Preinjury mild cognitive impairment (MCI) and dementia. Preinjury cognitive changes may be present in the older patient with TBI who is referred for a neuropsychological evaluation, and it is therefore important to enquire about this possibility. An analysis of CDC TBI Surveillance System data [1] found that falls were the leading cause of TBI in persons 65 years and older. Persons ≥85 years had a fall rate twice that of persons 65–84 years, and six times the rate of those 65–74 years. Risk factors for falls include MCI and dementia, with some studies demonstrating an association with specific cognitive abilities involving processing speed and executive functioning [4, 5]. Holtzer et al. [4] found that lower scores on a speed/executive function factor that comprise Trail Making [6], Digit Symbol [7], and Block Design [7] were associated with a greater risk for single or recurrent falls. Individuals in the low scoring group were almost four times more likely to fall than individuals in the high scoring group (OR = 3.9, 95 % CI = 1.5–10.1, p = 0.006). In contrast, a factor examining episodic [8] and semantic [9] memory was not related to increased risk. More recently, Nagamatsu et al. [5] found that persons ≥65 years who were at-risk for falls performed more poorly than a nonrisk group on a virtual reality administered task requiring divided attention of crossing the street while talking on the phone. The at-risk group was significantly slower and had more “collisions” with oncoming traffic.

Information concerning preinjury status should ideally be gathered from a significant other in order to reduce the possibility of unreliable estimates by the patient as a result of anosognosia or memory deficit. A number of informant questionnaires are available to identify MCI and dementia [6], and the wording can be modified to clarify that the respondent is being asked about functioning prior to the injury. In our research on older adults with TBI [10–13], we used the Blessed Dementia Scale [14] to enquire about cognitive and personality changes. The Blessed Dementia Scale includes items evaluating changes in the performance of everyday activities (e.g., managing money, recalling recent events), habits (eating, dressing, sphincter control), and personality, interests, and drives (e.g., quality of social interactions, maintenance of hobbies, initiative). Another instrument, the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE), asks about changes over the last 10 years in learning and memory, orientation, financial awareness, and executive skills [15]. The IQCODE is a reliable and validated instrument for the detection of MCI and dementia, with translations available in multiple languages besides English, and evidence for cross-cultural sensitivity to impairment [16, 17]. The Concord Informant Dementia Scale (CIDS) [18] evaluates changes in everyday cognitive functioning over the previous 5 years. Many of the items were taken or adapted from the IQCODE. Domains assess memory, orientation, judgment and problem solving, language, involvement in community affairs and home and hobbies, and personal care. An additional rating form, the Functional Assessment Questionnaire (FAQ), evaluates independence in performing instrumental activities of daily living such as handling complicated financial matters and managing medications [19]. The FAQ was described by the Agency for Health Care Policy and Research as an effective means of identifying demented individuals, with sensitivities and specificities in the 85–90 % range [20].

Medical comorbidities. It is important to obtain a health history in order to gauge the independent effects of certain medical conditions on neurobehavioral outcome after TBI. In their surveillance study of over 17,000 hospitalized TBI patients ≥65 years old, Coronado et al. [1] noted that approximately 80 % of the patients had comorbid medical conditions, ranging from 1 (23 %) to ≥5 (6 %). The most frequent conditions included hypertension (39 %), cardiac arrhythmias (18 %), fluid and electrolyte disorders (17 %), and diabetes mellitus (15 %) [1]. Falls are associated with numerous risk factors including disorders of gait, balance, weakness, decreased vision, and peripheral neuropathy [21, 22]. Motor vehicle crashes, the second most common mechanism of TBI in the elderly, are associated with medical conditions including stroke, heart disease, and arthritis as well as certain medications such as benzodiazepines, nonsteroidal anti-inflammatory drugs, and angiotensin-converting enzyme (ACE) inhibitors [23]. Helms et al. [24] analyzed the relationship between chronic health conditions and neuropsychological test performance in 585 elderly participants in the Canadian Study of Health and Aging. The Cumulative Illness Rating Scale (CIRS) [25] was used to determine the presence of 14 chronic illnesses. An increase in the number of conditions predicted poorer performance on measures examining visuospatial and constructional abilities, verbal memory, timed visuomotor sequencing, and set shifting. As the investigators noted, the use of a total CIRS score may have masked additional relationships that existed between specific medical conditions and cognitive performance.

Certain disorders associated with subclinical vascular ischemic disease such as hypertension, diabetes, cardiac disease, and sleep apnea may influence the TBI patient’s clinical phenotype. In normal aging, these risk factors are associated with a neurobehavioral syndrome characterized by poorer attention, executive functioning, and information processing speed, as well as depression and personality changes including emotional lability [26–29]. Epidemiological studies indicate that these comorbidities are common in older adults. In the United States, the prevalence of hypertension is 67 % in persons ≥60 years old [30]. Patients should be asked whether they have certain vascular comorbidities. If they respond affirmatively, they should then be asked whether these conditions are well-controlled. The importance of adequate blood pressure control on cognitive performance was demonstrated by Waldstein et al. [31] in a cross-sectional study of community residing older adults. Irrespective of a prior diagnosis of hypertension, persons with elevated BP (systolic ≥140 mmHg or diastolic ≥90 mmHg), versus those with normotensive values at the time of neuropsychological testing, performed more poorly on measures of visual memory, motor speed, and visuomotor integration. Persons with both a prior diagnosis as well as elevated blood pressure levels were most vulnerable to poor performance.

Medical records, if available, may be able to verify the presence of vascular comorbidities because a reliance on self-reports can be inaccurate. The unreliability of self-report data is highlighted by findings from the National Health and Nutrition Examination Surveys which indicate that between 2005 and 2006, 7 % of the US adult population had elevated systolic blood pressure readings ≥140 mmHg or diastolic blood pressure readings ≥90 mgHg, yet these individuals had not been told by any health care provider that they had high blood pressure. Overall, only 78 % of hypertensive adults were aware they had this condition [30]. It is important to gauge as well whether vascular comorbidities have been adequately controlled. This can be determined by a review of available medical records to examine repeated measures such as blood pressure values. In persons with diabetes, the glycated hemoglobin (A1C) value can provide an index of the extent to which blood glucose levels have been controlled over the past several months [32].

Vitamin deficiencies are also important to assess since they may adversely affect the cognitive status of older persons. The prevalence of Vitamin B₁₂ deficiency is 10–15 % in persons >60 years [33]. Hin et al. [34] observed a two to three times risk of cognitive impairment, determined via the Mini-Mental State Examination [35], in persons ≥75 years who had the lowest levels of serum B₁₂. In a more recent study, Tangney et al. [36] found that specific metabolites indicative of B₁₂ deficiency were associated with poorer performance on measures of episodic and semantic memory and perceptual speed in persons ≥65 years. Low levels of vitamin D can be another source of cognitive impairment in the elderly, notably on measures of executive functioning and processing speed [37, 38]. Low vitamin D levels have also been identified as a risk factor for longitudinal cognitive decline. Llewellyn et al. [39] found that persons >65 years with extremely low levels of vitamin D exhibited a faster decline, relative to those with adequate levels, in their overall cognitive status and set shifting speed over a 6-year follow-up period.

Medications. Medication use and potential side-effects need to be evaluated in older adults because certain classes can negatively impact neurobehavioral outcome from TBI. Polypharmacy, i.e., multiple medication use, is highly prevalent in the elderly. Hajjar et al. [40] found that surveys of community-residing elderly reported that on average, two to nine prescription medications were taken daily by older adults. In a population-based study, Kaufman et al. [41] found that during a 1-week period, 57 % of persons 65 years and older took ≥5 drugs, whereas 12 % took ≥10 drugs. Medications for pain, colds and coughs, and nutrition were among the most frequently non-prescribed medicines.

Medications with anticholinergic properties are used to treat several common age-related conditions including hypertension, cardiovascular and pulmonary diseases, Parkinson’s disease, and incontinence. These medications can cause MCI, dementia, and delirium [42]. Moreover, some over the counter drugs such as antihistamines can have anticholinergic as well as sedating effects. In one study, Bottiggi et al. [43] examined the impact of chronic anticholinergic medication use in cognitively normal adults ≥60 years who received annual neuropsychological assessments. The investigators found that compared to a group not treated with these medications, persons taking anticholinergic medications exhibited a significantly greater longitudinal slowing on Trails A and Trails B. Another investigation [44] found longitudinal declines as a function of continuous use of anticholinergic medications over a 6-year follow-up. Compared to women who had never taken anticholinergic drugs, women in the continuous use group were at higher risk of cognitive decline on measures of timed verbal fluency (OR, 1.5; 95 % CI, 1.1–2.0; P = 0.02) and overall mental status (OR, 1.4; 95 % CI, 1.1–1.9; P = 0.02). Males with continuous anticholinergic use versus males without such use were at a higher risk for a decline in visual memory (OR, 1.9; 95 % CI, 1.2–3.3; P = 0.01) and timed set shifting (OR, 1.9; 95 % CI, 1.1–3.6; P = 0.03).

Rating scales exist to characterize drugs according to their anticholinergic properties. One of these, the Anticholinergic Cognitive Burden (ACB) Scale [45], classifies the effects of centrally acting medications according to the severity of their cognitive impact, ranging from mild, moderate, or severe. A copy of this measure, as well as the associated medication list, can be obtained from the following website: http://​www.​uea.​ac.​uk/​mac/​comm/​media/​press/​2011/​June/​Anticholinergics​+study+drug+list​.

Another similar scale, the Anticholinergic Drug Scale [46] uses a rating system ranging from 0 (no known anticholinergic properties) to 3 (markedly anticholinergic). The ADS has been validated against serum anticholinergic activity, with scores accounting for 9.5 % of the variance in serum activity.

Are the extent and persistence of neurobehavioral deficits reasonable given the severity of the TBI? The neuropsychologist is commonly faced with the diagnostic dilemma of determining whether the performance of the older adult with TBI is reasonable given the severity of their injury and the time elapsed since their injury. This is an important determination from the standpoint of making effective recommendations and helping the patient and family with future planning. However, there have been relatively few studies examining the neurobehavioral recovery of older persons to help with diagnostic issues and to estimate the time course of recovery, as much of the available literature has focused on global indices of mortality and functional outcome. At present, and consistent with what is known about outcome in young adults, the handful of studies [10, 11, 47–49] suggest that uncomplicated mild TBI (i.e., no intracranial findings) is associated with subtle deficits in the acute outcome stage, faster cognitive recovery than moderate injuries, and a return to normal cognitive functioning by 1 year. In contrast, patients with complicated mild TBI have outcomes more similar to those with moderate than uncomplicated mild TBI. The important caveat for all these studies is that they are limited to relatively healthy individuals screened for preinjury medical comorbidities, psychiatric conditions, and dementia, and therefore the results cannot be generalized to all patients who are seen in clinical practice.

In an early study, we [11] prospectively recruited patients who were ≥50 years old with uncomplicated mild head injuries (Glasgow Coma Scale (GCS) [50] scores of 13–15, loss of consciousness <20 min, and normal neurologic and neuroradiologic findings). Patients had sustained head injuries of sufficient intensity to produce a TBI (e.g., striking the head during a fall) and evidence of retrograde and/or posttraumatic amnesia (PTA). A second group of patients with moderate TBI (GCS scores 9–12, or 13–15 with intracranial complications) who had similar demographic features and injury-test intervals as the mild patients were included. Both groups received measures of attention, language, memory, and executive functioning at an average of 1 month postinjury. We found no significant differences in performance between the patients with mild TBI versus community residing controls on any cognitive measure, except for worse performance of the patients on a timed phonemic fluency task. In contrast, the mild patients and controls both performed significantly better than the moderate patients on measures of visuomotor processing speed and set shifting, verbal memory, confrontation naming, reasoning and hypothesis generation.

We subsequently examined possible differences in outcome between patients with and without intracranial complications but comparable GCS scores of 13–15 [10]. Patients were classified as having either uncomplicated mild TBI (GCS scores of 13–15, normal neuroradiologic findings), complicated mild TBI (GCS scores of 13–15, abnormal neuroradiologic findings), or moderate TBI (GCS scores of 9–12 with or without abnormal neuroradiologic findings). The uncomplicated patients with mild TBI performed significantly better on language (naming, fluency) and executive functioning (number of categories) measures than patients with complicated mild TBI. This latter group, in turn, performed similarly to patients with moderate TBI with the exception of faster set shifting ability.

Rapoport et al. [47] prospectively recruited patients ≥50 years old who sustained mild TBI (GCS score 13–15, PTA <24 h, loss of consciousness/confusion ≤20 min), and moderate TBI (GCS 9–12, PTA <1 week, or GCS 13–15 with intracranial complication). After controlling for demographic factors and severity of medical comorbidities, it was found that the patients with TBI performed significantly worse than community controls at the 1 year assessment on measures of overall cognitive status, verbal memory, timed letter fluency, visuomotor speed, and expressive language (naming). Post hoc comparisons demonstrated that these cognitive differences occurred between patients with moderate TBI and the controls, but not between mild TBI patients and the controls. In a follow-up study, Rapoport et al. [48] investigated the 2-year cognitive outcome of the same cohort of older TBI adults, and the investigators did not find any significant differences in cognitive performance between the two groups.

Extrapolating from these findings, the neuropsychologist evaluating a relatively healthy older person who has been screened for preinjury cognitive impairment should expect mild TBI patients to exhibit a good outcome in the first year, unless their injury was complicated by intracranial pathology on acute imaging or necessitated urgent surgery. Those with moderate TBI should be performing within normal limits in most domains by about 2 years postinjury. The cognitive outcome and recovery of patients with severe TBI (GCS scores of 3–8) has not been explored, but would be predicted to be less favorable based on what is known about their poor functional outcome. In two prospective series [51, 52], only 8–15 % of patients ≥56 years old with severe TBI, defined by GCS scores ≤8, exhibited a Good Recovery (resumption of normal activities with possible minor deficits) or Moderate Disability (disabled but independent in daily activities) at 6 months postinjury.

Is the cognitive pattern consistent with TBI or a neurodegenerative disorder such as Alzheimer’s disease? The differential diagnosis of cognitive impairments due to TBI versus Alzheimer’s disease (AD) is clinically challenging. Questions posed to informants about whether the patient exhibited preinjury cognitive changes may be difficult to accurately recollect or may be biased by the injury itself. Intracranial bleeds from TBI in some older adults can slowly evolve, and therefore informants may describe an insidious onset and progressive course of cognitive changes more characteristic of AD. In other cases, the patient’s pre-existing subtle cognitive deficits may be unmasked by the injury, and informants will attribute deficits to the TBI.

In a study to identify neuropsychological features that distinguish AD versus TBI, we [13] compared the cognitive profiles of older adults who sustained mild and moderate TBI or were diagnosed with probable AD. The groups were similar in demographics and overall cognitive status on the MMSE. Patients with TBI were screened for preexisting dementia and were recruited during their initial hospitalization or shortly after discharge while in an early stage of recovery. Both groups received the shortened form of the California Verbal Learning Test [53] requiring them to recall nine words over four trials, and then retain these words over time. Both patient groups demonstrated impaired recall of a word list relative to normal controls. However, those with AD also displayed poorer recall than patients with TBI. Whereas the patients with TBI and normal controls exhibited a nearly equal distribution of recall from the primacy, middle, and recency positions, the patients with AD recalled a significantly higher proportion of words from the end of the list. This latter finding could reflect more rapid forgetting of the earlier items in AD. Performance on letter and category fluency tasks also differentiated the patient groups. Patients with AD, in contrast to those with TBI or normal controls, did not show a normal facilitation in generating words belonging to categories compared with words beginning with specific letters.

Breed et al. [54] compared the cognitive profiles of patients with TBI and AD. Their patients with TBI were older than 55 years at the time of the study, they had been injured an average of 15.8 years previously, and determination of severity of injury was based on self-report information concerning length of loss of consciousness and PTA. The investigators replicated our findings of poorer timed letter fluency and memory functioning in the AD group. The patients with AD exhibited significantly lower percent retention scores for both verbal and visual material relative to TBI and normal controls, whereas the latter groups did not significantly differ from each other. This suggests that rapid forgetting is more characteristic of AD versus TBI.

What if the patient exhibits a delayed or a progressive deterioration in their neurobehavioral status?