Title: King-Devick Ocular testing and impact monitoring provide a relatively simple process for objectively identifying and managing concussion in sports participants
 
Running title: Randwick rugby union concussion study.
 
Authors: Cohen, Adrian MB.,BS., King, D., PhD
 
1.            NeckSafe
PO Box 509
Bondi 2026
Australia
 
 
Correspondence to:
Dr. Adrian Cohen
NeckSafe
PO Box 509
Bondi 2026
Australia
 
Email: ado@necksafe.com
 
Keywords: injury, linear, rotational, impact, rugby union, wireless head impact sensor, cognitive, baseline, King-Devick,
 
Submitted to:            TBA
Abstract:                      232 words
Manuscript:                 8111 words
References:                71
Tables:                        9
 

 
ABSTRACT
Background: Concussion in sport is becoming an increasingly controversial subject, and one where despite considerable advances in the last few years, a number of questions remained to be answered and a large degree of subjectivity clouds decision-making.
 
Purpose: To evaluate and quantify head impacts via a wireless head impact sensor acceleration for senior amateur rugby union players during match participation and analyse concussive impacts using the King-Devick Test of ocular pursuits.
 
Study Design: Descriptive observational cohort study
 
Methods: Data on impact magnitude and frequency were collected with a wireless head impact sensor worn by  senior amateur rugby players participating in the 2015 domestic season of matches. King-Devick Tests were performed pre-season and after significant impact scores (PLA > 100g, PRA > 10,000 rads/sec/sec). Participants were asked to undergo a one hour baseline assessment of balance, cognition and have transcranial magnetic stimulation testing performed. Following concussive episodes or significant head impacts (as measured by impact sensors), players were re-tested to elicit any differences and to assist in providing objective criteria for the return to play process.
 
Results: 20 concussions were recorded in all, of which only 3 were witnessed (2 players unconscious on the field, one with balance disturbance resulting in removal from play).
 
Conclusion: This study defined concussions as either witnessed or suspected (on the basis of failed performance against baseline pre-season data from the King-Devick Test in players tested as a result of high impact g force or rotational forces recorded). Based on the results, a large number of concussions would have been missed if impact monitoring had not been utilised, or King-Devick testing routinely implemented after each game and training.
 
A combination of regular King-Devick Ocular testing and impact monitoring provides a relatively simple process for objectively identifying and managing concussion in sports participants.
 
 

 
INTRODUCTION
Concussion in sport is becoming an increasingly controversial subject, and one where despite considerable advances in the last few years, a number of questions remained to be answered, and a large degree of subjectivity clouds decision-making.
 
A concussion typically occurs as a result of a direct impact to the head or from an indirect impact applied to the body that is transmitted to the head[i]. This results when the torso is either decelerated or accelerated rapidly[ii]. Consequently the head sustains a combination of linear and angular acceleration. Direct impacts with the head (linear acceleration-deceleration) and inertial loading of the head (rotational acceleration-deceleration) have been postulated as the two major mechanisms of head related injuries such as concussion[iii]. Linear acceleration produces focal injuries while rotational acceleration produces both focal and diffuse injuries.
 
There is limited published research on the forces associated with concussion. The majority of the data has come from National Football League (NFL) participation[iv] by on-field monitoring and using impact reconstruction of concussive impacts. In laboratory reconstructions it was reported that concussive events occurred with face mask acceleration impacts of 78G ±18G to 94G ±27G but were greater on other areas of the head with an average of 107G to 117G[v]. Using a helmet telemetry system, real-time head acceleration measurements were recorded for high school football participants with peak linear accelerations varying from 55.7 to 136.7g for concussive events. [vi] This was extended further with the helmet telemetry system monitoring impacts in collegiate football participants with peak linear accelerations ranging from 60G to 120G for concussive events with the majority of peak accelerations being above 95G.[vii] [viii]
 
On author has previously studied head impacts in Rugby Union and League footballers using instrumented mouthguards[ix] and skin-mounted sensors[x].
 
In this study, we attempted to provide objective information as to the impact size, rotational components, frequency and effects from playing rugby, related especially to concussion events, using an accelerometer attached to the skin of the neck of player and compare data from this with a number of objective physical (ocular and balance) as well as cognitive test parameters.
 
Our definition of concussion was based on events that were “witnessed”, either players being rendered unconscious or staggering after impact, or “suspected” based on failed performance against baseline pre-season data from the King-Devick Test in players tested as a result of high impact g force or rotational forces recorded on their impact data.
 
Following a presentation to the Randwick District Rugby Union Football Club (RDRUFC) Board in February 2015, consent was given to undertake a pioneering study into concussion using the club’s first grade team.
The study aimed to collect pre and post injury data, and assess the severity of the impact and resultant diagnosis of concussion using objective testing measures of force (via impact sensors/accelerometers) as well as physical parameters (ocular and balance) and cognitive testing.
 
Participation in this study was voluntary and the research paid for by the charity NeckSafe[xi].
 
Ethics approval was sought and received from Auckland University of Technology (AUT)[xii] and Deakin University[xiii] with governing body approval through World Rugby (WR) and the Australian Rugby Union Concussion Advisory Group (CAG). In addition, Rugby insurers Gow Gates P/L were contacted to confirm that there would be no influence on player insurance, the Rugby Union Players Association (RUPA) were briefed and each opposing club was contacted to gain consent for their players to participate in matches where the sensors were being worn. RDRUFC provided legal indemnity to the ARU, who in turn provided it to World Rugby.
 
Players were asked to complete a pre-competition concussion history questionnaire, a sports concussion assessment baseline evaluation, undertake a physical examination and complete a number of computerised cognitive tests: Standardised Concussion Assessment Test version 3[xiv] (SCAT3), King-Devick Test[xv] (KD) and Axon Cogstate Test[xvi]. These questionnaires and tests are used to evaluate players in the event of a suspected concussion occurring, and are referred to as Baseline Tests.
 
During the season additional tests were added, including MyBrainSolutions[xvii] cognitive test and associated iPad training exercises.
 
Balance Testing was assessed using an IsoBALANCE Forceplate[xviii] (including CHOICE and MAZE Cognitive test software).
 
A research team from Cornell University under Dr. David Putrino also contributed pioneering balance software using an Oculus Rift 3D Virtual reality system[xix].
 
In addition, Transcranial Magnetic Stimulation (TMS) testing was introduced as another objective measure of change in concussion[xx].
 
King-Devick OCULAR testing
The King–Devick Test was developed in 1976 by Alan King and Steven Devick, as an indicator of saccadic performance as it relates to reading ability. For more than 30 years, the King–Devick Test has been a proven indicator of oculomotor inefficiencies regarding eye movements during reading.
 

The King-Devick Test is used as an objective remove-from-play sideline concussion screening test that can be administered in minutes. The test is an accurate and reliable method for identifying players with head trauma and  can help to objectively determine whether players should be removed from games and when they have returned to their pre-injury test status they can deemed safe to return to sport or practice (provided they have no symptoms and are not taking medication).
 
The K-D Test is a test of ocular saccades(quick, simultaneous movements of both eyes) that requires a player to read single digit numbers displayed on a tablet or on an iPad. The test cards become progressively more difficult to read due to variability of spacing between the numbers. Both errors in reading and speed of reading are included in deriving a score. Saccades are quick, simultaneous movements of both eyes.

 After suspected head trauma, the player is given the test and if the time needed to complete the test is any longer than the player’s baseline test time, the player should be removed from play and should be evaluated by a licensed professional.
 
The K-D Test has been extensively validated for use as a concussion screening tool. In 2011, researchers from the University of Pennsylvania Perelman School of Medicine published a study in Neurology of an investigation of the test as a potential rapid sideline screening test for concussion in a cohort of 39 boxers and mixed martial arts (MMA) fighters[xxi].
A pre-fight K-D test baseline was determined and a post-fight K-D test was also administered after bouts. Post-fight K–D time scores were significantly worse (higher) than pre-fight scores for participants who had sustained head trauma during the match. Additionally, statistical analysis showed that the K–D test had high test-retest reliability. The researchers concluded that the King–Devick Test is an accurate and reliable method for identifying players with head trauma, and is a strong candidate rapid sideline screening test for concussion.

There have been over 50 peer-reviewed studies published supporting King-Devick Test as a quick, objective indicator of suboptimal brain function and correlates with indicators of neurological conditions such as: concussion, reading disabilities, hypoxia, multiple sclerosis, Parkinson’s disease, and extreme sleep deprivation.

 
 
Validated and promoted by the prestigious Mayo Clinic, the King-Devick test is fast becoming the standard for assessment of pre-season, sideline and return-to-play ocular tracking, an essential component of concussion management.
 
 
 
 
IMPACT TESTING
Players wore an XPatch impact sensor (accelerometer) fitted behind their ear from X2 Biosystems[xxii], a leading medical tech company based in Seattle, Washington. X2 sensors and their software are used by National Football League (NFL), Major League Baseball (MLB), National Hockey League (NHL), Major League Soccer (MLS) in the US and by Saracens RUFC (UK Rugby) as well as in research initiatives based in New Zealand[xxiii].
 
Each week, the authors attended the fixture and personally fitted the sensors behind the ears of players. At the end of the match, these were then collected (sometimes requiring an hour scouring the ground for ones that had dislodged.
 
 
    
 
The data was downloaded overnight and analysed and a top-level report was emailed to coaching and medical staff highlighting players who had sustained high level impacts and rotational forces.
 
    
 
 At the next training (typically two days after the game), players with high g-force linear were assessed against the cognitive and physical baselines established earlier in the season.
 
In addition to these linear forces, rotational forces measures in rad/s2 were recorded. It is hyothesised that these forces may be even more important than direct linear g-force in the causation of brain injury and concussion[xxiv].
 
On the basis of these results, input was given to the coaching and medical staff on the concussion status of players which might never otherwise have been known.
 
RESULTS: IMPACT TESTING
 

Frequency of head impacts

 
Over the duration of the study there were 18,877 impacts to the head greater than 10g (range 10g to 170g) (see Table 1) recorded over the duration of the study.
 
The mean number of impacts per player over the duration of the season of matches was 932 ±623 resulting in a mean of 82 ±69 impacts to the head per player, per match over the duration of the season of matches. The mean impacts per match were 1,242 ±520 resulting in a mean of 91 ±45 impacts to the head per player position, per match. Forwards recorded more impacts than backs (RR: 1.63 [95% CI: 1.59 to 1.66]; p<0.0001) over the duration of the season of matches.
 

Magnitude of head impacts

 
The impact magnitudes for linear accelerations were skewed to the lower values (Sp=3.27 ±0.02; p<0.0001) with a mean (±SD) linear acceleration of 21 ±14g. Resultant rotational accelerations were also skewed to the lower values (Sp=2.21 ±0.02; p<0.0001) with a mean (±SD) linear acceleration of 3,107 ±3,002 rad/s2

 

Location of head impacts

 
The number of impacts to the areas of the head varied over the season of matches (see Table 3). The right side of the head recorded more impacts than the left side of the head (RR: 1.12 [95% CI: 1.09 to 1.14]; p<0.0001) for total impacts recorded.

 

Player positions and head impact characteristics

The number of head impacts, the number of linear and rotational accelerations, and the cumulative impact burden per player varied by player position (see Table 2).
In the forwards, the blind-side flanker recorded a mean of 166 ±56 impacts per match, and the hooker recorded a median resultant linear acceleration of 18 [13-28]g per match. For rotational accelerations, the hooker recorded a median of 2,597 [1,388-4,776] rad/s2. The hooker recorded the highest cumulative linear acceleration burden of 45,883g, and the highest cumulative rotational acceleration impact burden of 7,174,636 rad/s2. In the backs the second five-eight recorded a mean of 95 ±60 impacts per match, and the fullback recorded a median resultant linear acceleration of 19 [13-25]g per match. The centre recorded a median resultant rotational acceleration of 2,356 [1,329-4,111] rad/s2. The fullback recorded the highest cumulative linear acceleration burden of 28,717g, and the second five-eight recorded the highest cumulative rotational acceleration burden of 4,444,897 rad/s2.
 
The right side of the head recorded more impacts than the left side of the head for the backs (RR: 1.90 [95% CI: 1.84 to 1.96]; p<0.0001) over the duration of the match competition season. Forwards recorded more impacts to the left side of the head (RR: 1.21 [95% CI: 1.17 to 1.25]; p<0.0001) over the duration of the competition season.
 
The top left side of the head recorded the highest median linear accelerations (20 [14-31]g) for player position and for forwards (24 [17-33]g) while the left side of the head recorded the highest median linear accelerations for the backs (20 [14-29]g). Linear accelerations were higher on the left than the right side of the head for impacts recorded to the front (p<0.0001) and the top (p=0.0060) for player positions.
 
The top left side of the head recorded the highest median rotational accelerations (3,130 [1,818-5,896] rad/s2) for player positions, and this was similar for Forwards (4,081 [2,334-6,126] rad/s2). Forwards recorded higher rotational accelerations than backs for impacts to the front left (p=0.0422), back left (p<0.0001), left side (p<0.0001) and the top left (p=0.0388) of the head. There were 79 impacts (1%) above the linear injury risk limit and 2,376 impacts (15%) above the rotational injury risk limit (see Table 4). Forwards recorded more impacts in the moderate rotational acceleration (4,600-7,900 rad/s2) range than backs (p=0.0910).
 
The majority of impacts were in the low impact severity for linear (98%) and rotational (80%) data acquisition limits. The calculated mean and median RWE values varied by player position and positional groups (see Table 5). The calculated RWE varied by the risk function employed.
 
The values for the RWELinear, RWERotational and RWECP values for player positions ranged from 0.0001 to 0.2335, 0.0000 to 1.0000 and 0.0001 to 1.0000 respectively.
 
Two percent (394/18,877) of the RWERotational values were equal to 1 (100% risk of concussion) and less than 0.4% (7/18,877) of the RWECP values for the players were equal to 1. None of the RWELinear values exceeded 0.2335. Forwards recorded higher RWELinear (p<0.0001), RWERotational (p<0.0001) and REWCP (p<0.0001) than backs over the duration of the match competition season.