On-Field Predictors of Neuropsychological and Symptom Deficit Following Sports-related Concussion
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Clinical Journal of Sport Medicine, 13:222–229
© 2003 Lippincott Williams & Wilkins, Inc., Philadelphia
On-Field Predictors of Neuropsychological and Symptom Deficit
Following Sports-related Concussion
Michael W. Collins, PhD,* Grant L. Iverson, PhD,‡ Mark R. Lovell, PhD,*
Douglas B. McKeag, MD, MS,§ John Norwig, MA, ATC,㛳 and Joseph Maroon, MD†㛳
Departments of *Orthopaedic Surgery and †Neurological Surgery, University of Pittsburgh Medical Center, Pittsburgh,
Pennsylvania, U.S.A.; ‡Department of Psychiatry, University of British Columbia, Vancouver, Canada; §Department of Family
Medicine and Indiana University Center for Sports Medicine, Indiana University School of Medicine, Indianapolis, Indiana,
U.S.A.; and 㛳Pittsburgh Steelers Football Club, Pittsburgh, Pennsylvania, U.S.A.
Objective: Investigate the relationship between on-field Results: Odds ratios revealed that athletes demonstrating
markers of concussion severity and postinjury neuropsycholog- poor presentation at 2 days postinjury were over 10 times more
ical and symptom presentation in an athlete-specific population. likely (P < 0.001) to have exhibited retrograde amnesia fol-
Design: Case control study. lowing concussive injury when compared with athletes exhib-
Setting: Multicenter analysis of high school and college iting good presentation. Similarly, athletes with poor presenta-
athletes. tion were over 4 times more likely (P < 0.013) to have exhib-
Participants: A total of 78 athletes sustaining sports-related ited posttraumatic amnesia and at least 5 minutes of mental
concussion were selected from a larger sample of 139 con- status change. There were no differences between good and
cussed athletes. poor presentation groups in terms of on-field loss of conscious-
Assessment of Predictor Variables: On-field presence of ness.
disorientation, posttraumatic amnesia, retrograde amnesia, and Conclusions: The presence of amnesia, not loss of con-
loss of consciousness. sciousness, appears predictive of symptom and neurocogni-
Main Outcome Measures: ImPACT, a computerized tive deficits following concussion in athletes. Athletes present-
neuropsychological test battery, was administered pre-season ing with on-field amnesia should undergo comprehensive and
and, on average, 2 days postinjury. Good postinjury presenta- individualized assessment prior to returning to sport par-
tion (n ⳱ 44) was defined as no measurable change, relative to ticipation. Continued refinement of sports concussion grad-
baseline, in terms of both ImPACT memory and symptom com- ing scales is warranted in lieu of consistent findings that brief
posite scores. Poor presentation (n ⳱ 34) was defined as a loss of consciousness is not predictive of concussion injury
10-point increase in symptom reporting and 10-point decrease severity.
in memory functioning (exceeding the 80% confidence interval Key Words: concussion, sports, grading systems, severity,
for measurement error on ImPACT). Athletes failing to meet loss of consciousness, ImPACT
good or poor selection criteria (n ⳱ 61) were not included in
the analysis. Clin J Sport Med 2003;13:222–229.
INTRODUCTION consequences associated with concussion. The lack of
uniformity in the area is somewhat disconcerting given
The diagnosis and management of sports concussion is that conservative numbers estimate 300,000 concussions
likely the most elusive clinical condition facing the occurring per year in the United States.1
sports medicine practitioner. At the current time, there is The lack of consensus regarding the management
no consensus on the definitive diagnosis of concussion, of sports concussion is attributable, in part, to the lack
parameters regarding return to sport participation follow- of prospective data correlating outcome to initial
ing injury, and the short-term and long-term neurologic signs and symptoms of the injury. The lack of scientific
foundation in this area is at least partially responsible
for the numerous sports-related management scales that
Drs. Collins, Lovell, and Maroon are company shareholders of have been published to date (ie, a minimum of 17).
ImPACT Applications, LLC, the distribution company of ImPACT Each scale yields a numeric value that grades the sever-
(Immediate Postconcussion Assessment and Cognitive Testing). ity of injury, and most have corresponding return to
Received March 2002; Accepted March 2003. participation recommendations that withhold athletes
Reprints: Michael W. Collins, PhD, UPMC Sports Concussion Pro-
gram, Department of Orthopaedic Surgery, Center for Sports Medicine, from competition for varying degrees of time (see
3200 South Water St., Pittsburgh, PA 15203, U.S.A. E-mail: Collins et al2 and Johnston et al3 for a review of these
collinsmw@msx.upmc.edu issues).
222Neuropsychological and Symptom Deficit Following Sports-related Concussion 223
Severity Markers of Sports Concussion The pre-eminent focus of LOC in defining severity of
Examination of the various grading systems reveals concussive injury may also be related, at least in part, to
that classification of injury has been largely predicated animal model work conducted by Denny-Brown and
upon the presence and duration of postinjury disorienta- Russell13 in 1941 and later by Ommaya and Gennarelli14
tion, posttraumatic amnesia (PTA), and loss of conscious- in the 1970s and Gennarelli et al15 in the 1980s. In the
ness (LOC). Although multiple definitions of concussive latter work, various degrees of whiplash and rotational
injury exist, these immediate markers are traditionally types of injury were induced in squirrel monkeys, with
considered the hallmarks of injury.4–10 In defining these subsequent histologic analysis of brain tissue. Extrapo-
constructs, confusion or disorientation represents im- lating from this data, these researchers proposed 6 grades
paired awareness and orientation to surroundings, though of traumatic brain injury. In order of severity, postinjury
memory systems might not be frankly affected.4,5 PTA is animal presentation was consistent with isolated confu-
typically represented by the length of time between sion (grade 1), confusion and PTA (grade 2), confusion
trauma and the point at which the individual regains with posttraumatic and retrograde amnesia (grade 3), and
normal continuous memory functioning.5,6 Disorienta- LOC associated with progressively worse neurologic
tion and PTA are not mutually exclusive and can be outcomes (grades 4–6). Thus, severity was based upon
difficult to dissociate. To help clarify this issue, PTA level and content of consciousness. Moreover, these re-
represents a loss in memory from the point of injury until searchers felt that surface brain structures (ie, cortical)
the return of a full, ongoing memory process. Disorien- were involved with milder grades and that, proportion-
tation, in and of itself, is not associated with memory ately speaking, damage extended inward (ie, subcortical,
loss. Though given less focus, retrograde amnesia is also brain stem), further defining severity of injury.
an injury severity marker and is defined as the inability Within this historical context, it is understandable how
to recall events occurring during the period immediately LOC has been deemed the sine qua non characteristic of
preceding trauma.7–9 The length of retrograde amnesia concussive brain injury. Decades of work and research
will typically shrink over time.10 For example, as recov- may have been woven into the fabric of common knowl-
ery occurs, the length of retrograde amnesia may contract edge. Such work has been pioneering, and even revolu-
from hours to several minutes or seconds, though, by tionary, in terms of better understanding the biomechan-
definition, a permanent loss of memory preceding injury ics and general presentation of traumatic brain injury.
occurs. No current concussion grading system explicitly Importantly, however, extrapolating from this data to de-
includes retrograde amnesia as a criterion for injury se- scribe sports concussion aptly may potentially be erro-
verity. Lastly, LOC represents a state of brief coma in neous and misleading.
which the eyes are typically closed and the athlete is In fact, in subsequent writings, Gennarelli15 pointed
unresponsive to external stimuli.5,8,9 By definition, any out that although their animal model paradigm worked
individual experiencing LOC experiences concomitant well for moderate to severe traumatic brain injury, such
PTA, both for the period of LOC and potentially for the a model did not necessarily reflect predictors of outcome
subsequent time span until memory systems are func- associated with mild brain injury or concussion. For ex-
tional. An athlete may or may not experience retrograde ample, proper behavioral assessment of confusion and
amnesia with LOC. posttraumatic or retrograde amnesia was obviously im-
possible in the squirrel monkey. As such, these research-
Historical Context of Defining Severity of ers stated that their work had questionable relevance to
Sports Concussion the phenomenology of injury that is typical of concussive
With the exception of the revised Cantu grading sys- injury in humans. Nonetheless, this animal model re-
tem,11 all current sports concussion severity rating scales search is explicitly referenced as evidentiary support of
base an injury with positive LOC as more severe than the Colorado and American Academy of Neurology con-
concussive injury with positive confusion or amnesia cussion management guidelines.16,17 Moreover, the strik-
with no LOC.2,3 Inherent in this assumption is that ing similarity of this animal model work and correspond-
greater morbidity and worse outcome are associated with ing grading system to current sports concussion manage-
LOC than with other markers of concussion. The genesis ment parameters is unmistakable. However, no current
of these recommendations, at least in part, appears at- study correlates postinjury symptom presentation with
tributable to decades of research generalizing from more common markers of concussion severity in an athlete-
severe head trauma. For example, in the 1920s, Sy- specific population.
monds12 defined LOC as the hallmark of concussive in-
jury. In describing concussion, Symonds stated, Moving Forward
the patient is completely unconscious and in a state of
To date, the study of recovery from sports concussion
flaccid paralysis. In a severe case the respiratory and has been limited by several factors. Traditional neurodi-
cardiac functions may hardly continue. In a few minutes agnostic techniques, such as CT scan, MRI, and neuro-
recovery begins; the visceral reflexes are the first to logic examination, though invaluable in determining
return, and vomiting is common at this stage. The other more severe intracranial pathology (eg, skull fracture and
cerebral functions recover more gradually, and there
may be complaint of headache and giddiness, but at the hematoma), are generally insensitive to measuring the
end of 24 hours, in an uncomplicated case of concus- subtle effects of concussion.18 Importantly, however, the
sion, recovery should be complete.12 past decade has seen systematic advancements through
Clin J Sport Med, Vol. 13, No. 4, 2003224 Collins et al
the utilization of paper and pencil neuropsychological time composite scores were significantly related to
tests to delineate subtle aspects of injury. In fact, both the postinjury headache (at 7 days postinjury) in a sample of
National Football League 19 and National Hockey 110 concussed high school athletes.25
League20 currently implement baseline (preinjury) and ImPACT’s memory composite consists of 5 subtest
postconcussion neuropsychological testing protocols to scores measuring different aspects of the construct, in-
help manage sports concussion clinically. Moreover, the cluding verbal (word) learning and recognition memory,
current availability of computerized neuropsychological visual associative memory, visual working memory and
test platforms,21 though still in the preliminary state of recognition, and letter memory. This composite index
validation, allows wider use of this protocol at the high represents the average percent correct score for these
school and collegiate levels of competition.22 As a result tasks. All stimuli within the inventory are randomized
of these advances, our base of knowledge regarding during each administration to minimize practice effects
sports-related concussion is developing rapidly, and that are commonly associated with neuropsychological
more evidenced-based approaches to management may testing. ImPACT also includes the Postconcussion
be on the horizon. Symptom Scale (Table 1),26 which is a 22-item scale
The current study was designed to investigate the re- (graded 0 to 6 in terms of severity) that measures symp-
lationship between the on-field markers of concussion toms commonly associated with concussion. This scale
severity (disorientation, posttraumatic or retrograde am- has also been shown to be sensitive to the effects of
nesia, LOC) and postinjury neuropsychological and concussion.24–26
symptom status in an athlete-specific population. Administration of the computerized neuropsychologi-
cal test battery was supervised by a team of clinical
METHODS neuropsychologists, certified athletic trainers, and/or
physicians who were thoroughly trained in the adminis-
Program Protocol and Outcome Measures
tration of the measures. Training was completed at each
Appropriate review for research with human subjects
site through formal seminars presented by 1 of the lead
was granted through the University of Pittsburgh Medi-
authors (M. W. C. and M. R. L.). Given that ImPACT is
cal Center (UPMC) to conduct this study. All subjects
a self-administered test battery, all information is gath-
participated in the UPMC Sports Concussion Program.22
ered in a standardized manner. Further, data accumulated
High school and college athletes from the states of Penn-
during the administration process are generated auto-
sylvania, Illinois, and Maine were involved in the current
matically within a detailed clinical report. Thus, there
study. The UPMC Sports Concussion program helps fa-
was no variation in administration or data collection
cilitate the implementation of a baseline and postinjury
among participating sites.
neuropsychological testing protocol to help determine
return to play in athletes sustaining concussion. A cor-
nerstone of the program is the utilization of ImPACT
(Immediate Post Concussion Assessment and Cognitive TABLE 1. Postconcussion Symptom Scale26
Testing), a computer-based demographic, symptom, and Rating
neuropsychological test inventory.23
Symptom None Moderate Severe
Baseline Evaluation
All participating athletes underwent a baseline or pre- Headache 0 1 2 3 4 5 6
injury evaluation and were administered the computer- Nausea 0 1 2 3 4 5 6
Vomiting 0 1 2 3 4 5 6
ized test battery prior to the 2000 and 2001 athletic sea- Balance problems 0 1 2 3 4 5 6
sons. Baseline data were collected during the off-season Dizziness 0 1 2 3 4 5 6
(ie, preseason). ImPACT consists of a detailed symptom Fatigue 0 1 2 3 4 5 6
inventory and demographic questionnaire (eg, relevant Trouble falling asleep 0 1 2 3 4 5 6
sport, medical, and concussion history information) as Sleeping more than usual 0 1 2 3 4 5 6
Sleeping less than usual 0 1 2 3 4 5 6
well as 7 test modules that measure aspects of neurocog- Drowsiness 0 1 2 3 4 5 6
nitive functioning, including tests of memory, reaction Sensitivity to light 0 1 2 3 4 5 6
time, processing speed, and impulse control. For the cur- Sensitivity to noise 0 1 2 3 4 5 6
rent study, memory tasks were specifically used to help Irritability 0 1 2 3 4 5 6
Sadness 0 1 2 3 4 5 6
determine the postinjury status of the athlete. These Nervousness 0 1 2 3 4 5 6
memory scores derived from ImPACT have recently Feeling more emotional 0 1 2 3 4 5 6
been shown to be sensitive to the effects of sports-related Numbness or tingling 0 1 2 3 4 5 6
concussion.24 Specifically, when compared with nonin- Feeling slowed down 0 1 2 3 4 5 6
jured controls and with baseline levels, 64 athletes sus- Feeling mentally “foggy” 0 1 2 3 4 5 6
Difficulty concentrating 0 1 2 3 4 5 6
taining mild concussion had reduced memory function- Difficulty remembering 0 1 2 3 4 5 6
ing at 3 postinjury intervals (36 hours, days 4 and 7). Visual problems 0 1 2 3 4 5 6
Moreover, in the noninjured control group, ImPACT’s
memory composite scores revealed no significant prac- Total score 0 1 2 3 4 5 6
tice effects and were stable across the testing sessions. In Reprinted with permission from J Head Trauma Rehabil. 1999;9:
another recent study, ImPACT’s memory and reaction 193–198.
Clin J Sport Med, Vol. 13, No. 4, 2003Neuropsychological and Symptom Deficit Following Sports-related Concussion 225
Postconcussion Evaluation lowing injury, this form was faxed or e-mailed to the
All athletes underwent preseason testing and were re- UPMC Sports Concussion Program, where consultation
evaluated with ImPACT within 5 days of sustaining a was conducted regarding the athlete’s injury status and
concussion (mean, 1.7 days; median, 1 day). In-season return to play considerations (via neuropsychological test
concussion was diagnosed based upon the on-field pre- results and postinjury symptom presentation).
sentation of 1 or more of the following symptoms after a
Formulation of Outcome Groups
blow to the head or body: (1) any observable alteration in
A sample of 78 concussed amateur athletes partici-
mental status or consciousness; (2) a constellation of
pated in the current study. These athletes were selected
self-reported symptoms, such as posttraumatic headache,
from a larger sample of concussed athletes (n ⳱ 139).
photosensitivity, nausea or vomiting, dizziness, and so
Subjects for this study met specific criteria to create bi-
forth, and (3) LOC, disorientation, PTA, or retrograde am-
nary, cleanly defined outcome groups that served as the
nesia as identified by on-field examination. Initial diagnosis
dependent variable for the current study. The first group
of concussion was made by sports medicine practitioners
was asymptomatic at the follow-up evaluation and con-
who were present on the sideline at the time of injury.
sidered to have good postinjury presentation (n ⳱ 44)
Sports medicine practitioners at participating institu-
and the second group was overtly symptomatic at the
tions carefully documented information pertaining to
postconcussion evaluation and demonstrated poor imme-
postconcussion markers of injury. At the aforementioned
diate presentation (n ⳱ 34). Selection criteria were de-
training seminars, athletic trainers and/or physicians
fined a priori, and all subjects meeting these criteria were
were trained to identify the on-field markers of concus-
included in the current analysis. Specifically, poor
sion, including disorientation, PTA, retrograde amnesia,
postinjury presentation was defined as a decline in
and LOC. On-field disorientation was assessed by ques-
memory, as measured by the ImPACT memory compos-
tioning the athlete’s postinjury awareness and orientation
ite score, of 10 or more points and an increase in self-
to surroundings (eg, name, current stadium, city, oppos-
reported symptoms, as measured by at least a 10-point
ing team, current month and day). Athletes presenting
increase on the Postconcussion Symptom Scale, as com-
with any level of on-field disorientation in this regard
pared with baseline. In contrast, athletes with good
were classified in the positive disorientation group. On-
postinjury presentation demonstrated no measurable
field PTA was assessed through immediate and delayed
change, relative to baseline, in terms of both memory and
(eg, 0, 5, 15 minute) memory for 3 words (eg, girl, dog,
symptom reporting. In the poor presentation group, the
green). The presence of on-field PTA was also assessed
level of change represented nearly 1 pooled standard
at the postinjury follow-up evaluation by documenting
deviation drop in memory performance and approxi-
the athlete’s ability to recall specific events that occurred
mately 1⁄2 of a pooled standard deviation increase in
immediately subsequent to the trauma (eg, memory of
symptoms, based upon the standard deviations presented
returning to sideline, memory for subsequent plays, and
in Table 2. Moreover, a change of 10 or more points
so forth). Any loss of memory in this latter regard indi-
exceeds the 80% confidence interval for presumed mea-
cated positive presence of PTA. On-field retrograde am-
surement error, based on reliable change analyses of the
nesia was assessed by having the athlete recall events
ImPACT memory composite score.27 All athletes from
occurring just prior to trauma (eg, memory for play or
the larger sample were included in the current analysis if
plays preceding trauma, events in previous quarter, and
they met the good or poor symptom presentation criteria.
so forth). Retrograde amnesia was also documented at
Athletes from the larger sample exhibiting partial or
the postinjury evaluation by assessing the athlete’s abil-
equivocal adverse effects of concussion (failing to meet
ity to recall events just prior to trauma. Any loss of
good or poor outcome groups, n ⳱ 61) were not included
memory in this regard indicated positive presence of ret-
in the current analysis.
rograde amnesia. LOC was documented when an athlete
Demographic and baseline and postinjury ImPACT
was unresponsive to external stimuli and in paralytic
memory and symptom scores are presented for the out-
coma as reported by teammates and/or on-field evalua-
come groups in Table 2. These data, as defined by our
tion. By definition, athletes experiencing LOC also ex-
subject selection criteria, illustrate the pronounced neu-
perienced a concomitant PTA (ie, loss of memory for the
duration of the unconscious state). For the purposes of
this study, athletes with any degree of LOC were catego- TABLE 2. Demographic variables and ImPACT Memory
rized in the positive LOC group. An athlete who sus- and Symptom scores for the 2 outcome groups
tained an additional period of PTA was also categorized Good Poor
in the positive PTA group. Markers of injury were not Variable Presentation Presentation
mutually exclusive, as athletes commonly display a con-
Age 15.5 (2.5) 17.4 (1.8)
stellation of these on-field symptoms. Education 11.0 (1.8) 11.0 (3.0)
ImPACT contains a standardized evaluation form that Injury-to-testing interval, d 2.0 (1.3) 1.4 (1.2)
requires the test administrator to input data regarding the Total Symptoms Score: baseline 10.9 (15.1) 10.7 (15.1)
presence and duration of these specific concussion mark- Total Symptoms Score: follow-up 9.8 (12.7) 45.6 (23.6)
Memory Composite Score: baseline 83.2 (10.6) 85.6 (8.2)
ers at the first postinjury evaluation. This form is auto- Memory Composite Score: follow-up 85.7 (10.2) 62.2 (11.9)
matically detailed and printed within the clinical report
that becomes part of the athlete’s medical record. Fol- (Value in parentheses is standard deviation).
Clin J Sport Med, Vol. 13, No. 4, 2003226 Collins et al
rocognitive and symptom differences between the out- athletes in the good presentation group. Athletes with
come groups. poor postinjury presentation were 10 times more likely to
experience retrograde amnesia and 4 times more likely to
RESULTS experience PTA. The 2 groups did not differ in the pro-
portion of subjects who experienced traumatic LOC or
In regards to the total sample (n ⳱ 78), 88.5% of sideline-assessed disorientation without amnesia. If the 4
subjects were male. Their average age was 16.8 years primary on-field severity markers are considered simul-
(SD, 2.4; range, 14–22), and their average education was taneously (positive LOC, disorientation, PTA, and retro-
11.0 years (SD, 2.4; range, 8–17). The breakdown of the grade amnesia), only 2.5% of athletes with good presen-
sample by sport was as follows: football, 69.2% (n ⳱ tation evidenced at least 3 of 4 of these markers, whereas
54); soccer, 10.3% (n ⳱ 8); hockey, 7.7% (n ⳱ 6); 28.1% of the athletes with poor presentation evidenced at
basketball, 6.4% (n ⳱ 5); lacrosse, 2.6% (n ⳱ 3); and least 3 of 4 abnormal markers. Athletes with pronounced
baseball, 1.6% (n ⳱ 2). High school athletes made up the postinjury symptoms and memory deficits were approxi-
majority of the sample (77.9%, n ⳱ 61), with 22.1% (n mately 10 times more likely to demonstrate at least 3 of
⳱ 17) college players. Approximately half of the sample 4 abnormal on-field markers of concussion severity.
reported a history of at least 1 previous concussion A variable representing mental status change for 5 or
(52.6%, n ⳱ 41), and 30.8% (n ⳱ 24) reported no his- more minutes was also created. Athletes were included in
tory of concussion. Approximately 1/5 of the sample this group if they exhibited 5 or more minutes of on-field
reported a history of sustaining multiple concussions in disorientation. This time cutoff was used because it rep-
the past (20.2%, n ⳱ 16). Concussion history data was resents a common unit of time that can be tracked rela-
missing for 16.7% (n ⳱ 13) of the sample. In terms of tively easily on the athletic playing field. Data pertaining
injury severity, 56% (n ⳱ 44) of athletes demonstrated to this variable were also entered directly into the
postconcussive symptoms, mental status changes, and/or ImPACT postinjury evaluation form. Only 15% of con-
amnesia that resolved within 15 minutes, whereas 44% cussed athletes with good postinjury presentation had
(n ⳱ 34) had symptoms persisting beyond 15 minutes. prolonged postinjury disorientation at the time of injury
Finally, 19.2% (n ⳱ 15) of the sample demonstrated as compared with 43.8% of players with poor postinjury
on-field LOC. Of these 15 athletes, 7 experienced LOC presentation (P ⳱ 0.007, OR ⳱ 4.4).
that persisted for less than 30 seconds, 4 experienced
LOC lasting between 30 seconds and 1 minute, 3 expe- DISCUSSION
rienced LOC lasting between 1 and 2 minutes, and 1
experienced LOC lasting longer than 2 minutes. No ath- Though recent animal model research has revealed
lete in our sample was noted to experience seizure ac- metabolic and pathophysiology aspects of concussive in-
tivity following the respective in-season concussion. jury,28 limited information exists examining the behav-
The majority of subjects in both outcome groups were ioral phenomenology of concussive injury in humans. A
male (91% vs. 85%, P ⳱ 0.441). The good versus poor lack of data in this regard has become particularly prob-
immediate outcome groups did not differ in age (P ⳱ lematic in the management of sports concussion.29 Given
0.101), education (P ⳱ 0.963), or history of concussion the inherent pressures of competition, athletes sustaining
(P ⳱ 0.401; 60% of the good presentation group and concussion are often needed for, and will even request,
70% of the poor presentation group had a history of 1 or return to sport participation in which the risk of subse-
more concussions). quent head impacts and trauma is salient. This is a con-
Statistical comparisons between the 2 outcome groups cern given the paucity of data regarding cumulative and
were conducted via 2 analyses with on-field markers of potentially catastrophic effects of multiple concus-
concussion severity as dependent variables. As seen in sions.30 There is currently much debate surrounding the
Table 3, athletes with poor postinjury presentation were lack of scientific foundation in current sports concussion
significantly more likely to experience retrograde amne- management parameters. Recent recommendations from
sia (P ⳱ 0.001) and PTA (P ⳱ 0.013) relative to those an international meeting on the topic of sports concus-
TABLE 3. On-field concussion severity markers by good and poor presentation groups
Good Poor 95% Confidence
On-field Injury Marker N Presentation Presentation 2 P Odds Ratio Interval
Positive LOC 75 11.9% 21.2% 1.2 0.275 — —
Retrograde amnesia 72 5.0% 34.4% 10.4 0.001 10.0 2.0–49.2
Posttraumatic amnesia 72 12.5% 37.5% 6.2 0.013 4.2 1.3–13.7
Any disorientation 67 72.4% 71.4% 0.01 0.934 — —
3–4 Abnormal markers 72 2.5% 28.1% 9.8 0.002 15.3 1.8–128.3
5+ Minutes disorientation 72 15.0% 43.8% 7.3 0.007 4.4 1.4–13.4
The total sample was 78 athletes. Due to the normal difficulties with collecting on-field markers, there were varying degrees of missing data. The
number of subjects who had each marker coded ranged from 67 to 75. The N column represents the number of subjects for whom data were available
for each category. Markers of injury are not mutually exclusive.
LOC indicates loss of consciousness.
Clin J Sport Med, Vol. 13, No. 4, 2003Neuropsychological and Symptom Deficit Following Sports-related Concussion 227
sion31 have reinforced the need to re-evaluate all con- postconcussion sequelae in athletes. Specifically, within
cussion grading systems and return to play management our study, athletes exhibiting pronounced postconcussion
directives. symptoms and memory deficits at approximately 2 days
The current study is the first to evaluate the relative postinjury were over 10 times more likely to have expe-
predictive value of on-field markers of concussion se- rienced any degree of retrograde amnesia as compared
verity in a sport-specific population. Recent work exam- with athletes exhibiting good outcome from injury. Simi-
ining concussive injury in a group of 195 trauma patients larly, athletes experiencing any degree of PTA were 4
(mostly car accident victims)32 revealed that LOC might times more likely to exhibit poor postinjury presentation
not be a potent predictor of postinjury neuropsycholog- from concussive injury. Athletes who experienced dis-
ical deficits. These patients were placed into 3 groups orientation (without the presence of amnesia) for 5 or
based upon positive, negative, or equivocal LOC as de- more minutes were also 4 times more likely to be in the
termined by post hoc recording of medical records. In- poor presentation group. Notably, nearly 3/4 of our
clusion criteria defined concussive injury as having a sample demonstrated some level of disorientation, high-
Glasgow Coma Scale (GCS) score between 13 and 15 lighting this sign as a common consequence of injury. It
and normal day-of-injury CT scan of the brain. Analyses appears that brief disorientation (5 minutes) was pre-
processing, memory, executive functioning, and verbal dictive of postinjury memory and symptom deficit. Thus,
and visual memory) existed between groups at approxi- it may be important to dissociate these signs of concus-
mately 4 days postinjury. In a follow-up study,33 similar sive injury. Consistent with the aforementioned research
analyses were conducted with 383 trauma patients. To with trauma patients, brief LOC was not predictive of
maximize relevance to concussion in sports, exclusion postconcussion symptoms or neurocognitive deficits.
criteria for this latter study included (1) GCS score less Lastly, though not a specific focus of the current study,
than 14, (2) skull fracture, (3) intracranial abnormality on it is interesting to note that history of concussion did not
day-of-injury CT, and (4) age greater than 45 years. differ between the good and poor presentation groups.
Moreover, patients were young (mean age, 28 years), and The underlying neuropathology relating to our find-
the majority had GCS scores of 15. Once again, no dif- ings is difficult to ascertain. The presence of amnesia
ferences in neuropsychological test performance was in- likely represents metabolic or other dysfunction in the
dicated among those experiencing no, equivocal, or posi- hippocampal and/or temporal cortical areas “either
tive LOC. All injured patients were evaluated on neuro- through disconnection from other brain regions or from
psychological measures within 7 days of injury. Results disruption in their intrinsic circuitry.”34 Conversely, sub-
of these prior studies raise doubts regarding the impor- cortical involvement (eg, reticular activating system),
tance of LOC as a predictor of outcome following con- though debatable, may be more likely responsible for
cussion. LOC.9 The frontal and temporal lobes are particularly
When compared with these previous studies, the cur- vulnerable to the acceleration, deceleration, and rota-
rent study has specific methodological advantages that tional forces that are common in the biomechanics of
allow a finer discrimination of the frequency, duration, sports concussion. Greater sensitivity of the cortical sys-
and importance of injury severity markers in concussed tems to memory dysfunction and presence of postcon-
athletes. For example, even though all trauma patients cussion symptoms may potentially help account for our
were deemed to experience mild brain injury, some of findings. It should be noted that our current study defined
these injuries were likely more severe than that evi- poor presentation specifically based upon attenuated
denced in the current sports-specific study. Moreover, memory functioning (as measured by ImPACT) and the
unlike the prior studies, all subjects within our sample presence of postconcussion symptoms. A potential com-
underwent preinjury neuropsychological and symptom peting hypothesis is that LOC may be associated with
evaluation that provided a direct comparison to their disparate symptom patterns and effects on specific cog-
postinjury status. Further, the current study includes nitive domains other than memory (eg, reaction time or
analyses determining the relevance of LOC and other processing speed). Notably, however, the aforemen-
markers of concussion severity. Postinjury markers of tioned studies with trauma patients failed to find such a
injury were also more accurately collected within the relationship. Nonetheless, a finer analysis of these issues
context of the athletic contest. Specifically, certified ath- is certainly indicated in future studies.
letic trainers and physicians familiar with these con- It is important to note that the majority of patients in
structs were trained to collect and document this infor- our sample demonstrated LOC that lasted for less than 1
mation systematically during both on-field and postin- minute. Therefore, this study investigates the relative im-
jury evaluation periods. Although inherent difficulty portance of brief LOC in determining symptom presen-
arises in separating the constructs of disorientation, am- tation from concussive injury. In our experience and
nesia, and LOC, we attempted to operationalize the con- based upon our data, prolonged LOC is a relatively rare
structs clearly and collect the data in a standardized fash- phenomenon in sports concussion. At the current time,
ion. there is no feasible way to determine an appropriate cut-
Results of our current analysis suggest that amnesia off when LOC becomes predictive of symptomatic, neu-
and not LOC may be more predictive of pronounced rocognitive, or neurologic consequence. Nonetheless, the
Clin J Sport Med, Vol. 13, No. 4, 2003228 Collins et al
continued pre-eminent emphasis of LOC in current to date, the general lack of attention on this construct in
sports concussion management directives may be mis- the sports concussion literature. Though Cantu11 de-
guided and erroneous. Further, this issue transcends the scribes the construct of retrograde amnesia in a recent
arena of sports given that most physician offices and revision of his grading system, no scale explicitly uses
emergency or trauma centers predicate level and scrutiny this as a criterion to define severity of injury. To date,
of mild brain injury care based upon the LOC construct. rudimentary assessment of retrograde amnesia has been
Again, this may be deserved in individuals experiencing fairly well documented in available systematic on-field
more prolonged LOC, though symptoms such as amne- mental status exams.39,40 Continued research should fo-
sia, disorientation, and postconcussion symptoms are tra- cus on the refinement of these on-field evaluative tech-
ditionally given short shrift in determining subsequent niques to document severity markers of injury. Immedi-
management directives. ate and serial assessment of all concussion severity
Continued refinement of sports concussion grading markers and symptoms are warranted given that these
scales may be warranted in lieu of consistent findings indicators may evolve over time.35
that brief LOC may not be predictive of concussion in- In summary, given current results, it is our belief that
jury severity. At present, all existing grading scales, with presentation of postinjury amnesia, prolonged disorien-
the exception of the revised Cantu11 criteria, base in- tation, or presence of postconcussion symptoms should
creased severity of injury on the LOC construct. Current necessitate physician referral and more thorough and in-
results suggest a rather strong relationship between am- dividualized evaluation (beyond sideline or mental status
nesia and overt postinjury symptom presentation, though testing) prior to return to sport participation. Recent and
findings should certainly be replicated. Further studies in current research elucidates the sensitivity of baseline and
this regard may eventually lead to more evidence-based postinjury neuropsychological testing to help delineate
standards in terms of managing sports-related concus- underlying deficits associated with injury.24,25,35–38,41
sion. The continued refinement and validation of ImPACT and
Specific limitations of the current study deserve men- other computerized neuropsychological test batteries21
tion and should be considered carefully within the con- might increase the availability of these postinjury assess-
text of our current data. First, this study examined the ment tools to all levels of sport participation.
issue of memory and symptom outcome at 2 days postin-
jury, whereas previous sports-specific studies have re- REFERENCES
vealed similar and other neuropsychological deficits at 1. Centers for Disease Control and Prevention. Sports-related recur-
longer intervals.24,25,35–38 Researchers are encouraged to rent brain injuries: United States. MMWR Morb Mortal Wkly Rep.
replicate and expand upon our current findings at longer 1997;46:224–227.
intervals in the recovery process from concussion. Sec- 2. Collins MW, Lovell MR, McKeag DB. Current issues in managing
ond, our study defined good and poor postconcussion sports-related concussion. JAMA. 1999;282:2283–2285.
3. Johnston KM, McCrory P, Mohtadi NG, et al. Evidence-based
presentation based upon empirically defined change on review of sports-related concussion: clinical science. Clin J Sport
the ImPACT memory and symptom scale composite Med. 2001;11:150–159.
scores. The presence and duration of on-field markers of 4. Fisher CM. Concussion amnesia. Neurology. 1966;16:826–830.
concussion severity were reported for the 2 respective 5. Symonds CP. Concussion and its sequelae. Lancet. 1962;1:1–5.
groups. Though such methodology is a reasonable first 6. Russell WR, Smith A. Post-traumatic amnesia after close head
injury. Arch Neurol. 1961;5:16–29.
step, future studies will benefit from a prospective analy- 7. Yarnell PR, Lynch S. Retrograde memory immediately after con-
sis with a priori hypotheses established on the relative cussion. Lancet. 1970;1:863–864.
importance of each specific marker of concussive injury. 8. Cartlidge NEF, Shaw DA. Head Injury. London: WB Saunders;
Similarly, this study reports no data in athletes with mild 1981.
or equivocal postinjury concussive deficits. The current 9. Plum F, Posner JB. States of acutely altered consciousness. In:
Plum F, Posner JB, eds. The Diagnosis of Stupor and Coma. 3rd
study also defined postinjury presentation using the ed. Philadelphia, PA: FA Davis; 1882:3–5.
ImPACT memory and symptom composite scores. Fu- 10. Benson DF, Geschwind N. Shrinking retrograde amnesia. J Neurol
ture studies should replicate current findings using addi- Neurosurg Psychiatry. 1967;30:539–544.
tional computerized neuropsychological measures and/or 11. Cantu RC. Posttraumatic retrograde and anterograde amnesia:
pathophysiology and implications in grading and safe return to
other outcome variables and could consider symptom play. J Athletic Training. 2001;36:244–248.
and memory scores as separate variables. Lastly, the cur- 12. Symonds CP. The differential diagnosis and treatment of cerebral
rent study examined the relative importance of on-field states consequent upon head injuries. BMJ. 1928;4:829–832.
markers in a combined sample of high school and college 13. Denny-Brown D, Russell WR. Experimental cerebral concussion.
athletes. Given the potential of differential age-related Brain. 1941;64:93–163.
14. Ommaya AK, Gennarelli TA. Cerebral concussion and traumatic
response to concussive injury, future studies should be unconsciousness: correlation of experimental and clinical observa-
conducted in high school, college, and professional tions on blunt head injuries. Brain. 1974;97:633–654.
samples of athletes. 15. Gennarelli TA, Thibaut LE, Adams JH, et al. Diffuse axonal injury
Though preliminary, our results suggest that on-field and traumatic coma in the primate. Ann Neurol. 1982;12:564–574.
assessment of sports concussion consistently incorporate 16. Kelly JP, Rosenberg JH. Diagnosis and management of concussion
in sports. Neurology. 1997;48:575–580.
mental status testing that carefully elicits signs of amne- 17. Kelly JP. Loss of consciousness: pathophysiology and implications
sia. Particular focus should be considered in assessing in grading and safe return to play. J Athletic Training. 2001;36:
retrograde amnesia, especially given current results and, 249–252.
Clin J Sport Med, Vol. 13, No. 4, 2003Neuropsychological and Symptom Deficit Following Sports-related Concussion 229
18. Johnston KM, Prito A, Chankowsky J, et al. New frontiers in 30. Rabadi MH, Jordan BD. The cumulative effects of repetitive con-
diagnostic imaging in concussive head injury. Clin J Sport Med. cussion in sports. Clin J Sport Med. 2001;11:194–198.
2001;11:166–175. 31. Aubry M, Cantu R, Dvorak J, et al. Summary and agreement
19. Lovell MR, Collins MW. Neuropsychological assessment of the statement of the 1st international symposium on concussion in
head injured professional athlete. In: Bailes J, Day A, eds. Neuro- sport, Vienna 2001. Clin J Sport Med. 2002;12:6–11.
logical Sports Medicine: A Guide for Physicians and Athletic 32. Iverson GL, Lovell MR, Smith SS. Does brief loss of conscious-
Trainers. Rolling Meadows, IL: Neurosurgical Topics Book Series ness affect cognitive functioning after mild head injury? Arch Clin
of the American Association of Neurological Surgeons; Neuropsychol. 2000;15:643–648.
2001:24–37.
33. Lovell MR, Iverson GL, Collins MW, et al. Does loss of con-
20. Lovell MR, Burke CJ. The NHL Concussion Program. In: Cantu R,
sciousness predict neuropsychological decrements after concus-
ed. Neurologic Athletic Head and Spine Injury. Philadelphia, PA:
sion? Clin J Sport Med. 1999;9:193–198.
WB Saunders; 2000:32–45.
21. Schnirring L. How effective is computerized concussion manage- 34. Wilson JL, Teasdale GM, Hadley DM, et al. Post-traumatic am-
ment? Phys Sportsmed. 2001;29:11–16. nesia: still a valuable research. J Neurol Neurosurg Psychiatry.
22. Collins MW, Hawn KL. The clinical management of sports con- 1994;56:198–201.
cussion. Curr Sports Med Rep. 2002;1:12–21. 35. Echemendia RJ, Putukian M, Macklin RS, et al. Neuropsycholog-
23. Maroon JC, Lovell MR, Norwig J, et al. Cerebral concussion in ical test performance prior to and following sports related mild
athletes: evaluation and neuropsychological testing. Neurosugery. traumatic brain injury. Clin J Sports Med. 2001;11:23–31.
2000;47:659–672. 36. Collins MW, Grindel SJ, Lovell MR, et al. Relationship between
24. Lovell MR, Collins MW, Iverson GL, et al. Recovery from mild concussion and neuropsychological test performance in college
concussion in high school athletes. J Neurosurg. 2003;98:296–301. football players. JAMA. 1999;282:964–970.
25. Collins MW, Field M, Lovell MR, et al. Relationship between 37. Hinton-Bayre AD, Geffen GM, Geffen LB. Concussion in contact
headache and neuropsychological test performance in high school sports: reliable change indices of impairment and recovery. J Clin
athletes. Am J Sports Med. 2003;31:168–173. Exp Neuropsychol. 1999;21:70–86.
26. Lovell MR, Collins MW. Neuropsychological assessment of the 38. Macciocchi S, Barth JT, Alves W, et al. Neuropsychological func-
college football player. J Head Trauma Rehabil. 1998;13:9–26. tioning and recovery after mild head injury in collegiate athletes.
27. Iverson GL, Lovell MR, Collins MW, et al. Tracking recovery Neurosurgery. 1996;39:510–514.
from concussion using ImPACT: applying reliable change meth-
odology. Paper presented at: National Academy of Neuropsychol- 39. Collins MW, Hawn KL. The clinical management of sports con-
ogy Annual Conference; October 2002; Miami, FL. cussion. Curr Sports Med Rep. 2002;1:12–22.
28. Giza CC, Hovda DA. The neurometabolic cascade of concussion. 40. McCrea M, Kelly JP, Randolph C, et al. Standardized assessment
J Athletic Training. 2001;36:228–235. of concussion: on-site mental status evaluation of the athlete. J
29. McCrory P, Johnston K, Mohtadi NG, et al. Evidence-based re- Head Trauma Rehabil. 1998;13:27–35.
view of sports-related concussion: basic science. Clin J Sport Med. 41. Matser EJ, Kessels AG, Lezak MD, et al. Neuropsychological
2001;11:160–165. impairment in amateur soccer players. JAMA. 1999;282:971–973.
Clin J Sport Med, Vol. 13, No. 4, 2003You can also read
