- Email: [email protected]
Injury severity of restrained front seat occupants in car-to-car side impacts
Accid. Anal. and F’rev., Vol. 27, No. 1, pp. 105-110, 1995 Copyright 0 1995Elsevier Science Ltd Printed in the USA. All rights reserved 0001-4575/95$9.50 + .OO
Pergamon 0001_4575(94)E0019-H
INJURY SEVERITY OF RESTRAINED FRONT SEAT OCCUPANTS IN CAR-TO-CAR SIDE IMPACTS* E.
MILTNER
and
H.-J.
SALWENDER
Institute of Forensic Medicine, University of Heidelberg, Vossstrasse 2, 69115 Heidelberg, Germany (Accepted
2 November
1993)
Abstract-The influence of a number of factors, including age and particularly seating position, on the injury severity of restrained occupants was examined for 41 front-seat occupants seated adjacent to the impact (near side) and 38 sitting opposite the impacted side (far side) in car-to-car side collisions (center of impact: front door and B-pillar). Above an energy equivalent speed of 40 km/h all near-side occupants and about half of the far-side occupants sustained severe injuries. A logistic regression analysis showed that within range of 30-60 km/h (delta u 20-60 km/h) the probability of severe injuries increased dramatically from approximately 20% to more than 90%; in these cases, far-side occupants had the same injury probability as near-side occupants only when the speed was 10 km/h higher. The main cause of death for 27 occupants seated on both sides was polytrauma, this was accompanied in two-thirds of the cases by serious head injuries. The second most frequent cause of death was head injury.
INTRODUCTION
verity and injury severity. The total sample comprises 670 occupants in 428 cars in real front and side collisions. Of the 272 occupants hit in side collisions, 79 were front seat occupants restrained by seat belts. Car-to-car impacts occurred somewhere between the middle of the front door and the B-pillar. Accident data were taken from the files of the road accident division of Heidelberg’s police department for the years 1987 to 1990 and from the autopsy protocols of the University of Heidelberg’s Institute for Forensic Medicine from 1983 to 1990. Accidents were reconstructed by the usual methods. In fatal accidents the cars involved were normally examined. The total vehicle masses were calculated by combining their weight in a “roadworthy condition” with the estimated weight of the occupants (males 75 kg, females 60 kg). As it was not possible to determine in each case the fuel tank contents and vehicle load, an error of +I00 kg is possible. Deformation depth and the degree of overlapping were estimated by photogrametric means and from written damage descriptions; they were measured only in exceptional cases. The error made in estimation may amount to between 10% and 20% in cases of deformation less than 10 cm, and between 5% and 10% for deformations greater than 10 cm. The entire range of impact was not stated in side collisions, but rather the center of impact on either the driver’s or passenger’s sides. The centers
Depending on which sample of car accidents is studied, the proportion of side collisions in all car accidents lies between 15% and 40% (Otte 1982, 1990; Rouhana and Foster 1985; Langwieder and Hummel 1987, Ropohl 1990). Side collisions are particularly dangerous for near-side occupants (i.e. the impacted side) when the passenger cell is impacted close to the occupant’s position (Niederer, Walz, and Weissner 1980; Kallieris and Mattern 1984; Kallieris, Mattern, and HPdle 1986). Therefore attempts are being made to improve the protection of near-side occupants through passenger cell reinforcement and padding (Kallieris, Schmidt, and Miltner 1991), and as well through the employment of side airbags (Olsson and Skiitte 1989). Due to the vulnerability of near-side occupants, safety measures concentrate on near-side occupants and only indirectly on the far-side occupants (Jones 1982). This survey comments on the risk of injury for occupants on both sides and addresses the question of whether additional measures to protect far-side occupants are necessary. MATERIAL
AND METHOD
This report is part of a large-scale retrospective study of the interrelationship between accident se*Parts of this paper were presented at the 71,Jahrestagung der Deutschen Gesellschaft fiir Rechtsmedizin Berlin on 18 September 1992. 105
E. MILTNER and H.-J. SALWENDER
106
of impact were recorded as one of the following: front or rear fender, center of front or rear door, or A, B, and C pillars. The impact angle was defined relative to the longitudinal axis of the faster vehicle (OO).Thus the impact angle is the angle between the longitudinal axes of both vehicles. Impact speeds were evaluated according to the usual accident reconstruction methods, e.g. accident layouts drawn to scale (Burg and Rau 1981). The Energy equivalent speed (EES) was evaluated according to the normal methods. According to Burg and Zeidler (1980) the error made in calculation amounts to approximately 5 km/h, and in the worst case to 10 km/h. The change in velocity (delta u) resulting from vehicle impact is a variable related to occupant loading (Kramer 1980). This was also evaluated according to the usual methods (Slibar 1966; Burg 1973; Burg and Rau 1981; Burg and Zeidler 1980). The delta u calculation error is ?5 km/h. The injuries were coded according to the 1985 Abbreviated Injury Scale (AIS) due to its common use in international accident surveys (States et al. 1980). The Abbreviated Injury Scale is the most commonly used injury classification system in accident studies. AIS classifies injuries into 7 degrees of severity (from 0 = uninjured to 6 = immediately fatal) in 7 body regions (external, head, neck, thorax, abdomen, spine, extremities). The total injury severity (MAIS) is the degree of the most severe individual injury. The border between severe almost fatal and fatal injuries is at AJS 4, thus the sample was separated into two groups AIS O-3 and AIS 4-6. RESULTS The sample was 41 near- and 38 far-side occupants, i.e. 52 drivers and 27 passengers (46 male and 33 female) (Table 1). One passenger was struck as a near side occupant while sitting alone in a parked car. A total of 14 car brands were involved. On the
Table 1. Distribution Driver onlv Near-side driver Near-side passenger Far-side driver Far-side passenger
11 0 11 0
of occupants’ positions Passenger onlv 0 1 0 0
Notes: Belt-restrained front seat occupants; 17 far side in car-to-car side collisions. Center of impact: front door or B-pillar.
Two front seat oassenaers 13 16 17 10 15 near side and
near side the average EES amounted to 38.3 km/h with an average delta u of 31.5 km/h and for the far side 34.9 km/h and 27.6 km/h, respectively. The average deformation depth was 52 cm on the near side and 47 cm on the far side. In 57 cases the collision angle was between 80” and 100”. The average ratio of vehicle masses, impacting car to impacted car, was 1.14. The average occupant age was 37 near side and 35 years far side. There were 27 fatalities. The causes of death were polytrauma (12 cases), skull-brain trauma (5 cases), shock syndrome (4 cases of nonhemorrhagic shock), hemorrhage (4 cases), and spinal trauma (2 cases). The distribution of fatal injuries for far- and near-side occupants was almost equal. For fatally injured near-side occupants with the diagnosis polytrauma, 5 out of the 9 cases experienced also AIS 4-6 head injuries. With cases of fatal hemorrhaging there was: one case of thorax AIS 5, two cases of thorax AIS 6, one case of abdomen AIS 4 and two cases of abdominal AIS 5. The diagnosis shock comprises cases with traumatic shock and adult respiratory distress syndrome. For cases of fatal shock there were: one case of thorax AIS 4, two cases of thorax AIS 5, one case of abdominal AIS 4 and one case of abdominal AIS 5. Thirtysix percent of the cases with severe thorax injuries experienced severe head injuries as well. In 19 cases, survival time was less than one day, in 11 cases survival time was less than 6 hours. The longest survival time was 28 days. Figure 1 shows the distribution of the total injury severity MAIS for drivers depending on the EES and the number of front seat occupants. There is no evident difference in injury severity between those cases in which a near-side occupant is present and no near side occupant is present. Because the group numbers are too small to derive further conclusions from that result, we combined both groups for the following investigations. Figure 2 shows the MAIS distribution plotted against EE.S. When EES was below 40 km/h, severe injuries occurred only in exceptional cases. Above an EES and delta u of 40 km/h, all near-side occupants sustained severe injuries. Although injuries of AIS- are defined as severe but usually not fatal, all near-side occupants with AIS- injuries died when EES and delta u were at least 40 km/h. Under the same impact severity only half of the far-side occupants sustained severe injuries MAIS 4-6. In the EES range of 20 to 39 km/h occupants on both sides experienced approximately the same number of MAIS l-3 injuries. Figure 3 shows the probability of severe and fatal injuries (MAIS 4-6) against EES. In the multi-
107
Injury severity in car-to-car side impacts
10
-f$
8
.._...
1
i
(
8
. ..i. ._..........__..
l-3
??+AlS
.._...._.....,....
i
.............. . ...._......
.i...
~--.“““‘...--...i ._______.._................ j-.___.....................
+.. _.j. ......._____............
i
3
3
4-6
.j.. j I
3
_i_
i-
0 driver
only
driver
and
passenger
20-39 km/h driver
20-39 km/h only
driver
km/h and
passenger
driver
driver
km/h only
driver
and
only
driver
and
passenger
passenger
Fig. 1. Distributions of total injury severity MAIS for drivers depending on EES and number of front seat occupants (77 belt-restrained front seat occupants in car-to-car side collisions).
trunk injuries, followed by head injuries. Far-side occupants also receive mainly trunk injuries, followed by head and limb injuries. But even within the EES range of 20 to 39 km/h, head, trunk and spinal cord AIS 1-3 injuries can be expected for farside occupants. Table 2 shows that for selected individual injuries, often several simultaneous, but independent, factors influence the injury severity. Besides the interrelationship between injury and occupant position, near or far side, a relationship between age and bone injuries can clearly be detected.
variate logistic regression model, EES (p < .OOl) and occupant position (p < .OS>independently influence the injury severity. Within the ESS range of 30 to 60 km/h the probability of MAIS 4-6 rises steeply from approximately 20% to over 90%. The slope of this curve is similar for near- and far-side occupants although the latter is shifted 10 km/h to the right, i.e. the same EES, lower MAIS 4-6 probability. Plots for delta u 20-60 km/h are similar. Figure 4 shows injury severity distribution for all body locations during 40- to SPkm/h EES collisions. Near-side occupants receive predominantly 18
1I
I
’O-l 9 km/h ’O-l 9 km/h ‘20-39 km/h ‘20-39 km/h ‘40-59 km/h ‘40-59 kmh ’> 59 km/h ’> 59 km/h ’ near skfe far side near side near side far side near side far side far side Fig. 2. Distributions
of total injury severity MAIS depending on EES and occupant’s position (77 belt restrained front seat occupants in car-to-car side collisions).
108
E. MILTNER and H.-J. SALWENDER
Prob[MAIS4-61
Table 2. Influential factors on individual injuries
1,O
Skull fractures
0,9 -
X
EES
08 -
On impact! opposite side Aee
0,7 036 -
Liver runtures
Pelvic fractures
Rib fractures
X
X
X
X
X
X
X
X
Notes: 77 belt-restrained front seat occupants collisions. Center of impact: front door or B-pillar. Multivariate analysis, logistic regression.
085 0,4 -
in side
0,3 02 081 0,O
I 20
1 10
0
I 30
I 40
I 50
I 60
t 70
I 80
1 90 km/h
When EES, occupant position, and age are used in combination as predictors of pelvic fracture, 66 of the 77 cases (85%) were correctly predicted. These were 65% of the cases with pelvic fractures and 93% of those without.
EES Fig. 3. Probability of injuries MAIS 4-6 depending on EES and occupant’s position (79 belt-restrained front seat occupants in car to car side collisions. Center of impact: front door or B-pillar).
By means of logistic regression, it was possible to calculate a prediction formula for pelvic fractures: -6,5293074
+ 0,08094332 x EES + 0,0452515 match
X Age +
on near side/far side:
- 1,0108663,
when “far side”
1,01086635,
when “near side”
.)
otherwise.
DISCUSSION In Fildes and Vulcan’s study (1990) of 150 side collisions, near-side front seat occupants were injured mainly by interior vehicle structures. In general, far-side occupants were less frequently injured: 36% due to contact with other occupants, 27% by interior vehicle surfaces, 18% by safety belts, and 18% by the dashboard. Rouhana and Foster (1985) described similar results. The behavior of front seat car occupants in side collisions can be studied in experimental dummy tests. In experimental 90” side collisions with beltrestrained dummies, Faerber (1982) observed that
12
0
I
’abdomen abdomen head near' head far ’ thorax ’ thorax near side far side near side far side side side
I
I
spine near spine far side side
I
extrem. near side
Fig. 4. Distribution of injury severity on all body locations at an EES of 40-59 km/h (belt-restrained front seat occupants; and 10 far side in car to car side collisions. Center of impact: front door or B-pillar).
I
extrem. far side
10 near side
Injury severity in car-to-car side impacts
the near-side dummy was thrown into the middle of the passenger cell and against the far-side dummy. The near-side dummy usually received one impact from the car door and one impact when hitting the far-side dummy, the far-side dummy received only one impact, namely by the near-side dummy. Occupant-to-occupant contact generated dummy loads that reached nearly the same level in each of the interacting body regions, but the loads were only 25% to 50% of those measured in the primary impacts. In our study, we found no increase in injury severity in far-side drivers, when a passenger was present. However, we could not differentiate between injuries caused by contact with the other front seat occupant and injuries caused by internal car structures, especially by the other front seat. The risk of life-threatening injuries for near-side occupants is two to three times higher than that on the far side according to: Riley and Radley 1976; Walz et al. 1977; Norin, Nilsson-Ehle, and Gustafsson 1982; Danner and Langwieder 1976; Rouhana and Foster 1985. Rouhana and Foster (1985) detected not only a three times higher risk of severe injuries for near-side occupants, but also observed a 1.6 times greater risk for near-side passengers than for near-side drivers. Near-side passengers sustained fatal injuries three times more often than nearside drivers. Near-side passengers bear the greatest risk of fatal injury, while far-side passengers have the lowest risk. However, most of the passengers in Rouhana’s survey had not fastened their seat belts and some of them were thrown out of the car. In our survey a clear correlation between the total injury severity and EES for occupants seated on both sides could be detected; the graph for farside occupants is shifted to the right by approximately 10 km/h. Within the critical range of 30-50 km/h for near side occupants, and 40-60 km/h for far-side occupants, the probability of severe injuries MAIS 4-6 increased from approximately 20% to more than 90%. Within the range of 40-59 km/h, starting at an EES of 40 km/h, all near-side occupants were killed; far-side occupants also sustained severe injuries, however, only half as often as the near-side occupants. Otte, et al. (1984a) detected similar results when regarding delta u with impacts over the entire occupant compartment. For near-side occupants the risk of MAIS 5-6 injury from a delta u of 30 km/h increased drastically; however, for far-side occupants this trend was far less. Rouhana and Foster (1985) found the median delta u for severely injured occupants in side collisions was 27 km/h; it was 50 km/ h for those fatally injured. In this study in the EES speed range 40-59 km/h
109
the most frequently, severely injured body locations were the trunk and head. As in the case of total injury severity, the injury risk was only half as great for far-side as for near-side occupants. These results concur with Langwieder 1975 and Fildes and Vulcan 1990; however, in their studies all speed ranges were regarded. The influence of occupant position, near and far side, on resulting injury was less for head injuries than trunk injuries. Otte et al. (1984b) found that the difference in injury frequency between the vehicle occupant positions was highest for the lower trunk (22.7% near side and 9.5% far side). Principally, far-side occupants are less endangered than those adjacent to the impact. Nevertheless, within the critical EES range of 40-50 km/h there is still a risk of severe injury for far-side occupants-25% compared to 50% for the near side. Thus all seating positions must be considered when occupant protection measures are formulated. Without precisely defining the side-impact location, McCoy, Johnstone, and Kenwright (1989) observed pelvic fractures in near- and far-side occupants, which arose at a delta u as low as 15 km/h. With regard to the EES, pelvis fractures occurred on the near side at an EES less than 20 km/h and on the far side at an EES of 40 km/h. Our survey revealed that the probability of pelvic fractures, as well as other bone injuries, does not depend solely on the energy involved in the accident, but as well on occupant age. This observation corresponds to large scale statistical surveys (Evans 1985, 1988) and also to experimental biomechanical tests (Vetre 1977; Mattern et al. 1979; Schmidt 1979). Evans 1988 determined that the risk of sustaining fatal injuries is lowest at 20 years of age and three times higher at 70 years. Thus the validation of occupant protection measures should also take occupant age into consideration. Viano et al. 1990 detected that 64% of near-side occupants were over 50 years old and 36% over 70 in fatal multivehicle side impacts. This result contrasts the young age of the occupants in our sample (37 years near side, 35 years far side). The difference originates from the kind of the samples. Viano’s work is an epidemiological study, while we tried to collect primarily severe accidents. In this study a regression analysis was performed to predict the probability of pelvic fracture based on EES, age, and occupant position. This relationship predicted correctly the existence of pelvic fracture in 85% of the cases. Thus, through the employment of extensive real accident analyses, it is possible to develop predictors for injuries and injury severity with a high degree of accuracy.
110
E. MILTNERand H.-J. SALWENDER
Acknowledgements-Many thanks to the following for the reconstruction of accidents: Klaus Peter Wiedmann, Burkhard Leutwein, Hans Peter Hepp, and Rainer Fischer.
REFERENCES Burg. H. Das Antriebs-Balance-Diagramm als optimales Der Verkehrsunfall, Heft 2 Mittel der Unfallanalyse.
(1973). Burg, H.; Rau, H. Handbuch der Verkehrsunfallrekonstruktion. Verlag Information Ambs GmbH, Kippenheim (1981). Burg, H; Zeidler, F. EES-Ein Hilfsmittel zur Unfalrekonstruktion und dessen Auswirkungen auf die Unfalforschung. Der Verkehrsunfall Heft 4 (1980). Danner, M.; Langwieder, K. Car/vehicle side impacts-a study of accident characteristics and occupant injuries Proceedings of the 6th International Conference on Experimental Safety Vehicles, 1976; 516-540. Danner, M.; Langwieder, K.; Hummel, T. H. Experience from the analyses of accidents with a high belt usage rate and aspects of continued increase in passenger safety. 1 lth International Technical Conference on Experimental Safety Vehicles, Washington, May 12-15, 1987; 201-210. Evans, L. Involvement rate in two-car crashes versus driver age and car mass of each involved car. Accid. Anal. Prev. 16:387-405; 1985. Evans, L. Risk of fatality from physical trauma versus sex and age. J. Trauma 28:368-378; 1988. Faerber, E. Interaction of car passengers in frontal, side, and rear collisions. Proceedings of the 26th Stapp Conference, Ann Arbor, MI, October 20-21, 1982; 113, 335-352. Fildes, B. N.; Vulcan, A. P. Crash performance and occupant safety in passenger cars involved in side impacts. International Conference on the Biomechanics of Impacts (IRCOBI), Bron-Lyon, France, 1990; 121-140. Jones, I. S. Injury severity versus crash severity for front seat car occupants in front and side impacts. Proceedings of the 26th Annual Meeting of the American Association for Automotive Medicine, Ottawa, Canada, October 4-6, 1982. Kallieris, D.; Mattern, R. Belastbarkeitsgrenzen und Verletzungsmechanik der angegurteten Fahrzeuginsassen beim Seitaufprall (1). FAT-Schriftenreihe Nr. 36 (1984). Kallieris, D.; Mattern, R. ; HLdle, W. Belastbarkeitsgrenzen und Verletzungsmechanik der angegurteten Fahrzeuginsassen beim Seitaufprall(2). FAT-Schriftenreihe Nr. 60 (1986). Kallieris, D.; Schmidt, G.; Miltner, E. Verminderung der Verletzungsschwere kollisionsnaher Frontinsassen bei der 90 Grad Pkw/Pkw-Seitenkollision durch fahrzeugkonstruktive Massnahmen. Hefte Unfall- und Sicherheitsforschung Strassenverkehr, Kongressbericht der Deutschen Gesellschaft fur Verkehrsmedizin e.V.; 202-205 (1991). Kramer, F. Einfluss der mittleren Fahrzeugverzdgerung auf die Insassenbelastungswerte bei der Frontalkollision. Automobil-Industrie 1, 115 (1980). Langwieder, K. Aspekte der Fahrzeugsicherheit anhand einer Untersuchung von realen Unfallen. Dissertation. Berlin; 1975. Mattern, R.; Barz, J.; Schulz, F.; Kallieris, D.; Schmidt,
G. Problems arising when using injury scales in biomechanical investigation with special consideration of the age influence. Proceedings of the 4th IRCOBI Conference, Goteborg, Sweden, September 5-7, 1979; 223-23 1. McCoy, G. F.; Johnstone, R. A.; Kenwright, J. Biomechanical aspects of pelvic and hip injuries in road traftic accidents. J. Orthop. Trauma 3: 113-123; 1989. Niederer, P.; Walz, F.; Weissner, R. Verletzungsursachen beim Pkw-Insassen, Verletzungsminderung durch modeme Sicherheitseinrichtungen. Unfallheilkunde 83, 326-340; 1980. Norin, H.; Nilsson-Ehle, A.; Gustafsson, C. Volvo traffic accident research system as a tool for side impact protection evaluation and development. Proceedings of the 9th International Conference on Experimental Safety Vehicles, Kyoto, November 1-4, 1982; 439-448. Olsson, J. A.; Skotte, L.-G. Air bag system for side impact protection. Proceedings of the 12th Conference on Experimental Safety Vehicles, Goteborg, Sweden, September 5-7, 1989. Otte, D.; Ktihnel, A.; Suren, E. G.; Weber, H.; Gotzen, L.; Schockenjoff, G.; Vu Han, V. Erhebungen am Unfallort. Unfall- und Sicherheitsforschung Strassenverkehr 37, l-100 (1982). Otte, D.; Suren, E. G.; Appel, H.; Nemzow, J. Vehicle parts causing injuries to front-seat car passengers in lateral impact. Proceedings of the 28th Stapp Conference, Chicago, 1984a; 13-24. Otte, D.; Suren, E.-G.; Appel, H.; Nehmzow, J. Bewertung von Einflussparametern fur die Verletzungen von Pkw-Insassen bei Seit-Kollisionen. Verkehrsunfall und Fehrzeugtechnik, 41-48 (1984b). Riley, B. S.; Radley, C. P. Traffic accidents to cars with directions of impact from the front and side. Proceedings of the 6th International Conference on Experimental Safety Vehicles, Washington, October 12-15, 1976; 563-573. Ropohl, D. Die rechtsmedizinische Rekonstruktion von Verkehrsunfallen. DAT-Schriftenreihe Technik, Markt, Sachverstandigenwesen, Band 5,90-92 (1990). Rouhana, S. W.; Foster, M. E. Lateral impact-an analysis of the statistics in the NCSS. Proceedings of the 29th Stapp Conference, 1985; 79-98. Slibar, S. Die mechanischen Grundsatze des Stossvorganges und ihre Anwendung. Archiv fur Unfallforschung, GUFU Baden-Baden (1966). Schmidt, G. The age as a factor influencing soft tissue injuries. Proceedings of the 4th IRCOBI Conference, Goteborg, September 5-7, 1979; 143-150. States, J. D.; Huelke, D. F.; Baker, S. P.; Bryant, R. W. et al. The abbreviated injury scale (AIS), 1985 revision. Arlington Heights, IL: American Association for Automotive Medicine; 1985. Vetre, T. Die dynamische Biegefestigkeitsprtifung von menschlichen Rippen. Med. Diss., Heidelberg (1977). Viano, D. C.; Culver, C. C.; Evans, L.; Frick, M. Involvement of older drivers in multivehicle side-impact crashes. Accid. Anal. prev. 22:177-188; 1990. Walz, F.; Niederer, P.; Zollinger, U.; Renfer, A. belted occupants in oblique and side impacts. Proceedings of the 3rd IRCOBI Conference, Berlin, September 7-9, 1977; 194-205.
Pergamon 0001_4575(94)E0019-H
INJURY SEVERITY OF RESTRAINED FRONT SEAT OCCUPANTS IN CAR-TO-CAR SIDE IMPACTS* E.
MILTNER
and
H.-J.
SALWENDER
Institute of Forensic Medicine, University of Heidelberg, Vossstrasse 2, 69115 Heidelberg, Germany (Accepted
2 November
1993)
Abstract-The influence of a number of factors, including age and particularly seating position, on the injury severity of restrained occupants was examined for 41 front-seat occupants seated adjacent to the impact (near side) and 38 sitting opposite the impacted side (far side) in car-to-car side collisions (center of impact: front door and B-pillar). Above an energy equivalent speed of 40 km/h all near-side occupants and about half of the far-side occupants sustained severe injuries. A logistic regression analysis showed that within range of 30-60 km/h (delta u 20-60 km/h) the probability of severe injuries increased dramatically from approximately 20% to more than 90%; in these cases, far-side occupants had the same injury probability as near-side occupants only when the speed was 10 km/h higher. The main cause of death for 27 occupants seated on both sides was polytrauma, this was accompanied in two-thirds of the cases by serious head injuries. The second most frequent cause of death was head injury.
INTRODUCTION
verity and injury severity. The total sample comprises 670 occupants in 428 cars in real front and side collisions. Of the 272 occupants hit in side collisions, 79 were front seat occupants restrained by seat belts. Car-to-car impacts occurred somewhere between the middle of the front door and the B-pillar. Accident data were taken from the files of the road accident division of Heidelberg’s police department for the years 1987 to 1990 and from the autopsy protocols of the University of Heidelberg’s Institute for Forensic Medicine from 1983 to 1990. Accidents were reconstructed by the usual methods. In fatal accidents the cars involved were normally examined. The total vehicle masses were calculated by combining their weight in a “roadworthy condition” with the estimated weight of the occupants (males 75 kg, females 60 kg). As it was not possible to determine in each case the fuel tank contents and vehicle load, an error of +I00 kg is possible. Deformation depth and the degree of overlapping were estimated by photogrametric means and from written damage descriptions; they were measured only in exceptional cases. The error made in estimation may amount to between 10% and 20% in cases of deformation less than 10 cm, and between 5% and 10% for deformations greater than 10 cm. The entire range of impact was not stated in side collisions, but rather the center of impact on either the driver’s or passenger’s sides. The centers
Depending on which sample of car accidents is studied, the proportion of side collisions in all car accidents lies between 15% and 40% (Otte 1982, 1990; Rouhana and Foster 1985; Langwieder and Hummel 1987, Ropohl 1990). Side collisions are particularly dangerous for near-side occupants (i.e. the impacted side) when the passenger cell is impacted close to the occupant’s position (Niederer, Walz, and Weissner 1980; Kallieris and Mattern 1984; Kallieris, Mattern, and HPdle 1986). Therefore attempts are being made to improve the protection of near-side occupants through passenger cell reinforcement and padding (Kallieris, Schmidt, and Miltner 1991), and as well through the employment of side airbags (Olsson and Skiitte 1989). Due to the vulnerability of near-side occupants, safety measures concentrate on near-side occupants and only indirectly on the far-side occupants (Jones 1982). This survey comments on the risk of injury for occupants on both sides and addresses the question of whether additional measures to protect far-side occupants are necessary. MATERIAL
AND METHOD
This report is part of a large-scale retrospective study of the interrelationship between accident se*Parts of this paper were presented at the 71,Jahrestagung der Deutschen Gesellschaft fiir Rechtsmedizin Berlin on 18 September 1992. 105
E. MILTNER and H.-J. SALWENDER
106
of impact were recorded as one of the following: front or rear fender, center of front or rear door, or A, B, and C pillars. The impact angle was defined relative to the longitudinal axis of the faster vehicle (OO).Thus the impact angle is the angle between the longitudinal axes of both vehicles. Impact speeds were evaluated according to the usual accident reconstruction methods, e.g. accident layouts drawn to scale (Burg and Rau 1981). The Energy equivalent speed (EES) was evaluated according to the normal methods. According to Burg and Zeidler (1980) the error made in calculation amounts to approximately 5 km/h, and in the worst case to 10 km/h. The change in velocity (delta u) resulting from vehicle impact is a variable related to occupant loading (Kramer 1980). This was also evaluated according to the usual methods (Slibar 1966; Burg 1973; Burg and Rau 1981; Burg and Zeidler 1980). The delta u calculation error is ?5 km/h. The injuries were coded according to the 1985 Abbreviated Injury Scale (AIS) due to its common use in international accident surveys (States et al. 1980). The Abbreviated Injury Scale is the most commonly used injury classification system in accident studies. AIS classifies injuries into 7 degrees of severity (from 0 = uninjured to 6 = immediately fatal) in 7 body regions (external, head, neck, thorax, abdomen, spine, extremities). The total injury severity (MAIS) is the degree of the most severe individual injury. The border between severe almost fatal and fatal injuries is at AJS 4, thus the sample was separated into two groups AIS O-3 and AIS 4-6. RESULTS The sample was 41 near- and 38 far-side occupants, i.e. 52 drivers and 27 passengers (46 male and 33 female) (Table 1). One passenger was struck as a near side occupant while sitting alone in a parked car. A total of 14 car brands were involved. On the
Table 1. Distribution Driver onlv Near-side driver Near-side passenger Far-side driver Far-side passenger
11 0 11 0
of occupants’ positions Passenger onlv 0 1 0 0
Notes: Belt-restrained front seat occupants; 17 far side in car-to-car side collisions. Center of impact: front door or B-pillar.
Two front seat oassenaers 13 16 17 10 15 near side and
near side the average EES amounted to 38.3 km/h with an average delta u of 31.5 km/h and for the far side 34.9 km/h and 27.6 km/h, respectively. The average deformation depth was 52 cm on the near side and 47 cm on the far side. In 57 cases the collision angle was between 80” and 100”. The average ratio of vehicle masses, impacting car to impacted car, was 1.14. The average occupant age was 37 near side and 35 years far side. There were 27 fatalities. The causes of death were polytrauma (12 cases), skull-brain trauma (5 cases), shock syndrome (4 cases of nonhemorrhagic shock), hemorrhage (4 cases), and spinal trauma (2 cases). The distribution of fatal injuries for far- and near-side occupants was almost equal. For fatally injured near-side occupants with the diagnosis polytrauma, 5 out of the 9 cases experienced also AIS 4-6 head injuries. With cases of fatal hemorrhaging there was: one case of thorax AIS 5, two cases of thorax AIS 6, one case of abdomen AIS 4 and two cases of abdominal AIS 5. The diagnosis shock comprises cases with traumatic shock and adult respiratory distress syndrome. For cases of fatal shock there were: one case of thorax AIS 4, two cases of thorax AIS 5, one case of abdominal AIS 4 and one case of abdominal AIS 5. Thirtysix percent of the cases with severe thorax injuries experienced severe head injuries as well. In 19 cases, survival time was less than one day, in 11 cases survival time was less than 6 hours. The longest survival time was 28 days. Figure 1 shows the distribution of the total injury severity MAIS for drivers depending on the EES and the number of front seat occupants. There is no evident difference in injury severity between those cases in which a near-side occupant is present and no near side occupant is present. Because the group numbers are too small to derive further conclusions from that result, we combined both groups for the following investigations. Figure 2 shows the MAIS distribution plotted against EE.S. When EES was below 40 km/h, severe injuries occurred only in exceptional cases. Above an EES and delta u of 40 km/h, all near-side occupants sustained severe injuries. Although injuries of AIS- are defined as severe but usually not fatal, all near-side occupants with AIS- injuries died when EES and delta u were at least 40 km/h. Under the same impact severity only half of the far-side occupants sustained severe injuries MAIS 4-6. In the EES range of 20 to 39 km/h occupants on both sides experienced approximately the same number of MAIS l-3 injuries. Figure 3 shows the probability of severe and fatal injuries (MAIS 4-6) against EES. In the multi-
107
Injury severity in car-to-car side impacts
10
-f$
8
.._...
1
i
(
8
. ..i. ._..........__..
l-3
??+AlS
.._...._.....,....
i
.............. . ...._......
.i...
~--.“““‘...--...i ._______.._................ j-.___.....................
+.. _.j. ......._____............
i
3
3
4-6
.j.. j I
3
_i_
i-
0 driver
only
driver
and
passenger
20-39 km/h driver
20-39 km/h only
driver
km/h and
passenger
driver
driver
km/h only
driver
and
only
driver
and
passenger
passenger
Fig. 1. Distributions of total injury severity MAIS for drivers depending on EES and number of front seat occupants (77 belt-restrained front seat occupants in car-to-car side collisions).
trunk injuries, followed by head injuries. Far-side occupants also receive mainly trunk injuries, followed by head and limb injuries. But even within the EES range of 20 to 39 km/h, head, trunk and spinal cord AIS 1-3 injuries can be expected for farside occupants. Table 2 shows that for selected individual injuries, often several simultaneous, but independent, factors influence the injury severity. Besides the interrelationship between injury and occupant position, near or far side, a relationship between age and bone injuries can clearly be detected.
variate logistic regression model, EES (p < .OOl) and occupant position (p < .OS>independently influence the injury severity. Within the ESS range of 30 to 60 km/h the probability of MAIS 4-6 rises steeply from approximately 20% to over 90%. The slope of this curve is similar for near- and far-side occupants although the latter is shifted 10 km/h to the right, i.e. the same EES, lower MAIS 4-6 probability. Plots for delta u 20-60 km/h are similar. Figure 4 shows injury severity distribution for all body locations during 40- to SPkm/h EES collisions. Near-side occupants receive predominantly 18
1I
I
’O-l 9 km/h ’O-l 9 km/h ‘20-39 km/h ‘20-39 km/h ‘40-59 km/h ‘40-59 kmh ’> 59 km/h ’> 59 km/h ’ near skfe far side near side near side far side near side far side far side Fig. 2. Distributions
of total injury severity MAIS depending on EES and occupant’s position (77 belt restrained front seat occupants in car-to-car side collisions).
108
E. MILTNER and H.-J. SALWENDER
Prob[MAIS4-61
Table 2. Influential factors on individual injuries
1,O
Skull fractures
0,9 -
X
EES
08 -
On impact! opposite side Aee
0,7 036 -
Liver runtures
Pelvic fractures
Rib fractures
X
X
X
X
X
X
X
X
Notes: 77 belt-restrained front seat occupants collisions. Center of impact: front door or B-pillar. Multivariate analysis, logistic regression.
085 0,4 -
in side
0,3 02 081 0,O
I 20
1 10
0
I 30
I 40
I 50
I 60
t 70
I 80
1 90 km/h
When EES, occupant position, and age are used in combination as predictors of pelvic fracture, 66 of the 77 cases (85%) were correctly predicted. These were 65% of the cases with pelvic fractures and 93% of those without.
EES Fig. 3. Probability of injuries MAIS 4-6 depending on EES and occupant’s position (79 belt-restrained front seat occupants in car to car side collisions. Center of impact: front door or B-pillar).
By means of logistic regression, it was possible to calculate a prediction formula for pelvic fractures: -6,5293074
+ 0,08094332 x EES + 0,0452515 match
X Age +
on near side/far side:
- 1,0108663,
when “far side”
1,01086635,
when “near side”
.)
otherwise.
DISCUSSION In Fildes and Vulcan’s study (1990) of 150 side collisions, near-side front seat occupants were injured mainly by interior vehicle structures. In general, far-side occupants were less frequently injured: 36% due to contact with other occupants, 27% by interior vehicle surfaces, 18% by safety belts, and 18% by the dashboard. Rouhana and Foster (1985) described similar results. The behavior of front seat car occupants in side collisions can be studied in experimental dummy tests. In experimental 90” side collisions with beltrestrained dummies, Faerber (1982) observed that
12
0
I
’abdomen abdomen head near' head far ’ thorax ’ thorax near side far side near side far side side side
I
I
spine near spine far side side
I
extrem. near side
Fig. 4. Distribution of injury severity on all body locations at an EES of 40-59 km/h (belt-restrained front seat occupants; and 10 far side in car to car side collisions. Center of impact: front door or B-pillar).
I
extrem. far side
10 near side
Injury severity in car-to-car side impacts
the near-side dummy was thrown into the middle of the passenger cell and against the far-side dummy. The near-side dummy usually received one impact from the car door and one impact when hitting the far-side dummy, the far-side dummy received only one impact, namely by the near-side dummy. Occupant-to-occupant contact generated dummy loads that reached nearly the same level in each of the interacting body regions, but the loads were only 25% to 50% of those measured in the primary impacts. In our study, we found no increase in injury severity in far-side drivers, when a passenger was present. However, we could not differentiate between injuries caused by contact with the other front seat occupant and injuries caused by internal car structures, especially by the other front seat. The risk of life-threatening injuries for near-side occupants is two to three times higher than that on the far side according to: Riley and Radley 1976; Walz et al. 1977; Norin, Nilsson-Ehle, and Gustafsson 1982; Danner and Langwieder 1976; Rouhana and Foster 1985. Rouhana and Foster (1985) detected not only a three times higher risk of severe injuries for near-side occupants, but also observed a 1.6 times greater risk for near-side passengers than for near-side drivers. Near-side passengers sustained fatal injuries three times more often than nearside drivers. Near-side passengers bear the greatest risk of fatal injury, while far-side passengers have the lowest risk. However, most of the passengers in Rouhana’s survey had not fastened their seat belts and some of them were thrown out of the car. In our survey a clear correlation between the total injury severity and EES for occupants seated on both sides could be detected; the graph for farside occupants is shifted to the right by approximately 10 km/h. Within the critical range of 30-50 km/h for near side occupants, and 40-60 km/h for far-side occupants, the probability of severe injuries MAIS 4-6 increased from approximately 20% to more than 90%. Within the range of 40-59 km/h, starting at an EES of 40 km/h, all near-side occupants were killed; far-side occupants also sustained severe injuries, however, only half as often as the near-side occupants. Otte, et al. (1984a) detected similar results when regarding delta u with impacts over the entire occupant compartment. For near-side occupants the risk of MAIS 5-6 injury from a delta u of 30 km/h increased drastically; however, for far-side occupants this trend was far less. Rouhana and Foster (1985) found the median delta u for severely injured occupants in side collisions was 27 km/h; it was 50 km/ h for those fatally injured. In this study in the EES speed range 40-59 km/h
109
the most frequently, severely injured body locations were the trunk and head. As in the case of total injury severity, the injury risk was only half as great for far-side as for near-side occupants. These results concur with Langwieder 1975 and Fildes and Vulcan 1990; however, in their studies all speed ranges were regarded. The influence of occupant position, near and far side, on resulting injury was less for head injuries than trunk injuries. Otte et al. (1984b) found that the difference in injury frequency between the vehicle occupant positions was highest for the lower trunk (22.7% near side and 9.5% far side). Principally, far-side occupants are less endangered than those adjacent to the impact. Nevertheless, within the critical EES range of 40-50 km/h there is still a risk of severe injury for far-side occupants-25% compared to 50% for the near side. Thus all seating positions must be considered when occupant protection measures are formulated. Without precisely defining the side-impact location, McCoy, Johnstone, and Kenwright (1989) observed pelvic fractures in near- and far-side occupants, which arose at a delta u as low as 15 km/h. With regard to the EES, pelvis fractures occurred on the near side at an EES less than 20 km/h and on the far side at an EES of 40 km/h. Our survey revealed that the probability of pelvic fractures, as well as other bone injuries, does not depend solely on the energy involved in the accident, but as well on occupant age. This observation corresponds to large scale statistical surveys (Evans 1985, 1988) and also to experimental biomechanical tests (Vetre 1977; Mattern et al. 1979; Schmidt 1979). Evans 1988 determined that the risk of sustaining fatal injuries is lowest at 20 years of age and three times higher at 70 years. Thus the validation of occupant protection measures should also take occupant age into consideration. Viano et al. 1990 detected that 64% of near-side occupants were over 50 years old and 36% over 70 in fatal multivehicle side impacts. This result contrasts the young age of the occupants in our sample (37 years near side, 35 years far side). The difference originates from the kind of the samples. Viano’s work is an epidemiological study, while we tried to collect primarily severe accidents. In this study a regression analysis was performed to predict the probability of pelvic fracture based on EES, age, and occupant position. This relationship predicted correctly the existence of pelvic fracture in 85% of the cases. Thus, through the employment of extensive real accident analyses, it is possible to develop predictors for injuries and injury severity with a high degree of accuracy.
110
E. MILTNERand H.-J. SALWENDER
Acknowledgements-Many thanks to the following for the reconstruction of accidents: Klaus Peter Wiedmann, Burkhard Leutwein, Hans Peter Hepp, and Rainer Fischer.
REFERENCES Burg. H. Das Antriebs-Balance-Diagramm als optimales Der Verkehrsunfall, Heft 2 Mittel der Unfallanalyse.
(1973). Burg, H.; Rau, H. Handbuch der Verkehrsunfallrekonstruktion. Verlag Information Ambs GmbH, Kippenheim (1981). Burg, H; Zeidler, F. EES-Ein Hilfsmittel zur Unfalrekonstruktion und dessen Auswirkungen auf die Unfalforschung. Der Verkehrsunfall Heft 4 (1980). Danner, M.; Langwieder, K. Car/vehicle side impacts-a study of accident characteristics and occupant injuries Proceedings of the 6th International Conference on Experimental Safety Vehicles, 1976; 516-540. Danner, M.; Langwieder, K.; Hummel, T. H. Experience from the analyses of accidents with a high belt usage rate and aspects of continued increase in passenger safety. 1 lth International Technical Conference on Experimental Safety Vehicles, Washington, May 12-15, 1987; 201-210. Evans, L. Involvement rate in two-car crashes versus driver age and car mass of each involved car. Accid. Anal. Prev. 16:387-405; 1985. Evans, L. Risk of fatality from physical trauma versus sex and age. J. Trauma 28:368-378; 1988. Faerber, E. Interaction of car passengers in frontal, side, and rear collisions. Proceedings of the 26th Stapp Conference, Ann Arbor, MI, October 20-21, 1982; 113, 335-352. Fildes, B. N.; Vulcan, A. P. Crash performance and occupant safety in passenger cars involved in side impacts. International Conference on the Biomechanics of Impacts (IRCOBI), Bron-Lyon, France, 1990; 121-140. Jones, I. S. Injury severity versus crash severity for front seat car occupants in front and side impacts. Proceedings of the 26th Annual Meeting of the American Association for Automotive Medicine, Ottawa, Canada, October 4-6, 1982. Kallieris, D.; Mattern, R. Belastbarkeitsgrenzen und Verletzungsmechanik der angegurteten Fahrzeuginsassen beim Seitaufprall (1). FAT-Schriftenreihe Nr. 36 (1984). Kallieris, D.; Mattern, R. ; HLdle, W. Belastbarkeitsgrenzen und Verletzungsmechanik der angegurteten Fahrzeuginsassen beim Seitaufprall(2). FAT-Schriftenreihe Nr. 60 (1986). Kallieris, D.; Schmidt, G.; Miltner, E. Verminderung der Verletzungsschwere kollisionsnaher Frontinsassen bei der 90 Grad Pkw/Pkw-Seitenkollision durch fahrzeugkonstruktive Massnahmen. Hefte Unfall- und Sicherheitsforschung Strassenverkehr, Kongressbericht der Deutschen Gesellschaft fur Verkehrsmedizin e.V.; 202-205 (1991). Kramer, F. Einfluss der mittleren Fahrzeugverzdgerung auf die Insassenbelastungswerte bei der Frontalkollision. Automobil-Industrie 1, 115 (1980). Langwieder, K. Aspekte der Fahrzeugsicherheit anhand einer Untersuchung von realen Unfallen. Dissertation. Berlin; 1975. Mattern, R.; Barz, J.; Schulz, F.; Kallieris, D.; Schmidt,
G. Problems arising when using injury scales in biomechanical investigation with special consideration of the age influence. Proceedings of the 4th IRCOBI Conference, Goteborg, Sweden, September 5-7, 1979; 223-23 1. McCoy, G. F.; Johnstone, R. A.; Kenwright, J. Biomechanical aspects of pelvic and hip injuries in road traftic accidents. J. Orthop. Trauma 3: 113-123; 1989. Niederer, P.; Walz, F.; Weissner, R. Verletzungsursachen beim Pkw-Insassen, Verletzungsminderung durch modeme Sicherheitseinrichtungen. Unfallheilkunde 83, 326-340; 1980. Norin, H.; Nilsson-Ehle, A.; Gustafsson, C. Volvo traffic accident research system as a tool for side impact protection evaluation and development. Proceedings of the 9th International Conference on Experimental Safety Vehicles, Kyoto, November 1-4, 1982; 439-448. Olsson, J. A.; Skotte, L.-G. Air bag system for side impact protection. Proceedings of the 12th Conference on Experimental Safety Vehicles, Goteborg, Sweden, September 5-7, 1989. Otte, D.; Ktihnel, A.; Suren, E. G.; Weber, H.; Gotzen, L.; Schockenjoff, G.; Vu Han, V. Erhebungen am Unfallort. Unfall- und Sicherheitsforschung Strassenverkehr 37, l-100 (1982). Otte, D.; Suren, E. G.; Appel, H.; Nemzow, J. Vehicle parts causing injuries to front-seat car passengers in lateral impact. Proceedings of the 28th Stapp Conference, Chicago, 1984a; 13-24. Otte, D.; Suren, E.-G.; Appel, H.; Nehmzow, J. Bewertung von Einflussparametern fur die Verletzungen von Pkw-Insassen bei Seit-Kollisionen. Verkehrsunfall und Fehrzeugtechnik, 41-48 (1984b). Riley, B. S.; Radley, C. P. Traffic accidents to cars with directions of impact from the front and side. Proceedings of the 6th International Conference on Experimental Safety Vehicles, Washington, October 12-15, 1976; 563-573. Ropohl, D. Die rechtsmedizinische Rekonstruktion von Verkehrsunfallen. DAT-Schriftenreihe Technik, Markt, Sachverstandigenwesen, Band 5,90-92 (1990). Rouhana, S. W.; Foster, M. E. Lateral impact-an analysis of the statistics in the NCSS. Proceedings of the 29th Stapp Conference, 1985; 79-98. Slibar, S. Die mechanischen Grundsatze des Stossvorganges und ihre Anwendung. Archiv fur Unfallforschung, GUFU Baden-Baden (1966). Schmidt, G. The age as a factor influencing soft tissue injuries. Proceedings of the 4th IRCOBI Conference, Goteborg, September 5-7, 1979; 143-150. States, J. D.; Huelke, D. F.; Baker, S. P.; Bryant, R. W. et al. The abbreviated injury scale (AIS), 1985 revision. Arlington Heights, IL: American Association for Automotive Medicine; 1985. Vetre, T. Die dynamische Biegefestigkeitsprtifung von menschlichen Rippen. Med. Diss., Heidelberg (1977). Viano, D. C.; Culver, C. C.; Evans, L.; Frick, M. Involvement of older drivers in multivehicle side-impact crashes. Accid. Anal. prev. 22:177-188; 1990. Walz, F.; Niederer, P.; Zollinger, U.; Renfer, A. belted occupants in oblique and side impacts. Proceedings of the 3rd IRCOBI Conference, Berlin, September 7-9, 1977; 194-205.