashes, with significant chest deflection and minimal muscle tension. The magnitude of chest compression recommended by the American Heart Association (38-51 mm) corresponded closely to the chest deflection seen in FMVSS 208 tests (64 mm) at the time of the study, though they noted that the rate of loading in CPR was an order of magnitude slower than those seen in frontal crashes . Consequently, this review incorporates findings from quasi-static and dynamic thoracic loading studies to provide a basis for understanding the thorax response and ribcage behaviour under compressive forces like those experienced during CPR.  

Studies of dynamic thoracic impact provide insight into how and when rib fractures occur under rapid loading. Duma et al. instrumented cadavers with strain gauges to examine rib fracture timing under a simulated belt-loading condition. They found that all fractures occurred within the first 35% of chest compression on the left side, where the shoulder belt passed over, occurring first, then followed by fractures occurring in the upper right ribs . This illustrates that local load concentration strongly influences fracture initiation, and that injury can occur at relatively small fractions of maximum compression. When the study compared their results to NHTSA thoracic injury criteria for Hybrid III frontal impact dummies (63 mm for a 50th percentile and 52 mm for a 5th percentile female), the rib fractures in cadavers occurred at considerably lower chest deflections than the dummy-based thresholds, suggesting that the ATD-based thresholds may overpredict the deflection required to cause injury under dynamic loading . For CPR, this suggests that although the loading rate is lower than in a crash, local stress concentrations such as the rescuer’s hands over specific ribs may have the potential to cause injuries early in the compression cycle, before the chest reaches the full compression depth that is recommended as a standard. Thus, the location and distribution of applied force during CPR must be considered during delivery when considering injury risk.  

When compared to CPR, the recommended chest compression depth guidelines of 5–6 cm fall within the range of chest deflections associated with early rib fractures in the cadaveric study. Although the loading rate in CPR is much slower than in a car crash (2 compressions per second or 120 compressions per minute), the study highlights the initial failure points of the thorax under anterior–posterior compression. In CPR, the slower, repetitive nature of compressions introduces fatigue as another factor not covered by most impact studies. Regardless, this still supports the idea that rib and sternal fractures can occur early in the compression cycle, particularly in regions of higher local stress, providing a biomechanical basis for CPR-related injuries.

Similarly, Kindig et al. showed that rib response is highly rate-dependent: dynamically loaded ribs were stiffer and fractured at lower deflections compared to quasi-statically loaded ribs, although the total energy required for fracture was similar across loading rates . This supports the idea that rib fracture is a progressive event, with peak reaction forces occurring before complete structural failure. Together, these studies highlight that under rapid or localized loading, injury may initiate earlier than deflection metrics alone would suggest. Although CPR loading is slower, its repetitive nature may introduce fatigue mechanisms not captured in single-impact experiments. In the context of CPR delivery, the strong rate dependence implies that the lower loading rates used in resuscitation likely reduce instantaneous stiffness and allows greater deflections before fracture. However, because there are hundreds of repeated cycles, sub-fracture loads may accumulate effects from fatigue that are unexplored in current literature and uncaptured by single-impact type experiments.  

Given that CPR involves this repeated loading mentioned, an important question is whether such single injuries meaningfully change the thorax’s mechanical response. Shaw et al. investigated the response of five male cadaver torsos to both quasi-static and dynamic anterior loading. They loaded each subject 2-3 times per site, which introduced the potential for ribcage characteristics changing from successive loading (e.g., permanent tissue deformation).  Dynamic loading cases produced higher peak forces and greater chest deflection than quasi-static loading, while quasi-static force-deflection responses remained approximately linear between 18 and 25 mm. The sternum exhibited greater stiffness than the upper and lower ribcage sites, suggesting that rib structures deform more readily under anterior loading; a finding consistent with the prevalence of rib fractures identified from CPR from studies that use imaging. Interestingly, isolated rib fractures did not substantially reduce the ability of the chest to resist anterior loading, suggesting that a single injury does not substantially change the thorax’s biomechanical properties as a whole. However, the effects of rib fractures became more evident when many bicortical fractures (i.e., the fracture line completely severs the bone into separate pieces) were present and when the lower rib cage was loaded. Through autopsy, 61 fractures were identified for the 5 subjects in total with most (25) being bicortical non-displaced fractures (i.e., disruption of both the outer and inner surfaces of the rib). Overall, these findings indicate that the thoracic response is rate dependent and that cumulative rib fractures (e.g., multiple bicortical lower-rib fractures) can reduce the thorax’s stiffness and change the compression force-depth relationship during prolonged CPR. Since CPR can produce isolated rib fractures, it is important to note that such single injuries do not necessarily compromise chest mechanics or the effectiveness of compressions; rather, significant changes in thoracic stiffness typically require the accumulation of multiple bicortical lower-rib fractures.

Arbogast et al. compared anterior-posterior force-deflection characteristics between live CPR patients and PMHS. At a 40 mm deflection, live human CPR generated substantially lower peak forces (286 N) compared to PMHS loading (588 N), and no rib or sternal fractures were observed during clinical CPR . The living thorax exhibited greater hysteresis, underscoring its more viscoelastic behaviour.  

The PMHS data used for comparison was from Kent et al., where controlled anterior loading was applied to the PMHS using a hydraulic cylinder to develop thoracic response corridors representative of impact conditions similar to a 48 km/h sled test from different loading configurations (blunt hub, double diagonal belt, single diagonal belt, and distributed loading). Specifically, the hub-based response was used for comparison.

Direct extrapolation from cadaveric loading tests to CPR must be approached with caution. The loading regimes in impact and quasi-static studies differ substantially from the cyclic, moderate-rate, cumulative characteristics of CPR. Repeated loading at the same sites during cadaver experiments can introduce progressive tissue damage, but not in a manner that fully reflects the higher number of cycles during prolonged CPR. Overall, while dynamic and quasi-static cadaver studies offer boundary information on thoracic response, no experimental regime fully replicates the loading characteristics of CPR. Integrating these findings suggests that CPR-related injuries most likely arise from a combination of substantial anterior–posterior deflection, localized stress concentrations, and repetitive cyclic loading.

Injury Risk Factors related to Manual and Mechanical CPR

Demonstration of the LUCAS mechanical chest compression device applied to a CPR training manikin

The delivery of chest compressions, whether manual or mechanical, represents a critical biomechanical intervention where force is applied to the thorax in order to keep oxygen-rich blood flowing when the heart stops beating. Mechanical devices like the LUCAS 2 provide consistent compressions, and research into their use can provide insight into chest loading patterns and potential deformation and failure under repeated loading.

The LUCAS 2 device is a piston-based mechanical CPR machine designed to deliver consistent chest compressions at a preset depth and rate. Its advantages include maintaining consistent compression quality over prolonged resuscitation and reducing rescuer fatigue. Limitations include potential variation in patient-specific chest compliance, as well as the inability to adapt dynamically to anatomical differences, and risk of thoracic injury under high compression force and/or prolonged use. The device is typically configured to compress at a frequency of approximately 102 ± 2 per minute, with a target depth of 2.1 ± 0.1 inches for a nominal adult patient . It operates with a 50 ± 5% compression/decompression duty cycle and is intended for patients with a sternal height between 6.7 and 11.9 inches and a maximum chest width up to 17.7 inches. Depth is adjusted linearly for patients with smaller sternums.

Because the available biomechanics literature comparing manual and mechanical CPR is limited (and often clinically-based), we chose to focus on studies that offer quantifiable thoracic responses, mechanical loading parameters, and factors such as compression force variation, age, and compression location.

Azeli et al. investigated the mechanical behavior of the thorax during automated chest compressions delivered by the LUCAS 2 device, focusing on the relationship between compression force variation (CFV) and CPR-related injuries. CFV was defined as the difference between the maximum and minimum compression forces exerted by the device during the first 30 minutes of administering mechanical CPR via the LUCAS 2 device. Sternal fractures were present in 42.3% of cases, and at least one rib fracture was observed in 63.5% of patients. Although this data originates from clinical and forensic assessments rather than controlled biomechanical testing, they highlight the prevalence of injury under repeated loading and its correlation with CFV.

Higher CFV was associated with the occurrence of ribcage injuries, including sternal and bilateral rib fractures. A bilateral rib fracture occurs on both the left and right sides of the body. The occurrence of bilateral rib fracture was correlated with subjects who experienced longer total CPR durations, suggesting a cumulative fatigue effect from repetitive loading. Biomechanically, the primary resisting forces during chest compression are from the elastic properties of the osseous-cartilaginous (both bone and cartilage components) system of the ribcage and the damping effects of the internal organs within the thoracic cavity. As the chest wall materials fatigue under repeated loading, structural integrity decreases, leading to the development of fractures. Once these fractures occur, the chest wall loses some of its elastic behaviour, leading to more unpredictable responses under subsequent compressions. Compression depth and total CPR duration were identified as some of the main risk factors.  

These findings are consistent with the clinical study by Friberg et al., which reported higher rates of sternal and rib in cardiac arrest non-survivors who received LUCAS-administered CPR.  

Moreover, as CPR compressions are related to anterior-posterior thoracic deflection, the findings of Kent and Patrie can be used to better understand rib fracture risk during CPR. While their study focused on different loading conditions (blunt hub, airbag, seatbelt, etc.), the goal of developing thoracic injury functions based on maximum chest deflection, age, and loading conditions can be applicable to CPR biomechanics. Of the 93 tests, 71 resulted in at least one rib fracture, and 39 involved more than six fractures, making this dataset relevant to both the onset and progression of fracture severity during repetitive chest loading seen in CPR. Injury risk was not dependent on loading conditions but strongly influenced by age. For example, they quantified a 50% risk of any rib fracture occurring for 35% chest deflection in a 30-year-old, compared to 13% chest deflection in a 70-year-old, indicating that older adults have substantially lower chest deflection tolerance and are more likely to sustain thoracic injuries under comparable compressive loads. Integrating these age-dependent tolerance data into CPR methods, such as mechanical compression device settings, could help balance effective CPR administration with reduced injury risk.

As CPR requires somewhat localized compression, compression location effects must be considered. Suazo et al. found that caudal displacement of the compression site along the sternum increases effective compression depth, especially over ribs 4 to 6. The study involved evaluating standardized points on the sternum (P1 to P5), with P5 near the sternal tip. Regions P2 and P4 achieved the target depth with lower forces, whereas P3 and P5 required higher forces. This highlights that compression location in these points could increase the risk of rib fractures and sternocostal separations.  

Taken together, these studies indicate that chest compression biomechanics is influenced by multiple factors such as compression force variation (CFV) and patient-specific characteristics such as age, and the location of applied force. Mechanical devices like the LUCAS 2 provide consistent compressions, but evidence shows that repetitive loading can still result in rib and sternal fractures, particularly when CFV is high or compressions are prolonged. Age dependence and regional variation in compression effectiveness further demonstrate that both manual and mechanical CPR can carry injury risk if applied without consideration of thoracic biomechanics and patient-specific characteristics.

Discussion

CPR-related thoracic injuries are reported at high frequencies across the literature, with rib and sternal fractures commonly observed and often exceeding more than half of cases. Age, chest stiffness, and variability in compression force consistently emerge as major factors influencing injury risk, yet there remains substantial uncertainty about how these injuries affect clinical outcomes or whether they reflect effective compressions rather than preventable harm. Studies also differ in their conclusions regarding technique, loading characteristics, and the translation of cadaveric findings to living patients. While the studies discussed have established the fundamental relationships between compressive forces, patient anatomy, and injury patterns, the highly variable nature of research methods and limited clinical data mitigate the ability to develop concrete and comprehensive injury prevention strategies. Despite advancements in experimental modelling, the field remains challenged by the gap between controlled biomechanical data and real-world clinical outcomes.

Strengths

Current research is strengthened by having a diversity of methods in pursuit of advancing the understanding of CPR biomechanics. The integration of cadaver and medical device-based research and live clinical data provides reinforcing perspectives on thoracic responses under CPR-like compressive loading conditions.

Cadaver-based research, similar to that performed by Shaw et al. and Kent and Patrie, offers controlled, repeatable experimental conditions that would be unethical or impractical to replicate in living human subjects. These studies successfully quantified fundamental biomechanical properties, including force-deflection relationships, rate-dependent stiffness variations, and age-related tolerance thresholds.

Limitations

Methodological Limitations

Despite these strengths, substantial limitations restrict clinical applicability. The most significant is the scarcity of direct biomechanical measurements from real CPR scenarios. The study by Arbogast et al. involved 91 CPR patients and 13 PMHS but lacked detailed information about compression technique variability or long-term outcomes. The majority of biomechanical data derives from cadaveric studies or anthropomorphic test devices rather than actual clinical CPR.

Cadaveric studies introduce inherent limitations related to post-mortem tissue changes. Arbogast et al. noted that PMHS data showed reduced hysteresis and altered viscoelastic properties compared to living tissue, due to the loss of perfusion and physiological changes that occur after death . Shaw et al. acknowledged during their study that repeated loading of cadaveric specimens risks progressively altering tissue properties through cumulative damage, potentially influencing results.

The design of these studies also limits generalization. For example, Shaw et al. loaded torsos only 2-3 times per site, while actual CPR practices can involve thousands of compressions. Additionally, Azeli et al. measured force variation only during the first 30 minutes despite identifying cumulative fatigue as a key mechanism. Additionally, the studies cited do not exactly reflect the characteristics of CPR and any conclusions must be drawn with caution. The loading regimes differ from the cyclic, lower-rate compression characteristics of CPR. The repeated loading at the same sites in the cadaver studies could change ribcage properties through cumulative damage, but not necessarily in a manner that reflects cumulative mechanisms from the higher number of compressions that CPR requires.

Outcome-Related Limitations

Although many studies document fracture prevalence, few link biomechanical loading to long-term clinical outcomes such as how these injuries affect survival, neurological recovery, or long-term quality of life.

Injury detection methods vary considerably across studies. Baumeister et al. noted that X-ray imaging likely underestimates fractures compared to methods such as CT or autopsy, complicating cross-study comparisons. Because the study included patients who received advanced life support, with survivors undergoing imaging such as X-ray and/or CT while non-survivors were examined by autopsy, the possible underestimation of fractures in survivors assessed only by imaging is a limitation. Additionally, there was no data on how manual compressions were administered before mechanical CPR, and the data for CFV was also from the first 30 minutes, thereby limiting conclusions about longer-duration loading effects.

Population-Related Limitations

Current research also exhibits significant gaps in population representation, predominantly involving adult males, with limited data on children, women, or individuals with anatomical variations.

Controversies

Areas of ongoing debate reflect fundamental uncertainties. Baumeister et al.’s study found no significant difference in rib fracture rates (injuries in 19/20 manual and 24/24 mechanical cases), while Azeli et al. linked higher compression force variation to increased bilateral and sternal fractures. The clinical significance remains debatable since most patients receive manual CPR before mechanical devices are applied.

Additionally, while rib and sternal fractures are common (63.5% and 42.3% respectively in Azeli et al.), the relationship between these injuries and post-resuscitation complications remains unclear. Debate arises from the argument on whether CPR-related fractures are inevitable consequences of lifesaving intervention or are avoidable through better understanding and development of injury-reducing techniques that don’t compromise effectiveness. While Suazo et al. found lower sternal compressions require 17% less force for equivalent depth. Whether these findings translate to reduced clinical injury risk remains up for debate.

Most notably, the application of cadaveric data to living patients is up for controversy as Arbogast et al. demonstrated clear differences between live and post-mortem responses, but the extent to which these affect injury prediction remains uncertain.

Another ongoing controversy involves whether the existence of fracture truly corresponds with worse clinical outcomes. High fracture rates are reported in a number of imaging based studies, but some studies found no clear difference in short term survival between individuals with and without sternal or rib fractures . Meaning that, fractures might just be a sign that CPR was strong and effective, not necessarily that the patient is doing worse. Conversely, other studies have shown that fractures can sometimes cause problems after the patient is resuscitated. Research found that broken ribs make it harder to breathe, which may increase the chance of pneumonia . Therefore, it is not certain whether CPR fractures are harmless side effects of good compressions or whether they make recovery harder. This highlights a gap between injury data and clinical prognosis.

There is also debate about whether compression technique or compression location can prevent injuries. Lower sternal positioning reduces the force that is required, whereas some studies reported very high fracture rates even with CPR guidelines are being used . Questioning if good hand placement really reduces injuries or does patient anatomy matter much more than technique.

This leads to questions about how age affects chest toughness and its tolerance. It was reported that older adult’s fracture at much lower chest deflections , when compared to younger people. When looking at the CPR guidelines, they apply uniform compression depths to all adults. Raising concerns whether using one standard depth might cause more fractures in older adults who are also a substantial portion of cardiac arrest cases.

Overall, these controversies show that CPR related injuries are more complex than we think. Where some studies have disagreed whether fractures are unavoidable and how much technique or patient anatomy contribute to injury risk. There is major uncertainties about applying cadaver data to living patients and there is still a big debate and whether current CPR guidelines should be used for all young and older adults. This highlights the need for further research and gaps within the delivery of CPR needs to evaluate more in ways that needs to be both effective and safe for all.

Future Work

Despite knowledge about CPR biomechanics growing, the current research is still fragmented and leaves major unanswered questions. To truly optimize CPR and reduce preventable injuries, future research should focus on key areas such as direct CPR measurement from real CPR cases and linking the injuries to clinical outcomes to advance understanding and enable evidence-based optimization of CPR delivery.

Lack of Direct CPR Biomechanical Data

A major challenge is the limited availability of direct measurements collected during real clinical resuscitations. Most in vivo data stem from a small number of studies, notably Arbogast et al., involving instrumented load cells and accelerometers during actual cardiac arrest resuscitation. Large studies across multiple hospitals, using built-in sensors to measure force and chest deflection in real time, would provide accurate data from a wide range of patients of different ages, sizes, and medical conditions to establish more representative empirical force-deflections relations. In specific, studies could resolve discrepancies between the mechanical forcing observed in controlled laboratory settings and actual clinical practice, potentially identifying whether the protective effects of living tissue, demonstrated by Arbogast et al., persist across diverse populations.

Linking Injuries to Clinical and Functional Outcomes

Moreover, understanding the consequences of CPR-related injuries also requires more deliberate and systematic study. Current literature reports the frequency of rib and sternal fractures but rarely establishes how these injuries affect neurological function, or long-term recovery. Prospective studies that collect high-quality CPR biomechanical data, apply consistent imaging methods to document injuries, and follow patients over time would help clarify how injury patterns influence outcomes. For example, methodological improvements such as cross-over simulations studies comparing mechanical and manual CPR devices would further support clearer links between injury patterns and clinical trajectories. Overall, this evidence is essential for determining whether modifying CPR techniques would improve or compromise resuscitation efficacy.

Modeling and Understanding Population-Specific Mechanical Responses

Additionally, certain populations like children and seniors remain unstudied, even though their chest structure and mechanical response differ significantly from an average adult. The development of age-specific FEA models is needed to account for pediatric and geriatric anatomical and material response variation. Validating these models against PMHS data would improve our understanding of how CPR affects these vulnerable groups. While there is currently no dedicated CPR-specific PMHS dataset for validating cyclic chest compression, rib fracture and organ injury across age groups, there is relevant PMHS data from blunt impact and automotive crash studies that can be used for reference, rather than specific benchmarking.  

This gap exists largely because, unlike automotive safety research, CPR biomechanics is not embedded within a regulated framework such as the FMVSS 208 nor emphasizes the development of standardized injury criteria. In contrast, CPR research operates under more clinically driven conditions, as reflected by the predominance of retrospective clinical studies rather than controlled biomechanical testing. True clinical CPR data can also only be collected opportunistically during resuscitation events, where measurement priorities should not interfere with patient care. Consequently, these datasets can also be more likely to be inconsistent and lack controlled experimental variability.

Establishing Standardized Protocols and Novel Technology

In parallel, CPR-based chest stiffness data from clinical populations and pediatric ATD corridor work can bridge the gap between impact-derived PMHS response corridors and the unique intermediate-rate loading seen in CPR, particularly for younger and older chests. Eventually, these integrated models and relationships could support the development of tailored pediatric and geriatric CPR guidelines that better balance safety and effectiveness.

Beyond immediate injuries and population variability, the relationship between CPR-related thoracic injury and post-resuscitation complications requires closer attention. Existing research shows that rib and sternal fractures can contribute to pneumonia, prolonged dependence on mechanical ventilation, and extended ICU stays, yet the reasons behind these associations remain uncertain. Investigating how pain, breathing instability, and physiological disruption contribute to these complications would support more informed decisions about management strategies, including whether surgical stabilization could benefit select patients .

However, progress in these fields also depends on improving consistency in how studies measure biomechanical forces and classify injuries. Variability in measurement tools, modeling approaches, and injury definitions makes it difficult to compare findings across studies or draw unified conclusions. Establishing widely accepted standards for data collection, modeling assumptions, and injury classification would create a stronger foundation for future research and enable more reliable meta-analysis. For instance, real-time biofeedback systems integrating CPR force and deflection measurement with dynamic compression guidance represent an underexplored research frontier. Current CPR feedback systems typically emphasize the rate and depth of application but do not provide feedback regarding force consistency, or force trajectory shape. Training systems that provide real-time feedback to rescuers regarding compression force and the frequency of beats during resuscitation could help reduce injury during CPR.