Basket | Login



Diagnosis and treatment of dementia

Diagnostic Criteria

Principles of application of diagnostic criteria

Currently, to fulfil diagnostic criteria for ICD-10 or DSM4 ‘dementia’, increased impairment of social or occupational functioning is necessary. This definition has advantages and disadvantages. On the one hand, it introduces a threshold based on the practical impact of the disease. On the other hand, the impact of neurodegeneration on social and occupational functioning varies widely depending on factors which are unrelated to the extent of neurodegeneration. These include the complexity of intellectual demands, the level of premorbid functioning, educational attainment and social support. Furthermore, although some instruments designed to measure degree of functional impairment have cut-offs which have been validated for use in routine practice (against the gold standard of ‘expert clinician’ diagnosis), none has become established in this role.

These limitations have prompted a proposal for a new set of diagnostic criteria for AD: as well as deficits in episodic memory, the presence of abnormalities in quantitative structural imaging, amyloid imaging, or CSF analysis of Ab and tau is required (Dubois).

Routine quantitative evaluation of invasive or expensive biomarkers will be justified if either of two circumstances are fulfilled:

  • the biomarker correlates closely with a disease process which influences symptoms and is modifiable by treatment
  • the biomarker adds significantly to the prediction of decline from MCI to dementia, or of rate of decline, afforded by the application of a panel of less invasive or expensive, but adequately reliable tests (eg episodic memory, age, apoE, functional impairment)

However, neither of these criteria has been demonstrably fulfilled yet (I). Another, perhaps more tractable, current limitation in our knowledge is that there are no simple, validated, operationalised algorithms which can be applied in routine clinical practice to predict symptom progression.

Clinicians should therefore continue to refer to standard diagnostic criteria when making a diagnosis of dementia. These are enumerated in Appendix B.

Disclosure of diagnosis, prognosis and implications

Disclosure of diagnosis and prognosis should be done tactfully (I). Often, this is done when the most likely cause of dementia has been established, rather than at the point of initial presentation. This gives time for families to recognize the likely outcome of the investigations.

Families sometime ask for the diagnosis to be withheld from the patient. This causes a difficult ethical dilemma because it can be more difficult for patients to come to terms with their illness or plan for the future if they are not told their diagnosis. One way of approaching this (III) is to ask the patient whether he or she wants to know

a) what is causing the cognitive impairment

b) whether it is going to get worse

c) whether there is any treatment.

Often, patients who do not want to be confronted with their diagnosis will simply elect to hear about treatment rather than aetiology.

Criteria for Dementia

DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, fourth edition; or ICD-10, International Classification of Diseases, 10th revision;

Criteria for Alzheimer’s Disease

National Institute of Neurologic, Communicative Disorders andStroke–AD and related Disorders Association

Criteria for Vascular Dementia

National Institute of Neurological Disorders and Stroke Association Internationale pour la Recherche et l’Enseignement en Neurosciences

Criteria for dementia with Lewy bodies

International Consensus criteria for dementia with Lewy bodies

Criteria for Frontotemporal dementia

Lund-Manchester criteria, or

NINDS criteria for frontotemporal dementia

Frontal Behavioural Inventory

Middelheim Frontality Score

Criteria for Mixed Dementia

Instead of the term ‘mixed-dementia’, list the suspected dementia subtypes and treat as for the most prominent features.


Currently, treatment decisions and management of patients with dementia are based on clinical signs and symptoms. Investigations are usually currently orientated at the identification of treatable pathologies, establishing the extent of cerebrovascular disease, and identifying rarer causes dementia. More recently research has been directed at identifying tests which predict cognitive decline or distinguish between its causes.


Blood tests should be conducted to rule out common co-morbid conditions and to identify potentially reversible causes of dementia (I).

Routine blood tests to be included in the diagnostic work-up are:

  • Blood sedimentation rate or CRT
  • Complete blood cell count
  • Electrolytes
  • Calcium
  • Glucose
  • Renal function tests
  • Liver function tests
  • Thyroid function tests
  • Serum vitamin B12 and folate
  • Cholesterol

A wide range of blood tests may be considered if the above are uninformative and specific aspects of the history or examination are suggestive (III). These tests, which are rarely needed and very rarely informative, include:

  • Serology
    • Syphilis
    • HIV
    • Borelliosis
    • Herpes
    • Lead and mercury levels
    • Vitamin and hormone levels (B1, B6, niacin, cortisol, parathyroid hormone),
    • Selenium/bismuth levels if corresponding preparations are ingested.
    • 24h urinary copper clearance


A urine specimen should routinely be dip-tested, even in the absence of urinary symptoms (III). This is because urinary infection commonly causes abrupt deteriorations of mental state. It is also commonly present without causing impairments – deterioration may then be incorrectly attributed to longstanding bacteruria.

Genetic Testing

Early onset dementia: Patients and caregivers should be alerted to the possible genetic implications in cases of dementia onset before the age of 60 years where another first degree family member has had early onset dementia (I). Referral of family members to a clinical genetics service is recommended when desired (II). Patients and caregivers should be alerted to the possibility of ‘banking’ blood which can be available for future genetic analysis (II). Genetic testing for autosomal dominant dementias should not be carried out outside specialist centres with informed consent and counselling of all involved (I).

Late onset dementia: ApoE4 is associated with memory decline in normal ageing and brings forward the time of onset in those who are destined to develop dementia. Its impact is most relevant to those in their 60s and 70s. Patients with MCI who have two copies of ApoE4 are also more likely to progress to AD than those without. However, up to 50% of cases who are homozygous and live beyond the age of 90 do not develop dementia and about two thirds of those who develop dementia have no e4 allele. There are no validated algorithms either for the differential diagnosis of dementia, or for the prediction of progression from MCI to dementia, which incorporate ApoE status. ApoE4 testing is therefore not currently recommended in MCI or dementia (I).

Although ApoE status predicts long-term cognitive outcome following cerebral infarction, haemorrhage or trauma and possibly following coronary surgery, there is no evidence either way that different management in those with differing ApoE status alters outcome. ApoE4 testing is therefore not recommended following cerebral insults or before major surgery (I).

CSF Analysis

No CSF biomarker has a license in any jurisdiction for the prediction of AD or dementia, or for the differential diagnosis of dementia. However, examination of CSF (with routine cell count, protein, glucose, and protein electrophoresis) is recommended when CJD, inflammatory disease, vasculitis or demyelination is suspected (I). The 14-3-3 protein, a marker of acute neuronal loss and brain damage, is associated with CJD.

Experience of routine use of lumbar puncture outside research centres is uncommon. In such centres, the side-effect profile of lumbar puncture in older patients with dementia or MCI is generally benign (II), although rates of headache requiring bedrest of 2% have been reported. The acceptability and feasibility of routine lumbar puncture will vary according to setting and culture.

Longitudinal research suggests that high CSF levels of phosphorlylated tau, high CSF-tau and low CSF Ab42 may predict conversion from MCI to AD. The ratio of tau (or phospho-tau) to Ab42 may enhance the specificity and sensitivity of this prediction to above 80%. There are early indications that CSF measurements of p-tau and Abeta42 can assist in the differential diagnosis of AD, vascular dementia and frontotemporal dementia though these studies do not have post-mortem validation. Reliable cut-offs and standardisation of assays across centres have yet to be agreed. There are no data suggesting that CSF biomarkers significantly improve the prediction of progression from MCI to AD over and above that of more readily accessible risk factors (age, severity and multiplicity of cognitive deficits) are sparse.

Routine CSF estimation of tau, phospho-tau and Ab42 is not recommended until these issues have been addressed (I).


Structural neuroimaging

Structural imaging is recommended as part of the diagnostic process (II). A CT or MRI scan should routinely be used for the following reasons

  • to identify rare potentially operable lesions (eg meningiomas, subdural haematoma, normal pressure hydrocephalus)
  • to help distinguish dementia subtypes
  • the establish the presence, location, severity and characteristics of cerebrovascular disease
  • to help caregivers and patients understand that dementia is associated with brain pathology

CT or MRI scanning should always be used routinely in young onset dementia (<60 years), and dementia of brief duration (<2 years) to exclude non-degenerative causes (I).

For late onset dementia, the greater availability, speed and tolerability of CT make it a satisfactory choice when the patient presents with clearcut dementia of moderate severity (II).

However, MRI is recommended in equivocal or mild dementia (II). Semiquantitative assessment of hippocampal atrophy can be done reasonably reliably and this practice is something to which centres of excellence should aspire (III). The sensitivity and specificity of a single examination at the point of expert clinician diagnosis of mild or moderate AD is 85% and 88% (35), which is a modest improvement on the specificity of the NINCDS_ADRDA criteria. The accuracy of hippocampal atrophy for a confirmed diagnosis of mild Alzheimer’s disease is 67-100 (36). The use of automated software for volumetric analysis of MRI depends on careful calibration of scanners and has not yet reached the stage where it can be recommended for routine practice (I).

The better resolution of MRI makes it is more useful than CT for the assessment of vascular pathology (II). However, the inter-rater reliability of vascular assessment is moderate at best (37).

Functional neuroimaging


SPECT imaging of cerebral blood flow (99mTc-HMPAO or 133Xe) may be useful as an adjunct to CT or MRI scans where diagnostic uncertainty still exists. It is recommended as a option for distinguishing AD from FTD where the sensitivity and specificity are 71% and 78% (38) (II).

Dopamine transporter imaging (DAT-scan) has a license for distinguishing AD from dementia with Lewy bodies. It improves the prediction of a final clinical diagnosis of DLB in patients with isolated symptoms of parkinsonism, fluctuation, visual hallucinations or REM-sleep behaviour disorder (39). Its use is therefore recommended in patients with ‘Possible’ DLB (II).

Fluorodeoxyglucose FDG PET

The incremental diagnostic value over and above consensus clinical criteria at initial evaluation in routine secondary care clinical populations of techniques employing PET has been established (40) (41). The value is roughly equivalent to the extra clinical knowledge gained over 4 years of followup (42). When available, its use can be recommended when early diagnosis is important (II)

Amyloid ligand PET

Amyloid imaging is currently well advanced as a research tool but is not yet available outside research centres. Several ligands with a short half life are available. None has a license. Flourine-based ligands, which have a longer half life, are currently being evaluated. Early evidence suggests that amyloid imaging can aid in the prediction of progression from MCI to dementia (43)(II). However, routine clinical deployment of amyloid imaging is not recommended unless future therapies which reduce amyloid deposition are shown to affect the course of dementia (III).


The value of EEG in routine evaluation of dementia lies in excluding pathology (III): a normal EEG has a negative predictive value of 82% for the subsequent presence of AD at pathology (44). Conversely, an abnormal EEG results in a diagnostic gain (38%) only when the prior probability is low (45). The fluctuating attention of DLB is associated with EEG changes which may precede a definite diagnosis of DLB and be of diagnostic utility (46). The presence of periodic sharp wave complexes predict neuropathological CJD with 66% sensitivity and 74% specificity (47)) and are part of the diagnostic criteria for sporadic CJD. An EEG is therefore indicated when CJD is suspected (II). EEG is also diagnostic in cases where focal temporal lobe seizures cause transient epileptic amnesia (I).

Other Investigations

Chest X-Ray and ECG are not routinely needed as part of diagnostic work-up unless there are specific indications. Patients with bradycardia (eg <60bpm) should usually have an ECG before starting a cholinesterase inhibitor (II).



Last Updated: Thursday 08 October 2009


  • Acknowledgements

    The EuroCoDe project received financial support from the European Commission. Neither the European Commission nor any person acting on its behalf is responsible for any use that might be made of the following information. Alzheimer Europe also gratefully acknowledges the support it received from Fondation Médéric Alzheimer for this project.
  • European Union
  • Fondation Médéric Alzheimer