The authors have declared that no competing interests exist.
Conceived and designed the experiments: CN. Performed the experiments: CN. Analyzed the data: CN MLW KN AS SV. Wrote the paper: CN. Critical review of the manuscript: CN MLW KN AS SV.
Frontotemporal dementia (FTD) is considered to be a mainly early-onset neurodegenerative disorder with a strong hereditary component. The aim of the study was to investigate age-related incidence and family history in FTD compared to other dementia disorders, especially Alzheimer's disease (AD).
The Swedish Dementia Registry (SveDem) registers all new cases of dementia diagnosed by the participating centres, including data on demographics, diagnosis, and investigations used. Data for the period 2008–2011 were extracted and compared with age-related population data on a regional and national level.
There were 20 305 patients registered in SveDem during 2008–2011, whereof 352 received a diagnosis of FTD. Mean age at diagnosis for FTD was 69.6 years and almost 70% of FTD cases were 65 years or older at the time of diagnosis. Both FTD and AD showed an increased incidence with age, which reached a maximum in the age group 80–84 years at 6.04 and 202 cases per 100 000 person-years, respectively. The proportion of cases with a positive family history was significantly lower in FTD than in AD.
Contrary to general opinion within the field, data from SveDem show that the incidence of FTD increases with age, and that the majority of cases are diagnosed after the age of 65 years. In addition, data from SveDem might suggest that the importance of hereditary factors in general is similar in FTD and AD. The recognition of these findings has important consequences for the diagnosis, treatment and care of patients with FTD.
The frontotemporal dementias (FTD) are a group of neurodegenerative disorders affecting primarily the frontal and temporal lobes of the brain, leading to various combinations of behavioural, cognitive and motor symptoms
Although the disorder was first described by Pick more than 100 years ago, formal clinical and neuropathological criteria were not developed until the mid-90s
Many studies have reported that a high proportion of FTD cases have a positive family history for dementia
The aim of the present study was to test the hypothesis that the incidence of FTD increases with age and is higher in the elderly, similarly to other neurodegenerative disorders. Furthermore, the proportion of FTD cases with a positive family history was compared to family history in AD. For this purpose we used data extracted from the Swedish Dementia Registry (SveDem) for the period 2008–2011, encompassing 20 305 new cases diagnosed with dementia, including 352 cases of FTD. The present study supports the hypothesis of increasing incidence of FTD with age, similar to other major neurodegenerative disorders, and that late-onset FTD is more common than early-onset FTD.
The Swedish Dementia Registry (SveDem;
Diagnosis is registered in SveDem by choosing from a list of diagnoses (early-onset AD, late-onset AD, vascular dementia (VaD), Mixed AD and VaD, FTD, Dementia with Lewy bodies (DLB), Parkinson disease dementia (PDD), Dementia of other causes and Dementia not specified). In addition, the primary and secondary ICD-10 codes are registered. In Sweden, the ICD-10 codes used for the diagnosis of different forms of dementia are made according to a consensus document
It is important here to explain the nature of the Swedish health care quality registries, such as SveDem. Data are entered in a standardized fashion to enable follow up of diagnostic procedures, treatment and management, with the purpose of standardizing and improving health care nationally. Extracted data for scientific purposes are anonymised and cannot be combined with the individual patient's medical records. Adherence to the former
The presence or absence of a family history of dementia in first- and second-degree relatives is noted in the SveDem registration form. For classification purposes, the responses were divided into three categories: Positive family history (Yes), negative family history (No/None known) and Not known (reserved for when questions about family history have not been asked alternatively no response registered).
Incidence rates were calculated by dividing the number of diagnosed cases with the number of person-years in the population at risk, multiplied by 100 000. This was done for each age cohort as well as for the whole population in defined geographical regions. Data on population size in each age group and in each local council was obtained from publicly available census figures for 2011 (Statistics Sweden;
Differences in gender distribution for each diagnosis, as well as the distribution between early-onset and late-onset cases of FTD, were calculated by an exact binomial test. T-test was used for statistical comparisons of age at diagnosis in AD and FTD. For comparison of the proportion of a positive family history in first and second-degree relatives between cases with AD and FTD we used Fisher's exact test. In cases with FTD, the proportion of cases with a positive family history in different age groups was compared using Fisher's exact test and Mantel-Haenszel chi-square.
The statistical analyses were performed in R version 2.15.2 (The R Foundation for Statistical Computing,
The project was approved by the Regional Ethics Committee of Lund University (Permit number 2012/137) and the SveDem Steering Committee. All registered individuals were informed about SveDem orally and in writing at the time of diagnosis and gave their consent to registration, with the possibility of declining participation, in accordance with Swedish legislation.
In total, 20 305 cases of dementia were registered in SveDem during the period 2008–2011. The number of cases and proportion of the different diagnoses registered are summarized in
Diagnosis | Cases (%) | Age (range) | F/M (p-value) |
AD | 6766 (33.3) | 77.4 (27–99) | 65.8/34.2 (p<0.0001) |
AD+VaD | 4064 (20.0) | 80.9 (52–100) | 59.3/40.7 (p<0.0001) |
VaD | 3652 (18.0) | 79.8 (33–103) | 50.5/49.5 (p = 0.53) |
FTD | 352 (1.7) | 69.6 (39–96) | 55.4/44.6 (p = 0.049) |
DLB | 477 (2.3) | 76.8 (53–94) | 38.2/61.8 (p<0.0001) |
PDD | 316 (1.6) | 75.0 (49–94) | 38.6/61.4 (p<0.0001) |
Other | 500 (2.5) | 74.4 (30–95) | 52.5/47.5 (p = 0.30) |
Dementia NS | 4178 (20.6) | 80.7 (40–101) | 63.2/36.8 (p<0.0001) |
Total | 20305 (100) | 79.0 (27–103) | 59.6/40.4 (p<0.0001) |
Data are n (%) for diagnoses, mean age (range) and the female (F)/male (M) distribution percentage (p-value). Statistical differences in gender distribution was calculated by a binomial test, see Methods. AD = Alzheimer's disease, VaD = Vascular dementia, FTD = Frontotemporal dementia, DLB = Dementia with Lewy bodies, PDD = Parkinson disease dementia, Other = Dementia of other causes, Dementia NS = Dementia not specified.
There were 352 cases of FTD registered during the investigated time period. Diagnoses of FTD were made almost exclusively in specialist clinics (95%). The majority of centres that reported cases of FTD registered less than 6 cases (range 1–43). More than one third (37%) of the cases were registered in only 5 centres. While FTD accounted for 1.7% of all cases overall, the proportion was greater in centres with a higher number of FTD cases. For example, in our own centre (Lund) the proportion of FTD was 4.2%.
In Sweden, certain basic diagnostic procedures are considered mandatory for a diagnosis of dementia and have to be performed before referral to a specialist centre. These basic procedures include history from patient and caregiver, physical examination, blood chemistry, cognitive screening tests and computer tomography (CT) of the brain. Among all recorded FTD cases, 95.7% had performed cognitive testing of some kind (MMSE, other screening tests, and/or examination by a neuropsychologist). The main reason given for not performing cognitive tests was that it was not possible due to the patient's condition. Structural neuroimaging (CT or MRI) was performed in 96.3% and lumbar puncture in 64.4% of cases with FTD.
For calculation of the total incidence of FTD, we chose to compare the smaller regions of Lund and Uppsala, with populations of 289 550 and 335 000 inhabitants, respectively, with the larger Stockholm region with 2 091 357 inhabitants. While the Lund and Uppsala regions each are dominated by a specialist centre located at the respective university hospital, both with research on FTD, there are 10 non-university specialist centres in the Stockholm region in parallel to the memory clinic at Karolinska University Hospital. The total annual incidence of FTD in these three regions was 1.17 (Stockholm), 2.07 (Lund) and 3.21 (Uppsala) per 100 000 inhabitants.
The distribution of age at diagnosis in FTD cases is summarized in
(A) Age at diagnosis in FTD cases registered in SveDem 2008–2011. (B) The proportion of FTD cases in each age group in the total FTD cohort (blue bars), compared to those cases diagnosed in specialist centres where investigations included cognitive testing and structural or functional imaging (red bars) and cases that had performed lumbar puncture with CSF analysis (green bars).
Symptoms of frontal lobe dysfunction occur both in VaD and AD. To minimize the risk for incorrect diagnosis of FTD, especially in older age groups, we performed further analysis of the age distribution after excluding cases that did not fulfil the following criteria: (1) diagnosis at a specialized centre, (2) use of structural (CT/MRI) or functional (SPECT/PET) neuroimaging, (3) use of cognitive testing (MMSE, other screening tests, and/or examination by a neuropsychologist). Out of the 352 FTD cases, 314 cases fulfilled the above criteria. In addition, out of these 314 cases, the 206 cases that had performed lumbar puncture with CSF analysis of AD biomarkers were analysed separately. The age distribution in these three groups was very similar except for a slightly lower proportion of cases in the age group 80–84 years and a higher proportion of cases in the age group 65–69 years among cases that had performed lumbar puncture (
Considering that the younger age cohorts are larger, the age-related incidence of FTD was calculated and compared to that of patients with AD (cases of mixed dementia were excluded). Both for FTD and AD, the highest incidence occurred in the age group 80–84 years at 6.04 and 202 cases per 100 000 person-years, respectively (
The figures show the incidence of FTD (A; blue bars) and AD (B; red bars) in relation to age (years). Values for incidence are given as cases per 100 000 person-years in each 5 year age cohort.
The number of FTD cases with a family history of dementia was recorded and compared to AD cases (
The data shown in (A) represents the response to questions on the presence of dementia in first-degree and second-degree relatives of patients diagnosed with AD (blue bars) and FTD (red bars). In (B), the responses in FTD cases are divided into age cohorts: 39–64, 65–80, and 81–96 years. For classification purposes, the responses were divided into three categories: Positive family history (Yes), negative family history (No/None known) and Not known (reserved for when questions about family history have not been asked alternatively no response registered).
In the largest study of age-related incidence to date, using a large community-based registry, we demonstrate that the incidence of clinically diagnosed FTD increases with age as has been previously reported in smaller populations
While it is well documented that AD affects more women than men, and the reverse is true for DLB/PDD, there is conflicting data on the gender ratio in FTD
The presence of a first-degree relative with dementia is common in both AD and FTD, suggesting that genetic risk factors are important in both these disorders
The results of the present study confirm a high degree of positive family history in first-degree relatives of cases with AD and FTD. Cases with second-degree relatives were uncommon and autosomal dominant mutations have more recently been estimated to account for only around 10% of cases with AD and FTD
The presence of a positive family history in FTD might, however, have been underestimated, especially in second-degree relatives. The questions used in SveDem do not specifically address the symptoms specific for FTD, and a significant proportion (15–25%) of responses were either “Not known” or data was missing. On the other hand, previous estimates of a high prevalence of positive family history in FTD might have been influenced by referral bias and research interest in familial cases as well as the high prevalence of dementia in the general population
The main limitation of the present study is that the clinical data leading to diagnosis is not available. Consequently, it is not possible to control the validity of the diagnosis in each individual case. Furthermore, neuropathological confirmation of the diagnoses was not available. However, almost all diagnoses of FTD in the SveDem cohort were made at specialist centres. High clinicopathological concordance in early-onset dementia, with up to 97% specificity for bvFTD, has recently been demonstrated in a highly specialized centre
The number of FTD diagnoses differed greatly between participating centres and interest and experience in FTD appears to be of greater importance for the diagnosis of FTD than the diagnostic procedures used. It is thus likely that FTD is still underdiagnosed in Sweden and this might be even more pronounced in the elderly. This is also supported by the relatively low incidence of FTD seen compared to previously published estimates
There are several possible reasons why FTD might be underdiagnosed in the elderly: First, the 1998 diagnostic criteria lists onset before 65 years of age as one of the supportive criteria and late-onset cases are stated as being rare
In summary, data from SveDem suggest that increasing age is an important risk factor in FTD, as for other neurodegenerative disorders. The increased recognition of FTD in the elderly has important consequences for dementia care. Compared to AD, patients with FTD often require other strategies for psychosocial support and nursing
We would like to thank Helene Jacobsson, R&D Centre Skåne, for statistical support and The Swedish Dementia Registry, SveDem (