Geriatric Traumatic Brain Injury: Epidemiology,Outcomes,Knowledge Gaps,and Future Directions

Prevalence and incidence

TBI is extremely common, with a lifetime prevalence of up to 40% among adults. ⁴⁰ Whereas incidence of TBI peaks three times over the life course—in childhood, adolescence, and older adulthood—the highest incidence of TBI occurs in older adults (Fig. 1). More than 1 in 200 Americans ages 65–74 years and more than 1 in 50 Americans ages ≥75 years experienced a TBI-related ED visit, hospitalization, or death in 2013, with the next highest incidence occurring in children ages 0–4 years (incidence 1,591 in 100,000). ¹ In 2013, adults ages ≥75 years accounted for 26.5% of all TBI-related deaths and 31.4% of all TBI-related hospitalizations in the United States. From 2006 to 2008, 800,000 adults ages ≥65 years were evaluated in U.S. EDs for TBI, among whom the average age was 80 years. ⁴¹ Incidence of admission to acute inpatient rehabilitation for a primary diagnosis of TBI also increases with age, with the highest incidence occurring among those ages ≥80 years. ⁵

OPEN IN VIEWER

FIG. 1.

Annual incidence of traumatic brain injury (TBI) emergency department (ED) visits, hospitalizations, and deaths 2002–2013 by age. Annual incidence of TBI-related ED visits, hospitalizations, and deaths per 100,000 U.S. population are shown for the periods 2002–2006 (A), 2007 (B), and 2013 (C). Over time, incidence of TBI has shown the greatest increase among adults ≥75 years of age, with most of this increase attributed to increased ED visits. Data adapted from the Centers for Disease Control. ^1,3

Incidence of TBI-related ED visits, hospitalizations, and deaths is increasing among older adults, whereas incidence of TBI-related hospitalizations and deaths is declining among children and adults <55 years of age ⁴² (Fig. 1). From 2007 to 2013, TBI-related ED visits among those ages ≥75 years doubled and TBI-related hospitalizations increased more than 25%. ¹ The rapid rise in TBI-related hospital visits among the oldest segment of the U.S. population exceeded population growth during this time frame. ⁹ This high and increasing incidence of TBI-related ED visits, hospitalizations, and deaths among older adults has been confirmed in multiple epidemiological studies in individual U.S. states and nation-wide databases ^{4,43 –45} as well as in higher-income countries around the globe including Spain, ⁴⁶ the United Kingdom (UK), ⁴⁷ Scotland, ⁴⁸ the Netherlands, ⁴⁹ Austria, ⁵⁰ Finland, ⁵¹ Canada, ^{52 –54} and Australia. ⁷ Despite this high incidence, older adults may be less likely to seek medical attention for TBI ⁵⁵ and are also less likely to be accurately diagnosed even when medical attention is sought. ⁵⁶ These findings suggest that TBI incidence among older adults likely exceeds published reports.

Mechanisms and demographics

The majority of TBIs sustained by older adults are attributed to low-level or same-level falls from standing height or less, ^7,57 even among those requiring surgical treatment of traumatic intracranial hemorrhage (Fig. 2). ⁵⁸ Mechanisms of TBI are biologically important because fall-related TBIs more commonly result in mass lesions, such as subdural hemorrhage, while the motor vehicle accident–related TBIs experienced by teens and younger adults more commonly result in diffuse axonal injury. ⁵⁹ Physically active older adults may be at elevated risk for specific sport-related mechanisms. For example, adults ages 55–64 years have the highest incidence of skiing-related TBI. ⁶⁰ Bicycle-related injuries, including TBI, in older adults most commonly occur while mounting or dismounting the bicycle. ⁶¹ Compared to younger individuals with TBI, older adults are less likely to engage in alcohol or drug abuse. ⁵ Finally, the majority of the oldest old patients with TBI are female and white (Fig. 3), ^9,41,51,56 mirroring U.S. nation-wide demographics of aging. ⁶² Although few population-based or multi-center studies have compared the distribution of TBI severity (e.g., mild, moderate, or severe) across age categories, at least one U.S. nationally representative study reported a similar distribution of TBI severity as defined by the Glasgow Coma Scale (GCS) and Abbreviated Injury Scale (AIS) across the age spectrum. ⁹

OPEN IN VIEWER

FIG. 2.

Major mechanisms of traumatic brain injury (TBI) by age (2007–2010). Falls are shown in black; motor vehicle accidents (MVA), in dark gray; and other mechanisms, in light gray. Mechanism of TBI among older adults is predominantly falls whereas mechanism among younger individuals is predominantly MVA. ⁹

OPEN IN VIEWER

FIG. 3.

Sex and race of adults with traumatic brain injury (TBI) by age (2007–2010). The demographics of TBI shift toward increasing prevalence of white race (dark gray bars) and female sex (black bars) with increasing age such that by age 85 years+ the majority of patients with TBI are female. ⁹

Healthcare cost and utilization

Compared to younger patients with TBI, those ages ≥65 years who seek emergent care are nearly 3 times more likely to receive a head computed tomography (CT) or MRI in the ED setting and 4 times more likely to be admitted to the intensive care unit (ICU), step-down, or surgical unit. ⁴¹ In the inpatient rehabilitation setting, an improvement of one point on either the Functional Independence Measure (FIM) or the Disability Rating Scale was associated with double the healthcare expenditures in older versus younger adults. ¹⁸ In a large, nation-wide sample of older adults treated in hospitals for TBI, ⁶³ the average annual treatment cost per person ranged from $73,000 to $78,000. ⁶⁴ In this study, older-old patients had significantly lower rates of outpatient injury-related clinic visits and significantly higher rates of rehospitalizations, home healthcare visits, and weekly hours of unpaid care from friends and family compared to younger-old patients, suggesting possible age-related disparities in coordinated care after hospital discharge. ⁶⁴

Clinical assessment

The GCS, although the most widely used clinical assessment to determine TBI severity at the time of initial presentation, may lack the nuance required to accurately assign TBI severity in older adults. Older adults with pre-existing dementia may have an abnormal GCS at baseline, ⁹⁴ others may have comorbid medical conditions or medication side effects that may complicate accurate diagnosis, ⁹⁵ and in others the burden/evolution of TBI may not be adequately captured by the initial GCS. ^96,97 As an example, age-related atrophy may provide space for an intracranial hemorrhage to expand substantially before leading to clinically apparent signs or symptoms that would be detected by the GCS. Thus, there is an urgent need for objective biomarkers to aid in the diagnosis, management decisions, and recovery monitoring of older adults with TBI, particularly those with pre-existing medical or neurological conditions.

Neuroimaging

Head CT is an important diagnostic tool in the acute evaluation, management, and outcome prediction for patients across the age spectrum. CT evidence of neurotrauma both increases with age and is associated with worse outcomes. ^36,98,99 The types of neurotrauma observed on head CT differ by age: Prevalence of extradural hematoma declines with age whereas prevalence of midline shift and subdural hematoma increases. ¹⁰⁰ Among adults ages ≥65 years admitted to hospitals with any severity of TBI, up to 45% may have subdural hematoma apparent on head CT. ⁴⁷ Among adults ages ≥65 years presenting to the ED with mild TBI (GCS 13–15), 11–21% may have evidence of intracranial trauma (compared with 5% in younger adults). ^{101 –104} Even those with normal GCS (GCS = 15) are at high risk: 17% of adults >60 years of age with a normal GCS had a positive head CT in one large Canadian study, ¹⁰² and 57% of adults >60 years of age who were found to have an intracranial hemorrhage on head CT had presented with a normal GCS in a large Swedish study. ¹⁰¹ This dramatically higher prevalence of CT evidence of neurotrauma among older versus younger patients is hypothesized to result from several factors: age-related changes in vasculature and white matter rendering vessels more vulnerable to rupture and white matter tracts more susceptible to shear injury, weakened musculature in the neck and trunk; such that even ground-level falls are not well braced by the body, pre-existing conditions, and medications such as antithrombotics. ^10,11,13

Structural MRI has been shown to improve prognostic modeling post-TBI by identifying evidence of neurotrauma that may be missed on head CT. ¹⁰⁵ Only a few studies have investigated the impact of age on MRI findings in TBI. One prospective study of 98 patients across the spectrum of age and TBI severity that performed MRIs, on average, 2.3 years post-TBI identified an association between older age and both larger lesion volumes and smaller gray matter volumes. ¹⁰⁶ Emerging neuroimaging technologies for TBI include advanced structural MRI sequences such as diffusion tensor imaging or 7-Tesla MRI, functional MRI, ¹⁰⁷ and a number of PET ligands that bind (with varying specificity) to amyloid-beta, tau, and markers of inflammation. ^108,109 Using these emerging technologies in older adults deserves further study. Evidence of tau or amyloid deposition on PET in older adults with TBI, however, will need to be interpreted carefully given evidence from multi-site nonselected autopsy studies that have demonstrated that high-grade amyloid and tau neuropathology increase approximately 10–15% per decade beginning in the sixth decade of life, rising from 1% to 2% among those ages 61–65 years to nearly 40% among those ages 91–95 years. ¹¹⁰

Emerging proteomic and blood biomarkers

Recent advances in biochemical assays of serum and cerebrospinal fluid (CSF) have identified a number of promising candidate blood-based and proteomic biomarkers of apoptosis and neuronal injury (neuron-specific enolase, tau, and amyloid-beta 1–42), glial injury (glial fibrillary acidic protein [GFAP], S100B, and excitatory amino acids), demyelination (myelin basic protein), proinflammatory cytokines (interleukins-1, −6, and −8), and gene expression of microRNAs (miR-16, miR-26a). ¹¹¹ Very few studies have evaluated the impact of aging on the kinetics or clinical performance of serum or CSF-based biomarkers of TBI. ^{112 –117} For example, although serum S100b showed initial promise in identifying low-risk patients who do not require head CT, ¹¹⁸ this biomarker may have reduced specificity among adults ages ≥65 years. ¹¹³ Additional emerging evidence suggests that there may be differences in CSF levels of excitatory amino acids and cytokines in older versus younger patients with TBI that may be independent of TBI severity. ¹¹⁴ Of note, circulating cytokine levels are affected by a range of comorbidities that are common among older adults, including hypertension, diabetes, dementia, Parkinson's disease, and osteoporosis. ^119,120 Studies of TBI biomarkers in rodent models have identified compelling differences between aged and younger animals on TBI-associated proteomic signatures, ¹¹⁵ peri-TBI levels and temporal sequences of mRNA expression of microglial (CD11b and ionized calcium binding adaptor molecule 1), and astrocytic (GFAP and S100b) activation markers, ¹¹⁶ and peripheral monocyte recruitment post-TBI. ¹¹⁷

Guidelines

Older adults with TBI are more likely to experience an interfacility transfer to a level 1 or 2 trauma center compared to younger adults, suggesting higher rates of either inappropriate initial triage or delayed deterioration necessitating transfer. ¹²¹ The GCS is widely used to assign TBI severity, but is poorly predictive of morbidity and mortality in older adults who frequently have better initial GCS scores than younger individuals with the same injury severity. ¹²² For this reason, researchers in Ohio have been working to develop and validate geriatric trauma-field triage criteria to optimally identify older adults with TBI who require emergent transfer to a trauma center, for example, a GCS cutoff of ≤14 (vs. ≤13 for younger individuals). ^{123 –125}

Most, but not all, ¹²⁶ existing rules and validation studies support the routine use of head CT for all patients >60 ¹⁰² or >65 years of age ^{103,104,127,128} presenting with mild TBI even after rapid return to baseline. The American College of Emergency Physicians recommends considering head CT in all patients ages ≥65 years who present with TBI, even mild injury without loss of consciousness (LOC), and recommends obtaining a head CT in all patients >60 years of age with TBI and LOC. ¹²⁹ The Canadian CT Head Rule identifies ages ≥65 years as a high-risk factor for intracranial trauma needing neurosurgical intervention among patients presenting with TBI and a GCS of 13–15, regardless of LOC. ¹⁰³ A large Austrian study later confirmed age >65 years as a high-risk factor for CT evidence of neurotrauma, but also identified novel high-risk comorbidities, including history of dementia or ischemic stroke. ¹²⁸ Further development and validation of geriatric TBI neuroimaging guidelines are therefore critically important.

Efforts are underway to predict which geriatric TBI patients may be managed safely on a standard medical or surgical ward rather than an ICU. ^130,131 One Swedish trauma center developed a protocol for geriatric neurosurgical decision making based on pre-injury functional status and predicted 1-year survival. ²⁹ While this center reported favorable outcomes in 41% of patients >65 years of age who underwent craniotomy for acute subdural hemorrhage, further research is needed to determine whether this approach may be overly conservative. ²⁹ According to the Eastern Association for the Surgery of Trauma, clinicians are encouraged to limit further aggressive treatment for older adults with severe TBI who do not improve within 72 h of admission. However, at least one retrospective observational study found that although lack of improvement by 72 h was significantly associated with increased in-hospital mortality, it was not associated with functional status at discharge or 12-month survival among those who survived to discharge. ²⁷

There remain, however, few to no evidence-based national or international consensus guidelines to inform acute inpatient management or long-term outpatient follow-up of older adults with TBI. This is, in large part, attributed to the paucity of dedicated Class I prospective clinical trials of treatments for older adults with TBI. In the absence of evidence-based guidelines, some centers have proposed strict age cutoffs for offering aggressive treatment such as neurointensive care admission or neurosurgical intervention for older adults with severe TBI, ^25,26,132 whereas others have a policy to admit all older adults with any TBI to the neurointensive care unit for serial neurological assessments and head CTs—a practice that may prove overly conservative. ¹³³

Neurocritical care and neurosurgical management issues

Acute care and prognosis of older adults with moderate-to-severe TBI are particularly challenging as demonstrated by dramatic variability in practice and outcomes across centers (Supplementary Table; see online supplementary material at http://www.liebertpub.com/neu). A few recent observational, largely retrospective studies have assessed the value of various acute neurosurgical interventions including intracranial pressure (ICP) monitoring, ^134,135 craniotomy, ¹³⁶ and decompressive craniectomy ^132,136 in older adults with moderate-to-severe TBI. However, these studies may be limited by confounding by indication and other potential sources of bias such as the physician's opinion of a patient's prognosis, which may impact treatment decisions.

Whether ICP monitoring improves outcomes in this population remains controversial, with some studies supporting its use ¹³⁴ and one study finding no evidence of meaningful benefit. ¹³⁵ The surprisingly low prevalence of ICP monitoring in the negative study, however, suggests that there may have been unmeasured confounding variables. A single-center study of continuous bedside neuromonitoring data found that although older patients have lower ICP and higher cerebral perfusion pressure compared to younger patients, factors typically associated with better outcomes, ¹³⁷ they also had worse vascular pressure reactivity and autoregulation compared to younger patients, factors that may contribute to worse outcomes. ¹³⁷ Together, these findings suggest that although ICP monitoring likely improves outcomes in some older adults, given the tendency for older adults to have lower ICPs on average than younger adults, it is possible that older adults may benefit from the development of geriatric-specific clinical criteria to determine whether ICP monitor placement is appropriate.

There is substantial between-study heterogeneity in outcomes after craniotomy or craniectomy in geriatric TBI. ^132,138 The method of surgical intervention may impact outcome as demonstrated by a retrospective cohort study that utilized propensity scores to mitigate confounding by indication; patients who were treated with decompressive craniectomy had worse 6-month Glasgow Outcome Scale (GOS) outcomes compared to those treated with craniotomy. ¹³⁶

The contribution of antithrombotic therapies to poor outcomes in older adults with TBI is contested. Some studies have identified increased mortality and worse outcomes for all classes of antithrombotic agents. ^139,140 Others report that anticoagulant agents such as warfarin (but not antiplatelet agents) are associated with increased mortality ¹⁴¹ (particularly if warfarin is at a therapeutic level ¹⁴² ) and need for neurosurgical intervention. ¹⁴³ Still others have found that early aggressive treatment can mitigate any negative impact of antithrombotics on mortality or outcomes. ^144,145 Thus, although antithrombotic therapies may not be associated with worse post-TBI outcomes in a setting in which patients receive rapid and aggressive treatment for hemorrhagic complications and in analyses that adjust for initial TBI severity, it is clear that anticoagulant therapy is associated with worse initial TBI severity, such as acute subdural hemorrhage after relatively minor trauma. ^{141,142,145,146} This conclusion is further supported by a study of 80 patients (all ages) with acute subdural hemorrhage reporting that initial hematoma volume and GCS are better predictors of hematoma expansion and outcome than age or antithrombotic therapy if patients are rapidly given appropriate agents to reverse coagulopathy. ¹⁴⁷ The decision to restart anticoagulants such as warfarin after a TBI in older adults at high risk of thrombotic events or ischemic stroke is complex. One large, retrospective study of over 10,000 Medicare beneficiaries hospitalized for TBI who had been taking warfarin during the month preceding admission reported that restarting warfarin after discharge was associated with a 51% increased risk of all hemorrhagic events and only a 23% reduction in all thrombotic events over the subsequent year. However, when stroke risk was assessed in isolation, there was a net benefit of a 17% reduction in risk of combined hemorrhagic or ischemic stroke. ¹⁴⁸

The controversial role of age

It is well established that, on average, older adults with TBI have higher mortality, ^{4,14 –16} slower rates of functional and cognitive recovery, ^{17 –20} and worse functional outcomes post-TBI compared to their younger counterparts. ^{5,17,21 –24} In one of the largest studies to date to investigate the association between age and 6-month functional outcome score on the GOS among 8719 patients with moderate-to-severe TBI contained in the IMPACT database, there was a strikingly linear relationship observed between age and outcome. ¹⁴⁹ Thus, some have argued that there is unlikely to be a specific age beyond which outcomes precipitously worsen. ¹⁵⁰ Other studies, however, have reported an “inflection point” in the fourth or fifth decade of life at which trauma mortality appears to increase steeply. ^151,152 Regardless of the presence or absence of an inflection point, it is clear that a substantial number of older adults with TBI may recover well (Table 1, Table 2, and Supplementary Table; see online supplementary material at http://www.liebertpub.com/neu), including some with severe TBI who receive aggressive neurosurgical management. ^{29,92,138,153–154}

The role of provider attitudes as well as patient and family preferences must also be considered when interpreting outcomes in older adults with TBI, who may be more likely to have care electively withdrawn. ¹⁵⁵ The issue of provider attitudes was highlighted in a large UK study of patients with TBI and cerebral contusions, reporting that increasing age was associated with longer delays in obtaining an initial head CT, lower likelihood of being transferred to a neurotrauma center, and lower likelihood of review by a senior (vs. junior) physician. ¹⁵⁶ Similar findings suggestive of agism in TBI management decisions were reported in one Scottish study, ¹⁵⁷ but a multi-center, prospective U.S. study did not identify any evidence of age-related neurosurgical intervention bias. ⁹¹

Outcome assessment

In the United States, the GOS and the Glasgow Outcome Scale-Extended (GOSE) are the most widely used and widely cited functional outcome measures in TBI clinical research ¹⁵⁸ and are included in the National Institute for Neurological Disorders and Stroke TBI Common Data Elements (CDEs). ¹⁵⁹ Neither the GOS nor the GOSE were developed or validated in older adults and may not adequately quantify TBI-related functional impairment in a geriatric population, particularly in those with pre-existing functional impairment. For example, a multi-center, prospective study of older adults with severe TBI found that although these patients experienced significant improvement in physical function over 1 year according to the Health Related Quality of Life Measure, this functional improvement was not detected by the GOSE. ¹⁶⁰ Additionally, many in the field rely upon the 1998 GOS and GOSE administration and scoring guide by Wilson and colleagues. ¹⁶¹ Yet, even in this comprehensive guide, scoring of patients with pre-injury disability is described as “problematic.” ¹⁶¹ Thus, a patient with severe baseline disability who fully recovers to their severely disabled baseline status may be scored as a GOSE 3 (severe disability) by one study and a GOSE 8 (good recovery) by another study. Additionally, the GOS and GOSE do not systematically distinguish between effects of brain versus body trauma or between cognitively versus physically mediated function. ¹⁶¹

We report nearly 40 studies that assessed functional outcome after all-severity or moderate-severe TBI in older adults (Table 2 and Supplementary Table; see online supplementary material at http://www.liebertpub.com/neu); functional outcome after mild TBI in older adults is relatively understudied (Table 1). One small U.S. study of 40 adults ≥65 years of age with mild TBI reported that by 6 months post-injury, 88% had achieved their pre-injury functional status based on the modified FIM. ¹⁷ To achieve a perfect score on the modified FIM, however, a patient need only communicate intelligibly and fluently, eat in a “customary manner,” and walk 150 feet. ¹⁷ Thus, this instrument may suffer from substantial ceiling effects in a mild TBI population.

There is increasing recognition that the use of the GOS, GOSE, or FIM as the primary endpoint for TBI clinical trials does not adequately capture the complex, multi-dimensional, and evolving nature of TBI, thus historically limiting the success of these trials. ¹⁶² To address these limitations, efforts are underway to develop improved TBI endpoints. ^163,164 As part of this effort, it will be critically important to either evaluate the performance of existing multi-modal measures in geriatric TBI populations (as was done in this Taiwanese study, ¹⁶⁵ these Canadian studies, ^166,167 this meta-anlaysis, ¹⁶⁸ and this review/opinion piece ¹⁶⁹ ) or develop novel multi-modal geriatric-specific TBI endpoints to optimize success of future TBI clinical trials in older adults.

Mortality

When considering mortality after geriatric TBI, it is important to distinguish between short-term mortality (during initial hospitalization or rehabilitation) and longer-term mortality (over months or years after TBI). Short-term mortality post-TBI is high among older adults, particularly those with severe TBI (Supplementary Table; see online supplementary material at http://www.liebertpub.com/neu), with several studies reporting in-hospital mortality rates as high as 70–80% in this population. ^98,99 There is, however, substantial variability across centers that may be attributed to a combination of variability in clinical practice and study design. Factors such as aggressive treatment, ^138,170 pre-morbid independent function, and good pre-injury health status ^{37,92,171,172} have been associated with lower short-term mortality among older adults with TBI and most in-hospital deaths in this population occur following decisions to withdraw care. ^155,173,174 Other factors associated with short-term mortality in older adults with TBI include older age and CT evidence of brainstem or diencephalic injury, whereas injury mechanism and GCS may be less important predictors. ⁹⁸ Among older adults who survive the initial hospitalization and rehabilitation period post-TBI; however, several studies have reported that the observed higher mortality among older versus younger individuals may predominantly be accounted for by expected age-related mortality as is observed in the general aging population. ^175,176

Post-traumatic neurological disorders: Epilepsy, stroke, and neurodegenerative disease

Compared to younger individuals, older adults are at increased risk for post-traumatic epilepsy ^177,178 and are more likely to present with delayed rather than early seizures post-TBI. ¹⁷⁹ Pre-existing conditions such as Alzheimer's dementia (AD) further increase the risk of epilepsy in older adults. ¹⁸⁰ The choice of drug for both short-term post-traumatic seizure prophylaxis and long-term post-traumatic epilepsy treatment deserves further study in older adults. First-generation antiepileptic agents such as phenytoin may be suboptimal in older adults because of nonlinear pharmacokinetics, propensity for drug-drug interactions, and cognitive side effects, whereas certain newer agents such as lamotrigine or levetiracetam may be preferable. ^181,182

TBI may be an independent risk factor for both ischemic and hemorrhagic stroke. ^183,184 Emerging evidence suggests that initiation of treatment with a selective serotonin reuptake inhibitor (SSRI) among older adults hospitalized for TBI was associated with increased risk of hemorrhagic, but not ischemic, stroke. ¹⁸⁵ This finding was further supported by a recent population-based study that identified an association between SSRI initiation and spontaneous hemorrhagic stroke among patients taking oral anticoagulants. ¹⁸⁶ Further research is needed to determine the mechanism of post-TBI stroke and SSRI-associated hemorrhage in order to inform safe management of post-TBI depression (which is common and undertreated in older adults ¹⁸⁷ ) as well as optimal post-TBI stroke prevention strategies.

Although TBI is now a well-established risk factor for dementia and Parkinson's disease, ^{188 –191} few past studies have assessed risk of dementia or PD specifically after geriatric TBI (e.g., TBI sustained in the fifth decade or beyond). ^{34,35,192 –194} One small, prospective study of falls and risk of dementia among individuals ages ≥70 years concluded that fall-related TBI was associated with earlier onset of dementia and that presence of an apolipoprotein E epsilon 4 allele acted synergistically with fall-related TBI to further increase risk of earlier onset of dementia. ¹⁹² An important methodological concern in epidemiological studies assessing risk of a neurodegenerative disease after geriatric TBI is the possibility of reverse-causation—that the patient fell and sustained a TBI because of early symptoms of a neurodegenerative disease rather than the reverse. This concern was highlighted by a population-based Danish study reporting that the association between TBI and PD was almost entirely attributed to fall-related TBIs sustained within the 3 months preceding initial PD diagnosis. ¹⁹⁴ To mitigate this potential for reverse causation, two large studies using California-wide data from the Healthcare Cost and Utilization Project compared risk of dementia and PD in adults ages ≥55 years at least 1 year after sustaining a TBI versus fracture. In these studies, incident geriatric TBI was associated with a 44% increased risk of PD and a 26% increased risk for dementia over the subsequent 5–7 years. Even mild TBI was associated with a 24% increased risk for PD for those ages ≥55 years and a 21–25% increased risk for dementia among those ages ≥65 years (but not for those ages 55–64). ^34,35

Repeated TBI in older adults may be associated with greater risk for neurodegenerative outcomes than single TBI. ³⁴ Repeated concussive and subconcussive injuries have been associated with chronic traumatic encephalopathy (CTE), a unique degenerative tauopathy, primarily described among contact-sport athletes, blast-exposed military personnel, and victims of domestic violence. ^{195 –197} Whether CTE may result from repeated fall-related TBIs or repeated fall-related subconcussive injuries in older adults is unknown. However, a case series of 139 cases of autopsy-proven multiple system atrophy (MSA)—a rare degenerative movement disorder that causes dysautonomia, parkinsonism, and cerebellar ataxia—identified CTE pathology in 8 cases (6%). Of these 8 cases, only 4 had a history of contact sports generating the hypothesis that repeated falls attributed to MSA may have precipitated the CTE pathology. ¹⁹⁸ An analysis of over 1700 brains from the Mayo Clinic brain bank, however, did not identify any cases of CTE in patients exposed to a single fall-related TBI. ¹⁹⁹

Other chronic psychosocial and cognitive impairments

Several studies have assessed psychosocial sequelae after geriatric TBI. ^{24,200 –203} Based on a systematic review of post-TBI depression, prevalence of depression is 1.8–8.9% in older community-dwelling adults, 25% in skilled nursing facilities, and 21–37% among older adults with TBI. ²⁰¹ TBI in older adults has been associated with 11% increased risk of new-onset depression and 50% increased risk of new-onset PTSD. ²⁰⁴ Risk of depression among older adults hospitalized for TBI may be highest immediately after discharge and then decline over the subsequent 12 months. ²⁰² Whereas this reduction in incident depression with increasing time post-injury may be explained by declining rates of post-TBI longitudinal follow-up and resultant ascertainment bias, at least one study has found that depressive symptoms decline over time post-injury only among older, and not younger, adults. ²⁰³ Similarly, a small study of 26 older adults with mild TBI found that, compared to younger patients, older adults reported less psychosocial impairment, psychological distress, and physical symptoms 1 month after their TBI, but these findings were largely mediated by employment status. ²⁴ This finding suggests either that older adults are less prone to some of the psychosocial sequelae of mild TBI or that the outcome assessments capturing psychosocial sequelae of TBI are less sensitive among retirees. Together, these findings suggest that careful screening for mood symptoms among older adults with TBI is important at any time post-injury, but particularly within the first few months to years post-injury when risk and symptom burden may be highest.

Cognitive symptoms and impairment after TBI in older adults are common. ²⁰⁵ Just as in younger patients, prevalence and severity of cognitive sequelae in older adults tends to increase with increasing TBI severity. ²⁰⁶ Older adults with TBI, however, experience slower recovery of cognitive function during rehabilitation ¹⁸ and over the year after TBI compared to younger patients. ¹⁹ Most studies of cross-sectional cognitive outcomes after TBI have reported that older patients have worse cognitive outcomes compared to younger patients. ²⁰⁷ The few studies that have adjusted for expected age-related decline in cognitive function, however, have reported that older adults may have equivalent or even better cognitive outcomes compared to younger individuals with TBI. ²⁰⁷ This compelling finding highlights the importance, as in the evaluation of post-TBI mortality described above, of accounting for expected age-related changes that are unrelated to the injury when comparing outcomes across age categories. Other important modifiers of cognitive outcome in older adults with TBI include pre-injury factors, such as pre-existing comorbidities that may be independently associated with worse cognitive function, as well as the deleterious cognitive impact of trauma and hospitalization more generally, versus the specific impact of TBI. For example, one small study of mild TBI in older adults reported that although 3-month cognitive outcomes were worse among those with mild TBI compared to community-based controls, outcomes were equivalent to orthopedic controls, ²⁰⁸ suggesting that cognitive impairment after mild TBI in older adults may be partially attributed to overall trauma or predisposition to injury. ²⁰⁸

Prognostic models

As the global incidence of geriatric TBI continues to rise, the need for geriatric TBI prognostic models has become increasingly urgent. Recently two prognostic models have incorporated age as well as clinical and CT indicators of TBI severity and have subsequently each been studied in large, population-based samples of older adults with TBI. The CRASH-CT prognostic model is a predictor of 14-day mortality and 6-month unfavorable outcome on the GOS and includes only variables of age, GCS, pupil reactivity, extracranial injury, and evidence of trauma on head CT. ²⁰⁹ The model does not take into account baseline comorbidities or pre-injury function. The model showed adequate discrimination and calibration in older adults with all-severity TBI (ages 65–84 years) in a U.S. study, ²¹⁰ but performed very poorly with dramatic overprediction of mortality (observed 50%; predicted 81%) and unfavorable outcome (observed 72%; predicted 95%) in older adults with severe TBI (ages 65–92 years) in a Norwegian study. ²¹¹ The IMPACT prognostic model predicts 6-month mortality and unfavorable outcome on the GOS. ^212,213 The IMPACT model includes age, GCS, pupillary reactivity, hypoxia, hypotension, CT evidence of trauma, and blood glucose and hemoglobin levels, but, again, does not include baseline comorbidities (other than those captured by baseline glucose and hemoglobin) or pre-injury function. It showed adequate discrimination, but poor calibration, among older adults (ages 65–84 years) with evidence of substantial underprediction of mortality. ²¹⁴ The poor performance of these models in older adults with TBI is not surprising given that the development cohort for both the CRASH-CT and IMPACT models consisted of pooled trials and observational cohort studies that largely excluded older adults with pre-existing comorbidities. Further studies are now needed to determine whether incorporating comorbidities, polypharmacy, baseline function, or measures of frailty into these models will improve their prognostic value, as has already been demonstrated for geriatric trauma outcome prediction models. ^86,87