Article Text
PDF
Abstract
Objectives To evaluate prevention strategies, their unintended consequences and modifiable risk factors for sport-related concussion (SRC) and/or head impact risk.
Design This systematic review and meta-analysis was registered on PROSPERO (CRD42019152982) and conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.
Data sources Eight databases (MEDLINE, CINAHL, APA PsycINFO, Cochrane (Systematic Review and Controlled Trails Registry), SPORTDiscus, EMBASE, ERIC0 were searched in October 2019 and updated in March 2022, and references searched from any identified systematic review.
Eligibility criteria Study inclusion criteria were as follows: (1) original data human research studies, (2) investigated SRC or head impacts, (3) evaluated an SRC prevention intervention, unintended consequence or modifiable risk factor, (4) participants competing in any sport, (5) analytic study design, (6) systematic reviews and meta-analyses were included to identify original data manuscripts in reference search and (7) peer-reviewed. Exclusion criteria were as follows: (1) review articles, pre-experimental, ecological, case series or case studies and (2) not written in English.
Results In total, 220 studies were eligible for inclusion and 192 studies were included in the results based on methodological criteria as assessed through the Scottish Intercollegiate Guidelines Network high (‘++’) or acceptable (‘+’) quality. Evidence was available examining protective gear (eg, helmets, headgear, mouthguards) (n=39), policy and rule changes (n=38), training strategies (n=34), SRC management strategies (n=12), unintended consequences (n=5) and modifiable risk factors (n=64). Meta-analyses demonstrated a protective effect of mouthguards in collision sports (incidence rate ratio, IRR 0.74; 95% CI 0.64 to 0.89). Policy disallowing bodychecking in child and adolescent ice hockey was associated with a 58% lower concussion rate compared with bodychecking leagues (IRR 0.42; 95% CI 0.33 to 0.53), and evidence supports no unintended injury consequences of policy disallowing bodychecking. In American football, strategies limiting contact in practices were associated with a 64% lower practice-related concussion rate (IRR 0.36; 95% CI 0.16 to 0.80). Some evidence also supports up to 60% lower concussion rates with implementation of a neuromuscular training warm-up programme in rugby. More research examining potentially modifiable risk factors (eg, neck strength, optimal tackle technique) are needed to inform concussion prevention strategies.
Conclusions Policy and rule modifications, personal protective equipment, and neuromuscular training strategies may help to prevent SRC.
PROSPERO registration number CRD42019152982.
- Sport
- Brain Concussion
- Preventive Medicine
- Risk factor
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Primary prevention strategies in sport can reduce the high burden of concussion such as policy eliminating body checking in child ice hockey.
More evidence is needed to support the protective effect of mouthguards, additional padding in American football helmets, appropriate helmet fit in collision sport, policy limiting contact practice in adolescent American football, head contact rule enforcement in contact sports (eg, American football, ice hockey), and training strategies targeting modifiable intrinsic risk factors.
WHAT THIS STUDY ADDS
Mouthguards are associated with a 28% lower rate of sport-related concussion in ice hockey.
Policy disallowing bodychecking in child and adolescent ice hockey is associated with a 58% reduced concussion rate, without unintended consequences associated with reduced bodychecking experience when subsequently participating in bodychecking leagues.
Strategies limiting contact practice in American football are associated with an overall 64% lower practice-related concussion rate.
A neuromuscular training warm-up programme in rugby is associated with a 32%–60% lower concussion rate.
Current concussion management strategies may reduce recurrent concussion rates.
HOW THIS STUDY MIGHT AFFECT RESEARCH PRACTICE OR POLICY
Mouthguard use should be supported in child and adolescentice hockey.
Policy disallowing bodychecking should be supported for all children and most levels of adolescent ice hockey.
Strategies limiting contact practice in American football should inform related policy and recommendations for all levels.
Neuromuscular training warm-up programmes are recommended based on research in rugby, while more research is needed for females and other team sports. The focus should be on exercise components targeting concussion prevention.
Policy mandating optimal concussion management strategies to reduce recurrent concussion rates is recommended.
Introduction
Primary prevention of sport-related concussion (SRC) is a priority that can have significant public health impact in reducing SRC rates and their potential long-term consequences. The fifth International Conference on Concussion in Sport defined SRC as a traumatic brain injury (TBI) induced by biomechanical forces.1 A 2017 systematic review focused on SRC prevention informing the fifth International Conference on Concussion in Sport highlighted three targets for prevention including personal protective equipment, rules/policy changes and training strategies.2
Globally, there is a one in five lifetime risk of concussion.3 An estimated 3 million people (50% children and adolescents) sustain a concussion in North America annually, 30% are recurrent and 30% remain symptomatic for more than 1 month.3–5 SRC reportedly accounts for 36%–60% of concussions in children and adolescents.6 7 In Canada, one in nine adolescents sustain a concussion annually.8
The strongest and most consistent concussion prevention evidence reported demonstrated a protective effect of policy disallowing bodychecking in youth ice hockey.2 Meta-analyses suggested potential protective effect of mouthguard use in collision sport; however, additional research was needed.2 Additional promising prevention strategies identified in the previous review included thicker mandibular helmet padding and proper helmet fit in American football, rule enforcement to reduce head contact in soccer, larger international ice surface size in elite adult ice hockey and visual training strategies in adult American football players, but required further evaluation. Future research recommendations included rigorous evaluation of SRC prevention strategies using valid injury surveillance with consideration of modifiable risk factors, potential confounders (eg, sex, previous concussion), consistent SRC definitions and exposure data to accurately measure SRC rates.8 Psychological and sociocultural considerations were highlighted for implementation in the uptake and maintenance of SRC prevention strategies.2
The specific research questions for this systematic review and meta-analysis included: (1) What SRC prevention strategies reduce concussion and/or head impact risk (eg, equipment, policy/rules, training strategies)?; (2) Are there unintended consequences of SRC prevention strategies? and (3) What modifiable risk factors are associated with SRC risk?
Methods
Equity, diversity and inclusion statement
The multidisciplinary author group is composed of 15 men and 5 women and represent a blend of junior, mid-career and senior researchers and clinicians. The authors represent several institutions within Canada (11 authors), USA (2 authors), UK (4 authors), South Africa (1 author) and Australia (2 authors). This systematic review included sport participants from any nationality, sex and/or gender, age group and performance level (see the Selection of studies section). Where possible, analyses included consideration of age group, sex and/or gender, and parasport versus able-bodied (see the Data extraction and risk of bias (ROB) assessment section).
Data sources and search strategy
This systematic review reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) guidelines (online supplemental files 1 and 2).9 The protocol was registered on PROSPERO: https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=152982
Supplemental material
Supplemental material
Relevant studies were identified through eight databases:
OVID MEDLINE (R) and Epub Ahead of Print, In-Process & Other Non-Indexed Citations and Daily.
CINAHL Plus with Full Text (Ebsco).
APA PsycINFO (OVID).
Cochrane Databases for Systematic Review (OVID).
Cochrane Central Register of Controlled Trials Registry (OVID).
SPORTDiscus with full text (Ebsco).
EMBASE (OVID).
ERIC (Ebsco).
A search focused on three identified main concepts (ie, concussion/head impacts, sports, prevention/modifiable risk factors) was performed in October 2019 and updated in March 2022. Details of the search strategy are summarised in the methodology for the Amsterdam International Consensus on Concussion in Sport.11 For this review, the search was pilot tested with five identified seed articles and relevant studies from the fifth International Conference on Concussion in Sport,1 then translated to all databases. Searches were limited to 2001–2022. Reference lists of selected systematic reviews were also handsearched to identify additional papers. Only peer-reviewed literature manuscripts were included. The search strategies for all databases are available as supplementary content and the Medline search is annotated (online supplemental file 3).
Supplemental material
Selection of studies
The full text of all potentially relevant studies was independently reviewed by one of two lead authors (CE (senior anchor author) or PE (first author)) and one other author (ATK, MPP, SWW, SBailey, LM, AMB, SBroglio, GAD, BEH, JDS, KAS, MT, RT, NW, RZ and KJS) to determine final study selection. Study inclusion criteria were as follows: (1) contained original human research data full-text studies only; (2) investigated an outcome of SRC or head impacts; (3) evaluated an SRC prevention intervention (eg, protective equipment, rules/policy, training) to reduce SRC and/or recurrent SRC and/or head impacts or modifiable risk factor; (4) participants competing in any sport (excluding recreational activities) including all nationalities, sex and/or gender, age groups and performance level; (5) analytical study design including a comparison group (eg, randomised controlled trial (RCT), quasi-experimental, cohort, case–control, cross-sectional); (6) systematic reviews were included to identify original data manuscripts in reference search and (7) peer-reviewed. Exclusion criteria were as follows: (1) review articles, pre-experimental, ecological, case-series or case-studies and (2) not written in English.
Data extraction and ROB assessment
Data extracted included study design, duration, year, country, participants (eg, sport, level, sex, age), concussion definition, intervention/control or level of modifiable risk factor, concussion incidence rate (IR) or prevalence by study group, and effect estimate (eg, incidence rate ratio (IRR), risk ratio (RR), hazard ratio (HR), odds ratio (OR)) (online supplemental file 4). Effect estimates are reported based on describing a protective effect (prevention intervention) or increased risk (modifiable risk factor). Where not reported and data were available, an effect estimate was calculated. Data were extracted by two authors for each paper (CE or PE) and one additional coauthor. Either consensus was achieved, or a third author (CE or PE) discussed discrepancies. Two authors (CE or PE and other coauthor) independently assessed ROB as per data extraction based on the Downs and Black checklist for methodological quality12 and the study design-appropriate Scottish Intercollegiate Guidelines Network (SIGN) critical appraisal checklists13 that were adapted for the Amsterdam International Consensus on Concussion in Sport process with further detail on the available in the accompanying methodology paper.11 Only studies deemed to be high quality (‘++’) or acceptable (‘+’) based on SIGN criteria were included in results and meta-analyses. Analyses included consideration of child (5–12 years) vs adolescent (13–18 years) vs adult (>18 years) where applicable. Sex and/or gender and parasport versus able-bodied considerations were included where possible. Quality of evidence and grading strength for key recommendations for each research question was assigned (PE and CE) using the Grading of Recommendations Assessment, Development and Evaluation system (online supplemental file 5).14
Supplemental material
Supplemental material
Meta-analyses
Data for MAs were synthesised with summary estimates. RRs, IRRs and HRs were considered comparable and kept for analysis.15 The consolidation of effect estimates was made by recalculating them when they were reported as ORs to reflect IRRs or RRs. When effect estimates were not available, they were calculated based on information provided. RRs were derived using concussion frequency and population from each study group (eg, policy allowing bodychecking vs no bodychecking). When person-time data were reported (eg, player match-hours, athlete-exposures), IRRs were computed. Articles were excluded from meta-analyses if they did not include number of concussions (or head impacts) over either person-time or total number in cohort. A random effects model using the DerSimonian-Laird method computed measures of heterogeneity and adjusted summary effect estimates. All analyses were completed by using Stata V.17.16 Standard errors (SE) were computed using previously described methods.17 18 Forest plots were examined by sport and age.
Results
The search yielded 16 121 studies (figure 1). In total, 220 studies (6/220 (3%) female focused; 115/220 (52%) child and/or adolescent focused) were included for data extraction and ROB assessment and categorised by sport and prevention or modifiable risk factor. Results for unintended consequences are presented with the relevant prevention strategy. Results and discussion for modifiable risk factors are included as online supplemental file 6.
Supplemental material
Study identification PRISMA flow diagram. mTBI, mild traumatic brain injury; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
ROB assessment
A total of 28 studies (13%) had low methodical quality (‘−’) based on SIGN criteria. Of the remaining 192 studies, most (169/192; 88%) were sufficient quality (‘+’) with the remaining (23/196; 12%) excellent quality (‘++’). The median Downs and Black ROB assessment for included studies was 13/33 (range: 6–24). Several common limitations were inadequate reporting on adverse events, description of population representativeness, lost to follow-up, validity/reliability of outcome measures, a priori sample size and insufficient adjustment for potential confounders.
Prevention strategy evaluation studies
Personal protective equipment
Helmets
Comstock et al 19 demonstrated a significantly lower SRC rate from stick/ball contact in male adolescent lacrosse players with mandated hard shell helmets with full facial protection compared with girls allowed (but not mandated) to wear flexible headgear (RR 0.38; 95% CI 0.31 to 0.49).19 Emerging evidence suggests proper helmet fit in adolescent American football20 may reduce concussion symptom severity (p<0.01) and duration of symptoms (p=0.04), as well as reduce SRC odds (OR 0.37; 95% CI 0.15 to 0.96) in children and adolescent ice hockey.21 Five studies examined different helmet types in American football.22–26 Collins et al 22 reported a 31% lower SRC rate (RR 0.69; 95% CI 0.5 to 0.96) in high school football when comparing a helmet with thicker padding over the zygoma/mandible area to traditionally designed helmets. Rowson et al 23 also reported a 46% lower rate of SRC with greater padding (RR 0.54; 95% CI 0.24 to 0.72) in college football players. Two studies did not find a difference in the SRC rate, head impact characteristics or time loss following SRC by helmet type in adolescent football,24 25 however, a 19% lower SRC risk (OR 0.81; 95% CI 0.68 to 0.96) was reported in professional football players wearing National Football League approved helmets compared with players wearing unapproved helmets.26 Greenhill et al 20 compared different helmet liner types (air bladder vs foam or gel) and found differences in individual concussion symptoms, but not in the total number of concussion symptoms reported. One study examined helmet age in high-school American football and found no association with SRC rates.24
Headgear
Fifteen studies examined headgear use in rugby, soccer, lacrosse, Australian football and boxing (online supplemental file 6).27–41 Based on combining data from team-based collision sports, a meta-analysis suggests a potential protective effect between headgear use and SRC rates (IRR 0.84; 95% CI 0.67 to 1.04) (figure 2) but this was not statistically significant. When stratified by sport, headgear was protective against SRC in soccer (IRR 0.64; 95% CI 0.44 to 0.92) but not lacrosse or rugby.
Forest plot based on meta-analysis evaluating headgear use with no headgear by sport (left) and age group (right). IRR, incidence rate ratio; SIGN, Scottish Intercollegiate Guidelines Network.
Face-shields/faceguards
Five studies in ice hockey examined the impact of full, half or no face-shields (online supplemental file 6).42–46 Three studies demonstrated that full-face shields were not associated with lower SRC rates compared with half-visors.42–44 Benson et al 42 reported that players who sustained an SRC while wearing a half-visor had more time-loss than players wearing full-face shields. No association was found between SRC odds with half-visor use compared with no visor in professional ice hockey players (OR 1.34; 95% CI 0.72 to 2.48).45
Eyewear
Protective eyewear was examined in three studies across lacrosse and field hockey.47–49 Lincoln et al 47 compared IRs before and after policy mandating protective eyewear in female high school lacrosse players. Despite lower eye and overall head/face IRs after policy change, the SRC rate increased (IRR 1.6; 95% CI 1.1 to 2.3). Two studies examined protective eyewear and eye injuries among high school field hockey players and reported a reduction to head/face injuries, but no reduction in SRC rates (IRR 0.96; 95% CI 0.57 to 1.59; IRR 0.77; 95% CI 0.58 to 1.02).48 49
Mouthguards
The protective effect of mouthguards was evaluated in eight studies across several sports with conflicting results.27 28 31 34 50–53 Five studies included in a meta-analysis demonstrate that mouthguard use was associated with a combined 26% lower SRC rate (IRR 0.74; 95% CI 0.64 to 0.85) (figure 3). This protection, however, was only significant in ice hockey (IRR 0.72; 95% CI 0.60 to 0.87) but not in rugby (IRR 0.80; 95% CI 0.51 to 1.27) when stratified on sport (figure 3). Protection was significant across mixed age groups (children, adolescents and adults). Dentist fit mouthguards were not associated with additional SRC protection compared with off-the-shelf types,24 50 54 nor were other specialised mouthguards.55
Forest plot based on meta-analysis evaluating mouthguard use with no mouthguards by sport (left) and age group (right). IRR, incidence rate ratio; SIGN, Scottish Intercollegiate Guidelines Network.
Jugular vein compression collars
Two studies examined the use of a jugular vein compression collar and head impacts in adolescent ice hockey and American football players demonstrating that head impact frequency and severity were not reduced by wearing a collar.56 57
Policy, rule or law changes
Eight studies examined the effectiveness of policies disallowing bodychecking in child and adolescent ice hockey and subsequent SRC rates.58–65 Combining studies with individual level injury and exposure data and a study using hospital-based surveillance demonstrated a 58% lower SRC rate where policy disallowed bodychecking (IRR 0.42; 95% CI 0.33 to 0.53) (figure 4). Further, prior bodychecking experience in games was not associated with lower SRC rates in leagues permitting bodychecking, suggesting no unintended consequences.66 67
Forest plot based on meta-analysis evaluating bodychecking policy change in youth ice hockey. IRR, incidence rate ratio; SIGN, Scottish Intercollegiate Guidelines Network.
Fair play rules (additional points for not exceeding a predetermined number of penalty minutes) in 11–14 years old ice hockey led to a reduction in head impacts (RR 0.24; 95% CI 0.07 to 0.78) but not SRC rates.68 No association was found between team penalty minutes per game and their opponents’ game-related SRC rate.69 However, implementation of game suspension for exceeding a threshold for penalty minutes was associated with lower odds of SRC (OR 0.44; 95% CI 0.23 to 0.85).70
The application of a rule that made targeting an opponent’s head illegal in professional ice hockey (Rule 48; National Hockey League) was not associated with a reduced SRC incidence in the NHL.71 In ages 11–14, a zero tolerance for head contact rule did not reduce SRC rates (IRR (range) 1.85–7.91)) as rates were higher following this change.72 Despite this rule change, there was no difference in primary (direct player-to-player) (IRR 1.05; 95% CI 0.86 to 1.28) and secondary head contact rates (head contacts to the boards, glass, net or ice surface) (IRR 0.74; 95% CI 0.5 to 1.11), or the proportion of primary head contact penalties (<14%).73 SRC rates decreased (IRR 0.74; 95% CI 0.62 to 0.88) in the seasons after policies removed the two-line pass rule (a rule that disallowed direct passing across the defending teams blueline and redline) and inclusion of stricter rule enforcement to prevent obstruction in professional play.74
Policy enforcing red cards for high elbows or intentional elbow to head contact in professional soccer was associated with a non-significant reduction in SRC (IRR 0.71; 95% CI 0.46 to 1.09) and significant reduction in overall head injury (RR 0.81; 95% CI 0.67 to 0.99).75 76 After heading the ball was banned in 2015 by the US Soccer Federation in players under 10 and in games for players aged 10–13, an increase in SRC relative to other injuries (OR 1.29; 95% CI 1.09 to 1.52) was seen in emergency departments in players aged 10–13.77 The heading ban also included an initiative to improve concussion education and implementation of more uniform concussion management.77 Rules minimising intentional contact to the head or neck and the use of bodychecking in adolescent male lacrosse led to lower bodychecking-related concussion rates during practices (IRR 0.29; 95% CI 0.12 to 0.70) and matches (IRR 0.51; 95% CI 0.29 to 0.91).78 In professional baseball, a rule limiting collisions between the base runner and the catcher at home plate was associated with a significant reduction in catcher concussion rates (RR 0.31; 95% CI 0.11 to 0.85).79 80 A rule change in rugby league limiting the frequency of interchange replacements did not reduce SRC rates (IRR 0.59; 95% CI 0.04 to 9.48).81 Policy reducing the maximum height of the legal tackle in rugby union from the line of the shoulders on the ball carrier to the line of the armpits did not reduce SRC rates (IRR 1.31; 95% CI 0.85 to 2.01) (online supplemental file 6).82
Studies comparing head impacts between youth tackle football and flag football where tackling was not permitted are summarised as supplemental content (online supplemental file 6).83–87 Several policy/rule change initiatives have been examined to reduce head impacts and SRC rates in American football. Kerr et al 88 demonstrated that SRC rates did not differ when child and adolescent level players were grouped based on age and weight rather than just age only (RR 0.6; 95% CI 0.3 to 1.4). Restricting the frequency and/or duration of collision practices in adolescents reduced head contact and practice-related concussion rates.89–92 Similar policy changes have not been successful at the collegiate level (online supplemental file 6).93–95 A meta-analysis combining studies indicated a 64% reduction in practice-related concussion rates when policy and non-policy approaches to limiting contact in practices were implemented across adolescent and adult leagues (IRR 0.36; 95% CI 0.16 to 0.80) (figure 5). The meta-analysis examining all strategies to reduce practice-related head impacts in adolescents indicated a 53% reduction, but this was not significant (figure 6, online supplemental file 6). After the kickoff line was moved up and touchback line moved back (aimed to increase kickoffs landing in the end zone and the likelihood of more touchbacks), a significant reduction in SRC rate was seen at the collegiate level (IRR 0.19; 95% CI 0.04 to 0.61).96
Forest plot based on meta-analysis evaluating strategies (policy and non-policy) (left) aimed to reduce practice-related concussion risk in American football and combined by age group (right). IRR, incidence rate ratio; SIGN, Scottish Intercollegiate Guidelines Network.
Forest plot based on meta-analysis evaluating targeting rule changes in adult American football. IRR, incidence rate ratio; SIGN, Scottish Intercollegiate Guidelines Network.
Five studies examined various targeting rules that have been implemented at all levels of American football play.97–101 Aukerman et al 97 reported a higher SRC rate during plays in which a targeting penalty was called versus a non-targeting play at the collegiate level (IRR 36.9; 95% CI 22.4 to 60.7). Hanson et al 98 reported a 32% reduction in weekly concussion reports among professional defensive players after implementing the crown of the helmet rule (penalising players intentionally initiating contact using the top of their helmet). Baker et al 99 demonstrated a 40% lower SRC rate (RR 0.60; 95% CI 0.50 to 0.73) after the targeting rule was broadened.99 After implementing targeting rules, a reduction was found in adolescent SRC rates (p=0.04) and concussions caused by helmet to helmet contact (p=0.03) presenting to emergency departments.100 Westermann et al 101 reported a higher SRC rate in seasons after implementing targeting rules (IRR 1.34; 95% CI 1.08 to 1.66).101 When considering potential unintended consequences, Hanson et al 98 and Westermann et al 101 reported increased lower extremity IRs in professional and collegiate football after targeting rules were implemented. This was contrary to Baker et al 99 who did not report any increased lower extremity IR at the professional level, but did note an increase in games missed from lower extremity injury. A meta-analysis including three of these studies suggests that targeting rules were not associated with reduced SRC rates (IRR 0.77; 95% CI 0.38 to 1.56) (figure 7).97 99 101
Forest plot based on meta-analysis evaluating strategies aimed to reduce head impacts in adolescent American football. IRR, incidence rate ratio; SIGN, Scottish Intercollegiate Guidelines Network.
Training strategies
Examining off-field training strategies, Clark et al 102 demonstrated an 85% reduction in SRC risk (RR=0.15, p<0.001) in American football players following vision training. Training strategies targeting head impact outcomes are summarised in the supplemental material (supplemental file 6).103–105 A 10-week training programme including exercises focused on increasing core strength was associated with lower SRC rates in adolescent American football, soccer, and volleyball players (RR 0.15; 95% CI 0.06 to 0.42).106
Cluster-RCT evaluation of on-field training strategies demonstrated efficacy of a neuromuscular training (NMT) warm-up strategy (eg, balance, whole body resistance, static neck contractions, plyometric training, landing/cutting manoeuvres).107 NMT was associated with 59% lower SRC rates in school-boy (ages 14–18) rugby players (RR 0.41; 90% CI 0.17 to 0.99) when completed ≥3 times/week, compared with standard practice warm-up.107 Attwood et al 108 evaluated an NMT programme compared with a standard practice warm-up in adult men’s community players demonstrating a 60% lower SRC rate (RR 0.4; 90% CI 0.2 to 0.7).108 An NMT evaluation in players aged 12–19 showed significant reductions in training and match-injury IR in those completing the NMT three or more times per week, but did not show any reduction in SRC rates specifically.109 Community players who trained <3 hours/week were more likely to sustain an SRC sooner than those who practised ≥3 hours/week (HR 0.68; 95% CI 0.48 to 0.94).110 Adult players with poorer dynamic balance performance had higher SRC odds than players with optimal balance performance (OR 3.63; 95% CI 1.20 to 10.97).111
Kerr et al 112 (aged 8–15 years) and Shanley et al 113 (adolescent) demonstrated that child and adolescent American football players exposed to a comprehensive coach education programme (ie, proper equipment fitting, tackling technique, strategies for reducing player contact, concussion awareness) had significantly lower practice-related head impacts and game and practice-related concussion rates (RR 0.67; 95% CI 0.19 to 0.91), relative to players in leagues that did not participate.112 113 When the education programme was coupled with instituted guidelines restricting contact in practices, there was an 82% lower practice-related concussion rate in players aged 11–15 (IRR 0.18; 95% CI 0.04 to 0.85) but not players aged 5–10 (IRR 0.82; 95% CI 0.15 to 4.48) compared with those who did not have any education or contact restriction.114 The addition of a player safety coach whose responsibility was to ensure other coaches adhered to proper safety protocols was associated with a reduction in practice-related concussions (IRR 0.12; 95% CI 0.01 to 0.94) but not game-related concussions (IRR 0.14; 95% CI 0.02 to 1.11).115
American football training strategies to reduce head impacts are summarised as supplemental file 6.116–133
Concussion management strategies
After concussion laws were enacted in the USA (eg, mandatory removal from play, requirements to receive clearance to return to play from a licensed health professional, and education of coaches, parents, and athletes), an initial increase in recurrent SRC rate trends was seen across adolescent sports, but then a decrease was seen 2.6 years after the laws went into effect.134 135 Arakkal et al 135 found that in US states where the category of healthcare provider was specified for return to play clearance, recurrent SRC rates were lower than in states where the healthcare provider was not specified (1.59%/standardised month; 95% CI −0.22 to 3.42); however, this was not significant. When examining multiple design elements of the concussion laws (ie, strength of law, number of law revisions, speed of law adoption), Yang et al 136 demonstrated lower recurrent SRC rates when States had more law revisions ( ≥ 2 vs §amp;lt; 2) and adopted laws later. Increasing strength of law (based on 13 discrete evidence-based concussion law provisions) did not reduce recurrent SRC rates.
Across adolescent and adult sports, a symptom-free waiting period after sustaining an SRC did not reduce clinical recovery time or reduce risk of recurrent SRC.137 However, over the past 15 years, improved concussion protocols in collegiate American football players (eg, preseason concussion education, preparticipation assessments, structured plan for concussion diagnosis, postinjury management and return to play) have shown significantly longer symptom durations, symptom-free waiting periods, and return to play, with a significantly lower risk of recurrent SRC.138 Charek et al 139 suggested that players who reported continuing to play for more than 15 min after an SRC took longer to recover than those that continued to play for fewer than 15 min or were removed immediately. This was contrary to the findings by Zynda et al 140 where adolescent players presenting to a paediatric sport medicine clinic experienced similar recovery times when they reported continuing to play following an SRC compared with those that did not. Zynda et al 140 also demonstrated a longer time before presentation at the clinic was associated with a prolonged recovery time. This finding was consistent with a study in youth and adolescent ice hockey (ages 11–17 years), where those that delayed seeing a physician (>7 days) also had a longer clinical recovery time.141 SRC recovery was not significantly different between adolescent ice hockey players that played in a bodychecking league and those who did not.141
When examining SRC rates at the professional level in American football following initiatives to reduce concussions (eg, targeting rule changes, eliminating specific practice drills and in-game blind-side blocks) as well as improve concussion detection and diagnosis (eg, introduction of a centralised clinical electronic health record, Athletic Trainer spotter programme, unaffiliated neurotrauma consultants), a 23% decrease in game-related concussions was observed (IRR 0.76; 95% CI 0.65 to 0.88).142 Teramoto et al 143 did not find an association between SRC rates in professional American football players and the number of days of rest, game location or timing of the bye week. Number of days of rest was also not related to risk of repeat SRC.143 Similarly, Gardner et al 144 did not find any association between the rate of SRC in professional rugby league players and the number of days rest between matches or the match location.
Discussion
This comprehensive systematic review and meta-analysis includes original data studies evaluating primary and secondary SRC prevention strategies to reduce concussion, recurrent concussion and/or head impact rates in various sports. Further, studies evaluating unintended consequences of SRC prevention strategies and studies examining potential modifiable risk factors for SRC were included. Concussion prevention strategies include personal protective equipment, policy/rule changes, training strategies, environmental targets and management strategies targeting recurrent concussion. Potential modifiable risk factors have also been identified for future prevention strategies development, implementation and evaluation (including figure 8, online supplemental file 6).
Forest plot based on meta-analysis evaluating turf to grass fields by sport (left) and age group (right). IRR, incidence rate ratio; SIGN, Scottish Intercollegiate Guidelines Network.
Personal protective equipment
Studies evaluating helmet design and/or materials including flexible panels and helmet fit remain an opportunity for SRC prevention. Cohort studies have indicated that thicker padding over the zygoma/mandible area may reduce SRC rate in American football.22 23 Two studies have identified that secure helmet fit may reduce SRC rates and severity.20 21 Biomechanical studies with appropriate controls remain an opportunity to support more rigorous helmet standards for manufacturing and establishing sport-specific helmet fit criteria.
Studies evaluating headgear report mixed findings with regard to SRC protection. When data were combined across studies in lacrosse, rugby and soccer in an MA, headgear did not reduce SRC rates although the point estimate did suggest an 18% reduction overall (IRR 0.82; 95% CI 0.65 to 1.03). By sport, headgear use was associated with lower SRC rates in soccer but not rugby or lacrosse. There is insufficient evidence to currently recommend the use of headgear in soccer and further evaluation of different headgear design and materials is warranted and potentially considering differences between positions of play. Headguards were not protective against stoppages due to head contact in one study evaluating their use in boxing. Policy mandating headguards in male boxing was removed prior to the 2016 Rio Olympics (online supplemental file 6).41
Ice hockey studies evaluating face shielding suggest that full face shielding does not offer significant protection against SRC over half visors.42–44 Limited evidence suggests that full face shielding may offer protection against SRC severity based on time loss.42 Full facial protection does provide superior protection against orofacial injuries compared with half visors.145 Eyewear use has been recommended in lacrosse and field hockey to reduce head and face injury but does not appear to reduce SRC rates.47–49
Mouthguards are well established in protecting against orofacial injury across sports,146 but their use as an SRC prevention measure has been controversial. A meta-analysis combining ice hockey and rugby studies demonstrated mouthguard use was associated with an overall 26% reduction in SRC rates. While this reduction was found when combining studies, a large majority (83%) of the weight came from one study in ice hockey due to the precision of the estimates. A previous meta-analysis examining mouthguard use suggested a similar point estimate that was not statistically significant (IRR 0.81; 95% CI 0.6 to 1.1).2 When stratified by sport, the effect of mouthguards was significant for adolescent ice hockey but not for adult rugby, potentially suggesting mouthguard use is a marker of safety behaviour or previous concussion in elite rugby but not adolescent ice hockey. Further, the accelerations and impact forces based on the speed of skating (and changes of direction) compared with running in rugby require further consideration when comparing the differential effects of mouthguards in these two sports. Results from this meta-analysis suggest mouthguards should be worn in ice hockey and their use is strongly recommended in other collision sports given the potential concussion protection in addition to orofacial protection. Future studies with rigorous injury surveillance methodologies and consideration of potentially confounding covariables are still recommended to further the understanding of mouthguard and concussion across sports, particularly in children and adolescents. RCTs are likely unethical in some collision sports where their use is already mandated but case–control approaches may be considered.2 147
Jugular vein compression collars, purported to lead to physiologic distension of the superior jugular veins and encourage cerebral venous engorgement similar to the effects produced at higher altitude (online supplemental file 6), has been hypothesised to increase brain resistance to movement or inertia and protect the brain from head impacts and microstructural changes.56 57 Currently, there is not sufficient evidence to recommend the use of such devices to reduce SRC risk or head impact frequency or severity.56 57
Policy, rule or law changes
The meta-analysis assessing the effectiveness of rule changes disallowing bodychecking in children and adolescent ice hockey shows an overall 58% reduction in SRC rates. Surveillance following policy restricting bodychecking demonstrated no unintended injury consequences with fewer years of body checking experience.66 67 148 A recent video-analysis study has also suggested no player performance deficits associated with disallowing bodychecking.149 Head contact rule changes in ages 11–14 and adult professional level have not shown reduced SRC risk.71 72 Referral patterns, referee behaviours, surveillance methods, and increased media attention and concussion awareness may all contribute to reducing the effectiveness of head contact policies.2 Given the evidence suggesting continued high rates of head contacts occurring at the adolescent level even after the introduction of head contact policy,73 greater referee training in sports that disallow head contact may be an avenue for future research examination. Limiting head contacts in soccer, lacrosse and baseball have led to lower concussion or head impact rates.75 76 78–80
Policy limiting the number and duration of contact practices in American football has led to reduced SRC and head impact rates in adolescents.89–92 Limiting the number of contact practices did not have as much success in terms of reducing SRC risk or head impacts at the collegiate level as teams were noted to run longer duration practices and with more intense contact (online supplemental file 6).93–95 Further restrictions on limiting practice duration may help in decreasing head impacts and SRC risk at the collegiate level. Based on the results of the meta-analysis examining targeting rules in American football (eg, prohibiting initiating contact to an opponent above the shoulders, lowering the head or initiating contact with the crown of the helmet, targeting of defenseless players in the head/neck area), these policy changes did not significantly reduce SRC rates (IRR 0.77; 95% CI 0.38 to 1.56). It is unclear whether the implementation of targeting rules led to increased lower extremity IRs.98 99 101 Moving the kickoff line up significantly reduced SRC rates at the collegiate level in American football.96 Similarly, Ruestow et al 150 examined the effect of the free kick rule in professional football and found a non-significant reduction in head injuries (IRR 0.33; 95% CI 0.09 to 1.21). Other concussion initiatives (eg, targeting rule changes, eliminating specific practice drills and in-game blind-side blocks) at the professional level are associated with decreased game-related concussions, however, unintended consequences should always be considered.142
Training strategies
Studies across sports examining vision/cognitive training programmes have reported mixed findings with regard to lowering SRC and head impact risk (online supplemental file 6).102–104 151 152 Potential differences between studies may include training programme components and differences between sports such as rules (eg, tackling vs bodychecking) and positions of play. Exercise warm-up programmes that include several components (eg, balance, resistance, landing and cutting) have been shown to reduce SRC rates in rugby and are recommended.107 108 The effect of NMT programmes in reducing concussion rates specifically has not been assessed in other sports. While there is extensive evidence to support the effectiveness of NMT warm-up programmes in reducing all injury and lower extremity injury in sport, more research is needed for NMT warm-up programmes in females and other team sports specifically targeting exercise components aimed to reduce concussion rates. Comprehensive coach education that included several other components such as strategies to reduce player contact has been shown to reduce SRC and head impacts in child and adolescent American football.112–115 Many studies support limiting contact and equipment during practice drills and improving tackling and blocking techniques to reduce SRC and head impact kinematics.116–129 131–133 Across sports, concussion education programmes without additional strategies have been shown to improve concussion knowledge and promote potential behavioural changes, yet there is a paucity of research evaluating whether these programmes reduce SRC rates.153 154 Future studies in other sports evaluating similar exercise programmes with additions of sport-specific components are warranted.
Other strategies to reduce concussion risk, head impacts or severity
Child and adolescent ice hockey leagues that have fair play programmes help reduce the number and severity of penalties,155–159 but it is unclear whether these programmes also help reduce SRC risk.68 70 Initial evidence suggests players are at lower risk of overall injury when venues use a flexible board/glass system rather than a traditional system, which may extend to a lower SRC risk as well.160 See online supplemental file 6 for further discussion on secondary prevention.139–141 161
Strengths and limitations
This comprehensive systematic review and meta-analysis evaluated prevention strategies and modifiable risk factors for SRC, head impacts and SRC severity. Some papers that were included in the previous systematic review that informed evidence based prevention strategies for the fifth International Conference on Concussion in Sport were not included in this review.1 2 This is due to stricter inclusion criteria such as limiting publication years, a focus on sport (not recreational activities) and only studies of stronger methodological quality.12 13 Studies must have been published in English, introducing potential language bias. Measurement bias (including self-report) was prevalent in the many studies. Any measurement bias with regard to concussion definition was likely non-differential and equal across study groups between the probability of a concussed player being classified as non-injured and a non-injured as concussed. Small samples have limited the ability to examine age, sex/gender effects and parasports. Not all studies controlled for potentially confounding variables or clustering effects in team sports. Our results are limited in that studies assessing head injury broadly or TBI were excluded if they did not specify concussion. Studies that primarily considered all injury as the outcome of interest may have been missed based on our search strategy. Several included papers commented on how increased media attention, awareness of concussion and concurrent concussion education programmes may have influenced concussion reporting rates which may have affected individual study results. There is also the potential for under-reporting of concussion symptoms, which may also affect study results.
See online supplemental file 6 for further discussion regarding head impacts.162–168
Conclusions
Some of the strongest evidence for SRC prevention is through policy and strategies restricting body checking or contact across several child, adolescent and adult sports. Continued research examining prospective rule changes and associated biomechanical investigation is recommended as is research examining helmet fit and types. Mouthguards are associated with a lower overall risk of SRC and should be worn in ice hockey. Neuromuscular warm-up programmes have a protective effect in reducing SRC in rugby, with future research required to consider other sport contexts and greater attention to concussion-targeted training components. Certain modifiable risk factors such as neck strength require further evaluation to elucidate their role in SRC prevention. The continued evaluation of SRC and head impact prevention strategies targeting sport-specific extrinsic (eg, rules) and intrinsic (eg, concussion history) risk factors are required. Appropriate evaluation designs (eg, RCTs, cohort, case–control) using validated injury surveillance methodologies, consideration of potential confounding variables (eg, concussion history) and with common concussion definitions consistent with consensus definitions are needed. Video-analysis and instrumenting players (eg, mouthguards) support concussion surveillance evaluation approaches. Consideration of individual player exposure data (ie, player participation) to measure IRs and clustering effects for team-based sports is also important. Psychological and sociocultural factors continue to be important considerations in the uptake and maintenance of SRC prevention strategies. Future research examining prevention strategies should target understudied populations (eg, women/girls, para athletes).
Key recommendations
1. What sport-related concussion (SRC) prevention strategies reduce concussion and/or head impact risk (eg, equipment, policy/rules, training strategies)?
Mouthguard use should be supported in child and adolescent ice hockey (Grading of Recommendations Assessment, Development and Evaluation, GRADE quality rating: low).
Policy disallowing bodychecking should be supported for all children and most levels of adolescent ice hockey (GRADE quality rating: high).
Strategies limiting contact practice in American football should inform related policy and recommendations for all levels (GRADE quality rating: low).
Neuromuscular training warm-up programmes are recommended, based on research in rugby, while more research is needed for females and other team sports. The focus should be on exercise components targeting concussion prevention (GRADE quality rating: moderate).
Policy mandating optimal concussion management strategies to reduce recurrent concussion rates is recommended (GRADE quality rating: very low).
2. Are there unintended consequences of SRC prevention strategies?
Prior bodychecking experience in ice hockey games was not associated with lower concussion rates when adolescent players played in leagues permitting bodychecking, suggesting no unintended consequences of policy disallowing bodychecking to refuse policy recommendation above (GRADE quality rating: moderate).
Future research should consider evaluation of unintended consequences of concussion prevention strategies across all contexts.
3. What modifiable risk factors are associated with SRC risk?
Lower concussion rates have been demonstrated in certain sports when matches are played on an artificial turf field compared with a natural grass field. Further research should target detailed understandings of playing surface and associated mechanisms of injury prior to concussion prevention strategy recommendations (GRADE quality rating: low).
Further prospective analytical research designs examining neck strength as a potential modifiable risk factor for concussion are needed to inform future development of related concussion prevention strategies (GRADE quality rating: very low).
Future sport-specific research evaluating optimal tackle technique to reduce concussion risk in rugby is necessary before informing related prevention strategy targets (GRADE quality rating: low).
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
Ethics statements
Patient consent for publication
Acknowledgments
The authors would like to acknowledge the work of librarians Drs. Zahra Premji (search strategy), Diane Lorenzetti (PRESS review) and Shauna Rutherford (reference retrieval and organisation in Covidence) for their contributions to the work. The authors also acknowledge Paul Ronksley for assisting with the review process.
References
Footnotes
Twitter @westy160991, @stujohnbailey, @aacademic, @SportswiseUK, @Kat_Schneider7, @CarolynAEmery
Contributors PE and CE conducted the rapid screen. PE, ATK, MPP, SWW, SBailey, LM and CE reviewed study titles and abstracts. Reference lists of selected systematic reviews were hand-searched by PE, SWW, ATK, MPP or CE. The full text of all potentially relevant studies was independently reviewed by one of two lead authors (CE (senior anchor author) or PE (1st author)) and one other author (ATK, MPP, SWW, SBroglio, LM, AMB, SBailey, GAD, BEH, JDS, KAS, MT, RT, NW, RZ and KJS) to determine final study selection. Data were extracted by two authors for each paper (CE or PE) and one additional coauthor (ATK, MPP, SWW, SBroglio, LM, AMB, SBailey, GAD, BEH, JDS, KAS, MT, RT, NW, RZ and KJS). Either consensus was achieved, or a third author (CE or PE) discussed discrepancies. Two authors (CE or PE and other co-author) independently assessed risk of bias. Quality of evidence and grading strength for key recommendations for each research question was assigned (PE and CE) using the Grading of Recommendations Assessment, Development and Evaluation system. J-MG led all meta-analyses. GAD and RZ led paediatric considerations. NW led parasport considerations. AH led the development of the search process. All authors critically reviewed and edited the manuscript before submission. CE is the guarantor. Please see the Summary of the Methodology for the Amsterdam 2022 International Consensus on Concussion in Sport for additional detail.11
Funding Education grant from the Concussion in Sport International Consensus Conference Organising Committee through Publi Creations for partial administrative and operational costs associated with the writing of the systematic reviews.
Competing interests PE: Data consultant to the National Hockey League. Received an honorarium for the administrative aspects of the concussion consensus review. JMG: No conflicts of interest. ATK: Research funding for PhD received from Canadian Institutes of Health Research and University of Calgary Eyes High Doctoral Recruitment Scholarship. Youth Council Member for the Canadian Institutes of Health Research Institute of Human Development, Child, and Youth Health’s. MPP: No conflicts of interest. SWW: Research funding received from World Rugby. SBailey: PhD Research was funded by Scottish Rugby, the national governing body for rugby union in Scotland. LM: No conflicts of interest. AMB: Peer-reviewed research funding from the Social Sciences and Humanities Research Council, Sport Information Resource Center board member, Canadian Athletic Therapists Association committee member. Received an honorarium for the administrative aspects of the concussion consensus reviews. SBroglio: Research funding from the National Institutes of Health; Centers for Disease Control and Prevention; Department of Defense-USA Medical Research Acquisition Activity, National Collegiate Athletic Association; National Athletic Trainers’ Association Foundation; National Football League/Under Armour/GE; Simbex and EA. He has consulted for US Soccer (paid), US Cycling (unpaid), University of Calgary SHRed Concussions external advisory board (unpaid), medicolegal litigation, and received speaker honorarium and travel reimbursements for talks given. He is coauthor of 'Biomechanics of Injury (3rd edition)' and has a patent pending on 'Brain Metabolism Monitoring Through CCO Measurements Using All-Fiber-Integrated Super-Continuum Source' (US Application No. 17/164490). He is on the and is/was on the editorial boards (all unpaid) for Journal of Athletic Training (2015 to present), Concussion (2014 to present), Athletic Training & Sports Health Care (2008 to present), British Journal of Sports Medicine (2008 to 2019). GAD: Member of the Scientific Committee of the 6th International Consensus Conference on Concussion in Sport; an honorary member of the AFL Concussion Scientific Committee and has attended meetings organised by sporting organisations including the NFL, NRL, IIHF and FIFA; however, has not received any payment, research funding, or other monies from these groups other than for travel costs. BEH: No conflicts of interest. JDS: No conflicts of interest. KAS: Employed (part-time) by the Rugby Football Union, the national governing body for rugby union in England. Research funding received from World Rugby, Rugby Football Union, Premiership Rugby, Football Association Premier League, England and Wales Cricket Board, and British Racing Foundation. MT: No conflicts of interest. RT: Employed as a consultant by World Rugby, the body that regulates the sport of Rugby Union globally. The role includes research into prevention of concussion through various interventions. NW: International Paralympic Committee Medical Committee. RZ: Current or past competitively funded research grants from Canadian Institutes of Health Research (CIHR), National Institutes of Health (NIH), Health Canada, Ontario Neurotrauma Foundation (ONF), Ontario Ministry of Health, Physician Services Incorporated (PSI) Foundation, CHEO Foundation, University of Ottawa Brain and Mind Research Institute, Ontario Brain Institute (OBI), and Ontario SPOR Support Unit (OSSU), and the National Football League (NFL) Scientific Advisory Board. I hold Clinical Research Chair in Pediatric Concussion from University of Ottawa, and I am on the concussion advisory board for Parachute Canada (a non-profit injury prevention charity). I am the cofounder, Scientific Director, and a minority shareholder in 360 Concussion Care (an interdisciplinary concussion clinic). AH: No conflicts of interest. KAS: Kathryn Schneider has received grant funding from the Canadian Institutes of Health Research, National Football League Scientific Advisory Board, International Olympic Committee Medical and Scientific Research Fund, World Rugby, Mitacs Accelerate, University of Calgary) with funds paid to her institution and not to her personally. She is an Associate Editor of BJSM (unpaid) and has received travel and accommodation support for meetings where she has presented. She is coordinating the writing of the systematic reviews that will inform the 6th International Consensus on Concussion in Sport, for which she has received an educational grant to assist with the administrative costs associated with the writing of the reviews. She is a member of the AFL Concussion Scientific Committee (unpaid position) and Brain Canada (unpaid positions). CE: Carolyn Emery has received external peer-reviewed research funding from Canadian Institutes of Health Research, Canada Foundation for Innovation, International Olympic Committee Medical and Scientific Committee, National Football League Play Smart Play Safe Program, and World Rugby. She is an Associate Editor of BJSM (unpaid) and has received travel and accommodation support for meetings where she has presented. She is an external advisory board member for HitIQ.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.