Abstract
In order to develop effective restraint systems for the pregnant occupant, injury criteria for determining fetal injury risk must be developed. This study presents computer simulations of a 30 week pregnant occupant that illustrate the importance of local uterine compression on the risk of fetal injury. Frontal impact simulations with a range of velocities and belt positions were used to identify the best correlation between local uterine compression and peak strain measured at the uterine-placental interface. It is suggested that future pregnant dummy development and specifically pregnant injury criteria should be based on local uterine compression relative to the placental attachment location.
Automobile crashes are the largest single cause of death for pregnant females (Attico et al., 1986) and the leading cause of traumatic fetal injury mortality in the United States (US) (Weiss, 2002). Each year, 160 pregnant women are killed in motor-vehicle crashes (MVCs) and an additional 800 to 3200 fetuses are killed when the mother survives (Klinich et al., 1999b) in the US. Unfortunately, fetal injury in motor vehicle crashes is difficult to predict due to the fact that real world crash data is limited and cadaver studies are not feasible.
In the non-pregnant female, the uterus is a muscular organ the size of a lemon located within the abdominal cavity. As the fetus grows during pregnancy, the abdomen stretches to the size of a watermelon. The internal volume increases from 0.005 L to 5 L and as much as 10 L (Rupp et al., 2001). The uterine wall is uniform prior to delivery, with a thickness of about 1 cm. The uterosacral and round ligaments extend from the uterus to the pelvis and act to support the uterus. The interior of the uterus contains the fetus, which is surrounded by amniotic fluid and the placenta (Figure 1). The placenta is a vascular organ that acts as a permeable membrane that exchanges oxygen, nutrients, and waste products between the mother and fetus via the umbilical cord. It is a flat, roughly circular structure 2 cm thick in the center and is typically located in the upper half of the uterus (Fried, 1978).
Figure 1.
Anatomy of a 40-week pregnant woman (ligaments not shown).
In an effort to reduce the risk of injury to pregnant occupants in car crashes, a pregnant anthropometric test dummy (ATD) has been developed at the University of Michigan Transportation Research Institute (Rupp et al., 2001). The Maternal Anthropomorphic Measurement Apparatus Version 2B (MAMA-2B) is a second-generation prototype ATD that is a retrofit Hybrid III small female dummy. One of the primary limitations of the pregnant dummy is the lack of injury criteria for the fetus. The MAMA-2B was designed to measure anterior and posterior pressure in the fluid-filled abdomen insert as well as the strain on the perimeter of the insert. However, only the anterior pressure measurements were repeatable (Rupp et al., 2001). Therefore, it would be beneficial to have an injury criterion for the pregnant dummy that utilizes currently established ATD measurement methods. One leading example would be to measure overall abdominal compression in a similar manner that used to measure chest compression. For example, this could be done by using a string potentiometer as is done in the chest.
The most common cause of fetal death from motor vehicle accidents is placental abruption, which is the premature separation of the placenta from the uterus (Klinich et al., 1999b). Both the pregnant dummy and the pregnant model used in this study utilize this injury mechanism to predict fetal outcome (Moorcroft et al., 2003a). However, due to the difficulties in measuring this mechanism in the pregnant dummy, such as strain and pressure, another type of measurement is desired that can accurately predict fetal injury risk. The purpose of this paper is to identify the best correlation between uterine compression that could be measured by a dummy and the risk of fetal injury as predicted by the strain at the UPI.
METHODOLOGY
A total of 15 computer simulations of a 30 week pregnant occupant in a pure frontal impact were performed using a model previously developed and validated by Moorcroft et al. (2003a, 2003b, 2003c). The model is a modified MADYMO small female human model with a pregnant abdomen (Figure 2). The abdomen consists of the uterus, placenta, and amniotic fluid. The uterus is supported by two pairs of ligaments and surrounded by fat.
Figure 2.
Simulation configuration of pregnant occupant in passenger position with three-point belt.
Four techniques were used to validate the pregnant model. First, a global biofidelity response was evaluated by using a seatbelt to compress dynamically the pregnant abdomen (Moorcroft et al., 2003b). The force versus compression results were within the published corridors from scaled cadaver tests (Hardy et al., 2001). Second, a similar validation procedure was performed with a rigid bar (Moorcroft et al., 2003b) and these results were also consistent with previous data (Hardy et al., 2001). The third technique involved validating the model against real-world crashes in order to investigate the model’s ability to predict injury. Using fatal crashes from pregnant occupants (Klinich et al., 199b), the model showed strong correlation (R2 = 0.85) between peak strain at the uterine-placental interface (UPI) as measured in the model compared to risk of fetal demise as reported in the real-world crashes over a range of impact velocities and restraint conditions (Moorcroft et al., 2003a). The forth method compared the physiological failure strain from placental tissue tests to the failure strain measured in the model. Tissue tests by Rupp et al. (2001) suggested approximately a 60% failure strain for UPI tissues which is in agreement with the model’s prediction of 75 % risk of fetal loss at a 60% strain in the UPI. In summary, the global, injury, and tissue level validation techniques all indicate the model is good at predicting injurious events for the pregnant occupant.
The right-front passenger seat was chosen for all simulations because the potentially confounding effects of the steering wheel are not present and therefore the occupant loading is dominated by the restraint system (Kent et al., 2001). In addition, the airbag was not deployed in any simulations in order to focus on the uterine compression. As with the omission of the steering wheel, the lack of airbag deployment allows for the lap belt to load the abdomen, thus giving a consistent loading pattern and allowing for the location of the abdominal measurement to be assessed. This is effectively a rigid bar impact to the abdomen, but by using lap belt, the load conforms more realistically around the abdomen and a better contact interaction is achieved between the two finite element objects.
The passenger interior is a typical MADYMO passenger compartment with a seatback, seat cushion, floor, dashboard, and windshield. The setback distance of the passenger seat is consistent with that of a typical passenger compartment and representing a fully aft seat position with an abdominal clearance of 40 cm. This positioning of the pregnant occupant was based on the seated anthropometry of a pregnant woman in her 30th week of pregnancy as defined by Klinich et al. (1999a). The MADYMO finite element belt was constructed of nonlinear triangular membrane elements. The material model was isotropic with hysteresis and the loading curve had an approximate slope of 1.75 GPa and an unloading slope of approximately 200 MPa. The belt was 5 cm wide.
The test matrix consisted of simulations with five different lap belt locations (Figure 3). The initial lap belt location is representative of the recommended position. The top belt placement represents a lap belt that has slid over the top of the uterus. Three intermediate belt locations were also defined, with nominal locations, with respect to the initial belt location, of +0.03 m, +0.07 m, and +0.09 m. Positioning of the belts was accomplished by placing a straight belt at a specific vertical location and then pulling the belt across the abdomen until the belt was taut. The initial, or recommended location, is based on the recommended position for all occupant which has the belt as low as possible on the abdomen to facilitate loading of the pelvis (Rupp et al., 2001).
Figure 3.
Five belt locations relative to pregnant uterine model.
Abdominal measurements were taken at five locations, named AB1, AB2, AB3, AB4 and TH1 to correspond to the names of the flexible layers in MADYMO (Figure 4). The first abdominal layer, AB1, is at the same height as the pelvis. The fourth abdominal layer, AB4, is directly in line with the placenta. The first thoracic layer, TH1, is above the uterus. For each layer, a central node on the skin, both on the anterior and posterior surfaces of the abdomen, was selected. The magnitude of the distance between the two points was outputted. The magnitude was chosen to account for the rotation of the body. The initial distance ranged from 217 mm to 248 mm. Peak UPI strain was also recorded for each simulation.
Figure 4.
Local uterine compression measurement locations.
Each of the five lap belt positions was evaluated at three velocities (13 kph, 35 kph, 55 kph) for a total of 15 simulations. The applied sled pulse was a half-sine wave imposed for a duration of 100 ms. Linear egression analysis was used to correlate the peak uterine compression values to the peak UPI strain and subsequent risk of fetal injury.
RESULTS
The peak uterine strain increased with crash speed and lap belt height except for the top belt location (Table 1). However, overall compression, which is the peak compression value for all locations, was constant over the range of lap belt locations for a given speed; 20% for 13 kph, 35% – 40% for 35 kph, and 50% for 55 kph. In nearly all simulations, the peak compression occurred at the same time as the peak strain. Therefore, the same event is being recorded by strain and compression. Next, the compression at each location was evaluated (Table 2). The measurement locations each showed a bias towards localized loading. The bottom two locations and the top measurement location responded minimally to remote loading.
Table 1.
Simulation matrix, peak results and timing.
| Run Number Belt Location. | Speed | Peak Strain(%) | Time of peak strain (ms) | Overall Peak Compression (%) | Time of peak compression (ms) |
|---|---|---|---|---|---|
| 1.Recommended | 13 | 15.8 | 78 | 17.0 | 80 |
| 2.Location 1 | 13 | 23.7 | 76 | 20.2 | 89 |
| 3.Location 2 | 13 | 32.8 | 89 | 22.6 | 93 |
| 4.Location 3 | 13 | 29.7 | 91 | 18.0 | 93 |
| 5.Top Location | 13 | 13.7 | 88 | 15.9 | 89 |
| 6.Recommended | 35 | 33.7 | 81 | 35.1 | 90 |
| 7.Location 1 | 35 | 49.5 | 93 | 40.7 | 87 |
| 8.Location 2 | 35 | 96.9 | 80 | 45.1 | 80 |
| 9.Location 3 | 35 | 83.1 | 77 | 35.4 | 77 |
| 10.Top Location | 35 | 35.2 | 88 | 31.4 | 75 |
| 11.Recommended | 55 | 37.4 | 81 | 44.6 | 99 |
| 12.Location 1 | 55 | 51.5 | 90 | 49.8 | 97 |
| 13.Location 2 | 55 | 145.4 | 122 | 50.8 | 113 |
| 14.Location 3 | 55 | 97.5 | 77 | 51.4 | 117 |
| 15.Top Location | 55 | 41.8 | 84 | 36.2 | 94 |
Table 2.
Peak abdominal compression for each uterine location.
| Run Number Belt Location. | AB 1(%) | AB 2 (%) | AB 3 (%) | AB 4 (%) | TH 1 (%) |
|---|---|---|---|---|---|
| 1.Recommended | 3.6 | 8.8 | 17.0 | 5.6 | 2.8 |
| 2.Location 1 | 1.6 | 9.5 | 20.2 | 12.6 | 4.5 |
| 3.Location 2 | 0.0 | 2.7 | 16.8 | 22.6 | 11.7 |
| 4.Location 3 | 0.0 | 0.4 | 7.9 | 18.0 | 16.7 |
| 5.Top Location | 0.0 | 0.4 | 1.3 | 5.5 | 16.0 |
| 6.Recommended | 10.0 | 16.3 | 35.1 | 11.9 | 4.7 |
| 7.Location 1 | 0.0 | 18.1 | 40.7 | 17.3 | 6.3 |
| 8.Location 2 | 0.0 | 5.7 | 37.8 | 45.1 | 17.9 |
| 9.Location 3 | 0.0 | 0.4 | 16.9 | 35.2 | 35.4 |
| 10.Top Location | 12.7 | 0.4 | 2.0 | 9.7 | 31.4 |
| 11.Recommended | 12.6 | 18.4 | 44.7 | 14.9 | 5.2 |
| 12.Location 1 | 0.0 | 19.4 | 49.8 | 19.5 | 6.5 |
| 13.Location 2 | 0.0 | 5.9 | 46.2 | 50.8 | 20.8 |
| 14.Location 3 | 0.0 | 0.4 | 18.7 | 51.4 | 43.9 |
| 15.Top Location | 0.0 | 0.4 | 5.7 | 10.6 | 36.2 |
The kinematics of the simulations follow the same general trend showing submarining of the occupant below the lap belt occurring for belt positions above the recommended position (Figure 5). Additionally, large abdominal and subsequently uterine compression resulted from this submarining. The danger of this loading is evident in strain values which exceeded the 60% tissue limit in many of the simulations at 35 kph and 55 kph. The localized loading is evident in examination of the uterus itself (Figure 6). Loading at belt location 1 maximally compressed the uterus at the AB 3 location; however, minimal compression is observed at AB 4 and minimal deformation of the upper half of the uterus as well, thereby resulting in a moderate value of peak UPI strain.
Figure 5.
Simulation belt location 2 pre-test (A) and peak overall uterine compression (B).
Figure 6.
Example uterine compression peak at location AB 3.
A linear regression analysis of the data was performed for each measurement location (Table 3). The best correlation was between AB 4 and peak UPI strain for all three impact velocities (Table 3). Moreover, the AB 4 peak compression results in the best correlation coefficient (R=0.93) for all simulations compared when to the overall peak compression (R=0.75) (Figure 7).
Table 3.
Correlation coefficient values for each uterine compression measurement versus peak UPI strain.
| Measurement Location | 13 kph | 35 kph | 55 kph | All speeds |
|---|---|---|---|---|
| AB 1 | 0.48 | 0.62 | 0.69 | 0.53 |
| AB 2 | 0.50 | 0.69 | 0.77 | 0.65 |
| AB 3 | 0.24 | 0.45 | 0.72 | 0.05 |
| AB 4 | 0.92 | 0.97 | 0.92 | 0.93 |
| TH 1 | 0.20 | 0.14 | 0.33 | 0.26 |
Figure 7.
Overall peak compression and AB4 peak compression versus peak UPI strain.
DISCUSSION
The AB 4 measurement location correlated the best because the strain is measured in the uterus at the placental location, which is at approximately the same height as the AB4 measurement location. As noted above, the greatest risk for fetal injury results from placental abruption, which occurs when the strain in the uterus near the placenta exceeds the tissue limit. Therefore, in simulations where the abdominal loading was remote from the placenta, low strains were recorded and relatively low compression was recorded at the AB4 level as compared to the level of the loading. Although placental location varies from pregnancy to pregnancy, and can vary during pregnancy, as many as 95% of placenta’s are located in the upper half of the uterus (Fried, 1978). Therefore, it is suggested as best practice to measure abdominal loading at the uppermost uterine location in an attempt to predict injury to the UPI and risk of fetal demise.
Predicting fetal injury from abdominal deflection is loosely analogous to using chest deflection to predict thoracic injury. As a simple comparison, chest deflection for the small female is limited to 52 mm by federal safety standards (Eppinger et al., 1999). A chest deflection of 52 mm is approximately 35% compression which corresponds to approximately 40% risk of an AIS 3 or greater injury (Mertz et al., 1991). Given the obvious anatomical differences between the thorax and pregnant uterus, it is interesting that 35% compression of the uterus at AB 4 is also the higher limit of injury. The abdominal deflection could be measured in the same manner as chest deflection, using a string potentiometer, chestband, or through processing of digital video. It is important to note that the measurements need to be taken from a pregnant dummy, which has the correct anthropometry and abdominal force-deflection response as a pregnant woman.
Although this study indicates a local compression measure is needed to predict injury to the fetus, it may be difficult to measure this in the MAMA-2B test dummy. However, the results of this study should be utilized to help design future versions of the pregnant dummy in order to measure local compression values. This is analogous to the progress that has been made in thoracic injury dummy measurements. A single chest compression measure is used currently in the HIII dummy to predict chest injuries; however, the THOR next generation male test dummy includes sensors to predict several local chest deflections that may provide more accurate thoracic injury information.
It is important to note that previous simulations indicate that for all frontal impacts it is safest for the pregnant occupant to ride in the passenger seat while wearing a three-point belt and utilizing the frontal airbag when appropriate (Moorcroft et al., 2004). However, for the purpose of this paper and examining uterine compression, the airbag was not utilized.
As with all computational models, this model is limited by the accuracy of input and simplifications made. The tissue data, from which the failure strain is derived, is sparse and simplifications are made to use that data as a material model. Additionally, the boundary conditions and geometry can and should be improved in future generations of the model. Furthermore, the model only looks at injury at the UPI. In cases with very large deflections, direct injury to the fetus may occur at injury rates different then those for placental abruption. It is recommended that the methods in this paper be applied to future generations of the pregnant occupant model to provide a continually improving understanding of pregnant occupant injury risk prediction.
CONCLUSIONS
Computer simulations of a 30 week pregnant small female were performed using a range of belt locations and impact velocities in order to evaluate a new injury predictor for the pregnant occupant. The results illustrate that local uterine compression correlates with peak uterine strain and is a better predictor then the overall global peak compression. It is critical for future pregnant dummy development and for overall protection of the pregnant occupant that injury criteria be evaluated. Abdominal compression is a simple measurement that should yield viable results when predicting the likelihood of injury to the fetus as a result of motor vehicle crashes,
ACKNOWLEDGEMENTS
The authors would like to thank Kathy Klinich at the University of Michigan Transportation Research Institute for her assistance in the development of the pregnant model.
REFERENCES
- Attico NB, Smith RJ, III, Fitzpatrick MB, Keneally Automobile safety restraints for pregnant women and children. J Reprod Med. 1986;31(3):187–92. [PubMed] [Google Scholar]
- Eppinger R, Sun E, Bandak F, Haffner M, Khaewpong N, Maltese M, Kuppa S, Nguyen T, Takhounts E, Tannous R, Zhang A, Saul R. Development of improved injury criteria for the assessment of advanced automotive restraint systems – II. NHTSA; 1999. Docket No. 99-6408-5. [Google Scholar]
- Fried AM. Distribution of the bulk of the normal placenta. Review and classification of 800 cases by ultrasonography. American Journal of Obstetrics and Gynecology. 1978;132(6):675–680. doi: 10.1016/0002-9378(78)90863-3. [DOI] [PubMed] [Google Scholar]
- Hardy WN, Schneider LW, Rouhana SW. Abdominal impact response to rigid-bar, seatbelt, and airbag loading. Stapp Car Crash Journal. 2001;45:1–32. doi: 10.4271/2001-22-0001. [DOI] [PubMed] [Google Scholar]
- Kent RW, Crandall JR, Bolton J, Prasad P, Nusholtz GS, Mertz HJ. The influence of superficial soft tissues and restraint condition on thoracic skeletal injury prediction. Stapp Car Crash Journal. 2001;45:183–204. doi: 10.4271/2001-22-0008. 2001-22-0008. [DOI] [PubMed] [Google Scholar]
- Klinich KD, Schneider LW, Eby B, Rupp J, Pearlman MD. Seated anthropometry during pregnancy. 1999a UMTRI-99-16. [Google Scholar]
- Klinich KD, Schneider LW, Moore JL, Pearlman Investigations of crashes involving pregnant occupants. 1999b UMTRI-99-29. [PMC free article] [PubMed] [Google Scholar]
- Mertz HJ, Horsch JD, Horn G, Lowne RW. Hybrid III sternal deflection associated with thoracic injury severities of occupants restrained with force-limiting shoulder belts. SAE Technical Paper 910812; 1991. pp. 105–119. [Google Scholar]
- Moorcroft DM, Duma SM, Stitzel JD, Duma GG. Computational Model of the Pregnant Occupant: Predicting the Risk of Injury in Automobile Crashes. American Journal of Obstetrics and Gynecology. 2003a;2003;189(2):540–544. doi: 10.1067/s0002-9378(03)00519-2. [DOI] [PubMed] [Google Scholar]
- Moorcroft DM, Duma SM, Stitzel JD, Duma GG. A finite element and multi-body model of the pregnant occupant for the analysis of restraint effectiveness. SAE Technical Paper; 2003b. 2003-01-0157. [Google Scholar]
- Moorcroft DM, Stitzel JD, Duma SM, Duma GG. The Effects of Uterine Ligaments on the Fetal Injury Risk in Frontal Automobile Crashes. Journal of Automobile Engineering. 2003c;217:1049–1055. Part D. [Google Scholar]
- Moorcroft DM, Duma SM, Stitzel JD, Duma GG. The effect of pregnant occupant position and belt placement on the risk of fetal injury. SAE Technical Paper; 2004-01-0324.2004. [Google Scholar]
- Rupp JD, Klinich KD, Moss S, Zhou J, Pearlman MD, Schneider LW. Development and testing of a prototype pregnant abdomen for the small-female Hybrid III ATD. Stapp Car Crash Journal. 2001;45:61–78. doi: 10.4271/2001-22-0003. [DOI] [PubMed] [Google Scholar]
- Pearlman MD, Ashton-Miller JA, Dyer T, Reis P. Data acquisition for development to characterize the uteroplacental interface for the second-generation pregnant abdomen. 1999 Submitted to NHTSA. [Google Scholar]
- Weiss HB, Strotmeyer S. Characteristics of pregnant women in motor vehicle crashes. Injury Prevention. 2002;8(3):207–214. doi: 10.1136/ip.8.3.207. [DOI] [PMC free article] [PubMed] [Google Scholar]