Common crash sequence across disciplines

The athletes impacted the snow surface with their skis or board first, followed by the upper or lower extremities, buttocks/pelvis, back and, finally, the head (n=48, 84%). This information is important to increase the ecological validity of future head impact injury reconstructions. For example, a previous laboratory reconstruction of snowboarding back edge catches with anthropomorphic test devices presented them as being flipped up in the air after the edge catch, with the hips and spine in full extension and landing directly onto the head.17 This is not a realistic reconstruction of a snowboarding back-edge catch event, based on our findings (see figure 4). The identification of this crash sequence may also be important for further development of wearable ski-racing airbags, specifically in relation to airbag deployment, that is, the triggering algorithm. Airbags were first used in official FIS WC races in the 2015/2016 season. However, further design improvements may be possible, particularly with respect to protecting the cervical spine and head in backward pitching falls, as described in the current paper.

Sideways falls common in alpine skiing

Among alpine skiers, sideways falls were common (45%). Two common patterns were observed. The athletes were either mainly out of balance in the frontal plane (roll) in air during flight, falling to the left or right hand side, impacting the side of the helmet, or the athletes landed mainly out of balance in the transverse plane (yaw) after flight, subsequently catching the ski edge and tripping. Being tripped, the athlete then fell sideways, also impacting the side of the helmet (figure 3).

Our findings are slightly contradictory to a recent study investigating head injury mechanisms in recreational skiers and snowboarders, where hospital data were combined with a survey based on sketches depicting the crash and impact locations.13 Bailly et al reported that ‘Falling head first’ while skiing was the most common injury mechanism (28%), followed by ‘Falling sideways (catching the ski edge)’ representing 19% of skiers’ falls.13 The two main head impact locations were the frontal (57%) and facial (41%) areas.13 However, for the sideways falls, they reported that 28% of head impacts were to the occipital region, which is similar to our findings.13 This indicates that mechanisms of falling may be somewhat different between recreational and WC skiers, with recreational skiers having more impacts to the front/face and falling head first, although catching the ski edge and falling sideways has been identified as a common injury mechanism at both levels.

Backwards pitching falls common in freestyle and snowboard

Backwards pitching falls were the most common among snowboarders and freestyle skiers in our study. We observed two previously described types of backward pitching falls among our freestyle and snowboarding cases: ‘slapback’ injuries and back-edge catches.19 32 The gross injury mechanism in 62% of our snowboarding cases was a ‘back-edge catch’ (opposite edge catch), which is previously described as a common head injury mechanism in snowboarders.32 33

Backward pitching falls were frequently observed in alpine skiing as well (35%). Bailly et al reported that falling backwards represented 14% of falls in recreational skiers, with the impact location being the occipital region in 73% of the backwards falls.13 Backwards pitching falls may therefore be more common among WC alpine skiers compared with in recreational skiers.

Helmet ejections in alpine skiing: cause for concern

Among the alpine skiing cases, there were three helmet ejections at head impact (10% of cases). From the visual analysis, the cause of the helmet ejections cannot be determined. It could be that the helmet did not fit adequately, was not securely fastened or that the loads of the crash exceeded the stability of the helmet/strap.

One of the main requirements of a helmet is to provide and maintain appropriate and adequate coverage to the head, and a helmet that is poorly fitted or fastened may become displaced during normal use or even ejected during a crash.34 Among cyclists, a recent study investigated the fit of helmets and reported that bicycle helmets worn by recreational and commuter cyclists are often the wrong size and often worn and adjusted incorrectly.35 In addition, among motorcyclists, helmet type and wearing correctness were among the factors that affected the loads at which helmets became displaced.34 However, the athletes in the current study were supported by professional teams and therefore likely received optimal advice and optimally fitted helmets from their equipment suppliers. Therefore, the helmet ejections observed represent a concern.

Many cases of inappropriate gate contact

In the alpine skiing cases, over 40% of the athletes had inappropriate gate contact, which threw the skier out of balance and ultimately led to the crash. In most cases, the gate contact resulted from a personal mistake of the skier (misjudging the turn/skiing line or having an inappropriate course line) and therefore hooking the gate with the upper extremity, impacting the gate panel, or straddling the gate with the inner ski. This is supported by previous video analysis of WC alpine skiing injuries in general, where in 30% of cases inappropriate gate contact caused the injury situation.29 From the 2010/2011 season, FIS enforced the use of release gate panels, which must release from the pole when the athlete collides with the gate,11 yet further design improvements may be possible.

Head impacts and impact location

Although most athletes experienced one head impact, many athletes (28%) experienced two, and in alpine skiing, some (21%) even more than two impacts. Alpine ski helmets have been demonstrated to provide protection against low-severity repetitive impacts, such as impacting slalom gates.36 However, ski helmet liner materials exhibit degradation in performance for substantial repetitive impacts,37 which may be an important consideration for helmet manufacturers with respect to helmet design and construction, although we did not detect an association between the number of head impacts and injury severity. However, we do not know whether the helmets used had suffered previous impacts.

Few impacts were to the front of the helmet or the face (16%); however, the face is mostly unprotected. In fact, we observed two cases in freestyle aerials where the athletes did not crash, but impacted their face onto their own knees, one suffering an orbital blow-out fracture. In contrast, at the recreational level, facial bone fractures and dental injuries are reported among male snowboarders and skiers to occur most frequently after falls or collisions with other persons.38 39

Head impact location, mainly to the back and side of the helmet, and impacting snow/ice (83%), may be important information for helmet manufacturers, as at the recreational level collisions with stationary objects or other skiers/riders might be more common.40–43

Helmets continue moving postimpact

In half of the cases, the helmet slid along the surface postimpact. A variable to evaluate helmet rebound motion up from the snow surface postimpact was not included in the video analysis form. In previous reconstructions of skiing and snowboarding head impact injuries, both linear and angular velocity changes indicated that there was a rebound phase immediately postimpact, which might not be anticipated in an impact with a compliant surface such as snow.24 Although helmet rebound was not specifically investigated in this study, the helmet was observed to not stop moving postimpact in most cases.

Methodological considerations and limitations

The current study sample was derived from a systematic, prospective collection of injury videos over a 10-year period (2006–2016) based on the FIS ISS. We managed to acquire videos of 85% (29/34) of all WC alpine head and face injuries, ensuring that our sample of alpine injury videos is representative. However, we could only obtain videos of 28% (13/47) and 36% (15/42) of snowboarding and freestyle skiing head and face injuries, respectively, for the same period. This was mainly due to injuries not being videotaped by the television producer or the injury situation was not visible on the video. In addition, many head and face injuries in snowboard and freestyle skiing occur during qualification runs, which are not broadcasted. Therefore, the data from freestyle skiing and snowboarding should be interpreted with caution. Nevertheless, our findings parallel previous epidemiological literature.

We did not detect any difference in injury severity between the cases with and without video available; this suggests that the sample we were able to analyse is representative.

The injury recording was through interviews with athletes, medical personnel or coaches. Recall bias is a challenge with retrospective interviews. However, a methodological study found that in the WC setting, retrospective interviews was the best method compared with prospective injury registration by team medical personnel or FIS Technical Delegates.28 Interview forms based on the race schedules were used to help the interviewee recall the date, location and circumstances of injury.28 However, a limitation is that we did not have access to more detailed medical information, for example, the results of imaging studies done or standard severity scores such as Glasgow Coma Scale or Abbreviated Injury Scale.

A greater problem could be that concussions are not recognised by athletes, coaches or medical personnel, and therefore are under-reported. Athletes might not self-report an injury they do not recognise as being harmful or dangerous at the time of competition.44 45 From other sports, it is known that concussions are under-reported to a large extent.46–49

Video quality and available camera views represent a challenge when determining the head impact frame and when assessing the gross injury mechanisms. It should be noted, however, that the assessment was consistent across analysts.

We included videos where the head impact frame was not visible. However, we could still perform an accurate analysis of the gross body biomechanics leading up to the head injury, which provides novel and valuable information.

Further perspectives

Based on the three helmet ejections we observed, it seems prudent to ensure that FIS WC athletes have optimally fitting helmets, which are fastened correctly, as this could potentially be an area of improvement with respect to athlete safety.

Our observation that the helmet continues moving postimpact, combined with findings from previous papers24 25 describing a linear and angular rebound motion up from the snow surface, could be an important consideration for helmet manufacturers. Both the helmet and the snow impact surface may contribute to rebound, and future helmet standards could potentially address these issues.24 25

Based on information about real gross head injury mechanisms, future biomechanical studies could reconstruct realistic crash sequences, as this might help our understanding of the comparability of laboratory reconstructions or computer modelling and real head impact injuries on a snow surface.

FIS has developed gates with panels/poles offering less resistance or with an optimised release mechanism when hooking. This effort should continue based on the high number of inappropriate gate contacts that lead to head injuries (and knee injuries)29 in alpine skiing.

In snowboarding and freestyle skiing, most head injuries occurred during landing from a jump or when crashing while passing an element. The primary focus for course design should therefore be on safe jump and landing constructions, and on the design of elements, such as banked turns. Several previous studies using computer modelling techniques have investigated if the creation of safer terrain park jump designs that reduce the risk of impact injuries is possible.50–53 In particular, it has been discussed if the severity of impact risk can be characterised by equivalent fall height, a measure of jumper impact velocity normal to the slope.51 The thought is that the smaller the equivalent fall height, the smaller the probability of serious injury resulting from impacts normal to the snow surface.52 However, the crash sequence we described, with the skis/board having initial contact, followed by the extremities, buttocks, back and lastly the head, could mean that not only the normal-to-slope equivalent fall height could be  of importance to the impact severity, but this pitching motion could possibly also contribute to the severity of head impact injury. We therefore reiterate the necessity of future biomechanical studies to reconstruct crash sequences realistically.

In alpine skiing, safe course design in general, and not only for jumps, must be a priority. Further investigations into the reasons athletes make mistakes during turning, and into the causes of inappropriate gate contact, are therefore warranted, in addition to addressing jump safety. Spörri et al  reported that the main perceived risk factors among alpine expert stakeholders were system ski, binding, plate and boot; changing snow conditions; physical aspects of the athletes; speed and course setting aspects and speed in general.54 Gilgien et al reported that in fall or crash situations, the magnitude of speed is of particular importance since speed determines the kinetic energy that has to be dissipated during a crash impact.55 In technically demanding sections such as jumps, rough terrain and turns, anticipation and adaptation time decrease with speed and mistakes might be more likely to occur.55 Simulation models of jump landings in WC downhill skiers suggested that limited preparation time, high take-off speeds, steep take-off angles and landings in flat terrain had the most influence on landing impact injury risk.56 It therefore seems reasonable to suggest that reducing skier speeds especially during turns and terrain transitions, and focusing on optimal safety jump design would reduce injury risk.