Article Text
PDF
Abstract
Background Concussion is the most common injury in youth ice hockey. Whether mouthguard use lowers the odds of concussion remains an unanswered question.
Objective To determine the association between concussion and mouthguard use in youth ice hockey.
Methods Nested case–control design. Cases and controls were identified from two prospective cohort studies using valid injury surveillance methods. Cases were players concussed during a game or practice; controls were players who sustained a non-concussion injury during a game or practice. The primary exposure was mouthguard use at time of injury; mouthguard type (dental custom fit or off the shelf) was a secondary exposure. Physician-diagnosed or therapist-suspected concussion was the primary outcome. Dental injury was a secondary outcome. Multilevel logistic regression with random effect at a team level was used to obtain ORs for the mouthguard effect, adjusted for level of play, age group, position, concussion history, mechanism of injury, cohort, session type and body checking policy.
Results Among cases, 236/315 (75%) were wearing a mouthguard at time of injury, while 224/270 (83%) controls were wearing a mouthguard at time of injury. Any mouthguard use was associated with an adjusted OR for concussion of 0.36 (95% CI 0.17 to 0.73). Off-the-shelf mouthguards were associated with a 69% lower odds of concussion (adjusted OR: 0.31; 95% CI 0.14 to 0.65). Dental custom-fit mouthguards were associated with a non-significant 49% lower odds of concussion (adjusted OR: 0.51; 95% CI 0.22 to 1.10). No dental injuries were identified in either cohort.
Conclusion Mouthguard use was associated with lower odds of concussion. Players should be required to wear mouthguards in youth ice hockey.
- concussion
- Injury prevention
- Ice hockey
Statistics from Altmetric.com
Introduction
Among youth ice hockey players, concussion is the most frequently sustained injury.1 Mouthguards are commonly worn in youth ice hockey and were developed in the 1890s by a British dentist to prevent orofacial injury during boxing matches.2 3 Mouthguards lower the incidence of dental and maxillofacial trauma in sport.4–11
A 2017 systematic review and meta-analysis suggested a protective, though not statistically significant (incidence rate ratio 0.81; 95% CI 0.60 to 1.10), effect of mouthguards on concussion.12 The literature on the protective effects of mouthguards is of varying methodological quality and the findings are inconsistent.4 5 13–18 Stenger’s case series, which suggested that mouthguards protect against concussion, is widely cited in the dental literature, but is limited by sample size and the lack of a control group.19 Additionally, mouthguards may either be dentist custom fit or purchased off the shelf (ie, boil and bite). The effect of mouthguard type on concussion risk has been investigated in a few studies.2 20–23 Only two studies suggested a lower risk of concussion with custom-fit mouthguard use.2 22
The current published literature has limitations with respect to methodological quality and a high risk of bias. Specifically, many studies lacked a control group and those with comparison groups often did not simultaneously adjust for important relevant covariates that are known to affect concussion risk including prior concussion, competitive play, body checking rules, player age and sex and level of play.2 6 12 19 24 Further, no evidence exists examining the relationship between mouthguard use and the risk of concussion and dental injuries specific to youth ice hockey. In 2013, referees in the Calgary, Alberta area, and across Canada, stopped enforcing a mandatory mouthguard use policy,25 while the governing body of Hockey Calgary still mandated their use.26 The challenges with enforcement leading up to and after the rule change allowed an opportunity to evaluate the effect of mouthguards on concussion risk in youth ice hockey players.
The primary objective of this study was to examine the association between concussion and mouthguard use in youth ice hockey. Our secondary and exploratory objectives were to (1) examine the association of mouthguard type on the outcomes of concussion and dental injury, (2) the association between mouthguard use at time of concussion on time to full return to play, and (3) the association between mouthguard use at time of injury and dental injuries.
Methods
Design, setting and participants
A nested case–control design was used. Cases and controls were identified from two prospective cohort studies employing the same validated method of injury surveillance in youth ice hockey.1 The first cohort (2011–2012) was a single season cohort of Bantam (13–14 years of age) and Midget (15–17 years of age) ice hockey players, all in a higher level of play (n=778); the second cohort was a three-season cohort from 2013 to 2016 which comprised Pee Wee (11–12 years of age), Bantam and Midget players of all levels of play (n=1577). Cases were defined as those who sustained a suspected concussion during a game or practice. Controls were players who sustained a non-concussion (eg, trunk or extremity orthopaedic) injury.
The primary research objectives for both original cohort studies were different from that of the current study on mouthguard use and concussion risk27–29 However, the data from both cohorts provided an opportunity to examine the issue of concussion and mouthguard use as the data on mouthguard use were available, consistently collected and the timing was fortuitous in light of the policy climate. As such, there was no a priori sample size calculation.
Procedures
A team designate, typically the team safety coach/manager, responsible for recording player participation and outcome information was present at most games/practices and completed the injury report form (IRF) at the time of injury. However, the study therapist reviewed each IRF and followed up in the case of missing data at time of physician assessment (within 72 hours of injury) or by telephone call if the player did not see a study physician.
Exposure
Mouthguard use at time of injury was the primary exposure. After a player injury, the team designate completed an IRF that contained questions on mouthguard use and type at time of injury, mechanism of injury (MOI), healthcare utilisation and time loss. Player baseline questionnaires (PBQ) included information on demographics, medical and concussion history, hockey experience and self-reported mouthguard use and type at the start of the season.
Outcome measures
Concussion definition was based on the third (2012 cohort) and fourth (2013–2016 cohort) Consensus Statements on Concussion in Sport, and was operationally defined as a physician-diagnosed or therapist-suspected concussion.30 31 Injury was defined as an event in which the player either sought medical attention, could not complete the session in which the injury occurred, or missed at least 1 day of sporting activity. Players sustaining a concussion were offered follow-up with a study sport and exercise medicine physician and return to play protocols were standardised. All IRFs were verified by a physical or athletic therapist.
Statistical analysis
All original data were entered into and managed using either StudyTRAX (2011–2012) or the REDCap electronic data capture platform (2013–2016) hosted at the University of Calgary and were analysed using Stata V.14 and V.15, and R.32–35 If a subject sustained more than one injury during the study period, only the first injury was included as either a case or control. Those who sustained injuries outside of on-ice practices or games, or those injuries without an inciting acute traumatic event were excluded (eg, chronic overuse injuries). When mouthguard use, mouthguard type or player position at time of injury was not available, medical interview or PBQ data were used. Per cent agreement between PBQ and IRF reported mouthguard use and type were calculated. Characteristics were described with frequencies and percentages by case/control status and cohort. The proportion of cases and controls wearing a mouthguard at time of injury was calculated. For the primary objective, ORs and 95% CIs were estimated based on the odds of cases wearing and not wearing a mouthguard at time of injury, using a complete-case multilevel, multivariable logistic regression with a random effect at a team level, adjusting for level of play, age group, position, concussion history, clinician-determined primary MOI, cohort, body checking policy at time of injury and session type. Covariates were chosen based on known risk factors for concussion in youth ice hockey.1 27 28 36–39 A similar analysis was performed for type of mouthguard and for dental injury outcomes. Time to medical clearance was described for concussed players wearing and not wearing a mouthguard at the time of injury with Kaplan-Meier survival curves, median, IQR, minimum and maximum. If medical interview time loss data were missing, self-reported time loss or time loss recorded as part of exposure data was used.
Two sensitivity analyses were performed for the primary objective: (1) multiple imputation by chained equations (MICE); and (2) ‘worst case scenario’: assigning missing information from cases to ‘mouthguard use’ and from controls to ‘no mouthguard use’.
Patient and public involvement
Integrated knowledge translation was a priority for the research programme in concussion and injury prevention in youth ice hockey. Community stakeholders from Hockey Calgary and Hockey Canada, as well as youth ice hockey players having personal experience with concussion, were engaged as knowledge brokers in all aspects of the research process for the large cohort study. Knowledge brokers supported the planning, recruitment, data collection and dissemination of results in the youth ice hockey community.
Results
Concussion case and control characteristics are summarised in table 1 by cohort. In the 2013–2016 cohort, 199 concussion cases and 137 non-concussion injured controls were identified; in the 2012 cohort, 116 concussion cases and 133 non-concussion injured controls were identified.
Case (concussion) and control (non-concussion injury) characteristics
Among cases and controls where both baseline and injury values were available, per cent agreement between baseline and injury reported mouthguard use was 72.0% (36/50) at practices and 84.1% (343/408) at games. Per cent agreement between reported mouthguard type at baseline and time of injury was 77.5% (158/204).
Results of the complete-case multivariable logistic regression model with a random effect for team are summarised in table 2. After adjustment for covariates, compared with no mouthguard use, mouthguard use was associated with 64% lower odds of concussion (OR: 0.36; 95% CI 0.17 to 0.73) (n=513).
Results of a complete-case multilevel, multivariable logistic regression with a random effect for team, estimating the effect of mouthguards on the odds of concussion adjusted for covariates (n=513)
Using a similar multivariable modelling approach with a random effect for team (n=492), the OR for dentist custom-fit mouthguards was 0.51 (95% CI 0.22 to 1.10) and the OR for off-the-shelf mouthguards was 0.31 (95% CI 0.14 to 0.65), both compared with no mouthguard use. The OR comparing off-the-shelf with dentist custom-fit mouthguards was 0.61 (95% CI 0.36 to 1.01).
The median number of days of time loss from participation was 17 (IQR: 11–24.5 days; range: 4–78 days) for concussed players not wearing a mouthguard (n=72). Among concussed players wearing a mouthguard (n=229), the median number of days of time loss from participation was 15 (IQR: 9–24 days; range: 0–135 days). Time loss was missing for 14 players. Based on visual assessment of the Kaplan-Meier survival curves, the curves cross and there is substantial overlap of the CIs for each curve, suggesting no evidence of a difference in the time to recovery after concussion between those who wore a mouthguard and those who did not at time of injury (figure 1 and online supplementary figure 1). Since few players had not returned within 50 days of their concussion (n=15), online supplementary figure 1 is truncated to this point while figure 1 shows the complete time loss data.
Supplemental material
Kaplan-Meier survival curves for concussed players wearing and not wearing a mouthguard at time of injury. MG, mouthguard.
No dental injuries occurred in either cohort; therefore, further analysis of the association between this outcome and mouthguard use was not possible.
Sensitivity analyses
(1) When analysing the data with the MICE technique, the adjusted OR for concussion with mouthguard use was 0.28 (95% CI 0.14 to 0.53). (2) When we assumed the most extreme situation where all cases missing mouthguard data were wearing mouthguards, and all controls missing mouthguard data were not wearing mouthguards, the adjusted OR for concussion with mouthguard use was 0.39 (95% CI 0.19 to 0.79) (n=517).
Discussion
In our study, mouthguards were associated with 64% lower odds of concussion in a youth ice hockey population. The direction of the point estimate of our OR was consistent with recent systematic reviews and meta-analyses that reported a non-significant reduction in the incidence of concussion with mouthguard use.12 40 The challenge with applying the findings of past literature to our results on youth ice hockey players is that many previous studies involved adult athletes and were implemented in a variety of sports (ie, basketball, football, rugby and ice hockey).12 Many past investigations were cohort studies. Although cohort studies are often touted as superior to case–control studies, we argue that an injured control group allows a more unbiased evaluation of the effect of mouthguards during an aetiologically relevant time period (ie, use at the time of the injury event). Our study is among the first to report that mouthguards protect against concussion in youth ice hockey.
How might mouthguards lower the risk of concussion?
Several plausible biomechanical mechanisms could explain how mouthguards might lower the risk of concussion. Stenger et al suggested mouthguard use increases the distance of the mandibular condyle from its fossa, thereby reducing force transmission through the temporal bone to the brain when the jaw is hit.19 Having a mouthguard in place may activate neck musculature and reduce head acceleration. Mouthguards may attenuate force transmission following a blow to the jaw and reduce mandibular deformation.41–43 Arguing against these mechanisms, Viano et al found no difference in head injury criterion score using an instrumented artificial mandible skull model with and without a mouthguard in place.44 More biomechanical modelling research is necessary to better inform the mechanism of mouthguard protection of concussion.
Does the type of mouthguard matter? Custom fit versus off the shelf?
Our results suggest off-the-shelf mouthguards are associated with a significant 69% lower odds of concussion. The reduction in the odds of concussion with dental custom-fit mouthguards was not statistically significant; however, if we focus on the 95% confidence limits, there may be up to an 88% reduction, but a maximum of a 10% increase, in the odds of concussion, which could be considered to be clinically relevant. Further, the comparison of off-the-shelf with dentist custom-fit mouthguards was not statistically significant either, but when focusing on the 95% confidence limits, there may be a 64% reduction, but a maximum of a 1% increase, in the odds of concussion. In a study on high school football players, custom mouthguard use was associated with a greater hazard of concussion relative to generic devices (HR=1.69; 95% CI 1.20 to 2.37).20 A 2005 randomised controlled trial found no difference in the incidence of concussion between Canadian collegiate football players who were randomised to use a ‘brain pad’ mouthguard or their standard mouthguard of choice (either custom or off the shelf).21 Based on our primary objective, we recommend that a mouthguard should be worn in youth ice hockey. Based on our secondary objectives, off-the-shelf mouthguards can also be recommended in youth ice hockey.
Baseline and injury exposure agreement
Agreement between baseline and injury reported mouthguard type was substantial, suggesting baseline-reported values may be a valid alternative when injury data are not available. Baseline-reported mouthguard use at practice had lower agreement with mouthguard use reported at time of injury compared with games.
Dental injuries
No dental injuries were reported in either cohort during the study period. These results are consistent with the findings of a recent systematic review indicating that mouthguards significantly reduce the risk of orofacial injuries.44 Dental injuries have previously been identified as prevalent in European ice hockey players of all ages—a population where mouthguard use is low.44
Mouthguard use and dental injury have not previously been investigated in a youth ice hockey population. The absence of dental injuries may also have resulted from mandatory full-face cage or shield policies in place in both cohorts.
Strengths and limitations
Our study has several strengths. The design of this study was retrospective in nature; however, all data were collected as part of prospective cohort studies. Our nested case–control approach allowed us to gather comprehensive covariate data at an aetiologically relevant time period (ie, during the injury event). Since both case and control groups sustained an injury, potential confounding by characteristics of the individual that would make them more or less likely to be injured (eg, risk-taking personality, more aggressive play, MOI) would be similar. In addition, the risk factors collected from the athletes and included in the multivariable model represent those that are established concussion risk factors in the literature. While mouthguard exposure status was self-reported, recall bias was minimised given that the IRFs were completed at the time of injury, at the physician visit or with therapist follow-up call. Risk of selection bias was reduced by identifying cases and controls from the same source population. As such, this study contributes to the literature given its methodological strengths and efforts to mitigate biases relative to other existing literature on mouthguard use and concussion in sport.
We acknowledge several limitations. The cohort studies from which the cases and controls were drawn were not primarily powered to directly evaluate the effect of mouthguard use or type on concussion or dental injuries. Social desirability and the acute effects of concussion may have affected reporting of mouthguard use at time of injury. However, the team designate (or replacement safety coach) completing the IRF should have been present at the time of injury but may not have completed the IRF. If the IRF was not complete the study therapist followed up in the case of missing data at time of physician assessment (within 72 hours of injury) or by telephone call if the player did not follow-up with a study physician. In any case, if present, over-reporting of mouthguard use among concussion cases and non-differential misclassification of mouthguard use among cases and controls would have underestimated the protective effect we found. While some argue protective equipment use may alter risk tolerance among wearers, this would likely have had a similar effect on the concussion case group, and the non-concussed, but still injured control group.44 Finally, proper mouthguard fit and mouthguard quality were not assessed at time of injury or at baseline.
The differences we reported in the protective effects of off-the-shelf and dentist custom-fit mouthguards may be the result of confounding related to player behaviour (eg, more aggressive players or players with a history of concussion may choose dentist custom-fit mouthguards). However, we did adjust for level of play and history of concussion in the analysis. In addition, there was a similar proportion of players with no mouthguard (47%), dentist custom-fit (49%) and off-the-shelf mouthguards (44%) reporting a concussion history making any substantial confounding unlikely.
Conclusion
Mouthguard use was associated with lower odds of concussion in youth ice hockey. Players who used off-the-shelf mouthguards had lower odds of concussion than did players not using mouthguards. Based on the CI, the potential benefit and protective effect of custom-fit mouthguards outweighs the potential harm. We believe that use of a mouthguard, custom fit or off the shelf, should be used in youth ice hockey. Our data add to the case for policy requiring mouthguards in youth ice hockey.
What are the findings?
As part of a case–control study nested within two prospective cohort studies, the adjusted odds of concussion were 64% lower for mouthguard wearers (adjusted OR 0.36; 95% CI 0.17 to 0.73).
Off-the-shelf mouthguards also had a 69% lower odds of concussion compared with not using mouthguards.
How might it impact on clinical practice in the future?
Given that mouthguard use was associated with lower odds of concussion, we recommend policy requiring mouthguard use in youth ice hockey.
Acknowledgments
The authors are grateful to the research coordinators, research assistants and trainees of the Sport Injury Prevention Research Centre, without whom this research would not have been possible. The authors also thank Hockey Calgary, Hockey Alberta, Hockey Canada, study therapists and physicians, team therapists, team designates, coaches, players and parents for their support and contributions.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @aacademic, @Kat_Schneider7, @CarolynAEmery
Contributors DAC contributed to data collection, entry, analysis and manuscript preparation. AMB contributed to data collection and provided data analysis and interpretation support. LPD provided data analysis and data interpretation. KJS contributed to acquisition of funding, study design, and led all aspects of the 2012 cohort. PHE contributed to data analysis and interpretation. CAE contributed to study design, funding acquisition, analysis and interpretation of the results. BEH led the design, analysis and interpretation of data. All authors critically reviewed the manuscript.
Funding This study was supported by funding from the Canadian Institutes of Health Research (CIHR), Alberta Children’s Hospital Research Institute (ACHRI) and Alberta Innovates Health Solutions (AIHS). DAC was funded through the AIHS Summer Studentship Program and the Markin Undergraduate Student Research Program for Health and Wellness Research. The University of Calgary Sport Injury Prevention Research Centre is one of the International Research Centres for Prevention of Injury and Protection of Athlete Health supported by the International Olympic Committee.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. Deidentified participant data are held by Dr Carolyn Emery and the Sport Injury Prevention Research Centre, Faculty of Kinesiology, University of Calgary.