Survival in Huntington’s disease and other young‐onset dementias

Abstract Objectives To compare survival and risk factors associated with mortality in common young‐onset dementias (YOD) including Huntington's disease. Methods This retrospective cohort study included inpatients from an Australian specialist neuropsychiatry service, over 20 years. Dementia diagnoses were based on consensus criteria and Huntington's disease (HD) was confirmed genetically. Mortality and cause of death were determined using linkage to the Australian Institute of Health and Welfare National Death Index. Results There were 386 individuals with YOD included. The dementia types included frontotemporal dementia (FTD) (24.5%), HD (21.2%) and Alzheimer's disease (AD) (20.5%). 63% (n = 243) individuals had died. The longest median survival was for those who had HD, 18.8 years from symptom onset and with a reduced mortality risk compared to AD and FTD (hazard ratio 0.5). Overall, people with YOD had significantly increased mortality, of 5–8 times, compared to the general population. Females with a YOD had higher standardised mortality ratio compared to males (9.3 vs. 4.9) overall. The most frequent cause of death in those with HD was reported as HD, with other causes of death in the other YOD‐subtypes related to dementia and mental/behavioural disorders. Discussion This is the first Australian study to investigate survival and risk factors of mortality in people with YOD. YOD has a significant risk of death compared to the general population. Our findings provide useful clinical information for people affected by YOD as well as future planning and service provision.


| INTRODUCTION
Young-onset dementia (YOD) refers to a dementia where symptom onset occurs at less than 65 years of age. 1 A recent Delphi study reported high consensus that the most common causes of dementia in younger people include Alzheimer's disease (AD), frontotemporal dementia (FTD), vascular dementia (VaD) and Huntington's disease (HD). 2 While AD, FTD and VaD have both sporadic and genetic causes, HD is relatively rare in comparison, with a reported prevalence of 5.7 per 100,000, 3 and is an autosomal dominant neurodegenerative condition caused by a trinucleotide CAG repeat expansion on chromosome 4, causing abnormal accumulation of the huntingtin protein (HTT). 4 The characteristic triad of symptoms are neurological, cognitive and psychiatric, with cognitive and psychiatric symptoms preceding chorea by up to 15 years. 5 Individuals with 36-39 CAG repeats will be at increased risk of developing symptoms related to HD and those with ≥40 CAG repeats will have 100% complete penetrance and develop HD. 5 A higher number of CAG repeats is associated with development of symptoms at a younger age, faster progression and earlier age of death. 6 Dementia is a leading cause of death in older people and 7 investigating survival in dementia has important implications for societal costs and disease burden, and in YOD, is essential for future planning and disease progression due to its onset in midlife. 8,9 There has not been any previous work comparing survival in these major types of YOD in Australia. Survival in non-genetic causes of YOD has been reported as 22 years for VaD, 16.4 years for FTD and 15.6 years for AD, from symptom onset. 10 Shorter survival duration was reported in a larger Dutch cohort, with 7 years survival from age at diagnosis for FTD and AD. 11 Survival in HD has ranged from 14.5 12 to 35 years from symptom onset. 13 Our previous work in this area found that the mean survival time in YOD overall was 12.7 years, double that of older-onset dementia (6.3 years) 14 but excluded individuals with HD. In this study, the main objective was to determine survival of individuals with HD compared with other YODs, in a wellcharacterised cohort seen over a 20-year period. Secondary objectives were to identify factors associated with survival, to compare mortality rates to Australian population norms, and to investigate causes of death.

| Participants and study design
This retrospective study was performed at Neuropsychiatry, Royal Melbourne Hospital, Australia. 'Neuropsychiatry' is a tertiary specialist centre providing assessment, diagnosis and follow-up of people with a range of neuropsychiatric conditions, including YOD.
Neuropsychiatry also has a HD-specific service including a predictive testing clinic for those at risk of HD. 15 Inpatients are admitted for multidisciplinary assessment (neuropsychiatry, neurology, allied health and nursing) and investigations (neuroimaging, lumbar puncture and blood tests). The detailed methods have been previously described. 14 We reviewed all inpatient summaries for the period 1992-2014 inclusive and identified patients diagnosed with any dementia diagnosis during this period. Individuals were included if they had symptoms determined to be related their dementia at less than 65 years old. Data collection included: � Demographics; � Dementia diagnoses: Alzheimer's disease (AD), 16 Table 2

| Predictors of mortality in YOD
The Cox regression model (Table 4)  CVRFs nor alcohol were significantly associated as predictors of mortality.

| Cause of death
The underlying/primary cause of death (Figure 2)  Abbreviation: SMR, standardised mortality rates. This is the first study to examine mortality in the commonest causes of YOD in Australia, using the NDI to obtain complete, reliable information and minimise attrition. 11,20 We found that HD had longest survival (18.8 years) compared to the other dementia subtypes. All YOD subtypes were associated with increased mortality (5-8 times greater than the population norms). Our study identified three additional findings. First, having HD compared to AD and FTD conferred lower mortality risk. Second, presenting with cognitive symptoms as part of a dementia was a risk factor for mortality, compared to presenting with psychiatric symptoms, which we had previously found. 14 Finally, females with YOD had higher SMR compared to males. Our findings emphasise the need for early identification of dementia and provide clinicians with evidence-based information for the important questions frequently asked by people with YOD and their families.
Our study reported that those with HD had the longest survival, which has been reported elsewhere. 12 (C9orf72) mutation. 23 Comparing genetic and sporadic forms of FTD and survival suggests that having a pathogenic mutation confers increased mortality 24 but there is variation within these mutations.
The microtubule-associated-protein tau (MAPT) mutation appears to be associated with shorter survival compared to the progranulin and C9orf72 mutation. 25 The C9orf72 mutation is a cause of FTD, MND and FTD-MND 23 and while there is consistency that individuals with FTD-MND phenotype have a short survival, 2-3 years, 26 there are increasing reports that there are others who have a range of presentations 27 and outcomes, with some people living for a very long time. 28 For genetic forms of AD, there is variation in symptom onset, with individuals who had a presenilin 1 (PSEN1) mutation having onset on average 7.1 years earlier than those with amyloid precursor protein (APP) mutation. However, their median survival was similar for both pathogenic mutations, 11.4 and 12.5 years, respectively. 29 Atypical presentation, that is, non-amnestic symptom onset was associated with longer median survival in both mutations, despite similar age of onset, particularly for those with PSEN1 mutation. 29 Similarly, we found that those with a non-cognitive presentation in our cohort had longer median survival. It could be that as a psychiatric service, we are inherently biased towards psychiatric symptoms, which may place undue emphasis and interpretation of psychiatric and behaviour changes as part of a dementia syndrome. We were unable to analyse the mortality risks for the different clinical presentations within the major dementia subtypes due to small numbers. Causes of death in HD were unsurprisingly attributed to HD and respiratory disorders. 13,32 Suicide is also a common cause of death in HD, 13,32 though we only had one individual with HD who had this as their reported cause of death. The first 3 months of a diagnosis of dementia in someone aged less than 65 years is a particularly vulnerable period, with reports that there is a 3� risk of suicide, 33 highlighting the need for close follow-up post-diagnosis. Amongst our cohort, there were only nine individuals with 'external' or 'unknown' causes of death, which might be attributed to suicide.

| Limitations
Diagnostic investigations and diagnostic criteria evolved over the study period. The study used the contemporaneous diagnosis at the time of initial assessment rather than re-classifying diagnoses based on current criteria. We did not have pathological confirmation of the dementia diagnoses and while younger people with dementia are known to have diagnostic change, due to follow-up of our patients, we opt to minimise this change by using the most 'recent' dementia diagnosis. We were able to genetically diagnose HD. Retrospective studies are dependent on the information contained within file review, dependent on collateral history for symptom onset type and F I G U R E 2 Underlying causes of death for young-onset dementia subtypes. age of onset, and thus subject to interpretation. Finally, this study was undertaken in a single inpatient setting, where the sample may be more complex or severe, restricting the generalisability of the findings. The sample size of 386 is relatively small, but is the largest Australian study to date. This limitation is potentially a strength, given that we have been able to access a relatively rare group of patients, including with HD, over a long period of time and describe the largest number of YOD patients linked to the Australian NDI.

| CONCLUSIONS
In summary, we studied survival, predictors of mortality and causes of death, in a large cohort of individuals with common types of YOD. We provide survival information which will be of value to patients, families and clinicians for future planning and service provision, and highlight the need for early identification. Supporting these individuals to maintain quality of life during their disease is very important.