Home > > Dementia: A NICE-SCIE Guideline on... > PREVENTION, EARLY IDENTIFICATION,...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Mental Health (UK). Dementia: A NICE-SCIE Guideline on Supporting People With Dementia and Their Carers in Health and Social Care. Leicester (UK): British Psychological Society; 2007. (NICE Clinical Guidelines, No. 42.)



Prevention of dementia syndromes would have a huge impact on large numbers of individuals and on society as a whole. Primary prevention, to avert early pathological changes, or secondary prevention, to delay pathological processes, are strategies that seem worthwhile pursuing. To do this, we must understand the factors that increase the risk of developing dementia and those that appear to be protective. Earlier identification of dementia may allow secondary prevention interventions, as well as early mobilisation of support and resources. To achieve earlier identification, we need to understand the ways in which cognitive function in dementia syndromes diverges from normal ageing processes. Diagnosis of dementia and its subtypes is often a complex process in which practitioners need to consider personal and informant histories, cognitive function testing and exclusion of other organic and psychological disorders. The importance of structural and functional neuroimaging is debated, as are the respective roles of primary- and secondary-care specialists in the diagnostic process. This step-wise process of reaching a diagnosis of dementia may involve assessment by a number of different professionals over a period of time. The process may raise difficult questions about how to discuss the possible or actual diagnosis, with whom and when. The experience of this assessment process may influence the way in which the person with dementia and his or her family or other carers assimilate and accommodate themselves to the diagnosis, and professional skills in managing the assessment process are therefore of great importance.


6.2.1. Introduction

Prevention of dementia must remain an ultimate goal. Even a delay in onset of dementia would effectively be a preventive strategy, since it has been estimated that delaying the onset of dementia by 5 years would half its prevalence (Jorm et al., 1987). Possible preventive strategies for dementia require consideration of:

  • knowledge about risk factors for dementia and its subtypes
  • the extent to which such risk factors are modifiable
  • evidence that modification of these risk factors does indeed result in subsequent reduction in the incidence of dementia.

Current practice

There is no systematic public-health strategy for the prevention of dementia in the UK.

The evidence base/limitations

The evidence base for the prevention of dementia is largely restricted to observational case control and cohort studies. The only prospective randomised controlled trials in prevention are of antihypertensive medication, hormone replacement therapy and statin therapy.

6.2.2. Non-modifiable risk factors for dementia


Advancing age remains the single biggest risk factor for developing AD, VaD and DLB, though it should be noted that some less common causes of dementia (for example, FTD, CJD and Huntington’s disease) are more common in mid life rather than at older ages (Harvey et al., 2003). Prevalence rates for AD double every 4.5 years after the age of 65 and those for VaD double every 5.3 years, leading to a steady increase with advancing age (Jorm & Jolley, 1998). Rates may stabilise at around 50–60% when people reach their mid 90s, but there are insufficient studies of ‘very old’ people to know this with any degree of certainty. Clearly age is not a modifiable risk factor.

Learning disabilities

The ageing process for people with learning disabilities may begin much earlier than for the general population; people with Down’s syndrome are at risk of developing a dementia of Alzheimer type about 30–40 years earlier than the rest of the population, and most researchers have found the prevalence of dementia in people with learning disabilities without Down’s syndrome is also raised to two or three times that expected in those aged over 65 years.


Prevalence studies show higher rates of dementia in women than men, especially for AD (Rocca et al., 1991). Rates for VaD are higher in men, though women tend to catch up at older ages. In part, the substantial difference (approximately 2:1) in prevalence between women and men in AD is due to women generally living longer than men. However, most, though not all, incidence studies also confirm higher rates of new cases in women, suggesting additional factors are involved. The reasons for the apparent increased susceptibility of women to develop AD are not clear, though a number of theories, including loss of oestrogens, have been proposed (but see section below on hormone replacement therapy) (Geerlings et al., 2001).


Several autosomal dominant forms of young-onset AD have been described, including mutations in the amyloid precursor protein, presenilin 1 and presenilin 2 genes. Such cases are rare (accounting for only about 1% of all AD) and characteristically have an age of onset below 55 years (Morris, 2005), although this may vary depending on the specific site of mutation (Lippa et al., 2000). Genetic testing after appropriate counselling can be provided for such individuals and for non-affected members of their families. In relation to late-onset AD, the main genetic risk factor described thus far is possession of the apolipoprotein E4 allele, which shows a dose effect and has a major impact in terms of bringing forward the age at which dementia develops (Cedazo-Minguez & Cowburn, 2001). It is also likely to be a risk factor for VaD and DLB (Hebert et al., 2000; Singleton et al., 2002). However, at least 50% of late-onset AD develops without the apoE4 allele, making it difficult to interpret the results of such tests for diagnostic purposes). Undoubtedly, genetic factors are important in late-onset AD and may explain up to half the liability to develop the disorder (Pedersen et al., 2004). Many other putative genetic risk factors have been described, but require further replication before they can definitively be accepted. There is much interest in potential therapeutic avenues that may be based on an understanding of the molecular changes caused by genetic factors associated with dementia, but apart from the possibility of pre-natal genetic diagnosis for known autosomal dominant cases, no modification of genetic risk factors is possible at the current time.

6.2.3. Potentially modifiable risk factors for dementia

Alcohol consumption

Excessive alcohol consumption is an established risk factor for dementia (Saunders et al., 1991). There is consistent evidence from epidemiological studies that moderate alcohol consumption is associated, both cross-sectionally and in prospective studies, with lower rates of cardiovascular disease, cerebrovascular disease and dementia (including Alzheimer’s and vascular disease) than either heavy drinking or abstinence (Letenneur, 2004). There is no consistent evidence that one type of alcoholic drink, for example, red wine, is more protective than another (Letenneur, 2004). These associative studies are not a basis on which to advise consumption of alcohol, though they do imply that drinking within the recommended range (up to 14 units per week for women and up to 21 units for men) is unlikely to increase risk of dementia.


Large prospective epidemiological studies have established that smoking is a risk factor for dementia in general and AD in particular (Ott et al., 1998), contradicting cross-sectional studies that had suggested that smoking may be protective (Wang et al., 1999). Smoking is clearly a major risk factor for cardiovascular and cerebrovascular disease, and as such increases the risk of stroke and VaD.


A number of recent prospective studies have supported an association between raised body mass index in mid life and subsequent increased risk of dementia in general and AD in particular (Gustafson et al., 2003; Kivipelto et al., 2005). Obesity also puts individuals at increased risk of developing type 2 diabetes, which is itself a risk factor for cerebrovascular disease and subsequent development of dementia (Ott et al., 1999; Biessels et al., 2006). No prospective studies have been undertaken to examine whether reducing obesity lowers risk of dementia.


Hypertension is a major risk factor for cardiovascular and cerebrovascular disease, including stroke, and antihypertensive treatment has been shown to reduce the incidence of both (Chalmers et al., 2003; Sacco et al., 2006; Williams et al., 2002). Longitudinal epidemiological studies have also established hypertension in mid life to be a risk factor for the subsequent development of VaD and AD (Skoog et al., 1996). Mid-life hypertension has also been associated with increased AD pathology at autopsy (Petrovitch et al., 2000). Because of this, a number of prospective randomised controlled trials have investigated antihypertensive treatments in terms of possible reduction of dementia risk. A meta-analysis of such studies is reported by Feigin and colleagues (2005). Their analysis included four studies, details of which are provided in Appendix 15a. Study inclusion criteria varied; for example, the Perindopril Protection Against Recurrent Stroke Study (PROGRESS) (Tzourio et al., 2003) included those with a history of stroke or transient ischaemic attack, whilst the other three studies included those with hypertension. Results are shown in Figure 1. Individually, the only study reporting a positive outcome was the SYST-EUR study (Forette et al., 1998) of a dihydropyridine calcium-channel blocker (nitrendipine), and this showed an approximate halving of the number of dementia cases (mostly a reduction in AD) in the treatment group compared to placebo, though numbers developing dementia were small (21 cases in the treatment group and 43 in the control group). Overall, when the four studies are combined, the effect size shows a non-significant trend (relative risk [RR] = 0.80, 95% confidence interval [CI], 0.63–1.02) to a reduction in dementia in treated subjects. This suggests that antihypertensive treatment may be a promising avenue for prevention of dementia, including AD and VaD, but that further studies are required. It will also be important for future studies to distinguish between potential specific pharmacological effects of the agent under consideration (for example, an action on calcium channels) and the effects of lowering blood pressure itself. It should also be remembered that there are already many evidence-based reasons for treating hypertension apart from reducing dementia risk, including reducing cardiovascular and cerebrovascular events.

Figure 1. Risk of dementia or cognitive decline reported in trials of antihypertensive drugs (data from Feigin et al., 2005).

Figure 1

Risk of dementia or cognitive decline reported in trials of antihypertensive drugs (data from Feigin et al., 2005).


Raised cholesterol level has been implicated as a risk factor in the development of dementia in some (Jick et al., 2000) but not all (Rea et al., 2005) studies. Not only is it a recognised risk factor for stroke, itself a major risk factor for VaD, but raised cholesterol has also been associated with the development of AD. Scott and Laake (2001) reviewed epidemiological studies examining the association between statin therapy and a reduced incidence of AD, but could not exclude bias as responsible for the observed associations. In terms of biomarkers, statin use has not been demonstrated to have any effect on serial brain measurement of hippocampal volume or amyloid biomarkers in plasma (Hoglund et al., 2004). A community-based prospective cohort study (Li et al., 2004) found that, while cross-sectional analysis revealed an apparent protective effect of previous statin use, prospective data showed no effect of statin use on subsequent development of either dementia in general or specifically AD. A similar finding was reported from the Canadian Study of Health and Aging (Rea et al., 2005). The PROSPER study (Shepherd et al., 2002) compared pravastatin to placebo and found a significantly reduced risk of coronary disease in treated subjects but no effect on cognitive function or stroke. In the heart protection study (HPSCG, 2002), there were no differences between simvastatin and placebo groups after 5 years’ treatment in terms of rates of dementia or rates of cognitive impairment. A preliminary study suggests a possible beneficial effect in established AD (Sparks et al., 2005) and further trials are ongoing. NICE recently published guidance on the initiation of statin therapy in adults with clinical evidence of cardiovascular disease and in adults considered to be at risk of cardiovascular disease (NICE, 2006).

Head injury

A number of epidemiological studies have shown that head injury sufficient to cause loss of consciousness increases the risk (an approximate doubling) of subsequent dementia (Guo et al., 2000; Plassman et al., 2000). A possible biological mechanism has been suggested by the finding that after acute head injury, cerebrospinal fluid (CSF) and brain amyloid levels are elevated (Olsson et al., 2004), though not all studies have supported this (Jellinger, 2004) and a meta-analysis of 15 epidemiological studies concluded that the risk was only apparent for males (Fleminger et al., 2003).

Low folate and raised homocysteine levels

Raised homocysteine levels, associated with low folate intake and low folate levels, have been associated with increased risk of cardiovascular and cerebrovascular disease and increased risk of dementias, including AD (Seshadri et al., 2002). There is much current interest in whether supplementation with B12 and/or folate (to reduce homocysteine levels) may prevent the development of dementia and perhaps slow its progression. Although results to date have been negative (Malouf et al., 2003), further studies are ongoing and results are awaited with interest.

Hormone replacement therapy

The apparent increased vulnerability of women to develop AD, together with findings from a number of epidemiological studies that past use of hormone replacement therapy (HRT) was associated with a decreased risk of dementia and AD, have prompted great interest in the possibility that HRT may delay or prevent the onset of dementia in general and AD in particular. A Cochrane review on the effects of HRT on cognitive function in cognitively intact post-menopausal women found little evidence that oestrogen therapy enhanced overall cognitive functioning (Hogervorst et al., 2002). More recently, Low and Anstey (2006) conducted a meta-analysis of 26 studies (17 cross-sectional, five longitudinal observational and four RCTs) on the association between HRT and cognition in cognitively intact post-menopausal women. A further analysis was made of 18 studies on HRT and risk of dementia. The results suggest that there is no consistent benefit in favour of HRT and there may be an inverse relationship between study design and effect size. For example, a large prospective RCT (the Women’s Health Initiative Memory Study) involving 4,532 post-menopausal women over the age of 65 and randomised to combined oestrogen or placebo found, not only that hormone replacement was not protective, but that it substantially increased the risk of dementia of any cause and cognitive decline (Shumaker et al., 2004). Rates of dementia were approximately doubled in the group receiving HRT.


The relationship between depression and dementia is a complex one. Rates of depression are increased in those with dementia, while depression can sometimes be an early prodrome or manifestation of a dementing illness (Jorm, 2000). In addition, those with depression have a variety of cognitive impairments, sometimes severe enough to mimic dementia, whilst depression has also been examined as a risk factor for dementia. A meta-analysis, which included seven case controls and six prospective studies, found an approximately doubling of subsequent risk of dementia in those with a history of depression (Jorm, 2000). However, there was insufficient evidence to determine the hypothesis by which this relationship occurred, which may be a) that depression is an early prodrome of dementia, b) both depression and dementia have shared risk factors (environmental or genetic), c) depression itself, or some changes associated with it (inflammatory, cortisol toxicity), increase subsequent risk of dementia. The MIRAGE study (Green et al., 2003) showed that depression does not simply appear to represent a prodrome of dementia, in that even depressive episodes occurring 25 years before the onset of cognitive impairment increase the risk of dementia. There are no prospective studies that have examined whether reducing depression subsequently reduces dementia risk.

Non-steroidal anti-inflammatory drugs

A number of studies, prompted by the inflammatory hypothesis of AD (McGeer & McGeer, 1999), have investigated whether the use of non-steroidal anti-inflammatory drugs (NSAIDs) may prevent subsequent development of dementia in general and AD in particular. A meta-analysis by De Craen and colleagues (2005) included 25 case control and cohort studies and showed that studies including prevalent and incident dementia cases showed a decreased risk of dementia in those using anti-inflammatory drugs, but not when cognitive decline was used as the clinical endpoint. The risk reduction of NSAIDs in preventing dementia or cognitive impairment was 50% in retrospective studies involving existing dementia cases, 20% in prospective studies examining new dementia cases and was absent when cognitive decline alone was used as the endpoint. To explain this, the authors suggested that many of the reported beneficial effects of NSAIDs may result from various forms of recall, prescription and publication biases. Other reviews (for example, Szekely et al., 2004) have suggested that duration of exposure may be important, with studies showing a duration of 2 years or more affording most benefit. The MIRAGE study (Yip et al., 2005) suggested a particular benefit among apoE4 carriers. A systematic review and meta-analysis of nine studies of NSAIDs showed significant reduction in subsequent risk of AD, especially for long-term use (over 2 years) (Etminan et al., 2003a). There was no effect in eight studies of aspirin. However, NSAIDs can have significant side effects, especially in older people. Whilst NSAIDs may yet have a role in prevention, further research is needed to establish dose, drug and duration of potential benefits, with careful consideration of potential risks. Effects need to be confirmed in prospective double blind, randomised trials.


Because oxidative damage has been associated with pathological changes of dementia and AD (Retz et al., 1998), it has long been hypothesised that antioxidants may provide some protection to the ageing brain in terms of reducing radical oxygen species and the damage they can cause. This has led to studies of antioxidants, particularly vitamin C (ascorbic acid) and vitamin E (D-alpha-tocopherol acetate in doses of 400 iu or more daily), both for AD and as a possible preventative measure. However, large-scale randomised controlled trials have not been undertaken and prospective observation studies examining this issue have come to differing conclusions (Boothby & Doering, 2005; Zandi et al., 2004; Luchsinger et al., 2003). Other evidence suggests that use of high doses of vitamin E for more than a year may be associated with an increase in all-cause mortality (Miller et al., 2005) and increase the incidence of heart failure in those with diabetes mellitus or pre-existing vascular disease (Lonn et al., 2005). The conclusion of a recent comprehensive review is that, in the absence of prospective, randomised controlled trials documenting benefits that clearly outweigh risks, vitamin E should not be recommended for primary or secondary prevention of AD or other dementias (Boothby & Doering, 2005). It further concludes that, although the risks of taking high doses of vitamin C are lower than those for vitamin E, there is no consistent efficacy data supporting its use.


Exercise has many recognised benefits, including beneficial effects on the cardiovascular system and bone density and effects on well-being and mood, as well as potentially increasing social interaction when undertaken in a group setting. Some evidence suggests an association between physical activity in mid life and decreased risk of dementia and AD later in life. A population-based study over 21 years showed that physical activity at least twice a week (defined as 20–30 minutes in duration and causing breathlessness and sweating) was associated with a reduced risk of dementia and AD (roughly a halving of risk) (Rovio et al., 2005). However, no randomised controlled trials have yet been undertaken to investigate whether starting to engage in exercise for those who currently do not will reduce their risk of developing dementia. In summary, engagement in physical activity lasting 20–30 minutes at least twice a week in mid life has been associated with decreased subsequent risk of dementia and AD. However, there is insufficient evidence to recommend physical activity specifically as a preventive measure for dementia, though there are many other reasons to encourage moderate exercise in everyone.

Education and mental stimulation

There is evidence that low educational attainment is associated with subsequent development of dementia (Valenzuela & Sachdev, 2005), but interpretation of this finding remains unclear, because whether this represents a true increased risk, or alteration of the threshold at which dementia becomes apparent, is uncertain. Of greater interest, is whether participation in cognitively stimulating activities can reduce the subsequent incidence of AD or other dementias (that is, the ‘use it or lose it’ hypothesis). A prospective study over 4.5 years showed that people participating in common cognitive activities had a decreased rate of subsequently developing AD (Wilson et al., 2002). However, such observations would also be compatible with the ‘cerebral reserve’ hypothesis, whereby those of higher ability, who would be more likely to engage in mentally stimulating activities, would also be at decreased risk of subsequently developing dementia. Randomised controlled trials have demonstrated cognitive benefits of cognitive training, which last up to 2 years, but without effect on measures of everyday functioning (Ball et al., 2002). Verghese and colleagues (2003) also found that participation in cognitively demanding leisure activities may protect against dementia. Over a median follow-up period of 5.1 years, the authors found that reading, playing board games, playing musical instruments and dancing were associated with a reduced risk of dementia. Similar results were found when AD and VaD were analysed separately. Whilst there are many reasons to encourage engagement in such activities, further studies to assess possible protective effects of cognitive activities on risk of dementia are needed.

6.2.4. Other risk factors for dementia

Other factors have been suggested as potential risk factors for dementia but await more definitive evidence. These include atrial fibrillation (Ott et al., 1997) and the consumption of saturated fats (Mental Health Foundation, 2006).

6.2.5. Evidence summary

Established non-modifiable risk factors for dementia in general and AD in particular include advancing age, genotype, female gender and having a learning disability. Established risk factors that are potentially modifiable include hypertension, excessive alcohol consumption, diabetes, depression and head injury. Other potentially modifiable risk factors may include obesity, raised homocysteine levels and raised cholesterol levels. Risk factors for VaD overlap with AD and include age, vascular risk factors (stroke, hypertension, diabetes and smoking) and apoE4 genotype. Protective factors for dementia may include prior long-term use of NSAIDs control of vascular risk factors, regular exercise and engagement in leisure and cognitively stimulating activities. However, thus far prospective randomised controlled trials have not clearly demonstrated that modification of risk factors leads to a reduction in dementia rates. Four RCTs of antihypertensive therapy showed a non-significant trend towards reduced dementia rates in treated subjects, two RCTs of statins found no effect and one study of HRT unexpectedly found increased dementia rates in treated people.

6.2.6. Health economics evidence

No evidence was identified by the systematic literature search on the cost effectiveness of interventions that can prevent or delay the onset of dementia.


6.3.1. Identifying dementia early

Population studies of ageing and cognition suggest that impairment in multiple cognitive domains is observable several years before a diagnosis of AD and other dementias is made. In these epidemiological studies, the observed cognitive deficit is not qualitatively different from that seen in ‘normal’ ageing, suggesting continuity rather than discontinuity in the shift from ‘normal’ ageing to pre-clinical dementia (Bäckman et al., 2004). Global cognitive deterioration occurs early, affecting episodic memory, executive functioning, verbal ability, visuospatial skills, attention and perceptual speed (Bäckman et al., 2004). There is no evidence as yet that these changes are detectable in individuals in clinical encounters.

Precise diagnosis of AD and other dementias in their early phases would be aided by an understanding of the initiating neurobiological events and the development of specific cognitive paradigms or biomarkers (DeKosky, 2003).

Longitudinal studies suggest that the magnitude of cognitive impairment may remain relatively constant for a period of several years. This phase corresponds to the clinical concept of ‘mild cognitive impairment’ (MCI), in which the individual has subjective symptoms (predominantly of memory loss) and measurable cognitive deficits but without significant impairment in usual activities of everyday life. There is a considerable overlap in cognitive performance between ‘normal’ ageing and this stable phase (Small et al., 2003).

At this stage, stringent tests of episodic memory are the best current neuropsychological predictors of subsequent conversion from MCI to AD at group level.

Imaging techniques can identify early brain changes, both structural and metabolic, but no single technique if used as a screening test can accurately identify individuals with MCI who will subsequently develop AD or other dementias (Nestor et al., 2004).

A combination of neuropsychological testing and neuroimaging improves the diagnostic accuracy of predicting cognitive decline in people in the MCI phase over that achieved with either modality alone (Chong & Sahadevan, 2003). However, the tools for identifying the early changes of AD and other dementias are outpacing the therapeutic options and so the usefulness of such very early ‘pre-clinical’ diagnosis currently remains uncertain (Chang & Silverman, 2004).

Although an evidence-based approach to case finding and case management can be described, a systematic review of the literature on early identification of dementia states that there is insufficient evidence of benefit to justify population screening in primary care (Boustani et al., 2003).

In AD, the stable phase ends with a precipitous decline in cognitive function, lasting between 2 and 5 years, in which semantic memory (the store of facts and general knowledge) and implicit memory (the non-conscious influence of past experience on subsequent performance) also become degraded (Spaan et al., 2003). Other dementia subtypes will follow different paths, for example, difficulties in attention and executive function will be prominent in VaD, language disturbances in types of frontotemporal dementia, and motor features and psychosis in DLB. People presenting with language disturbance may benefit from access to specialist speech and language therapy assessment and therapy within the multidisciplinary team at this early stage.

There is some evidence from the United States that early recognition and active therapy at this point (when early dementia can be diagnosed) delays the subsequent need for nursing home care and reduces the risk of misdiagnosis and inappropriate management (Chang & Silverman, 2004).

Earlier recognition of dementia syndromes can assist people with dementia and their families by dispelling anxiety about changes in memory, thinking, mood or behaviour and allowing mobilisation of resources that will be needed in the future. However, there are multiple obstacles to earlier diagnosis, both in terms of public awareness and professional understanding of dementia, which are still less than optimal (Iliffe et al., 2002). Earlier recognition of dementia syndromes would have major resource, training and organisational implications for multidisciplinary working, services and inter-agency working if the development of ‘care gaps’ is to be avoided (Manthorpe et al., 2003).

Personal awareness of changes in cognitive function in people with AD and other dementias is associated with better treatment outcomes from cognitive rehabilitation, but awareness can be difficult to assess, since individuals with AD may deny problems in one context but report awareness of them in another (Clare, 2004).

An evidence-based educational intervention using adult learning principles and conducted in the practice can improve diagnostic rates of dementia syndromes in general practice. Decision-support software designed to assist diagnostic and management thinking also improves recognition rates for dementia (Downs et al., 2006).

6.3.2. Evidence summary

The shift from ‘normal’ ageing to pre-clinical dementia appears to occur slowly and there is no evidence as yet that this change is detectable in individuals in clinical encounters. The magnitude of cognitive impairments that do emerge may remain relatively constant for a period of several years. The individual does have subjective symptoms (predominantly of memory loss) and measurable cognitive deficits, but without significant impairment in usual activities of everyday life. At present, tests of episodic memory are the best neuropsychological predictors of subsequent conversion from pre-clinical to clinical dementia. Imaging techniques can identify early brain changes, both structural and metabolic, but no single technique if used as a screening test can accurately identify individuals with pre-clinical dementia who will subsequently develop clinical dementia. The means of identification of the early changes of dementia syndromes are developing more rapidly than the therapeutic options and so the usefulness of such very early pre-clinical diagnosis currently remains uncertain. There is insufficient evidence of benefit to justify population screening in primary care. The stable pre-clinical phase ends with a sharp decline in cognitive function, its length being dependent on the subtype of dementia. There is some evidence that early recognition and active therapy at this point (when early dementia can be diagnosed) delays the subsequent need for nursing home care and reduces the risk of misdiagnosis and inappropriate management.

6.3.3. Health economics evidence

No economic evidence was identified regarding the early identification of dementia.


6.4.1. Introduction

Many conditions apart from dementia can present with cognitive impairments, most commonly delirium, depression, side effects of medication, other psychiatric illnesses, substance misuse and medical conditions like hypothyroidism, or intracerebral infections or space-occupying lesions. A careful and comprehensive assessment with appropriate investigations is therefore necessary to arrive at a diagnosis of dementia.

Current practice

Diagnosis of a dementia syndrome may often be appropriately made in primary care, though the knowledge and expertise to make a subtype-specific diagnosis resides in specialists. Most specialists undertake a broadly similar assessment in terms of history, mental state and use of routine blood tests, but use of other diagnostic tests (for example, imaging) varies considerably. Several different sets of diagnostic criteria exist, which are used to varying extents by different clinicians.

The evidence base/limitations

Current practice is largely based on cross-sectional and cohort studies that have assessed the usefulness of investigations in the clinical diagnosis of dementia and the accuracy of clinical diagnostic criteria as judged by autopsy verification. Definitive conclusions regarding the ‘added value’ of certain diagnostic tests over others are limited by the paucity of studies that assess investigations against autopsy in a truly blinded fashion and the relatively few studies that have directly compared one diagnostic test against another.

6.4.2. Diagnosis

As detailed in Chapter 4, dementia is a clinical diagnosis that is made when acquired cognitive deficits in more than one domain, representing a decline from a previously higher level of functioning, interfere with social and/or occupational functioning. Other features, including behavioural changes and symptoms such as depression, delusions and hallucinations, are commonly present. Dementia can result from a number of single or combined causes; it is usually progressive and can in some circumstances be reversible (for example, when due to a space-occupying lesion or infective or metabolic process).

Several different diagnostic criteria for dementia exist and have been shown to have good reliability (Knopman et al., 2001). The Diagnostic and Statistical Manual of Mental Disorders, third edition revised (DSM-IIIR; American Psychiatric Association, 1987) definition of dementia includes as an absolute requirement an impairment in memory, whilst the tenth edition of the International Classification of Diseases (ICD-10; World Health Organization, 1992) definition requires multiple cognitive disturbances, including memory. Such definitions are often appropriate for the dementia seen in AD, when memory is nearly always affected, but are not always appropriate for other disorders including VaD, FTD and DLB when memory is not predominantly impaired (O’Brien et al., 2003; Neary et al., 1998; McKeith et al., 1996, 2005).

As a result, a person may not meet ICD-10 or DSM-IIIR criteria for dementia because of an absence of memory impairment, despite having multiple acquired cognitive domains representing a decline from a previous level of functioning and causing social and/or occupational impairment.

6.4.3. Diagnostic criteria for subtypes of dementia

A number of different clinical diagnostic criteria have been proposed for the main subtypes of dementia.

Alzheimer’s disease

Several sets of diagnostic criteria for AD exist, including:

The most widely studied in terms of predictive ability are the DSM and the NINCDS/ADRDA criteria, while the NINCDS/ADRDA criteria have almost universally been adopted for entry into therapeutic trials. Both the DSM-IIIR criteria for dementia of Alzheimer type and the NINCDS/ADRDA criteria for probable AD have reasonable sensitivity (mean 81%) and specificity (mean 70%), while the NINCDS/ADRDA diagnosis of possible AD has higher sensitivity (mean 83%) at the price of loss of specificity (mean 48%) (Knopman et al., 2001). The diagnostic accuracy of the NINCDS/ADRDA criteria, their ability to subdivide into probable and possible AD and their widespread use in clinical therapeutic trials make them the preferred criteria for clinical diagnosis.

Vascular dementia

Criteria for the diagnosis of VaD include:

Criteria for subcortical ischaemic VaD have also been proposed (Erkinjuntti et al., 2000). As judged by pathology, sensitivity with all criteria tends to be low, around 50%, while specificity is good, around 85% (Gold et al., 2002). Therapeutic studies have generally used the NINCS/AIREN criteria, making them the preferred criteria for clinical diagnosis. However, neither these nor criteria for AD deal adequately with the issue of mixed dementia, which may be the most common subtype in clinical practice. In assessing people with possible dementia, clinicians will often be faced with difficulty, since many people may not fulfil strict criteria for a single subtype such as VaD or AD.

Dementia with Lewy bodies

Consensus diagnostic criteria for DLB exist (McKeith et al., 1996) and have been prospectively validated (McKeith et al., 2000b). In common with other criteria for diagnosis of subtypes of dementia, the criteria for DLB have low sensitivity (around 50%) but high specificity (Litvan et al., 2003; McKeith et al., 2005). Sensitivity for possible DLB criteria appears higher, but has been the subject of few investigations. These criteria have recently been reviewed and amended based on clinical expert discussion and review of the latest research evidence (McKeith et al., 2005). The consensus criteria have been almost universally adopted for use in all clinical research and therapeutic studies and are the criteria of choice. It should be noted that DLB may form part of a spectrum of Lewy body dementias, which would include dementia seen in Parkinson’s disease (PDD). DLB and PDD are separated in current diagnostic criteria by the duration of parkinsonian features (if present), which has to be more than one year before onset of dementia for PDD and less than one year for DLB.

Frontotemporal dementia

Two sets of diagnostic criteria for FTD have been proposed:

The Lund-Manchester criteria recognise three main syndromes, FTD, progressive nonfluent aphasia and semantic aphasia. The NINDS Work Group criteria recognise two main presentations, frontal (behavioural) and temporal (language). Neither set has been prospectively validated against autopsy.

Mixed dementia

Many people with dementia will have mixed disease; indeed the community-based Cognitive Function and Ageing Study showed that this was common in older people (MRC/CFAS, 2001). In such cases, until further evidence emerges to suggest otherwise, it is pragmatic to consider the clinical condition that best fits. For example, someone with mixed dementia whose dementia is predominantly thought to be due to AD would likely best be supported and managed as someone with AD. This is the approach taken by NICE in its technology appraisal of drugs for AD47.

6.4.4. Assessment

Diagnosis of a dementia syndrome can often be made in primary care, though if diagnosis is in doubt, referral to a specialist (old age psychiatrist, neurologist, physician in healthcare of older people or specialist GP, as deemed appropriate) should be undertaken. In most cases, subtype-specific diagnosis of the type of dementia will be required and people should be referred to a specialist with expertise in the differential diagnosis of the condition. Other reasons for referral to a specialist at any stage of dementia may include the presence of other behavioural changes and symptoms, especially if severe and non-responsive to initial treatment, or the need to access expertise or special investigations, services or treatments (for example, medication for the treatment of dementia) only available within secondary care. Agreed referral pathways between primary and secondary care should be in place for people with dementia, as recommended in the National Service Framework for older people (Department of Health, 2001b).


Diagnosis of dementia can only be made after a thorough assessment consisting of history, cognitive and mental state examination, physical examination and appropriate investigations. A detailed history from the person and, as the person may not be able to give a fully accurate history, a history from a relative or someone who knows the person well should be obtained, including the history of the presenting complaint, past medical and psychiatric history, medication use, drug or alcohol history, family medical and psychiatric history, a history of changes in personality or behaviour and assessment of changes in abilities to undertake everyday tasks. Useful standardised informant-administered assessment measures include the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) (Jorm & Jacomb, 1989) and measure of activities of daily living such as the Bristol Activities of Daily Living Scale (BADL) (Bucks et al., 1996).

Physical examination

A physical examination (including basic neurological examination) can detect evidence of physical disorders that can cause cognitive impairment, as well as other features that are important in making an accurate diagnosis, including the presence of focal neurological signs and motor features such as parkinsonism.

Mental state examination

A mental state examination should examine for the presence of psychiatric disorders that may cause diagnostic confusion with dementia (in particular depression) and also for associated non-cognitive symptoms that may be associated with dementia, including delusions, misidentifications and hallucinations that may have impact on diagnosis and management (Royal College of Psychiatrists, 2005b); Scottish Intercollegiate Guidelines Network, 2006).

Cognitive testing

Brief cognitive assessment

In order to make a clinical diagnosis of a dementia, cognitive assessment is essential. The aim of this is to determine which, and to what extent, different cognitive domains are affected. As dementia represents a change from the person’s previous level of function, life-long intellectual and educational attainment provides the context for current performance. Cognitive assessment includes testing attention and concentration, orientation, immediate and delayed memory, and higher cortical function including praxis, perception, language and executive function.

There are several standardised screening tests for the initial assessment of dementia that may be helpful adjuncts. The most widely used is the 30-item Mini Mental State Examination (MMSE), widely used in old age psychiatry and which takes around 5 minutes to complete. Others include the Newcastle Mental Test Score; the 7-minute screen (Meulen et al., 2004); the clock drawing test, useful for assessing praxis and executive function (Sunderland et al., 1989); the General Practitioner Assessment of Cognition (GPCOG) (Brodaty et al., 2002), which helpfully includes a brief informant rating as well as cognitive items; and the 6-Item Cognitive Impairment Test (6-CIT) (Brooke & Bullock, 1999), used in general practice. Advantages of standardised scales include the facts that norms are available, that severity of cognitive impairment can be quantified and easily communicated to others, that they serve as a useful baseline from which future change (for example, deterioration in the case of suspected dementia or improvement on medication or when delirium/depression resolves) can be accurately determined. Potential limitations, particularly of the MMSE, include associations with education level and sensitivity to depression (Orrell et al., 1992). There is some variance in terms of reliability, particularly when standardised instructions and procedures are not followed. These screening measures may be misleading with those with either high or low premorbid ability and with certain types of dementia (especially FTD and VaD). They do not generally adequately reflect important cognitive functions, such as executive function. There are also difficulties in their use with people with language impairment or sensory difficulties, or who are being tested in a second language. Where English is not the person’s first language, the use of an interpreter should be considered; where a screening test has been translated into the person’s first language, this should be utilised, so the interpreter may focus on the person’s responses (see for example, Gangulai et al., 1995; Kabir & Herlitz, 2000; Lindesay et al., 1997; Rait et al., 1997; Rait et al., 2000; Rowland et al., 2006). Published cut-off points to assist in the diagnosis of dementia should be interpreted cautiously, in view of the many other factors involved and the possibility of high false-positive rates (White et al., 2002). Cognitive impairment is not a good predictor of behavioural disturbance and carer stress, so only forms a part of the assessment needed.

Neuropsychological assessment

For those with mild or questionable impairment, and in other selected cases to assist with specific subtype diagnosis and differential diagnosis and to inform management, more comprehensive standardised cognitive assessments (for example, Cambridge Cognitive Examination – Revised (CAMCOG-R) (Roth et al., 1998; Williams et al., 2003), Addenbrooke’s Cognitive Examination (ACE) (Mathuranath et al., 2000), Alzheimer’s Disease Assessment Scale cognitive subscale (ADAS-cog) (Rosen et al., 1984), Middlesex Elderly Assessment of Mental State (MEAMS) (Golding, 1989) and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (Randolph, 1998)) may be needed. This would usually be undertaken as part of specialist referral. These more detailed assessments might form the first part of a full and detailed neuropsychological assessment by a clinical psychologist. Where there is also a significant impairment of language, an assessment by a speech and language therapist will contribute to the overall neuropsychological assessment. Such testing may provide important information regarding diagnosis and management, with specific comparisons made with predicted life-long levels of attainment and ability, but, if diagnosis is unclear, also provides a baseline against which any future cognitive change can be measured. This may be particularly important in cases where cognitive change has been identified but does not meet the diagnostic criteria for dementia, as in MCI.

Hentschel and colleagues (2005) provide evidence that a neuropsychological assessment (using the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) battery of tests) adds to the basic neuropsychiatric evaluation, with the initial diagnosis being changed in a significant number of cases. This occurs mainly at the borderline between ‘no dementia’ and ‘dementia’. In contrast, magnetic resonance imaging (MRI) scanning tends to change the subtype of dementia, rather than influence decisions about its presence. This suggests that neuropsychological assessment and MRI scanning may be complementary in the added value they provide to diagnostic assessment.

Blood tests and other investigations

There is no universal consensus on the appropriate diagnostic battery that should be undertaken in those with suspected dementia. However, a review of 14 guidelines and consensus statements found considerable similarity in recommendations (Beck et al., 2000). The main reason for undertaking investigations in a person with suspected dementia is to exclude a potentially reversible or modifying cause for the dementia and to help exclude common misdiagnoses including delirium. However, it should be noted that in recent studies the prevalence of reversible dementias is low. Clarfield (2003) reviewed 39 studies of over 7,000 cases and found potentially reversible causes in 9%, though only 0.6% of cases actually reversed. A recent consensus document from the Royal College of Psychiatrists recommends full blood count, erythrocyte sedimentation rate (ESR) or C-reactive protein (CRP), vitamin B12, folate, thyroid function, urea and electrolyte, calcium, liver function and glucose tests, with blood tests for syphilis, lipids and HIV listed as ‘optional’ (Royal College of Psychiatrists, 2005b). The recent Scottish Intercollegiate Guidelines Network (SIGN) dementia guideline (SIGN, 2006) does not recommend any specific blood tests and indicates that they should be selected ‘on clinical grounds according to history and clinical circumstances’.

Other investigations may include neuroimaging and cerebrospinal fluid analysis, but these are likely to be undertaken as part of further assessment for subtype diagnosis by a specialist with experience of differential diagnosis of the condition.

Structural neuroimaging

The two main forms of structural imaging are computed tomography (CT) and MRI. Local access to each varies considerably, while some forms of scanning are appropriate for some people and not others (for example, some find MRI scanning claustrophobic or have contraindications such as pacemakers or metallic implants). MRI is more costly than CT but superior in terms of anatomical visualisation of structural lesions, infarcts and white-matter pathology.

There are two main reasons for undertaking structural imaging in people with suspected dementia. The first is to exclude an intracerebral lesion (for example, a space-occupying or subdural lesion, or normal pressure hydrocephalus) as a cause for the cognitive impairment. Systematic reviews have suggested that between 2.2 and 5% of cases with suspected dementia had conditions that required structural neuroimaging to assist with diagnosis (Chui & Zhang, 1997; Clarfield, 2003).

Though such lesions can sometimes be suspected on clinical grounds by factors such as atypical history, early neurological signs, seizure, disturbance and short duration – factors that may prioritise those who undergo imaging if resources are limited (Royal College of Psychiatrists, 2005b) – a systematic review of six different clinical prediction rules for neuroimaging in dementia showed that most had poor sensitivity and all low specificity (Gifford et al., 2000).

The second use of structural imaging is to inform the subtype-specific diagnosis of dementia, in particular differentiating AD from VaD and FTD. Reported sensitivities and specificities in this regard vary depending on the method (that is, visual rating, volumetric or voxel based) and the anatomical area studied. A systematic review of over a hundred studies found several structural imaging measures, which clearly separated groups with AD, including early-stage, from controls, with overlap between groups of less than 6% for the hippocampus (Zakzanis et al., 2003). Similar changes in hippocampus and/or entorhinal cortex have been found, on a group basis, to differentiate those with mild impairment who subsequently progress to AD from those who remain stable (Jack et al., 1999; Stoub et al., 2005). Serial MRI scanning may also identify early brain changes before clinical onset of dementia (Scahill et al., 2002). Such brain imaging changes are not absolutely diagnostic in any individual but may be used to inform clinical judgement regarding diagnosis and likelihood of progression.

However, such structural imaging changes are less helpful in distinguishing AD from other types of dementia, including VaD and DLB, where atrophy of the hippocampus also occurs, albeit to a lesser extent than in AD (Barber et al., 1999). In FTD, frontal lobe atrophy may be seen on CT and MRI but, while this is a fairly specific marker, it can lack sensitivity. VaD is associated with a number of cerebrovascular changes, including cortical infarcts, lacunes and extensive white-matter lesions (Roman et al., 1993). Many lesions can be seen on CT, but MRI has greater sensitivity to detect small vascular lesions and subcortical white-matter change.

There is much variability in methods of analysis and reporting of such MR images, and some techniques (such as volumetric) are very time consuming to apply in clinical practice; further work is needed to harmonise and standardise methods on analysis and reporting of clinical CT and MRI studies before such methods could be widely adopted clinically.

Functional imaging

Single-photon emission computed tomography (SPECT) and positron emission tomography (PET)

Both perfusion hexamethylpropyleneamine oxime (HMPAO) SPECT and 2-[18F] fluoro-2-deoxy-D-glucose (FDG) PET can offer valuable information in the assessment and diagnosis of those with suspected dementia. A systematic review of diagnostic accuracy of HMPAO SPECT found pooled sensitivity of 77.1% and specificity of 89% in the separation of AD from normal comparison groups, with sensitivity of 71% and specificity of 76% separating AD from VaD, and sensitivity of 71% and specificity of 78% in separation from FTD (Dougall et al., 2004). In a comparison with clinical criteria (as judged by pathological verification), clinical criteria were found to be more sensitive in detecting AD than SPECT (81% versus 74%), but SPECT provided higher specificity against other types of dementia than clinical criteria (91% versus 70%) (Dougall et al., 2004). This supports the view that SPECT can be helpful in selected cases in the differentiation of AD, in particular from FTD and VaD. Other studies have suggested that perfusion SPECT is particularly helpful when there is diagnostic uncertainty, for example, in cases of possible as opposed to probable AD (Jagust et al., 2001).

Another form of SPECT imaging, iodine I 123-radiolabeled 2beta-carbomethoxy-3beta-(4-iodophenyl)-N-(3-fluoropropyl) nortropane (FP-CIT) has been licensed for the investigation of suspected parkinsonism and shows high sensitivity and specificity in separating cases of Parkinson’s disease from disorders such as essential tremor (Benamer et al., 2000). DLB is associated with nigrostriatal degeneration, even in people without clinical parkinsonism, and studies have suggested high sensitivity and specificity (one against autopsy verification) in terms of distinguishing AD from DLB (Walker et al., 2002; O’Brien et al., 2004; Ceravolo et al., 2004). Detection of degeneration of the dopaminergic system in people with dementia has been incorporated into the latest consensus diagnostic criteria for DLB (McKeith et al., 2005). As such, FP-CIT SPECT may give useful information in people suspected of having DLB.

PET scanning has been shown to improve the sensitivity and specificity of clinical criteria in much the same way as SPECT, and sensitivities of around 90% and specificity of 70% have been reported in pathological verification studies (Mosconi, 2005; Patwardhan et al., 2004). FDG PET may show some superiority over perfusion SPECT in detecting AD (Mielke & Heiss, 1998) but currently PET is not widely available in the UK and remains an expensive and invasive investigation.

Other neuroimaging techniques

Functional MRI, MR spectroscopy, diffusion weighted MR and other techniques have been investigated in the assessment and diagnosis of those with dementia. Other specific ligands and imaging markers are being developed, such as PET amyloid imaging (Klunk et al., 2004). Group differences between AD and other types of dementia have been reported for many of these modalities. However, their high cost, limited availability and lack of evidence base mean they cannot be recommended for clinical use at the current time.

Role of cerebrospinal fluid and other biomarkers

There is much interest in the possible development of tests, both in cerebrospinal fluid (CSF) and blood, for use in AD to allow earlier and more accurate diagnosis. Most studies to date that have been undertaken on CSF biomarkers include tau, phosphorylated tau and beta-amyloid 1–42. These three biomarkers have high sensitivity to differentiate AD from normal ageing, depression and some forms of dementia (for example, alcohol related) but have much lower specificity against other dementias, including FTD, DLB and VaD (Andreasen & Blennow, 2005). Some recent reports have suggested high sensitivity and specificity in relation to determining progression of those with mild cognitive impairment to AD but not other types of dementia. However, the majority of studies come from specialist centres, making widespread interpretation of results difficult, and there remain concerns and difficulties about reliability and standardisation of assays between different laboratories (Wiltfang et al., 2005). It also remains to be seen whether lumbar puncture to obtain CSF would be widely acceptable to people with dementia as a routine investigation. The 14-3-3 protein is associated with rapid neuronal loss and in most studies has been shown to have high sensitivity and specificity (both >90%) for CJD (Hsich et al., 1996).


Many studies have investigated the ability of the electroencephalogram (EEG) to separate AD from normal ageing (Jonkman, 1997) and, less commonly, other causes of dementia (Walker et al., 2000). Many rely on complex quantitative techniques, which are not applicable clinically. The resting EEG is often diffusely abnormal in dementia and may not be useful as a routine investigation. However, its use in selected cases can be helpful. Abnormal EEGs have been described in delirium (gross slowing) (Jacobson & Jerrier, 2000) and CJD (triphasic spikes) (Poser et al., 2000), while in FTD, EEG is reported to be normal (Neary et al., 1998).

Brain biopsy

In highly selected cases, a brain biopsy, usually of non-dominant frontal cortex, may sometimes be considered necessary when a treatable disorder, such as an infective, metabolic or inflammatory condition, is suspected but which cannot be diagnosed by other means. A retrospective study showed that this produced a diagnosis in 57% of cases, with 11% of people having complications such as bleeding or seizures (Warren et al., 2005). However, it should be noted that even in this highly selected group referred for biopsy, yields of reversible causes of dementia were low (around 10%).

6.4.5. Evidence summary

Standardised and widely accepted criteria exist for the diagnosis of subtypes of dementia including AD, VaD, DLB and FTD. These generally have high specificity but sensitivity can be low, and none deals adequately with those who have mixed causes of dementia. In those with suspected dementia, the most frequent reasons for misdiagnosis include delirium, depression, metabolic and endocrine disorders and other intracerebral pathologies. A standardised cognitive assessment tool is a useful adjunct to cognitive testing. Routine screening for syphilis and HIV is not indicated. CT scanning can detect most gross intracerebral pathology, but MRI has superior sensitivity and is preferred where available. Neuropsychological assessment can be helpful, especially in early cases, to help determine whether dementia is present or not. Blood-flow SPECT or FDG PET can detect functional changes in AD and be useful in differentiating AD, FTD and VaD. Dopaminergic SPECT or PET can detect nigrostriatal degeneration in vivo and can differentiate DLB and Parkinson’s disease from AD and VaD. CSF examination shows changes in AD, including increased levels of tau and phosphorylated tau and reduced levels of Abeta 1–42, though the role of these measurements in diagnosis and monitoring of AD and other dementias remains to be determined. The 14-3-3 protein shows a high sensitivity and specificity for the diagnosis of CJD. The resting EEG shows non-specific abnormalities in most types of dementia but is relatively normal in FTD and shows specific abnormalities in some cases of CJD. Brain biopsy may be considered appropriate in highly selected cases and can contribute in a minority of cases to accurate diagnosis that alters management.

6.4.6. Health economics evidence

Five studies were identified that addressed the cost effectiveness of neuroimaging tests for the diagnosis of dementia. Four studies, three in a US setting and one in a European setting, compared a range of neuroimaging tests to a standard examination that involved medical history, cognitive and functional assessment, laboratory tests and an MRI (McMahon et al., 2000 & 2003; Silverman et al., 2002; Moulin-Romsee et al., 2005). One study, LaFrance and colleagues (1998), compared dynamic susceptibility contrast MRI to SPECT. Characteristics and results of all studies are presented in the form of evidence tables in Appendix 18.

The diagnostic procedures and primary outcome measures were considered not relevant to the UK setting.


6.5.1. Introduction

The diagnosis of dementia is often a step-wise process in which the person with dementia is assessed by different professionals. The experience of this assessment process may influence the way in which the person with dementia and his or her family or other carers assimilate and accommodate to the diagnosis. Professionals involved in assessment and diagnosis of dementia need to understand the potential impact of the ways in which they respond to questions, offer and share information and provide support during the assessment process.

Professionals working with people with dementia can find the process of reaching a diagnosis difficult, because they may have to respond to anxiety, distress and disbelief on the part of the individual and those around him or her. Many professionals express concerns about how and when to convey the diagnosis, and to whom. This review of qualitative research evidence gives some guidance to assist practitioners in these tasks.

6.5.2. Sharing the diagnosis

The majority of people with mild dementia wish to know of their diagnosis (Pinner & Bouman, 2003; De Lepeleire et al., 2004), and all practitioners should assume that the diagnosis will be discussed with the person with dementia, unless there are clear reasons not to do so. The benefits of sharing a diagnosis include ending uncertainty, confirming suspicions, increasing understanding of problems, giving access to support, promoting positive coping strategies, facilitating planning and fulfilment of short-term goals (Bamford et al., 2004; Pratt & Wilkinson, 2003; Husband, 1999, 2000; Smith & Beattie, 2001).

People who have been informed of their diagnosis sometimes report that this disclosure process is badly handled, that little information is provided or that little or no follow-up occurs (Clare, 2004). A systematic review of the literature about sharing the diagnosis of dementia suggests that non-sharing of information or vague information about it is experienced as confusing, upsetting and difficult for some people with dementia and their families (Bamford et al., 2004).

However, there is also some evidence that clinicians do discuss disclosure carefully with people with suspected dementia and their families, to establish the best approach and what the person wishes to be told. During the assessment and diagnostic process, people should routinely be asked if they wish to know the diagnosis and with whom this should be shared (Pratt & Wilkinson, 2003; Pinner & Bouman, 2003).

Adjustment to the diagnosis of dementia is easier if the social context is supportive. Continuing support for people receiving a diagnosis is the responsibility of primary care. Necessary skills may include being able to anticipate and respond to shock, fear and grief.

People with severe dementia are often not given their diagnosis, although family carers are likely to receive it (Fahy et al., 2003). There may be issues around lack of insight and difficulties in retaining information, which make conveying information about diagnosis difficult if not impossible in such situations. However, this should be assessed and, wherever possible, responded to on an individual basis. There should be no automatic assumption that diagnosis should not be conveyed to the individual simply because of the perceived severity of dementia.

People with learning disabilities should be told their diagnosis of dementia; those supporting them should have access to specialist clinical advice and support about information sharing (Kerr & Wilkinson, 2005).

Clinicians may find the work of sharing the diagnosis with the individual and providing support difficult and may require their own support with such therapeutic work (Arber & Gallagher, 2003).

6.5.3. Confidentiality

Issues regarding confidentiality (that is, the question of with whom information is shared) should be addressed with people with dementia at the time of diagnosis and throughout their subsequent assessments or reviews (also see Section 4.9.5). Records and care plans should routinely note information about sharing agreements, outlining interprofessional exchange and communications with others, including carers, and the rare circumstances when such steps might need to be breached (for example, with regard to issues such as driving, safety and risk) (Pinfold et al., 2004; Tracy et al., 2004).

6.5.4. Support and information

For people with newly diagnosed dementia and members of their family, accurate details of local support should be available in primary and secondary care, social care and voluntary and community settings (Audit Commission, 2000; Cornwall County Council, 2005). Information is most useful if people with dementia have been involved in its compilation and presentation, and examples of good practice should be adapted for local situations (McKillop & Wilkinson, 2004; Scottish Action on Dementia48).

Family members should have access to information and advice from sources that are able to address their different concerns, which may be different to those anticipated by clinicians (Gely-Nargeot et al., 2003). People with financial problems may find care-giving particularly difficult and may benefit from early referral to financial or debt advice agencies (Schneider et al., 1999). People with dementia who also have another mental health problem, such as depression, may need particular and individualised help to access support (Manthorpe & Iliffe, 2005). Information for children in families where a member has dementia should be available, particularly for children and young people who act as carers. Voluntary sector organisations are a major source of information for people with dementia and their families, but their resources are often limited and capacity-building is necessary at local level (Manthorpe et al., 2003).

When working with people from minority communities where terms like dementia may not be widely known, practitioners should develop or draw on specialist support and publications (Bowes & Wilkinson 2003; Adamson, 2001).

6.5.5. Assessment and support

Assessment experiences can induce shame and distress and can lead to people trying to maintain their sense of identity, and this can be perceived as ‘covering up’ or loss of insight (Keady et al., 1995; Cheston et al., 2003a, 2003b). Support groups for people with newly diagnosed dementia promote well-being. These may be experiential group therapy, allowing people to explore the experience of dementia in a safe, supportive and secure setting (Cheston, 1996; Cheston & Jones, 2000; Cheston et al., 2003a, 2003b). Family therapy services show promise at helping people with dementia and their families where psychiatric problems exist (in the person with dementia or his or her family) that are beyond the skills of primary care practitioners (Benbow et al., 1993).

Delays in referrals to diagnostic and assessment services are experienced as stressful by relatives of people with dementia (Sperlinger & Furst, 1994). Continuity of care matters and services that facilitate continuity of contact and support are appreciated by people with dementia and their carers, who dislike discontinuity associated with hospital-based services and prefer community-based assessment services (Timlin et al., 2005).

At all stages of contact with services, advocacy for people with dementia should be available but at present is widely variable in scale and scope (Cantley et al., 2003).

6.5.6. Evidence summary

The majority of people with mild dementia wish to know of their diagnosis. All practitioners should assume that the diagnosis will be discussed with the person with dementia, unless there are clear reasons not to do so. Failure to share or provision of vague information about the diagnosis is experienced as confusing. There should be no automatic assumption that diagnosis should not be conveyed to the individual simply because of the perceived severity of dementia. Adjustment to the diagnosis of dementia is easier if the social context is supportive and continuing support for people receiving the diagnosis is the responsibility of primary care.

Family members should have access to information and advice from sources that are able to address their different concerns, which may be different to those anticipated by clinicians. Accurate details of local support should be available in primary and secondary care, social care and voluntary and community settings for people with newly diagnosed dementia and members of their family. Assessment experiences can induce shame and distress and can lead to people trying to maintain their sense of identity, and this can be perceived as ‘covering up’ or loss of insight. Services that facilitate continuity of contact and support are appreciated by people with dementia and their carers.


6.6.1. Risk factors, screening and genetic counselling

General population screening for dementia should not be undertaken. [For the evidence, see section 6.3]

In middle-aged and older people, vascular and other modifiable risk factors for dementia (for example, smoking, excessive alcohol consumption, obesity, diabetes, hypertension and raised cholesterol) should be reviewed and, if appropriate, treated. [For the evidence, see sections 4.5.2 and 6.2]

Healthcare professionals working with people likely to have a genetic cause for their dementia (for example, familial autosomal dominant AD or FTD, CADASIL, or Huntington’s disease) should offer to refer them and their unaffected relatives for genetic counselling. [For the evidence, see sections 4.5.1 and 6.2.2]

Regional genetic services should provide genetic counselling to people who are likely to have a genetic cause for their dementia and their unaffected relatives. [For the evidence, see sections 4.5.1 and 6.2.2]

If a genetic cause for dementia is not suspected, including late-onset dementia, genotyping should not be undertaken for clinical purposes. [For the evidence, see sections 4.5.1 and 6.2.2]

6.6.2. Preventive measures

The following interventions should not be prescribed as specific treatments for the primary prevention of dementia:

For the secondary prevention of dementia, vascular and other modifiable risk factors (for example, smoking, excessive alcohol consumption, obesity, diabetes, hypertension and raised cholesterol) should be reviewed in people with dementia and, if appropriate, treated. [For the evidence, see sections 4.5.2 and 6.2.3]

6.6.3. Early identification of dementia

Primary healthcare staff should consider referring people who show signs of MCI49 for assessment by memory assessment services to aid early identification of dementia, because more than 50% of people with MCI later develop dementia. [For the evidence, see sections 4.1.1 and 6.3]

Those undertaking health checks as part of health facilitation for people with learning disabilities should be aware of the increased risk of dementia in this group. Those undertaking health checks for other high-risk groups, for example those who have had a stroke and those with neurological conditions such as Parkinson’s disease, should also be aware of the possibility of dementia. [For the evidence, see sections 4.1.1, 4.1.4, 4.5.2, 4.6.3, 6.2.2 and 6.2.3]

Memory assessment services that identify people with MCI (including those without memory impairment, which may be absent in the earlier stages of non-Alzheimer’s dementias) should offer follow-up to monitor cognitive decline and other signs of possible dementia in order to plan care at an early stage. [For the evidence, see sections 4.1.1, 4.6.1, 4.6.2 and 6.3]

6.6.4. Recognition

A diagnosis of dementia should be made only after a comprehensive assessment, which should include:

  • history taking
  • cognitive and mental state examination
  • physical examination and other appropriate investigations
  • a review of medication in order to identify and minimise use of drugs, including over-the-counter products, that may adversely affect cognitive functioning. [For the evidence, see sections 4.1.1, 4.1.2, 4.6.1, 4.6.2, 4.6.3 and 6.4]

People who are assessed for the possibility of dementia should be asked if they wish to know the diagnosis and with whom this should be shared. [For the evidence, see section 6.5.2]

Clinical cognitive assessment in those with suspected dementia should include examination of attention and concentration, orientation, short and long-term memory, praxis, language and executive function. As part of this assessment, formal cognitive testing should be undertaken using a standardised instrument. The MMSE has been frequently used for this purpose, but a number of alternatives are now available, such as the 6-CIT, the GPCOG and the 7-Minute Screen. Those interpreting the scores of such tests should take full account of other factors known to affect performance, including educational level, skills, prior level of functioning and attainment, language, and any sensory impairments, psychiatric illness or physical/neurological problems. [For the evidence, see section 6.4.4]

Formal neuropsychological testing should form part of the assessment in cases of mild or questionable dementia. [For the evidence, see section 6.4.4]

6.6.5. Investigation

A basic dementia screen should be performed at the time of presentation, usually within primary care. It should include:

Testing for syphilis serology or HIV should not be routinely undertaken in the investigation of people with suspected dementia. These tests should be considered only in those with histories suggesting they are at risk or if the clinical picture dictates. [For the evidence, see section 6.4.4]

A midstream urine test should always be carried out if delirium is a possibility. [For the evidence, see section 6.4.4]

Clinical presentation should determine whether investigations such as chest X-ray or electrocardiogram are needed. [For the evidence, see section 6.4.4]

Cerebrospinal fluid examination should not be performed as a routine investigation for dementia. [For the evidence, see section 6.4.4]

6.6.6. Diagnosis of subtypes

A diagnosis of subtype of dementia should be made by healthcare professionals with expertise in differential diagnosis using international standardised criteria (see Table 6). [For the evidence, see section 6.4]

Structural imaging should be used in the assessment of people with suspected dementia to exclude other cerebral pathologies and to help establish the subtype diagnosis. MRI is the preferred modality to assist with early diagnosis and detect subcortical vascular changes, although CT scanning could be used. Specialist advice should be taken when interpreting scans in people with learning disabilities. [For the evidence, see section 6.4.4]

HMPAO SPECT should be used to help differentiate AD, VaD and FTD if the diagnosis is in doubt. People with Down’s syndrome may show SPECT abnormalities throughout life that resemble those in Alzheimer’s disease, so this test is not helpful in this group. [For the evidence, see section 6.4.4]

If HMPAO SPECT is unavailable, FDG PET should be considered to help differentiate between AD, VaD and FTD if the diagnosis is in doubt. [For the evidence, see section 6.4.4]

FP-CIT SPECT should be used to help establish the diagnosis in those with suspected DLB if the diagnosis is in doubt. [For the evidence, see section 6.4.4]

CSF examination should be used if CJD or other forms of rapidly progressive dementia are suspected. [For the evidence, see section 6.4.4]

EEG should not be used as a routine investigation in people with dementia. [For the evidence, see section 6.4.4]

EEG should be considered if a diagnosis of delirium, FTD or CJD is suspected, or in the assessment of associated seizure disorder in those with dementia. [For the evidence, see section 6.4.4]

Brain biopsy for diagnostic purposes should be considered only in highly selected people whose dementia is thought to be due to a potentially reversible condition that cannot be diagnosed in any other way. [For the evidence, see section 6.4.4]

Table 6. Diagnostic criteria for dementia.

Table 6

Diagnostic criteria for dementia.

6.6.7. Mixed dementias

Many cases of dementia may have mixed pathology (for example, AD and VaD or AD and DLB). Unless otherwise stated in this guideline, such cases should be managed according to the condition that is thought to be the predominant cause of dementia. [For the evidence, see section 6.4.3]

6.6.8. Specialist services for dementia assessment

Memory assessment services (which may be provided by a memory assessment clinic or by community mental health teams) should be the single point of referral for all people with a possible diagnosis of dementia. [For the evidence, see section 4.6.2]

Memory assessment services should offer a responsive service to aid early identification and should include a full range of assessment, diagnostic, therapeutic, and rehabilitation services to accommodate the needs of people with different types and all severities of dementia and the needs of their carers and family. [For the evidence, see sections 4.6.2, 4.8 and 6.4]

Memory assessment services should ensure an integrated approach to the care of people with dementia and the support of their carers, in partnership with local health, social care, and voluntary organisations. [For the evidence, see sections 5.2.2 and 5.2.3]

6.6.9. Addressing needs that arise from the diagnosis of dementia

The experience of the diagnosis of dementia is challenging both for people with dementia and family members and for healthcare professionals, so healthcare professionals should make time available to discuss the diagnosis and its implications with the person with dementia and also with family members (usually only with the consent of the person with dementia). Healthcare professionals should be aware that people with dementia and family members may need ongoing support to cope with the difficulties presented by the diagnosis. [For the evidence, see sections 4.1.2, 4.1.3, 4.1.4, 4.2, 4.3, 4.6.1, 4.6.2, 4.7.2, 4.8, 4.10.2, 6.1, 6.3, 6.5, 9.2 and 9.5]

Following a diagnosis of dementia, health and social care professionals should, unless the person with dementia clearly indicates to the contrary, provide them and their family with written information about:

  • the signs and symptoms of dementia
  • the course and prognosis of the condition
  • local care and support services
  • support groups
  • sources of financial and legal advice, and advocacy
  • medico-legal issues, including driving
  • local information sources, including libraries and voluntary organisations.

Any advice and information given should be recorded in the notes. [For the evidence, see sections 4.8 and 6.5]

Healthcare professionals who regularly diagnose dementia and discuss this with people with the condition and carers should consider mentoring or providing clinical supervision to less experienced colleagues. [For the evidence, see sections 4.8, 6.5.1 and 6.5.2]



For further information see www​.nice.org.uk/guidance/TA111.


MCI is a syndrome defined as cognitive decline greater than expected for an individual’s age and education level, which does not interfere notably with ADLs. It is not a diagnosis of dementia of any type, although it may lead to dementia in some cases.

Copyright © 2007, The British Psychological Society & The Royal College of Psychiatrists.

All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Enquiries in this regard should be directed to the British Psychological Society.

Cover of Dementia
Dementia: A NICE-SCIE Guideline on Supporting People With Dementia and Their Carers in Health and Social Care.
NICE Clinical Guidelines, No. 42.
National Collaborating Centre for Mental Health (UK).
Leicester (UK): British Psychological Society; 2007.

NICE (National Institute for Health and Care Excellence)

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...