Consequences of Repetitive Head Impacts and Multiple Concussions (original) (raw)

In recent years there has been an increase in research on the cognitive and neuropathological consequences of repetitive head impacts and multiple concussions in athletes. Given the frequency of head impacts in contact sports, the public health implications of these consequences may be significant. This chapter addresses those elements of the committee's statement of task that concern the effects of “subconcussive” head impacts (i.e., head impacts that do not result in symptoms consistent with a diagnosis of concussion) and multiple concussions. The chapter reviews the clinical manifestations, neuroimaging features, risk factors, and animal studies related to repetitive head impacts and multiple concussions. It also discusses the possible long-term neuropathological consequences associated with repetitive head impacts and multiple concussions, including chronic traumatic encephalopathy (CTE), an emerging diagnostic entity associated with retired athletes with a history of head injury as well with as military personnel exposed to repeated brain injury from blast and other causes. The goals of this chapter are to provide a comprehensive review of the current literature, to clarify controversies, and to point out important directions for future research.

NEUROPSYCHOLOGICAL AND NEUROPHYSIOLOGICAL CONSEQUENCES

Studies of Repetitive Head Impacts

As with much of the clinical literature on the consequences of concussions in sports, the generalizability of many studies of the effects of repetitive head impacts is limited by methodological weaknesses. For example, helmet-based head impact recording devices are typically set to record only impact forces over a minimum threshold (e.g., 10 g of linear acceleration; see Duma et al., 2005) and, therefore, do not record all impacts to the head. Although recent advances in technical, statistical, and clinical knowledge have helped to improve research on repetitive head impacts, earlier findings have to be viewed in the context of history: Their importance lies more in their groundbreaking attempts to quantify relevant variables and not necessarily in their specific findings.

Findings from Soccer Studies

In soccer, athletes experience repetitive head impacts from using their heads to strike the ball for passing and shooting. Older research involving amateur and professional soccer players indicated an association between cumulative heading and neuropsychological impairments (see, for example, Matser et al., 1998, 1999, 2001; Sortland and Tysvaer, 1989; Tysvaer and Lochen, 1991). One study of 37 former professional soccer players found mild to severe deficits in the areas of attention, concentration, memory, and judgment in 81 percent of the players. The authors speculated that this finding could be indicative of permanent organic brain damage resulting from repeated traumas from heading the ball (Tysvaer and Lochen, 1991). In another study involving 53 active professional soccer players, impairments in memory, planning, and visuo-perceptual tasks were observed and compared with those in non-contact-sport athlete controls. Among the soccer players, performance on these tasks was inversely related to the frequency of heading the ball (Matser et al., 1998). Computed tomography scans of 33 former professional soccer players identified central brain atrophy in one-third of study participants, although scans were only visually inspected, and there were no baseline or control comparisons (Sortland and Tysvaer, 1989).

Several other studies, including more recent ones, involving youth soccer players have found no effect of heading on neurocognitive performance (Broglio and Guskiewicz, 2001; Guskiewicz et al., 2002; Kaminski et al., 2007, 2008; Kontos et al., 2011; Stephens et al., 2010; Straume-Naesheim et al., 2005). For example, Guskiewicz and colleagues (2002) found no differences in neurocognitive function or Scholastic Aptitude Test scores between collegiate soccer players (n=91) and groups of athletes from other contact and non-contact sports (n=96) or between the collegiate soccer players and non-athletic controls (n=53), suggesting that soccer players are not differentially affected by soccer playing (and, by extension, heading).

Furthermore, studies that have directly assessed changes in cognition related to heading a soccer ball have failed to establish any relationship between heading and neurocognitive changes. Following detailed observation of heading frequencies by 63 high school soccer players and the administration of Immediate Post-Concussion Assessment and Cognitive Testing (ImPACT), Kontos and colleagues (2011) found no differences in neurocognitive performance or symptoms among low-, moderate-, and high-exposure header groups. Similarly, Putukian and colleagues (2000) found no changes in neuropsychological test scores between pairs of collegiate soccer players who headed the ball in practice for 20 minutes compared with those who did not head in practice. Kaminski and colleagues conducted two studies with 71 (2007) and 393 (2008) female collegiate soccer players. Using heading counts as the independent variable and pre- and post-season balance and neuropsychological tests to determine neuropsychological changes, they found no significant relationships on any test measure.

Early research using magnetic resonance imaging (MRI) and cognitive tests found no significant cognitive impairments or differences on MRI scans among soccer players, boxers, and track and field athletes (Haglund and Ericksson, 1993). In a recent study using diffusion tensor imaging (DTI) scans on 12 German soccer players and 11 swimmers, several group differences in brain white matter were noted, including increased radial diffusivity and axial diffusivity in soccer players compared to swimmers (Koerte et al., 2012). Although the authors of the study suggest that heading in soccer may lead to neurophysiological changes in the brain, this study's generalizability is limited because of the small number of participants and because it did not include a baseline scan. Furthermore, it is not clear what the functional significance of such findings would be.

In summary, studies of the consequences of heading in soccer have obtained mixed results, with more recent studies showing no relationship between heading and neuropsychological impairment. The positive findings of some older studies may have been due in part to the more frequent use in the 1980s and 1990s of soccer balls that absorbed more water, increasing the weight of the ball by up to 20 percent and potentially making them more dangerous for heading (Smodlaka, 1984). Today, players use waterproof synthetic soccer balls that absorb less water (Kirkendall and Garrett, 2001). The DTI study (Koerte et al., 2012) appeared to show neurological differences in a very small sample. Due to small sample sizes and other methodological limitations, caution is required in interpretation of these studies' findings.

Findings from Football and Ice Hockey Studies

As is the case with soccer players, football and ice hockey players can incur repetitive head impacts (Brainard et al., 2012; Crisco et al., 2010, 2011, 2012). For example, a lineman in football who tackles another player with his head in successive plays experiences a series of repetitive head impacts. Hockey players may experience repetitive head impacts from collisions with the board and with other players.

McAllister and colleagues (2012) examined repetitive head impacts over a single season in collegiate football and ice hockey athletes and compared those athletes with a group of athletes who played a non-contact sport on a variety of measures. Contact athletes wore accelerometer-instrumented helmets and took pre- and post-season ImPACT tests. A subset from one of the three Division I universities also completed a paper-and-pencil neuropsychological battery and had preseason and postseason neuroimaging. There were no group differences on cognitive tasks over a sport season. The researchers also examined baseline neurocognitive tests scores across three sport seasons and found no differences on baseline assessments among the sport groups, suggesting that previous exposures to contact did not affect test scores negatively. However, the researchers did report that a higher percentage of the contact sport athletes performed worse than those in the non-contact group on a measure of new learning (California Verbal Learning Test), with no ImPACT composites showing significant change. Furthermore, the authors found that impact exposure above the 95th percentile in frequency during the last week of the season was related to poorer performance on the Trail Making test, a measure of visual attention and task switching, and that the peak linear acceleration for the season was related to slower ImPACT reaction times. A relationship among recent biomechanical exposures, brain white matter integrity, and lower scores was also found, although the absolute value of (significant) test score decline did not reach impairment level.

Other studies of the effects of repetitive head impacts in high school and collegiate football players have found no association with neurocognitive impairment or physiological changes (see, for example, Broglio et al., 2011; Gysland et al., 2012; Miller et al., 2007). However, these studies all have methodological weaknesses. Only McAllister and colleagues (2012) used non-contact controls and adjusted neurocognitive test scores for practice effects, baseline levels, regression to the mean, and relevant demographic factors, while also comparing seasonal exposure and recent exposures across biomechanical measurements of magnitude and frequency.

A few small studies of high school and collegiate football and hockey players have looked at DTI, neurocognitive test scores, and biomechanical data; these have found axonal changes but mixed neuropsychological findings (Bazarian et al., 2012; Breedlove et al., 2012; Talvage et al., 2010). Evidence for the effects of repetitive head impacts on diffuse axonal injury in humans comes largely from DTI studies that measure directionality (fractional anisotropy, or FA) and regularity (mean diffusivity, or MD) of white matter tracts. This technique showed pre-season to post-season changes in a small sample of high school athletes (n=10, 16 to 18 years old, hockey or football) relative to controls (n=5, 16 to 35 years old) following self-reported repetitive head impacts (Bazarian et al., 2012). While pre- and post-season FA and MD changes (calculated as the percentage of white matter voxels showing either a significant increase or a significant decrease) were largest in a concussed player with greater than 3 percent change, those athletes with repetitive head impacts had intermediary changes of more than 1 percent on average, while the controls had insignificant changes of less than 0.5 percent on average. These findings warrant further investigation in larger samples with same-aged comparison groups.

Traumatic brain injury may result in disruption of the blood-brain barrier (i.e., increased permeability of the brain vasculature) (Neuwelt et al., 2011). Marchi and colleagues used DTI and serum measurements of S100B (a protein secreted by cells in the central nervous system and used as a marker of blood-brain barrier disruption [Blyth et al., 2009, 2011; Marchi et al., 2004]) and S100B auto-antibodies, to evaluate whether head impacts below the threshold for a diagnosis of concussion can disrupt the blood-brain barrier. Sixty-seven college football players were enrolled. In a subset of players (n=15) for whom pre- and post-game blood samples were available, only those players with the most subconcussive head impacts based on self-report and post-game review of the game film had detectable serum levels of S100B and elevated levels of auto-antibodies against S100B. Serum S100B antibodies predicted lasting changes in mean brain white matter diffusivity in a subset of players (n=10) who had preseason and postseason and 6-month follow-up DTI scans. Post-season S100B auto-antibodies also correlated with impulse control and balance problems. Although the study sample is too small to make firm inferences, this research provides preliminary evidence that repetitive head impacts that do not result in a diagnosis of concussion may disrupt the blood-brain barrier. It is important to note that there are sources of serum S100B outside of the central nervous system (e.g., fat cells), although this does not preclude use of S100B as a biomarker of brain injury (Marchi et al., 2013).

Functional magnetic resonance imaging (fMRI) studies have begun to examine blood-oxygen-level dependent brain activity following repetitive head impacts. Talvage and colleagues (2010) prospectively followed 11 male high school football players ages 15 to 19 both preseason and postseason. A negative association was observed between the number of subconcussive repetitive impacts to the front of the head and activity in the prefrontal cortex, as indicated by blood-oxygen-level dependent signals on the fMRI during the performance of a working memory test. These players also exhibited neurocognitive deficits as measured by ImPACT scores on visual memory.

Together these studies suggest changes in cognitive function the brain following repetitive head impacts in football and hockey players. However, the types of cognitive impairment and brain changes that are observed vary by study, and the results are based on small sample sizes, which raises questions about the reliability of these studies.

Findings from Boxing Studies

Many youth continue to participate in boxing even though several medical groups have called for its discontinuation due to the incidence of brain injury (Purcell and LeBlanc, 2012). Although many youth and amateur boxers wear protective gear and follow rules that are different from those for professional boxers, a primary goal of boxing is to attack the head and face of the opponent, which often results in a concussion or more severe brain injury (Jordan, 1987). Indeed, the association of boxing and traumatic brain injury (TBI) is very well-recognized in the medical literature. The so-called punch-drunk syndrome was first recognized as early as 1928 (Wilberger and Maroon, 1989), and it is associated with personality disturbances, dysarthria, or Parkinson-like disturbances.

There is ample evidence supporting the association of boxing with chronic traumatic brain injury. Several researchers have found brain abnormalities in professional boxers (see, for example, Casson et al., 1984; Drew et al., 1986; Kaste et al., 1982; Morrison, 1986; Ross et al., 1983). Jordan and colleagues (1997) examined boxers who had a high exposure to head contact (defined as having had 12 or more professional bouts) to boxers with low exposure to head contact (defined by less than 12 professional bouts) on neurocognitive performance, symptoms, and genetic testing. The authors reported that athletes with a high exposure to head contact had lower cognitive function than did those with low head contact exposure.

The largest DTI studies to date on repetitive head impacts are those involving professional boxers (ages 20 to 52), with sample sizes ranging from 24 to 81. Across these studies, microstructural abnormalities were found as indicated by increased regional and whole brain diffusion and decreased FA in boxers relative to control subjects (Chappell et al., 2006; Zhang et al., 2003, 2006). These findings are further supported by a high-resolution MRI study of 100 boxers (85 with complete data, ages 19 to 42 years) showing a significant correlation between years of boxing and diffuse axonal injury (Orrison et al., 2009).

Together these studies suggest that boxing is associated with possible long-term cognitive decline and axonal injury. Although boxing is an extreme example of a contact sport, the neuropsychological and imaging findings from studies of boxers supplement those of athletes who play other contact sports such as football and hockey.

Studies of Multiple Concussions

As discussed in Chapter 2, there is some evidence from both animal studies and research involving humans that the brain is at increased risk while recovering from a concussion. Thus, a repeat injury while recovering from a prior concussion may occur with less force, take longer to resolve, and in rare cases lead to catastrophic results (e.g., second impact syndrome) (Bey and Ostick, 2009; Simma et al., 2013, Slobounov et al., 2007). Indeed, this is the purpose of the advice “When in doubt, sit them out” (McCrory et al., 2013). A related concern is the effect of a history of concussions on cognition and brain physiology. In a retrospective survey of 223 high school athletes, 20 percent reported a history of at least one concussion (Moser et al., 2005), suggesting that many youth sustain multiple concussions over the course of their athletic careers.

The committee reviewed 16 studies that attempted to answer various questions regarding the effects of multiple concussions. Four additional studies focused on professional and adult athletes and so were not included. Most of the 16 studies assessed the neurocognitive function and symptom load of “stable” athletes (those not currently concussed) and compared groups based on the reported histories of previous concussions. This cross-sectional approach has many limitations, as noted by Iverson and colleagues (2012). A primary difficulty is finding enough athletes with a history of three or more concussions to provide sufficient statistical power. Concussions are a relatively low-base-rate phenomenon, which means that obtaining a large enough sample of individuals who have sustained multiple concussions is particularly difficult.

High School–Age Athletes

Five studies of high school athletes compared symptom presentations, and three compared neurocognitive findings. Schatz and colleagues (2011) compared baseline symptoms of 251 high school athletes who had no reported concussions with 260 athletes who had had one concussion and 105 athletes who had had two or more. The athletes with a history of two or more concussions had significantly more cognitive problems, physical symptoms, and sleep problems at the time of pre-season baseline evaluation than those with no history of concussions. On the other hand, in a study of 867 male high school and college athletes with no (n=664), one (n=149), and two (n=54) previous concussions, Iverson and colleagues (2006) found no group differences in neuropsychological test performance or symptom reporting.

Collins and colleagues (2002) attempted to study the effect of previous concussions in high school athletes by characterizing the on-field signs and symptoms of a subsequent concussion. Sixty athletes with no concussion history were compared with 28 athletes who had had three or more concussions. Those with a history of three or more concussions were significantly more likely to suffer loss of consciousness, anterograde amnesia, and mental status changes lasting longer than 5 minutes.

Among studies that examined neuropsychological test scores, two had conflicting results (Iverson et al., 2006; Moser et al., 2005), and two had too few subjects to have confidence in the results (Elbin et al., 2012; Moser and Schatz, 2002). Moser and colleagues (2005) compared 82 athletes with no concussions to 56 with one concussion, 45 with two or more (although they had had no injuries for at least 6 months), and 40 recently concussed athletes. Athletes were compared on neuropsychological test scores (Repeatable Battery for the Assessment of Neuropsychological Status, Trail Making). While there was a significant difference between groups, the only post hoc result reported was that those with two or more concussions were not different from the recently concussed. Inspection of the data indicates that scores of the recently concussed athletes were lower than those of the no- and one-concussion groups. In contrast, Iverson and colleagues (2006) compared baseline ImPACT scores for 664 athletes with no history of concussion, 149 with a history of one previous concussion, and 54 with a history of two. After controlling for education level, no significant differences were found between the groups.

Overall the findings from studies of the effects of multiple concussions on high school athletes are mixed. In addition to the number of concussions, the interval between concussions may be an important recovery-related factor to consider.

College-Age Athletes

Ten studies of college-age athletes were reviewed. Four studies used symptom presentation as the dependent variable (Collins et al., 1999; Covassin et al., 2008; Guskiewicz et al., 2003; Iverson et al., 2012). Collins and colleagues (1999) found increased symptoms among those athletes with more concussions at baseline, while Guskiewicz assessed the time to symptom resolution of a current concussion based on the number of previous concussions. Neither Covassin and colleagues (2008) nor Iverson and colleagues (2012) found a relationship between symptom levels at baseline testing and the number of previous concussions. However, among those studies that looked at neuropsychological test scores, four found significant differences between previously concussed and non-concussed athletes (Collins et al., 1999; Covassin et al., 2008, 2010; De Beaumont et al., 2009), while four failed to find differences (Broglio et al., 2006; Bruce and Echemendia, 2009; Guskiewicz et al., 2002; Iverson et al., 2012). Three studies had too few cases to be considered (De Beaumont et al., 2007; Elbin et al., 2012; Killam et al., 2005).

In 1999, Collins and colleagues compared baseline symptom totals among 179 athletes with no concussion history, 129 with one, and 78 with a history of two or more. Significant differences were found for symptoms. The researchers also noted that baseline symptom scores increased with the frequency of previous concussions. In a large sample of 184 college football players who suffered concussions, Guskiewicz and colleagues (2003) found that 30 percent of the athletes with more than three concussions took longer than 1 week for symptoms to resolve, compared to only 15 percent of those with one concussion taking more than a week to resolve.

Collins and colleagues (1999) also compared baseline neuropsychological test scores. Significant differences were found on two tests of processing speed. However, the average neuropsychological test scores for the group with two or more concussions were within normal limits. In 2002, Guskiewicz and colleagues compared collegiate soccer players with histories of concussion to soccer players, other athletes, and non-athletes with no histories of concussion. He found no significant differences at baseline on any of a battery of neuropsychological tests. Broglio and colleagues (2006) compared baseline test scores of 163 athletes with no history of concussions to 43 athletes with one previous concussion, 18 with two, and 11 with three. The researchers administered both the Concussion Resolution Index (CRI), an Internet-based neurocognitive test that assesses “simple” and “complex reaction time and process speed, and ImPACT and found no differences in any scores. The small number of athletes with two or more concussions limited the study.

Covassin and colleagues (2010) found no differences on any ImPACT composite test between athletes with no previous concussions and those with one previous concussion. However, both the two-concussion (n=50) and three-plus concussion (n=48) groups scored significantly lower on the verbal memory test than did the no-concussion group (n=50); furthermore, visual memory scores were significantly lower for those with a history of three or more concussions (n=48) than those with no concussion history. The authors concluded that they had demonstrated a partial dose-response relationship for those tests.

In a prospective study Guskiewicz and colleagues (2003) followed a large sample of collegiate football players over 3 years; of those athletes, 184 experienced concussions during the study. The researchers reported that athletes with a history of three or more previous concussions had three times the risk of getting a subsequent concussion than those with no previous history.

In a study of the effects of concussion history on recovery after a subsequent concussion, Covassin and colleagues (2008) compared post-injury neurocognitive test scores 1 day out and 5 days out from the injury between 56 athletes with no previous concussion and 21 with a history of two or more concussions. While they found significant differences in all four composite scores at day 1 and in two composite scores at day 5 (verbal memory and reaction time), they did not control for baseline scores or compare the number of significantly changed scores (CRI) between groups. There were no symptom level differences at any point.

Methodological differences make comparisons among these studies difficult. Differences in how concussion history is measured (e.g., two and more, three and more, etc.) may obscure an important threshold. Variable controls, small sample sizes, and cross-sectional designs make generalizations difficult. Group-level studies are mixed in terms of results, and better methodologies are needed to identify the timing and cumulative effects of multiple concussions. Longitudinal studies and studies highlighting individual differences are lacking. Collaborative studies may be needed to accumulate larger samples of multiply concussed athletes.

MULTIPLE CONCUSSIONS AND DEPRESSION AND SUICIDE

Surveys of retired professional athletes provide some evidence that a history of multiple concussions increases risk for depression (Didehbani et al., 2013; Guskiewicz et al., 2007; Kerr et al., 2012). In a survey of more than 2,500 retired professional football players, 269 of the respondents (11.1 percent) reported having had a prior or current diagnosis of clinical depression. After controlling for parameters such as age, number of years since retirement, number of years played, physical condition, and diagnosed comorbidities such as osteoarthritis, coronary heart disease, stroke, cancer, and diabetes (but not substance abuse or intervening psychosocial issues), the authors found an increasing linear relationship between history of concussion and diagnosis of lifetime depression (p < 0.005). Compared with retired players with no history of concussion, retired players reporting three or more previous concussions (24.4 percent) were three times more likely to have been diagnosed with depression; those with a history of one or two previous concussions (36.3 percent) were 1.5 times more likely to have been diagnosed with depression (Guskiewicz et al., 2007). In another study of 30 retired professional football players with a history of concussion versus 29 age- and IQ-matched controls without a history of concussion, a significant correlation was observed between number of lifetime concussions and current cognitive symptoms of depression as measured by the Beck Depression Inventory II. These findings suggest that the number of concussions an individual has sustained may be related to later depressive symptomology (Didehbani et al., 2013). Imaging research is beginning to explore the relationship between depression symptoms and brain white matter abnormalities in retired athletes (Hart et al., 2013; Strain et al., 2013).

Athletes who retire as a result of suffering multiple concussions may experience distress and reduced quality of life, similar to outcomes reported following other serious athletic injuries (Caron et al., 2013; Kuehl et al., 2010; Mihovilovic, 1968). Social support has been identified as important to psychological recovery following more severe brain injuries (Gan et al., 2006) as well as within the sport injury and rehabilitation process (Bianco, 2001; Clement and Shannon, 2011; Wiese-Bjornstal et al., 1998). Individuals who have sustained concussions are also likely to benefit from social support. However, little is known about the role of social support in managing athletes' concussion symptoms and related psychosocial outcomes. The effect of support may be complicated in situations where norms encourage athletes to play through their injuries and when athletes fear being stigmatized by peers as lacking toughness (Safai, 2003; Young et al., 1994). Qualitative interviews with five retired National Hockey League players who had retired due to symptoms following multiple concussions revealed that they were significantly affected by their injuries in their postathletic careers and in their personal relationships. They continued to feel debilitated by post-concussive symptoms and experienced symptoms of anxiety and depression. Three of the participants reported thoughts of suicide in the months immediately following their retirement. Though it is difficult to distinguish whether the experiences of these former athletes were a result of multiple concussions or the end of their careers in professional ice hockey, these findings indicate a need for professional support for athletes when they are recovering from concussions and during the transition to their post-athletic careers (Caron et al., 2013).

Recently, after several highly publicized suicides by professional athletes who showed evidence of CTE, there has been growing interest in understanding the relationship between multiple concussions and suicide (Omalu et al., 2006; Reider, 2012). Though there is some indication of a relationship between number of previous concussions and risk of developing depression, very little research has evaluated the relationship between concussions and suicidal thoughts and behaviors. There are certainly theoretical reasons why individuals who have sustained concussions might be predisposed to suicidal ideation and behavior. For example, there is growing evidence that individuals who attempt suicide, particularly those who engage in high lethality attempts, show deficits in attention, working memory, and risk assessment, which overlap with the neurocognitive residua of concussions, both in the short term and, for those with longer-lasting postconcussive symptoms, in the long term as well (Bridge et al., 2012; Jollant et al., 2005; Keilp et al., 2001, 2013). Thus, the deficits associated with a concussion may lower the threshold for a person with suicidal thoughts to act on them. In addition, some of the associated symptoms of concussion, namely pain, depressive symptoms, and sleep impairment, are common antecedents of suicidal behavior (Goldstein et al., 2008; Wong et al., 2011).

There is a growing literature on the relationship between more severe TBI and suicidal behavior. Oquendo and colleagues (2004) examined a clinical sample of depressed patients and found that while a past history of TBI was a risk factor for suicidal behavior, this increased risk for suicidal behavior was explained by the higher rates of substance abuse, cluster B personality disorder, and higher self-reported aggression and hostility in those with TBI. Because this was a cross-sectional study, it was not possible to determine if these characteristics antedated—and perhaps contributed to—the TBI or if they were sequelae of the TBI. Mainio and colleagues (2007) found that among suicide victims, those with TBI were more likely to have been hospitalized for a psychiatric disorder and to have a substance abuse disorder. In a review paper Simpson and Tate (2007) found that pre-morbid psychiatric disorder and substance abuse, a previous suicide attempt, and severe hopelessness were all related to patients with TBI making a suicide attempt. Similar correlates of suicidal ideation were found in community-dwelling adults with a history of TBI (Tsaudousidies et al., 2011), namely, a current psychiatric diagnosis of depression, anxiety, or posttraumatic stress disorder (PTSD) and a history of pre-morbid substance misuse. In this sample there was no relationship between injury severity and suicidal ideation, with 32 percent of those with moderate to severe TBI and 25 percent of those with mild TBI (mTBI) reporting suicidal ideation. Conversely, among patients undergoing treatment for substance abuse, those with a history of TBI were more likely to have had parental loss and childhood conduct disorders and, as an adult, to have made a suicide attempt (48 percent versus 37 percent) (Felde et al., 2006).

In military populations, those under care in the Department of Veterans Affairs (VA) system with a history of TBI were 1.55 times more likely to die by suicide than were those without a history of TBI (Brenner et al., 2011). Greater injury severity was associated with a greater risk of suicide. Among deployed military personnel with mTBI, increased suicidality was associated with depression and with its interaction with the presence of PTSD (Bryan et al., 2013). Similarly, in veterans with TBI, PTSD is a significant risk factor for suicide attempts (Brenner et al., 2011). A study of military personnel (n=161) provides preliminary evidence of a dose-response relationship between the number of TBIs (none, single, multiple), including concussions, an individual has sustained during his or her lifetime and suicidal thoughts or behavior (Bryan and Clemans, 2013). These results persisted even after controlling for the effects of depression, PTSD, and TBI symptom severity. Further research is needed to confirm these findings.

In summary, the long-term effects of repeated concussions have been linked to risk for depression in retired professional football players, although suicidal ideation and behavior have not been reported in these samples (Didehbani et al., 2013; Guskiewicz et al., 2007). In both military and non-military samples, while there appears to be an increased risk for suicidal behavior and suicide associated with TBI, the increased risk has been reported to occur with injuries of greater severity than concussion in most studies (Brenner et al. 2011; Tsaudousidies et al., 2011). In military samples, depression and PTSD are strong contributors to suicidal risk in those with TBI (Brenner et al., 2011; Bryan et al., 2013), and one recent study found that TBI makes a unique contribution to the risk for suicidal ideation or behavior even after controlling for depression and PTSD (Bryan and Clemans, 2013). Prospective studies that examine changes in depression and suicidal ideation behavior pre- and post-concussion will be required to address whether there are increased risks for suicidal behavior in individuals who have suffered a sports-related concussion. Only one intervention study for suicide prevention in individuals with TBI has been reported. In a pilot study Simpson and colleagues randomly assigned 17 patients with TBI and moderate to severe hopelessness or suicidal ideation, or both, to cognitive psychotherapy or usual care, and they found significant decreases in hopelessness in the treatment group versus the usual care group (Simpson et al., 2011). Clearly more work needs to be done on identifying individuals with concussion who are at risk for suicide as well as in developing effective interventions to reduce suicidal risk.

EXPERIMENTAL MODELS

Behavioral and Cognitive Consequences

Several experimental models of repeat TBI have demonstrated changes in behavioral outcome in the absence of overt pathology (DeFord et al., 2002; Prins et al., 2010; Shitaka et al., 2011). The Morris water maze (MWM) task has been traditionally used to quantify memory and learning impairments after moderate TBI and has been used to detect acute deficits after mTBI. In two models of repeat weight drop injury and controlled cortical impact (CCI) injury in the adult mouse, repeat TBI produced significant latency deficits in the MWM 7 days post injury in the absence of histopathology as measured (Creeley et al., 2004; DeFord et al., 2002; Shitaka et al., 2011). Among the studies addressing the cognitive consequences of repeat TBI in the younger brain, MWM deficits at 2 weeks post injury were not detected in 11-day-old rats that received one to three closed-skull CCI injuries (Huh et al., 2007). Because deficits detected by the MWM task following mTBI can be subtle and short-lived, other behavioral assessments have been used to quantify cognitive dysfunction. In a juvenile repeat closed-head CCI injury model, the novel object recognition (NOR) task was used to evaluate transient memory impairments, in the absence of overt cell death, in rats (Prins et al., 2010). Thirty-five-day-old rats were given sham, single, or two injuries (over a 24-hour interval) and were tested in the NOR task 24 hours after the last injury. All groups were able to recognize the novel object when the interval between familiar objects and novel object was 1 hour. Increasing the interval to 24 hours made the task more difficult and resulted in both injured groups showing significant impairments that were alleviated after 3 days of recovery only in the once-injured group (Prins et al., 2010). Findings from this study demonstrate the value of using other testing paradigms to characterize the nature of behavioral and cognitive deficits after mTBI and repeat mTBI. Collectively, experimental models have been able to mimic the acute memory deficits often clinically reported in the absence of gross pathology in both the adult and younger age groups. Future studies need to focus on age and gender differences following multiple concussions.

Acute Cellular Pathology

Experimental models have also been able to show axonal pathology following repeat TBI. Huh and colleagues (2007) used a modified convex silicone 5-millimeter tip to deliver an impact to 11-day-old rats held within a stereotaxic frame. The effects of a single impact were compared to those of sham impacts and of two or three injuries delivered at 5-minute intervals. Histological samples were collected at 1, 3, or 7 days post injury, and cognitive function was assessed 14 days post injury with the MWM. While single injury did not result in gross damage at 7 days, multiple (two or three) impacts caused ventricular enlargement and white matter atrophy. Both reactive astrocytosis throughout the cortical layers and axonal swellings increased with repeat TBI. There were no latency differences between the sham group and any of the injury groups. While the majority of experimental repeat TBI models are rodents, one research group has developed a novel rotational piglet injury model (Raghupathi and Margulies, 2002) that has been used to address axonal injury and cognition after repetitive injury. Piglets (3-5 days old) were given one injury or two injuries 15 minutes apart and were histologically assessed at 6 hours post injury. There were no physiological responses to the injury, with the exception of a mild decrease in blood pressure. While the density of the injured axons did not differ between once- and twice-injured groups, the number of axonal swellings per axon did increase in the twice-injured brains (Raghupathi et al., 2004). In the next study, piglets were a given single rotational injury, two injuries at a 1-day interval, or two injuries at a 1-week interval. Animals given repeat TBI at the 1-day interval showed 43 percent mortality and poorer cognitive composite scores. Those with repeat TBI at the 1-week interval showed greater βAPP staining. No differences were observed between groups on open field testing, T-maze testing, or glass barrier task (Friess et al., 2009).

More recently the CCI injury has been used in a juvenile rat to mimic some of the common pathophysiological processes described after mTBI and concussions, including mild or transient memory impairment, white matter and axonal dysfunction, and the absence of overt cell death (Prins et al., 2010). Adolescent (35-day-old) mice were given either a sham injury, a single injury, or two injuries 24 hours apart, and axonal damage and astrocytic reactivity were histologically assessed 24 hours post injury. The βAPP immunohistochemical labeling was positive in the ipsilateral white matter and was significantly greater in animals exposed to repeat TBI than in the single-injury or sham-injury animals. GFAP labeling revealed slight increases in the ipsilateral gray-white matter junction in single-injury animals. In contrast, the animals with repeat injuries showed a bilateral increase in GFAP with clear morphological changes in the astrocytes. The study is currently the only research addressing the additive effects of repeat TBI in the “adolescent” brain. Experimental models of repeat TBI that report a lack of gross histological pathology have also shown evidence of axonal damage and astrocytic reactivity in both the adult and young developing brain.

Acute Metabolic Dysfunction

As discussed in Chapter 2, one sort of metabolic dysfunction that is known to follow a TBI is a change in the cerebral metabolic rate of glucose consumption (CMRglc). Shortly after an injury the brain enters a prolonged period of glucose metabolic depression. This decrease in CMRglc has been observed in various types of experimental injury models (Andersen and Marmarou, 1992; Chen et al., 2004; Kawamata et al., 1992; Prins and Hovda, 2009; Richards et al., 2001; Sutton et al., 1994; Yoshino et al., 1991, 1992) and in human TBI (Bergsneider et al., 1997; O'Connell et al., 2005). The magnitude and duration of CMRglc depression increases with the severity of the injury and correlates with behavioral dysfunction (Hovda et al., 1994; Moore et al., 2000; Queen et al., 1997). This relationship between CMRglc and injury severity has also been observed in human TBI patients (Hattori et al., 2003). TBIs that do not cause overt cell death or gross pathology can still produce dysfunction. Instances of mTBI, which may be associated with axonal damage or dysfunction as detected from DTI, also show measurable decreases in CMRglc (Gross et al., 1996; Humayun et al., 1989).

Changes in cerebral metabolism have also been examined following repeat TBI at different intervals in adults and in juveniles. Although most of the studies compare single and repeat injury to sham, few studies have incorporated a design in which the interval between injuries was varied. Vagnozzi and colleagues (2007) used a varied interval injury design in adult rats to examine the effects of multiple mild injuries on mitochondrial function and oxidative damage. A weight drop injury was delivered between 1 and 5 days after the primary injury. The greatest cumulative effects on adenosine triphosphate, _N_-Acetylaspartic acid (NAA), redox, and oxidative damage were all seen with the 3-day injury interval. This research is the first to examine the temporal window of cerebral vulnerability after a mild primary TBI. A more recent study conducted in adolescent rats showed decreases in brain glucose metabolism after a single concussion, with CMRglc recovering in 3 days. The duration of the change in brain glucose metabolism was prolonged if the second concussion was delivered within the first 24 hours, but the effects were not cumulative if the second blow was delivered after the rat had recovered from the first concussion (Prins et al., 2013). These results demonstrate that the window of vulnerability for the adult and adolescent brain may be related to post-concussive metabolic derangements, which could be used as a biomarker. Given that the duration of metabolic depression varies with age (Prins and Hovda, 2001; Thomas et al., 2000; Yoshino et al., 1991), it is likely that the metabolic window of vulnerability will also vary with cerebral maturation. These studies emphasize the need for establishing age-appropriate biomarkers for cerebral vulnerability to help inform return-to-play guidelines.

The human findings concerning the effects of repetitive head impact on brain physiology are supported by controlled animal studies which show increased vulnerability to axonal damage and cognitive impairment with repeated mild head injury (Barkhoudarian et al., 2011; Laurer et al., 2001) that are amplified when injuries occur within a day apart (Prins et al., 2010). Moreover, animal studies show that the pathologies differ in myelinated versus unmyelinated fibers (Reeves et al., 2005), which may suggest that myelin provides some protection against concussive injury so that the immature brain with less myelin may be more vulnerable to brain trauma (Shrey et al., 2011). Although this research involved rodents, epidemiological evidence is consistent with children having poorer outcomes than adolescents or adults following mTBI. There is not yet evidence from imaging studies of greater or more sustained diffuse axonal injury in children relative to older individuals.

BIOMARKERS AND RISK FACTORS

TBI biomarkers have been well researched, with hundreds of published articles. However, there are significantly fewer articles on neurochemical or serum markers for mTBI or concussions and even fewer on issues related to children. There are currently no serum biomarkers that have been shown to be related to the risk of subsequent concussions following the first concussion.

Magnetic resonance spectroscopy (MRS) has been used to examine the effects of multiple concussions on brain metabolites. This research shows that changes in NAA, a marker of neuronal injury, take longer to resolve following a second concussion in nonprofessional athletes—an average of 45 days (n=13) versus an average of only 30 days following a first concussion (n=10) (Vagnozzi et al., 2008). Parallel MRS and cerebral glucose metabolism studies in rodents appear to be consistent with these findings in suggesting that the number of concussions and the interval between concussions both play a role in recovery (Longhi et al., 2005; Prins et al., 2013; Tavazzi et al., 2007; Vagnozzi et al., 2005, 2008). Nonetheless, larger studies may be needed to verify the effects of a second concussive event during this recovery curve.

There has been only one study to date that has addressed the issue of vulnerability of the developing brain after repeat concussions, and this was in an animal model. In this study cerebral glucose metabolism (i.e., CMRglc) was measured in 35-day-old (adolescent) rats that had been exposed to repeat concussive injuries (Prins et al., 2013). The closed-head, mild concussive injury model used in the experiment has been shown to generate mild axonal injury without overt pathology and to produce measurable cognitive dysfunction (Prins et al., 2010). Following a single closed-head mild injury, CMRglc was decreased at 24 hours and had recovered by 3 days post injury. When a second injury was introduced during the metabolic depression, both the magnitude and the duration of CMRglc depression were exacerbated. However, when the second injury was delivered after the metabolic depression had receded, CMRglc did not change significantly. These results demonstrate that the window of vulnerability for the adolescent brain may be related to post-concussive metabolic derangements, which could be used as a biomarker. Given that the duration of metabolic depression varies with age (Prins and Hovda, 2001; Thomas et al., 2000; Yoshino et al., 1991), it is likely that the metabolic window of vulnerability will also vary with cerebral maturation. These studies emphasize the need for establishing age-appropriate biomarkers for cerebral vulnerability in the development of return-to-play guidelines.

Beckwith and colleagues (2013) documented the time-course of diagnosis of a subset of concussions from a large dataset of football players at six Division I universities and related the time-course of diagnosis to biomechanical parameters. Of 105 concussions documented during a 6-year period involving more than 1,200 athletes, 45 concussions were diagnosed within the same day of injury, and athletes were removed immediately from the competition. The 60 concussions diagnosed sometime after the contest (with the athletes not removed from competition) were found to have a different pattern of accelerometer readings (Beckwith et al., 2013). Compared to the immediate diagnosis group, this delayed-diagnosis group sustained more impacts on the day of injury (33 compared to 17 impacts above 15 g) and also during the week before injury (70 impacts compared to 50 impacts). The study concluded that concussions diagnosed immediately were related to hits with a greater force, while those diagnosed later were preceded by a higher number of impacts.

LONG-TERM NEURODEGENERATIVE CONSEQUENCES

In general there is a paucity of literature on the long-term neuropathological consequences of repeated or chronic traumatic brain injury in athletes. However, the recent interest in the effects of repeat head injury on professional athletes, many of whom began playing sports in their youth, has prompted a series of studies aimed at identifying the neuroanatomical and neuropathological substrates that underlie the behavioral outcomes in these athletes. Notwithstanding these efforts, little is known about how TBI, either in acute or chronic form, affects the developing brain at the critical periods when circuit formation and synaptic connectivity are active.

Dementia Pugilistica

Much of the current understanding of the neuropathology of chronic traumatic brain injury in athletes comes from classical studies of professional boxers who experienced repeated impacts to the head throughout their careers (Corsellis et al., 1973). The committee recognizes that the nature of brain injury in boxers is quite different from—and its extent much more severe than—that experienced in other contact sports, so that the long-term neuropathological consequences in boxers may be quite different from those in athletes in other contact sports. However, it is possible that the available data on neurodegenerative features in boxers may provide insights for understanding less severe injury conditions.

Dementia pugilistica, also known as “punch-drunk syndrome,” is a chronic progressive traumatic encephalopathy that has been detected in some professional boxers (Corsellis, 1989; Roberts et al., 1990; Tokuda et al., 1991). It exhibits distinct neuropathological features such as cerebral atrophy, thinning of the corpus callosum, enlarged ventricles, and large cavum septi pellucidi with multiple fenestrations, which presumably are caused by tearing of the septa. Microscopically, dementia pugilistica shares certain histological features with Alzheimer's disease, including tau-positive neurofibrillary tangles (NFTs) and diffuse Aβ amyloid plaques, although the distribution of NFTs and plaques in dementia pugilistica shows a greater abundance in the brainstem and the superficial layers of the neocortex. In addition, NFTs in boxers are often clustered as multifocal patches in the dorsolateral frontal cortex, temporal cortex, and orbital gyri, and they show unique perivascular distributions. The tau-positive protein aggregates have also been identified in glial cells and in neurites in the subcortical white matter (McKee et al., 2009; Saing et al., 2012).

As is seen in Alzheimer's disease, the presence of apolipoprotein E (APOE) e4 allele in boxers is associated with an increased risk of cognitive impairment. In particular, high-exposure boxers (those with 12 or more professional bouts) with an APOE e4 allele have been shown to have significantly greater (i.e., worse) chronic traumatic brain injury scores (mean, 3.9; standard deviation [SD], 2.3) than high-exposure boxers without APOE e4 (mean, 1.8; SD, 1.2) (p=0.04) (Jordan et al., 1997). These results were extended by several other studies which showed that, in moderate and severe TBI, APOE e4 allele carriers tend to perform worse on neuropsychological tasks that are presumed to be related to temporal lobe, frontal lobe, and white matter integrity (Ariza et al., 2006), and experience poorer clinical outcomes (Chiang et al., 2003; Friedman et al., 1999). While the underlying mechanisms that account for the increased risks and poor outcome in APOE e4 allele carriers after TBI remain unclear, studies on animal models show that transgenic mice expressing APOE e4 have increased propensity for Aβ amyloid deposits following traumatic brain injury (Hartman et al., 2002; Nicoll et al., 1995).

Chronic Traumatic Encephalopathy

The identification of tau-positive NFTs in boxers with dementia pugilistica raises the question of whether this is an early and a consistent diagnostic feature in repeated traumatic brain injury. In a study focusing on the neuropathological features in a 23-year-old boxer, Geddes and colleagues showed that NFTs, but not amyloid plaques, indeed could be identified in the orbitofrontal and temporal cortex (Geddes et al., 1996). This finding was verified in a follow-up study, in which argyrophilic, tau-positive NFTs, and neuropil threads were consistently identified in two boxers, two patients with repeated head injury due to seizures, and one amateur football player, all of whom died at young ages (Geddes et al., 1999). Although the mechanisms for the abnormal tau-positive NFTs remain unclear, it is postulated that repeated traumatic brain injury may cause mechanical injury in axons, leading to hyperphosphorylation and the formation of abnormal tau protein aggregates. NFTs in these cases tend to be more accentuated in the perivascular regions for reasons that are poorly understood.

CTE is a form of brain neurodegeneration that is thought to result from the sort of repeated head injuries that occur in many contact sports (Gavett et al., 2011). Clinical features of CTE include the progressive decline of memory and cognition, depression, suicidal behavior, poor impulse control, aggressiveness, Parkinsonism, and dementia (Stern et al., 2011). The term chronic traumatic encephalopathy first emerged in two case reports that described neuropathologic changes in two National Football League (NFL) players who suffered from a wide range of neuropsychological disorders after long careers playing football in high school and college and professionally (Omalu et al., 2005, 2006). Gross neuropathological examinations in these two index cases showed no evidence of brain atrophy or fenestrations in the septum pellucidum. Microscopically, the consistent findings were the presence of tau-positive NFTs and neuropil threads, most prominently in frontal, parietal, and temporal neocortex. The results from these two studies were similar to those reported by Geddes and colleagues (1996, 1999), and they strongly suggest that tau-positive NFTs are indeed a consistent and early feature in repeated traumatic brain injury. This notion has been extended by a series of studies that have also identified tau-positive NFTs as a consistent diagnostic neuropathology feature in athletes in professional American football (Goldstein et al., 2012; McKee et al., 2009, 2013; Omalu et al., 2010).

There are a few limitations to this research that should be noted. Most of the studies on CTE have been case reports that lacked proper controls, rendering the results difficult to interpret. Furthermore, unlike neurodegenerative diseases such as Alzheimer's disease and frontotemporal dementia, the diagnostic criteria for CTE are based on a relatively small sample size and have not been universally accepted in the field. Clinical history and comorbid factors, such as complications from other medical conditions and exposures (e.g., substance use) may make it difficult to determine whether the features of CTE are a result of head impacts or of other factors, which complicates the development of diagnostic criteria for CTE.

An important question is whether CTE represents a single “disease entity” that can be graded based on the severity and distribution of tau pathology (McKee et al., 2013), or is part of a spectrum of disease manifestations that happen to share a common finding of tau pathology.

One emerging finding is the high prevalence of patients diagnosed with CTE who also show clinical and neuropathological features of motor neuron disease, such as amyotrophic lateral sclerosis (ALS). These results raise the possibility that the mechanical impacts in repeated head injury trigger a pathological process similar to that reported in frontotemporal lobar degeneration (FTLD), which shares similar signatures of abnormal protein aggregates involving tau and TDP-43 (Mackenzie et al., 2010; McKee et al., 2010). Recent evidence indicates that abnormal tau protein aggregates can propagate from cell to cell in experimental models of neurodegeneration (de Calignon et al., 2012; Kfoury et al., 2012; Liu et al., 2012). In addition, it is possible that cases diagnosed with CTE represent a selected group of individuals who have a higher propensity to develop a spectrum of neurodegenerative disorders that are phenotypically similar to FTLD.

There is limited evidence that APOE e4 is a risk factor for CTE. In an analysis of 10 CTE cases verified by autopsy and where APOE status was known, APOE e4 was overrepresented in those with CTE versus its prevalence in the general population (McKee et al., 2009). A few studies of athletes found an association between APOE e4 and poorer clinical outcomes such as more severe concussion symptoms (Teasdale et al., 2005; Terrell et al., 2008) and lower neurocognitive performance (Kutner et al., 2000), but they do not provide any indication of the role of APOE e4 in the development of CTE.

Other Long-Term Consequences

Several studies indicate that head injury is a risk factor for the development of Alzheimer's disease and other dementias (Bazarian et al., 2009; Fleminger et al., 2003; Guskiewicz et al., 2005; Mortimer et al., 1985, 1991; Plassman et al., 2000; Reitz et al., 2011; Schofield et al., 1997). Plassman and colleagues showed that both moderate and severe head injuries sustained during early adulthood are associated with increased risk of Alzheimer's disease, whereas the relationship between mild head injury and Alzheimer's disease was inconclusive (Plassman et al., 2000). A meta-analysis of 75 published studies found that dementia of the Alzheimer's type was associated with moderate and severe TBI but not with mTBI unless there was loss of consciousness (Bazarian et al., 2009).

Within the past year a study of retired NFL players ages 45 to 73 years (n=5) with histories of mood and cognitive symptoms was published that used positron emission tomography (PET) following intravenous injection with FDDNP1 (FDDNP-PET) to index both tau tangle and amyloid plaque deposition in vivo (Small et al., 2013). This technique was first described by Shoghi-Jadid and colleagues (2002) who observed higher FDDNP signals in brain regions where tau tangles accumulate in Alzheimer's disease and in later research was used to differentiate Alzheimer's from mild cognitive impairment as well as to predict later cognitive decline (Small et al., 2006, 2012). In the current study, FDDNP-PET signals in subcortical and cortical regions in retired athletes were similar to controls (n=5) of comparable age, education level, body mass index, and family histories of Alzheimer's disease, but greater depressive symptoms were seen in the retired athletes as measured by the Hamilton Rating Scale for Depression. The results indicated higher FDDNP signals in athletes compared with controls in regions such as the amygdala that have been shown to produce tau deposits following trauma. Unfortunately, a small sample and lack of postmortem tissue to confirm the findings limit the implications of this study. The trend for higher FDDNP signals in the two oldest athletes (64 and 73 years) is consistent with previous reports of elevated FDDNP binding in geriatric depression (Kumar et al., 2011).

Some research indicates that TBI may be associated with higher incidence of ALS. For instance, two case-control studies found an increased risk of ALS among veterans of the armed forces who had experienced head injuries during the last 15 years (Schmidt et al., 2010) and among soccer players with multiple head injuries (Chen et al., 2007). The association of ALS with both severe and repeated traumatic brain injury is supported by another case-control study from a population-based registry (Pupillo et al., 2012). However, a systematic review of studies found insufficient evidence to support an association between TBI and ALS (Bazarian et al., 2009).

The presence of abnormal TDP-43 protein aggregates in certain CTE cases suggests similar proteinopathy might increase the propensity of ALS in individuals with CTE (McKee et al., 2010), although the exact nature of the relationship between these disease entities requires further research. Similar to the case with ALS, TBI and repeat TBI are associated with an increased risk for the development of Parkinson's disease (Bazarian et al., 2009; Goldman et al., 2006), but because α-synuclein pathology is not a common feature in CTE, the mechanisms that connect Parkinson's disease and CTE will require future larger-scale studies to confirm. Equally important will be the ability to accurately determine if the underlying neuropathology leading to Parkinson's disease in CTE patients is caused by the α-synucleinopathy seen in sporadic Parkinson's disease or the tauopathy seen in individuals with FTLD with atypical Parkinsonism.

Experimental Models of the Long-Term Consequences of Repeat TBI

It is important to note that, at present, all experimental models of repeat TBI addressing long-term neurodegenerative consequences involve injuries sustained in adulthood. Models of repeat injury have been developed for use in transgenic mice to determine whether repeat TBI increases susceptibility for Alzheimer's disease. Under pentobarbital anesthesia, 8- to 10-week-old mice were placed in a stereotaxic frame, and a 6-mm silicone tip was used to deliver a CCI injury to the exposed skull (Laurer et al., 2001). Mice given either a single injury or two injuries 24 hours apart were given motor, cognitive, and histological assessments. While no significant neuroscore or cognitive deficits were seen, the repeat TBI group showed impairments in the rotorod task. The repeat TBI group also showed significantly greater blood brain–barrier breakdown and axonal injury than the single-injury group. No βAPP or tau deposits were observed in any group at 56 days post injury. The same injury paradigm was applied to transgenic mice that expressed human Aβ precursor protein (Uryu et al., 2002). While the injured groups showed no motor deficits, the repeat group showed latency deficits in the MWM and increased cortical Aβ deposits at 16 weeks post injury. Treatment of these transgenic mice with vitamin E– enriched food for 4 weeks prior to the two injuries given 24 hours apart decreased the levels of lipid peroxidation and the number of Aβ deposits and improved cognitive performances relative to standard-fed transgenic mice (Conte et al., 2004). CCI injury has also been delivered to transgenic mice expressing the human tau isoform (Yoshiyama et al., 2005). Among all the published repeat TBI models, this is the only study that delivered impacts to both hemispheres. A 9-mm silicone tip delivered four injuries to the exposed skull with two injuries per hemisphere at 20-minute intervals. This series of four injuries was repeated once a week for 4 consecutive weeks. Neurobehavioral tests conducted at 6 months post injury showed no difference between wild-type and transgenic repeat TBI groups. The study reported that one transgenic mouse showed extensive neurofibrillary tangles and had significant behavioral deficits.

The effects of TBI on the development of Aβ and tau pathology have also been recapitulated in triple-transgenic Alzheimer's disease mice harboring mutant genes for Aβ, presenilin-1, and tau (P301L) (Oddo et al., 2003). Similar to the results from other studies, CCI in the triple-transgenic Alzheimer's disease mice also results in intra-axonal Aβ accumulation and phospho-tau immunoreactivity at 24 hours and up to 7 days after injury (Tran et al., 2011). Treatment with compound E, a γ-secretase inhibitor, successfully blocks the posttraumatic Aβ accumulation but not the tau protein pathology. Results from the Aβ and tau transgenic mice studies provide some of the first evidence that repeat TBI can increase accumulation of these proteins and increase the risk of neurocognitive complications.

In addition to the implications of Aβ and tau in the pathogenesis of TBI, there is emerging evidence that reactive astrogliosis and microgliosis may contribute to neuronal injury following TBI. Indeed, microarray analyses of gene expression profiles in wild-type adult mouse brains at multiple time points following TBI show dysregulations of multiple gene ontology categories, including trophic factors, transcription factors, inflammationrelated factors, and many glial markers (Kobori et al., 2002). One of the astroglial genes, S100A4—which is markedly up-regulated in, and released by, white matter astroglia—has been shown to protect neurons from apoptosis during TBI (Dmytriyeva et al., 2012; Kozlova and Lukanidin, 2002). Genetic deletion of S100A4 exacerbates neuronal loss after TBI because of increases in oxidative cell damage and down-regulation of neuroprotective protein metallothioneins. These results raise the intriguing possibility that, in addition to the potential influences from neurodegeneration-related genes, there are many critical molecular and cellular responses that can contribute to the pathogenesis of neuronal injury following TBI. Many of these factors could serve as potential therapeutic targets to mitigate both the short-term and the long-term consequences of TBI.

FINDINGS

The committee offers the following findings on the consequences of repetitive head impacts and multiple concussions:

REFERENCES

1

FDDNP refers to 2-(1-{6-[(2-[F-18]fluorethyl)(methyl)amino]-2-napthyl}ethylidene)malononitrile, a chemical marker injected prior to brain imaging to help identify accumulation of abnormal protein deposits.