Background The paper presents a novel laboratory method for assessing boxing headguard impact performance. The method is applied to examine the effects of headguards on head impact dynamics and injury risk.
Methods A linear impactor was developed, and a range of impacts was delivered to an instrumented Hybrid III head and neck system both with and without an AIBA (Association Internationale de Boxe Amateur)-approved headguard. Impacts at selected speeds between 4.1 and 8.3 m/s were undertaken. The impactor mass was approximately 4 kg and an interface comprising a semirigid ‘fist’ with a glove was used.
Results The peak contact forces were in the range 1.9–5.9 kN. Differences in head impact responses between the Top Ten AIBA-approved headguard and bare headform in the lateral and forehead tests were large and/or significant. In the 8.3 m/s fist-glove impacts, the mean peak resultant headform accelerations for bare headform tests was approximately 130 g compared with approximately 85 g in the forehead impacts. In the 6.85 m/s bare headform impacts, mean peak resultant angular head accelerations were in the range of 5200–5600 rad/s2 and almost halved by the headguard. Linear and angular accelerations in 45° forehead and 60° jaw impacts were reduced by the headguard.
Conclusions The data support the opinion that current AIBA headguards can play an important role in reducing the risk of concussion and superficial injury in boxing competition and training.
Statistics from Altmetric.com
Boxing is a combat sport that is associated with head impact and head injury risks. In 2013, the International Boxing Association (AIBA, Association Internationale de Boxe Amateur), which is responsible for setting the competition rules for boxing at the Olympic games, banned the use of headguards in selected competitions.1 Headguards are soft padded helmets with no hard shell. AIBA do not specify impact performance tests for boxing headguards nor do they mandate any standard, rather they specify headguard dimensions, for example, mass <450 g.2 This paper presents a novel method for assessing boxing headguard impact performance and examines the effects of headguards on head impact dynamics.
Head impact dynamics, that is, impact force vector and head linear and angular accelerations, are understood to be mechanically related to head injuries that can occur in combat sports, for example, superficial injury, orofacial fractures and brain injury.3–11 According to Zhang et al,10 the tolerance levels for mild traumatic brain injury (mTBI) are 6 krad/s2 and 240 for angular acceleration and the Head Injury Criterion (HIC15), respectively. Rowson et al 11 noted a 75% concussion likelihood for a resultant angular acceleration of 6.9 krad/s2, which is similar to that reported by McIntosh et al. 12 McIntosh et al 12 reported 50% and 75% concussion likelihood for resultant linear head acceleration as 65 and 89 g, respectively.
Although research has been conducted in which boxers have punched headforms, such tests do not offer the level of experimental control and repeatability required to assess headguards.4 Some boxing and combat sports headguard tests have been conducted with pendulum impactors.13 , 14 A literature review (see online supplementary appendix A) identified that mean impact glove speeds in boxing ranges from 3.0 to 11.9 m/s, and peak impact force in gloved punches ranges from 1.4 to 4.8 kN and varied by punch type. The literature also demonstrates that punches delivered in competition or in combination during laboratory experiments have approximately half the impact force of single maximal-effort punches.7 , 15 , 16
Boxing headguards have the potential to reduce the impact force by attenuating the impact energy of the punch and distributing the impact force, but must perform over multiple head impact exposures in training and competition.3 , 17 Although helmet drop tests are a reliable and repeatable method for testing helmets, in a relatively novel area such as boxing headguards, it may be challenging to interpret the test results in the framework of boxing impacts and related injury risks.17 , 18 A second limitation is that the head's angular kinematics cannot be measured in standard drop tests.3 Therefore, it was decided to design and build a novel linear impactor (punch machine) that could be used to deliver punches to the head of an Anthropometric Test Device. The punch machine was used to: compare the performance of two AIBA-approved headguards; compare headguard performance against bare headform impacts; and, using the punch machine with a glove interface, compare head impact dynamics both with and without an AIBA-approved headguard.
A spring driven linear impactor was developed and commissioned through a series of repeatability tests. The impactor is guided by linear bearings and winched back against the resistance of the springs. The displacement of the springs determines the impact speed. Preliminary tests showed that the punch machine delivered repeatable and reliable impacts. A detailed description of the punch machine and system tests is presented in online supplementary appendix B. Two impact interfaces were used: a ‘fist-glove’ and ‘disc-pad’ (figure 1). The total impactor mass was 3.880 kg for the disc interface and 3.885 kg for the fist-glove impacts, including the glove
Hybrid III head and neck
A calibrated Hybrid III head and neck was used in all tests (figure 2). The head and neck were mounted on a massive stand that permitted vertical and rotational orientation of the head and neck with respect to the impactor. The impactor height and angle were adjustable.
Instrumentation, data acquisition and signal conditioning
The head was instrumented with a triaxial linear accelerometer (aHdx,y,z), three angular velocity sensors (ωHdx,y,z) and a six-axis upper neck load cell. The head angular accelerations (αHdx,y,z) were derived by differentiating the filtered angular velocity time histories. For a posteriorly directed frontal impact, the main angular motion is extension, that is, +y angular displacement, velocity and acceleration.19 For a right directed left lateral impact, the main angular motion is right lateral flexion, that is, +x angular displacement, velocity and acceleration. An impact to the left jaw will result in initial axial rotation to the right, that is, +z angular displacement, velocity and acceleration.
The impact force was measured using a Kistler 9331B uniaxial force link mounted between the shaft and impact interface. This force is referred to as the ‘measured force’ (Fm). An estimate of the contact force (Fc) was derived from Fm, where Fc is what the boxer would ‘feel’ when punched (see online supplementary appendix B)
All data were acquired at 20 kHz with a TDAS (Seal Beach, California, USA) data acquisition system. The following signals were filtered with a SAE CFC 1000 filter: aHdx,y,z; Fm; aI; and FNx,y,z. MNx,y,z were filtered with a SAE CFC 600 filter. Angular velocity and acceleration (ωHdx,y,z and αHdx,y,z) were filtered with a CFC 180 filter. Resultant linear and angular accelerations, respectively, RaHd and RαHd, were calculated. The 15 ms limited HIC15 was calculated. Neck loads were measured but not reported. The signal conditioning processes conformed to SAE J211.20 A timing gate was positioned to measure the velocity of the impactor just prior to contact.
Test matrix and headguards
The minimum number of planned tests is presented in table 1. Large-sized Top Ten and Adidas AIBA-compliant headguards were used in all tests. The headguard thickness was in the range of 20–26 mm, density approximately 80 kg/m3, and mass approximately 0.3 kg. All tests were conducted at the Roads and Maritime Services Crashlab in Sydney. Each headguard was tested once only at an impact location. The impact orientations were: centre-front forehead, lateral, 45° forehead and 60° jaw impacts (see online supplementary appendix C).
Descriptive statistics were calculated for the main independent variables: peak linear and angular headform acceleration, peak measured and contact force, and HIC15. t Tests were performed to assess differences between the headguards and between the headguard and bare headform tests.
In total, 64 tests were performed of which three were discarded completely because of test system failures. The coefficient of variation for impact velocity—intended versus obtained—was in the range 1–4%. All results with the disc-pad interface are presented in the online supplementary appendix C.
In total 37 tests were conducted using the fist-glove interface both with and without the Top Ten headguard. Exemplar time-histories for the head impact responses are presented in online supplementary appendix D. Differences in head impact responses between the Top Ten AIBA-approved headguard and bare headform in the fist-glove lateral and centre-front forehead tests were large and/or significant (p<0.05; table 2).
At 4.11 m/s, the headguard tests resulted in slightly lower head impact response values than in the bare headform tests (see online supplementary appendix E). For example, peak RaHd (g) was 24 and 22 g, respectively, for centre-front forehead and lateral headguard impacts compared with 35 and 29 g for bare headform impacts. Peak αHdy for centre-front impacts was 1484 rad/s2 for headguard tests and 1694 rad/s2 for bare headform tests. Peak αHdx for lateral impacts was 1215 rad/s2 for headguard tests and 1750 rad/s2 for bare headform tests.
Lateral jaw and 45° forehead impacts at 8.34 m/s were conducted. The tests were multiplanar and emphasised the head z-axis angular kinematics (see online supplementary appendix D). Significance tests were not performed because of the small sample size (table 3). Angular acceleration in the z-axis was reduced with headguards in both test configurations. The bare headform jaw impacts resulted in a mean peak RαHd of 8605 rad/s2 and mean peak RαHdz of 8333 rad/s2. With the headguard, jaw impacts were reduced to 4335 and 3941 rad/s2, respectively. The bare headform 45° forehead impacts resulted in a mean peak RαHd of 8365 rad/s2 and mean peak RαHdz of 5619 rad/s2. With the headguard, these were reduced to 4335 and 2620 rad/s2, respectively.
The linear impactor built for these tests delivered repeatable impacts. Average peak Fcc by test condition for all the glove tests was in the range of 1.9–5.9 kN, which overlaps with the target range of 1.4–4.8 kN in similar punch speed ranges. A linear impactor could become the basis for a technical specification for boxing headguards and/or gloves as an alternative to a helmet drop test or pendulum impactor.3 , 13 , 14 , 21 , 22 Although the Hybrid III's biofidelity in all possible impact situations is unknown, it has been used to study many impact situations including helmets, boxing and concussive impacts.13 , 19 , 23–25
Injury risk reduction
In general, the results showed that peak impact force, and linear and angular head accelerations were substantially reduced by headguards compared to the bare headform condition; often at least halved. On the assumption that the system biofidelity is meaningful from the perspective of the boxer, the results of the glove tests were interpreted with respect to the following concussion-oriented injury assessment reference values: peak resultant linear acceleration <75 g, HIC15<240, and peak resultant angular acceleration <6000 rad/s2.10–12
In the 8.34 m/s glove impacts, the mean peak resultant headform accelerations for bare headform tests exceeded 75 g for lateral (133 g) and centre-front impacts (131 g). With the headguard, mean peak resultant headform accelerations were 86 g in the lateral impacts and 88 g in the centre-front impacts. Mean HIC15 exceeded 240 for the 8.34 m/s impacts and was less than 240 with the headguard. This suggests that the struck boxer without a headguard would be concussed, and a proportion of those wearing headguards might be concussed. In the 6.85 m/s glove impacts, the mean peak resultant headform accelerations for bare headform tests exceeded 75 g for lateral impacts (86 g) and centre-front impacts (88 g). By contrast, with the headguard, mean peak resultant headform accelerations were less than 75 g. In the 6.85 m/s impacts, mean HIC15 was less than 240 for both headguard and bare headform conditions. This suggests that a proportion of those not wearing headguards would be concussed, and boxers wearing headguards would not be concussed in equivalent punches.
In the lateral and centre-front 8.34 m/s bare headform impacts, peak resultant angular head accelerations exceeded 6000 rad/s2. With headguards, peak resultant angular accelerations were below 4500 rad/s2. In the 6.85 m/s bare headform impacts, peak resultant angular head accelerations were slightly below 6000 rad/s2 (mean 5200–5600 rad/s2), and was almost halved by the headguard. Therefore, the angular acceleration results indicate a much lower likelihood that boxers wearing headguards would be concussed compared to those not wearing headguards.
In the lateral and centre-front 4.99 m/s fist-glove impacts, the IARVs were not exceeded for the headguard or bare headform conditions. This suggests that concussion is unlikely in these impacts.
Average peak Fc for bare headform impacts was in the range 3.7–5.9 kN, depending on location and velocity, compared to 1.9–4.2 kN for headguard tests. Fc is applied over a large surface area because of the glove and, therefore, Fc is unlikely to correspond directly to experimental data on facial fracture forces.26–29 The force data (Fc) suggest that the likelihood and severity of laceration would be reduced by the headguard plus glove, as reflected in practice.
Differences indicative of a protective effect of headguards in the 8.34 m/s jaw and 45° forehead impacts were also observed. In the bare headform, jaw and forehead impacts, 75 g was exceeded but not with the headguard. By contrast, with the headguard, mean peak resultant headform accelerations were close to 83 g, or <75 g. In the bare headform, jaw and forehead impacts a HIC15 of 240 was exceeded but not with a headguard; 6000 rad/s2 was clearly exceeded for both jaw and 45° forehead bare headform impacts. With headguards, peak resultant headform angular acceleration was 7173 rad/s2 in the jaw impact and 4335 rad/s2 in the 45° forehead impact. This strongly suggests that the boxer without a headguard, who was punched on the forehead or jaw in equivalent punches would be concussed, and a proportion of those wearing headguards might be concussed. It is important to note that the z-axis (yaw) angular acceleration was reduced by approximately 50% with a headguard in the jaw impacts. The test results do not show that headguards will increase the risk of head and brain injury.
In totality, the data support the opinion that current AIBA headguards can play an important role in reducing the risk of concussion and superficial injury in boxing competition and training. The results indicate that for slower punches, that is, <5 m/s with the punch machine, the benefits offered by a headguard over and above a glove are small. In the range of punch speeds between 5 and 9 m/s, an AIBA-approved headguard, in combination with a glove, will offer a large level of protection to the boxer's head. The tests in the range of 5–9 m/s correspond well with observed punch speeds and energies.
What are the new findings?
Laboratory impact tests show that a boxing headguard in combination with a glove offers a level of protection to the head and brain in a wide range of impacts.
The optimal benefits of current AIBA (Association Internationale de Boxe Amateur)-compliant headguards are realised in midrange speed punches (5–8.5 m/s) where the impact test results suggest that the headguard will reduce the likelihood of concussion and superficial head wounds.
In low-speed punches (<4 m/s) the addition of a headguard may have only a limited benefit, and in high-speed punches (>9 m/s) the headguard effects in terms of reducing the likelihood of concussion may be limited.
The headguard reduced the magnitude of angular head accelerations, including in impacts to the lateral jaw.
The authors would like to thank the IOC for funding these projects, in particular Dr Torbjorn Soligard, who has been our point of contact; Cherine Fahmy and Professor Lars Engebretsen; RMS Crashlab staff, in particular—Drew Sherry, Colin Jackson, David Corless and Bruce Parker; and David Felice from Zoe's martial arts and boxing supplies in Sydney. Their assistance is gratefully acknowledged.
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.
- Data supplement 1 - Online supplement
Competing interests None declared.
Provenance and peer review Not commissioned; internally peer reviewed.
Contributors ASM and DAP were responsible for data collection, analysis and authorship. All assistance is acknowledged.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.