Severity of sports-related concussion and neuropsychological test performance

Method.

We recruited 175 athletes from two national rugby league clubs followed over three consecutive seasons. Twenty-nine athletes were concussed; however, eight were excluded from analysis. Of those athletes excluded, four had noncomparable preinjury test exposure, having previously acted as controls; three had follow-up outside desired intervals; and one player voluntarily declined to participate. All but three of the remaining 21 concussed athletes had a previous history of concussion. Concussed and control athletes were equivalent in terms of age, education, number of seasons played, number of previous concussions, and preseason psychometric performance.

We assessed three cognitive measures of psychomotor speed and speed of information processing, including the Speed of Comprehension test, the Digit Symbol test, and the Symbol Digit test.6 Nationally accredited sports trainers completed injury reports for each concussion. Concussions were diagnosed and graded by sports medicine physicians using standardized clinical interviews and procedures. Diagnoses were made in accordance with a broad definition involving any change in mental status or the presence of physical signs or other symptoms.2 Four athletes were graded retrospectively as the injury was not witnessed. Athletes were retested at 2 and 10 days following trauma.

Results.

The Reliable Change Index with adjustment for practice allows the clinician to detect a significant change in the individual. Correction for practice ensures that the effect of concussion is not masked by natural improvement.7 An athlete’s retest score is compared to a preinjury score adjusted by the average change of a control group. The difference is then divided by an error term accounting for test reliability and variability, resulting in a standardized (Z) score. Cognitive impairment was defined as a decline from preseason to post-trauma performance on any one test (Z < −1.645).

Table 2 presents the number of athletes displaying deteriorated performance from preseason to 2 days and 10 days post-trauma. Test sensitivity at 2 days was 86%; specificity was 81%. The percentage of athletes showing deteriorated performance 2 days after trauma was similar across grades, irrespective of system. However, when retested 10 days following trauma, more athletes with AAN and Cantu grade 1 concussions were impaired compared to those with grade 2 concussions. Under the Colorado guidelines a comparable number of athletes with grade 1 and grade 2 concussions were impaired at 10 days. No athletes with grade 3 concussion (LOC) under the AAN and Colorado systems recorded impairment at 10 days postinjury, whereas many athletes with grade 2 concussions did. The three athletes with an AAN grade 1 concussion who demonstrated impairment 2 days post-trauma also reported delayed symptoms following the game. These athletes were still classified with AAN grade 1 concussions as their mental status was clear and they were symptom-free within 15 minutes. The three Colorado grade 1 concussed athletes all demonstrated deteriorated performance 2 days after trauma and also reported symptoms lasting from several hours to days. The presence of cognitive impairment did not appear to be related to the duration of symptoms or a history of concussion in the previous year.

Discussion.

No relationship was observed between concussion severity grade and cognitive impairment for the systems investigated. Cognitive deficits existed within 2 days of concussion regardless of injury severity. The percentage of athletes impaired at 10 days post-trauma was consistent across severity classifications. The only salient injury predictor of impairment at 2 days was PTA, yet neither its presence nor duration was related to duration of impairment. Athletes with LOC did not appear more likely to demonstrate impairment. Furthermore, athletes with no mental status aberration were impaired as late as 10 days post-trauma.

The concussion guidelines investigated place a considerable emphasis on LOC, particularly the Colorado and AAN systems, where any LOC is a severe concussion. The current findings suggest that a brief LOC in contact sport does not necessarily lead to prolonged cognitive impairment.8 Nonetheless, an athlete with LOC should not participate further that day and should be monitored carefully for more emergent intracranial problems. Yet once serious complications can be ruled out, the decision to arbitrarily withhold athletes from participation may be unwarranted. A larger sample, particularly with more cases of brief LOC, would help to clarify this point.

Six athletes with grade 1 concussions with PTA less than 1 minute demonstrated cognitive impairment. Three athletes, reporting that their symptoms had cleared within 15 minutes (AAN grade 1), also demonstrated cognitive impairment at 10 days post-trauma. Present guidelines would permit these impaired athletes to return to participation. The effects of the mildest concussions may be underestimated. The risks associated with participating while cognitively impaired are not known. These athletes may have returned prematurely and further exposure to subconcussive trauma may partly explain the prolonged impairment. As most athletes had a history of multiple concussions this may also partially explain the impairment in otherwise minor concussions. The impact of repeated sports-related head trauma is clearly in need of further research. Moreover, a better understanding of the risk factors for second impact syndrome will determine the utility of severity gradings and cognitive impairment criteria for safe return to play.

The use of psychometric testing in the management of concussion must fit into a more comprehensive clinical assessment. Classifications of severity may assist in the emergent management of concussion, but may not be useful in determining readiness to return to participation. The clinician must consider whether deteriorated psychometric performance alone is sufficient reason for recommending abstinence from sport. Further research is required to develop and validate concussion management guidelines.

The implications of this study must be qualified as neuropsychological tests are indirect measures of brain function and integrity, yet more sensitive and reliable technology is not readily available. It is also possible that severity grades were related to other brain functions. Although impaired speed of information processing is the most reliable mental effect of concussion, a broader battery would help clarify the nature and predictability of cognitive impairment.