Symptom-based assessment of the severity of a concussion

Transcription

1 J Neurosurg 98: , 2003 Symptom-based assessment of the severity of a concussion DAVID ERLANGER, PH.D., TANYA KAUSHIK, PSY.D., ROBERT CANTU, M.D., JEFFREY T. BARTH, PH.D., DONNA K. BROSHEK, PH.D., JASON R. FREEMAN, PH.D., AND FRANK M. WEBBE, PH.D. Department of Neuroscience and Education, Columbia University, New York; HeadMinder, Inc., New York, New York; Department of Psychiatry, Albert Einstein College of Medicine, Bronx, New York; Department of Surgery, Emerson Hospital, Concord, Massachusetts; Department of Psychiatric Medicine, University of Virginia, Charlottesville, Virginia; and Department of Psychology, Florida Institute of Technology, Melbourne, Florida Object. Current grading systems of concussion and return-to-play guidelines have little empirical support. The authors therefore examined the relationships of the characteristics and symptoms of concussion and the history of concussion to three indicators of concussion severity number of immediate symptoms, number of symptoms at the initial follow-up examination, and duration of symptoms to establish an empirical basis for grading concussions. Methods. Forty-seven athletes who sustained concussions were administered alternate forms of an Internet-based neurocognitive test until their performances were within normal limits relative to baseline levels. Assessments of observer-reported and self-reported symptoms at the sideline of the playing field on the day of injury, and at follow-up examinations were also obtained as part of a comprehensive concussion management protocol. Although loss of consciousness (LOC) was a useful indicator of the initial severity of the injury, it did not correlate with other indices of concussion severity, including duration of symptoms. Athletes reporting memory problems at follow-up examinations had significantly more symptoms in general, longer durations of those symptoms, and significant decreases in scores on neurocognitive tests administered approximately 48 hours postinjury. This decline of scores on neurocognitive testing was significantly associated with an increased duration of symptoms. A history of concussion was unrelated to the number and duration of symptoms. Conclusions. This paper represents the first documentation of empirically derived indicators of the clinical course of postconcussion symptom resolution. Self-reported memory problems apparent 24 hours postconcussion were robust indicators of the severity of sports-related concussion and should be a primary consideration in determining an athlete s readiness to return to competition. A decline on neurocognitive testing was the only objective measure significantly related to the duration of symptoms. Neither a brief LOC nor a history of concussion was a useful predictor of the duration of postconcussion symptoms. KEY WORDS brain concussion athletic injury sports neuropsychology computerized cognitive testing C Abbreviations used in this paper: AAN = American Academy of Neurology; CRI = Concussion Resolution Index; LOC = loss of consciousness; PA = posttraumatic amnesia; SD = standard deviation; SE diff = standard error of difference. J. Neurosurg. / Volume 98 / March, 2003 URRENT guidelines for grading sports-related concussions base their return-to-play recommendations largely on two parameters: the severity of the injury and the patient s history of concussion. 11,18,21,24 The two most widely used guidelines are those of the AAN 1 and those of Cantu 4,5 (Table 1). In the assessment of injury severity, both grading systems take into account the nature and duration of key injury characteristics. The AAN guidelines emphasize the qualitative importance of LOC, whereas Cantu guidelines distinguish between brief and extended LOC, and draw attention to the duration of PA. Injuries are classified as Grade I (mild), Grade II (moderate), or Grade III (severe). Each grade is associated with a recommendation regarding the amount of time that should elapse following symptom resolution before an athlete may be exposed again to contact. More severe grades require longer interim periods. In addition to the grading systems, both guidelines include consideration of an athlete s history of concussion, especially within the same season, in determining his or her readiness to return to competition. This interaction of concussion grade (severity) and concussion history is shown in Table 2. Research on the validity of the variables used for these grading systems is scant. McCrea and colleagues 28 found that, in the immediate aftermath of concussion, on average, athletes who sustained LOC displayed greater decreases in performance in a brief assessment of general mental status than those who did not. In addition, they found that athletes who experienced PA performed more poorly (although not 477

2 D. Erlanger, et al. TABLE 1 Concussion severity grading guidelines Grade AAN Guidelines Cantu Guidelines I symptoms 15 mins all symptoms 30 mins & no LOC & no LOC II symptoms 15 mins symptoms 30 mins, but 7 days & no LOC and/or PTA 30 mins, but 24 hrs &/or LOC 1 min III any LOC symptoms 7 days &/or PA 24 hrs &/or LOC 1 min as poorly as athletes experiencing LOC) than those patients who did not have PA. By 24 hours postinjury, however, distinctions between groups could no longer be ascertained using tests of mental status. This finding lent empirical support to the following: 1) the value of considering both LOC and PA in an assessment of concussion severity; and 2) demonstrating the limitations of brief tests of mental status in predicting concussion severity and recovery. In a survey of athletic trainers, Guskiewicz and colleagues 17 reported an incidence of 88.9, 10.6, and 0.4% for Grade I, II, and III concussions, respectively, by using the Cantu scale in 1003 athletes. Athletes with Grade II concussions who experienced either a brief LOC ( 30 seconds) or a period of PA that lasted from 20 minutes to 24 hours, reported more overall symptoms than those with Grade I injuries, and these symptoms had significantly longer durations. In this study, PA was assessed by a retrospective estimate of the duration of memory loss. In a prospective study of college football players, Collins, et al., 6 reported an incidence of 68, 21, and 11% for concussion Grades I, II, and III, respectively, when applying the AAN guidelines to 19 injured athletes. These athletes were reported to have performed more poorly than controls on selected neuropsychological tests, including those of learning and memory. In that study, however, no comparisons between concussion grades or other statistical analyses were possible because only four athletes experienced Grade II concussions and only two Grade III injuries. Moreover, the authors reported classifying concussions solely according to the persistence of abnormalities in mental status, with no mention of noncognitive postconcussion symptoms, which likely would have affected the classifications. Beyond the examination of signs and symptoms of concussion, the AAN and Cantu guidelines recommend consideration of an athlete s history of concussion in the decision of whether he or she should return to play. Although research findings suggest that those with a history of concussion do not necessarily suffer greater neurocognitive impairment following a new head injury, 14,17,26 these athletes have been found to be at increased risk for subsequent concussion 7,13,14,17,32 and to be more likely to experience persistent postconcussion symptoms. 27,29 Furthermore, athletes with more than one previous concussion were found to report significantly more concussion-related symptoms on a self-report scale following a new concussion than athletes with only one 30 or no previous concussion. 22 The goal of the present study was to assess prospectively athletes who sustained concussions from an initial injury until resolution of symptoms to establish an empirical basis for classifying concussion severity. As suggested by Guskiewicz and colleagues 17 we defined the severity of the injury according to the number and duration of symptoms. We also chose variables for consideration that had been established by previous researchers. Thus, in keeping with McCrea, et al., 28 we considered LOC and PA. Similarly, we took into account a significant decrease in cognitive function, because this was identified by Collins, et al., 6 as possibly indicative of concussion severity. Finally, in keeping with Mrazik and colleagues 30 and with Moser and Schatz, 29 we included history of concussion in our study. TABLE 2 Interaction of concussion severity and concussion history in return-to-play guidelines Classification of Concussion 1st Concussion 2nd Concussion 3rd Concussion Grade I AAN return to play when return to play when return to play when asymptomatic for 15 mins asymptomatic for 1 wk asymptomatic for 1 wk Cantu return to play when return to play in 2 wks terminate season; may asymptomatic when asymptomatic return next season for 1 wk (consider terminating season) Grade II AAN return to play when return to play after return to play after asymptomatic for 1 wk 2 wks if asymptomatic 2 wks if asymptomatic Cantu return to play when return to play after 1 terminate season; may asymptomatic for 1 wk mo if asymptomatic return next season for 1 wk (consider terminating season) Grade III AAN return to play when return to play when return to play when asympasymptomatic for 1 2 wks asymptomatic for 1 mo tomatic for 1 mo Cantu return to play 1 mo terminate season after injury if asymptomatic for 2 wks 478 J. Neurosurg. / Volume 98 / March, 2003

3 Symptom-based return-to-play decision making Clinical Material and Methods Measures Used in the Study The CRI (Concussion Resolution Index; HeadMinder, Inc., New York, NY) 10 is a brief Internet-based neurocognitive assessment battery. The tests can be administered in 20 to 25 minutes by using any computer with an Internet connection. Statistical analyses of injured athletes test performances are adjusted for test retest reliability and for practice effects, 12 and alternate forms are provided automatically. Reports are generated immediately following completion of the posttrauma test. The CRI has been shown to be sensitive to cognitive symptoms evident postconcussion and to be useful in monitoring resolution of decrements in reaction time relative to preseason baseline measurements. 12 Six cognitive tests are administered at baseline and again at each posttrauma evaluation. The results of these six tests constitute three speed test indices and two error scores. Animal Decoding is a test in which athletes are instructed to type numbers keyed to pictures of animals, and Symbol Scanning is a test in which athletes are instructed to determine rapidly whether identified sets of symbols are present among a set of distractors; scores obtained on these two tests comprise the processing speed index. When Reaction Time is tested, athletes press the spacebar as a target shape appears on the screen, and when performing Cued Reaction Time, athletes press the spacebar as a target shape appears immediately after a cue shape; these comprise the simple reaction time index. An error index is also calculated based on total false-positive and false-negative answers on these two tests. Visual Recognition 1 and Visual Recognition 2 present series of pictures, some of which are repeated. Athletes are instructed to press the spacebar as quickly as possible whenever they recognize a picture from an earlier exposure and the response latency is measured. These latency measurements comprise the complex reaction time index. An error index is also calculated based on the total falsepositive and false-negative responses on these two tests. Cognitive impairment was defined as a decrease of more than SE diff (p 0.05) from baseline on one or more CRI factors at the first follow-up evaluation. Clinically, this translates into decreases of at least 106, 145, and 356 msec on the simple reaction time, complex reaction time, and processing speed indices, respectively. Therefore, athletes who exhibit a significant decrease in speed on at least one CRI index are deemed cognitively impaired. At baseline administration of the CRI, the tests are preceded by a short questionnaire designed to gather demographic information, history of concussion, and other pertinent medical information that is useful in return-to-play decision making. Following trauma, the cognitive tests are preceded by questions about the injury (LOC and confusion or disorientation), the answers to which are supplied by the athletic trainer or team physician who witnessed the injury. Information concerning the presence and intensity of postconcussion symptoms was garnered from the athlete s self-report. All symptoms were surveyed at each postconcussion assessment by using a symptom questionnaire. Symptoms were recorded as either absent, mild, moderate, or severe. For purposes of data analysis in the present study, these scales were truncated to the dichotomy of present or absent. J. Neurosurg. / Volume 98 / March, 2003 TABLE 3 Demographics of the concussion sample Factor No. of Injured Athletes (%) total no. of injured athletes 47 (100) sex male 27 (57.4) female 20 (42.6) age (yrs) (21.3) (27.7) (31.9) (17.1) 22 1 (2.1) education high school 23 (48.9) college 24 (51.1) sport soccer 14 (29.8) football 18 (38.3) wrestling 3 (6.4) field hockey 3 (6.4) basketball 2 (4.3) ice hockey 1 (2.1) other 6 (12.8) no. of past concussions 0 15 (31.9) 1 12 (25.5) 2 9 (19.1) 3 7 (14.9) 4 1 (2.1) 5 1 (2.1) 6 1 (2.1) 8 1 (2.1) Participants and Procedures Baseline CRI assessments were administered in groups to 1603 athletes in computer laboratories by using Internetenabled desktop computers at nine US high schools, colleges, and sports organizations as part of an ongoing research project initiated in the spring of The majority of athletes were engaged in high-risk sports such as football and hockey. Approval was obtained from the Institutional Review Boards of seven institutions; two institutions did not require such approval. Informed consent was obtained from all adult athletes and parental permission was obtained for all athletes younger than 18 years of age. Following a concussion, athletes were administered follow-up tests according to the clinical judgment of the athletic trainer, team physician, or psychologist in charge of their care, typically at 1- to 2-day intervals, until all symptoms had resolved. These follow-up tests were performed in a clinical office environment. Results Forty-seven athletes sustained concussions and were followed up until all the symptoms had resolved. Of these, 27 athletes were male (57%) and 20 were female (43%). The mean age of the athletes was years ( SD) with a range from 14 to 22 years. Thirty-two (68%) of the athletes had a history notable for previous concussions (Table 3). All 47 injured athletes were seen for an initial follow-up 479

4 D. Erlanger, et al. TABLE 4 Mean scores and effect sizes for the CRI speed indices Mean Decrease Effect Size Speed Index SD (msec) Range SE diff (z Score) simple reaction time to complex reaction time to processing speed to assessment (mean time days [ SD]. Athletes who displayed no symptoms at any given follow-up examination were not evaluated further. Subsequently, 33 athletes were administered a second follow-up evaluation, 16 a third, eight a fourth, two a fifth, and one athlete a sixth follow-up test. All postconcussion symptoms had resolved in all participants by Day 16 postinjury. A majority (55.3%) of athletes performed significantly slower on at least one of the CRI speed or error indices. The mean decreases for the CRI speed indices and effect sizes, expressed as the mean SE diff are provided in Table 4. In keeping with Guskiewicz, et al., 17 we considered larger numbers and longer durations of symptoms to be useful indicators of concussion severity. The frequency of symptoms observed at the sideline of the playing field on the day of injury and at the initial follow-up evaluation are listed in Table 5. In the immediate aftermath of concussion there was a mean of ( SD) and a range of one to seven symptoms. At the initial postconcussion follow-up examination, there was a mean of ( SD) and a range of 0 to 11 symptoms. Overall, the mean duration of symptoms was days ( SD) with a range of 0 to 16 days. Data were unavailable for three athletes. To determine which symptoms were the strongest predictors of concussion severity, a two-step procedure was implemented. First, correlations between each symptom observed at the sideline on the day of injury (that is, at the sideline evaluation) and the total number of symptoms identified at that time were computed. This measurement assumes a positive relationship between the severity and number of symptoms. Evidence of an LOC, dizziness, nausea, or headache at the sideline evaluation were all significantly correlated with the total number of symptoms reported. Second, an LOC, dizziness, nausea, and headache at the sideline evaluation were entered into a general linear model, with the total number of symptoms apparent at the same evaluation as the dependent variable. This measure allows a representative weighting of each symptom s contribution to the total number. The overall model was significant (F = , adjusted r 2 = 0.564), and all four variables remained significant predictors. Table 6 contains the F statistics, probability values, and the weights for the four variables, as well as the variable means and SDs. History of concussion was not associated with the number of symptoms at the sideline assessment. Using similar statistical procedures, evidence of cognitive impairment, memory problems, dizziness, nausea, and headache at the first follow-up examination were all identified as significantly associated with the number of symptoms identified at that time. The number of symptoms at the first follow-up evaluation was not normally distributed and, therefore, a square-root transformation was performed. TABLE 5 Frequency of symptoms observed at the sideline evaluation and observed or reported at the initial follow-up examination Symptoms of Injured Athletes Percentage sideline evaluation headache 93.6 dizziness 85.1 confusion/disorientation 83.0 nausea 53.2 LOC 25.5 retrograde amnesia 13.0 vomiting 4.3 initial follow-up examination headache 57.4 cognitive impairment 55.3 fatigue 44.0 memory problems 37.2 nausea 31.9 concentration problems 29.8 dizziness 23.4 weakness 17.0 irritability 17.0 impaired vision 14.9 sleep problems 14.9 sensitivity to light 12.8 depression 10.6 nervousness 8.5 vomiting 0.0 other sensory problems 0.0 Subsequently, the factors noted previously were entered into a general linear model with the transformed total number of symptoms at the first follow-up evaluation as the dependent variable. The overall model was significant (F = 4.59, adjusted r 2 = 0.281); however, only memory problems and headache at the first follow up proved to be significant predictors. Table 7 contains the F statistics, probability values, and weights for the five variables as well as the untransformed variable means and SDs. Loss of consciousness and history of concussion were not associated with the number of symptoms at the first follow up. The role of individual symptoms in predicting the total duration of symptoms was next assessed. Cognitive impairment (as defined by CRI scores), dizziness at the sideline evaluation, and both symptomatic memory problems and irritability at the first follow-up assessment correlated significantly with the duration of symptoms. These factors were entered into a general linear model with duration of symptoms as the dependent variable. The overall model was significant (F = 9.768, adjusted r 2 = 0.449); however, only cognitive impairment and symptomatic memory problems remained significant. Table 8 contains the F statistics, probability values, and weights for the four variables, along with the variable means and SDs. Loss of consciousness and history of concussion were not associated with the overall duration of symptoms. Evidence supporting the validity of athletes self-reports of memory dysfunction was identified using two methods. First, a one-way analysis of variance was performed to compare the mean decreases (from baseline to the initial postconcussion evaluation) on the CRI memory factor of athletes who reported memory problems with those who reported no problems with memory. The former had significantly larger decreases in scores than the latter 480 J. Neurosurg. / Volume 98 / March, 2003

5 Symptom-based return-to-play decision making TABLE 6 Type of symptom by total number of symptoms observed at the sideline evaluation Symptom at Sideline Evaluation No. of Athletes Mean SD Weight F Statistic p Value dizziness present absent LOC present absent nausea present absent headache present absent (mean 2.64 SE diff compared with 0.88 SE diff, F = 5.86, p = 0.02). Second, a chi-square analysis showed a significant association between reports of memory problems and objective evidence of memory dysfunction on the CRI memory factor ( 2 = 4.36; p = 0.036, one-tailed Fisher exact test). Discussion In this prospective study, we examined the relationship of concussion symptoms and objective neurocognitive measurement to three potential indicators of concussion severity: 1) number of symptoms observed at the sideline evaluation; 2) number of symptoms noted at the first follow-up assessment; and 3) duration of symptoms. In keeping with the findings of McCrea and colleagues, 28 we found that LOC was associated with initial estimates of concussion severity, as indicated by the number of symptoms observed and reported in the immediate aftermath of injury. Loss of consciousness, however, was not associated either with the total number of symptoms at the initial follow-up evaluation or with the overall duration of symptoms. Moreover, the presence of nausea and dizziness also were significant indicators of the severity of the injury in the immediate aftermath of concussion. Therefore, although LOC, especially when prolonged, clearly should not be ignored for the sideline assessment of concussion severity, 5,23 evidence of its usefulness in establishing return-to-play guidelines was weak, a finding that agrees with those of previous studies in which insignificant correlations of LOC with indicators of concussion severity were found. 8,25 In contrast, an athlete s self-report of memory problems at an initial follow-up assessment predicted the number of additional symptoms observed at the same time and the overall duration of symptoms. Furthermore, we found that these self reports were significantly associated with objective evidence of anterograde memory dysfunction, as indicated by a decreased performance on the CRI memory factor. This finding at an average of 2 days following injury foreshadows the later significant relationship between symptoms of memory dysfunction and objective measures of decline 1 year postinjury. 2 Finally, we found that cognitive impairment at the initial follow-up evaluation, relative to a baseline measurement, is a significant predictor of the duration of postconcussion symptoms. This objective predictor of the duration of symptoms should prove useful in conjunction with the consideration of the athlete s self-report, the athlete s medical history, and observations of a trainer familiar with the athlete for formulating recommendations for the returnto-play schedule. Furthermore, this is an important finding TABLE 7 Type of symptom by total number of symptoms at the first follow-up examination Symptom at 1st Follow Up No. of Athletes Mean SD Weight F Statistic p Value cognitive impairment present absent memory problems present absent dizziness present absent nausea present absent headache present absent J. Neurosurg. / Volume 98 / March,

6 D. Erlanger, et al. TABLE 8 Type of symptom by duration of symptoms Symptom No. of Athletes Mean SD Weight F Statistic p Value cognitive impairment present absent memory problems at 1st follow-up examination present absent dizziness at sideline present absent irritability at 1st follow-up examination present absent because it provides an objective predictor of the presence of postconcussion symptoms, which may be particularly useful in situations in which an athlete is motivated to underreport subjective symptoms to hasten the return to competition. An LOC associated with a mild concussion is not caused by activity in the cerebral cortex, but rather to brainstem nonpathological activation of an endogenous pontine cholinergic inhibitory system. Thus, LOC is not directly related to postconcussion cognitive symptoms such as memory dysfunction. 19 These cognitive symptoms are thought to be related to subsequent changes in cerebral cortex glucose metabolism, unrelated to the brainstem, with prominent concentrations of activity in the mesial temporal and frontal lobes. In animal models of concussion, these changes generally resolve within 10 days. 20 This dual-mechanism model helps explain the finding of McCrea and colleagues 28 that there are more severe initial alterations in mental status among athletes experiencing LOC. That is, the initial deficits in mental status would be due to sequelae of transient changes in the brainstem s cholinergic system, whereas persisting postconcussion symptoms would be due to developing changes in cerebral glucose metabolism. This model might also help explain findings of the study of Echemendia and colleagues, 9 in which athletes performed better on cognitive tests of verbal memory administered 2 hours postconcussion than on tests administered 24 hours postinjury. By describing two different mechanisms that account for LOC and developing neurocognitive symptoms, this model also helps explain our findings that memory dysfunction TABLE 9 Comparison of symptom frequencies across two studies at three points in time Barth, et al. Present Study Barth, et Days Symptom (% of patients) (% of patients) (% of patients) headache memory problems nausea dizziness weakness at 24 hours, but not a brief LOC, predicts a more severe sports-related concussion if severity is defined in terms of duration of both self-reported postconcussion symptoms and positive findings on objective neurocognitive tests of memory. We did not find a significant relationship between history of concussion and the initial presentation of symptoms or the persistence of postconcussion symptoms. Previous studies have presented contrasting findings in this regard. Our results are consistent with the research conducted by Macciocchi and colleagues 26 and by Webbe and Ochs, 31 but not with that of Matser, et al., 27 or Moser and Schatz. 29 It is possible that some of the differences among these studies may be explained by the proximity and severity of prior concussions, which were not variables under consideration in the present study. In light of these conflicting findings, we recommend a continued consideration of the history of concussion in return-to-play decision making until findings from larger studies with longer follow-up periods and more detailed information regarding previous concussions are available. Applying our findings to sideline concussion grading guidelines for return-to-play decisions is somewhat problematic because the critical datum of memory dysfunction (presumably based on the cortical glucose mechanism) cannot be established at the time of the sideline assessment. The AAN guidelines, which emphasize LOC, and the Cantu guidelines, which stress PA, other symptoms, and LOC to a lesser extent as the critical elements, appear reasonable for describing the initial severity of injury. Based on the current results, however, the AAN guidelines appear to be weak in predicting the severity of injury, as measured by the number or duration of symptoms. By emphasizing PA (however it is measured) and other symptoms apparent both at the sideline evaluation and at 24 hours postinjury, the Cantu guidelines more closely target the symptomatic, memory-based factors that we find are most predictive of the severity and duration of injury. More generally, applying terminology originally used in contexts other than sports-induced concussions appears to be problematic. The vast majority of sports-related concussions are mild, and even severe sports-related concussions may be relatively mild in comparison with concussions sustained in automobile accidents. 11 The use of the term PA in 482 J. Neurosurg. / Volume 98 / March, 2003

7 Symptom-based return-to-play decision making the Cantu guidelines, even in its broadest context, would not include the self-reported memory problems studied in the present investigation. Similarly, the more traditional term anterograde memory deficit is not typically used to describe subjective complaints of memory problems. Nevertheless, these subjective complaints were significantly related to objective findings on the CRI memory tests, indicative of an apparently mild, transient, yet functionally significant deterioration in memory efficiency. We propose that the term post concussion memory dysfunction be used to encompass both subjective complaints and objective symptoms, inclusive of PA and anterograde memory deficit. Our results suggest that self-reported memory problems and/or significant declines shown on cognitive testing, relative to baseline, should be topics of consideration in return-to-play decisions. In our sample of athletes engaged in a number of sports at various levels of competition, the denial of ongoing memory problems appeared to be a simple and useful method for ruling out the presence of ongoing postconcussion memory dysfunction. This may not be the case in other settings, in which the pressure the athlete experiences for an early return to play may tempt him or her to underreport subjective symptoms (National Athletic Trainers Association Research and Education Foundation, unpublished data). When such underreporting is suspected, and there is no objective evidence of sequelae (for example, receiving a low score on a mental status test or brief neurological examination) or reports of symptoms of postconcussion syndrome, findings on objective neurocognitive measures of memory as well as reaction time and processing speed could provide valuable information for determining the athlete s readiness for return to competition. Alternative, noncognitive measures such as the athlete s performance on balance testing 15,16 also could be used to support these decisions. Our investigation was based on a sample of 47 high school and college athletes from a variety of contact sports. Some caution should be exerted in generalizing these results to other settings. For example, our rate of LOC was higher (25.5%) than those of many other studies. 6,17,28 This may be due to the inclusion of high school athletes, in whom LOC has been noted to occur somewhat more frequently than in college athletes. 17 In contrast, our sample appeared similar in the frequency of reported symptoms to that of an earlier study 3 (Table 9), suggesting that our athlete s injuries were similar in postconcussion symptom patterns, even if they differed in initial injury characteristics. Conclusions Further prospective investigations should be conducted on larger numbers of athletes. In addition, other potential parameters for determining concussion severity should be investigated, including the type, location, and severity of the blow, and the time interval between the current and most recent trauma, if the athlete has a history of concussion. Any of these could be useful indicators of the severity of injury. If so, they would provide the responsible health professional with a more complete model for describing concussion severity and predicting the course of recovery. J. Neurosurg. / Volume 98 / March, 2003 Disclosure David Erlanger is the president of HeadMinder, Inc., and Tanya Kaushik is an employee of the company. HeadMinder, Inc., is the publisher of the CRI. References 1. Anonymous: Practice parameter: the management of concussion in sports (summary statement). Report of the Quality Standards Subcommittee. Neurology 48: , Arcia E, Gualtieri CT: Association between patient report of symptoms after mild head injury and neurobehavioral performance. Brain Inj 7: , Barth JT, Alves WM, Ryan TV, et al: Mild head injuries in sports: neuropsychological sequelae and recovery of function, in Levin HS, Eisenberg HM, Benton AL (eds): Mild Head Injury. New York: Oxford University Press, 1989, pp Cantu RC: Guidelines for return to contact sports after a cerebral concussion. Physician Sportsmed 14:76 79, Cantu RC: Posttraumatic retrograde and anterograde amnesia: pathophysiology and implications in grading and safe return to play. J Athl Train 36: , Collins MW, Grindel SH, Lovell MR, et al: Relationship between concussion and neuropsychological performance in college football players. JAMA 282: , Delaney JS, Lacroix VJ, Gagne C, et al: Concussions among university football and soccer players: a pilot study. Clin J Sport Med 11: , Echemendia RJ, Cantu R: Return to play following cerebral concussion, in Barth J, Collins M, Lovell MR, et al (eds): Sports Neuropsychology. The Netherlands: Swets-Zeitlinger (In press) 9. Echemendia RJ, Putukian M, Mackin RS, et al: Neuropsychological test performance prior to and following sports-related mild traumatic brain injury. Clin J Sport Med 11:23 31, Erlanger DM, Feldman DJ, Kutner KC: HeadMinder: Concussion Resolution Index (CRI): Professional Manual. New York: HeadMinder, Erlanger DM, Kutner KC, Barth JT, et al: Neuropsychology of sports-related head injury: Dementia Pugilistica to Post Concussion Syndrome. Clin Neuropsychol 13: , Erlanger DM, Saliba E, Barth JT, et al: Monitoring resolution of postconcussion symptoms in athletes: preliminary results of a web-based neuropsychological test protocol. J Athl Train 36: , Gerberich SG, Priest JD, Boen JR, et al: Concussion incidences and severity in secondary school varsity football players. Am J Public Health 73: , Guskiewicz KM, Marshall SW, Broglio SP, et al: No evidence of impaired neurocognitive performance in collegiate soccer players. Am J Sports Med 30: , Guskiewicz KM, Perrin DH, Gansneder BM: Effect of mild head injury on postural stability in athletes. J Athl Train 31: , Guskiewicz KM, Riemann BL, Perrin DH, et al: Alternative approaches to the assessment of mild head injury in athletes. Med Sci Sports Exerc 29 (Suppl 7):S213 S221, Guskiewicz KM, Weaver NL, Padua DA, et al: Epidemiology of concussion in collegiate and high school football players. Am J Sports Med 28: , Harmon KG: Assessment and management of concussion in sports. Am Fam Physician 60: , Hayes RL, Lyeth BG, Jenkins LW: Neurochemical mechanisms of mild and moderate head injury: implications for treatment, in Levin HS, Eisenberg HM, Benton AL (eds): Mild Head Injury. New York: Oxford University Press, 1989, pp Hovda DA, Prins M, Becker DP, et al: Neurobiology of concussion, in Bailes JE, Lovell MR, Maroon JC (eds): Sports-Related 483

8 D. Erlanger, et al. Concussion. St. Louis: Quality Medical Publishing, 1998, pp Johnston KM, McCrory P, Mohtadi NG, et al: Evidence-based review of sport-related concussion: clinical science. Clin J Sport Med 11: , Jordan SE, Green GA, Galanty HL, et al: Acute and chronic brain injury in United States National Team soccer players. Am J Sports Med 24: , Kelly JP: Loss of consciousness: pathophysiology and implications in grading and safe return to play. J Athl Train 36: , Leclerc S, Lassonde M, Delaney JS, et al: Recommendations for grading of concussion in athletes. Sports Med 31: , Lovell MR, Iverson GL, Collins MW, et al: Does loss of consciousness predict neuropsychological decrements after concussion? Clin J Sport Med 9: , Macciocchi SN, Barth JT, Alves W, et al: Multiple mild head injury in collegiate athletes: clinical and methodological implications. Arch Clin Neuropsychol 14:781, 1999 (Abstract) 27. Matser JT, Kessels AGH, Jordan BD, et al: Chronic traumatic brain injury in professional soccer players. Neurology 51: , McCrea M, Kelly J, Randolph C: Standardized Assessment of Concussion: SAC: Manual for Administration, Scoring, and Interpretation, ed 2. Waukesha: CNS, Moser RS, Schatz P: Enduring effects of concussion in youth athletes. Arch Clin Neuropsychol 17:91 100, Mrazik M, Kral M, Hynd GW, et al: Neuropsychological assessment of collegiate football players. Arch Clin Neuropsychol 14: 781, 1999 (Abstract) 31. Webbe FM, Ochs SR: Recency interacts with frequency of soccer heading to predict weaker neuro-cognitive performance. Appl Neuropsychol (In press) 32. Zemper ED: Analysis of cerebral concussion frequency with the most commonly used models of football helmets. J Athl Train 29:44 50, 1994 Manuscript received June 11, Accepted in final form November 12, Address reprint requests to: David Erlanger, Ph.D., 3 East 65th Street, Suite 5B, New York, New York david@ headminder.com. 484 J. Neurosurg. / Volume 98 / March, 2003