Clinical manifestations

Presentation and course

The signs and symptoms of concussion are directly related to the disruption of brain physiology. Concussion symptoms can be immediate and sometimes delayed up to 72 hours. Most concussion symptoms are short-lived, lasting on average 10 to 14 days (98). As a rule, the longer the symptoms last, the more severe the concussion (228). Understanding the signs and symptoms and when they occur is essential in managing the concussion. Immediate signs and symptoms include vacant stare (befuddled facial expression), delayed verbal and motor responses (slow to answer questions follow instructions), confusion and inability to focus attention, disorientation (walking in the wrong direction, unaware of time, date, and place), slurred or incoherent speech, gross observable decrease in coordination (stumbling, trouble with tandem gait), emotions out of proportion to circumstances, memory deficits (repeatedly asking the same question that has already been answered, or inability to memorize and recall 3 of 3 words or 3 of 3 objects in 5 minutes), or any period of loss of consciousness (242; 333). Within the first few hours, the patient may experience intermediate symptoms, including headache (by far the most common symptom), dizziness, loss of equilibrium or vertigo, lack of awareness of their surroundings, and nausea or vomiting (111; 242; 372). As the concussion progresses, late signs and symptoms develop and can include persistent low-grade and sometimes even severe headache (may include photo-/phonophobia); lightheadedness; poor attention, concentration, and memory; heat intolerance; easy fatigability; sleep disturbances; irritability; anxiety or depressed mood, which are often unrecognized; sleep disturbances; and decreased reaction time, gait abnormalities, and balance (304), which are the last symptoms to resolve (41; 242). On physical examination concussed individuals may demonstrate the aforementioned neurocognitive findings along with decreased smooth pursuits on extra ocular movement testing, issues with near point convergence, exacerbation of symptoms with the aforementioned tests, as well as sway on Romberg and trouble with tandem gait and dual task switching (ie, tandem gate while counting backwards from 100).

Prognosis and complications

Most concussions resolve spontaneously within 10 to 14 days (243; 96). Major complications of mild traumatic brain injury are prolonged postconcussion symptoms (postconcussion syndrome is now defined by ICD-10 as a diagnosis of malingering), ie, symptoms greater than 6 months, and chronic posttraumatic headache. Prolonged postconcussion symptoms include headache, dizziness, sensitivity to light or noise, difficulty sleeping, and cognitive issues.

Individuals who suffer multiple or repeat concussions are likely at risk for neuropathological changes and may be at risk for early dementia and prolonged cognitive issues (160). Furthermore, a number of studies have shown an increased risk for the development of posttraumatic stress disorder. It is estimated that 15% to 30% of concussed individuals will go on to develop prolonged postconcussion symptoms (82); however, definitive studies are lacking. Impaired neurovascular unit function has been hypothesized to contribute to persistent symptoms (20). Larger prospective studies have concluded that the risk for prolonged postconcussion symptoms is increased in older children with loss of consciousness, headache, nausea, or vomiting. Other variables include female gender, Glasgow Coma Scale of less than 15, prior history of depression, alcohol intoxication at time of injury, etiology of assault (33), duration of posttraumatic amnesia (149), prior history of migraine, or psychiatric issues (135). Other studies have noted that initial dizziness may predict prolonged postconcussion symptoms (392). Posttraumatic headache appears more likely to develop following mild traumatic brain injury (concussion) compared with moderate or severe traumatic brain injury (356).

Carlson and colleagues found that 80% of Operation Enduring Freedom and Operation Iraqi Freedom veterans with traumatic brain injury had a psychiatric diagnosis (46). Moreover, mild traumatic brain injury is associated with elevated rates of posttraumatic stress disorder in veterans serving in Iraq and Afghanistan (258). Other studies have demonstrated elevated rates of major depression among soldiers returning from deployment to Iraq who sustained mild traumatic brain injury with loss of consciousness (22.9%) compared with those with other injuries (6.6%). In a study by Hoge and colleagues, there was not a significantly elevated risk of depression among those with mild traumatic brain injury with alterations in, but not loss of, consciousness (156). Among civilians, the incidence of new-onset depression was 18% within the first year following mild traumatic brain injury (298). Mayers looked at concussed college athletes from 1998 to 2011 (235). A total of 98 concussions occurred in 95 athletes; 69% were male and 31% female. The mean number of days with symptoms was 6.9, and the mean number of days until return-to-play was 14.6. There was no significant gender effect on symptom duration or time interval until return-to-play. Forty-three percent had experienced a prior concussion. Of the athletes suffering persistent short-term memory loss (n=8), only one recovered after 8 months.

Chronic traumatic encephalopathy. Chronic traumatic encephalopathy is defined as a progressive neurodegenerative syndrome caused by single, episodic, or repetitive impacts to the head (274). The syndrome was first described in 1928 by Martland who published a paper in the Journal of the American Medical Association titled “Punch drunk” (232). In it he described a group of what he termed “poor fighters” who tended to “take considerable head punishment”. According to the author, early symptoms included occasional clumsiness, slight ataxia, and periods of confusion. He noted that many never progress beyond this stage whereas others would go on to develop tremors, dysarthria, deafness, physical slowing, “dragging legs while walking,” and mental deterioration to a point where some required institutionalization. Some would go on to develop a progressive neurologic syndrome leading to mental or physical helplessness. In 1937, Millspaugh in describing effects in navy boxers coined the term dementia pugilistica (255). The term chronic traumatic encephalopathy first appeared in the literature in the mid-1960s, and in 1973 Corsellis, Bruton, and Freeman-Browne described three stages of clinical deterioration in chronic traumatic encephalopathy (65). The disorder gained widespread national attention after a 2005 published report by Dr. Bennett Omalu showed neuropathological changes consistent with chronic traumatic encephalopathy in a retired national football league player (275) and again after a series of case reports by researchers at Boston University in 2009 (245; 328). Chronic traumatic encephalopathy has also been seen in military veterans exposed to blast injury and other individuals with head trauma (343; 274; 350).

Many researchers have attempted to classify a clinical syndrome of chronic traumatic encephalopathy (245; 343; 274). The duration between precipitating event or events and development of clinical symptoms of chronic traumatic encephalopathy (ie, the latency period) varies widely among reported cases and ranges from weeks to months and to up to 40 years (274). Early symptoms include chronic headaches and dizziness as well as mood disorders (mainly depression), paranoia, agitation, social withdrawal, poor judgment, and aggression. As the disease progresses, neurocognitive symptoms and Parkinsonian symptoms become more predominant (343; 385; 274). The largest study to date, by Stern and colleagues, looked at 36 adult male athletes with a confirmed neuropathological diagnosis of chronic traumatic encephalopathy (343). History of clinical presentations were obtained through postmortem telephone interviews with the patients’ next of kin. Three subjects were asymptomatic at the time of death; 11 of the symptomatic cases had changes in neurocognitive function prior to the development of behavioral or mood disturbances. Initial changes in behavior and mood prior to the development of neurocognitive disturbance were seen in 13 subjects. None of the subjects had motor disturbances as their initial feature. Ten subjects were diagnosed with dementia, and all were rated as having stage IV chronic traumatic encephalopathy. The authors hypothesized that there could be two different clinical presentations of chronic traumatic encephalopathy. The first begins at an early age with patients exhibiting behavioral and mood changes, and the other at a later age with individuals exhibiting neurocognitive impairment (343). Case reports have suggested a possible higher risk of suicide (245; 274). A systemic literature review found 85 preliminary abstracts with only seven meeting the criteria for review; four of those were excluded as single case studies, and one did not include suicide as an outcome, leaving only two case series. In looking at the case series, the evidence regarding a relationship between chronic traumatic encephalopathy and suicide was felt to be insignificant (385). To date, there have been no published epidemiological, cross-sectional, or prospective controlled studies demonstrating any correlation between the described neuropathological and clinical changes (242). As a result, chronic traumatic encephalopathy can only be classified as a pathological syndrome.

According to the current consensus, requirements for pathological diagnosis of chronic traumatic encephalopathy include abnormal perivascular accumulation of Tau in neurons; astrocytes and cell processes in an irregular pattern at the depths of the cortical sulci; and confirmed tau immunoreactivity in an irregular, focal, perivascular distribution at the depths of the cortical sulci (248). Gross examination of the brain from individuals with chronic traumatic encephalopathy often shows a reduction in brain weight; pituitary, hippocampal, thalamic, and midbrain mammillary atrophy; enlargement of the lateral and third ventricles; thinning of the corpus callosum; cavum septum pellucidum with fenestrations; scarring and neuronal loss of the cerebellar tonsils; and atrophy of the frontal, temporal and parietal lobes (245; 248; 274). There appears to be a divergence on the microscopic and immunohistochemical characteristics reported by McKee and colleagues as compared to Omalu. McKee and colleagues found a predominance of perivascular foci of phosphorylated tau, immunoreactive astrocytic tangles, and neurofibrillary tangles (245). There was an irregular cortical distribution of phosphorylated tau, immunoreactive neurofibrillary tangles, and astrocytic tangles. Astrocytic tangles tend to be clustered in the subpial and periventricular regions (sub cortical U fibers, external capsule, anterior and posterior commissures, and corpus callosum) in the cerebral cortex, thalamus, basal ganglia, and brainstem. This tends to correlate with findings seen in diffuser tensor imaging studies in patients with mild traumatic brain injury and traumatic brain injury. Neurofibrillary tangles in the cortex have been seen in the more superficial layers (245). Alternatively, Omalu found sparse to frequent neurofibrillary tangles and neutrophilic threads in the cerebral cortex, basal ganglia, subcortical nuclei, and brainstem with or without neurofibrillary tangles and neutrophil threads (274). The neuronal neurofibrillary tangles may exhibit what he called a distinctive "skip phenomenon" in the cortex, whereby the neurofibrillary tangles were haphazardly located, being present and absent in adjacent regions of the cortex within the same lobes of the brain. In addition, cerebral amyloid angiopathy can accompany the tauopathy in chronic traumatic encephalopathy (274). Finally, McKee and colleagues proposed a staging criterion for chronic traumatic encephalopathy, ie, four stages based the observed pathological cases and clinical syndromes reported through postmortem interviews with patients’ families (245).

Whether or not chronic traumatic encephalopathy is its own entity or a precursor or subtype or Alzheimer disease or other forms of cortical dementia remains to be seen. The incidence of dementia in the general population doubles every 5 years, with prevalence in the 30- to 64-year-old age group being 54/100,000 (242). Studies have shown evidence of Tau pathology in Alzheimer disease and frontotemporal dementia. Furthermore, deposition of beta amyloid in the form of diffuse amyloid plaques can be seen in 20% to 30% of pathologically diagnosed chronic traumatic encephalopathy cases. With advancing age, chronic traumatic encephalopathy and Alzheimer disease become pathologically indistinguishable, and Omalu has proposed a disease model whereby chronic traumatic encephalopathy could progress and become fully developed Alzheimer disease (274). An alternative hypothesis is that the Tau deposition seen early in the pathological process could represent early stages of a subtype of Alzheimer disease—ie, chronic traumatic encephalopathy is an early form of Alzheimer disease.

At present, there is still no direct evidence of causation. Most authors have attempted to link chronic traumatic encephalopathy with single or repeat concussions or subconcussive hits (245; 274). However, there are no published prospective crossover studies linking chronic traumatic encephalopathy with the above mechanisms. Animal models have suggested that repeated mild traumatic brain injury can result in tau immunoreactivity and astrocyte or microglia activation, especially in the superficial layer of the motor or somatosensory cortex and corpus callosum. However, common human markers for chronic traumatic encephalopathy, ie, neuritic threads or astrocytic tangles, were not seen (270). Furthermore, not all athletes with multiple concussions go on to develop neurocognitive issues and pathologically diagnosed chronic traumatic encephalopathy. In a small study by Hazrati and colleagues that examined six cases of Canadian Football League players with a history of multiple concussions and reported progressive neurologic decline, only three had pathologically diagnosed chronic traumatic encephalopathy (151). This suggests a possible genetic link. Stern and colleagues, in looking at the Boston University population of athletes with pathologically diagnosed chronic traumatic encephalopathy, found a predominance of APOE-4 homozygotes in the chronic traumatic encephalopathy group when compared to age-matched controls and proportionally more e4 homozygotes in the cognition group (343).

What is the pathogenesis by which concussion or subconcussive hits can result in chronic traumatic encephalopathy? Multiple cascades with multiple pathogenic mechanisms have been proposed (248). In the hyperphosphorylation cascade, mild traumatic brain injury via a metabolic cascade (ie, via increased membrane permeability and ionic shifts with calcium influx) can cause normal Tau to dissociate and become hyperphosphorylated, rendering it insoluble and too large to function in axons; as a result, it translocates from the axon. Tau then accumulates, and neurofibrillary tangles form and are secreted into extracellular space where they affect other neurons, ultimately leading to de-innervation and cell death (224; 274). Another possible mechanism involves activation of astrocytes (which provide support for neurons) in the white matter via proinflammatory cytokines. This can acutely occur immediately after head trauma, or initial head trauma can prime the microglia with subsequent injury, resulting in microglia activation. In turn, activated astrocytes release reactive oxygen species that indirectly result in excitotoxicity by inhibiting the ability of astrocytes to uptake glutamate (224; 248). Another possibility is through microglial activation. Microglia are inactive immune cells in the normal brain. It is hypothesized that repetitive head trauma can activate microglia. Short-term activation appears to be protective; however, long-term activation results in the formation of glial tangles (224). Abnormal protein folding has also been hypothesized. Following brain injury, the endoplasmic reticulum, which is responsible for the correct folding and sorting of proteins, becomes dysfunctional. A dysfunctional endoplasmic reticulum results in protein accumulation in the cell, which some have termed, “endoplasmic reticulum stress response.” Endoplasmic reticulum stress response via the enzymatic regulation of CHOP protein ultimately results in apoptosis and tau hypophosphorylation (224; 248).

At present chronic traumatic encephalopathy remains a pathological diagnosis with a lack of consensus about the microscopic and immuno-histochemical characteristics. Longitudinal, controlled, prospective studies are underway to further define if an associated clinical syndrome actually exists. Evidence-based studies are also needed to determine causation (association is not the same as causation) with single or multiple concussions or subconcussive hits and if a genetic predisposition exists. Research is also needed to determine whether chronic traumatic encephalopathy is its own pathological and clinical entity or a precursor to Alzheimer disease or other types of dementia. Preliminary case reports (with sampling bias) in athletes who have participated in high-impact contact sports have shown a higher incidence of neurocognitive deficits. Whether or not these neurocognitive deficits are directly related to traumatic brain injury or a precursor to a progressive neuropathological process remains to be seen.

Amyotrophic lateral sclerosis. Repetitive concussions also appear to be a risk for amyotrophic lateral sclerosis. Epidemiological studies have demonstrated an increased risk of head trauma and the development of amyotrophic lateral sclerosis (55). Other studies have demonstrated an increased risk of amyotrophic lateral sclerosis in Italian professional soccer players and NFL players in the United States. The most recent by Ettore Beghi was presented as an abstract at the American Academy of Neurology meeting in 2019 (27). The team reviewed trading cards of about 25,000 male professional soccer players who played in Italy from 1959 to 2000, recording birth date, place of birth for each player, as well as soccer position and their playing history with the team. They used news reports to determine which players developed amyotrophic lateral sclerosis. Researchers found 33 soccer players developed amyotrophic lateral sclerosis, ie, 3.2 cases per 100,000 annually. The rate of amyotrophic lateral sclerosis in the general Italian population is 1.7 cases per 100,000 people every year. This translates into a 2-fold increase in the development of amyotrophic lateral sclerosis among Italian soccer players. When looking just at soccer players age 45 and younger, the rate of developing amyotrophic lateral sclerosis was 4.7 times higher. Finally, the average age of developing amyotrophic lateral sclerosis among soccer players was 43 years compared to 63 years for the general population. A Swiss study also found a history of head trauma to be the only extrinsic risk factor for developing amyotrophic lateral sclerosis (107).

MeKee and colleagues examined 12 cases of chronic traumatic encephalopathy; in 10 of those cases, they found a widespread TAR DNA-binding protein of approximately 43kd (TDP-43) (245). Three of the athletes with chronic traumatic encephalopathy also developed signs and symptoms of progressive motor neuron disease; in these three cases, there were abundant TDP-43-positive inclusions and neurites in the spinal cord in addition to tau neurofibrillary changes, motor neuron loss, and corticospinal tract degeneration. The results suggest that TDP-43 proteinopathy seen in chronic traumatic encephalopathy can extend into the spinal cord and is associated with motor neuron disease (246). Rat percussion studies have also found long lasting deficits in motor function, increased incidence of tau deposition, decreased cortical thickness, atrophy of the corpus callosum, and earlier onset of amyotrophic lateral sclerosis symptoms in SOD1 rats. (358).

Alzheimer disease. Although there are currently no published studies directly linking single or multiple concussions or subconcussive hits in athletes with an increased risk for Alzheimer disease, there are numerous case reports showing progressive neurocognitive deficits in retired athletes with a history of concussion. The largest and best study to date retrospectively looked at 1.4 million military veterans (18). There was a 2- to 3-fold increased risk of being diagnosed with dementia after mild traumatic brain injury. Alzheimer disease has a long prodromal phase prior to presentation; therefore, mild traumatic brain injury and traumatic brain injury early in life would not impact the individual until decades later. In addition, 75% of all traumatic brain injuries are concussion-related, and a history of traumatic brain injury predisposes some individuals to Alzheimer disease (36; 103; 176). With that said, advancing age remains the biggest risk factor for the development of the disorder (36). The average neurologist is well aware of the underlying pathology of Alzheimer disease, ie, amyloid beta deposition (ABeta peptides), neurofibrillary tangles, widespread loss of cortical neurons, and inflammation of the brain’s glial support cells. Many may not be aware that neurofibrillary tangles and beta amyloid pathology is present in one third of traumatic brain injury patients, with beta amyloid developing in the acute phase of traumatic brain injury (neurofibrillary tangles are seen later in the disorder as the symptoms become more chronic) (36). Animal models have demonstrated that axonal swelling immediately after traumatic brain injury results in amyloid beta production as well as impaired protein transport and cytoskeletal alterations. Furthermore, studies using transgenic mice have shown that amyloid beta may also spread from the initial traumatic brain injury site to more distant brain regions (178), which is similar to what is hypothesized with tau in chronic traumatic encephalopathy. In addition to its presence in the acute phase of traumatic brain injury, beta amyloid (beta amyloid peptide AB42) has been shown to be present many years after traumatic brain injury. This protein is highly neurotoxic and predisposed to aggregation, which, in turn, can result in cell death (326). In addition to beta amyloid deposition, neuroinflammation and microglial activation likely play a role in the development of Alzheimer disease in patients with a history of traumatic brain injury. Animal models have shown an association between traumatic brain injury and microglial activation and the development of anti-inflammatory cytokines, the latter of which may be neuroprotective due to their ability to clear beta amyloid. However, activated microglia and proinflammatory cytokines can persist for many years after the initial traumatic event where they can have detrimental effects on brain parenchyma. Taken together, it is possible that a vicious cycle involving microglial and cerebral amyloid beta activation with spreading may result in the development of Alzheimer disease pathology in patients with a history of traumatic brain injury.

It is possible that trauma is not an independent risk factor; however, when combined with genetic risk factors, mild traumatic brain injury predisposes the individual to the development of Alzheimer disease pathology. Hayes and colleagues studied 160 Iraq and Afghanistan War veterans with a history of mild traumatic brain injury (150). Individuals with a high genetic (APO E4) risk for Alzheimer disease had reduced cortical thickness in Alzheimer disease vulnerable regions, the degree of which was directly related to the duration from time of initial injury. There are also studies refuting the traumatic brain injury-Alzheimer disease link. In a sample of 4761 autopsy patients, 453 with a remote history of traumatic brain injury, a self-reported history of traumatic brain injury was not related to Alzheimer disease neuropathic changes, other neurodegenerative disorders, and did not predict baseline dementia severity or cognitive function (347). Further research is needed to more clearly delineate the neuropathological and clinical phenotype of the progressive neurocognitive changes seen in a subpopulation of patients with mild traumatic brain injury. As Gardner and colleagues have proposed, it is possible that these individuals suffer from a separate, yet to be defined dementia (126).

Parkinson disease. The death of boxing great Muhammad Ali rekindled the debate about the risk of Parkinson disease with repeat head trauma (80; 146). Most experts, however, feel that Mr. Ali’s Parkinson disease was genetically related. With that said, repeated head trauma appears to be a risk for the development of Parkinson disease per a meta-analysis of 22 studies that reported a pooled odds ratio of 1.57 for the association between Parkinson disease and head trauma (169). The risk for Parkinson disease in mild traumatic brain injury is less clear; however, a study compared 11,000 patients with mild traumatic brain injury to over 113,000 patients with non-traumatic brain injury trauma. The researchers identified over 1,300 subsequent cases of Parkinson disease, noting that patients with mild traumatic brain injury were 24% more likely to develop Parkinson disease than those with non-traumatic brain injury trauma (122; 125). Another large epidemiological study conducted on almost 500 Italian patients with Parkinson disease also demonstrated a statistically significant risk of developing Parkinson disease after head trauma when compared to age-matched controls. (267). Other studies have revealed conflicting results (315). From a pathophysiology standpoint, it appears that the protein alpha-synuclein plays a major factor. This presynaptic protein plays a role in synaptic vesicle recycling and major histocompatibility complex II, which is involved in inflammation. The accumulation of alpha-synuclein has been shown to cause neuronal damage, likely via the disruption of cell membranes of dopaminergic neurons in the substantia nigra (02). Alpha-synuclein is elevated in cerebrospinal fluid of patients with traumatic brain injury compared to controls during the week following injury, and the degree of elevation is highly predictive of survival (257); deposition of this protein is often seen within injured axons in those who do not survive (368). Conversely, a small study (34 patients) using ioflupane I 122 injections ie, DaTscan and single photon emission tomography, to assess the incidence of Parkinson disease 1 to 7 years post mild traumatic brain injury did not demonstrate a significant increased risk of Parkinson disease in those with mild traumatic brain injury (300). Overall, research is lacking regarding the actual link and mechanism between head trauma and Parkinson disease, and until in vivo biomarkers (either neuroimaging, blood, or CSF) are available, significant questions will remain.

Clinical vignette

A 15-year-old male with no prior history of concussion or headache, nor family history of headache, was referred for 6 months of daily headache and postconcussion symptoms, including difficulty with attention and concentration, short-term memory loss, loss of equilibrium, and anxiety. The patient reported that the symptoms occurred after a high school football game during which he played the center position. He did not recall any significant head trauma; however, he sustained multiple subconcussive hits. Approximately 24 hours later, he began to experience headache, dizziness (described both loss of equilibrium and lightheadedness), fatigue, confusion (described as “being dazed”), and “brain fog.”

The patient continued to attend football practices, and over the next 2 weeks developed increased fatigue, worsening headaches, photophobia, phonophobia, trouble with attention and concentration, tinnitus, bilateral hand numbness, and worsening of his vestibular symptoms. He was initially seen by his primary care physician, who diagnosed him with vestibular neuritis and treated with high-dose steroids. The patient continued to practice when his symptoms would allow. His symptoms persisted for the next 6 weeks, during which time there was a marked decrease in his academic performance. Shortly thereafter, he experienced a syncopal episode while lifting weights and was sent for an MRI of the brain, which demonstrated a small Chiari malformation. He was referred to a pediatric neurosurgeon at a major academic institution who did not feel that the symptoms were related to the Chiari malformation; however, the neurosurgeon did not offer any alternative explanation. A psychiatrist was consulted and prescribed citalopram for underlying depression and anxiety. About 3 months after the onset of the symptoms, the patient was finally evaluated by a local neurologist who diagnosed him with prolonged postconcussion syndrome and posttraumatic headache. The patient was referred for vestibular rehabilitation and instructed to limit his physical activity. Over the next 4 weeks, he had mild and only short-term improvement in the symptoms.

On initial consultation, the patient was experiencing daily headaches, which included an ever-present underlying low-grade dull aching headache in a reverse baseball cap distribution. The headache would intensify on a daily basis to a severe disabling headache, which was holocranial in nature and described as throbbing, vise like, and pounding. The pain was rated a 7 to 8 out of 10 on a 10-point visual analog scale. There was associated photophobia, phonophobia, nausea, occasional vomiting, and blurred vision. The headache would worsen with bending and physical activity, the latter of which, along with prolonged cognitive activity, would trigger the more severe headache. He would obtain mild relief with over-the-counter medications, including Tylenol, Aleve, and ibuprofen, and prescription medications, including oxycodone. Prophylactic treatment had not been tried, and he was not overusing medications on a daily basis. The patient was also complaining of dizziness, which on detailed questioning was most consistent with a combination of disequilibrium and lightheadedness. The symptoms were constant and would wax and wane in intensity with physical and cognitive activity along with quick position changes, ie, lying to seated and standing, and rapid head movements. Persistent difficulties with attention, concentration, and retention were also present.

His neurologic exam was normal other than decreased smooth pursuits with sway on Romberg and difficulty with task switching (tested using a combination of heel to toe exam and rapid alternating movements). Musculoskeletal exam demonstrated pain on palpation bilaterally in the suboccipital muscles, cervical paraspinal muscles, trapezius, levator scapular, and over the greater and lesser occipital nerves, auricular temporal, and supraorbital nerves bilaterally.

His headache was classified as chronic posttraumatic headache secondary to mild head injury (5.2.2) along with prolonged postconcussion syndrome. He was treated with a prednisone taper (60 mg daily weaning by 10 mg every 3 days), frovatriptan 2.5 mg at breakfast and dinner for 10 days, and tizanidine 4 mg at bedtime, and he underwent supraorbital, supratrochlear, auricular temporal, and greater and lesser occipital nerve blocks. The latter resulted in a significant decrease in his headache symptoms and a significant improvement in his dizziness. He was also started on a physical therapy program that included massage, myofascial release, ultrasound, cervical traction and therapeutic exercise, and electrical stimulation. He was instructed to limit his physical and cognitive activity to tolerance.

Over the next 3 months, the patient continued to experience symptoms with headaches being the predominant issue. He was tried on prophylactic medications, including memantine (10 mg bid) and zonisamide (100 mg bid), which resulted in a slight decrease in overall headache intensity and a slight decrease in the number of severe disabling headaches. Transitional medications, including meloxicam (5 mg bid), and abortive medications, including sumatriptan (100 mg prn) and then dihydroergotamine (1 mg intramuscular daily as needed for headache), were also tried with some limited success. Eight months from the onset of the symptoms, the patient was still experiencing daily headaches rated at a 3 to 4 out of 10, which would intensify approximately 2 to 3 days per week into a severe, disabling headache rated at an 8 out of 10. He also continued to have significant vestibular and cognitive symptoms (especially with increased physical and cognitive activity). He underwent botulinum toxin injections using a modified PREEMPT injection paradigm incorporating a fix dose, fixed site, and modified follow the pain pattern. One month after the injection, the patient was completely headache free, with resolution of his cognitive and vestibular symptoms. The patient had also discontinued all abortive and preventative medications. Four months later, he continued to remain headache free and asymptomatic and had returned to his prior academic courseload with a 3.70 grade point average.

Biological basis

Etiology and pathogenesis

Definition. There is currently no standard definition of minor closed head injury. Some authors have attempted to separate the terms “mild traumatic brain injury” and “concussion,” whereas others use the terms synonymously (239; 240). The Centers for Disease Control and Prevention has recommended the use of a single term, ie, mTBI (225). This author does not agree with grouping concussion and mTBI into one definition. Therefore, one plausible definition of concussion is a prolonged transient alteration in neuronal function caused by a blow to the head, or body with transmission of force to the head, resulting in rotational or translational (ie, angular and lateral) movement of the head resulting in neurologic symptoms that resolve sequentially over time (64). For the purposes of this article, the term "minor closed head injury" will encompass concussion and mild traumatic brain injury. The latter will be defined as more than 6 months of symptoms, including headache; dizziness, ie, vertigo; lightheadedness; or disequilibrium as well as cognitive symptoms. Although headache is not the only symptom, there is no evidence of structural abnormalities on standard or advanced neuroimaging studies. Neuropsychological testing should show abnormalities in subtests, such as processing speed, attention, executive function, and memory. However, there are no signs of malingering or secondary gain. In this article, “traumatic brain injury” will include the aforementioned symptoms with signs of structural brain abnormalities on standard and advanced neuroimaging.

Mechanism. There are two primary mechanisms associated with traumatic brain injury: impact loading and impulse loading. Impact loading involves a direct blow transmitted primarily through the center mass of the head causing linear acceleration and results in extracranial focal injuries, such as contusions, lacerations, external hematomas, and skull fractures. Shock waves from blunt force trauma may also cause underlying focal brain injuries, such as cerebral contusions, subarachnoid hematomas, and intracerebral hemorrhages. Impulse (inertial) loading caused by sudden movement of the brain relative to the skull results in angular and rotational acceleration and produces subdural hematomas and concussion. Inertial loading at the surface can produce bridging vein rupture and resultant subdural hematoma or can affect deep white matter structures, producing diffuse axonal injury, ie, a form of traumatic brain injury.

There are several theories about the possible mechanism of concussion; however, none are universally accepted (325; 240c). The mechanism is felt to be similar in sports injuries, falls, and motor vehicle accidents (281; 28). It is generally accepted that both angular and rotational acceleration are the main biomechanical forces involved in concussion. The head can be directly impacted, or impact can be transferred to the head from other parts of the body; this, in turn, causes rotation of the cerebral hemispheres around the upper brainstem (239). Holbourn was the first to cite angular acceleration as an important mechanism in brain injury (157). Gennarelli, Thibault, and colleagues, in a series of studies using live primates and physical models, investigated the role of rotational acceleration in brain injury (129; 128; 357; 277). They concluded that angular acceleration contributes more than linear acceleration to the generation of concussive injuries, diffuse axonal injuries, and subdural hematomas. More recent studies also support the role of angular and rotational acceleration in concussion, more specifically their propensity to induce shear and strain on cortical and brainstem structures (280). Employing basic physics, Newton’s second law (F = ma) applies to the linear acceleration component and outlines the basics of head impacts by stating that as the magnitude of impact force increases, the greater the acceleration of the head of the individual absorbing the force (370; 37). When looking at a rotational or angular acceleration situation, the formula T = I x A applies, where T equals torque, I equals movement of inertia, and A equals angular acceleration. To what extent each linear versus angular acceleration plays a role in concussion remains to be seen. In addition, the cervical spine plays a role in the transmission of force. Contracting the cervical spine prior to impact appears to transmit some of the force back to the body. As part of the NFL’s investigations on concussions, Viano and colleagues reported that increasing neck tension resulted in a 35% decline in concussion risk (374). The authors reported that neck contraction resulted in a lower injury risk because the head and torso were coupled. A number of studies have attempted to identify possible biomechanical mechanisms (281; 72; 37). Headshots or helmet hits were the most common mechanism; shots to the lower side of the helmet and oblique portion of the face mask induce translational and rotational forces on the head. Blindside hits to the body, which can induce lateral flexion of the neck and rotation of the head, have also been demonstrated to induce concussion (281). In addition to the acute effects of high-impact collisions, there also appear to be chronic effects of repeated lower impact collisions (291; 140).

Although there is scientific consensus that establishes angular or rotational acceleration of the brain as an accepted mechanism for concussion, there is still no consensus on concussion threshold. Several studies have been performed toward identifying thresholds for concussion. Ommaya and Gennarelli published a landmark study on cerebral concussion and traumatic unconsciousness in 1974 (276). The researchers devised an inertial loading apparatus that produced pure translational (linear) or rotational (angular) loading on the heads of primates without inducing contact phenomena (impact). Pure translation produced focal injuries (such as contusions) whereas diffuse effects (including concussion and subdural hematoma) were only produced when the rotational loading was present. Some continue to suggest thresholds based on linear accelerations. Ono published a summary of research regarding thresholds for human skull fracture based on cadaver experiments (277). The heads were suspended while a series of 42 frontal, 36 occipital, and 58 temporal blows were delivered, during which linear accelerations were measured. Frontal and occipital impacts shared a skull fracture threshold of 250 g for a 3 meter-per-second (msec) impact duration, decreasing to 140 g for a 7-msec impact duration; skull fracture threshold for lateral impacts was 120 g over a 3-msec duration, decreasing to 90 g for a 7-msec duration. Results clearly indicate that skull fracture threshold is inversely related to impact duration. Pellman and colleagues published a series of studies in which they performed biomechanical reconstructions of concussive impacts during NFL football games (281). The researchers reported that concussions were related to linear and rotational accelerations in the order of 98 + 28 g and 6432 + 1813 rad/s2, respectively, whereas subconcussive impacts averaged 57 + 22 g and 4029 + 1438 rad/s2. Their results show that a concussion is possible at 45g/3500 rad/s2, whereas 5500 rad/s2 represented a 50% risk of concussive trauma. A similar reconstructive study by McIntosh on nonhelmeted Australian rules football players demonstrated a 50% and 75% likelihood of concussion with linear acceleration of 65.1 and 88.5 g respectively, and angular acceleration of 1747 rad/s2 and 2296 rad/s2 in the coronal plane. The mean maximum angular acceleration to induce concussion was 7951 rad/s2. Overall significant differences in inducing concussion were seen between angular acceleration in the coronal and transverse plane, with the coronal plane more likely to induce concussion (244). Rowson and Duma have published extensively on their laboratory and field-based biomechanical evaluations of head injuries in football (310; 313; 312; 311; 73). Their findings are based on a large database of football impacts recorded by a Head Impact Telemetry System (HITS), developed by Simbex, Inc. (Lebanon, NH), which incorporates an array of uni-axial accelerometers in the crown of a player’s helmet. Based on 62,974 recordings of subconcussive impacts and 37 diagnosed concussions, a concussion threshold of 104 + 30 g and 4726 + 1931 rad/s2 was computed. A significant value in this research comes from the identification of thresholds for subconcussive impacts, the cumulative effects of which some believe to be highly correlated with early onset of cognitive dementias. The top 50% of subconcussive impacts in the HITS dataset consisted of impacts with average accelerations of 38 + 20 g and 1528 + 984 rad/s2, whereas the top 25% of subconcussive impacts consisted of impacts with average accelerations of 52 + 21 g and 2036 + 1124 rad/s2 (311). It should be noted that the thresholds for concussive impacts calculated by Pellman and Rowson exceed scientific consensus of thresholds for subdural hematoma in football impacts. Studies by Lowenhielm, who performed mechanical tests on human cadavers, concluded that bridging vein rupture occurs if the maximal angular acceleration exceeds 4500 rad/s2 (220; 221; 222). Depreitere, on the other hand, found that bridging vein rupture occurred at approximately 10,000 rad/second2 (for pulse durations shorter than 10 msec) (87); however, the threshold decreased as the duration of impact increased (6,000 rad/s2, but perhaps as low as 3,000 rad/s2 after adjusting for standard error) (203). Depreitere’s bridging vein rupture threshold is in line with epidemiologic evidence presented by Forbes and colleagues (247), in which they reported an appreciable risk of subdural hematoma associated with head impacts resulting in angular accelerations of 5,000 rad/s2. If Pellman’s and Rowson’s data were correct (ie, concussion threshold greater than those required for bridging vein rupture), the epidemiologic data should reflect an epidemic of subdural hematomas, which, of course, is not the case. Therefore, one must seriously question the validity of the findings of these studies.

In addition to angular and rotational forces, impact location appears to play a role in whether or not an individual sustains a concussion. Most concussions occur when a player sustains an impact to the front or side of the head. Head-down impacts were also more likely to induce concussion, and hits to the top of the head are more likely to induce loss of consciousness (180). Location, however, does not appear to predict symptom severity or outcome. Sustaining a concussion in a motor vehicle accident may result in a longer recovery time when compared to sustaining a concussion while playing sports (329).

Blast injuries suffered by our military heroes involve primary, secondary, and tertiary components (271; 30; 156; 22; 10; 16). The blast wave, which travels faster than the speed of sound, is the main determinant of the primary blast injury (Blast wave + Blast Wind = Injury) (85; 84; 210; Long and Bentley 2009; 54; 216). There are primary and secondary mechanisms of injury. Primary mechanisms include direct interaction with the head through direct passage of the blast wave through the skull or causing acceleration or rotation of the head resulting in concussion; intracerebral hemorrhage; subdural hematoma; edema; and diffuse axonal injury (85; 84; 210; Long and Bentley 2009; 393; 205). The thoracic mechanism includes the transfer of kinetic energy from the blast wave through large blood vessels in the abdomen and chest to the central nervous system causing both morphological and functional damage to distinct brain structures resulting in infarction or microhemorrhage (09; 85; 84; 49). Secondary mechanisms include vagovagal reflex leading to apnea followed by rapid breathing, bradycardia, and hypotension as well as reduction of heart rate (bradycardia) and dilatation of the peripheral blood vessels with resulting lowering of the blood pressure and cerebral hypoperfusion (88; 190). Injury of peripheral soft tissue and internal organs can lead to hemorrhage and synthesis and release of autacoids, which, when released in systemic circulation, activate the neuroendocrine immune system and contribute to initiation of secondary brain injury mechanisms (207; 03; 349; 299).

Subconcussive injury. There is an increasing amount of evidence that subconcussive hits or repetitive mild traumatic brain injury can result in neurocognitive issues later in life. It has been hypothesized that this may result from changes in excitatory neurotransmission with resultant chronic neuroinflammation and neurodegeneration (ie, chronic metabolic dysfunction leading to the activation of intracellular proteases and eventually cell death) (132). In addition, accumulation of abnormal proteins secondary to oxidative stress (ie, via early and prolonged proteasomal activation) also likely plays a factor. Animal studies have provided the greatest insight into the long-term effects of subconcussive hits. One such study used a rat percussion model to look at the effects of a single or repeated mild head trauma on neurophysiological function, neurocognitive function, and neuropathological outcome (13). Repeated mild traumatic brain injury caused substantial neuronal loss in the activated microglia 28 days after the initial injury in the ipsilateral and contralateral hippocampus as well as attenuation of the NMDA receptor-mediated response. Repeated mild traumatic brain injury also resulted in cognitive impairments in various maze tests. The authors concluded that repeated mild traumatic brain injury can result in deficits in hippocampal function and chronic neuroinflammation and neurodegeneration.

There have also been attempts to study subconcussive hits in human subjects (especially children) using helmet sensors to document the number of hits a player sustains either in a game or throughout a season. Daniel and colleagues used accelerometers to measure impacts in seven youth football players (77). Fifty-nine percent of hits occurred during practice with an average of 107 hits per player per season, with a higher magnitude of hits associated with practice than games. (ie, 63 during practice and 44 during game situations). Wong and colleagues used helmet sensors (ie, Shockbox) to follow 22 Pop Warner football players for six games and five practices; the players were 12 to 13 years of age (383). They measured impacts over 30 g and used video to look at player position, impact location (ie, line of scrimmage versus open field), type of hit, and whether or not the hit was to the head or not. Players sustained a total of 480 head impacts (109 in games and 371 in practice) with an average linear acceleration of 46.7 g. There were on average 21.8 head impacts per practice and 61.8 per game. Players had an average of 3.7 head impacts per game and 1.5 per practice. Eleven of the impacts were greater than 80 g with two concussions diagnosed over the course of a season. Those involved with in open-field plays had the highest magnitude of impact. It should be noted that the researchers used a 30 g threshold for detection, where other studies have used a 10 g threshold; this would explain the lower numbers in the Wong study. Finally, a third study followed three youth football teams and found in teams that limited hitting in practice there was a statistically significant reduction in both the frequency and magnitude of impacts. As a result, there have been rule changes both in Pop Warner and at the high school and college levels limiting the amount of contact during practice. Whether or not this results in a decrease in neurocognitive impairment later in life remains to be seen; longitudinal studies are underway.

Proposed biology and phases. The biology of concussion is poorly understood, and genetic variability, gender differences, and age are all likely involved in risk, severity, and recovery. What is known is based on rat percussion models and a few neuroimaging studies. It is, however, widely accepted that postconcussive deficits are based on temporary neuronal dysfunction (131; 16). There are various phases of concussion: acute, subacute, and late. These phases involve variations in ion movement, neurotransmitter release, energy utilization, blood-brain barrier dysfunction, axonal and cytoskeletal injury, neuronal plasticity, neuronal inflammation, and, at times, cell death. Alterations in cerebral blood flow also likely play a major role (209) and are likely associated with recovery. Symptomatic children have an increase in cerebral blood flow and those who recover quickly have decreased cerebral blood flow. This suggests that clinical recovery precedes cerebral recovery (17).

The acute phase is characterized by metabolic and ion derangements. Angular and rotational acceleration of the head result in shearing and stretching forces that cause a temporary disruption in plasma membrane, which is also known as mechanoporation. There is an initial outflow of intracellular potassium and subsequent diffuse neuronal depolarization. The latter results in abrupt and indiscriminate release of neurotransmitters and unchecked ion fluxes. Excitatory neurotransmitters such as NMDA and glutamate trigger neuronal depolarization with an efflux of potassium and influx of calcium. Increased extracellular calcium triggers further neuronal depolarization and further release of excitatory neurotransmitters and still further release of potassium into the extracellular space, ie, a feedback loop of depolarization and hyperexcitability. Normally excessive extracellular calcium is taken up by surrounding glial cells; however, this mechanism is overcome in concussion. The massive excitation is followed by a wave of neuronal suppression (ie, spreading depression) that is similar to what is seen in migraine with aura. Early loss of consciousness, amnesia, and other cognitive deficits may be a result of posttraumatic spreading depression.

Membrane pumps become activated in an effort to restore homeostasis, which results in increased glucose utilization. Increased glycolysis leads to increased lactate production, which results in neuronal dysfunction through processes such as metabolic acidosis, membrane damage, alterations in blood-brain barrier permeability, and cerebral edema.

The intermediate phase is characterized by an energy crisis, with uncoupling of glucose metabolism and cerebral blood flow. Increased lactate appears to serve as an alternative fuel for the brain during this crisis. Calcium influx, mitochondrial dysfunction, and delayed glucose hypometabolism also occur (131; 16). Uncoupling causes a 50% reduction of blood flow, which creates an energy mismatch and crisis. There is a biphasic recovery of oxidative metabolism with a reduction on day 1, recovery by day 2, bottom out by day 5, and complete recovery by day 10. Calcium accumulation can persist for 2 to 4 days, and the excess calcium is sequestered in mitochondria resulting in impaired metabolism and energy failure. Cerebral glucose use is diminished by 24 hours and can last up to 2 to 4 weeks’ postinjury, with the average recovery at 10 days. Cerebral glucose metabolism and oxidative metabolism correlate with the average concussion recovery time of 10 days. Intracellular magnesium levels are immediately reduced and remain for up to 4 days, and magnesium level recovery is correlated with improvement in motor function.

In the later stages of concussion, persistent calcium and sodium alterations, as well as neurotransmitter alteration, continue. Persistent elevations in intracellular calcium can lead to overactivation of enzymes and free radical production, resulting in cell damage (337). Cell death can also be a result of the aforementioned alterations; however, it is usually seen in more severe traumatic brain injury. Postconcussion alterations in NMDA, adrenergic, cholinergic, and GABA neurotransmission can result in long-term deficits in memory and cognition, even in the setting of minimal anatomic damage. Loss of forebrain cholinergic neurons can lead to learning and spatial memory deficits. Loss of GABA can result in disinhibition of hippocampal structures (ie, easy distractibility) and increase the risk of seizures (16). Changes in GABA and glutamate concentrations appear to be region specific (131).

Repeat concussions during the postinjury period, when the cell is most vulnerable, can have catastrophic consequences. In the first 30 minutes, when the system is stretched to its maximum, the brain may be unable to respond to a second stimulus-induced increase in cerebral glucose metabolism (131; 16). An increase in intracellular calcium after a second physiological stimulus can lead to protease activation and programmed cell death (366). Alterations in NMDA receptor composition can persist for 1-week postinjury, and a second injury in this period can lead to further impairment of excitatory neurotransmission. In addition to ion fluctuations and neurotransmitter dysfunction, cerebral blood flow is known to decrease immediately following both traumatic brain injury and mild traumatic brain injury and can remain lowered for extended periods of time. It has been hypothesized that alterations in CO2 levels may play a significant role in the regulation of cerebral blood flow (209). Following severe traumatic brain injury, cerebral autoregulation is either lost or impaired, and younger patients may have issues in cerebral reactivity, which in itself, is likely mediation by alterations in the brain’s metabolic activity. Furthermore, cerebral oxygenation is significantly reduced (up to a 35% decrease) on day 1 following mild traumatic brain injury and appears to be unresolved up to 7 days following the injury (67). Finally, it has been proposed that there may be neuroautonomic cardiovascular dysregulation, ie, an uncoupling between the autonomic nervous system and the cardiovascular system, following mild traumatic brain injury, which, in turn, could result in alterations in cerebral blood flow.

Epidemiology

Lack of a uniform definition and variability in ascertainment methods are major barriers in obtaining the true incidence of mild traumatic brain injury. Despite these limitations, large reviews and reasonable well-considered estimates have been undertaken. The Centers for Disease Control and Prevention (CDC) estimates that 1.6 to 3.8 million concussions occur each year in the United States. Other studies estimate the incidence to be approximately 600 per 100,000 (381). Furthermore, Meehan and colleagues found that almost one third of all athletes seen at their tertiary concussion clinic have had concussions that went undiagnosed (249). According to the CDC, children aged 0 to 4 years, older adolescents aged 15 to 19 years, and adults aged 65 years and older are most likely to sustain a traumatic brain injury (106; 14). Falls are the leading cause of mild traumatic brain injury (35.2%), followed by motor vehicle accidents (17.3%), sports and sports-related activities (16.5%), and assault (10%) (106). In the workplace, the greatest number of mild traumatic brain injury cases occurred from truck driving, construction work, and farm work (381). Other causes include assault, slip and falls, and walking into stationary objects. Mild traumatic brain injury is now the signature injury seen in our military heroes returning from Iraq and Afghanistan, with studies estimating that as many as 14.9% to 22.8% of soldiers returning from combat deployment meet the criteria for concussion. A majority of these mild traumatic brain injuries were related to blast injury (381). Based on evidence-based studies, for males, college and high school football have the highest concussion rate followed by ice hockey, rugby, lacrosse, and soccer (133; 308; 286; 14). Lower rates are seen in volleyball, baseball, and cheerleading (286). Overall, there is a significant variation in risk of concussion across all levels of play, with sub-elite players having the greatest risk (124). According to Gardner and colleagues, tackling appears to be the highest risk factor. In females, high school soccer and college ice hockey pose the greatest risk. The risk of concussion among female soccer players aged 11 to 13 years is 2.09 (345). Because of the greater number of male participants in sports studied, the total number of concussions is greater for males than females for all sports; however, evidence shows that females have a higher risk of concussion (14). There is variability in other areas of the world; for example, in Australia rugby has the highest incidence (3.01-4.73 per 1000 hours in match play, 0.07 per 1000 hours in practice in males and .55 per 1000 hours in match play in females). Another study found a higher incidence of concussion in Rugby 7 versus Rugby 15 (which is consistent with the above results) (115). In addition, athlete-specific characteristics such as body mass index greater than 27 kg/m2 and training time less than 3 hours weekly likely increase the risk of concussion. Conversely, age, level of competition, and position in most major team sports do not pose a higher risk of concussion. With that said, concussion risk is probably greater among linebackers, offensive linemen, and defensive backs as compared with receivers in collegiate football. Baugh and colleagues looked at 730 NCAA football players during the 2012 football season and found that offensive linemen were more likely to go undiagnosed after having suffered a concussion and were more likely to return to play while still experiencing symptoms of concussion (21). Furthermore, there is strong evidence indicating that a history of concussion or mild traumatic brain injury is a significant risk factor for additional concussions, and prior concussion exposure is highly likely to be a risk factor for chronic neurobehavioral impairment across a broad range of professional sports (333). There is moderate evidence indicating that a recurrent concussion is more likely to occur within 10 days after a prior concussion.

Looking at elite European soccer players, Nordstrom and colleagues found that suffering a concussion placed the athlete a significant risk in the first year of suffering any type of sports-related injury when compared to other players who did not suffer concussion (269). Predictors of severe or prolonged early postconcussion impairment include a prior history of migraine headache, prior history of concussion fogginess, alteration in mental status, younger age, lower level of play, and early amnesia (382). Carney and colleagues reviewed the literature from 1980 to 2012, ultimately including 231 studies (47). They found that the most prevalent and consistent indicators of concussion are disorientation or confusion immediately after the event, impaired balance within 1 day of the event, decreased reaction time within 2 days of the event, and impaired verbal learning and memory within 2 days of injury. Predictors of prolonged concussion symptoms include depression, anxiety, photophobia, phonophobia, multiple injuries, and forgetfulness or memory loss at time of presentation (42) as well as older age and female gender (184). Furthermore, APOE e4 genotype and a prior history of learning disorders is likely to be associated with chronic cognitive impairment after concussion exposure. In addition, a history of ADHD appears to be a risk factor for concussion as demonstrated in a study of NCAA division 1 athletes, with 50% reporting a history of prior concussion (06). Vagt and colleagues looked at average concussion recovery time among division II NCAA student athletes (371). They found the average recovery from concussion was 50 days, and a decrease in baseline SAC score was the only significant predictor of prolonged recovery time. Finally, according to Lee and colleagues there was no difference between college and high school athletes in symptom prevalence, severity, and total symptoms (204). There was also no difference in time to return to baseline. An increase in public awareness and media attention has led to the enactment of new guidelines, which has resulted in an increase in incidence, mainly as a result of increased reporting. Kilcoyne and colleagues looked at the combined concussion incidence in three NCAA division schools and one military school, which double to 0.57 per 1000 athletic exposures when new NCAA guidelines were published (181).

Prevention

There is really very little one can do to prevent mild traumatic brain injury other than immobilizing the neck, which, of course, creates other obvious issues. There is, however, one FDA "cleared" device: Q-Collar, which is worn around the neck and applies light pressure to the jugular veins. It works on the hypothesis that an increase in blood volume in the brain’s venous structures reduce movement when angular and rotational forces are experienced. This is based on rat percussion studies that demonstrated that rats with jugular vein compression experienced a 30% decrease in intracranial and intraocular pressure and an 80% decrease in amyloid precursor protein positive axons, which are a marker for traumatic brain injury (338). In three separate studies on male and female high school students, wearing the Q-Collar device resulted in statistically significant differences in controls, both acutely and in longitudinal studies in diffusion tensor MRI (ie, mean diffusivity, axial diffusivity, radial diffusivity, and fractional anisotropy) and functional MRI (blood oxygen level–dependent response ie, a reflection of neuronal activity), when compared to athletes who wore the collar, ie, the athletes who wore the collar had a significant decrease in injury risk (390; 391).

Efforts to minimize traumatic brain injury have been successful. Seatbelts, airbags, and infant seats have all contributed to moderating the effect of motor vehicle accidents. Surprisingly, helmets do little to prevent concussion; however, they do play a significant role in the prevention of skull fracture and brain bleeds (62). Mouth guards and soccer head gear do little to prevent concussion (133). Knowledgeable coaches and trainers can help protect participants in youth athletics by developing and implementing cervical strengthening programs, teaching proper tackling and hitting/checking techniques, and reinforcing the need for athletes to report their symptoms. In the elderly, physicians should screen patients regularly for gait disorders, avoid or use lower doses of medications with side effects that increase the risk of falls, provide and encourage the use of ambulatory devices that prevent falls (walkers, canes), and refer patients for gait retraining whenever possible.

Diagnostic workup

When it comes to concussion, the greatest focus has been in the attempt to develop diagnostic testing to evaluate individuals who are suspected of having suffered a concussion. A majority of the research has come from work with athletes. Testing ranges from simple sideline evaluation tools to complex neuroimaging studies. However, the best way to assess an athlete or any individual who has sustained a concussion is still a comprehensive neurologic history and detailed neurologic examination performed by a properly trained physician. The FDA has not yet approved any standalone medical products that are intended to specifically diagnose or treat mild traumatic brain injury; however, it has cleared (ie, it is safe to use but does not have sufficient studies to make an accurate diagnosis) Eye Box, King Devick, and IMPACT for mild traumatic brain injury. Also on the market (however not cleared) are Trazer (looks at balance, vision, and reaction time and tests the individual in various stages of exercise, with potential use in treatment as well), Eye Sync, XLNT, balance systems such as Cobalt (uses a foam pad), Biosway and Biodex (which use computerized sensor based technology along with a foam pad), and Sway Balance (which uses a phone app). As a result, in April 2019 the FDA issued a statement warning the public not to use medical devices marketed to consumers that claim to help assess, diagnose, or manage head injury, including concussion, traumatic brain injury, or mild traumatic brain injury. Furthermore, the agency warned that “such tools — such as apps on a smartphone marketed to coaches or parents for use during sporting events — have not been reviewed by the FDA for safety and efficacy and could result in an incorrect diagnosis, potentially leading to a person with a serious head injury returning to their normal activities instead of getting medical care”.

Sideline and battlefield concussion assessment tools. The most widely used sideline evaluation tool in sports is the is the Sideline Concussion Assessment Tool, SCAT2 /updated SCAT3, which is also available in a version for children under 12 years of age.

Changes from SCAT2 to SCAT3 include initial assessment of injury severity using the Glasgow Coma Scale (GCS), immediately followed by observing and documenting concussion sign, and waiting 15 minutes prior to conducting neurocognitive and balance testing (241). Changes from the SCAT3 to SCAT5 include:

	• Declaration that the complete SCAT5 cannot be appropriately completed in less than 10 minutes.
	• Inclusion of an Immediate/Acute Assessment section, including indications for emergency management and observable signs of possible concussion.
	• Clarified instructions that the Symptom Checklist should be completed by the athlete in a resting state.
	• Different instructions for completing the symptom checklist at baseline and postinjury have been added.
	• Addition of questions that compare the athlete’s postinjury presentation with preinjury behavior.
	• The SAC immediate and delayed word recall lists include an option to use 10 words instead of five to minimize ceiling effects.
	• All six versions of the SAC word lists are now presented with alternate stimulus sets for the word list and digits backwards. Their administration should be randomized at baseline and serially postinjury.
	• A notation of when the last trial of the word list was administered is required (the delayed recall should not be administered sooner than 5 minutes after the immediate memory subtest).
	• Digits Backwards now contains six versions of the digit strings, which should be randomized at baseline and serially postinjury.
	• A Rapid Neurologic Screen has been included.
	• A section has been added that includes affirmation that the SCAT5 was used or supervised by a healthcare professional and whether a concussion was diagnosed.
	• The Instruction section has been enhanced to include all of the modifications described above.
	• The Return to Sport progression emphasizes that the initial period of physical and cognitive rest should typically only last 24 to 48 hours.
	• A Return to School progression has been added, including possible academic accommodations.
	• The SCAT5 specifically indicates that written clearance by a healthcare professional is necessary prior to returning to play/sport.

The test is usually given at the start of the season as a baseline and repeated if an athlete sustains a concussion; however, it can be used at any time. The test has many of the same components of the standard neurologic exam, with eight subsections, including symptoms, physical signs, Glasgow Coma score, Maddock score, cognitive assessment, balance assessment, coordination, and a standardized concussion assessment (Glasgow Coma score and Maddock score do not apply if using the test as a baseline assessment). Many concussion experts do not use the Glasgow Coma Scale because loss of consciousness is present in less than 10% of all athletes. Portions of the test have not been validated, and, therefore, SCAT should never be used as the sole measure to return an athlete to play, or as a substitute for a physician-performed neurologic evaluation (241). The test is also not likely appropriate for use in the office setting (133). Overall, the most accurate component of the SCAT appears to be the symptoms assessment component (121).

Other tests include the Standardized Assessment of Concussion and Maddock Scale, both of which assess orientation, immediate memory, concentration, and delayed recall. The Standardized Assessment of Concussion also assesses neurologic function and symptoms on exertion (114).

In 2006, the Department of Defense developed the Military Acute Concussion Evaluation (MACE) for use in assessing soldiers (114). The test has both a history and evaluation component. The history component can confirm the diagnosis of mild traumatic brain injury after establishing that trauma has occurred. The evaluation component, designed to be easily used by medics and corpsmen, can be administered within 5 minutes. It utilizes the Standardized Assessment of Concussion to preliminarily document neurocognitive deficits in four cognitive domains: orientation, immediate memory, concentration, and memory.

Computerized testing. Most professional, college, and some high school sports teams use computerized testing such as IMPACT, Headminder Concussion Resolution Index (CRI) (New York, NY), and CogSport (CogState, Melbourne, Australia). The military uses the Automated Neuropsychological Assessment Metric (ANAM) (Center for the Study of Human Operator Performance, The University of Oklahoma, Norman) and CNS Vital Signs. Athletes and soldiers are tested by a trainer or neuropsychologist prior to the start of the season/deployment and then during the season or during deployment if a concussion is reported. Despite being validated with formal neuropsychological testing, computerized testing should never be used as the sole means for returning an athlete/soldier to play/combat, and it should never be substituted for a comprehensive history and neurologic examination. Most tests measure multiple aspects of cognitive functioning, including attention span, working memory, sustained and selective attention, response variability, and nonverbal problem solving. A study looked at four of the more common computerized tests (ANAM4, CNS Vital Signs, CogState, and ImPACT) and compared them to standard neuropsychological testing. There was significant variability and poor convergent and discriminative validity between neuropsychological testing and computerized testing (61). Issues with computerized testing include lack of initial effort (ie, preseason) whereby the athlete/soldier will actually score better after sustaining a concussion, inability to address associated symptoms (ie, headache, sleep disorders), and marginal reaction time testing that does not simulate game or battlefield conditions (166; 296). The tests are also influenced by symptoms such as headache and lack of sleep, which are very common with concussion (297; 71; 303; 307). A study by Sandel and colleagues found that a history of migraine has no influence on computerized neuropsychological testing (319). Furthermore, the tests do not appear to accurately reflect the metabolic recovery of the injured brain, as most high school and collegiate athletes displayed functional/cognitive recovery at approximately 6 days following a sports-related concussion, whereas professional athletes displayed recovery within 3 days (69). This is inconsistent with physiological recovery, which takes between 14 and 28 days (166). Finally, as the tests are often administered on consecutive days, improvements in concussed athletes may be partially associated with learning effects and not with injury recovery (70; 19; 142).

Neuropsychological testing. Formal paper and pencil neuropsychological testing is the best method to evaluate cognition in individuals who have sustained concussion. The tests have been validated in studies with nonathletes (183; 252). Surprisingly, there are very few studies in concussed athletes (259; 182; 219). Due to cost, length of testing, and access, neuropsychological testing is mainly used for individuals with prolonged cognitive symptoms. Testing tends to focus on assessment of sustained and divided attention, reaction time, visual and auditory processing speed, and working memory. Intelligence, problem solving, and language are also tested.

Balance, agility, and reaction time testing. After concussion, communication between sensory systems is lost in the majority of individuals, causing moderate to severe postural instability in either the anterior-posterior direction, medial-lateral direction, or both (139; 137; 138; 252; 309; 159; 264; 265; 316). Decrease in postural stability persists for up to 3 days following injury, after which time the individual gradually recovers to approximately the scores of matched control subjects by day 10 postinjury (137; 138; 145). Vestibular deficits are related to problems with sensory integration, whereby the concussed individual fails to use their visual and vestibular systems effectively. Studies have demonstrated difficulties with task switching in concussed athletes (138; 102; 161; 162; 202; 353; 66). A number of tests are currently available or in development, including the Balance Error Scoring System (BESS), Sensory Organization Testing (SOT), Gait Testing, Virtual Reality Testing, Instrumented Agility Task Testing (137; 138; 237; 56), and CaneSense. The latter uses what is called the postconcussion excursion index (PCEI) as a measure of postural stability and is expressed as a percentage of symmetry between the lower limbs. A study compared CaneSense to the BESS and found significant differences in PCEI between concussed individuals when compared to their baseline. In addition, the BESS did not detect these differences and in fact there was actually a slight improvement in BESS performance postconcussion (108). Another study by Purkayastha and colleagues measured center of pressure (ie, postural stability) using forced plate technology with eyes open and closed (294). They found that postural sway normalized in the concussion group by day 21 whereas the postural complexity index did not and persisted beyond day 21.

In addition to the complex tests listed above there are components of the neurologic exam including Rhomberg (ie, sway), finger-nose-finger, fine finger testing, rapid alternating movement testing, and tandem gait alone or in conjunction with a task, ie, dual task. Evidence has shown adolescents with concussion perform poorly on dual task tests, especially speed. This has led to the suggestion that tasks requiring greater motor coordination are likely more effected in concussion (163).

Eye movement testing. Frontal-parietal and subcortical nuclei are quite susceptible to concussion. These circuits are prone to neurophysiological changes resulting in visual symptoms, oculomotor dysfunction, and deficient cognitive control of eye movement (373; 60). Individuals with concussion often have eye-movement-specific neurologic exam findings, including abnormalities with smooth pursuit, convergence, accommodation, photosensitivity, vestibular-ocular reflex, and saccades. Tests of saccadic eye movement have the ability to assess attention, memory, and executive function. Unfortunately, they require computer and video-ocular test apparatus and are not practical in the clinical setting. Smooth pursuits and convergence abnormalities are much more practical and evaluate attention and memory. Approximately 60% of individuals with concussion have difficulties with smooth pursuits, and 47% to 64% have abnormalities on convergence (44). Combining tests with other measures of neurocognitive function have shown that elite male and female hockey players with a near point fixation disparity equal to or greater than 15 cm, comprehension rate less than 85%, and abnormalities on an ADHD questionnaire were on average 10.72 times more likely to have suffered a concussion (288). In a second study, a combination of vestibular/ocular motor screening (VOMS) tests was highly correlative to a concussed athlete’s postconcussion symptom score, with 61% reporting provocation of concussion symptoms after undergoing at least one VOMS test and the vestibular ocular reflex being most predictive (260). Galetta and colleagues developed the King-Devick test (K-D) to assess eye movement abnormalities in concussed athletes (116; 117). Unlike computerized testing, which takes over 20 minutes, the K-D test takes approximately 1 to 2 minutes and requires a subject to rapidly read a set of numbers on three sets of test cards. The sum of the 3-time scores and the number of errors are scored, with the baseline score being the total time in seconds and number of errors. The test is hypothesized to measure impairment of eye movements, attention, and language and mimics anticipatory saccade testing and evaluates the dorsolateral prefrontal cortex, brainstem, cerebral cortex, and cerebellum (ie, areas involved in attention, language, and reading) (373; 168; 116; 117). Galetta and colleagues initially administered the test to a cohort of boxers (n=27) and mixed martial arts fighters (n=12), pre- and post-fight (116). Post-fight K-D time scores were significantly higher (worse) for participants who had head trauma (59 vs. 41 seconds), and those with loss of consciousness had higher post-fight K-D scores than those without loss of consciousness (65.5 vs. 52.7 seconds). Abnormal post-fight scores also correlated with abnormal MACE scores. The authors suggested that the test could serve as a rapid sideline screening test. The study has not been fully validated, and there are issues with the timing of the post-testing, ie, including the time immediately after the fight, when the athlete is physically fatigued and dehydrated. Fatigue and dehydration are well-known to affect cognitive function and cognitive testing (238). In addition, giving the test immediately after the fight, when the athletes’ motivation may not be as high as pre-fight and there are significant external distractions, can affect outcome. Further studies by the same group compared the utility of the King-Devick with that of the SCAT-2 and SAC in 27 professional ice hockey players (119). On baseline testing, lower scores on the SAC and SCAT-2 correlated with worse times on the K-D. Two athletes were assessed immediately after concussion and had worsening of their K-D scores by 4.2 and 6.4 seconds respectively. A second study by the same group looked at youth and collegiate athletes (n=332) with the hypothesis that adding the K-D to the SAC and a timed tandem gait task would increase the ability to detect concussion (118). Baseline K-D tests were better (ie, lower times) with increasing age. In the 12 athletes who sustained concussion, K-D scores increased by an average of 5.2 seconds, and the probability of detecting concussed versus nonconcussed athletes was 92% with the K-D. Others have compared the K-D with computerized neuropsychological testing (IMPACT) and the postconcussion symptom scale (PCSS). Thirty-five concussed individuals (no baseline K-D) were evaluated during four or more visits (360). During that time, improvement in K-D scores paralleled improvement in PCSS and IMPACT scores. Silverberg and colleagues compared K-D and SCAT-2 in the emergency room setting (334). Twenty-six individuals with concussion and 33 controls were evaluated; however, no differences in K-D scores were seen. There were statistically significant changes on the SCAT-2. Keeping with measures of rapid automatized naming, a newer test, The Mobile Universal Lexicon Evaluation System (MULES), uses rapid picture naming to assess concussion. It uses a series of pictures, ie, 54 grouped photographs of fruits and animals administered on an 8.5 x 11–inch double-sided laminated card. The test integrates saccades, color perception, and contextual object identification. The test was studied in six athletes with concussion who underwent preseason baseline testing. All athletes showed worsening of their MULES scores from baseline after concussion. (05).

Video-oculography is another emerging technology that looks at eye tracking and gaze tracking. Patients with concussion show increased saccadic latencies, saccadic dysmetria, errors in predictive target tracking, and changes in vergence (04). Another study compared 64 normal, healthy, noninjured control subjects and compared the findings to 75 trauma subjects, ie, positive head CT scan (n = 13), negative head CT (n = 39), or nonhead injury (n = 23), to determine whether eye tracking would reveal the disconjugate gaze associated with both structural brain injury and concussion (318). Tracking metrics were then correlated to the clinical concussion measure Sport Concussion Assessment Tool 3 (SCAT3) in trauma patients. Five out of five measures of horizontal disconjugacy were increased in positive and negative head CT patients relative to noninjured control subjects. Only one of five vertical disconjugacy measures was significantly increased in brain-injured patients relative to controls. Horizontal disconjugacy negatively correlated with SCAT3 symptom severity score and positively correlated with total Standardized Assessment of Concussion score. Abnormal eye-tracking metrics improved over time toward baseline in brain-injured subjects observed in follow-up. Eye tracking may also help quantify the severity of ocular motility disruption associated with concussion and structural brain injury. “Eye box,” which involves having the patient watch a short video while their eye movements are monitored, has received FDA clearance. Initial studies compared 30 concussed participants (age: 14.4 ± 2.2 years, mean ± SD, 50% female) to 30 controls (age: 14.2 ± 2.2 years, 47% female) (272). Subjects completed eye tracking and gait assessments. Symptoms were collected using the Post-Concussion Symptom Scale (PCSS), and gait speed was measured with triaxial inertial measurement units. Eye movements were evaluated using a "BOX score," a metric of pupillary disconjugacy, with scores below 10 classified as normal and scores of 10 or more as abnormal. There was a significant association between total PCSS score and BOX score in the concussion group (β = 0.16, p = 0.004, 95% confidence interval [CI]: 0.06‒0.27) but not in the control group (β = 0.21, p = 0.08, 95% CI: –0.03‒0.45). Overall, the concussed group with impaired eye tracking reported higher total symptom severity as well as worse symptom severity across the five PCSS symptom domain profiles. This technology also appears to be useful in separating acute concussion from chronic mTBI. Jain and colleagues used the technology to compare 180 uninjured adolescents to 224 concussed adolescents with acute or persistent symptoms (170). Concussed adolescents had larger left and right mean, median, minimum, and maximum pupil size than uninjured controls. Concussed adolescents had greater differences in mean, median, and variance of left and right pupil size. Twelve metrics distinguished female concussed participants from uninjured controls.

Others have used the Nintendo WiiFit soccer-heading game and a monocular eye-tracking device to assess athletes with concussion. They found significant gaze deviations from center and negative correlation between gaze and balance in post-concussed athletes (264; 265). Finally, combining ocular testing may improve concussion diagnosis. One study looked at acutely concussed military personnel (n=100) combining King-Devick, pupillary light reflex, near point convergence, and convergence insufficiency symptom survey. Concussed individuals showed statistically significant difference between pre- and post-concussion (43).

Electrophysiological testing. Mild traumatic brain injury results in electrophysiologic abnormalities visible in electroencephalography (EEG) recordings. However, EEG can reveal significant differences in brain activity up to 12 months, suggesting postconcussive electrophysiological deficits last much longer then the clinical recovery (263). There appears to be no significant risk of epilepsy during the first 5 years after injury (380). Quantitative EEG (QEEG) makes use of quantitative techniques to analyze EEG characteristics such as frequency, amplitude, coherence, power, phase, and symmetry over time independently or in combination (144). QEEG allows for identification of subtle changes or tendencies in the patterns of source EEG data (196) and can be predictive of long-term working memory deficits (278). It may also be useful in the identification, tracking, and classification of individuals with mild traumatic brain injury (208). Two consistent electrophysiological trends seen with QEEG include the reduction in mean alpha frequency and an increase in theta-alpha frequency ratios. QEEG has been shown to be a sensitive indicator of the presence of brain injury after mild head trauma (352; 290; 289). QEEG features appear to be sensitive for postconcussion symptoms and can predict recovery of function at 1-year postinjury (94). It may also be a sensitive indicator of brain dysfunction after mild head injury due to blast concussion (365). McCrea and colleagues used QEEG in a prospective, nonrandomized study of 396 high school and college football players, which included a cohort of 28 concussed athletes and 28 matched controls (238). All underwent preseason baseline testing, including postural stability testing, cognitive testing, and QEEG. They concluded that the duration of physiological recovery may last longer than observed clinical recovery. Dupuis and colleagues found that concussed athletes had a decrease in the P300 amplitude and concluded that this may reflect alterations in attention and concentration (95). Finally, Donaldson and colleagues (using QEEG and a CNS questionnaire) found that F8 and T6 (right side of the head) and O1 and O2 (back of the head) were the most common sites injured in a small cohort of Canadian junior hockey players, indicating that QEEG could be used to predict the probability and severity of concussion (92).

Biomarkers. In February of 2018 the FDA approved the first blood test to detect traumatic brain injury. The Banyan BTI (brain trauma indicator) measures levels of two protein biomarkers -- ubiquitin carboxy-terminal hydrolase-L1 and glial fibrillary acidic protein -- that are released from the brain into blood within 12 hours of head injury. Levels of these blood proteins after traumatic brain injury can help predict which patients may have intracranial lesions (especially intracranial and other types of brain bleeds), thus, negating the need to screen via a CAT scan. The test is not able to diagnose concussion. The biotechnology industry and military are currently looking at a number of biomarkers to measure mild traumatic brain injury, both on the field/battleground and in the clinic. Apolipoprotein E (APOE), APOE promotor gene, neurofillament light (NF-L), total tau, microtubule associated protein (MAT)/tau exon 6 Ser53Pro, MAPT/tau Hist47Tyr, IL-6RASP358ALA, alpha synuclein, serum microRNA, catechol-o-methyltransferase (COMT), dopamine D2 receptor (DRD2) (ANNK1 gene), interleukin p53, angiotensin converting enzyme (ACE), CACNA1A, and SB100 have been or are currently being evaluated. The two most widely studied are apolipoprotein and SB100, with the latter showing minimal efficacy in detecting concussion in sports (26; 306; 134; 188; 133; 317). Apolipoprotein has been looked at in a number of studies in both athletes and nonathletes and has been correlated as a marker for chronic injury. The largest looked at 1056 college athletes and found those with an APOE-4 polymorphism were 40% less likely to suffer a concussion whereas those with IL-6R had a 3 times greater concussion risk (354). Zhou and colleagues performed a meta-analysis and looked at 14 cohort studies (395). There was no correlation of the ε4 allele with initial injury severity. However, the ε4 allele was associated with poorer outcome at 6 months after injury. Terrell and colleagues looked at 195 college athletes, mainly football and soccer players (355). The cross-sectional study investigated the association between APOE, APOE promoter, and tau polymorphisms and a self-reported history of concussion over a prior 8-year period. There was a 3-fold increase in risk of concussion in those with the TT genotype of G-219T polymorphism relative to the GG genotype, and a 4-fold increased risk in those a with self-reported history of concussion associated with loss of consciousness. There was, however, no association with APOE or tau genotypes. In a prospective cohort study of 318 collegiate athletes in various sports, Kristman and colleagues compared concussion rates in athletes with and without APOE ε4 allele (193). They found no association between ε4 allele and sustaining a concussion. Finally, Tierney and colleagues looked at 196 college athletes (163 male football and 33 female soccer players) in a multicenter cross-sectional study evaluating the association of carrying one or more APOE rare (or minor) alleles (APOE ε2, APOE ε4, and T allele of G-219T APOE promoter polymorphism) and a self-reported history of concussion (359). Athletes carrying all three rare alleles were 9.8 times more likely to report a previous concussion. Athletes carrying the T allele of the APOE promoter gene were 8.4 times more likely to report multiple concussions, and the authors concluded that carriers may be at greater risk for multiple concussions. Preliminary results from a large multicenter NCAA study where athletes underwent blood sampling throughout the course of a football season demonstrated increases in concentrations of NF-L throughout the season as the number of head impacts increased whereas levels of tau decreased over the course of the season (273). Similar findings were seen in a study by Rubin and colleagues, who found the frequency and magnitude of head impacts were associated with increases in NF-L and s100B levels but not tau (314). In addition, changes seen did not correlate with the number of years of football experience or concussion history.

Others have looked at alpha synuclein with a small study, demonstrating that those with lower levels experienced increased concussion symptoms and depression (386). Finally, one of the newer players, microRNAs, ie, tiny epigenetic molecules that are expressed throughout the body, can cross the blood-brain barrier and are transported from cell to cell where they regulate gene expression (12). In a study of NCAA division 1 college players, athletes with the lowest SAC score had the highest levels of microRNAs. Athletes with declining neurocognitive function over the course of the season had increases in microRNA concentrations. Subtypes of microRNAs also correlated with balance issues and decreased reaction time (279).

A more recent study looked at autoantibodies in the saliva, ie, immunoglobulin A (IgA) autoantibody, of 167 male and female athletes 24 to 48 hours post-concussion or who had sustained subconcussive hits (287). There was a significant increase in the prevalence of IgA toward protein fragments representing 5-hydroxytryptamine receptor 1A (HTR1A), serine/arginine repetitive matrix 4 (SRRM4), and FAS (tumor necrosis factor receptor superfamily member 6) after concussion and subconcussion. The authors suggest that concussion and subconcussion induce similar physiological effects, especially in terms of immune response, and saliva IgA autoantibody has a potential use as a biomarker for concussion and subconcussive hits.

Neuroimaging. Individuals who sustain concussion do not require routine imaging. Exceptions include those with loss of consciousness, increasing lethargy, and focal neurologic findings on their neurologic exam. If the individual requires imaging, CT and conventional MRI are both useful in detecting intracranial and subdural bleeds; however, they are usually without findings. Conventional MRI with gradient echo is also useful in detecting microbleeds and diffuse axonal injury in more severely injured patients and conventional MRI with arterial spin labeling can show changes in cerebral blood flow. More advanced imaging such as positron emission tomography (PET), functional MRI (fMRI), magnetic resonance spectroscopy (MRS), and diffusion tensor imaging (DTI) hold the most promise for quantitative assessment of sports-related concussion. Combining techniques may add to detection and improve diagnostic accuracy. One such technique, magnetoencephalography, which in essence combines MRI and EEG, detects the origin of brain activity and allows for the measurement of ongoing brain activity on a millisecond-by-millisecond basis. Ideally, imaging needs to provide quick, reliable, and longitudinal capabilities, and be easily employed in the community setting.

Conventional MRI. Helmer and colleagues used susceptibility-weighted imaging (SWI) to look for micro-bleeds (ie, small hypointensities that may signal chronic small vessel damage both in concussive and subconcussive injury) in 45 university-level male and female hockey players before and after a single season (153). Eleven subjects sustained concussion and were imaged at 72 hours, 2 weeks, and 2 months. A statistically significant increase in hypointensity burden relative to baseline was observed in the male subjects at 2 weeks. There was no statistically significant increase in burden for nonconcussed individuals; however, nonconcussed males had a higher overall burden when compared to females. In a similar, study Lannsjo and colleagues used conventional MRI, SWI, DWI, and FLAIR imaging as well as computer-aided volume analysis and compared the results to various clinical outcome scales (198). Nineteen patients underwent MRI, postconcussion assessment questionnaires, and depression scales at 2 or 3 days and at 3 months. One patient had positive MRI findings at initial evaluation (hippocampal edema, T2 hyperintensity, DWI abnormalities) with loss of hippocampal volume at follow up. Three other patients had volume abnormalities at follow up. Other studies have used voxel-based MRI to assess patients with mild traumatic brain injury and traumatic brain injury (214). Significant differences if the gray and white matter (atrophy in the right internal capsule and right ventrolateral prefrontal cortex) were found between patients 6 months after suffering a traumatic brain injury and controls. Those with a history of multiple mild traumatic brain injuries showed decreased density in the temporal lobes, parahippocampal gyri, prefrontal cortex, external capsule, and cerebellum. An emerging MRI technique, muticomponent driven equilibrium single pulse observation of T1 and T2 (mcDESPOT), which looks at acute and chronic white matter changes (ie, myelin water fraction), was used to study college football and rugby players 3 months postinjury (all were asymptomatic). When compared to a control group of noncontact sports athletes, the contact sport players had significant increases in myelin water fraction in the basal ganglia and deep white matter regions, which suggests that acute and chronic head injury can result in alterations in white matter integrity possibly secondary to axon neuropathy versus disorderly remyelination/hypermyelination (339). Preliminary results from the ongoing multicenter NCAA/DOD project of the natural history of concussion using arterial spin labeling MRI showed decreased cerebral blood flow in several brain regions that correlated with tests of cognitive function and balance (376).

Evidence has suggested that the hippocampus is vulnerable to both traumatic brain injury and mild traumatic brain injury (362). Singh and colleagues looked at 25 high school players with concussion and 25 players without as well as 25 age-matched controls (335). They found players with or without concussion had smaller hippocampal volumes when compared to controls. In addition, players with concussion had smaller hippocampal volumes when compared to those without concussion. Tremblay and colleagues used quantitative neuroimaging (voxel-based MRI, hippocampal volume, and cortical thickness proton MRI) as well as APOE and neuropsychological testing to assess former athletes who sustained their last concussion more than thirty years prior to testing (363). Concussed individuals demonstrated enlarged lateral ventricles (which correlated with trouble with episodic memory), cortical thinning (frontal, temporal, and parietal regions), and neurometabolic abnormalities in the medical temporal lobes when compared with controls who had never suffered a concussion.

Positron emission tomography (PET). It is widely accepted that a single severe traumatic brain injury increases the risk of Alzheimer disease with an increase in CNS amyloid (256). PET is useful in measuring quantitative brain glucose uptake and regional oxygenation and, therefore, can demonstrate metabolic disturbances after brain injury (01). There are little data on athletes who have sustained concussion. One study looked at 19 boxers and eight normal controls (292). The study demonstrated hypometabolic areas, ie, decreased glucose uptake in the bilateral posterior parietal lobes that extended to the lateral occipital lobes, bilateral frontal lobes, bilateral cerebellar hemispheres, and posterior cingulate cortex. There are also a few studies in patients with mild traumatic brain injury that demonstrated correlation between metabolic dysfunction and neuropsychological performance (29; 284). Conversely, athletes with multiple concussions (mild traumatic brain injury) and possibly subconcussive hits appear to increase the risk of early dementia characterized by the accumulation of Tau protein in the CNS, ie tauopathy. Three PET tracers; florbetaben F18 (beta-amyloid neuritic plaque density), florbetapir F 18 (beta amyloid), and flutemetamol F18 (amyloid) are approved to detect Alzheimer disease. In addition, new PET ligands FDDNP and F-T807are being investigated to detect Tau.

A study utilized FDDNP PET and neuropsychological testing to evaluate five retired NFL players (336). The retired players had increased PET signals in the subcortical brain regions and the amygdala when compared to five age-matched controls. A similar study by Mitis and colleagues used florbetapir and F-T807 PET imaging to assess a retired NFL player with multiple concussions and cognitive decline similar to Alzheimer disease as well as a 59-year-old physician who suffered a traumatic brain injury 10 months prior to evaluation and was experiencing cognitive decline and personality changes (256). Florbetapir imaging was negative and F-T807 imaging was positive (globus pallidus and substantia nigra) in the NFL player who was also noted to have a decrease in hippocampal volume on MRI. His neuropsychological testing demonstrated decreased processing speed, fine motor function, verbal comprehension, naming, and immediate and delayed verbal recall. His overall presentation was consistent with chronic traumatic encephalopathy. The physician with traumatic brain injury demonstrated a negative florbetapir study for amyloid and increased uptake of F-T807 in the site of impact, with evidence of an old frontal subdural hematoma and atrophy in the frontal and temporal lobes on conventional MRI. His neuropsychological testing demonstrated significant deficits in memory, language, and executive function. His overall clinical and radiological presentation was felt to be consistent with frontal temporal dementia. Potential disadvantages include study duration, arterial sampling for quantitative studies, lack of available units in the community, and the need to produce isotopes onsite or locally. The most important disadvantage is that in patients with mild traumatic brain injury, PET imaging studies appear to remain abnormal after the recovery phase. A more recent study, also involving NFL players (n=14), attempted to measure translocator protein 18 kDa (TSPO), a marker of activated glial cell response (68). The NFL players showed higher total distribution volume in 8 of the 12 brain regions examined (P <  .004) as well as a change in white matter fractional anisotropy and mean diffusivity in 13 players compared with 15 control participants. The results suggest that localized brain injury and repair, indicated by higher TSPO signal and white matter changes, may be associated with NFL play.

Magnetic resonance spectroscopy (MRS). MRS uses metabolite data from regions of the brain to provide an assessment of neurochemical alterations after brain injury. Metabolites typically include N-acetylaspartate NAA (neuronal specific metabolite and a marker for neuronal health), myoinositol (glial marker), choline (marker of inflammation), lactate (indirect marker for ischemia and hypoxia), and creatinine and phosphocreatinine (stable brain metabolite and marker of cellular energy status). The technology is highly experimental and there are issues with consensus on the optimal protocol (185). It does, however, have the potential to provide a better understanding of the underlying neuropathology and neurophysiology of acute mild traumatic brain injury, to more accurately detect and diagnose mild traumatic brain injury, and to predict those individuals who are more likely to suffer long-term disability (209). Studies looking at acute mild traumatic brain injury have shown biochemical changes, including decreased NAA, reflective of neuronal injury, increased Cr and Glucose in white matter, and decreased glucose in gray matter, reflective of energetic and neurotransmitter disturbances. Ultimately, longitudinal MRS studies can also be used to measure injury outcome or monitor therapeutic response. One such example by Dean and colleagues used MRSPECT to look at eight patients with persistent postconcussion symptomatology for more than 12 months (82; 83). They compared patients with mild traumatic brain injury with and without postconcussion symptomatology and patients with no head injury. They found a decrease in the creatine/choline ratio in both groups with mild traumatic brain injury, concluding that patients with mild traumatic brain injury may have persistent metabolic changes for up to 12 months. Another interesting study utilized 3-dimensional proton MRS (H-MRSI), which is highly specific for cell status, neuronal integrity, cellular energy, membrane turnover and astro-glial proliferation (186). The technology utilizes NAA, creatine, choline and myoinositol as biomarkers that are highly sensitive in detecting changes in the brain’s gray and while matter. The results demonstrated a decrease in global white matter NAA in concussed individuals when compared to controls. The findings lead the researchers to conclude that lower white matter NAA levels were a sign of neuronal dysfunction and not cell death and is consistent with the proposed pathophysiology of concussion. Studies in athletes by Vagnozzi and colleagues used proton MRS to evaluate 11 concussed athletes and 11 age- and sex-matched controls at 3, 15, 30, and 45 days post-injury (369). Concussed athletes demonstrated an increase in the NAA/creatine ratio at 3 and 15 days postinjury, which was followed by a decreased and subsequent normalization by day 30. In addition, NAA/choline levels were decreased at 3, 15, and 30 days after concussion when compared to controls, with no difference observed at 45 days. Finally, there were significant differences in the choline/creatine ratio at 3 days, with no differences at 35 and 45 days between concussed individuals and controls. Taken together, the results indicate that brain metabolism appears to normalize within 30 to 45 days after concussion. Finally, Raji and colleagues evaluated the clinical utility of single photon SPECT in mild, moderate, and severe traumatic brain injury (295). A systematic literature review and meta-analyses demonstrated 19 longitudinal and 52 cross-sectional studies. In the longitudinal analysis, SPECT was superior (level IIA evidence) to CT and MRI in detecting lesions. There was also a correlation between SPECT and neurologic and neurocognitive outcome as well as treatment outcome. They concluded that with newer, more advanced SPECT techniques, SPECT should be able to identify subacute and chronic traumatic brain injury and that SPECT holds the potential as a biomarker for assessing the effectiveness of treatment. Perhaps the most important aspect of the analysis was SPECT’s near 100% negative predictive value. More recent studies have looked at retired NFL players, female rugby players, high school contact sport athletes, and military personnel with persistent concussion symptoms who were treated with hyperbaric oxygen. In retired NFL players, a higher incidence of repetitive head injury correlated with reduced cellular metabolism (ie, lower creatine) and MR-SPECT chemicals associated with neuroinflammation correlated with behavioral symptoms (07). In female rugby players combined evaluations with MR-SPECT and DTI MRI pre- and post-season demonstrated reduced glutamine immediately after and at 3 months after concussion and reduced glutamine/creatinine at 3 months. In nonconcussed athletes decreased fractional anisotropy and radial diffusivity (which are consistent with possible remyelination) correlated with changes in glutamine and glutamine/creatine, suggesting decreased oxidative metabolism (322). In addition, a group of high impact high school athletes (male football and female soccer players) demonstrated marked deviations in neurometabolism in the absence of any significant symptoms, suggesting that subconcussive impacts may have unrecognized consequences on neuronal health (15). Finally, active-duty or veteran military personnel and normative controls underwent MRS outcome measures at baseline, 13 weeks (mTBI group only), and 6 months (48). Surprisingly, there were no observable differences between participants with mTBI and normative controls, nor any observable changes over time in the NAA/Cr (area), Cho/Cr (area), and Cho/NAA (area) ratios.

Functional MRI (fMRI). fMRI measures changes in regional blood oxygenation that are usually quantified based on blood oxygen level dependent activity (BOLD). Following injury, decreases in blood flow are speculated to represent impaired functional capacity (197). The study usually requires a patient to perform a task while being imaged. Fewer than 50 papers have been published investigating cognitive functioning after adult mild traumatic brain injury (civilian and military) using fMRI. They have focused on the acute and subacute injury period (ie, first 1 to 2 months). It appears that individuals with mild traumatic brain injury may exhibit different and symptom dependent alterations in regional blood oxygenation (158). Furthermore, whole brain resting state fMRI changes appear to be delayed to around day 8 postconcussion (179). A majority of studies have utilized cognitive tasks in which patients with mild traumatic brain injury show comparable performance to healthy controls on aspects of frontal lobe/executive functioning (eg, working memory and attention) (227). These studies demonstrated both decreases and increases in frontal lobe activation. Messe and colleagues used resting state fMRI (at 1-3 weeks and 6 months) comparing 17 patients with persistent postconcussion symptoms with 38 mild traumatic brain injury patients with no symptoms and 34 health controls (250). Alterations in functional brain networks were found in both groups with mild traumatic brain injury, specifically in the temporal and thalamic regions during the subacute phase and in the frontal regions in the late phase. In contrast, resting state fMRI (ie, where the subject is not required to perform a task), which measures functional connectivity, demonstrated abnormalities in the anterior cingulate and posterior cingulate cortex in acutely concussed individuals (173). However, Churchill and colleagues showed both increased and decreased functional connectivity in distinct but various brain regions (ie, frontal, temporal, and insular regions) (58).

Limited studies using fMRI in sports-related concussion have demonstrated abnormal BOLD activity, which correlated with symptom scores and neuropsychological testing (218). The studies also demonstrated improvement in BOLD activity in patients whose symptoms had resolved, implicating fMRI as a tool to possibly assess recovery from concussion. In addition, athletes who displayed hyperactivation on a cognitive task in the acute phase had prolonged recovery times relative to those athletes who demonstrated typical activation in the acute phase implicating fMRI could be used to predict who will recover more quickly. fMRI has also been used to assess specific areas affected by sports concussion. Jantzen and colleagues found increased activation in the areas of the parietal, lateral frontal, and cerebellar regions (171). Furthermore, oculomotor dysfunction has long been associated with all stages of concussion. Using fMRI with oculomotor tests (saccades and eye fixation), Johnson and colleagues looked at nine recently concussed athletes (six male and three female) and nine age-matched controls (175; 174). Significant differences during the acute phase in performance between normals and concussed individuals were seen on 3 out of 7 tasks (horizontal anti-saccades, memory-guided saccades, and self-paced saccades, ie, complex testing). fMRI revealed widespread increased activation of multiple brain areas in the concussed individuals during oculomotor testing. Hammeke and colleagues used fMRI and balance testing to compare 12 concussed high school athletes and 12 age-matched controls (141). They found injured athletes showed decreased activation of right hemisphere attentional networks during the acute period with a reversed pattern of activation (injured > controls) in the same networks at 7 weeks following injury. The changes coincided with a decrease in self-reported postconcussive symptoms and improvement on cognitive test performance. Preliminary results from the ongoing multicenter NCAA/DOD project of the natural history of concussion using fMRI and advanced arterial spin labeling MRI showed significant changes in cortical blood flow in several brain regions in concussed athletes. Correlations were found between decreased cortical blood flow and the Balance Error Scoring System total score as well as scores on the Immediate Post-Concussion Assessment, Cognitive Test Memory Composite, and Impulse Control Composite (376).

Magnetoencephalography (MEG). MEG is a brain mapping technique that provides an estimation of the brain’s electrical activity (196). When compared with fMRI, MEG has a higher degree of temporal resolution; it is less distorted by the skull and scalp as compared to EEG. MEG has been shown to be sensitive for mild traumatic brain injury in military and civilian populations suffering injuries from both blast and nonblast causes (165; 164). The technology is still evolving, and more detailed studies are being published including a study by Kaltiainen and colleagues, who recorded resting-state magnetoencephalography data in both eyes-open and eyes-closed conditions from 26 patients (11 females and 15 males, aged 20-59) with mild traumatic brain injury 6 days to 6 months after the injury, and compared their spontaneous oscillatory activity to corresponding data from 139 healthy controls (177). The results suggested that aberrant theta-band oscillatory activity can provide an early objective sign of brain dysfunction after mild traumatic brain injury. Other studies have demonstrated decreases in global activation (cue and target-evoked P300m and contingent magnetic variation) in concussed patients (285), as well as abnormal hypo- and hyper-activation patterns in the frontal-parietal ventral occipitotemporal, temporal, and subcortical areas when using tests of visual working memory (330). Finally, resting state MEG using cross frequency coupling and dynamic functional connectivity is currently being looked at as a potential biomarker for concussion (08).

Near infrared spectroscopy (NIRS). NIRS is a noninvasive technique that evaluates cerebral blood volume and oxygenation. The technology is based on the transmission and absorption of near-infrared light as it passes through tissue; however, it is limited by its ability to only detect changes in cortical tissue. NIRS is more portable than most functional neuroimaging techniques and can also be used on moving subjects. Cote and colleagues monitored cerebral hemodynamics during acute exercise following concussion in 14 male university hockey players and found cerebral oxygenation was reduced up to 35% on day 1 following concussion (67). Blood volume increased immediately following a concussion at rest and during exercise at day 1 and returned to baseline by day 7. The authors concluded that there is an increased demand for oxygenated blood following concussion. This technology uses a handheld scanner and has the potential to be used in the office as a screening technique.

Diffusion tensor magnetic resonance imaging (DTI/MRI)/diffusion kurtosis imaging (DKI). DTI/MRI uses state of the art high-field MRI (1.5 T and 3.0 T) to evaluate the speed and direction of water movement within axons, which is termed “fractional anisotropy” (81). DTI is based on the diffusion of water molecules. Water tends to move faster along nerve fibers rather than perpendicular to them. In healthy individuals, white matter diffusion is more organized in a specific direction; this is known as anisotropy. Fractional anisotropy is believed to reflect many factors, including the degree of myelination and axonal density/integrity. Fractional anisotropy values range from 0 to 1, where 0 represents isotropic diffusion or lack of directional organization and 1 represents anisotropic diffusion or diffusion restricted to one direction (81). DKI/MRI extends conventional diffusion tensor imaging by estimating the kurtosis of the water diffusion probability distribution function. The extra information provided by DKI can also resolve intravoxel fiber crossings and, thus, be used to improve fiber tractography of white matter. A positive kurtosis means the distribution is more strongly peaked and has heavier tails than a Gaussian distribution with the same variance. Water diffusion in biological tissues is non-Gaussian due to the effects of cellular microstructure (eg, cell membranes and organelles). This is particularly evident in the brain, where water diffusion is strongly restricted by myelinated axons. A large diffusional kurtosis suggests a high degree of diffusional heterogeneity and microstructural complexity (172).

Technically, DTI is a modification of diffusion-weighted imaging and determines white matter integrity and is sensitive to changes in white matter microstructure. It is also specific to each individual scanner and, therefore, prior to institution, each imaging center must develop their own set of normal values and specific DTI protocol, which the operator and radiologist will then use to manually outline the specific regions of interest (81). Most imaging centers use region of interest analysis where white matter fractional anisotropy is measured in specific regions of the brain. The measured fractional anisotropy is then compared to a database of normal patients. Region of interest analysis is useful for testing hypotheses regarding the relation of white matter integrity in a specific neuroanatomical region to an outcome variable, and a 1.5 to 2 standard deviation in the fractional anisotropy is considered abnormal (81). There are a number of radiology facilities performing DTI analysis with tractography, ie, the radiologist subjectively evaluates the white matter tracts. This technique has not been studied sufficiently in head trauma and is more often used to evaluate patients with white matter disorders, such as multiple sclerosis.

Over the past decade there have been a number of small and a few large studies demonstrating DTI/MRI as a possible technology in both detecting permanent injury and, when used longitudinally, as a possible method to gage recovery from mild traumatic brain injury. To date, a majority of the studies using DTI have been in patients with long-term neurologic complaints (283). In 2007, Kraus and colleagues published a study that is considered the landmark paper for DTI methodology in evaluating traumatic brain injury (192). Region of interest analysis included the anterior and posterior corona radiata, corticospinal tracts, cingulum, external capsule, forceps major and minor, genu, corpus callosum, inferior fasciculus, and superior longitudinal fasciculus. The primary objective was to characterize white matter integrity utilizing DTI across the spectrum of chronic traumatic brain injury of all severities. Secondary objectives included examining the relationship between white matter integrity and cognition. In the study, 20 mild and 17 moderate/severe traumatic brain injury patients, and 20 controls, underwent DTI and neuropsychological testing with fractional anisotropy being the primary measure of white matter integrity. Moderate to severe traumatic brain injury patients showed decreased white matter fractional anisotropy in all region of interest analyses. Patients with mild traumatic brain injury showed decreased fractional anisotropy in the corticospinal tracts, sagittal stratum, and superior longitudinal fasciculus.

A second study by Mac Donald and colleagues hypothesized that blast-related traumatic brain injury causes traumatic axonal injury (227). The authors looked at 63 U.S. military personnel who had a clinical diagnosis of mild traumatic brain injury secondary to blast exposure plus a second nonrelated blast injury compared to 21 controls with blast exposure without traumatic brain injury (according to the results of clinical screening for traumatic brain injury). Inclusion in the study was based on individuals meeting the standard military criteria for traumatic brain injury. None of the patients had findings on conventional MRI, and both groups underwent DTI/MRI within 90 days of injury (mean 14 days) using region of interest analysis. Abnormalities consistent with traumatic axonal injury were defined as values for relative anisotropy that were more than 2 standard deviations below the mean of the values for controls. Subjects with traumatic brain injury showed marked abnormalities in the middle cerebellar peduncles, cingulum, and in the right orbitofrontal white matter. Furthermore, mean diffusivity and radial diffusivity were higher in subjects with traumatic brain injury than in controls on the initial scans, but normalized on follow-up scans. Axial diffusivity did not differ significantly between groups on the initial scans but was lower in the subjects with traumatic brain injury than in controls on follow-up scans. These findings are consistent with an evolution of injury, and the pattern of abnormalities on the initial scans was most consistent with axonal injury plus a cellular inflammatory response and edema. Follow-up DTI scans in 47 subjects with traumatic brain injury 6 to 12 months after enrollment showed persistent abnormalities that were consistent with evolving injuries, ie, the pattern of abnormalities was most consistent with persistent axonal injury plus resolution of the edema and cellular inflammation. The authors noted that abnormalities were found in regions of the brain not known to be commonly injured in civilian cases of mild traumatic brain injury, but were predicted to be vulnerable to blast on the basis of computational simulations, which is consistent with the current hypothesis that the mechanism of blast-related mild traumatic brain injury may be different when compared to mild traumatic brain injury related to other causes, ie, sports, falls, and motor vehicle accidents. The most significant conclusion was that as DTI can be performed relatively quickly on a majority of the MRI scanners at U.S. military facilities and civilian hospitals, DTI-based assessments may be useful in diagnosis, triage, and treatment planning in clinical practice.

As a rule, fractional anisotropy is decreased in patients with traumatic brain injury and can vary in patients with mild traumatic brain injury. Decreases in fractional anisotropy correlate with loss of white matter integrity (23). However, some studies, especially those where DTI was performed shortly after the concussive event, have found elevated fractional anisotropy and decreased mean diffusivity (123; 331; 34). Perez and colleagues looked at 13 patients with closed-head injury and positive findings consistent with diffuse axonal injury on conventional MRI and compared them to 10 healthy individuals at 1 day and 17 months post-injury (283). In the acute stages, mean, axial, and radial diffusivity were statistically increased, and fractional anisotropy was unchanged. In the chronic stages, patients with closed-head injury had disproportionally high axial and radial diffusivity with resultant decreases in fractional anisotropy. Increased mean diffusivity and decreased fractional anisotropy may indicate vasogenic (extracellular) edema that resolves over time. Moreover, increased fractional anisotropy and decreased mean diffusivity may represent cytotoxic edema (intracellular) secondary to axonal swelling and restricted diffusion, which, in turn, could indicate poor prognosis for recovery from mild traumatic brain injury. A number of studies have attempted to correlate abnormalities in fractional anisotropy and mean diffusivity with the clinical signs and symptoms of concussion. Miles and colleagues were able to demonstrate that an increase in mean diffusivity correlates with slowing of reaction time, and decreased fractional anisotropy correlates with neuropsychological dysfunction (ie, decreased executive function) (254). In two separate studies on patients with subacute and chronic mild traumatic brain injury, Niogi and colleagues found that decreases in fractional anisotropy were predictors of neurocognitive dysfunction (decreased attention) and decreases in reaction time correlated with the number of damaged white matter structures (268). In a separate study looking at patients with subacute and chronic mild traumatic brain injury, Maruta and colleagues found that abnormalities in the right corona radiate, left superior cerebellar peduncle, and genu of the corpus callosum correlated with eye movement abnormalities (233). Lipton and colleagues, in looking at patients with acute concussion (< 2 weeks), found that decreased fractional anisotropy in the dorsal lateral prefrontal cortex significantly correlated with poor executive function (213). Concussion can also be associated with persistent depression (10- to 14-year-old patients 6 months postconcussion) which can be measured using DTI, with changes seen in the middle and anterior corpus callosum, left middle frontal gyrus, and anterior cingulated cortex (226).

There are a few small DTI studies looking at athletes during the various stages of concussion, but none are longitudinal. Cubon and Putukian assessed white matter fiber tract integrity using tract-based spatial statistics in 10 varsity college athletes and controls (Cubon and Putukian 2011). They also included patients with moderate and severe traumatic brain injury and controls. Athletes still symptomatic 1 month after sports-related concussion demonstrated increased mean diffusivity in several white matter tracts in the left hemisphere, including the inferior and superior longitudinal fasciculus, fronto-occipital fasciculi, retrolenticular part of the internal capsule, and posterior thalamic and acoustic radiations. There was no difference in fractional anisotropy between athletes and controls; however, fractional anisotropy decreased with the level of severity. This could be interpreted to mean that the athletes’ fractional anisotropy was abnormal as the study’s controls actually had moderate to severe traumatic brain injury. The researchers concluded that fractional anisotropy may be more sensitive in detecting severe injury, and mean diffusivity may be more sensitive in detecting mild injury. A more recent study using tract-based spatial analysis found that radial diffusivity and mean diffusivity increased 2 days postinjury in white matter regions involving both hemispheres, which supports the hypothesis of a diffuse pattern of injury after concussion (74). After 2 weeks persistently elevated radial diffusivity and mean diffusivity were observed in the prefrontal portions of the white matter fiber tracts which remained elevated at 2 months, with recovery of white matter fiber tracts following a posterior to anterior trend that parallels the posterior to anterior pattern of white matter maturation in the normal population.

In a similar study, Henry and colleagues investigated the effects of sports concussion on white matter integrity using three different diffusion tensor imaging measures (fractional anisotropy, axial diffusion, and mean diffusivity) (154). A group of 10 nonconcussed athletes was compared with a group of 18 concussed athletes of the same age (mean 22.5 years) and education (mean 16 years) using a voxel-based approach within both the acute and chronic postinjury phases, ie, at 1 to 6 days postconcussion and again 6 months later. Fractional anisotropy was increased in dorsal regions of both cortical spinal tracts and in the corpus callosum in concussed athletes at both time points. Axial diffusivity at both time points was elevated in the right cortical spinal tracts. Mean diffusivity values were decreased in concussed athletes in the cortical spinal tracts and corpus callosum at both time points. Although there was some limitation in the technique used to image large fiber tracts, the authors concluded that sports concussions result in changes in diffusivity in the corpus callosum and cortical spinal tracts of concussed athletes.

In a study, diffusion kurtosis imaging (DKI) was used to monitor the effects of sports-related concussion on brain recovery in 96 football players at less than 48 hours and at 8, 15, and 45 days post-injury (261). The concussed athletes were compared with a matched group of uninjured players (n = 82). The concussed group reported significantly higher symptoms within 48 hours after injury than controls, which resolved by the 8-day follow-up. The concussed group also demonstrated poorer performance on balance testing at less than 48 hours and at 8 days than controls. There were no significant differences between the groups on the Standardized Assessment of Concussion (SAC), a cognitive screening measure. DKI data were acquired with 3 mm isotropic resolution and analyzed using tract-based spatial statistics. Additionally, voxel- and region of interest–based analyses were also conducted. At less than 48 hours, the concussed group showed significantly higher axial kurtosis than the control group. These differences increased in extent and magnitude at 8 days, then receded at 15 days, and returned to the normal levels by 45 days. Kurtosis fractional anisotropy exhibited a delayed response, with a consistent increase by days 15 and 45. The results indicate that changes detected in the acute period appear to be prolonged compared with clinical recovery.

Zhang and colleagues employed voxel-wise whole brain analysis and region of interest analysis to assess 15 student athletes (mean age 20.8); using an outdated concussion grading system, they found no significant change in fractional anisotropy (394).

The NCAA is now also partnering with the Department of Defense. In a preliminary study using DTI MRI, mean diffusivity was significantly higher in brain white matter of concussed athletes, particularly in frontal and subfrontal long white matter tracts. In the concussed group, axial diffusivity was significantly correlated with the Brief Symptom Inventory and there was a similar trend with the symptom severity score of the Sport Concussion Assessment Tool. In addition, concussed athletes with higher fractional anisotropy performed better on the cognitive component of the Standardized Assessment of Concussion (266).

In a prospective cohort study where athletes underwent pre- and postseason screening (3-month interval), Bazarian and colleagues found intermediate, but not statistically significant, changes in fractional anisotropy and mean diffusivity in eight athletes with subconcussive blows (range 26 to 399) when compared to controls (24). One subject sustained a concussion during the study and was scanned within 72 hours. That individual had an initial increase in fractional anisotropy and decrease in mean diffusivity in the right corona and right inferior longitudinal fasciculus. The values subsequently reversed (decrease in fractional anisotropy, increase in mean diffusivity) over time.

McAllister and colleagues attempted to correlate regional strain associated with a diagnosed concussion with brain injury using DTI imaging and helmet technology (HIT system) designed to record linear acceleration of the head (it does not measure rotational acceleration, which is required to induce concussion, which is a study design flaw) following impact (236). Participants underwent pre- and postseason DTI imaging and wore the helmet technology during games and practices. Ten football and ice hockey players, all male, sustained a concussion were imaged within 10 days of diagnosis. Maps of change in fractional anisotropy and median diffusivity were generated for the corpus callosum of each athlete to correlate strain with change in fractional anisotropy and mean diffusivity. Mean and maximum strain rate correlated with change in fractional anisotropy and mean diffusivity. Change in mean diffusivity correlated with injury-to-imaging interval, whereas change in fractional anisotropy did not. All scans were read as normal. The authors suggested that mean diffusivity may be more sensitive than fractional anisotropy in detecting microstructure injury. Furthermore, the corpus callosum is a common site for axonal injury of varying severities, including mild injury; however, brain regions other than the corpus callosum might also be subject to relatively high strains and contribute to the injury process.

A similar study by Davenport combined sensor technology, DTI MRI, and computerized neuropsychological testing to look at athletes after a season of high school football (78)). None of the athletes studied experienced clinical symptoms consistent with concussion. However, there was a linear correlation between the total number of impacts/risk-weighted cumulative exposure and all DTI measures. There was also a strong correlation between DTI measures and change in verbal memory subscores on computerized neuropsychological testing. Another study evaluated subconcussive events in concussion-naïve elite-level German club soccer players (n=12) using DTI and compared the results with age-matched controls (swimmers, n=11) (191). The authors used automated whole-brain, tract-based spatial statistics to measure fractional anisotropy, mean diffusivity, radial diffusivity, and axial diffusivity. Widespread differences between the groups were found, with increased radial diffusivity in soccer players in a number of brain regions. Axial diffusivity was higher in the corpus callosum in soccer players, and there were no differences in fractional anisotropy or mean diffusivity. The authors suggested that the changes in radial diffusivity and axial diffusivity are consistent with demyelination, which could be secondary to subconcussive brain trauma. Finally, a study by Fakhran and colleagues looked at sex difference in individuals with mild traumatic brain injury (105). Male patients with mild traumatic brain injury had significantly decreased fractional anisotropy values in the uncinate fasciculus bilaterally compared with female patients with mild traumatic brain injury and control subjects (P < .05). These individuals also took longer to recover.

Researchers are now combining DTI with other types of advanced neuroimaging (51). A study of 15 school-age children at 3 to 12 months after sports concussion and 15 age-matched controls used DTI MRI, MR SPECT, and perfusion-weighted MRI to assess functional and structural abnormalities and concussion recovery (251). Individuals with mild traumatic brain injury showed reduced cortical blood flow and relative cerebral blood flow in the bilateral thalamic regions. They also demonstrated decreased NAA/creatine and NAA/choline in the corpus callosum and bilateral parietal white matter. There were no group differences in fractional anisotropy in patients with concussion and controls; however, patients with cognitive symptoms showed increased radial diffusivity in the right anterior limb of the internal capsule when compared to controls. Individuals who sustained concussion but did not exhibit prolonged cognitive symptoms demonstrated increased fractional anisotropy and decreased radial diffusivity in the posterior limb of the right internal capsule when compared to controls and decreased radial diffusivity and increased fractional anisotropy in the right anterior and posterior limbs of the internal capsule when compared to concussed individuals with cognitive symptoms. Finally, an increase in apparent diffusion coefficient was seen in concussed individuals with behavioral abnormalities. Similar changes were seen in concussed individuals on perfusion CT (3.6 hours) and DTI (160 days) with decreased fractional anisotropy in the bilateral frontal white matter and increased mean diffusivity in the pareito-temporal regions, with cerebral blood volume correlating significantly with fractional anisotropy. Conversely, Fakhran and colleagues looked at white matter integrity (ie, fractional anisotropy) in patients who did not have abnormalities on conventional MRI (104). DTI was conducted on 64 patients with concussion and 15 controls. All patients underwent a total concussion symptom score assessment of verbal memory, visual memory, processing speed, and reaction time. Patients with and without sleep disturbances were also compared. There was a statistically significant correlation between a high total concussion score and reduced fractional anisotropy in the gray-white junction, most prominent in the auditory cortex. Fractional anisotropy in the parahippocampal gyri was significantly decreased in patients with sleep and wake disturbances. The parahippocampal abnormalities mirror the abnormalities seen in Alzheimer dementia. Ling looked at mild traumatic brain injury in 26 patients at 14 days and 4 months and correlated with neuropsychological testing (211). Regions of interest (mean diffusivity and fractional anisotropy) in both cortical and subcortical structures were assessed. During the semiacute phase, concussed patients showed increased fractional anisotropy in the bilateral superior frontal regions, with the left superior frontal cortex remaining elevated at 4 months. There was no difference in neuropsychological testing and gray matter atrophy between concussed individuals and controls.

Not all studies support the efficacy of DTI in detecting mild traumatic brain injury. Pothole analysis looks at clusters of voxels with fractional anisotropy values more than 2 STD below the normal control values. Patients with mild traumatic brain injury show significantly more potholes than controls. However, Watts and colleagues found that in cross validation studies, there was a 30% false-positive rate (377). Using a tract-based spatial analysis, Ilvesmaki and colleagues did not find abnormalities between patients with mild traumatic brain injury and controls, even when injury severity was taken into account (167). None of the DTI measures were associated with postconcussion syndrome either acutely or at 1 month. Statistically significant age-related decreases in fractional anisotropy were seen in patients aged 41 to 60 years when compared to the control group in general and decreases in apparent diffusion coefficient values were seen in patients 51 to 60 years of age; therefore, age should be considered in white matter analysis.

Combination and other forms of MRI. Combining techniques, although not practical in the clinical setting, has provided significant input into the complex physiological aspects of concussion in the research setting. Using fMRI and arterial spin labeling in conjunction with a respiratory challenge in the early stages of concussion, Churchill and colleagues showed SRC was associated with symptomatic-dependent greater reductions in BOLD activity during the early phase of the respiratory task in the frontal and prefrontal areas (arterial spin labeling cerebral blood flow was unaffected) when compared to nonconcussed controls in which a robust whole brain response was observed (59). Longitudinal studies combining DTI and resting state fMRI have demonstrated changes in white matter structure and functional network connectivity that persist beyond clinical recovery (229). Similar results (ie, diffusion abnormalities in multiple white matter tracts along with decreases in choline) were seen months after the initial concussion by the same group when they combined DTI with MR-SPECT in a cohort of Bantum hockey players (230).

Dynamic contrast-enhanced MRI, which looks at the quantitative disruption of the blood-brain barrier, was used in patients with prolonged postconcussion symptoms. Patients with prolonged concussion symptoms (when compared to normal controls) had significant higher quantitative values (ie, permeability) that correlated with delayed recall on neuropsychological testing. The authors hypothesized persistent blood-brain barrier disruption as a possible mechanism for patients with prolonged concussion symptoms (388). A similar study in children and adolescents with concussion using arterial spin labeling in conjunction with symptom checklists and cognitive testing showed regional differences in cerebral blood flow (ie, hypoperfusion in the posterior and inferior regions and hyperperfusion in the anterior/frontal and temporal regions) when compared to nonconcussed controls (39).

Finally, structural MRI (ie, increased ventricular brain ratio) has been shown to be a biomarker for traumatic brain injury. In a longitudinal study at 3 and 8 years post mild traumatic brain injury, military veterans did not show signs of gross brain atrophy as reflected by abnormally high ventricular brain ratio (79).

Management

Concussion symptoms tend to be short-lived and, therefore, do not usually require treatment (243; 96). The one exception is headache, which tends to be present from the onset (133). Over the past few years, most concussion experts have begun to move away from cognitive rest (ie, having the individual do nothing) (130; 320; 305) to physical and cognitive activity to tolerance (332). One study found that starting exercise by day 5 can actually improve outcome (200). Headache, depression, and sleep disturbances are the most common symptoms requiring pharmacological management. Vertigo or disequilibrium and musculoskeletal pain are best treated with physical medicine (136). However, it has been this author’s and other concussion experts’ experience that ocular and vestibular rehabilitation tends to worsen concussion symptoms and actually slow recovery. Despite the prevalence of the disorder, there is little in the way of evidence on the management of concussion symptoms; many physicians rely on clinical experience, applying principals used in the management of nonconcussed patients. For example, medications should be easily titrated and weaned and used at the lowest effective dose. Attempt to use one medication to treat multiple symptoms, ie, headache, sleep disturbances, or depression.

In athletes and active-duty military personnel, the most important principal that the physician must realize is that shortening the duration of symptoms can have catastrophic effects as medications can mask unresolved concussion symptoms. Prophylactic medications need to be discontinued and the athletes/soldiers monitored off of all such medications for sufficient time so that the medication is completely out of their system before any return to play/duty decisions are made. Finally, pharmacologic treatment of professional and elite amateur athletes needs to take into account governing bodies’ rules and regulations for banned substances; one should always check with the governing body to see if the medication is allowed. Most governing bodies, including the National Hockey League, use the World Anti-Doping Associations (WADA) guidelines, which are the strictest. The National Football League (NFL) and Major League Baseball (MLB) have their own guidelines. All of the above guidelines are very similar and include such substances and anabolic steroids, amphetamines and other stimulants, glucocorticoids, beta agonists, beta blockers, peptide hormones and growth factors, hormone antagonists and modulators (ie, masking agents), diuretics and other masking agents, blood/oxygen transport enhancers, narcotics, and cannabinoids.

Depression. Depression is a common and often unrecognized symptom of concussion. Behavioral therapy is the mainstay in the treatment of depression, with antidepressants reserved for the most chronic and refractory cases. Tricyclic antidepressants such as nortriptyline are usually used as first-line therapy and are also effective in the prevention of headache. In addition, they have sedative properties making them effective sleep agents, and titration and weaning schedules are much shorter than selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs). The one exception is amitriptyline, which has a higher incidence of anticholinergic activity and can worsen cognitive symptoms. Bupropion hydrochloride is an option in patients with prolonged significant daytime fatigue and decreased energy, and in patients with significant issues with attention and concentration, as it is often used off-label for attention deficit hyperactivity disorder. The medication has the potential to lower seizure threshold and should be avoided in patients with structural abnormalities on MRI or epileptic activity on EEG. SSRIs and SNRIs should be avoided secondary to prolonged titration and weaning schedules. They can also be sedating and have a higher risk of suicide in adolescents and late teens. If an SSRI/SNRI needs to be considered, ie, in a patient with prolonged symptoms who has failed tricyclic antidepressant, venlafaxine would be the preferred drug as it has stimulating properties and has been shown to be effective as a migraine preventative. Monoamine oxidase inhibitors should also be avoided due to drug interactions, especially with antimigraine medications. Finally, atomoxetine, which is FDA approved for the treatment of attention deficit hyperactivity disorder, could be considered in patients with significant cognitive symptoms (ie, attention and concentration).

Sleep. There appears to be a higher incidence of sleep disturbances (ie, maintenance, initiation, and early morning awakening) in mild traumatic brain injury compared to traumatic brain injury. This could be a result of differences in metabolic activity or recovery or may be secondary to under reporting (40). Poor sleep appears to prolong recovery (ie, increased duration or headache and cognitive symptoms) in individuals with sports-related concussion (262). There is, however, limited data on the effects of repetitive mild traumatic brain injury on sleep. Military studies using a self-reporting Insomnia Severity Index have shown a higher incidence and severity of insomnia in personnel who have sustained multiple concussions. Sleep is best treated with natural, over-the-counter remedies to prevent dependency and rebound insomnia. Compounds such as diphenhydramine (25 to 50 mg), valerian root, and melatonin (3 to12 mg) can be used alone or in combination. Diphenhydramine is also effective in aborting migraine and other headaches and can also be used as a short-term headache preventative (64). Melatonin acts to maintain sleep and has been shown to be a possible effective migraine preventative. If a medication is required, then tricyclic antidepressants would be considered first-line due to their ability to treat associated symptoms. Trazodone, which is chemically similar to tricyclic antidepressants, is another alternative. Sedative hypnotics such as zolpidem and eszopiclone, which can cause rebound insomnia and worsen postconcussion symptoms of headache, cognitive symptoms, and dizziness, should be avoided, as should benzodiazepines and barbiturates.

Headache. Headache is the most common symptom of both traumatic (71%) and mild traumatic brain injury and concussion (91%) (223; 62). One year after initial injury, 44% of patients with traumatic brain injury and 54% of patients with mild traumatic brain injury will continue to experience headache (223), with many meeting the International Headache Society (IHS) diagnosis requirements for chronic posttraumatic headache and refractory posttraumatic headache. Active soldiers with a history of head trauma and persistent headaches are 4 times more likely to have a medically related discharge (112). A majority of patients with mild traumatic brain injury meet the IHS criteria for migraine with and without aura. Exacerbation of primary headaches, such as tension headache, cervicogenic headache, and even cluster headache, can occur alone or in combination (234; 101).

Migraine and mild traumatic brain injury share a number of physiological mechanisms of which an in-depth discussion is too large for this review. Both migraine with aura and concussion involve cortical spreading and depression. However, migraine without aura does not. It is possible that the physiological cascade associated with concussion could activate yet to be determined “migraine generators” triggering the well-defined trigeminovascular cascade resulting in both peripheral and central sensitization. This cascade is mediated by the inflammatory neuropeptide calcitonin gene related peptide (CGRP). Evidence (ie, rat models) suggest a role for CGRP and even more so for mast cells in the development of posttraumatic headache (35). Furthermore, ongoing studies are currently looking at monoclonal antibodies to CGRP.

Another emerging field is imaging of posttraumatic headache. One study looked at regional volumes, cortical thickness, surface area, and brain curvature in patients with persistent posttraumatic headache, those with migraine, and controls. Differences were seen in the right lateral orbitofrontal lobe, right supramarginal gyrus, and left superior frontal lobe in patients with persistent posttraumatic headache. No differences were seen when comparing migraine patients to controls (323). A second study used DTI-MRI to compare the white matter tracts of migraine patients, persistent posttraumatic headache, and controls. There were significant differences in mean diffusivity and radial diffusivity between migraine subjects and posttraumatic headache subjects. Furthermore, there was a positive correlation between fiber tract alteration and headache frequency in both migraine patients and posttraumatic headache patients; however, the locations were quite different (57).

Just as with other primary headache disorders, treatment involves abortive and prophylactic therapy; however, most patients do not develop chronic posttraumatic migraine (253) and only require abortive medication, which should be instituted when the pain is mild and headache treated completely. When treating individuals, especially athletes with acute postconcussive headache, where the physiology involves a transient process, the preferred approach should be transitional--combining physical and cognitive activity to tolerance and interventional procedures, such as nerve blocks and physical medicine, prior to the institution of preventative medications that often take weeks to work and have significant side effects. Over-the-counter medications, including acetaminophen and NSAIDs, have usually been tried before the patient comes to the attention of the physician. In these patients, triptan medications would be the next logical choice. If the patient does not respond consistently, the triptan can be combined with an over-the-counter NSAID such as naproxen. For patients in whom the headache develops rapidly or for those who wake up with a headache, nasal sprays or injection is usually more effective as these medications bypass migraine-induced gastric stasis. Dihydroergotamine (DHE-45) is an excellent alternative for triptan nonresponders, as is oral or intramuscular ketorolac, which can be added to triptans or DHE-45 or used alone in patients who have contraindications to the use of the above medications. As is the case with any headache type, narcotics and butalbital- or caffeine-containing compounds should be avoided due to the risk of medication overuse headache (155).

For patients who do not respond to the abortive medications or who prefer not to take medications, peripheral nerve blocks are an excellent option and are one of the most widely used interventional procedures to treat posttraumatic headache. Common sites include the greater occipital nerve (GON), lesser occipital nerve (LON), auriculotemporal nerve (ATN), supraorbital nerve (SON), supratrochlear nerve (STN), and sphenopalentine ganglion (SPG). Interventions include blocking a single nerve unilaterally or bilaterally or blocking multiple nerves (usually bilaterally). The SON, STN, and ATN are branches of the trigeminal nerve whereas the GON and LON are derived from the dorsal and ventral rami of C2-4 (32). From a mechanistic perspective, it is hypothesized that local anesthetic to these nerves results in afferent feedback to the trigeminal nucleus caudalis, “shutting down the headache generator.” This can occur directly via trigeminal branches or indirectly via nociceptive afferents from the C 1-3 spinal nerves and the GON, which converge onto second-order neurons that also receive afferents from adjacent cervical nerves and from the first division of the trigeminal nerve via the trigeminal nerve spinal tract. Typical anesthetics include bupivacaine (0.25 to 0.75%) or lidocaine (2%), with volumes ranging from 0.5 to 2 cc per site. Local anesthetics inhibit conduction through nerve fibers by reversibly inhibiting sodium channels and can act on the unmyelinated C-fibers and the thinly myelinated Aδ fibers that mediate pain.

Local anesthetics can be given alone, combined with each other, or combined with a steroid, usually triamcinolone. The addition of steroids in nerve blocks for noncluster headaches has never been shown to be efficacious. In a single-blinded randomized controlled trial, Ashkenazi and colleagues compared the effect of lidocaine with triamcinolone versus lidocaine alone in patients with transformed migraine (11). No statistically significant differences were seen in any of the outcome measures between the two groups. Steroids should never be used in the forehead area due to the risk of tissue necrosis and unwanted adverse cosmetic side effects.

A number of studies have looked at the use of peripheral nerve blocks in posttraumatic headache, but none are randomized double-blind placebo-controlled studies, with almost all being case reports or retrospective case analysis. Most studies involved only blockade of the greater occipital nerve, with many being focused on the pediatric population (127; 152; 113; 361). One such study looked at 15 patients (mean age 15.47 years; range: 13-17) who received occipital nerve blocks for posttraumatic headache (median number of days per month 30, range 10 to 30) (327). Fourteen of the 15 patients were seen in follow-up at 5.57 months post-injury (standard deviation = 3.52 months). Sixty-four percent reported long-term response to the occipital nerve blocks, with associated improved quality of life and decreased postconcussion symptom scores (P < .05). A second study, also a retrospective analysis, looked at 28 patients (younger than 18 years of age) receiving either GON, SON, or LON blocks for posttraumatic headache secondary to concussion lasting more than 2 to 3 weeks (83% reported migraine like features) (93). Sixty-two nerve blocks were performed at a mean of 70 days post-injury. Seventy-one percent reported immediate relief of their headaches, with all patients reporting some reduction in headache intensity (average reduction was 94%). Finally, 93% reported good therapeutic effect, which was defined as sustained relief for more than 24 hours or request for a repeat block; 26% reported that the nerve blocks “cured their headaches.” Another possible target in treating posttraumatic headache is the sphenopalatine ganglion block. Advances in equipment have allowed for the intranasal blockade of the sphenopalatine ganglion. A case report demonstrated resolution of persistent posttraumatic headache in a 17-year-old boy with a sports-related concussion after treatment with an intranasal sphenopalatine bloc (348).

Patients who continue to experience posttraumatic migraine headache for more than a month, and at a frequency of greater than 6 to 8 headaches per month, may benefit from prophylactic treatment. There are very few well designed controlled studies looking at posttraumatic headache prevention. The mainstay of treatment is the tricyclic antidepressants, as they also treat associated symptoms. Conversely, antiepileptic medications, which are the mainstay of preventative migraine treatment for most neurologists, are not good options in concussed patients due to potential side effects, including sedation and cognitive or psychomotor retardation as well as prolonged titration and weaning schedules. If the physician were to consider using an antiepileptic drug, zonisamide 50 to 100 mg twice daily is an option. One study retrospectively looked at the use of gabapentin and tricyclic antidepressants in the treatment of posttraumatic headache (76). Neither medication had a significant effect on longitudinal improvements in outcome scores. There was, however, some short-term improvements with gabapentin and more sustained improvements with tricyclic antidepressants. Beta blockers and calcium channel blockers should be avoided in athletes, as both can cause a significant decrease in pulse rate in these highly conditioned, low resting pulse rate individuals. Furthermore, beta blockers are on a majority of the banned substance lists. Finally, they can increase lethargy and can have cognitive side effects, which most concussed patients already experience. Another interesting possibility is memantine, which has been shown in retrospective studies to be an effective migraine prophylactic (53). The drug is an NMDA receptor antagonist and appears to inhibit cortical spreading and depression, seen both in concussion and migraine. It may also be protective and improve associated concussion symptoms. Another frequently used drug in traumatic brain injury, amantadine (100 milligrams twice a day for 2 months), was shown to significantly reduce (ie, 80%) headache frequency and intensity in patients with prolonged concussion symptoms. Interestingly, time of initiation from the initial head trauma did not affect outcome (45). Finally, there is evidence that certain vitamins, including magnesium oxide 200 to 400 mg per day, whole leaf feverfew, riboflavin, and petasites root, alone or in combination are effective in preventing migraine. Natural/holistic therapy is favored by athletes due to a low side-effect profile and lack of banned substance issues (64).

Approximately 15% of patients will go on to develop chronic posttraumatic headache attributed to mild head injury (Conidi, Poster International Headache Society Meeting, Berlin, June 2011). These patients require a multidisciplinary approach incorporating neuromuscular rehabilitation, pharmacological management, infusion therapy, behavioral modification techniques, and interventional techniques (nerve blocks, trigger point injections, epidural injections). These patients are most appropriate for academic and nonacademic tertiary headache centers (64). Furthermore, some patients may benefit from onabotulinum toxin type A. However, due to the toxin’s prolonged duration of action, roughly 3 months, and potential to weaken the cervical musculature, it should be used with caution in individuals who are at risk for further concussion, ie, athletes and active-duty military personnel.

There are a few case reports showing efficacy of botulinum toxin in the treatment of chronic posttraumatic headache (212; 63; 387). The most extensive study to date retrospectively looked at the charts of 64 male servicemen aged 20 to 50 years who presented to the Concussion Care Clinic of the Womack Army Medical Center in Fort Bragg, North Carolina between 2008 and 2012 (387). The investigators reviewed 111 charts; 63 patients who received a diagnosis of chronic posttraumatic headache were included in the study. Of those patients, 36 (56.3%) had more than one type of headache, 10 (15.6%) had more than two headaches, and 48 subjects (75%) had continuous headache. Individuals were injected using the current FDA-approved protocol of 31 fixed-site injections (5 units per site), with additional injections being placed in a fixed-dose, follow-the-pain paradigm. Some patients had a combined diagnosis of cervical dystonia and chronic posttraumatic headache and underwent injections in a follow-the-pain pattern. The mean number of injection cycles was 3.3, with 30 patients receiving only one set of injections. Forty-one patients (64%) reported that they were “better” after treatment, 18 (28%) were unchanged, two (3%) were worse, and three patients were lost to follow-up. In addition, 18 (29%) continued to undergo treatment whereas 26 (43%) discontinued therapy for lack of efficacy. About half of the patients continued on active duty or were redeployed, and the other half either retired or were discharged (all but one for medical reasons). Common side effects included headache and neck pain. The biggest variable for lack of efficacy appeared to be the presence of continuous headache prior to treatment. Finally, the authors suggest that the discrepancy in efficacy when compared to the phase-3 FDA studies may be issues with secondary gain in the treatment population. Another modality, transcranial magnet stimulation, which is FDA-approved to treat headache, was shown to significantly reduce headache intensity in 30 military veterans with persistent daily headaches (206). Finally, monoclonal antibodies directed against calcitonin gene related peptide (CGRP) have been approved to treat episodic and chronic migraine. Posttraumatic headache likely shares pathophysiological mechanisms with migraine; as such CGRP may also a play role in posttraumatic headache. Bree and Levy used a rat percussion model to study posttraumatic headache (35). They found that inhibiting CGRP by using a mouse antiCGRP monoclonal antibody can reduce the headache and pain-related symptoms evoked by mild traumatic brain injury, suggesting the involvement of CGRP in mediating posttraumatic headache pain. It is, therefore, plausible that the current FDA-approved monoclonal antibodies on the market may also be useful in aborting and preventing posttraumatic headache. Studies are currently in process to assess that theory.

Conversely, a population group with no secondary gain are concussed athletes with chronic posttraumatic headache. In this author’s experience these patients have shown significant improvement in their headaches and associated concussion symptoms after treatment with botulinum toxin (387). The case report below outlines a typical patient and typical outcome. In addition, some patients with posttraumatic headache may also have suffered concomitant trauma to their cervical spine, whereby the underling cervical pathology may be a “trigger” for their headaches. These patients may benefit from interventional procedures, such as trigger-point injections, epidural injections, and facet blocks.

Finally, a small number of patients will become refractory, these individuals are excellent candidates for peripheral nerve stimulation. Carefully selected patients, ie, those without significant secondary gain or psychological issues and who meet the criteria above may be candidates for peripheral nerve stimulation. Peripheral neurostimulation is an adaptation of an FDA-approved spinal nerve stimulator where leads are placed over various peripheral nerves located in the head (as opposed to the dorsal root ganglion). Patients typically undergo a 5- to 14-day trial (with internal lead placement and external device placement), and if the device is felt to be efficacious (typically a greater than 50% reduction in headache days, pain, or disability), permanent implantation follows. Original studies looked at stimulation of the greater occipital nerve. One such study, a retrospective case analysis by Schwedt and colleagues looked at 15 patients with medically intractable or refractory headache (2 had posttraumatic headache) (324). There were significant reductions in headache frequency, severity, disability, and depression. Surprisingly, 8 of 15 patients required lead revision, which is significantly higher than seen in other trials. Elahi and Reddy report on two cases of posttraumatic headache responding to neurostimulation (99; 100). A 57-year-old male with intractable headaches (who did not try botulinum toxin) had a greater than 90% reduction in headache frequency and weaned off all preventative and prescription abortive medications after implantation and stimulation of the greater auricular nerve (a branch of C2,3). A second patient, a 40-year-old male with chronic posttraumatic headaches refractory to numerous medications and treatments, including botulinum toxin, also had a 90% decrease in “headache pain” after stimulator placement in the high cervical region (C1). Our group has extensive experience with peripheral neurostimulation with approximately 500 implanted patients for a variety of headache disorders including refractory, migraine, cluster, hemicrania, hemiplegic migraine, and posttraumatic headache (301; 2015; 99; 100; 91). Almost all received dual neurostimulation, typically the SON and GON, with some receiving only stimulation of the occipital nerves (unilateral and bilateral) and others receiving stimulation of the GON, SON, and ATN. An abstract presented at both the 2014 American Headache Society and American Academy of Neurology annual meetings looked at a random subgroup of 163 patients (129 female; 34 male) who underwent placement of a combined SON and GON stimulator (mean time since placement was 14 months). On average, patients experienced a 73% decrease in headache days (many were headache-free), a 76% reduction in MIDAS scores, and a 56% reduction in VAS pain scores; 71% of patients decreased medication usage by over 50%, and 38% completely discontinued routine headache medication; almost all would recommend the procedure. Lead migration rates were less than 5%, and infection rates were approximately 1%. Despite the numerous treatment options available, one significant question remains: in individuals for whom daily headache is the only remaining symptom, is the headache a continuation of the concussion? Or has the patient developed chronic posttraumatic migraine? Answering this question could have significant implications for treatment, and even return to play, work, or duty.

Physical medicine and rehabilitation. Neck pain is a common complaint in patients with mild traumatic brain injury and coincides with the mechanism of injury. The cervical spine can act as a significant trigger of posttraumatic headache. There are very few studies that specifically look at physical therapy, massage therapy, spinal manipulation, and mobilization as a treatment for concussion. Numerous studies have assessed the aforementioned modalities in cervicogenic headache and include patients with mild traumatic brain injury and traumatic brain injury. Anecdotal evidence from thousands of patients from our clinic (most of which have episodic and chronic posttraumatic headache secondary to mild traumatic brain injury or concussion) who were treated with neuromuscular or physiotherapy (ie, massage, myofascial release, traction, ultrasound, electrical stimulation, heat, ice and therapeutic exercise) in conjunction with medications and other interventional techniques, shows a significant improvement in patients who undergo adjunctive neuromuscular or physiotherapy when compare to those who do not. It is also our experience that most, if not all, patients with posttraumatic headache have pain or spasm in the cervical muscles, suboccipital muscles, and over the temporalis region, which when palpated can reproduce the patient’s headache symptoms. These findings are also consistent with the observation of other headache specialists and studies (38; 109; 110; 375). In the Von Piekartz study, patients underwent physiotherapy by unblinded physical therapists with the primary endpoint being the change in headache intensity. The patients were randomly assigned to receive either manual therapy for the cervical region (usual care group) or additional manual therapy techniques to the temporomandibular region. Patients in the treatment group experienced statistically significant decreases in headache intensity at 3 and 6 months when compared to the usual care group (38). A similar study targeted the sternocleidomastoid muscles (31). Twenty patients (10 active and 10 controls) with cervicogenic headache and trigger points in these regions underwent manual therapy (myofascial release, stretching, and massage) or simulated manual therapy. Patients receiving manual therapy had an increase in cervical range of motions and a greater reduction in headache and intensity of neck pain when compared to controls. Combination therapy (ie, cervical and vestibular) has also shown to be beneficial in patients with persistent concussion symptoms. Schneider and colleagues randomized 31 patients ages 12 to 30 with sports-related concussion (Schneider et at 2014). Seventy-three percent of those receiving physical therapy had significant improvement in their symptoms and were returned to play by 8 weeks, compared to only 7% in the control group.

Two other modalities that have received considerable study (especially with respect to cervicogenic headache) are spinal manipulation and mobilization. Cervical spinal manipulation is a technique often used by chiropractors and involves high velocity, low amplitude, localized force directed at cervical joint segments, whereas spinal mobilization involves low velocity, low amplitude movements with the patient’s range of motion and is often used by osteopathic physicians and physical therapists (389). When addressing the cervical spine, mobilization techniques are safer than manipulation techniques, which can be associated with adverse effects (ie, disc herniation and arterial dissection). Furthermore, in individuals with concussion, the high velocity forces used in manipulation can actually reproduce the angular and rotational forces that result in concussion, thus, potentially resulting in a second concussion. Despite this, there are a number of studies looking at cervical manipulation as a modality to treat primary headache disorders. A majority are prospective or retrospective case analyses (50; 109; 110). Chaibi and Russell performed an extensive literature search and identified six studies employing spinal manipulative therapy to treat cervicogenic headache (50). There were numerous methodological flaws in all of the studies, ranging from participants experiencing intermittent headaches, lack of treatment in the control group, small numbers of participants, and a lack of serial treatments. The authors conclude the spinal manipulation may be an effective treatment for cervicogenic headache; however, better studies are needed. Conversely, there are few studies looking at manipulation. Youssef and Shanb compared the efficacy of spinal mobilization with that of massage therapy in patients with cervicogenic headache (389). Thirty-six subjects were randomized, with 18 receiving passive spinal mobilization for 30 to 40 minutes and the other 18 receiving massage therapy, myofascial release, traction, and stretching exercises. Both groups were treated for 12 sessions (2 times per week for 6 weeks). Outcome measures included decrease in headache intensity, frequency, and duration as well as improvement in cervical pain and range of motion. Both groups experienced significant improvements in all measured variables, with the mobilization group experiencing statistically significant reductions in all variables when compared to the massage group. Finally, a case report by Channell and colleagues showed a multidisciplinary approach including medications and osteopathic manipulation to be effective in treating a 38-year-old female with chronic posttraumatic headache (52).

Exercise. A major change in concussion treatment is the early addition of a light, low-impact, nonaerobic, sub-symptom threshold exercise, usually beginning around day 3 or when the athlete is able to tolerate. Lawrence and colleagues evaluated 253 concussed individuals and found that earlier time to aerobic exercise was associated with faster return to sport and school/work (199). In addition, for each successive day in delay to initiation of aerobic exercise, individuals had a less favorable recovery trajectory. Furthermore, individuals who are in better physical shape appear to recover quicker as demonstrated by Leddy and associates, who found that individuals with a lower heart rate threshold took longer to recover from concussion (201).

Hyperbaric oxygen (HBO2). Preliminary studies are beginning to look at the effects of treatment with HBO2 in individuals with concussion, mTBI, and traumatic brain injury. These studies have mainly focused on military personal. A randomized, double-blind, sham-controlled trial enrolled military personnel with mTBI and persistent post-concussive symptoms (378). Participants were randomized to receive 40 HBO2 (1.5 atmospheres absolute [ATA], ⟩99% oxygen, 60 minutes) or sham chamber sessions (1.2 ATA, room air, 60 minutes) over 12 weeks. Participants and evaluators were blinded to allocation. Outcomes assessed at baseline, 13 weeks, and 6 months included symptoms, quality of life, neuropsychological, neurologic, electroencephalography, sleep, auditory, vestibular, autonomic, visual, neuroimaging, and laboratory testing. Seventy-one randomized participants received HBO2 (n=36) or sham (n=35). At baseline, 35 participants (49%) met posttraumatic stress disorder (PTSD) criteria. By week 13, the HBO2 group had improved Neurobehavioral Symptom Inventory scores (mean change -3.6 points, P=0.03) compared to the sham group (+3.9 points). In participants with PTSD, the change with HBO2 was more pronounced (-8.6 vs. +4.8 points with sham, P=0.02). PTSD symptoms also improved in the HBO2 group, and more so in the subgroup with PTSD. Improvements regressed at 6 and 12 months. Hyperbaric oxygen improved some cognitive processing speed and sleep measures. Participants with PTSD receiving HBO2 had improved functional balance and reduced vestibular complaints at 13 weeks.

A systematic review and pooled analysis of four Department of Defense studies looked to summarize available evidence for HBO2 in mTBI-associated post-concussive symptoms ± post-traumatic stress disorder (147). Participant-level data (n=254) were grouped into pooled HBO2 and sham intervention groups. Changes from baseline to post-intervention on post-concussive symptoms, post-traumatic stress disorder, and neuropsychological measures were assessed. The results demonstrated improvement in patients receiving HBO2 for post-concussive symptoms (Rivermead Total Score: -2.3, 95% CI [-5.6, 1.0], p=0.18) and post-traumatic stress disorder (PTSD Checklist Total Score: -2.7, 95% CI [-5.8, 0.4], p=0.09) as well as significant improvement in verbal memory (CVLT-II Trial 1-5 Free Recall: 3.8, 95% CI [1.0, 6.7], p=0.01). A dose-response trend to increasing oxygen partial pressure was also found, with a greater HBO2 effect in mTBI-associated post-traumatic stress disorder suggested.

Supplements and vitamins. At present no supplement has been approved by the FDA to treat concussion. Animal studies have shown promise in severe traumatic brain injury patients; however, these did not translate into human studies. Compounds such as omega 3 fatty acids, N acetyl cystine, caffeine, vitamins C, E, D, B2, magnesium, nicotinamide, ribose, melatonin, branch chain amino acids, curcumin, resveratrol, alpha lipohillic acid, creatine, and S baicalensis are currently being studied or are in planned studies to prevent concussion, improve concussion symptoms, and prevent long-term neurocognitive effects (364).

Return to play/duty/work. There have been a number of guidelines published to assist physicians and other clinicians in determining the readiness of athletes to return to play after sports concussion. These include The Cantu Guidelines (1986), which have been adopted by the American College of Sports Medicine; Colorado Medical Guidelines (1991), which were the basis for the 1997 AAN guidelines and have been adopted by the National Collegiate Athletic Association (NCAA); American Academy of Pediatrics; The American Academy of Neurology Guidelines (1997, 2013); and international guidelines from Vienna (2001), Prague (2004), and Zurich (2008, 2012). It is extremely important that any individual involved in returning an athlete to play understand that all current return to play guidelines are based almost solely on consensus opinion and not evidence-based studies. The most widely used return to play guidelines are the Zurich guidelines, which have their roots in the initial Cantu guidelines and are the basis for most of the guidelines listed above. All current updated guidelines and position statements require any athlete who is suspected of sustaining a concussion to be immediately removed from play. Return to play protocol should then follow a stepwise process based on a full clinical and cognitive recovery before the athlete is allowed to return to play. With this stepwise progression, the athlete should continue to proceed to the next level if asymptomatic at the current level. Each step should take 24 hours, so an athlete would take approximately 1 week to proceed through the full rehabilitation protocol. If any postconcussion symptoms occur while in the stepwise program, the patient should drop back to the previous asymptomatic level and try to progress again after a further 24-hour period of rest has passed (239; 240; 133).

Unlike athletes, there are no specific guidelines on returning a concussed individual to work. The current American College of Occupational Medicine practice guidelines have no recommendations (ie, evaluation, management, and return to work). However, new evidence-based guidelines are in process (Conidi, draft) and should answer this question. There are a few nonevidence-based recommendations such as those from the Official Disability Guidelines (384). Patients diagnosed with a mild concussion should return to work in 3 to 7 days. Those with a severe concussion without cognitive symptoms should return to a modified work schedule in 14 days, and those with a severe concussion with cognitive symptoms should return to work within 84 days.

The military has updated their guidelines, which are in line with what is seen in sports. The regulations require mandatory reporting of all suspected concussions and medical clearance prior to returning to duty. In addition, any service member who has been involved in a blast event, collision, or roll over; been within 50 meters of a blast; been exposed to more than one blast event; or suffers a direct blow to the head should be evaluated using an IED checklist (I=injury, E=evaluation, D=distance from the blast). Any member with a “yes” answer on the IED is then referred for further testing, including the MACE, along with a complete neurologic exam. Furthermore, any potentially concussive event requires 24 hours’ rest. A first concussion requires at least 24 hours’ rest and further if the symptoms persist, and a second diagnosed concussion within a 12-month period requires a delay of return to duty for 7 days after the symptoms resolve. Finally, three or more concussions in a 12-month period require a recurrent concussion evaluation, which entails a comprehensive neurologic exam, possible neuroimaging, neuropsychological assessment, functional assessment, and a duty status determination. Determination of duty status is usually conducted by a neurologist who reviews the above testing; once the neurologist is convinced the service member is completely asymptomatic, an appropriate return to duty schedule is decided (90).

Special considerations

Pregnancy

There are no reports on the influence of pregnancy on minor head injury. For protection of the fetus, shielding should be used whenever radiographs or CT are performed.

Anesthesia

Concussions and the National Football League (NFL)

In 2009, investigators released landmark studies on the brains of retired NFL players and other high-impact sports athletes, documenting a rare neuropathological or neurodegenerative disorder ie, chronic traumatic encephalopathy (245; 275). Since that study, researchers have begun to look at the long-term effects of concussion and subconcussive hits on retired NFL alumni. One such study commissioned by the NFL found that 6.1% of all players over 50 years of age were likely to receive a diagnosis of dementia, Alzheimer disease, or other form of dementia (which is over 5 times what is expected in those of similar age in the general population). More shocking were the numbers on the 50 years and younger age group who were 20 times more likely to carry one of the aforementioned diagnoses when compared to age-matched non-alumni (379). Overall, retired NFL players are 3 times more likely to die of a neurodegenerative disorder and 4 times more likely to die of Alzheimer disease or amyotrophic lateral sclerosis.

From an assessment and management standpoint, the NFL now has the most comprehensive concussion management program of any professional sports league. The program is administered by the Head, Neck, and Spine Medical Committee, which is chaired by two neurosurgeons. As a result of the collective bargaining agreement, any recommendations from the committee must be cleared by the NFLPA Mackey/White TBI committee, which currently has four member neurologists. Players are required to undergo preseason evaluations, including a comprehensive neurologic evaluation with concussion history and baseline evaluation, ie, a tablet-based NFL sideline concussion tool along with computerized neuropsychological testing (most teams use IMPACT). All teams have in their locker rooms an educational poster about sports concussion. In 2012, the NFL enacted a new rule that during each game a nonaffiliated athletic trainer must be situated in the press box to monitor the field and sidelines for potential concussions. If a concussion is suspected, the athletic trainer informs the team’s medical staff, and the player must be evaluated per current concussion protocol. Any player suspected of suffering a concussion is evaluated by the medical staff either on the sideline or in the trainer’s room. The player is given an initial screening containing six signs for obvious disqualification along with a neurologic screen for cervical spine injuries or serious brain trauma. If the player passes this screening, they are then evaluated (tablet-based) with the Sports Concussion Assessment Tool, version 3 (SCAT-3), and the scores are immediately compared to baseline. In addition, in 2013 the NFL instituted a new rule requiring an unaffiliated “neurotrauma consultant,” ie, neurologist, ER physician, or neurosurgeon with an affiliation with a local hospital (the NFLPA recommended this be a neurologist) be stationed on each sideline to provide a second opinion on the assessment of any potentially concussed players. When a player has been deemed to have suffered a concussion, his is immediately removed from play. During the recovery process, the player is usually followed by the team’s medical staff and neuropsychologist (who administers follow up computerized and, in some cases, paper and pencil neuropsychological testing). Once the player is asymptomatic and off all medications, with a normal neurologic examination, and computerized neuropsychological and SCAT-3 testing have returned to baseline, the player is cleared to undergo a graded return to play protocol as outline in the Zurich Consensus Statement (240). After the athlete has completed the above, they are then required to undergo an evaluation by an independent neurologist who has been approved by the Head, Neck, and Spine Committee. When the consulting neurologist, medical team, and neuropsychologist are all in agreement, the player is then cleared for full-contact practice and subsequently for game play.

Perhaps surprising to an outside observer is the lack of studies in active players. However, most NFL players and professional athletes in general are reluctant to participate in clinical studies for fear the results may be used against them to limit participation, terminate their career, or affect contract negotiations. There are a few retrospective or observational studies, including one by Kumar and colleagues who examined the short-term effects of concussion on player performance after returning to play (195). The data were obtained from the NFL (ie, league profiles) and included injury reports, player age, career experience, games misses, and Pro Football Focus performance scores from 2008 to 2012 (interesting as a significant change in NFL concussion management occurred in 2009). One hundred and twenty-four players qualified for the study (defensive secondary, wide receiver, and offensive line were the positions that were most at risk for concussion), and surprisingly 55% missed no games after sustaining a concussion. Players who missed at least one game were younger and less experienced; they were also more likely to have sustained a second or repeat concussion (69.5%). Older players and players sustaining late-season concussions were more likely to return to play without missing any games, with the odds of returning within 7 days increasing by 18% for each career year, and by 40% for each game before the final game of the season. However, the aforementioned odds decreased by 85% after the league introduced new concussion guidelines in 2009, requiring evaluation by an independent neurologist. No difference in player performance was found whether the player did or did not miss a game prior to return to play.

Studies have emerged that support the retired players’ claims that years of playing professional football has resulted in many developing early dementia. Most have used advanced neuroimaging in conjunction with neuropsychological or behavioral testing and have found a higher incidence of cognitive issues, depression, and white matter abnormalities in retired NFL alumni. Researchers attempted to correlate cognitive impairment and depression with structural abnormalities on DTI-MRI and blood flow abnormalities using arterial spin labeling and phase contrast MRI (89; 148). They looked at 34 former NFL players (mean age 61.8 years); 20 were normal, four had a fixed cognitive deficit, eight had mild cognitive impairment, and two had dementia (1 classified as vascular dementia). Eight (24%) were diagnosed as having depression, and 3 of the 8 had cognitive difficulties that could not be attributed to depression. Interestingly, the researchers were unable to correlate concussions or years played with neuropsychological abnormalities. On conventional MRI (ie, total and deep white matter volumes), there was a significant difference between players with cognitive deficits and age-matched controls on tests of naming, word-finding, and episodic visual and verbal memory. On DTI imaging, NFL players with cognitive impairment and depression had statistically significant reductions in fractional anisotropy in the bilateral frontal or parietal regions as well as the corpus callosum and left temporal lobe when compared to the control group. Regional blood flow differences (left temporal pole, inferior parietal lobe, and superior temporal gyrus) in the cognitively impaired group corresponded to regions associated with impaired neurocognitive performance. Two other studies by the same group compared depressive symptoms with white matter dysfunction and the number of concussions (89; 344). Both studies used the Beck Depression Inventory version II (BDI-II) and looked at a 3-factor model of depressive symptoms (affective, cognitive, and somatic). The cognitive factor was the only variable that correlated with the number of concussions. However, when using a voxel-wise DTI protocol, there was a negative correlation with fractional anisotropy in the frontal region and total BDI as well as the cognitive and somatic subfactors, and a partial correlation with the affective subfactor. There were also negative fractional anisotropy correlations using region-of-interest analysis and total BDI, with fractional anisotropy abnormalities in the forceps minor differentiating depressed from nondepressed athletes. Our group has also found deficits on both conventional and DTI-MRI, neuropsychological testing, and depression inventories. Of 40 retired NFL players tested, 42% have shown significant abnormalities on DTI-MRI (ie, fractional anisotropies of greater than 2.5 standard deviations when compared to age-matched controls). Functional MRI (fMRI) analysis of retired players looking at brain activation patterns has found pronounced hyperactivation and hypoconnectivity in the dorsolateral and prefrontal cortices that correlated with the number of times a player reported being removed from play after head injury. Computerized neuropsychological testing only showed small differences in executive function in the retired alumni when compared to age-matched controls. The results suggest that the athletes have developed cortical compensatory mechanisms (ie, frontal lobe subregions work harder) to counter the neurologic impact of repetitive head trauma (143). Contrary to the above studies, Kuhn and colleagues used a multivariable analysis (ie, conventional and DTI-MRI), neuropsychological testing (paper and pencil and computerized), biomarker analysis (APOE4 status), and depression scales (BDI and Patient Health Questionnaire) with the hypothesis that in-depth analysis of retired NFL players was unlikely to detect objective clinical abnormalities in a majority of subjects (194). Of the 45 retired NFL players evaluated, four (9%) had microbleeds, with three (7%) having a large cavum septum pellucidum with brain atrophy (a possible marker for chronic traumatic encephalopathy). The number of concussions was associated with abnormal results of DTI-MRI. Neuropsychological testing revealed isolated impairments in 11 players (24%); however, none were felt to have dementia. The APOE4 allele (a marker for Alzheimer disease and possibly chronic traumatic encephalopathy) was present in 38% of the players (a larger number than would be expected in the general male population).

On July 7, 2014, federal Judge Anita Brody granted preliminary approval of a settlement between retired NFL players, their family or representatives, and the NFL and NFL properties. The lawsuit was finalized in April of 2015 and there have been numerous challenges and modifications since. There have has been claims of fraud and deceptive practices from both sides. The retired NFL players sued, accusing the NFL of not warning players and hiding the damages of brain injury. Players and their families are not required to prove that playing professional football caused their injuries. The proposed settlement provides three benefits: baseline medical exams (by a neurologist or neuropsychologist) for retired NFL players ($75 million); monetary awards ($765 million and possibly more in total) for diagnoses of amyotrophic lateral sclerosis, Alzheimer disease, Parkinson disease, dementia, and certain cases of chronic traumatic encephalopathy; and education programs and initiatives related to football safety ($10 million).

The settlement recognizes two subclasses of plaintiffs. Those diagnosed with amyotrophic lateral sclerosis, Alzheimer disease, Parkinson disease, level 2 cognitive impairment (moderate dementia), level 1.5 cognitive impairment (early dementia), or a pathological diagnosis of chronic traumatic encephalopathy before July 7, 2014 and those diagnosed after. Individuals diagnosed prior to the cutoff date will receive higher monetary awards. In order to receive monetary awards, retired players with at least one half of an eligibility season will need to undergo a baseline assessment through the Baseline Assessment Program.

The Baseline Assessment Program uses a nationwide network of qualified and independent medical providers who provide the initial baseline assessment as well as any further testing or treatment. The Baseline Assessment Program Administrator, appointed by the Court, establishes the network of medical providers, ie, neurologists and neuropsychologists. Providers must be approved by both the NFL lead consul and the co-consul for the retired players and cannot serve as an expert witness for those players who opt out, nor could they have served as an expert witness for the lawsuit. Criteria for selection includes education, licensure, credentials, the ability to see players in a timely manner, geographic location, and the ability to provide on-site evaluations per the Baseline Assessment Program. Interestingly, those who elect to forgo being evaluated through the BAP program are still eligible for monetary awards (which appear to be lower) through the MAF tract. However, they will need to be evaluated by board-certified neurologists, board-certified neurosurgeons, or board-certified neurospecialist physician or similarly qualified specialists approved by the claims administrator. Retired players over age 43 have 2 years from the start of the Baseline Assessment Program to undergo a baseline examination, and retired players under age 43 have 10 years.

Current BAP testing consists of a mini-mental status examination, clinical dementia rating scale (CDR), a pre-defined battery of neuropsychological tests, an estimated pre-morbid function test (Test of Premorbid Functioning; TOPF), and current IQ analysis. There are five domains of cognitive functioning tested; each domain contains several sub tests. For example, looking at the domain of complex attention and processing speed, there are six subtests/test scores (digit span, arithmetic, letter and number sequencing, coding, cancellation, symbol search, cancellation). Other domains include executive functioning (four subtests), learning and memory (six subtests), language (three subtests), vision and perception (three subtests). The scores are then compared to demographically adjusted normative data for Caucasians and African Americans and a standard deviation (SD) below expected norm; a player must have more than one low test score in each domain.

The basic principle in defining impairment is that the neuropsychologist or neurologist must determine that a pattern of performance decline is determined (1.5 SD below expected for level 1 impairment, 1.7 to 1.8 SD below expected for level 1.5 impairment, or 2 SD below for level 2 impairment). Complicating things further is that subtests are actually weighted based on estimated intellectual functioning, with those having a below average level of pre-morbid intellectual functioning needing more tests to be abnormal than those with above average functioning. Alumni also undergo freestanding-, embedded-, and regression-based performance validity metrics along with testing for depression (MMPI-2RF and M.I.N.I. Version 5.0.0). Interestingly, BAP physicians are unable to diagnose Parkinson disease, amyotrophic lateral sclerosis, and Alzheimer disease despite being board certified neurologists.

From a clinical and real-world medicine standpoint the lawsuit does not follow what most neurologists would consider to be standard of care for workup of dementia in the young (ie, a brain MRI and blood work for reversible causes), is filled with fictional diagnoses, and does not address the issue of traumatic brain injury, which is the likely cause of most players’ symptoms. Complicating things further are guidelines by the International Working Group (IWG) and National Institute on Aging and Alzheimer’s Association (NIA-AA), and more recently the American Academy of Neurology (AAN), which incorporate the use of biomarkers in the diagnosis (ie, PET imaging, CSF tau, and A-Beta protein). The NFL lawsuit has no such requirements, and many have argued that individuals with positive biomarkers have met the criteria for Alzheimer disease. Overall, the NFL diagnostic criteria most closely resemble the outdated NIA-AA classification, which breaks the diagnoses into a pre-clinical component, mild cognitive impairment, and Alzheimer disease dementia. Unlike the NFL criteria, which make the diagnosis when the retired athlete has exhibited significant cognitive decline, the IWG guidelines and AAN recommendations focus on the early diagnosis of Alzheimer disease. They also rely on a more focused and simplified neuropsychological battery focusing on impaired performance on episodic memory testing. The IWG recommendations have diagnostic criteria for atypical Alzheimer disease and its associated phenotypes (occipitotemporal, biparietal, frontal, logogenic) whereas the NFL does not.