Clinical Studies
Immediate Neurocognitive Effects of Concussion

PATIENTS AND METHODS

Subjects and study design

A total of 2385 varsity football players from 30 high schools and 15 colleges were exposed to two separate protocols as part of a larger study of the utility of the SAC in the assessment of sports-related concussion. One protocol enrolled players (n=1189) from 15 high schools and 4 universities, all of whom underwent baseline testing with the SAC before the start of the season. The second protocol included players (n=1196) from an additional 15 high schools and 4 universities but the subjects did not undergo preseason baseline testing. The no-baseline protocol was included in this study for assessment of the utility of derived cutoff SAC scores in detecting concussion in the absence of preinjury baseline data with which to compare the subject’s performance immediately after injury. This issue is not as relevant to the assessment of sports concussion (where baseline testing is feasible) but pertains to the eventual application of the SAC to clinical settings outside sports (e.g., in the emergency department or at the trauma scene), where baseline testing is not possible. This study was fully approved by the institutional review board for the study of human subjects at Waukesha Memorial Hospital.

All injured subjects were identified and enrolled in the study protocol by a certified athletic trainer who was present on the sidelines during the athletic contest or practice. Concussion was defined according to the American Academy of Neurology practice parameter (3) (i.e., trauma-induced alteration in mental status, with or without LOC) and the American Congress of Rehabilitation Medicine definition of MTBI (i.e., alteration in mental state, LOC for 30 min or less, and PTA for not more than 24 h) (4). Criteria contributing to the identification of an injured player by the certified athletic trainer included the mechanism of injury (e.g., acceleration or rotational forces applied to the head), symptoms reported or signs exhibited by the player (e.g., alteration in mental status, confusion, headache, dizziness, or memory problems) (3, 20), and reports by teammates and other witnesses regarding the condition of the injured player. The occurrence and duration of LOC and amnesia were documented immediately after injury by the athletic trainer who examined the player. Amnesia included retrograde amnesia and PTA, which were defined as the inability to recall events immediately before (e.g., aspects of the play or event and the score of the game) and after (e.g., exiting the field and aspects of the examination) the injury, respectively. Injured subjects were not evaluated by a physician and did not undergo neuroimaging studies.

A total of 91 injuries (3.82% of the total sample) were documented during the study, including 45 from the protocol involving baseline testing (3.79%) and 46 from the no-baseline protocol (3.85%). Of the 91 injured subjects, 76 (84.4% of the injured sample) sustained no LOC and exhibited no retrograde amnesia or PTA after concussion. Eight subjects (8.79%) displayed PTA but no LOC. The remaining seven subjects (7.69%) sustained observed LOC and exhibited PTA. Naturally, there were no subjects who sustained LOC without amnesia. The duration of LOC in this sample was quite short, with a maximum of less than 1 minute in all cases. The length of amnesia was similarly short, ranging from a few seconds to several minutes. For the purpose of presenting the results, the group with no PTA or LOC is referred to as the no LOC/no PTA group, the group with PTA but no LOC as the PTA group, and the group with LOC and PTA as the LOC group. No subjects in this study experienced recurrent concussion. No physical neurological abnormalities were exhibited by any of the injured subjects, and no neurosurgical complications, catastrophic outcomes, or cases of second-impact syndrome (34) were encountered in this study.

All players identified by the certified athletic trainer as having sustained a possible concussion, according to the study’s injury criteria, were tested with the SAC on the sideline immediately after injury. This immediate assessment was conducted within 5 minutes after the actual occurrence of injury in all cases. Follow-up testing with the SAC was then performed 15 minutes, 48 hours, and 90 days after injury. As noted, one-half of the subjects enrolled in the study did not undergo baseline testing because of the design of the study. There were cases of missing data at each of the follow-up assessment points. The cases of missing data from follow-up testing were not unexpected, given the demanding responsibilities of the athletic trainers during games and the large number of schools and players involved. The 15-minute follow-up assessments (41% of subjects tested) were most difficult for athletic trainers to complete during games, because of their competing clinical demands, and the 90-day follow-up assessments (43% of subjects tested) created complications because of a lack of access to athletes after the end of the season. In contrast, 86% of players underwent assessments at the intermediate 48-hour time point. Overall, 79 of the 91 injured players (86.8%) underwent evaluations at one or more of the follow-up assessment points (15 min, 48 h, and 90 d), and 30 subjects (33.0%) underwent assessments at all three data points.

Neurocognitive assessment

The SAC (31) is a brief screening instrument designed for the neurocognitive assessment of concussion by a non-neuropsychologist with no prior expertise in psychometric testing. Previous studies have demonstrated the psychometric properties and clinical sensitivity of the instrument in assessing concussion (29, 30). The SAC requires approximately 5 minutes for assessment of four domains of cognition, including orientation, immediate memory, concentration, and delayed recall, summing to a total composite score of 30 points (Fig. 1). Three equivalent alternate forms of the test were used in this study, to minimize practice effects from repeated administration.

Statistical analyses

Data were analyzed descriptively and with tests of significance both between groups (e.g., injured versus noninjured subjects, no LOC/no PTA versus PTA versus LOC group, and baseline versus no-baseline protocol) and between assessment points. Primary dependent variables were the SAC total score and the four SAC subtest scores (i.e., orientation, immediate memory, concentration, and delayed recall scores). After preliminary analyses comparing baseline data for noninjured and injured subjects, SAC scores for injured subjects were compared with the population mean baseline SAC scores for noninjured subjects. Paired-sample t tests were used to compare SAC scores after injury with baseline performance scores for injured subjects. Bonferroni correction was used to minimize Type I error rates associated with multiple comparisons. Analyses were also conducted to determine the presence of any selection bias affecting subject attrition rates at the 15-minute, 48-hour, and 90-day assessment points. Because of the small sample sizes, qualitative analyses were performed and figures were generated for comparison of SAC performance at follow-up assessment points by subjects with or without LOC and PTA. Data were analyzed with SPSS 10.0 statistical software (35).

Preliminary analysis

College players achieved a slightly higher SAC total score than did high school subjects during baseline testing, but this difference did not reach statistical significance [t(1,1187) = -1.87, P = 0.06]. There was no significant difference between the SAC scores of injured college and high school subjects immediately after concussion [t(1,89) = -1.62, P = 0.11] or at 15 minutes [t(1,35) = -0.94, P = 0.35], 48 hours [t(1,76) = -1.40, P = 0.17], or 90 days [t(1,37) = -0.91, P = 0.37] after injury. There was also no difference between high school and college injured subjects in terms of calculated changes in scores from baseline times to concussion [t(1,43) = -0.25, P = 0.80]. Therefore, injury data for high school and college subjects were combined for further analyses. There were no significant differences in SAC total scores immediately after injury for subjects tested with different forms of the SAC (A,B, and C) [F(2,88) = 0.12, P = 0.88].

Table 1 illustrates the similar characteristics of subjects from the baseline and no-baseline protocols in this study. The preseason baseline SAC total score for subjects who were subsequently injured during the study was not significantly different from that for the larger population of noninjured subjects. There were also no differences between SAC performance scores for injured subjects in the baseline protocol versus the no-baseline protocol at the time of injury or 15 minutes, 48 hours, or 90 days after injury (Table 1). Therefore, injury and follow-up data for the two groups were combined for further analysis. With respect to the possibility of injury severity systematically affecting subject attrition rate, there was no significant difference in SAC total scores at the time of injury for subjects who did or did not undergo an examination 15 minutes [t(1,89)= -0.20, P = 0.84], 48 hours [t(1,89) = 0.95, P = 0.35], or 90 days [t(1,89) = 0.26, P = 0.79] after injury. Furthermore, SAC scores for a subset of 30 subjects who underwent testing immediately after injury and at all assessment points did not differ from those for the larger group of injured subjects at the time of injury [t(1,89) = 0.08, P = 0.93] or 15 minutes [t(1,35) = 0.06, P = 0.95], 48 hours [t(1,76) = -0.64, P = 0.52], or 90 days [t(1,37) = -1.32, P = 0.20] after injury Therefore, group means for the time of injury and each follow-up assessment point were used to measure the immediate effects of injury and the slope of the postinjury recovery curve for all injured subjects.

Next section: Results