Some of the most extensive testing analyzed in the 1970s involved six-inch diameter rigid pendulum impacts to the sternum of unembalmed PMHS (Fig. 13.4). The data were presented by Kroell et al., Nahum et al., and Stalnaker et al. [55–60]. These data have also been analyzed by Lobdell et al. and Neathery [61, 62]. In the Kroell tests, a six-inch diameter impactor contacted the sternum at the level of the interspace between the fourth and fifth ribs. Figure 13.5 shows force- deflection curves for 4.02–5.23 and 6.71–7.38 m/s in the Kroell et al. tests.

The total chest deflection including the flesh overlying the sternum was included by Kroell et al. [55, 56]. Neathery developed corridors of skeletal chest deflection, based on these data [62]. Lobdell et al. showed that lower impactor masses resulted in lower deflections [61]. Patrick developed force-deflection curves under conditions similar to the Kroell unrestrained back conditions, but used himself as a volunteer and a padded (2.4 cm Rubatex R310V) six-inch diameter striker at velocities of 2.4–4.6 m/s [63]. Forces in the tensed condition were slightly greater than in the relaxed condition for the same impact velocity. Apparent initial stiffness were: 79 N/mm at 2.4 m/s for tensed, 57 N/mm at 2.4 m/s for relaxed and 250 N/mm at 4.6 m/s for tensed conditions. Peak skeletal deflections were 44–46 mm (16–17 % deflection of the rib cage).

The idealized force-deflection curves derived from these responses can be divided into a loading and an unloading phase (Fig. 13.6), with the loading phase having three components [64]. An initial rapid rise or apparent initial stiffness (A) is due in large part to the viscous properties of the thorax. The force plateau (B) is also due to a viscous response. At maximum deflection (C), the impactor and subject are moving at a common velocity and the forces are due to inertial forces caused by whole-body acceleration and the elastic forces due to tissue compression. The unloading portion of the curve (D), is due to unloading of the compressed tissues, and follows the elastic non-linear unloading of the thorax seen in quasi-static tests.

Beginning in the early 1960s, three-point restraint systems have been installed in some vehicles. However, in the 1987 model year all vehicle manufacturers had to certify that their cars would meet the 50th percentile, adult male protection requirements in the 48 km/h frontal, rigid-barrier test specified in FMVSS 208. Manufacturers began installing 3-point restraints with force-limiting shoulder belts and frontal airbags for the driver and right front passenger [33].

Fayon et al. calculated the resultant normal force to the thorax from the geometry and tension forces at belt anchorages and determined that sternal thoracic stiffness was 166.4 N/mm [65]. Walfisch came up with slightly smaller values of 70–116 N/mm (mean 119.4 N/mm) [5]. L’Abbe et al. conducted static and dynamic tests and belt loading and consequent chest deflection in human volunteers [66]. The dynamic loads were up to 3,600 N and produced mid-sternal stiffness of 137 N/mm, right 7th rib stiffness of 123 N/mm and 200 N/mm at left clavicle. The Hybrid III dummy tended to produce stiffer responses than relaxed volunteers, with better agreement to tensed volunteers, particularly at mid-sternum. Backaitis and St.-Laurent in similar tests showed greatest deflections at the seventh rib in volunteers and least in the shoulder, while Hybrid III showed maximum deflections at the shoulder [67].

Cheng et al. performed eleven PMHS tests with a VW two- point belt and knee bolster in the right-front passenger position [68]. The belt crossed from the right shoulder to the left side. There were more rib fractures on the left in seven cadavers. The number of rib fractures ranged from 0 to 14, with seven subjects having seven or more rib fractures. Peak upper shoulder belt load was 5.1–7.6 kN for 48 km/h impacts with peak sled deceleration of 22 g. The peak belt loads ranged from 3.6 to 9.7 kN for four runs with peak sled deceleration of 35 g. The ratio of lower to upper belt peak loads was approximately 0.9.

Kent et al. tested 15 PMHS in single and double diagonal belts and distributed, and hub loading on the anterior thorax [69]. Subjects were positioned supine on a table and a hydraulic master–slave cylinder with a high-speed materials testing machine provided controlled chest deflection at rates similar to sled tests. Non-injurious tests were followed by a final injurious test (40 % chest deflection). The distributed loading condition generated the stiffest response (3.33 kN at 4.6 cm), followed by the double diagonal belt condition (3.18 kN at 4.6 cm), single diagonal belt (2.28 kN at 4.6 cm) and hub (1.14 kN at 4.6 cm). Corridors were developed to a deflection level of 20 % of the 50th percentile male’s external chest depth. Kindig et al. studied of force-displacement and kinematic responses under a highly-localized loading condition using three PMHS (ages 44, 61, and 63 years) ribcages [25]. In general, bilateral loading produced an approximately symmetric deformation pattern, while unilateral loading resulted in approximately twice the resultant deformation on the ipsilateral side compared to the contralateral side.

Forman et al. [70] investigated a 3-point restraint system in the rear seat with a force-limiting (FL) and pretensioning (PT) shoulder belt in a series of 48 km/h frontal impact sled tests (20 g, 80 ms sled acceleration pulse) performed with PMHS [70]. The results of these tests were compared to matched PMHS tests, published in 2008, performed in the same environment with a standard 3-point belt restraint. The FL + PT restraint system resulted in significant (p < 0.05) decreases in peak shoulder belt tension (average ± standard deviation: 4.4 ± 0.13 kN with the FL + PT belt, 7.8 ± 0. 6 kN with the standard belt) and 3 ms-resultant, mid-spine acceleration (FL + PT: 34 ± 3.8 g; standard belt: 44 ± 1.4 g). The FL + PT tests also produced more forward torso rotation caused by decreased forward excursion of the pelvis and increased payout out of the shoulder belt by the force-limiter.

As three point belts and air bags are more frequently used, lower rate loading will become more important in frontal impact. The distributed loading to the ribs due to the airbag, and rib and clavicle loading due to the shoulder belt make the biomechanical response of the ribs and clavicle (in addition to the sternum) increasingly important. Thus, in addition to the dynamic pendulum data from sternal impacts, quasi-static chest loading data is also important. These data are described below.

Stalnaker et al. loaded the sternum of volunteers and PMHS with a 153-mm diameter rigid plate with the subject’s back against a rigid wall [60]. Stiffness averaged 12.2 N/mm for unembalmed PMHS, 40.2 N/mm for relaxed and 114 N/mm tensed volunteers. Lobdell et al. showed much less stiffness under similar loading conditions: 7 N/mm for relaxed volunteers and 23.7 N/mm for the tensed subject [61].

Tsitlik et al. [71] measured force versus deflection due to loading the sternum of eleven patients with a 48 × 64 mm rubber thumper [71]. The average stiffness was 9.1 N/mm (range: 5.25–15.9 N/mm) for deflections of 30–61 mm. At WSU, Cavanaugh et al. loaded the sternum and rib cage of three unembalmed PMHS with 25-mm strokes at quasi-static loading rates (described in [72]). The thorax stiffness were as follows: upper and mid-sternum, 8.6–12.3 N/mm; lower sternum, 5.7–11.4 N/mm; 2nd rib, 5.6–7.3 N/mm; 5th rib, 5.1–8.4 N/mm and 7th rib, 3.4–5.2 N/mm. Weisfeldt loaded the chest of volunteers at 60 Hz with a 22 × 64 mm pad [73]. The mean stiffness was 6.3 N/mm. On one subject loading the chest with a 153-mm diameter pad resulted in a stiffness of 21 N/mm. Melvin et al. after reviewing the quasi–static load–deflection literature, concluded that up to deflections of 41 mm the thorax has an approximate linear stiffness of 26.2 N/mm and for deflections greater than 76 mm, the linear stiffness is 120 N/mm [74]. Melvin also modeled these force–deflection relationships with a quadratic equation:
where K = 47.6 N/mm^2.

Fayon et al. demonstrated stiffness of 17.5–26.3 N/mm due sternal deflections up to 25 mm for static belt loading to supine volunteers [65]. This compared to 8.8–17.5 N/mm for disc loading up to 38 mm. At the 2nd and 9th ribs the estimated stiffness were 17.5–35 N/mm and 8.8–17.5 N/mm. Melvin and Weber estimated an apparent stiffness from the static belt tests of L’Abbe et al. [64, 66] (1982). These results were 67.5 N/mm at mid-sternum, 40 N/mm at right 7th rib, and 95 N/mm at left clavicle for deflections up to 10 mm and estimated normal forces up to 667 N. Melvin has suggested that these data may be significantly higher than that of Fayon et al. because spinal curvature may have reduced stiffness in the latter study. Arbogast et al. compared thoracic force-deflections measured during adult cardiopulmonary resuscitation (CPR) to that measured during hub-based loading of adult PMHS [75]. A load cell and accelerometer was integrated into a monitor-defibrillator to measure chest compression and applied force during live human CPR. Chest response during CPR demonstrated more hysteresis than the PMHS, indicating the viscoelastic nature of the thorax is more pronounced when tissues are naturally perfused and substantial tissue autolysis has not begun.

Human tolerance for severe chest injury is often stated as the peak spinal acceleration (sustained for 3 ms or longer) not to exceed 60 g in a frontal crash. The Hybrid III dummy is used for assessment of frontal impact crashworthiness per Federal Motor Vehicle Safety Standard 208 (Code of Federal Regulations, Title 49, Part 571.208, 1986). Early acceleration tolerance data was obtained by Stapp, who demonstrated the human tolerance to rocket sled acceleration when belt restraints were worn [78, 79]. Thoracic accelerations up to 40 g for 100 ms or less were tolerated. In one subject a maximum of 45 g was tolerated which resulted in a calculated pressure of 252 kPa under the harness. Peaks of 30 G reached at a rate of 1,000 g/s were not tolerated. Eiband in an analysis of the Stapp data showed that the acceleration tolerance decreased as duration of exposure increased [80]. Spinal acceleration is a general indicator of the overall severity of whole body impact, but is sensitive to changing impact conditions [81]. Mertz and Gadd studied the tolerance and response of a 40 year old male stunt man who took 16 dives from heights of 27–57 ft onto a thick mattress [82]. For each dive the stunt man executed three quarters of a turn and landed supine on the mattress. In ten dives chest acceleration was measured. The authors concluded that 50 g chest acceleration for a pulse less than 100 ms was within voluntary range of healthy adult males. A 60-g acceleration with duration less than 100 ms was recommended as the tolerance level until other the availability of data. This finding was used in the development of the FMVSS 208 chest injury criterion.

Bierman et al. tested young male volunteers with a drop-weight device that loaded a lap/double-shoulder belt harness (490 cm² in area) [83]. Painful reactions and minor injury occurred when loads exceeded 8.9 kN. When load area was increased to 1,006 cm², loads of 8.0–13.3 kN were sustained without injury. Patrick et al. performed a series of sled tests with embalmed PMHS to simulate the response of an unrestrained occupant [84]. The head, chest and knees impacted padded load cells. These studies were used as the basis for the design of the energy absorbing steering column. Gadd and Patrick, and Patrick et al. (1969) used a prototype energy absorbing steering column in unrestrained cadaveric sled tests [85]. A 3.3 kN hub load to the sternum and 8.8 kN distributed load to the shoulders and chest resulted in only minor trauma in centered impacts. The reactive force is due to the inertial, elastic and viscous components of the torso.

Eppinger and Marcus examined the results of 82 laboratory impact tests and concluded that severity of injury to the thorax was proportional to the amount of specific energy that the thorax must absorb [86]. The severity of injury was found to be inversely proportional to the impacted area and the duration of time over which the energy was transferred.

Kroell et al. analyzed a large number of blunt thoracic impact experiments (Fig. 13.4) and determined that chest compression correlated well with AIS (r = 0.730) while maximum plateau force did not (r = 0.524) [55, 56]. The linear equation relating AIS to compression was:
where C was chest deformation divided by chest depth, resulting in C = 0.3 (30 %) for AIS 2 and 0.4 (40 %) for AIS 4. Thus for the 230 mm chest depth of the 50th percentile male, 30 % compression or 69 mm deflection predicts AIS 2 and 40 % compression or 92 mm deflection predicts AIS 4.

The integral of spinal acceleration which is a measure of the velocity of deformation correlated almost as well as compression to injury severity [59]. In 5–7 m/s sternal impacts to PMHS, compressions > 20 % regularly produced rib fractures. Compressions of 40 % produced flail chest. Neathery et al. extrapolated the 40 % tolerance for severe injury to the Hybrid III frontal impact dummy to obtain maximum allowable compression of 75 mm [62]. Viano demonstrated that severe injury to internal organs occurred at an average maximum compression (C_max) of 40 % and recommended C_max of 32 % to maintain enough rib cage stability to protect internal organs [87]. The Federal Motor Vehicle Standards 208 (Code of Federal Regulations, 571.208), allowed a maximum 76 mm chest deflection in the 50th percentile Hybrid III dummy in frontal impact crashworthiness testing of automobiles. This was reduced to 63 mm in the FMVSS 208 standard upgrade, commencing in model year 2003, and including the entire vehicle fleet by model year 2006. The upgrade also included the 5th percentile Hybrid III dummy with a maximum allowable chest deflection of 52 mm [88].

Soft tissue injury is compression and rate dependent [81]. Lau and Viano in impact studies over the liver of anesthetized rabbits, found that when maximum compression (C_max) was held to 16 %, liver injury increased as impact velocity increased from 5 to 20 m/s [89]. In frontal thoracic impacts to anesthetized rabbits at 5, 10 and 18 m/s, severity of lung injury was found to increase with C_max at each level of velocity [90]. The alveolar region was more sensitive to the rate of loading than regions of vascular junctions. Data from 123 frontal impacts to anesthetized rabbits were used to define the viscous tolerance [91]. This lead to the development of the viscous criterion: VC_max, the maximum product of velocity of deformation and compression is an effective predictor of injury risk, and is a measure of the energy dissipated by the viscous elements of the thorax [92].

Kroell et al. verified the validity of the viscous criterion in blunt thoracic frontal impacts to anesthetized swine [30, 93]. In these studies, 23 swine (53.3 kg average mass) were impacted at 15 and 30 m/s with a 4.9 kg striker mass with a 150 mm diameter striker plate [30]. According to logistic analysis, VC_max and V_maxC_max were good predictors of the probability of heart rupture and thoracic MAIS > 3, while C_max was not. Lau and Viano concluded that the viscous criterion is the best indicator for soft tissue injury to many body regions for velocities of deformation of 3–30 m/s [81]. The velocity of impact of automobile occupants to various parts of the automobile interior is in this 3–30 m/s range, which is intermediate to the high velocity pressure waves of a pure blast (in which injury occurs with little compression) and the pure crushing injuries of quasi-static loading (in which injury is due to compression alone). In an analysis of the 39 unembalmed PMHS sternal impacts (average age 62 years) performed by Kroell and others, VC_max of 1.3 m/s was the value for 50 % probability of thoracic AIS > 3, and VC_max of 1.0 m/s for a 25 % probability of AIS > 3, based on Probit analysis [81, 92].

In 1998 the NHTSA proposed Combined Thoracic Index (CTI) as an injury criterion for frontal impact that combined chest compression and acceleration responses, with the aim of addressing both air bag and belt loading [94].
Where, A_max = Maximum observed acceleration

A_int = Maximum allowable intercept value of acceleration

D_max = Maximum observed deflection

D_int = Maximum allowable intercept value of deflection

The principles of CTI are explained by the differences in loading of the thorax by belt versus bag systems. For a given load, a belt system would apply greater pressure along its contact area than an air bag system, which has a larger contact area. The chest is more vulnerable to injury under the more concentrated belt loading. With the combined belt/bag system, the predominant loading could range from stiff belt/soft bag which would produce predominantly a line load, to soft belt/stiff bag which would produce a more distributed load. CTI was proposed as a criterion that reflects both the conditions and the various combinations in between. Peak chest acceleration is a measure of the magnitude of total forces applied to the torso, in proportion to the mass of the torso. Chest deflection is an indication of the belt loading contributing to the restraint system effect. The greater the deflection per unit of acceleration, the more the relative contribution of the belt system [94]. Figure 13.7a, b show the cross plot of chest deflection to chest accelerations for different AIS levels and the proposed regulation limits for CTI.

CTI was developed based on 71 PMHS tests reported by Kuppa et al. [95]. Five different restraint systems were used: 3-point belt, 2-point belt/knee bolster, 3-point belt/air bag, air bag/knee bolster and air bag/lap belt. Impact velocities ranged from 23 to 57 km/h. Peak g’s in the deceleration profiles ranged from 15 to 25 g’s. Instrumentation included tri-axial accelerometers at T1 and chestbands wrapped around the chest at the location of the 4th and 8th ribs. The 3 ms clip T1 resultant acceleration, maximum chest deflection (D_max), V_max and VC_max were analyzed statistically as injury functions using logistic regression analysis of the probability of thoracic AIS of 3 or greater. Univariate and multivariate analyses of the linear combination of these responses were performed. The combination with the best predictive value was -6.43 + 0.076As + 13.68D_max. The probability of AIS 3 or greater injury using this function is shown in Fig. 13.8. CTI was not incorporated into FMVSS 208 after review and comment to the Notice of Proposed Rulemaking (NPRM). However, stricter criteria for chest deflection were incorporated into FMVSS 208 commencing in model year 2003 as described above under Compression Criteria [88].

In a Canadian study of injury to 121 belted occupants, Dalmotas found that shoulder/chest injuries constituted 23 % of AIS 3 and greater injuries [96]. For those drivers who did not contact the steering assembly, injuries were skeletal fractures of clavicle, sternum and ribs in the belt line. No intra-thoracic or abdominal injuries were attributed to belt loading. Patrick and Anderson reported injury results in crash investigations of drivers wearing three-point belts in Volvos [97]. Fourteen of 169 (8.3 %) occupants had rib fractures. Barrier equivalent velocity (BEV) ranged from 1 to 24 m/s. At 0 to 4 m/s, 29 of 32 occupants had no injury. There were 14 rib cage injuries in the 1–24 m/s BEV. Schmidt et al. reported thoracic injury data in a test series with 49 fresh PMHS [98]. Tests were conducted at a crash velocity of 50 km/h and a stopping distance of about 40 cm. Three-point retractor belts were used on 30 and two-point belts plus knee-bar were used on 19 PMHS. Age range was 12–82 years. The fracture patterns followed the belt line and there was an approximate linear relationship between the number of rib fractures and age, with more than 20 fractures occurring in several cases. Injuries to internal organs of chest and abdomen were also documented.

Horsch et al. studied the APR and Volvo real world crash studies relating shoulder belt loading to chest injuries [99]. The analysis of the data indicated that belt loading of the Hybrid III dummy chest resulting in 40 mm chest compression represented about a 25 % risk of AIS 3 or greater thoracic injury. Injury due to belt loading appeared to be driven by chest compression rather than VC_max. In APR and Volvo test data rib fractures increased with age but more so in the APR data base. Crashes severe enough to result in more than 40 mm Hybrid III dummy chest compression were infrequent. Cesari and Bouquet used a diagonal shoulder belt to dynamically load the chest of unembalmed PMHS and dummies [100]. Using two chestbands and an array of linear displacement transducers, the authors determined that the dummy thorax is almost twice as stiff as that of the cadaver. The human thorax had complex deformations with maximum occurring away from the sternum at the lower region of the rib cage.

In 1995 the first generation programmed restraint system (PRS) was introduced in France with seat belt force threshold of 6 kN. Thirty-seven frontal accident cases involving this restraint system were investigated by Bendjellal et al. [101] Analysis of this data showed only two cases were related to AIS 3 level injury but the authors concluded that reducing the shoulder belt force to 4 kN was a necessary further step and required redesign of the air bag. The new system was called PRS II. The PRS II uses a pyrotechnic buckle pretensioner with a 4 kN belt load limiter and air bag deployment to the sides and from the top to bottom in order to reduce risk in out-of-position situations. Foret-Bruno et al. sought to establish a thoracic injury risk function for belted occupants as a function of age and loads applied to the shoulder level [102]. Fifty-percent probability of AIS 3+ occurred at a belt load of 6.9 kN. The authors concluded that a belt load limit of 4 kN combined with a specially designed air bag could protect 95 % of those involved in frontal impacts from AIS 3+ chest injuries. Foret-Bruno et al. compared thoracic injury risks for two occupant populations: belted occupants involved in accidents in which the vehicle was not equipped with a load limiter (378 cases with pyrotechnic pretensioners), and belted occupants involved in accidents in which the vehicles were equipped with 4 or 6 kN load limiters and pyrotechnic pretensioners (347 cases) [103]. A 4 kN load limit resulted in a reduction of thoracic injury risk for all AIS levels, compared to other belt configurations. There was 50–60 % reduction for AIS 2+, 75–85 % for AIS 3+ and a complete absence of AIS 4+ with a 4 kN load limiter.