Side Impact Rib Fracture Injury Analysis

A statistical analysis of injury outcome and biomechanical response was performed using data from 28 left side impact tests employing Heidelberg-type sleds and post-mortem human subjects, with the objective of advancing the development of thoracic injury criteria for lateral impact. Injuries were scored by the test centers according to AIS 90. Rib fractures accounted for the maximum AIS score in each case. Curvature data from chest band gauges were used for calculation of contours depicting the shape of the thorax at 1 ms time intervals following impact. Thoracic deformations were deduced from the contours. Risk factors studied included maxima of curvature, deflection, rib and spinal accelerations, the Thoracic Trauma Index (TTI), and Average Spinal Acceleration (ASA). Subject age at death was found to have a significant effect on injury outcome. Consequently, age was used as a confounder variable in logistic regressions for the prediction of dichotomous outcomes P(AIS≥3) and P(AIS≥4). Stepwise backward logistic regression indicated that subject age and maximum normalized curvature relative to initial curvature are the only surviving independent variables among all considered. Separate logistic regressions employing age and a single risk factor confirmed that age combined with maximum curvature difference yields the greatest statistical significance, and the highest-ranking goodness of fit. Results also showed that by employing the logarithm of curvature difference in the logit, goodness-of-fit can be improved, and the usual problem of a poor fit at low values of risk factor is eliminated. Thoracic deflection was found to be the second highest-ranking injury correlate for side impact, ranked above TTI and ASA in its ability to predict accurately and reliably the extent of side impact thoracic injury. INTRODUCTION uman fatalities due to side impact account for approximately one-third of all traffic fatalities (Cavanaugh, et al., 1993). The majority of side impact accidents occur in cross-traffic converging at intersections at relatively low speeds (e.g., 40-60 kph). The side of a vehicle is second only to the front as the most frequent impact location. The goal of minimizing side impact injuries will become increasingly important as the use of air bags reduces fatalities in vehicle collisions. A great deal of attention is now being given to the effective design and implementation of side air bags. H Injury Biomechanics Research 230 An automobile safety standard for side impact was established in October 1970, as the addition of Federal Motor Vehicle Safety Standard (FMVSS) 214, Side Impact Strength – Passenger Cars. This standard focused on increasing side door strength to minimize intrusion into the passenger compartment, and incorporated a quasi-static load test using a rigid cylinder placed against the side of a vehicle. FMVSS 214 resulted in the introduction of side door beams in all passenger cars by auto manufacturers. By the late 1970’s, it was realized that while side door beams were effective in reducing accidental death in side impact collisions with fixed objects, they proved inadequate alone in impeding intrusion into the passenger compartment in a severe impact with another vehicle. In October 1990, a new rule was appended to the Code of Federal Regulations, imposing an additional dynamic requirement, FMVSS 214: Side Impact Protection. This new rule set forth specific requirements for a dynamic test procedure simulating a 90° impact on a moving vehicle, to include measurements of acceleration at various locations on instrumented crash test dummies. The simulation of an automobile traveling at 48 kph (30 mph) impacting the left side of a target vehicle at 90° with speed 24 kph (15 mph) is accomplished by means of a “moving deformable barrier” (MDB), a 4-wheeled assembly of standard design, impacting a stationary test vehicle. Specially designed Side Impact Dummies (SID) are positioned in front and rear occupant positions on the side of the vehicle being impacted. The test data measured from SID instrumentation include the rib, spine and pelvic accelerations, which must not exceed certain threshold values for compliance with FMVSS 214. The rib and spinal accelerations are combined into a single measure denoted as the Thoracic Trauma Index, TTI(d). Specifically, TTI(d) is given by 1 ( ) ( ) 2 R LS TTI d G G = + where GR is the maximum of peak accelerations of the lower and upper rib, and GLS is the lower spinal (T12 vertebra) peak acceleration. The value of TTI(d) cannot exceed 85g for 4-door vehicles and 90g for 2-door vehicles. The pelvic acceleration is assigned an upper limit of 130g. TTI was developed in the 1980’s, as a result of analyses by Eppinger (1984) and Morgan (1986) of dynamic, kinematic and injury data from 84 sled tests employing post-mortem human subjects. The present U.S. safety standard for side impact does not include a reference to either lateral or frontal thoracic deflection. In the late 1970’s, Stalnaker et al. (1979) and Terriere et al. (1979) analyzed force-deflection data in a series of lateral drop tests onto unpadded and padded force plates, using post-mortem human subjects. Their conclusion was that chest compression correlated better with injury than thoracic acceleration. The European standard for side impact, EU Directive 96/27/EC, is different from FMVSS in several ways. The EU standard makes no attempt to simulate movement of the target vehicle, and employs the EUROSID dummy as a test subject. Furthermore, the European standard includes upper limits for the head injury criterion (HIC), the viscous criterion (VC), abdominal and pelvic forces, and rib deflection. An alternate thoracic injury criterion was proposed by Cavanaugh et al. (1993), as the result of a study of 17 Heidelberg-type sled tests using unembalmed cadavers. Cavanaugh found that chest compression and ASA gave better agreement with observed injury data than acceleration and force-based criteria. ASA is the average slope of the velocity vs. time signature, taken over a specified time interval, where velocity is obtained by integrating the measured T12 lower spinal lateral acceleration. In a study by Pintar et al. (1997), utilizing data from 26 Heidelberg-type side impact sled tests performed with post-mortem human subjects, it was found that TTI was a better predictor of thoracic injury than ASA or chest deflection. However, in a recent evaluation by Kuppa, et al. (2000) of 34 side impact sled tests, an extension of the study by Pintar, it was concluded that maximum normalized chest deflection and upper spinal resultant acceleration gave a better fit to injury data than TTI or ASA. Side Impact Rib Fracture Injury Analysis 231 It is apparent that there are conflicting findings with regard to what constitutes a suitable injury correlate for thoracic injury due to frontal or side impact. The acceleration-based criteria referenced in FMVSS 208 and FMVSS 214 are deemed by many to not have a firm biomechanical basis. The overwhelming majority of injuries sustained by the thorax in automobile accidents are rib fractures. Variables that can be related in some way to the fracture stress of bone can be considered to be biomechanically-based, such as chest deflection and curvature. These variables are the focus of the present study, which utilizes the same data employed by Kuppa, et al. (2000), but explores a wider range of biomechanically-based risk factors. Of particular interest is the curvature (or more precisely, the change in curvature relative to the initial value) of horizontal cross-sections of the thorax, a quantity that can be measured directly by means of utilizing existing chest band instrumentation. TEST DATA DESCRIPTION Side impact sled tests using post-mortem human subjects were performed at the Medical College of Wisconsin (MCW) and also at NHTSA’s Vehicle Research Test Center at Ohio State University (OSU). The sled apparatus at both test centers is of the Heidelberg design (Kallieris, et al. 1981), configured for left side impacts against rigid and padded walls. The post-mortem subjects in the MCW tests were unembalmed fresh and frozen human cadavers, while those at OSU were fresh cadavers. Cardiovascular systems of the subjects were pressurized to approximate invivo conditions. Pulmonary systems were pressurized prior to impact, but left open to atmospheric pressure subsequent to the impact event. Additional details may be found in Kuppa et al. (2000) and Pintar (1996). Both test centers utilized chest band instrumentation on test subjects, consisting of two 40channel chest bands at levels of rib 4 and rib 7, for measurement of curvature at approximately every 2.5 cm around the chest perimeter. The local measurement of curvature allows the time dependent shape of a transverse cross-section of the thorax to be determined. Other thoracic-based instrumentation included: (1) triaxial accelerometers affixed to T1 or T2 vertebra, T12 vertebra, and the sacrum, and (2) uniaxial accelerometers affixed to left region of ribs 4 and 8 to measure lateral acceleration, and to the pelvis for determination of anterior-posterior acceleration. Test subjects were examined and radiographed before and after testing, and necropsied subsequent to testing to identify both hard and soft tissue injury. The injury severity was quantified in accordance with the AIS 90 standard (Abbreviated Injury Scale, 1990). Most of the trauma to the thorax consisted of multiple rib fractures with occasional hemopneumothorax (pleural tears caused by fractured ribs). According to AIS 90, AIS=1 is characterized by only one rib fracture, AIS=2 results from 2-3 fractures, and AIS=3 is assigned when there are more than 3 fractures on only one side of the ribcage. A score of AIS=4 is assigned when there are more than 3 fractures on each side. The presence of a hemopneumothorax or a flail chest increases the AIS score by 1. Side impact tests were performed with two impact velocities, 24 kph and 32 kph, and with several target configurations: flat rigid wall, flat wall with 10 cm of Ethafoam LC200 padding, rigid or padded wall with pelvic load plate offset by 12 cm to simulate an armrest. The test matrix employed in the present study is shown in Table 1. There are 4 female subjects among a total of 28 sled tests. Approximately two-thirds of the tests were performed at MCW. An earlier study by Kuppa et al. (2000) considered a slightly larger set of this same test group, 34 tests from MCW and OSU. The present work is an extension of that study. A detailed examination of injury records in the NHTSA Biomechanics database indicated that for 5 of the 34 tests, pre-existing rib fractures were indicated in pretest examinations of post-mortem subjects. Those tests have been excluded from the present analysis. In addition, a sixth test was identified by NHTSA as having questionable data, and was also excluded. Injury Biomechanics Research 232 The average age at death among the 28 test subjects in Table 1 is 67.5 ± 13.6 years, with the youngest subject age 27, and the oldest age 84. The average age at death of subjects sustaining AIS ≥ 3 injury is 70.9 ± 9.2, while the average for those with AIS < 3 is only 56.5 ± 16.4. This implies that age is a confounder (an independent variable that is associated with both the dependent variable and a risk factor under consideration) among risk factors influencing thoracic injury due to side impact, a result also arrived at by Kuppa et al. (2000). Table 1. Side Impact Sled Tests with Post-Mortem Human Subjects Test Sex Age Mass (kg) Test center Test config. AIS rbfx 3120 M 73 89 MCW RLF 4 15 3122 M 27 72 MCW RLF 0 0 3155 M 55 76 MCW RLF 3 11 3276 M 70 71 MCW PLF 0 0 3277 M 56 64 MCW PLF 2 2 3278 M 50 93 MCW PHF 2 3 3320 F 82 74 OSU PHF 4 33 3321 F 75 42 OSU PHF 4 25 3322 M 73 72 OSU RHF 4 12 3323 F 59 81 OSU PHF 4 21 3324 M 77 75 OSU RHF 4 34 3325 M 63 61 OSU RHF 4 16 3422 M 44 83 MCW RHF 2 3 3423 M 49 62 MCW RHF 4 5 3535 M 78 88 MCW RLO 4 13 3536 M 84 76 MCW RLO 4 15 3537 M 79 93 MCW RHO 3 12 3538 M 74 77 MCW RHF 5 22 3580 M 75 56 OSU PHF 3 16 3586 M 79 146 OSU PHF 4 20 3587 M 63 100 MCW PHF 4 17 3588 M 72 66 MCW PLF 4 10 3589 M 67 76 MCW PHF 3 11 3661 M 74 51 MCW PLO 3 4 3662 M 59 73 MCW PLO 2 2 3664 F 67 74 MCW RLF 0 0 3700 M 86 67 MCW RLF 3 9 3719 M 79 53 MCW PLF 0 0 Test center: MCW=Medical College of Wisconsin, OSU= Ohio State University (Vehicle Research Test Center); Test configuration: R=rigid wall, P=padded wall, L=24kph impact, H=32kph impact, F=flat wall, O=offset wall; AIS=Abbreviated Injury Scale (1990); rbfx=number of rib fractures Side Impact Rib Fracture Injury Analysis 233 Table 1 indicates that about two-thirds (20) of the subjects sustained AIS ≥ 3 injury, while one-half (14) sustained an even higher level of AIS ≥ 4 injury. Hence, half of the subjects had 3 or more rib fractures on each side of the ribcage. Multiple fractures on the same rib were a common occurrence, with one subject incurring 34 rib fractures. It should be noted here that AIS does not distinguish between minor (hairline) rib fractures and displaced (disconnected) fractures. The AIS score for all of the 28 tests is attributable to rib fracture, and is the same as the maximum AIS (among all body regions) for a subject, which is denoted as MAIS. METHODS The selection of risk factors, independent variables that might have an effect on the probabilistic outcome variable (thoracic injury), is limited by the available measured data, or quantities that can be derived from them. The measured data consist of local curvature of the thoracic cross-section, impact load forces, and acceleration measured at the sternum, ribs, spine, sacrum and pelvis. The curvature data can be used to calculate deflection, a biomechanically-based risk factor of major interest. Deflection can be calculated in a relatively straightforward manner from chest band contours. The RBANDPC software module developed by NHTSA and included in NHTSA’s SIMon (Simulated Injury Monitor) computer program (Bandak, et al., 2001) was used to calculate chest band contours from the curvature data measured for each band. Contours were calculated at 1 ms intervals starting at 2 ms prior to the time of impact, over a total interval of 200 ms for each chest band. The time of impact was deduced from direct observation of raw curvature data, and from sample contours obtained for each test. A visual inspection of curvature and of each contour generated served as a useful check on the integrity of the data. Chest bands consist of 40 curvature gauges placed approximately 2.5 cm apart, along the entire length of a band. The bands are made with a fixed length to accommodate all possible test subjects, so some overlap is always present. Curvature gauges in the overlapped region were ignored in this analysis. Data from faulty gauges were omitted from the chest band calculations, an acceptable practice which merely decreases spatial resolution. The curvature data were not filtered by the post-processing software. Sample chest band contours for a typical left side impact are illustrated in Figure 1, where contours are shown at 6 ms intervals. The spine (“o” symbol) is located near the bottom center of the figure, and the sternum (“x” symbol) near the top center. The origin of each contour (X=0, Y=0) is chosen arbitrarily as the point at which the band crosses the spine, where X is the lateral direction and Y is the anterior-to-posterior direction. A lateral deflection deemed to be representative of side impact is defined as the maximum of the change in distance between three pairs of points located at specified fractional distances along the band length, proceeding in a clockwise direction and starting at the location of the spine. The three point pairs are defined at 20%-80%, 25%-75% and 30%-70% of the circumferential distance along the band, as shown in Figure 2. This definition of “lateral” deflection, which will be denoted as variable dmaxn, is consistent with that employed by Kuppa et al. (2000), and will also be referred to here as a “6-point” deflection. A value of dmaxn is calculated for each chest band contour, and is taken as the larger of the two values over all time for the upper and lower chest bands. Hence, dmaxn is the maximum of 6 distances, each representative of a local lateral deflection of the full thoracic cross-section. The complete set of risk factors evaluated for possible use as thoracic injury correlates is given in Table 2. In addition to dmaxn, three alternate representations for deflection are listed in Table 2: (1) maximum “single-point” deflection at the anterior-posterior left side mid-point (latdefln), (2) maximum “point-by-point” deflection along all points on all contours (crdefln), and (3) spatial average of maximum deflection over time at each point on contours (avdefmxn). The maximum “point-by-point” deflection is taken as the maximum of the largest distance between a Injury Biomechanics Research 234 point on a contour at time t and its initial position at t = 0, among all points on all contours over all time. Variable avdefmxn is a spatial average taken over the length of contours, of the maximum distance traversed by each point on a contour. This maximum distance will generally be nonzero for each point along a contour. Since avdefmxn is the average over all such points of this maximum, it accounts for deflection in all directions. Due to the arbitrary position in space of each contour relative to the preceding contour, a lateral shift in Y-coordinates is performed prior to the calculation of variables crdefln and avdefmxn, such that the anterior-to-posterior “mid-point” on the left side aligns on each successive contour. This mid-point is computed at the initial time and is tracked with each contour at later times. The mid-point is defined simply as the leftmost point on the contour just prior to the initial time of impact. It is acknowledged that the shift in Y is chosen arbitrarily, and that the actual shift is unknown. Current chest band technology allows determination of the cross-sectional shape of the thorax, but not the position of the contour in space. Table 2 also shows that the complete set of risk factors investigated includes spinal, rib and pelvic accelerations. However, it should be noted that acceleration is not considered to be a biomechanically-based risk factor for thoracic injury, since it alone cannot lead to the determination of the maximum stress experienced by the ribcage and internal organs of the thorax. Acceleration was included here for the sake of completeness and for comparison with results of similar studies. Figure 1: Chest band contours for lateral impact. Side Impact Rib Fracture Injury Analysis 235 Figure 2: Locations along chest band for lateral deflection computation. Table 2. Risk Factors Evaluated as Injury Correlates in Side Impact Analysis