Diagnosis and treatment of dementia
Who should be assessed?
All healthcare professionals should be attuned to the possibility of cognitive impairment in the elderly. In particular, complaints of memory or other cognitive problems should not be attributed to ‘ageing’ without further evaluation. Often, the problem is first recognised by family, friends or caregivers. It is unusual for family members to bring complaints to medical attention when there is no basis for the complaint (II)(1)
The following patients should therefore receive further evaluation:
- a patient who arouses the healthcare professional’s suspicion of cognitive impairment during interview (I)
- any patient who complains of a worsening memory or other cognitive symptoms (unless restricted to naming people) (I)
- a patient where the family or healthcare professional notices difficulties in any of
- remembering appointments
- remembering things which happened recently
- recalling conversations
- finding the right words
- taking medication accurately
- dealing with finances
- decline in personal grooming
- changes in personality
- social withdrawal, loss of interest or changes in mood (I)
- A patient over 65 years who may have been responsible for a car accident (III) (2)
- An elderly patient who needs to make an important financial decision (such as the sale of a house, nominating others to manage financial affairs, or making a will) and in whom mental competency is questioned (III)(3;4)
As well as advancing age, a history of any of the following increases the risk of subsequent dementia and should further raise healthcare professionals’ index of suspicion (I):
- Parkinson’s disease
- head injury
- diabetes mellitus
- other cardiovascular risk factors (high blood pressure, smoking, high cholesterol, renal failure)
- family history of dementia
Approach to evaluation
There are six elements to diagnostic evaluation: taking a history from the patient, interviewing a caregiver or family member, physical examination, brief cognitive tests, and laboratory tests (I). The majority of patients also require structural imaging (II).
Much of the diagnostic process can be undertaken routinely in primary care. However, whether each of these six elements occurs in the patient’s home, in primary care or in specialist services, and whether it is conducted by a doctor or other healthcare professional, is less important than that each element happens (I).
The diagnostic process rarely needs to be urgent. Indeed, it is often helpful if it takes place over several visits (I).
It is usually essential to speak with a family member of caregiver separately from the patient (I). This reduces the risk that information about functional deficits or ‘embarrassing behaviour’ will be withheld for fear of upsetting the patient. It is also an opportunity to assess the support that is available for the patient, and to start to engage that support.
The information should then be drawn together and discussed with the patient and his or her family at a meeting to discuss the results, diagnosis and implications (I).
Primary or secondary care?
As with all chronic diseases, a partnership between patient, carer, GP and multidisciplinary secondary care specialists is needed to achieve high quality evaluation, treatment and support. The long term, progressive nature of dementia means that continuity of care by professionals whom the patient and family grow to know and trust is an important element in determining the best quality of care (I). This is because education and support of the patient and their family has a significant impact on the need for subsequent emergency or institutional care.
The timing of referral will depend on the severity of the problem, the level of diagnostic uncertainty, and patient and family preference. However, most patients with dementia will require the attention of specialist services at some point in their illness (II). Referral should not be delayed because of anxiety about the consequences (II). For example, patients who fear that referral will precipitate institutionalisation often have carers for whom specialist support will have greatest impact. Thus, it is better that patients receive an early specialist evaluation with discharge back to primary care until problems supervene, than that referral only occurs in crisis (II).
In some countries, only specialists can recommend starting drug treatment for dementia. In such countries, referral should not be confined to patients who are thought to be suitable for drug treatment (II).
Specialists with an interest in dementia are variously located in neurology, geriatric and psychiatric services. Specialist assessment services should ideally include not only office-based memory clinics but also the possibility that patients can be assessed in their own home (I). The extent to which domiciliary assessment is available varies widely across Europe (Work Package 4 of the EuroCode project).
Nearly half of people over the age of 85 fulfil diagnostic criteria for dementia and many cases go unrecognised. Early recognition of dementia allows better understanding and treatment of symptoms, facilitates financial and care planning, enables access to research programmes and may improve medication adherence. However, systematic screening for cognitive impairment, either on a population basis or as part of systematic health checks for the elderly, is of unproven benefit. It is also potentially associated with significant harms. These include anxiety, stigma, restriction on driving and access to care homes, and societal costs. It is not recommended (II) (5-7) . More accurate biomarkers or screening instruments than those which are currently available, and better treatments, would be needed to justify population screening.
Nevertheless, where the prevalence of identified cases of dementia in a population is substantially less than the expected prevalence, then measures should be taken to improve the effectiveness of opportunistic case identification (III). This may, for example, include programmes directed at primary care physicians. Furthermore, patients over 75 years who are admitted to hospital should routinely have a cognitive examination because of the high prevalence of cognitive impairment in this group and because this facilitates early identification and treatment of delirium (II) .
The presenting symptoms should be elaborated in terms of their impact on the patient and those around them (I). The distinction between deficits in short and longer term memory, and between recognition memory, cued recall and uncued recall will often be mentioned by the family. If not spontaneously volunteered, active enquiry concerning deficits in memory, orientation, attention, word finding, sequencing and planning should be made (II).
The history should initially focus on the duration of symptoms and their pattern of progression (I).
Insidious onset and gradual progression suggest Alzheimer’s disease. Stepwise progression and onset following a vascular event suggest vascular dementia. Rapid progression over weeks should trigger investigation to exclude intracranial space occupying lesions, inflammation or infection cause (I). Wide fluctuations in functional and cognitive ability over short periods should increase diagnostic suspicion of the possibility of dementia with Lewy bodies (I). Other intercurrent illnesses such as systemic infections (I), and psychosocial stresses (II) are also common precipitants of more rapid decline. In people with high levels of cognitive reserve, memory difficulties can appear to progress rather rapidly once that cognitive reserve has been overcome. Educational level affects both risk and presentation of dementia and should be ascertained (I).
Concomitant illnesses and surgical procedures should be identified (I). The occurrence of worsening confusion or delirium following recent anaesthetics or illnesses should be enquired after (I).
Most common neurological impairments are likely to increase the risk of dementia. The nature, consequence and timing of cerebrovascular events in relation to cognitive decline should be established (I). Confusional episodes should not be definitely attributed to transient ischaemic attacks unless motor deficits occurred at the same time or there is good imaging evidence of relevant lesions (II). Substantial or repeated head injury, including a history of boxing, should be ascertained (I).
Diabetes mellitus and all other cardiovascular risk factors (high blood pressure, smoking, high cholesterol, renal failure) should be ascertained (I).
A complete list of the medication, especially any hypnotics or other psychotropic medication, should be taken (I). Attention should be paid to non-psychiatric and ‘over-the-counter’ drugs which have psychoactive effects (I). The ability of the patient to take medication reliably should be assessed (I).
Alcohol intake should be ascertained (I). If suspected, a history of illegal drug use should be taken (I).
The family history should include all first degree relatives (I). The presence of an informant is particularly useful to verify the accuracy of the information (I). Muddling of generations, particularly of children and grandchildren, is a common and useful sign of dementia (II).
The focus should be on a history of dementia, Parkinson’s disease, vascular disease and institutionalisation (I). The approximate age of onset of dementia in relatives should be ascertained because onset before the age of 70, and particularly before the age of 60, implies a greater genetic loading (II).
If the patient is to be adequately supported, those responsible for the patient’s care need to understand the support which has already been mobilised, and also the support which is potentially available (I). This includes ‘informal’ caregivers, typically spouses or children. Day care, domiciliary care, befriending services and services from the voluntary sector can all play a role at different points in dementia. Similarly, the advice and practical support of community-based nurses and other professionals allied to medicine can significantly alter the burden experienced by informal caregivers thereby enabling them to fulfil this role for longer and delay institutionalisation. These issues are discussed in more detail in Work Package 4 of the EuroCode project.
Symptoms are set in the context of previous level of occupational and social functioning, which should therefore be ascertained (I).
If an open-ended enquiry does not reveal any impact on day to day function, questions about the following functions (I), which tend to be affected early, are useful:
- remembering appointments
- remembering things which happened recently
- recalling conversations
- finding the right words
- taking medication accurately
- dealing with finances
Questionnaires completed by family, such as the Bristol ADL Scale, AD8, or CSADL can efficiently reveal the details of further deficits (II). Questionnaires which incorporate questions about driving safety and accuracy of taking medication are useful because they introduce topics which are potential threats to autonomy (I).
Behavioural and Psychological Assessment
Symptoms such as apathy, depression, anxiety, hallucinations, delusions, aggression and sleep problems are common but not invariable in dementia. Their occurrence varies with the stage of dementia. The importance of these non-cognitive symptoms (or ‘behavioural and psychological symptoms of dementia’) of dementia is such that assessment of each type of symptom is separately dealt with in more detail below, alongside consideration of the drug treatment of each. However, some principles are worth enumerating.
Principles of assessment of BPSD
Such symptoms are a source of burden to carers. They also often reflect and cause distress in the patient. To some extent, ‘a burden understood is a burden lifted’. Although carers often need no prompting to discuss such changes, an open-ended enquiry about changes in personality, behaviour and mood should therefore also be made both in the diagnostic work-up and regularly throughout the course of the illness (I). Carer-completed questionnaires may also be useful as a starting point for discussion (III). The brief caregiver-completed version of the Neuropsychiatric Inventory (NPI-CGA) and the similar Cambridge Behavioural Inventory (CBI) (8) cover the most important areas (II).
The characteristics of the behaviour and of any associated emotional state should be carefully described. This should include: nature, severity, duration, frequency, precipitants, consequences for the patient and consequences for others (I).
It is useful to bear in mind four models of the causes of changes in behaviour and mood when arriving at a formulation of the problem:
- Direct model: the changes result from brain changes
- Lowered stress threshold model : the patient is more susceptible, and liable to over-react to minor stressors
- Behavioural model (ABC): Antecedents result in a Behaviour which has Consequences. The consequences may or may not reinforce the behaviour
- Unmet needs model: A variation of ABC in which the antecedent is an unmet physical, emotional or idiosyncratic need and the behaviour is a cry for help to have that need fulfilled.
Physical health factors, which may generate or aggravate the behaviour, should be assessed including: urinary infection or retention, faecal impaction, possible undetected pain or discomfort and side effects of medication (I).
An empathic effort to understand why the patient is behaving as they are should also be made (I). This can include enquiry concerning: the approach taken by carers or nurses, boredom and level of stimulation, individual biography including religious beliefs and spiritual and cultural identity, and physical environmental factors, including restraint (III).
Particular attention should be paid to the patient’s mobility and gait (I). This will determine whether the dominant dementia-associated risk is of falls or getting lost. The differential diagnosis will be informed by the presence of Parkinsonian, vascular ‘petit pas’, and frontal apraxic gaits. Signs of stroke, including hyperreflexia, extensor plantar responses, and pseudobulbar palsy suggest vascular dementia.
Even mild parkinsonian tremor, facial masking, bradykinesia and rigidity are of diagnostic significance (I), suggesting dementia with Lewy bodies.
A slow pulse may increase the risk of symptomatic heart block with a cholinesterase inhibitor (II). High blood pressure is a risk factor for both Alzheimer’s disease and vascular dementia (I). Blood pressure declines in late Alzheimer’s disease (I). Carotid bruits are comparatively rare but, like other flow abnormalities, increase the risk of embolic stroke (I).
Brief Cognitive Tests
The gold standard of cognitive assessment is a comprehensive neuropsychological battery. However, the limited availability of time and of specialist neuropsychologists means that many brief cognitive instruments have been developed. These are inevitably less sensitive and specific, but are far faster and more accessible. On the other hand, brief tests which were developed to support case identification or diagnosis are sometimes inappropriately used as a substitute for diagnostic evaluation. They are also often inappropriately used to monitor change over short periods in patients with known impairment (9). Day to day fluctuations in cognitive function substantially limit the weight that can be placed on small changes in scores (I).
Care should be taken when using the term ‘screening’ in relation to the function of brief tests. ‘Primary screening’ applies to population-based strategies. ‘Secondary screening’ applies to the preliminary assessment of patients who have presented with a cognitive complaint, or in whom suspicions of impairment have been aroused.
The most important differences between the different tests lie in their sensitivity in the diagnosis of mild dementia or early cognitive impairment, the cognitive domains covered, and how long they take to administer. Several comprehensive reviews of brief tests exist (10-15)
Very brief tests which assess both memory and executive function and take less than 5 minutes are appropriately used for guiding the decision about which patients with suspected impairment require further evaluation (I). Multidomain tests lasting 5-15 minutes are best used to define the affected domains and to define overall level of impairment. However, they are unlikely to be sufficiently sensitive to change, in an individual, for monitoring short-term changes with treatment (I). Longer neuropsychological tests focussed on specific domains can then be appropriately used to further define deficits and for monitoring short-term changes (I). These include several computerised test batteries with varying degrees of utility for clinical and research application (16).
The interpretation of the results should always take account of overall prior functioning and educational attainment, areas of highly developed cognitive skills, sensory impairment and language, and the day-to-fluctuations which occur (I). In particular, deafness reduces the patient’s ability to concentrate in learning tasks and visual impairment affects tasks of constructional ability (I). Anxiety may impair performance, but can also arise because the patient recognises their inability to perform a difficult task (I).
Comprehensive accounts of each of these types of tests can be found elsewhere. Each test has its enthusiasts. Non-specialist clinicians should use and gain experience with the same instrument as their local specialists to facilitate communication (II).
Very brief tests taking less than 5 minutes
Memory function and executive function are the domains which are most commonly affected early. Often, one is affected without the other. Therefore, very brief assessments should, as a minimum, include assessment of both domains (II). However, since such brief tests are of imperfect sensitivity, brief questionnaires completed by family members or other informants are particularly useful in supplementing these very brief cognitive tests (I).
The GPCog (17;18) was developed for screening use in primary care and is more sensitive for the detection of dementia than the MMSE. It assesses memory and executive function. If the patient is unable to accurately recall an address and enter the numbers and hands on a clock, the caregiver is asked six questions about the patient’s functioning.
A combination of a clock drawing test, the Memory Impairment Screen and the IQCODE or AD8 would fulfil similar functions.
The widespread use of Clock Drawing Tests reflects good patient acceptability and speed (19;20). They assess the general executive functions of planning, sequencing and monitoring subserved by frontal lobes, and constructional ability. They lack sensitivity for the detection of mild Alzheimer’s disease but may be superior for detecting vascular dementia. Various scoring systems are used. The ‘CLOX’ system (21) has the advantage that it distinguishes executive and constructional impairments (Appendix) (22).
In the Memory Impairment Screen (23), subjects are given the names of items in four different categories (animal, city, vegetable, and musical instrument). After a short delay, the subjects are asked to name the items. Two points are scored for each item correctly recalled without the category cue, and one point is recalled after cueing. It is more sensitive than the MMSE for detection of Alzheimer’s disease but lacks sensitivity for the detection of vascular dementia.
The 16-item IQCODE (24) is completed by informants. It has been evaluated in multiple languages. Changes in the patient’s performance on memory and daily functioning items over the last 10 years are rated on a five point scale. Those who are ‘slightly worse’ on 6 of the 16 items will be screen-positive. The 8 item AD8 (25;26) evaluates change in 8 items over the last several years caused by thinking and memory problems. Decline in more than one area would be screen-positive. Where patients lack a carer and are only mildly affected, the AD8 may also be usefully selfadministered (27).
Multidomain tests taking 5-15 minutes
Sensitive brief measures of memory and executive function can be enhanced by further elements testing these and other domains (I). However, there are no brief cognitive tests that are sensitive to change throughout the whole range of impairment (II).
Like the GPCog, some longer multidomain instruments combine direct assessments of cognitive function with questions concerning day-to-day function. However, detailed assessment is now commonly split into separate domains.
A few multidomain instruments which can be reliably administered in under 15 minutes examine most of the domains which are commonly affected and are available in many European languages. These include the Montreal Cognitive Assessment (MOCA) (28), the 7-minute screen (29), the Demtect (30) and the modified MMS. An item analysis of instruments is presented in Appendix A.
The Minimental State Examination
The 30 point Minimental State Examination is not recommended for routine use either as a way of identifying cognitive impairment, or as a multidomain test for diagnosis (II). It lacks sensitivity for the diagnosis of mild dementia and does not include any measure of executive function. It takes 5-10 minutes to administer. It assesses orientation, concentration, language, memory and construction.
Although overall assessment of dementia is best done using instruments designed for the purpose (e.g. Global Deterioration Scale, Functional Assessment Staging Test), the MMSE has some merit for communicating dementia severity because it has good sensitivity and specificity for distinguishing moderate dementia from normal cognitive function (I). Importantly, it is universally used and understood, and often forms the basis of reimbursement decisions and recommendations of government agencies. Patients with a score of 20-26 are sometimes said to have ‘mild’ dementia although this terminology often under-represents the considerable impact on the patient’s life of impairment in this range (II). Those scoring 10 or less have ‘severe’ dementia.
Longer neuropsychological tests
Detailed assessment of visuoperceptual and language function tends to be the preserve of specialist neuropsychologists. However, the frequency of episodic and semantic memory deficits and executive function means that assessment of these should be within the range of dementia specialists (II).
The importance of episodic memory in the assessment of Alzheimer’s disease is such that specialists making a diagnosis of Alzheimer’s disease should routinely perform a more detailed assessment of this domain in patients with mild or equivocal dementia (I). Examples of this are word list learning tasks and story learning tasks. Suitable tests include an immediate supra-span recall task, a delayed recall task, and a recognition (or cued) memory task (I). Examples of these are the California Verbal Learning Task, the Hopkins Verbal Learning Test and Logical Memory Test from the Wechsler Memory Test battery.
Impairment in the ability to perform higher level activities of daily living correlates better with impairments in executive function than with impairment in episodic memory. Thus, irrespective of whether memory is impaired, testing of executive functions is needed (I). Suitable tests incorporate elements of planning, sequencing, and monitoring. Examples of these are Clock drawing, Trailmaking B, and the Tower of Hanoi. Performance on letter fluency (eg words beginning with F) and semantic fluency tasks (eg animal naming) also correlate with executive function.
Tests of the ability to name objects of progressively decreasing familiarity are sensitive to early disruption of the link between language and meaning. The Graded Naming Test and Boston Naming Test are examples.
Impairment in processing speed is commonly related to impairment in executive function. A suitable test is the Digit Symbol Coding test.
Several longer batteries which incorporate most or all of these domains exist, including the CAMCOG-R (31), the Addenbrookes Cognitive Examination (32), Repeatable Battery for Assessment of Neuropsychological Status (RBANS) (33), Consortium to Establish a Registry for Alzheimer’s Disease CERAD. Such batteries are mainly useful in clarifying whether or not dementia is present in equivocal cases (34).
Last Updated: Thursday 08 October 2009