This review examines the clinical and neuroradiological features of traumatic brain injury that are most frequently associated with persistent cognitive complaints. Neuropsychological outcomes do not depend solely on brain injury severity but result from a complex interplay between premorbid factors, the extent and nature of the underlying structural damage, the person’s neuropsychological reserve and the impact of non-neurological factors in the recovery process. Brain injury severity is only one of these factors and has limited prognostic significance with respect to neuropsychological outcome. We examine the preinjury and postinjury factors that interact with the severity of a traumatic brain injury to shape outcome trajectories. We aim to provide a practical base on which to build discussions with the patient and their family about what to expect following injury and also to plan appropriate neurorehabilitation.
- traumatic brain injury
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The recent Lancet Neurology Commission on Traumatic Brain Injury highlighted the scope and considerable public health challenge of traumatic brain injury, estimating that about half of the world’s population will experience one or more traumatic brain injuries in their lifetime.1 In particular, the Commission highlighted the ‘urgent need for (the) development, validation and implementation of prognostic models in traumatic brain injury, particularly less severe traumatic brain injury.’ About 80%–90% of all traumatic brain injury cases fall into less severe classifications.2
Traumatic brain injury classifications
The severity of a traumatic brain injury was traditionally determined by several clinical indices, including loss of consciousness, Glasgow Coma Scale scores,3 presence/duration of retrograde and post-traumatic amnesia, and neuroradiological evidence of cerebral damage (see refs4 5 for reviews). Several classification systems have combined these clinical metrics to give an overall measure of the severity of the traumatic brain injury (table 1).
While these classification systems have a high degree of inter-rater reliability in terms of classifying the severity of a traumatic brain injury, they have limited prognostic value in terms of long-term neuropsychological outcome. Neuropsychological outcomes following traumatic brain injury do not depend solely on the severity of a brain injury but result from the complex interplay between premorbid factors, the extent and nature of the underlying structural damage, the person’s neuropsychological reserve and the impact of non-neurological factors in the recovery process.
Pathoanatomic classifications of traumatic brain injury
Neuroimaging has increasingly allowed clinicians to classify traumatic brain injury on a pathoanatomic basis, by using both the clinical features and the radiological appearances to define the features and location of traumatic abnormalities. Pathoanatomic features may be a more important determinant of long-term neuropsychological outcome than classifications based on clinically defined severity alone.2 4
Injuries may arise as a primary consequence of the trauma (through direct contact and/or through acceleration-deceleration forces) or secondary to subsequent non-traumatic consequences arising from the primary insult. The injuries may be focal, diffuse or both. Focal injuries include skull fracture, haemorrhage and contusions. Diffuse injuries include diffuse axonal injury, hypoxic-ischaemic damage, meningitis and/or vascular injury.6
Intraparenchymal haemorrhage or intracerebral haemorrhage refers to significant bleeding within the brain parenchyma. Pathologically, contusions and intraparenchymal haematomas exist along the same continuum. In a contusion, blood is intermixed with brain tissue7; radiologically a contusion becomes an intraparenchymal haematoma once two-thirds or more of the lesion is blood.8 Up to one-third of all contusions enlarge in the subacute phase; thus scans taken immediately after injury may underestimate or miss the extent of damage. A metabolic reaction is triggered in adjacent tissue, which peaks about 5 days after injury. Mechanically, the inferior surfaces of both the frontal and temporal lobes are particularly vulnerable to contusional injury, due to the irregular bony surface floor of the frontal and middle cranial fossa. Parenchymal injuries in these regions can lead to persistent focal neuropsychological deficits.
Haemorrhage and haematoma
Small haematomas resulting from epidural and subdural bleeds tend to be associated with fewer long-term neuropsychological problems than with other severe traumatic brain injuries. These injuries usually have a good neuropsychological prognosis, if there are no significant pressure effects or an oedematous or inflammatory response in the underlying parenchyma, and if the bleed does not require surgical management. However, the risk of morbidity (and indeed mortality) increases significantly if the bleed requires surgical intervention, particularly if this is delayed.9
Some brain regions tolerate mass effects better than others. The mesial temporal lobes are particularly susceptible to damage during this process8 and persistent impairments in language and memory functions, particularly new learning, can become apparent once the acute symptoms have resolved. The chance of developing long-term neuropsychological sequelae following a subdural haematoma is significantly increased if the resulting haematoma is large enough to cause pressure effects, particularly if there is visible midline shift and surgical decompression is required.
Intraventricular haemorrhage (bleeding into the ventricular system) can result from trauma, most commonly developing from a primary traumatic subarachnoid bleed. Survivors of injuries associated with raised intracranial pressure have a high risk of being left with significant and widespread cognitive difficulties.
Diffuse axonal injury
The definition of diffuse axonal injury has evolved over time. Originally reported in animal and human postmortem studies, the concept of diffuse axonal injury described the mechanisms of injury in people with very poor outcomes (coma or death) following severe head injuries but without significant intracerebral haemorrhage or a mass lesion such as a haematoma. However, with the advent of MRI techniques sensitive to these mechanisms of injury,10 11 patterns attributed to traumatic axonal injuries can occur in patients with far less severe traumatic brain injuries.
In a systematic review and meta-analysis, van Eijck et al 12 considered the functional prognosis, as measured by the Glasgow Outcome Scale/Glasgow Outcome Scale-Extended, in 32 studies of adult patients with traumatic brain injury and different patterns of diffuse axonal injury. They found that 38% of patients had an unfavourable overall functional outcome associated with diffuse axonal injury, representing a threefold greater risk of unfavourable outcome compared with patients without diffuse axonal injury. Furthermore, each of the three MRI grades of severity of diffuse axonal injury was associated with a threefold increase of risk of a poor outcome. However, 62% of patients with diffuse axonal injury had a favourable outcome: thus one must not assume that its presence necessarily implies a poor outcome.
Diffusion tensor imaging
Since the first diffusion tensor imaging (DTI) study in traumatic brain injury (2002), the number of articles published in this field has increased exponentially.13 Hayes et al 14 looked at white matter abnormalities and concluded that traumatic brain injury is associated with altered structural and functional connectivity, characterised by decreased integrity of white matter pathways.
De Simoni et al 15 used functional MRI and DTI to look at the structural and functional basis of post-traumatic amnesia. Intriguingly, DTI showed widespread traumatic axonal injury in patients with post-traumatic amnesia compared with the control group. However, there is as yet no established direct causal link between post-traumatic amnesia and white matter damage.
Kinnunen et al’s landmark study used DTI to study white matter damage after predominantly moderate to severe traumatic brain injury.16 They concluded that widespread white matter abnormalities persist following traumatic brain injury. The location of DTI white matter abnormality partially predicted cognitive function. The structure of the fornices correlated with associative learning and memory across both patient and control groups, while the structure of frontal lobe connections showed relationships with executive function that differed between the two groups.
However, in their comprehensive review of the first 100 DTI studies published, Hulkower et al found little consensus regarding the neuropsychological correlates of the DTI changes.13 They identified 51 papers that had examined the relationship between DTI measures and performance on some measures of neuropsychological test performance, but found very few confirmatory studies reporting the DTI correlates of the same measure. Even among articles examining the same cognitive domain, the results were inconsistent. The authors hypothesised that the discordance between the findings in the cognitive studies may have been due to the task specificity of the neuropsychological measures employed and differences in the severity of the brain injury of the subjects who took part in the studies. They also noted that the cognitive impairment following mild traumatic brain injury could be subtle and not picked up on formal neuropsychological tests, indicating that these structural changes can occur without objective measures of neuropsychological impairment. DTI appearances probably also change over time. In their review of the DTI literature, Eierud et al 17 found some evidence to suggest that DTI scans in the acute phase following a mild traumatic brain injury showed increased fractional anisotropy scores—a measure of diffusion directionality and fibre integrity—while these scores were reduced in scans taken some time after the injury.
The discrepancies in the literature on the neuropsychological correlates of DTI abnormalities probably reflect a combination of all these factors. This is a fast-evolving field but at present there is no clear consensus on the significance of an abnormality on these imaging measures of axonal integrity and objective measures of neuropsychological function in patients whose other clinical indices indicate that they have sustained a mild traumatic brain injury. Clinicians must therefore remember to distinguish these patterns of axonal injury from the classic descriptions of diffuse axonal injury. Classical diffuse axonal injury is associated with significant neurological and neuropsychological morbidity whereas axonal injuries evident on DTI may have only subtle or absent neuropsychological correlates, particularly in those with mild traumatic brain injury. The appearances on specialist MRI sequences following mild traumatic brain injury may be more usefully described as ‘traumatic axonal injuries’ to distinguish them from the more serious diffuse axonal injury.2
There is a complex relationship between symptom presentation in mild traumatic brain injury and DTI abnormalities. Controlled studies suggest that these abnormalities are not significantly associated with greater symptom reporting in patients with mild traumatic brain injuries, and that these microstructural anomalies are also identified in 1 in 8 healthy controls.18 DTI abnormalities are not specific to brain trauma and also occur in psychiatric conditions including post-traumatic stress disorder resulting from childhood trauma, anxiety and depression.19–21 Medication can also influence DTI appearances, such that people with opiate addiction may show increased radial diffusivity (indicating myelin pathology).22
Secondary injuries refer to the cerebral damage caused by events following the primary injury; these may directly follow the pathophysiological cascade triggered by the primary injury, or may be the neurological consequences of other physical injuries sustained in the trauma. Secondary injuries can result from hypoxia, raised intracranial pressure, hypercarbia, hyponatraemia and seizures. The presence of any of these processes significantly increases the likelihood of developing long-term neuropsychological sequelae following a traumatic brain injury and the impact is cumulative: the more injuries the brain sustains, the less likely is a full neuropsychological recovery.
Non-brain injury factors associated with neuropsychological recovery
The authors of a large-scale, well-controlled study of the biological, psychological and social factors that govern recovery concluded that the extent to which someone reports persistent symptoms following a mild traumatic brain injury represents:
‘the cumulative effect of multiple variables, such as genetics, mental health history, current life stress, medical problems, chronic pain, depression, personality factors, and other psychosocial and environmental factors. The extent to which damage to the structure of the brain contributes to the persistence of postconcussion symptoms remains unclear.’18
NB: Since postconcussion symptoms appear both common and non-specific, the International Collaboration on Mild Traumatic Brain Injury Prognosis (ICoMP) recommends abandoning the diagnostic label of ‘post-concussion syndrome’ for the more appropriate ‘post-traumatic symptoms’23 (see also Sharp and Jenkins’ persuasive arguments along the same lines in ‘Concussion is confusing us all’24).
Although the literature clearly demonstrates the multifactorial basis of these symptoms, neuroradiological findings can dominate neurological follow-up, both in terms of diagnosis and treatment of persistent cognitive complaints. However, evidence-based practice should encompass an examination of the broader biopsychosocial factors that underpin these complaints.
Traumatic brain injury is a non-random event. Males are twice as likely as females to experience a traumatic brain injury and incidence rates vary significantly in different age groups, with the highest incidence in children aged less than 4 and in adolescents aged 15–19 years.25 Clinicians should carefully consider premorbid factors when assessing a person following head injury and predicting their eventual outcome. A history of previous head injury, prior drug and alcohol misuse, and older age predict a worse outcome.26–28 In their systematic review of prognosis following mild traumatic brain injury, the WHO ICoMP team showed that in mild traumatic brain injury, premorbid factors predicted long-term outcomes better than the severity of the injury.29
It is easy to overlook these non-organic factors in the neurology clinic, particularly when there has been a clear traumatic brain injury and both the patient and family may begin to view the past through a ‘halo’ lens. Neuropsychological difficulties may be presented as ‘new’ when really they are a continuation or exacerbation of a pre-existing problem. This is an important distinction. New problems may well have an organic basis; exacerbations may also have an organic basis but also may represent the impact of psychological stress or pain, since normal inhibitions are often reduced. In their systematic review of prognosis following mild traumatic brain injury, the ICoMP experts found that adult patients after mild injury tended to devalue and minimise pre-existing symptoms and misattribute these symptoms to the injury itself.23 29
People with pre-existing health anxieties or medically unexplained symptoms often have persistent and severe neuropsychological complaints, even following mild traumatic brain injuries with no evidence of underlying structural damage. Furthermore, a history of emotional distress and maladaptive coping early after injury predicts functional outcome at 6 months.10 Clinicians need to be alert to the fact that patients may recount inaccurate details of their accident and their subsequent injuries in the neurology clinic. These exaggerations may directly relate to litigation in medicolegal cases, but erroneous beliefs can also develop as a subconscious response to the pain, fear and psychological trauma someone has experienced in their accident. Obtaining a verified history—with reference to the original medical records of the accident and review of preaccident primary care records—should be the first step in planning the appropriate treatment, and before investigating increasingly unlikely explanations associated with their injury.
Postaccident recovery of function
There is now a considerable body of evidence to suggest that non-organic factors play the primary maintenance role in persistent neuropsychological complaints after an uncomplicated mild traumatic brain injury.18 30 However, even in those who have sustained more significant cerebral trauma, non-brain injury factors often significantly impact (either exacerbatory or palliatively) on the associated neuropsychological deficits leading to their everyday function. The cause of neuropsychological difficulties in this group is not a dichotomy (brain injury vs psychological) but rather exists along a continuum with contributions from many non-organic factors.
Many of the psychiatric conditions that often develop after traumatic brain injury are associated with neuropsychological compromise including major depression, anxiety disorder, adjustment disorder and post-traumatic stress disorder. However, it is not necessary to have a diagnosable psychiatric condition or irreparable brain injury to experience cognitive difficulties. Pain and sleep disruption can significantly impact on neuropsychological function. Similarly, the experience of simply being told that you have a brain injury can lead to ‘diagnosis threat’31 or ‘expectation as aetiology’,32 particularly in vulnerable patients. Patients can develop hypervigilance to normal cognitive failings, particularly those associated with normal age-related decline,33 creating a vicious cycle of increased anxiety and vigilance, leading to further difficulties. This process can be exacerbated in those pursuing compensation for their injuries, where lawyers may advise keeping a ‘diary’ of all cognitive failings to support their claim. The medicolegal process itself can be a source of significant stress in the claimants’ life following their traumatic brain injury, particularly when there are conflicting expert opinions about the severity of their injuries and likely prognoses.
The indications and limitations of a neuropsychological assessment in traumatic brain injury
There are many reasons why someone may fail, or do poorly on, a standardised neuropsychological test. Failure does not necessarily indicate underlying organic dysfunction. Neuropsychological test scores have limited value as stand-alone measures in the differential diagnosis of an organic versus non-organic basis of cognitive dysfunction. Failures increase with the number of tests administered in an assessment, and depend on individual differences in age, education, sex and estimated premorbid IQ.34 In one study of over 1000 neurologically normal adults who completed 25 or more neuropsychological tests, 87% produced at least one score that was 1 or more SDs below the norm. Over one-third of the sample produced abnormal scores on at least five of the tests. When using a more stringent cut-off of 2 SD below the norm, 28% of the sample still produced at least one ‘impaired’ score.35
People may also perform very poorly on tests due to high levels of anxiety, low mood and other functional disturbance. This is particularly so when evaluating executive functions, which can be challenging to assess in office-based assessments. While symptom validity tests can help identify people who may be deliberately exaggerating their cognitive difficulties, they do not differentiate between those with an organic versus a functional basis for their cognitive impairment.
In many patients following traumatic brain injury, particularly those with ‘red flag’ organic characteristics for persistent neuropsychological difficulties (table 2), their neuropsychological profile reflects a combination of functional and organic influences on function. It is important to recognise this functional component, since these factors can be successfully treated, significantly improving everyday function. A multidisciplinary approach involving neurology, neuropsychology and neuropsychiatry is critical to maximising their outcomes.
Multiple factors determine neuropsychological outcome following traumatic brain injury, and the outcome does not depend solely on the severity of the brain injury. There is increasing support for the biopsychosocial model that conceptualises outcome as the result of a complex interplay between premorbid factors, the extent and nature of the sustained structural damage, the person’s neuropsychological reserve and non-neurological factors impacting on the recovery process.
The neuropsychological sequelae of traumatic brain injury may depend more on the characteristics of underlying pathoanatomical injury than the severity of the injury, as defined by traditional classification systems. Many patients make a complete or near-complete recovery from uncomplicated intracerebral bleeds and small contusions. The chances of significant long-term neuropsychological deficits resulting from these injuries are significantly increased if the initial injury was complicated by an inflammatory response or other secondary injuries. The presence of contusions on an acute CT scan of head can be associated with focal neuropsychological deficits, in the context of otherwise preserved function. Traditionally defined diffuse axonal injury remains a severe traumatic brain injury associated with significant neuropsychological and neurological morbidity. There are no consistently demonstrated neuropsychological correlates for the axonal injuries evident on DTI in patients who have sustained mild traumatic brain injuries, although processing speed may be reduced. These abnormalities are not unique to patients with traumatic brain injury18 and it is therefore important not to conflate these imaging findings with the prognoses for neuropsychological outcome associated with traditional diffuse axonal injury.
There are several reasons why someone may develop cognitive problems following traumatic brain injury. An integral part of neurological follow-up of these patients should be to identify and examine the role of non-organic factors in causing and maintaining cognitive problems following traumatic brain injury. This is particularly important in those who report difficulties disproportionate to those expected given the nature of their injury, or in those whose expected recovery trajectory appears prematurely stalled or regressive. This patient group often has potentially treatable exacerbating comorbidities. Erroneously attributing these difficulties to ‘brain injury’ can lead to unnecessary medical interventions and chronic learnt helplessness.30 Identifying and rapidly treating these remediable issues can ensure that each patient maximises their recovery potential.
The severity of a traumatic brain injury can become exaggerated over time in a patient’s records; where possible, check the acute records before classifying or treating this as severe.
Neurologists should be alert to the presence of non-neurological factors in the presentation of persistent neuropsychological complaints following traumatic brain injury, even in the context of abnormal imaging.
Erroneously attributing persisting neuropsychological complaints to severe brain injury can have a significant detrimental impact on a person’s recovery.
Early referral for neuropsychological assessment and intervention can help identify non-organic and potentially remediable factors to optimise recovery following traumatic brain injury.
Contributors All authors contributed equally to the conception and authorship of this paper.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Commissioned; externally peer reviewed by Colette Griffin, London, UK, and Alan Carson, Edinburgh, UK.