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. To minimize inclusion of frontotemporal syndromes with ambiguous 
cause only cases with ICD-10 codes according to national consensus were used. More important, analysis of data from only those cases diagnosed in specialist centres and that included both cognitive testing and neuroimaging demonstrated identical age distribution compared to the whole FTD cohort. Neuroimaging greatly increases the specificity of a clinical diagnosis of FTD and pathological findings on structural or functional imaging is a requirement for a diagnosis of probable FTD in the international consensus criteria published in 2011. Furthermore, lumbar puncture for analysis of AD biomarkers was performed in a high proportion of cases which increases detection of cases with underlying AD pathology. 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, which might lead to bias against the diagnosis of FTD in elderly patients with behavioural symptoms. Second, the behavioural symptoms of FTD might be more disruptive and noticeable in occupational and family se ings, thereby a racting more clinical a ention in earlyonset cases. Third, many memory clinics have a focus on earlyonset cases which leads to referral bias. Fourth, as the incidence of AD increases very sharply with age and the ratio between cases of AD and FTD is much lower in early-onset dementia, there could be a greater recognition of FTD cases in younger age cohorts. Finally, there is accumulating Tubeimoside-I evidence that the clinical and pathological features of FTD in the elderly differs from that of early-onset FTD, with memory problems and hippocampal sclerosis being more common, and frontal lobar atrophy less pronounced, in older patients. In support of this, the cases that failed to meet the new international consensus clinical criteria in a validation study were significantly older than the patients that fulfilled the criteria. Taken together, symptoms of frontal lobe dysfunction in the elderly might often be a ributed to other causes than FTD, such as VaD or AD. Prospective cohort studies, including neuropathological confirmation of the diagnosis, will be needed to confirm the findings in this study. In summary, data from SveDem suggest that increasing age is an important risk factor in FTD, as for other Procyanidin-B1 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.