Table 3 Study characteristics of articles regarding concussion and helmet use
StudySportStudy designDuration (seasons)Study populationExposure measuresOutcome measuresResults
Mueller (2008)47Ski and SnowboardCC53 ski resorts, Western United States, (cases = 4224, controls = 17674)Helmet useHead injuries: treatment by ski patrollers (n = 2537)Injured skier controls (below neck): helmet use significantly reduced risk of a head injury OR = 0.85 (0.76 to 0.93)
Sulheim (2006)46Ski and SnowboardCC18 major Norwegian alpine resorts (cases = 3277, controls = 2992)Helmet use, injury typeHead injuries: treatment by ski patrollers (n = 578)Uninjured controls: helmet use significantly reduced risk of a head injury OR =  0.40 (0.30 to 0.55)Injured skier controls: Helmet use significantly reduced risk of a head injury OR =  0.45 (0.34 to 0.59)
Hagel (2005)45Ski and SnowboardCC119 ski areas in Quebec, Canada (cases  = 1082, controls  = 3295)Helmet useHead injuries: treatment by ski patrollers (n = 693)Injured skier controls: Helmet use significantly reduced risk of a head injury OR = 0.71 (0.55 to 0.92)
Fukuda (2007)48SnowboardCS4 (Dec ’99 to March ’03)Snowboarding head injury patients at the SMH, Minami-uonuma, Japan (n = 1190)Helmet, knit cap, no-capSerious head injuries from snowboarding: med attn (n = 549)No significant association between helmet or knit cap wearing and the occurrence of serious head injury (p = 0.056)
McIntosh (2001)49RugbyRCT116 U-15 rugby union teams in Sydney (n = 294)Headgear useConcussion: time-loss (n = 9)No significant reduction in risk between those wearing headgear and those not wearing headgearIRR  = 1.06 (0.22 to 5.10)*
McIntosh (2008)50RugbyRCT2Males from U-13, U-15, U-18 and U-20 (n = 4095)Standard headgear use, modified headgear useConcussion: game injury or time-loss (n = 199)No significant difference in injury rates was observed for concussion across the three arms of the study (p>0.05)
Kemp (2008)51RugbyPC313 English Premiership rugby union clubs (n = 757)Headgear useConcussion: CISG definition (one or more symptom) (n = 96)Incidence of concussion with headgear: 2.0/1000 player hrs (1.0 to 4.2) Incidence of concussion without headgear: 4.6/1000 player hrs (3.7 to 5.7) IRR  = 0.43 (0.21 to 0.92)*
Marshall (2005)40RugbyPC15 rugby clubs, 4 secondary schools (n = 304)Headgear useConcussion: timeloss/med Attn (n = 22)No significant reduction of risk with use of headgear.Unadjusted RR = 0.93 (0.34 to 2.58)Adjusted RR = 1.13 (0.40 to 3.16)
Kahanov (2005)53RugbyXSNot specified8 Men’s university teams in the United States (n = 131)Headgear useConcussion: self-report (n = 136)Headgear concussions: (n = 32)No headgear concussions: (n = 104)No association measures of effect
Delaney (2008)66SoccerXS1Travel teams, ages 12–17, apart of the OSC, Canada (n = 278)Headgear useConcussion: CISG definition (one or more symptom) (n = 133)Not wearing headgear significantly increased risk of concussion.
RR  = 2.65 (p = 0.001)
Collins (2006)55FootballPC3High school athletes from Western PennsylvaniaHelmet (new vs old)Concussion: symptoms assessed by team trainer (n = 136)New helmets significantly decreased in risk of concussion.
RR  = 0.69 (0.50 to 0.96)
Torg (1999)56FootballXS3Purchasers of the ProCap polyurethane helmet cover (n = 155)ProCap helmet coverConcussion pre/post ProCap helmet: self reportThe natural history of repeat concussion occurrences may not be affected by a polyurethane helmet cover
Alles (1979)54FootballPC353 high school football teams and 148 college/university teams in the US (n = 16000)Helmet type (13 different types or brands)Concussion: time-loss/med attn (n = 905)No helmet brand or type was associated with an increased risk of concussion
Brandenburg (2002)68RodeoRC1Purchasers of the Bull Tough Helmet (n = 81)Helmet useHead injury: self-report (n = 39)Helmet use significantly reduced risk of a head injuryHelmet: IR = 15.4/1000 ridesNo-helmet: IR = 8/1000 ridesIRR = 0.43 (0.26 to 1.02)*
Kraus (1970)67HockeyCC2Intramural ice hockey players at the University of Minnesota (64 teams in 1968, 73 teams in 1969)Mandatory helmet implementationHead injury: med attnHelmet use significantly reduced the risk of a head injury (p = 0.03)No mandatory helmet: IR = 8.3/100 gamesMandatory helmet: IR = 3.8/100 games
Wasserman (1990)65BicyclingRC5Readers of four bicycling magazines (n = 191)Helmet useConcussionHelmet use did not significantly decrease the risk of concussion.OR = 0.56 (0.29 to 1.07)*
Dorsch (1987)64BicyclingRC5Bicycle club members in Adelaide, Australia (n = 894)Helmet use, helmet typeHead injuryHelmets protected against head injury. Risk of death from head injury was 10 times greater in unhelmeted rider compared to a helmeted rider
Thompson (1989)58BicyclingCC1 (Dec ’86 to Nov ’87)Cyclists with head injuries resulting in an emergency room visit to one of 5 hospitals in Seattle (WA), USA (n = 668)Helmet useHead injury, brain injury: diagnosis (n = 235, n = 99)Injured cyclists controls: helmet use significantly reduced risk of a head injury, OR = 0.26 (0.14 to 0.49) and brain injury, OR = 0.19 (0.06 to 0.57).Population-based controls: helmet use significantly reduced risk of a head injury, OR = 0.15 (0.07 to 0.29) and brain injury, OR = 0.12 (0.04 to 0.40)
Finvers (1996)62BicyclingCS2 (April ‘91 to Sept ‘93)Injured child bicylists (aged 3–16) at ACH (n = 699)Helmet useSevere head injuries: (n = 76)Injured children controls: helmet use significantly reduced risk of a head injury OR = 0.32 (0.08 to 0.89)*
Heng (2006)63BicyclingCS1 (Sept ‘04 to May ’05)Patients presenting to an ED in Singapore for bicycling injuries (n = 160)Helmet useHead injury: (n = 58)Injured cyclists controls: helmet use significantly reduced risk of a head injury OR = 0.09 (0.002 to 0.65)*
Lewis (2001)70FootballLabN/ABall (soccer)-to-head3 Male volunteers ages 16–30: (∼35 miles/hr)Pendulum device (bowling ball):1 volunteer and 2 instrumented cadaver headsAccelerometers (IO and IC) within mouthpiece and helmet.Helmet useMean peak GsSignificant reduction in mean peak Gs in participants wearing a helmet, measured with IO accelerometer (p = 0.01)Significant reduction in mean peak Gs in cadavers wearing a helmet, measured with IO and IC accelerometer (p<0.001)
Viano (2006)69FootballLabN/AHelmet-to-helmet ImpactsHybrid III 50th percentileDummies: 60 laboratory testsEndevco 7264a-2000 accelerometer used.Helmet: old (VSR-4) vs new (5 types)Severity index (concussion risk formula:91Rotational acceleration, Translational accelerationAverage reduction in concussion risk (severity index) with newer helmets was 10.8% (range: 6.9 to 16.7%).Reduction in newer helmets was 9.5% (range: 6.5 to 13.9%) for translational acceleration and 18.9% (range: 10.6 to 23.4%) for rotational acceleration.
McIntosh (2004)71Rugby (union and league), Australian rules football.LabN/AHeadgear-to-surface impactSize c rigid magnesium alloy headformDrop heights were 0.3, 0.4, 0.5 and 0.6 m onto a flat anvil.Data captured at 10000 Hz.Headgear: Albion Headpro, Body Armour honeycombImpact energy attenuation testPolyethylene foam: thickness increase from 10 mm to 16 mm improves energy attenuation, density did not alter energy attenuation significantlyModified headgear: significant energy attenuation was achieved by increasing foam density and thickness
McIntosh (2000)72SoccerLabN/AHeadform-to-impact surface:Hybrid III headform.Anatomical location: temporal-parietal area.Drop height: 0.2 m to 0.6 mEndevco 7264 (or 7265) accelerometer used.Data captured at 10 kHz at −10°, 20° and 50°C.Headgear: 8 typesHead injury criteria and Maximum headform accelerationMagnitude of headform accelerations increased as drop height increasedThe foam material was completely compressed at an impact energy of 20J. Unlikely for headgear to reduce concussion or severe head injuries
Broglio (2003)74SoccerLabN/ABall-to-platform:50 trials of each conditionSpeed of ball: 35 mphData captured using a Kistler® force platform mounted vertically.Headgear: headers, headblast, protectorPeak force (N), time to peak force (s) and impulse (N*s)Significant reduction in peak force of impact (N) with all 3 headbands compared to control (p<0.001)No difference in peak force was found between headbands (p>0.05)
Hrysomallis (2004)73SoccerLabN/AHeadform-to-impact surface:7 soccer headgear trials4 different impact locationsImpact energy  = 56JAccelerometers within the headformHeadgear: 7 typesHead injury criteria: 1000 was injury thresholdOnly one headgear (thickness = 15 mm) scored a HIC thickness below 1000A significant correlation was found between headgear thickness and HIC scores
Withnall (2005)75SoccerLabN/ABall-to-headHuman volunteerUp to 8.4 m/s (32 tests)Surrogate head test form:10–30 m/s (16 tests)Head-to-headBiofidelic dummy headforms: 2–5 m/s (40 tests)Data collected at 10 kHz, with requirements of SAE J211-1Headgear: (Head Blast, Full 40 headgear, Kangaroo soccer headgear)Injury assessment functions related with MTBI: linear acceleration, HIPmaxBall-to-head: no significant differences seen in between bare head and various headgears (p>0.05)Head-to-head: headgear reduced peak linear acceleration by -32.8% (95%CI: -40.9% to -24.7%) and reduced HIPmax by -32.6% (95%CI: -40.7% to -24.5%)
Caswell (2002)76LacrosseLabN/AHeadform-to-impact surface: front drop 10 times, ear drop site 10 times, repeated at three heights; 1.44 m, 1.32 m, 1.12 mData collected at 1650 Hz, according to NOCSAE protocol.Helmets: (2 contemporary helmets, 2 traditional helmets)GSIThere were significant differences in GSI scores between helmet types (p<0.05). Contemporary helmet designs performed better in RD tests than traditional helmets.All helmets exhibited sharp increases in GSI values with repetitive impacts.
Bishop (1984)77BicyclingLabN/AHeadform-to-impact surface: front, read, left side and right front boss locations evaluated.Tested from 1.0 m or 1.75 mData collected at 6060.6 Hz, using a Endevco accelerometer.Helmets: (7 brands)Peak acceleration and GSISignificance differences were found in GSI scores between helmets and within helmets at various locations. Helmets with polystyrene liners were much superior to soft foam liners
McIntosh (2003)78Cricket, Baseball, HockeyLabN/ABall-to-headform (Hybrid III)Front, lateral and rear locations evaluatedBaseball, cricket ball and hockey puck fired using air cannon at 19, 27, 36 and 45 m/sHeadform-to impact surface: drop tests from 0.2 m to 1.2 m, three different impact anvils usedData collected at 16000 Hz, using a Endevco accelerometerHelemets: (5 cricket, 2 baseball, 2 hockey)Headform accelerationAt the lower impact speeds, all helmets produced a good reduction in headform accelerationCricket helmet performance is satisfactory for low impact speeds, but not for high impact speeds. Baseball and hockey offer better relative and absolute performance at the 27 m/s ball and puck impacts
  • CC, case-control; CISG, Concussion in Sport Group; CR, case report; CS, case series; GSI, Gadd severity index; HIC, head injury criteria; HIPmax, Head Impact Power index (maximum value of power for concussion using a regression equation); IR, incidence rate; IO, intraoral; IRR, incidence rate ratio; Lab, laboratory; OSC, Oakville Soccer Club; MTBI, mild traumatic brain injury; NOCSAE, National Operating Committee on Standards for Athletic Equipment; PC, prospective cohort; RC, retrospective cohort; RCT, randomised controlled trial; U, under; XS, cross-sectional.