Statistics from Altmetric.com
In the past decade, interest and concern for younger people with dementia have dramatically increased. Following sustained campaigning and support from the Alzheimer’s Society1,2 and other voluntary organisations3 in the UK, the needs of younger people with dementia and those who care for them have been increasingly addressed. The Royal College of Psychiatrists has recently issued guidelines recommending that every area in the country has a named old age psychiatrist responsible for younger people with dementia.4 This has been matched by the development of specialised support services, day care, and care homes for this group of patients.
However, delivering effective services relies on accurate epidemiological data to drive the planning and implementation process. In the past, estimates of the prevalence of dementia in younger people have relied on historical data5–8 drawn from hospital inpatient data and diagnostic registers. The inherent problem of studies based upon known, diagnosed cases is the potential for underestimating the real prevalence, particularly given the rarity and relatively small numbers involved. Unfortunately, because of the relatively low prevalence, conventional “door knocking” would be too costly and time consuming.
The aim of our study was to use a large, geographical catchment area and to use a combination of techniques to identify every known case of dementia starting before the age of 65 years. Having identified cases, we were also interested in exploring the prevalence of the different causes of dementia, and using the results obtained to extrapolate estimates of the numbers of younger people with dementia in the UK.
Ethical approval was obtained from the five local research ethics committees covering the catchment areas of the London Borough of Hillingdon (110 km2, population 232 000, of whom 81 184 were between 30 and 64 years of age; Jarman Index 8)9 and the London boroughs of Kensington, Chelsea, and Westminster (25 km2, population 335 500 of whom 112 309 were between 30 and 64 years of age; Jarman Index 21).
The aim of the first stage of sampling was to raise awareness of the project, and the need to identify and report any case, or suspected case of dementia, where the illness had started before the age of 65 years.
Personal letters and information were sent out to every general practitioner, psychiatrist, old age psychiatrist, neurologist, geriatrician, and general physician. In addition, information was also provided to departments of social services, voluntary organisations, day centres, home care, and community care teams through senior managers in the relevant organisations. The project ran in parallel with social services/voluntary sector initiatives in both areas,3,10 providing senior organisational leadership across health and social services and ensuring widespread engagement with the project. Clinicians and practitioners were asked to notify any cases under their care on a defined “census” day for the study.
In parallel, hospital information systems and case registers at the four main hospitals in the catchment areas were searched for cases falling in the target age range who had relevant ICD-9/10 diagnoses.
Personal contact was made with key neurologists and psychiatrists working in the area who allowed hand searching of copies of clinic letters, discharge summaries, and departmental databases.
An experienced psychiatrist (RJH) followed up notification of cases, requesting copies and/or access to all available hospital and GP case notes. The available information included clinical history, and investigations (including neuroimaging) were reviewed.
A hierarchical diagnostic algorithm was then applied. At the top level, subjects achieved a definite diagnosis if the dementia was associated with a known disease-specific genetic mutation, if there were neuropathological results available from a cerebral biopsy, or from autopsy if the patient had died after the “census” date.
At the second level of the algorithm, well recognised and validated clinical diagnostic criteria were applied to make the diagnoses of Alzheimer’s disease (NINCDS/ADRDA criteria)11, vascular dementia (NINDS/AIREN criteria)12, dementia with Lewy bodies,13 frontotemporal dementia (Lund and Manchester Criteria)14, alcohol related dementia (DSM-IV criteria for alcohol induced persisting dementia)15, or other dementias for which DSM-IV criteria are available. These criteria were not applied hierarchically, but rather on the basis of best clinical opinion.
At the third level of the algorithm, patients fulfilling the DSM-IV criteria for dementia, but not fulfilling criteria for one of the above diagnostic categories were assigned to a Dementia Not Otherwise Specified category.
The age at onset of disease was defined as the age of the patient at which the earliest conclusive dementia symptom was noted or documented in the medical notes or other correspondence.
For half of the patients identified in stage 2, a direct approach was made through the general practitioner seeking consent for a detailed clinical assessment to confirm diagnosis. Those patients and carers willing to participate were visited at home by RJH, who completed a semi-structured interview to collect demographic details and a detailed clinical history. The clinical assessment included a medical and neurological examination and completion of the Modified Hachinski Ischaemia Scale,16 Clinical Dementia Rating Scale ,17 the Behavioural Pathology in Alzheimer’s Disease Rating Scale,18 the Interview to Determine Deterioration in Daily Functioning,19 and the Cambridge Mental Disorders in the Elderly Examination Cognitive Assessment.20 The data collected were used to confirm the diagnosis.
Any diagnoses that were not clear at either stage 2 or stage 3 were discussed and agreed by consensus between RJH and MNR.
In total, 227 people with suspected young onset dementia were identified to the study. The sources of identification were psychologists (21%), databases and case registers (20%), neurologists (17%), psychiatrists (14%), social workers (9%), GPs (6%), physicians (6%), community nurses (4%), and other (3%).
Of the 227 referrals, 19 were excluded as not fulfilling DSM-IV criteria for dementia (six were “worried well”, seven had traumatic brain injuries, two had chronic schizophrenia, and four cases identified from hospital information systems were clearly miscoded when the notes were reviewed).
Of the remaining 208 cases, review of the notes showed an age of onset greater than 65 years of age in 23 cases who were further excluded from the study. The remaining 185 cases formed the population for the prevalence study, and of these, 87 cases (47%) underwent a detailed stage 3 personal assessment.
In the stage 3 assessment group, 69 patients (83%) had undergone neuroimaging with either CT or MRI, with 18 (21%) having undergone both CT and MRI. Twenty one patients (24%) had had an EEG.
The mean age of the population was 58.7 years (95% CI 57.4 to 60.1 years). Fifty eight percent (95% CI 50.9 to 65.6%) were male and 42% (95% CI 34.4 to 49.1%) were female, suggesting a trend towards a higher prevalence of dementia in males than females.
Table 1 summarises the age and gender specific prevalence rates for dementia in the study population on the “census” day; the pattern of prevalence approximately doubles with each five year increase in age.
Pragmatically, mixed diagnoses were avoided and we aimed to make a “highest confidence” single diagnosis in each case. In 168 (90%) of cases a single diagnosis was established using the diagnostic algorithm, 19 cases (10%) fitted more than one set of criteria and required a clinical consensus diagnosis to be made. Differential diagnoses in the population included Alzheimer’s disease (62 cases, 34%), vascular dementia (34 cases, 18%), frontotemporal dementia (23 cases, 12%), alcohol related dementia (19 cases, 10%), and dementia with Lewy bodies (12 cases, 7%). Other causes of dementia accounted for 35 cases (19%) and included Huntington’s disease (nine cases), dementia in multiple sclerosis (eight cases), corticobasal degeneration (two cases), prion disease (two cases), dementia in Down’s syndrome (three cases), dementia in Parkinson’s disease (two cases), and dementia not otherwise specified (nine cases). The very low prevalence of dementia in Down’s syndrome is likely to be an artefact due to sampling bias. Despite raising awareness of the project with learning disability services, no referrals of cases were made from these sources.
The age specific prevalence rates for the four most common causes of dementia in this group of younger people are summarised in table 2.
Estimates of numbers of cases in the UK
One of the objectives of the study was to generate robust estimates of the numbers of people in the UK with young onset dementia. These figures are important both for service planning and for campaigning for improvements in recognition and provision.
Using data from the 2001 UK Census, we applied the prevalence figures derived from our population to the population figures for the UK. Table 3 summarises the estimated numbers of younger people with dementia, and with specific diagnoses in the UK.
This is the first study to report the prevalence of dementia and differential diagnoses in a community based catchment area study of younger people with dementia. The prevalence of dementia in people between the ages of 30 and 64 years was 54 per 100 000 with the prevalence approximately doubling for each five year age band. Based on the prevalence figures derived from this study we estimate that there are 18 319 (15 296–21 758) people under the age of 65 years with dementia in the UK.
The main strength of the study is the combination of a large catchment area and the use of multiple and varied means of case identification, essentially attempting to leave no stone unturned. The very varied sources of case identification confirm the benefits of this methodology. The main weakness of the study is that despite using a catchment area with a population of more than 500 000 people, the rarity of the condition meant that only 185 cases were identified. For more robust estimates with smaller confidence intervals a much larger population would be helpful.
This study was unique in being large enough to consider a range of differential diagnoses with our disease specific figures and overall figures being supported by other studies of individual diseases.
Thus our prevalence figure for young onset Alzheimer’s disease of 35.0 per 100 000 in the 45–64 years age group is very similar to estimates for the Northern Region (34.6),6 Finland (32.7),21 the Framingham Study (31.8),22 and Cambridge (15.1).8
Similarly, the prevalence rates we found for vascular dementia were close to the estimates derived from an earlier study in Scotland,5 and the prevalence rate we identified for frontotemporal dementia (15.4 per 100 000) was almost identical to the rate identified in the Cambridge study (15.1).8
While none of these studies used an identical methodology, this study is probably most similar to the study from the Northern Region which used case ascertainment from hospital inpatient records, and the Cambridge study which identified cases from memory clinic databases and admission records. The Framingham study and the Finnish study both used ascertainment from large population diagnostic registers, but generate similar prevalence figures. One conclusion from this is that younger people with dementia inevitably come to medical attention and receive a diagnosis—what is less certain is what follow up and support these patients are given after the diagnosis.
Our study underpins and supports initiatives from the Alzheimer’s Society and the Royal College of Psychiatrists to secure and improve services for this group of patients. Even though dementia in younger people is uncommon when compared with dementia in older people, every case is a personal tragedy for the individual and their family, which often includes young children.
The study also highlights the differences between dementia in younger people and dementia in older people. Alzheimer’s disease is considerably less common in this group, accounting for only 34% of cases, whereas in the elderly it would be expected to account for up to 80% of cases. Moreover, unusual diseases such as frontotemporal dementia accounts for 12% of cases, although there is increasing evidence that it is a commonly missed diagnosis in older people.23
The prevalence data generated by this study can be applied to local population figures to derive reliable estimates of the potential numbers of cases in any area. This should support better planning and delivery of services.
Project funded by NHS Executive (North Thames) Responsive R&D Grant, RFG045. Richard Harvey is funded by the Alzheimer’s Society.
Funding was awarded in open competition and following peer review. The funder had no role in study design, data collection, analysis, interpretation of data, or in the writing of the report. The funder encouraged us to submit the paper for publication as part of good research governance.
Competing interests: none declared
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.