*Corresponding Author:
Prof. Ayman El-Menyar
Department of Surgery, Trauma Surgery Section, Hamad General Hospital, Doha, Qatar.
E-mail: [email protected]


Rollover crashes (ROCs) are responsible for almost a third of all highway vehicle occupant fatalities. Although ROCs are common and serious mechanism of injury, ROCs are under‑reported. To analyze the causes, mechanism, impact and prevention of ROCs, we reviewed the literature between 1984 and 2013. By utilizing the search engines PubMed, MEDLINE and EMBASE by using key words “ROCs” “Ejection” and “vehicle” the initial search yielded 241 abstracts, of which 58 articles were relevant. Most of the articles were either retrospective or experimental studies funded by automobile companies. All vehicles are susceptible to rollovers to certain extents. Despite continuing innovation in vehicles’ safety, human factor is pivotal in prevention of ROCs. Distracted driving, speeding and drinking escalate the chances of rollover crashes. Wearing a seatbelt greatly improves the chances of surviving a ROC.


Crashes, Ejection, Motor vehicle, Rollover


Motor vehicle crashes (MVCs) are the major cause of morbidity and mortality among young population. Rollover crashes (ROCs) involves one of the most fatal type of MVCs.[1] It is referred to a crash in which vehicle rotation of ≥ 90° occurs about any longitudinal or lateral axis.[2] As the majority of ROCs take places in the highways, it poses serious road safety concerns.[3] Around 220,000 light motor vehicles sustain ROCs in the US annually involving 350,000 vehicle occupants. These crashes reported 9000 occupants deaths, 14,100 serious injuries whereas, minor to moderate injuries were reported in 224,000 victims of these crashes.[3] Despite the fact that ROCs constitute only 2.2% of all MVCs; it represents about 33% of the annual injury costs in the US (around $40 billion).[3] It has been reported that 50% and 10% of the harm due to ROCs is related to the head and neck and spines injuries, respectively.

Moreover, previous data showed that the annual incidence of ROCs-related mortality accounted for 3.4/100,000 person.[4] A high fatality rate is observed among the occupants of all vehicles which also includes light trucks and sports utility vehicles (SUVs; 50%) as well as the heavy trucks (60%).[4] Although the rollover is a serious event, the association of occupant’s ejection has the potential to add significantly to the resultant injuries. Furthermore, ejection from the vehicle is more serious when the occupant experiences contact either with the vehicle (on the way out) or the ground.[5] This is a review of the literature, to study the mechanism, classifications, severity and types of “ROCs” and “Ejection”.

Methods of Literature Search

To define the incidence, characteristics, mechanism, risk factors and preventive strategies for ROCs, we reviewed the literature between 1984 and 2013. We utilized the search engines PubMed, MEDLINE and EMBASE by using key words “ROCs” “Ejection” and “vehicle”. The initial search yielded 241 abstracts, of which 58 articles were relevant. Only relevant review articles and research studies (retrospective and prospective) were included in this review. The study design of the selected articles was mainly confined to relevant reviews articles and research studies either retrospective or experimental studies funded by automobile companies.

Studies addressing the car ROCs with or without ejection, impact of ROCs on safety measures, clinical presentation and outcome (mortality and hospital length of stay) were mainly considered. However, abstract, unpublished data and non-English manuscripts were excluded.

ROCs and Fatalities

MVCs are a significant source of major trauma among all age groups.[6] Fortunately, with the introduction of advanced safety features in motor vehicles (seat belt restraints and air bags); crash-related fatalities have been remarkably reduced.[7] Despite that, ROC remains a common cause of MVCs fatalities in all continents. In general, 35% of occupants fatalities are related to the rollover initiation speed at initiation and number of turns. Moreover, an increased association of fatality was observed with head injuries in restrained occupant.[8] In Australia ROC is responsible for about one in every five fatalities.[9] The estimated fatalities involving ROCs in Europe is around one in every 10 fatalities.[10] Studies from the US reported a relatively higher fatality rate by ROC which corresponds to one in every 3-4 fatalities.[11,12] Though, the intensity of collision and inadequacy of occupant protection in vehicles are associated with risk of severe injury in ROCs. Lack of standards for the assessment of ROCs resulted in overrepresentation of fatalities associated with ROCs in many countries.[13]

Mechanism of Injury in ROCs and Ejection

There are four main injury mechanisms involved in ROCs. First is roof intrusion, in which structural elements either penetrates the body of the occupants or crush them altogether due to massive structural deformations and loss of the residual space. Another mechanism is a projection in which uncontrolled movement of the occupants inside the vehicle resulted in occupants’ body impact with the structural parts of the vehicle compartment.[14] In case of complete ejection, the passengers could be ejected from the vehicle during the rollover process. However, in partial ejection parts of the passenger’s body come in contact with the outside surface and can be strongly crushed during rollover process.[14] Usually, the standard restraints equipped for frontal collisions are not beneficial in rollovers, particularly with roof intrusion. In addition, there is a high risk of severe head and spinal injuries in restrained occupant who were partially ejected.[13] A study by Rechnitzer et al.[15,16] demonstrated the ineffectiveness of restraints and impact of roof crush in the severity of injury.

Mechanism and Classification of ROCs

Rollover of vehicles involves two patterns. The first one is referred to lateral rollover in which the vehicle rotates around a longitudinal axis being parallel with the main longitudinal axis of the vehicle (i.e. bus). It is the most common way of rollover and about 98% of all vehicle ROCs represents this group.[14] In contrast, the second type of rollover vehicle crash involves rotation around an axis being perpendicular to the vertical longitudinal central plane of the vehicle. This type of ROCs is uncommon and observed only in 1-2% cases.[14]

ROCs categories based on different initiation scenarios

• Trip-over occurred when the vehicle is suddenly slowed or stopped inducing a rollover

• Flip-over: When the vehicle is rotated along its longitudinal axis by a ramp-like object such as a turned down guardrail

• Bounce-over: When a vehicle rebounds off a fixed object and consequently overturns

• Turn-over: When centrifugal forces from a sharp turn or vehicle rotation are resisted by normal surface friction

• Fall-over: When the surface on which the vehicle is traversing slopes downward in the direction of movement of the vehicle’s center of gravity (COG) such that the COG becomes outboard of its wheels

• Climb-over: When the vehicle climbs up and over an object (e.g., guardrail, barrier) that is high enough to lift the vehicle completely off the ground

• Collision with another vehicle: When an impact with another vehicle causes the rollover

• End-over-end: When a vehicle rolls primarily about its lateral axis after crashing with a concrete barrier.[13]

ROCs of heavy vehicles based number of turns

• Turn on side (¼ rotation): The vehicle generally slips a certain distance on its side and finally stops

• Turn into a ditch: The rotation is somewhat between one-fourth and a half and the depth of the ditch is enough to stop further rotation.

• Rollover from the road: More than half rotation, but not more than two. The level difference between the road and the ground, where the vehicle finally stops is not more than 10 m

• Serious rollovers involve more than two rotations and the level difference between the road and the ground is more than 10 m, where the vehicle finally stops

• Combined rollover: The rollover followed by fire, or before the rollover a severe frontal collision occurred, or after the rollover, the vehicle falls into a water body. In general, more than 90% of all ROCs rotate about the longitudinal (roll) axis of the vehicle.[14]

Rollover of Heavy Vehicles

ROCs of heavy trucks are particularly serious and common subtype on highways. Unlike passenger vehicles, ROCs of heavy trucks are associated with significant damage to the vehicle and injuries to the occupants, even those who are restrained. It has been estimated that on an average 35% of incapacitating and fatal injuries occur during ROCs of heavy trucks annually in the US.[17]

Analysis of Rollover Injuries

Post-roll events are frequently (66%) associated with severe injury followed by pre-roll and under turn ROCs.[18] Moreover, the unrestrained occupants had severe injuries due to intrusion by rollover in the majority of cases. Complex injuries of head/neck and hemo/pneumothoraces were the most significant injuries sustained during serious rollover events.[18]

Impact: In light vehicle rollover event, the most common cause of injury for non-ejected occupants is the impact of the event rather than crushing. During a rollover, impact corresponds to relatively short duration and involves only the striking and the struck objects.

Crushing: Represents a slower process involving the object and two other surfaces between which the object is being crushed. Crushing can take all day; on the other hand, even a slow impact is still capable of quick bump or bang. The rollover event is largely governed by the upsetting force (depends upon geometry and weight of the vehicle) which should persist long enough, to rolls over the vehicle.[18]

Contributing factors determine the severity of ROCs

These factors include (1) Type of vehicle, (2) pre-crash speed, (3) restraints used, (4) number of turns, (5) intensity of the impact, (6) vehicle damage especially roof intrusion in relation to survival space (the physical envelope in which the motion of an occupant is contained during a crash), (7) single or multivehicle event, (8) type of rollover initiation, (9) vehicle design, (10) field triage models, (11) age of the occupant, (12) occupant size body mass index, (13) location of occupant in the vehicle and (14) whether occupant was ejected or confined.[13,19-21]

Furthermore, the roof impact and availability of survival space are important indicators of severity in a single vehicle crash and provides a uniform relationship between crash severity and injury risk.[19] For non-restraint occupants, the ejection risk increases with the number of quarter-turns. On the other hand, for non-ejected unbelted occupants, one quarter-turn is commonly associated with severe injuries in single vehicle crashes.[19] An increased injury risk was also observed in a crash with a stationary object (tree or wall) prior to rollover. Also, multivehicle rollover event possesses greater risk of injury compared with a single vehicle crash. The injury severity for children in a ROC was also significantly higher than for non-ROCs. In crashes involving children, a greater risk of rollover observed for pickups and SUVs compared with passenger cars and minivans.[22]

Roof strength as risk of injury in ROCs

Though, some studies on crash analyzes have not established a strong association between roof deformation and injury of occupant.[23] Vast majority of studies suggested that maintenance of “survival space” in the vehicle is an important factor for crash outcome. Earlier studies have demonstrated a significant correlation between roof intrusion (>10 cm) and increased injury severity.[24-27] Moreover, single unit increase in roof strength-to-weight ratio demonstrated 20-25% reduction in serious injury during rollovers. A study by Mandell et al.[28] showed that the risk of mortality and severe head/spinal injuries increases significantly with an increasing degree of roof crush during ROCs. The risk of mortality was twice with a roof crush of >15 cm which increases to six-fold for roof crush of >30 cm. Similarly, the risk of traumatic brain injury increases more than three-times and spinal injuries more than two-fold for roof crush of above 30 cm.[28] Burns et al.,[29] demonstrated that by controlling the severe roof intrusion, it is possible to minimize the spinal cord injuries and in turn can reduce substantial direct cost. Therefore, improving the roof strength of a vehicle might help in minimizing the severity of injury. In addition, Inamasu and Guiot [30] have emphasized the efficacy of seatbelt use in minimizing the severity of rollover-induced thoracolumbar junction injury. About one-third of the contained and restrained occupants suffered serious thoracic injuries in single-vehicle ROCs.

Vehicle design as risk of injury in ROCs

It is shown that the fatality rates vary significantly based on vehicle types and makes and are largely associated with the vehicle stability, which shows the importance of vehicle design role in minimizing the morbidity and mortality, resultant from rollovers.[31] Most crashes have a high velocity at the initial rollovers and the occupants had high impact with the roof and doors due to massive displacement. There is a significant risk of severe injuries in absence of roof integrity particularly seen in vehicles equipped with four wheel drive.[31] Some seatbelt designs may be deficient and unlatched during the rollover or provides little restraint against partial ejection.[31] Moreover, earlier studies have demonstrated that some occupants were died due to front airbags in low-severity crashes which was supposed to impact minor or no injuries.[32,33] Furthermore, improper padding of roof structures and framing results in severe head injuries, mainly involving scalp lacerations, fractures of skull and brain injury.[31]

Vehicle design-related factors such as mechanistics, arising from vertical and lateral roof intrusion and impact loading of the head with the ledge formed by the underside of the roof and door frame leads to severe head and spinal injuries.[31] In general, roof crush together with deformation of other vehicle structures indicates the severity of rollovers. However, only roof crush without other structural deformities reflects defect in vehicle design. O’Neill [34] showed that newer vehicles designed for crash avoidance (with electronic stability control [ESC]) have the potential in reducing casualties in single-vehicle crashes.

Rechnitzer and Lane have recommended certain modification in vehicle design for improved safety.[31] To prevent partial ejection from the vehicle, plastic glazing should be used to maintain the integrity of side window. Minimum standard for roof integrity should be maintained by increasing roof framing and A and B pillar strength.[31] To avoid head contact, interior energy absorbing padding over the roof should be used. Modify the design of door/roof to avoid the risk of head stuck against the frame. Improve the performance of seat belts to reduce vertical movements of occupants. Enhance the integrity of doors together with energy absorbing side padding. Figures 1 and 2 show types of roof deformation.[13,35]


Figure 1: Types of roof deformations: Front and rear view[13,35]


Figure 2: Types of roof deformations: Side view[13,35]

ROCs among Children and the Relation to Vehicle Type

Children involved in MVCs with rollover events have two-fold increased risk of morbidity or mortality.[4] Furthermore, MVCs particularly involving SUVs have 11 times increased risk of rolling over compared to the passenger cars. In a study by Rivara et al.[4] observed that around 60% of children involved in ROCs were travelling in SUVs. However, Evans and Frick [36] demonstrated an inverse relationship between mass of the vehicle and fatality risk. Hence, SUVs being heavier vehicles compared to passenger cars possess a lower risk of mortality in rollover events.[4] Hence, the overall risk of mortality for children undergone SUVs rollover may not increase in comparison to the passenger cars and may in fact be lower;[4] this might be due to better safety measures, separate protected seats for children and stronger frames. Rasouli et al.[37] observed MVCs to be the major cause of traumatic spinal injuries (29%) in children. Moreover, single vehicle crashes, rollover event and ejection from the motor vehicle are the major risk factors for MVC-related spinal injuries in children. Olsen et al.[38] showed that the driver seatbelt use was associated with child restraint compliance and reduced risk of ROCs and head-on collisions.

Ejection and Confinement and Rollovers

Ejection is the process in which an occupant is either partially or fully thrown away from a vehicle resulted in ground contact and other surfaces exterior to the vehicle.[13] Serious injuries occur due to entrapment of the occupant between the road surface and the vehicle structure. Berg et al.[35] in their study have reported that occupants who were ejected had torso and head injuries mainly due to secondary impact with the ground or by entrapment.

Whereas, the non-ejected occupants experiences head, spine and extremities injuries primarily by roof crush or impact of the interior on the occupant. Sustaining an intact survival space (minimal roof crush) together with proper seatbelt restraint would be the beneficial in controlling occupant movement and subsequent ejection.[39] Non-compliance of vehicle restraints is the major cause of ejection. An earlier study, observed 70% compliance of seat belts among injured non-ejected occupants. On the other hand, 51% of the partially ejected and only 3% of completely-ejected occupants used restraints.[40] Table 1 shows comparison between rollover with and without ejection.[3,35,40]

Variable Ejected Non-ejected
Severely injured cases[35]* (%) 50 5
Seatbelt use[40] (%) 3 70
Relative HARM 6.2 0.5
(injury cost×number of accident)[3]  

Table 1: Comparison between rollover with and without ejection

Ejection and Fatalities

In a rollover, there is a five-fold increased risk of mortality, if the occupant ejected during the crash. It was also suggested that the fatality rate could be reduced by 70% by effective controlling of ejection in rollover.[19] Another study reported that even in less severe ROCs, two-thirds of the mortalities were attributed to occupant ejection from the vehicle.[41] Deutermann [42] also found ejection as the primary cause of mortality (62%) in severe ROCs. According to the Advanced Glazing Project of National Highway Traffic Safety Administration, which analyzed ejection ROCs, a relative risk of fatality for ejected to non-ejected occupants was found to be 3.55 for drivers and 3.15 for front seat passengers. So, controlling ejection of drivers significantly helps in reducing the rate of morbidity by 58% and mortality by 72%, respectively.[43] Howard et al.[44] in their study have reported that ejection from the vehicle is frequently associated with mortality (29%) among young children. Similarly, Scheidler et al.[45] found ejection to be associated with increased morbidity and mortality among children of all age groups. Further, infants and young children (0-4 years) had the highest fatality rate and were mostly found to be either unrestrained or improperly restrained. Therefore, factors involving increased risk of ejection needs to be analyzed for controlling significant risk of injury and mortality in ROCs with ejection. The relative risk of mortality in ejected occupants ranged from about 1.5 to 8 depending upon the crash mode or type. In addition, single-vehicle ROCs with ejection have the highest increased risk of mortality.[46] Table 2 shows studies of ROC and outcomes.[2,22,30,44,45]

Authors Age(years) Type ofstudy ROCcases Ejection/unrestrained % Mortality % Findings
Aldaghlas >16 Retrospective 257 -/40 - Use of seatbelt and physiometric data could be useful
et al.[2]           for triage of non-critically injured rollover occupants
            which save the cost of trauma center
Kallan =15 Retrospective 2461 -/2.3 - The risk of injury to the child occupants in rollovers was
et al.[22]           significantly higher than in non-rollover crashes
Scheidler 0-16 Retrospective 2298 8.2/88 14.8 Most children ejected from MVCs were either
et al.[45]           unrestrained or improperly restrained. Head injuries
            were the most common cause of death in all age groups
Howard 0-16 Retrospective 1832 29/81 64 Ejection from the vehicle is common among fatally
et al.[44]           injured children. Shoulder straps alone may not prevent
            the ejection of toddlers during rollovers
Inamasu 17-64 Retrospective 22 55/55 - Seatbelt helps in reducing the severity of
and Guiot[30]           rollover-induced thoracolumbar junction injury

Table 2: Studies of rollover with and without ejection

Significance of Rollover Direction on Ejection and Injury Risk

The kinematics of vehicle occupants is considerable affected by the direction of the rollover, magnitude of forces and position of the occupant.[47] Rollovers involving more numbers of quarter turn and far-side occupants had increased risk of serious injuries and fatality for unrestrained ejected occupants.[48] An earlier study observed that irrespective of the restraint use, the potential risk of severe head and thorax injury was high for crashes involving two and more complete rolls.[49]

Parenteau et al.[48] have reported that occupants of far-side sustained serious injury compared to near-side occupants. They also concluded that head injury was more common in drivers of far-side than near-side. Head, lower extremity, thorax and upper extremity injuries were frequently observed in near-side drivers. In contrast, occupants of far-side sustained spine, head and thorax injuries. Drivers of near-side rollovers were mainly through the driver’s door window opening. On the other hand, drivers in far-side rollovers were ejected either from the driver’s door window opening or through opening of glass roof.[49]

Safety Belt Use and Ejection

Use of seatbelt potentially reduces the risk of complete ejection during a ROC. Moreover, the risk of partial ejection is also substantially minimized by restraints.[50] For unbelted occupants, the relative risk of severe injury increases by 20 times whereas; in restraint occupants the injury risk increases to 77-folds for complete ejection in comparison to non-ejected occupants.[51]

Prevention of ROCs

Every year, MVCs cause significant mortality and morbidity with 40,000 deaths and 4.3 million non-fatal injuries world-wide.[52] Several epidemiological studies have proposed effective strategies to address these preventable deaths and injuries which are mainly associated with socio-economic factors. However, occupant ejection remains a critical issue in ROCs. The important factors that play a crucial role in ejection prevention involves seatbelt use, lesser roof inversions, vehicle design, curtain airbag deployment, near-side seating position and small occupant size.[53]

Based on preliminary data, side-curtain airbag deployment appears to be a promising ejection mitigation counter measure. Further reductions in ejection could potentially be achieved by blocking ejection portals in the roof.[53] Currently, cars are equipped with ESC which poses pressure to individual wheel brakes during the situation of no response from the driver’s steering (a precursor to rollovers).

Rollover prevention technology also provides a method for estimating the COG position in real time.[54] The timely and accurate estimation of COG is critical for triggering the appropriate vehicle safety mechanisms.

Apart from technological advancement, safety measures such as seatbelt compliance, avoidance of panic maneuvers and loading too much gear onto a roof rack, excessive speed around ramps/curves in the highway and excessive speed in slippery conditions would help in preventing ROCs.

Moreover, public awareness through educational programs has enormous potential to reduce these serious crashes. Driving while engaged in other activities which include using a cell phone, texting, eating, or reading, compromise the safety of the vehicle occupants and other individuals on the road.[55] According to the United States Department of Transportation, text messaging while driving had 23 times higher risk of crash than driving without distraction.[56] Despite these statistics, 37% of drivers in the USA have sent or received text messages while driving and 18% admit doing so regularly.[57] Distraction while driving is broad and multidimensional concept and we just hinted on this issue as it is very common in our region and world-wide as it is one of the major causes of ROC.


Though, it is difficult to determine the exact correlation between the injury severity and ejection and whether the injuries sustained were before or after the ejection. Studies showed that vehicle ROCs with ejection are particularly associated with higher rate of morbidity and mortality. The major causes of occupant ejection include lack of seat-belt restraints, average speed and age of driver and occurrence of rollover. So, seatbelt use, lesser roof inversions, vehicle design, near-side seating position and small occupant size are the important factors for ejection prevention. Furthermore, the majority of vehicle ROCs is lateral and the frequent injury mechanisms are roofing intrusion, projection, complete and partial ejection. It is noteworthy that the roof impact and availability of survival space are important indicators of severity in a single vehicle crash. Despite continuing innovation in vehicles’ safety, human behavior (distracted driving, speeding and drinking) is an important factor that increases the chances of ROCs. Furthermore, the risk of fatality in ROCs largely depends upon the intensity of collision, occupant protection and ejection status. In order to minimize ROC with ejection, all the aspects should be addressed including car design, driver behavior, the occupants and road. Therefore, public awareness through educational programs has enormous potential to reduce these serious crashes. The current review addresses all these issues in a simplified language that can be easily available for the concerned physicians.

Source of Support: Nil.

Conflict of Interest: None declared.


20+ Million Readerbase
Abstracted/Indexed in

  • Include Baidu Scholar
  • CNKI (China National Knowledge Infrastructure)
  • EBSCO Publishing's Electronic Databases
  • Exlibris – Primo Central
  • Google Scholar
  • Hinari
  • Infotrieve
  • National Science Library
  • ProQuest
  • TdNet
  • African Index Medicus
Annals of Medical and Health Sciences Research The Annals of Medical and Health Sciences Research is a bi-monthly multidisciplinary medical journal. more >>
Submit your Manuscript