Introduction
Dementia is three to five times more prevalent in Aboriginal people than in other Australians (Li et al., Reference Li, Guthridge, Aratchige, Lowe, Wang, Zhao and Krause2014; Pollitt, Reference Pollitt1997; Radford et al., Reference Radford, Mack, Draper, Chalkley, Daylight, Cumming and Broe2015; Smith et al., Reference Smith, Flicker, Lautenschlager, Almeida, Atkinson, Dwyer and LoGiudice2008). To develop effective prevention programmes, it is important to understand the reasons for this difference.
In this paper, the term ‘Aboriginal’ is used respectfully as the preferred term used by Central Australian Aboriginal people. The terms ‘dementia ‘and ‘Major Neurocognitive Disorder’ (Major NCD) are used synonymously, to describe the syndrome of decline in one or more cognitive domains and described in the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) (American Psychiatric Association, 2013). This is consistent with Dementia Australia recommendations (Dementia Australia, 2018). NCD may be divided into either Mild or Major categories, depending largely on a person’s ability to function independently. Since all participants in this study are dependent upon their Residential Aged Care Facility (RACF) for their activities of daily living, all those found to have an NCD were categorised as having ‘Major NCD’.
Dementia diagnosis in Aboriginal people has been greatly helped by the development of the Kimberley Indigenous Cognitive Assessment (KICA), a validated tool for use with Aboriginal people in Western Australia, the Northern Territory (NT) and Queensland (LoGiudice et al., Reference LoGiudice, Smith, Thomas, Lautenschlager, Almeida, Atkinson and Flicker2006; Smith et al., Reference Smith, Flicker, Dwyer, Marsh, Mahajani, Almeida and LoGiudice2009). Unfortunately, the identification of causes of dementia in Aboriginal people remains. poorly understood, with Smith et al. (Reference Smith, Flicker, Lautenschlager, Almeida, Atkinson, Dwyer and LoGiudice2008) and Li et al. (Reference Li, Guthridge, Aratchige, Lowe, Wang, Zhao and Krause2014) citing frequencies of above 50% for dementia of ‘unspecified’ cause. This study seeks to identify the prevalence of factors that may be associated with the categories of dementia found in Aboriginal people living in RACFs in Alice Springs, the major town in Central Australia in the NT
Ethics approval was granted by the Human Research Ethics Committee for NT Department of Health and Menzies School of Health Research (HREC 2015-2462 on 6/1/2016) and the Central Australian Human Research Ethics Committee (HREC-15-336 on 7/12/2015).
Methods
Participants and setting
Eighty-four potential Aboriginal participants aged 45 years or greater were identified in the three RACFS in Alice Springs. Consent was obtained from the individual (n=40) or legal guardian (n=35). Nine residents with no legal guardian were excluded due to lack of consent (Figure 1). Ethnicity was determined from RACF records and confirmed with participants and/or carers prior to recruitment.
Participant flow through the study.
Data collection
The research team included an experienced NT general practitioner; geriatrician; senior nurse; four senior researchers with psychiatric, aged care and physician expertise; two allied health academics and representatives from Australian Regional and Remote Community Services who manage the residential facilities and Dementia Australia.
Interpreters for six different Central Australian languages worked with guidelines adapted from the Aboriginal Interpreter Service. Team members provided training to interpreters in the appropriate use of the KICA. Culturally adapted consent forms, study information sheets and a pictorial flip chart were developed using pictures, plain English and minimal text. Screening was performed by the general practitioner and hospital file audits were conducted by the geriatrician.
Cognitive screening and assessment
Of the 84 participants assessed for eligibility, 9 were excluded due to lack of informed consent; 28 who were unable to be assessed and who had a diagnosis of dementia or a history suggestive of cognitive impairment from dementia, proceeded straight to file audit with guardian or proxy consent or a waiver of consent, leaving 47 people screened (Figure 1). Of the participants screened, 30 proceeded to file audit with 58 participants audited in total. The KICA screen (n=47) confirmed eligibility for the study, while a follow-up KICA-cog (n=25) was used to confirm a diagnosis of cognitive impairment (Figure 1). Two residents who were unavailable for KICA-screen testing were later tested using only the KICA-cog. Those scoring above the cut-off [21/22 for KICA-screen and 33/34 for KICA-cog (LoGiudice et al., Reference LoGiudice, Strivens, Smith, Stevenson, Atkinson, Dwyer and Flicker2011)] were excluded from the study. If an initial assessment was incomplete, data were included for those who had attempted 80% or more of the items.
Interpreters were used in 32 (68%) of the KICA-screens and 20 (87%) of the KICA-cog tests. All interviews were held in participants’ bedrooms or in quiet places within the RACFs. KICA-screens were completed in January and March and KICA-cogs in April and June 2016. In recognition of the potential for practice effects, it is noted that no participants were found to have an improved KICA-cog score following a KICA-screen and all 23 participants assessed with the KICA-cog had an abnormal result.
File audit
The electronic hospital records of the 58 participants were reviewed for clinical assessments of cognitive capacity and diagnoses; brain imaging and reports; pathology reports, and prescriptions of relevant medications. CT images, or reports of images, were available for almost all (93%) of the participants in this study. The median age of the most recent imaging was 2 years. RACF files were also reviewed for current medication and diagnosis.
DSM-5 criteria for sub-types of Major NCD guided each diagnosis, using KICA-cog and screen results and analysis of file audit data. Many participants had extensive case notes over many years, allowing a timeline of cognitive deterioration to be obtained. When this information was not available, an assumption was made that participants had functioned at a higher level prior to admission to the nursing home review, unless there was a known alternative diagnosis such as developmental delay.
A likely causal link between key events was assumed where other evidence was lacking. For example, a participant with a history of head injury and extensive encephalomalacia and previous craniotomy visible on CT scan was diagnosed with Major NCD due to traumatic brain injury (TBI). The assumption was made that the impairment of cognition occurred following the injury. While TBI is also a risk factor for Alzheimer’s Disease, we classified people as having NCD due to TBI only when file audit and imaging supported brain injury itself as the primary cause of cognitive loss.
Clinical examination of participants was not conducted because of difficulties with obtaining assent, and the possibility that examination would cause discomfort and distress in this highly vulnerable population. Most records contained detail of multiple previous assessments, and the geriatrician in the study had assessed almost 50 percent of participants previously. Relevant information about participants was also sought from a senior RACF clinician (GP, nurse) or carer.
Results
Diagnoses
Fifty of the 58 participants were diagnosed as having a Major NCD. Thirty-nine (67%) participant files had a diagnosis of ‘dementia’ already documented, 85% of which were of unspecified cause (Table 1).
Demographics of Residents Enrolled After Initial KICA Screening (N = 58)
* Diagnosis recorded in NT public hospital case notes and/or in aged care facility files.
Diagnoses were divided into five groupings: Major NCD due to vascular disease (VaD) (30%), Major NCD due to Alzheimer’s Disease (AD) (26%), Major NCD due to Brain Injury (BI) (20%), a fourth group (Major NCD-Other) containing the less frequent diagnoses and a fifth group of participants without Major NCD (Table 2).
Associations with Major Neurocognitive Disorders (Major NCDs)
VaD: Vascular disease; AD: Alzheimer’s Disease; BI: Brain Injury. T2DM: Type 2 Diabetes Mellitus.
Denominators in rows labelled ‘Thyroid Disease’, ‘Vit D’, ‘HTLV-1’ and ‘Pos Syphilis serology’ are the total numbers with test results documented for each group.
Positive Syphilis serology: all results reflect previously treated infection. No active disease documented.
The group of Major NCD due to BI included both the DSM-5 classification of Major NCD due to TBI and the group with Major NCD due to neurosurgery – a group that is not classified separately in DSM-5. This is because many participants had both TBI and subsequent or related surgery, either of which could have been the cause of cognitive impairment. In addition, both groups of participants had similar long-term concerns such as epilepsy. The Major NCD-other group contained participants with Major NCD due to mixed AD and VaD (10%); Major NCD of unspecified cause (4%); Major NCD due to substance misuse (4%) – both patients with Wernicke/Korsakoff syndrome; Major NCD due to another medical condition (4%) and one diagnosis of Major NCD due to Frontotemporal Dementia.
No participants were classified as Mild NCD. This assessment took into account the KICA results, clinical file review and the evidence that admission to the residential facility resulted from the need for significant support in activities of daily living.
Associations
The prevalence of a variety of associated factors were reviewed for each of the diagnostic groups (Table 2). Hypertension was recorded in 59% of participants, a prevalence that was spread relatively evenly across all diagnoses. Type 2 Diabetes Mellitus (T2DM) was diagnosed in 60% of participants, four of whom had previous hypoglycaemic episodes recorded, and 23 (74%) had a lowest HbA1c recording of less than 6.5%, suggesting the possibility of undiagnosed hypoglycaemic events. The highest prevalence of T2DM was found in those with Major NCD due to AD and Major NCD due to VaD.
Twelve participants (24%) diagnosed with Major NCD had hypothyroidism, either as a recorded diagnosis (9) or with a record of at least one high TSH level (> 5.5 mIU/L) and regular treatment with Thyroxine (3). Seven of these twelve (58%) had Major NCD due to AD.
Twenty-four of all participants (41%) had a history of alcohol misuse documented in their hospital files, most commonly those classified as Major NCD due to BI (70%).
Of those with documented seizures (26%), most were diagnosed as Major NCD due to BI (60%), while 71.5% had a documented history of alcohol misuse.
Records of 17 of the 18 people with documented tobacco use described them chewing rather than smoking tobacco; however, many of these participants were also observed smoking.
Low vitamin D levels (< 50 nmol/L) were recorded in 68% of the 25 residents with documented levels. Most were recorded prior to admission to the facility suggesting that vitamin D deficiency was a pre-existing condition.
Of the 19 diagnosed with Major NCD who were tested for HTLV-1, 13 (68%) were positive for the disease.
No active syphilis infection was found in any participants.
Discussion
Understanding the reasons for the high prevalence of dementia in Aboriginal people is a prerequisite to developing effective prevention programmes. While diagnosis was difficult at times, the availability of imaging and/or imaging results in this study led to greater confidence of diagnosis and improved ability to understand some of the underlying causes of dementia.
DSM-5 uses a syndromic approach to diagnosis and it can be difficult to fit pathological diagnoses to this framework. For example, the appropriate classification of impairment due to neurosurgery is unclear. Similarly, Wernicke/Korsakoff’s syndrome is classified as Major NCD due to substance abuse, whereas thiamine deficiency is the underlying cause.
Compared to similar studies, this study found higher rates of Major NCD due to VaD and TBI and lower rates of Major NCDs due to AD and unspecified causes. It is important to note, however, distinct differences in methodology and the study population limit direct comparison between studies. For example, the NSW community study by Radford et al. (Reference Radford, Mack, Draper, Chalkley, Daylight, Cumming and Broe2015) found higher rates of AD, which may perhaps be explained by the insidious onset and progress of AD leading to a higher number of cases in the community.
Studies of Aboriginal populations undertaken prior to the introduction of the DSM-5 have reported few diagnoses of dementia due to BI, despite finding high rates of head injury in their populations. High rates of head injury in Aboriginal people with dementia have been previously reported by Smith et al. (Reference Smith, Flicker, Lautenschlager, Almeida, Atkinson, Dwyer and LoGiudice2008), LoGiudice et al. (Reference LoGiudice, Smith, Fenner, Hyde, Atkinson, Skeaf and Flicker2016) and Radford et al. (Reference Radford, Mack, Draper, Chalkley, Daylight, Cumming and Broe2015), but only Radford et al. diagnosed dementia due to BIin a small proportion of their dementia group.
Twenty years ago, Pollitt suggested that alcohol misuse was the major cause of dementia in Aboriginal people (Pollitt, Reference Pollitt1997). Only two participants (4%) in this study were diagnosed with Major NCD due to substance misuse (in both cases, Wernicke/Korsakoff syndrome). Alcohol misuse, however, continues to be commonly associated with dementia (LoGiudice et al. Reference LoGiudice, Smith, Fenner, Hyde, Atkinson, Skeaf and Flicker2016; Radford et al. Reference Radford, Mack, Draper, Chalkley, Daylight, Cumming and Broe2015; Smith et al. Reference Smith, Flicker, Lautenschlager, Almeida, Atkinson, Dwyer and LoGiudice2010) and with physical trauma such as motor vehicle accidents and assaults (Morgan and McAtamney, Reference Morgan and McAtamney2009). In the clinical file audit, it was common to find patients who had been reported as heavy drinkers. This information was included in the audit and was assumed to be absent when it was not documented. Alcohol misuse was noted in 80% of those with a history of a head injury requiring hospital involvement and in 70% of those diagnosed with Major NCD due to BI.
Aboriginal people are reported to have a younger onset of dementia than non-Indigenous Australians (Li et al., Reference Li, Guthridge, Aratchige, Lowe, Wang, Zhao and Krause2014); however, this study did not confirm this finding. The difference is unlikely to be attributable only to the age of our study population, as RACFs may accommodate Aboriginal people from the age of 50 years. Of the six people under 60 years of age in this study, none had AD, three had histories of major head injuries, one a history of Wernicke/Korsakoff’s syndrome, one a history of congenital intellectual impairment and one had suffered a cerebral tumour.
Smith et al. (Reference Smith, Flicker, Lautenschlager, Almeida, Atkinson, Dwyer and LoGiudice2010) identified factors associated with dementia in Aboriginal people, including advanced age, current smoking, previous Cerebrovascular Accident (CVA), epilepsy, head injury and a lack of formal education. While being unable to identify risk factors, our small study did uncover further possible associations. As expected, there was a high frequency of T2DM across the study population and the possibility of undiagnosed hypoglycaemia in many of these people. Even with improved chronic disease care in the NT, food security remains a concern in the elderly, and hypoglycaemia may be an underlying factor in diabetics with Major NCD (Seligman et al., Reference Seligman, Davis, Schillinger and Wolf2010).
The prevalence of hypertension was high and spread across the whole study group. Hypertension is a known risk factor for dementia (Hughes & Sink, Reference Hughes and Sink2016). These findings suggest that better management of cardiovascular risk factors may lower the incidence of Major NCD.
Hypothyroidism was most commonly found in participants diagnosed with Major NCD due to AD (58%). Clinical hypothyroidism is now recognised as a cause of both reversible and irreversible cognitive impairment (Breteler et al., Reference Breteler, Van Dujjn, Chandra, Fratiglioni, Graves, Heyman and Mortimer1991; Harper & Roe, Reference Harper and Roe2010; Tan et al., Reference Tan, Beiser, Vasan, Au, Auerbach, Keil and Seshadri2008; Yeap et al., Reference Yeap, Alfonso, Chubb, Puri, Hankey, Flicker and Almeida2012). There is also growing literature about the role of subclinical hypothyroidism in the cognitive dysfunction of elderly people (Pasqualetti, Pagano, Rengo, Ferrara & Monzani, Reference Pasqualetti, Pagano, Rengo, Ferrara and Monzani2015).
Previous research suggests that low Vitamin D levels may be associated with cognitive decline (Balion et al., Reference Balion, Griffith, Strifler, Henderson, Patterson, Heckman and Raina2012). Vitamin D deficiency (defined as a blood level less than 50 nmol/L) is common (38.7%) in Aboriginal people living in remote areas (ABS, 2016). The high frequency of vitamin D deficiency found in this study (62% of those with Major NCD) may reflect inadequate sun exposure in an elderly population both in their home communities and in the aged care facilities. Its role, if any, in Major NCD in this population needs to be clarified.
HTLV-1 seropositivity is common in Central Australian Aboriginal people (Gordon & Hewagama, Reference Gordon and Hewagama2012) and varies by geographical region. Among the 23 study participants tested, 65% were positive and in 68% of those with Major NCD. In a cross-sectional study of 104 participants in Brazil, HTLV-1 seropositivity was linked to cognitive impairment (Gascon et al., Reference Gascon, Casseb, Smid, Vidal, Fonseca, Paiva and Penalva de Oliveira2017). Whether HTLV-1 has a causal role in the development of Major NCD in Aboriginal people is currently unknown.
Twenty-one participants had both a KICA-screen and a KICA-cog, using interpreters and conducted by the same investigator. The correlation coefficient between the two sets of results was only 0.31. This underlines the importance of repeating cognitive tests with Aboriginal people.
Limitations of this study
This is a small, cross-sectional study in which causal inference cannot be derived with any certainty. A key limitation was the lack of systematic clinical interview of participants and/or their carers. We acknowledge resource limitations as the major reason for our choice of methodology. Our participants, resident in remote Alice Springs in Central Australia, were usually far from home and family and spoke English as a second or third language. These challenges to the research setting rendered comprehensive individual assessments beyond scope. This gap was filled, to some extent, by the rich detail obtained from the medical records. Additionally, the specific study setting of RACFs in Alice Springs limits the generalisability of the findings to wider populations and does not reflect the spectrum of severity of NCDs in the community.
While imaging is an important aid in classifying categories of Major NCD, our images were not all current, and imaging diagnoses were often not clear cut. Clinical and historical factors remain essential to confident diagnosis of cognitive impairment diseases. Addition of carer interviews with interpreter support, and review of case records by a panel of experts rather than a single clinician would have potentially strengthened our diagnostic certainty.
Further, only 58 of the original 84 potential participants were included in the file audit which may have introduced bias to the results.
Conclusions
This study of Aboriginal people living in RACFs in Central Australia reveals a different dementia profile to community-based studies, with higher rates of Major NCD due to VaD and BI and lower rates of Major NCD due to AD and unspecified cause disease. It demonstrates a link between alcohol, TBI and Major NCD. The different findings suggest that services and preventative strategies may need adaptation for different populations. The study also identifies several factors that might be associated with cognitive impairment, and that warrant further investigation, including hypothyroidism, hypoglycaemia, HTLV-1 and Vitamin D deficiency.
Finally, while observing the often-animated interactions between interpreters and previously withdrawn residents, the potential benefit of employing Aboriginal carers in these RACFs became evident, especially for those residents who are far from family and country.
Acknowledgements
The authors would like to acknowledge the contributions of the residents and staff at the Aged Care Facilities as well as staff of Australian Regional and Remote Community Services, and the contribution of interpreters from the NT Aboriginal Interpreter Service.
Financial support
This work was supported by financial and in-kind contributions from Menzies School of Health Research, the Centre for Remote Health and the Northern Territory Department of Health.
Conflict of interest
BH has no conflicts of interest to disclose. ML has no conflicts of interest to disclose. TN has no conflicts of interest to disclose. CP has no conflicts of interest to disclose. ML has no conflicts of interest to disclose. AF has no conflicts of interest to disclose. HJ has no conflicts of interest to disclose. AC has no conflicts of interest to disclose. KD has no conflicts of interest to disclose.
Ethical standards
Ethics approval was granted by the Human Research Ethics Committee for NT Department of Health and Menzies School of Health Research (HREC 2015-2462 on 6/1/2016) and the Central Australian Human Research Ethics Committee (HREC-15-336 on 7/12/2015).The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
