Accepted Manuscript The prevalence of dementia sub-types in rural Tanzania Stella-Maria Paddick, MRCPsych Anna Longdon, MRCP Aloyce Kisoli, MSc William K. Gray, PhD Catherine L. Dotchin, MD Ahmed Jusabani, MD Ahmed Iqbal, FRCR Julian Hughes, PhD Andrew Teordorczuk, MD Paul Chaote, MD Richard W. Walker, MD PII:
S1064-7481(14)00074-8
DOI:
10.1016/j.jagp.2014.02.004
Reference:
AMGP 343
To appear in:
The American Journal of Geriatric Psychiatry
Received Date: 2 October 2013 Revised Date:
5 February 2014
Accepted Date: 6 February 2014
Please cite this article as: S.-M. Paddick, A. Longdon , A. Kisoli , W.K. Gray , C.L. Dotchin , A. Jusabani , A. Iqbal , J. Hughes , A. Teordorczuk , P. Chaote , R.W. Walker, The prevalence of dementia sub-types in rural Tanzania, The American Journal of Geriatric Psychiatry (2014), doi: 10.1016/ j.jagp.2014.02.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
The prevalence of dementia sub-types in rural Tanzania Stella-Maria Paddick MRCPsych,1,2 Anna Longdon MRCP,3 Aloyce Kisoli MSc,4 William K. Gray PhD,1 Catherine L. Dotchin MD1,4 Ahmed Jusabani MD5, Ahmed Iqbal FRCR,6 Julian Hughes PhD,1 Andrew Teordorczuk MD,1 Paul Chaote MD,4 Richard W. Walker, MD1,7 Institute of Ageing and Health, Newcastle University, Newcastle upon Tyne, UK
2.
Northumbria Healthcare NHS Foundation Trust, North Tyneside General Hospital, North Shields,
UK
RI PT
1.
South Devon Healthcare NHS Foundation Trust, Torquay, UK
4.
District Medical Office, Hai District Hospital, Boma’ngombe, Kilimanjaro Region, Tanzania.
5.
Kilimanjaro Christian Medical Centre, Moshi, Tanzania
6.
Institute of Neurosciences, Southern General Hospital, Glasgow, UK.
7.
Institute of Health and Society, Newcastle University, Newcastle upon Tyne, UK.
M AN U
SC
3.
Correspondence to: Catherine Dotchin, Department of Medicine, North Tyneside General Hospital, Rake Lane, North Shields, Tyne and Wear, NE29 8NH, UK; telephone and fax: 0191 293 2709. [email protected]
TE D
Word count: Text 3777, Abstract 250, References 47, Tables 3. Key words: Dementia. Alzheimer’s disease. Vascular dementia. Prevalence. Tanzania. Africa
AC C
EP
Running Title: Dementia sub-types in Tanzania
1
ACCEPTED MANUSCRIPT Abstract Objectives: The prevalence of dementia is predicted to increase rapidly in developing countries .Vascular risk factors may contribute to this rise. Our aim was to estimate the proportions of Alzheimer’s disease (ADD) and vascular dementia
Design: A two-stage door-to-door dementia prevalence study.
SC
Setting: Hai district, Tanzania
RI PT
(VAD) in a prevalent cohort of dementia cases in rural Tanzania.
Participants: In Phase I, the Community Screening Instrument for Dementia (CSI-D)
M AN U
was used to screen 1198 community-dwelling people for dementia. In Phase II, 168/184 (91.3%) of those with poor performance, 56/104 (53.8%) of those with intermediate performance and 72/910 (7.9%) of those with good performance on CSI-D were interviewed and diagnoses were made using the DSM-IV criteria.
TE D
Measurements: For subtype diagnosis, DSM-IV dementia criteria plus NINCDSADRDA criteria were used for ADD and NINDS-AIREN criteria for VAD. Other dementias were diagnosed by international consensus criteria. Diagnoses were
EP
confirmed or excluded by computerised tomography where clinically appropriate.
AC C
Results: Of 78 dementia cases, 38 (48.7%) were ADD and 32 (41.0%) were VAD. The crude prevalence of ADD was 3.7% (95% CI 2.5 to 4.9) and of VAD was 2.9% (95% CI 1.9 to 3.9). The age-adjusted prevalence was 3.0% (95% CI 1.8 to 4.2) for ADD and 2.6% (95% CI 1.6 to 3.6) for VAD. A previous diagnosis of diabetes mellitus was independently associated with greater odds of having VAD than ADD. Conclusions: VAD accounted for a greater proportion of dementia cases than expected. Further investigation and treatment of risk factors is required in this setting. 2
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
3
ACCEPTED MANUSCRIPT Introduction The prevalence of dementia in sub-Saharan Africa (SSA) was, until recently, thought to be remarkably low compared to that in high-income countries.(1-3) This view has been challenged by a number of recent studies suggesting a prevalence of dementia
RI PT
similar to that seen in the USA and Western Europe.(4-6) Indeed, the majority of people with dementia worldwide are thought to live in low- and middle-income
countries.(7-8) Furthermore, with demographic transition it is projected that the
SC
prevalence of dementia in these countries will increase at three times the rate of that
and middle-income countries.(7)
M AN U
in the developed world. By 2040 over 70% of people with dementia will live in low-
Worldwide, Alzheimer’s disease dementia (ADD) is the most prevalent form of dementia, with vascular dementia (VAD) being the second most prevalent form.(1-2) In Europe it is estimated that 16% of dementias are VAD in those aged 65 years and
TE D
over.(9) However, considerable geographical and inter-ethnic variation in the relative contribution of VAD and ADD to the overall burden of dementia may exist. A
EP
review of worldwide data suggested that in developing countries, ADD represented 60% and VAD represented 26% of all dementias.(9-10) As populations in these
AC C
regions urbanise, their risk factor profile is likely to change. This may influence overall dementia prevalence and the relative importance of the various dementia sub-types. This is particularly true in the case of cardiovascular risk factors such as hypertension, smoking, physical activity, cholesterol and obesity. There is little current reliable data on the relative proportions of dementia sub-types in SSA. Dementia incidence and prevalence studies that used either the International Classification of Disease (ICD) or Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria for diagnosis are listed in Table 1. Across all of the 4
ACCEPTED MANUSCRIPT prevalence studies, ADD accounted for 124 of 167 dementia cases (74.3%) and VAD for 38 cases (22.8%). Accurate characterisation of dementia subtypes may be impeded by lack of access to radiology services. Nevertheless, given the projected increases in dementia prevalence in SSA, it follows that it is of paramount
RI PT
importance to characterise not only the prevalence of dementia, but also its subtypes in this population. This will assist in healthcare planning in this resource poor setting.
SC
We have previously reported the prevalence of dementia in a rural site in Tanzania,
M AN U
East Africa.(11) The aim of the current study was to report the relative proportion of dementia sub-types within the cases identified and to consider the role of risk factors for each sub-type.
TE D
Methods
This study was approved by the National Institute of Medical Research, Dar-esSalaam, Tanzania and by the Newcastle and North Tyneside Joint Ethics Committee
EP
in the UK. Signed informed consent was obtained from each participant. We obtained a thumbprint for those that could not read and write. The purpose and
AC C
implications of the study were verbally explained. In cases where patients were unable to give valid consent, written assent was obtained from a close relative. The study methods have been described previously and only a brief description is given here.(11) Setting The Hai district of Northern Tanzania is a rural area on the slopes of Mount Kilimanjaro. The majority of the population in the Hai district are subsistence farmers 5
ACCEPTED MANUSCRIPT although some families are able to sell produce such as coffee and tomatoes. For financial, diagnostic and supply reasons, few older adults are on regular medication.(12).
RI PT
Case identification The Hai district of Northern Tanzania contains a demographic surveillance site
(DSS) with regular population censuses. The most recent census finished collecting
SC
data at the end of May 2009. On the 1st of June 2009 the population of the Hai DSS was 161,119. The data presented here were collected as part of a larger dementia
M AN U
prevalence study; initial findings from which have already been published.(11) We aimed to interview and assess all those aged 70 years and over living in each of six villages, selected by a random number generator from the 52 villages within the DSS.
TE D
Dementia screening and assessment
Initial phase I screening for dementia was carried out using the Community Screening Instrument for Dementia (CSI-D).(13) The CSI-D includes an interview of
EP
the person suspected of having dementia and of an informant, usually a carer or close relative. The scores for the two interviews are entered into a model which
AC C
produces an overall discriminant function score, with a higher score indicating a greater likelihood of dementia being present. The CSI-D has been validated for use in Swahili.(14) We also carried out further assessments using the 10/66 research group’s protocol for assessment of cognitive disorders in developing countries.(2, 15) As part of the protocol, the Geriatric Mental State (GMS) examination and the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) 10 word task were administered in addition to the CSI-D.
6
ACCEPTED MANUSCRIPT The local enumerators for each village administered the CSI-D after attending two training workshops delivered by our team. The CSI-D ranks individuals into ‘good performance’, ‘intermediate performance’ and ‘poor performance’ categories. During phase II we aimed to clinically assess all those in the ‘poor performance’ category,
RI PT
50% of those in the ‘intermediate performance’ category and 5% of those in the
‘good performance’ category, in accordance with the 10/66 dementia research group protocol.(15)
SC
Interviews were carried out by two UK research doctors with an interest in dementia
M AN U
accompanied by a trained study nurse fluent in both English, and Swahili, who was familiar with the local area and culture. Dementia diagnoses were based on DSM-IV criteria.(16) Dementia sub-type diagnosis used the NINDS-AIREN (National Institute of Neurological Disorders and Stroke - Association Internationale pour la Recherche et l'Enseignement en Neurosciences) criteria and Hachinski score for VAD, and
TE D
NINCDS-ADRDA (National Institute of Neurological and Communicative Disorders and Stroke - Alzheimer's Disease and Related Disorders Association) criteria for ADD.(17) Dementia with Lewy bodies (DLB) was diagnosed using the modified
EP
McKeith criteria(18) and Parkinson’s Disease Dementia (PDD) using the DSM-IV
AC C
criteria.(16) Mild cognitive impairment (MCI) was defined using the Petersen international consensus criteria.(19) Although Magnetic Resonance Imaging (MRI) was not available locally, computerised tomography (CT) scanning was available to support diagnoses where clinically indicated. Twenty-nine patients underwent CT scan. We aimed to offer CT scanning to all those with a history suggestive of VAD, moderate to severe ADD and those where there were symptoms suggestive of another cause of dementia such as normal pressure hydrocephalus or chronic subdural haematoma. Due to the 7
ACCEPTED MANUSCRIPT distances involved and the frailty of some of the participants, a full series of CT scans was not feasible in practical terms or clinically appropriate. Standard scans were obtained using a Philips Tomoscan SR 4000 5mm-10mm 37 slices. These were reported locally in Tanzania and then reported and reviewed by a UK based
RI PT
neuro-radiologist blind to clinical diagnoses (A.I.). In cases of doubt regarding
dementia sub-type, a UK-based old age psychiatrist (A.T.) reviewed case notes. In addition, a random selection of 10 suspected VAD cases were reviewed by a second
SC
UK-based old age psychiatrist (J.H.). Cases of probable and possible ADD/VAD were categorized as ADD/VAD. In cases where the cause was unclear, ‘mixed
and alcohol induced dementia. Other assessment
M AN U
aetiology’ was recorded. This included two cases where there were features of ADD
Blood pressure was recorded on all participants in the seated position (three
TE D
measurements taken one minute apart after five minutes resting quietly, with an average taken of the last two readings) using an A&D UA-767TM (A&D Instruments
EP
Ltd, UK) monitor.(20) A cut-off of ≥ 140mmHg systolic blood pressure or ≥ 90mmHg diastolic blood pressure was used to define hypertension.(21) Height and weight
AC C
were recorded where possible, although for a small number of cases weight could not be recorded due to disability. Height was recorded without footwear using a steel tape measure, with demispan recorded where the subject could not stand. Weight was recorded on a set of digital scales with subject wearing light clothes, but without footwear. Body mass index (BMI) was calculated using standard methods and a cut-off of ≥ 25 used to indicate being overweight. Patients were asked if they had been previously diagnosed with diabetes mellitus, though no formal testing was conducted as part of the current study. Non-fasting blood samples were taken from 8
ACCEPTED MANUSCRIPT those who consented. The collected blood samples were centrifuged, frozen and transported back to the UK where they were analysed at North Tyneside General Hospital using an automated biochemical analyser. In accordance with the recommendations of the US National Cholesterol Education Program,(22) high
RI PT
serum total cholesterol was defined as ≥ 200mg/dL (5.18 mmol/L) and low levels of high density lipoprotein (HDL) cholesterol as < 40 mg/dL (1.04mmol/L). Although no universally accepted cut-off exists, high total cholesterol/HDL cholesterol ratio was
SC
defined as ≥ 5.0.
M AN U
Patients and relatives were asked if they had ever had a stroke that was diagnosed by a medical doctor or if they had experienced symptoms suggestive of stroke for those unfamiliar with the term. Neurological deficits were assessed in all patients using the standard 10/66 protocol examination including tone, deep tendon reflexes, visual fields, coordination and gait. Since this was not a study of stroke, a more
TE D
detailed examination and formal assessment for stroke, according to World Health Organization criteria (23) was not undertaken.
EP
Estimation of prevalence rates
A point prevalence estimate was used. The denominator population was defined as
AC C
those aged 70 years and over who were resident in the six villages on the prevalence date, 12th April 2010. The numerator for the study was calculated based on cases identified by the study team between 12th April and 30th September 2010. Collecting accurate information on patient age can be difficult in SSA (12). Age was calculated from birth year and confirmed using memory prompts where the date of birth was in doubt. These methods have been validated in SSA (24).
9
ACCEPTED MANUSCRIPT Figure 1 depicts the study recruitment process. The six villages selected had a census population of 34,078 (21.1% of the Hai DSS population), of whom 1277 (3.7%) were aged 70 years and over. There were 169 exclusions (death since census, moved away, incorrect age recorded in census, duplicates etc.) and 152
RI PT
additions (moved to area since census, or turned 70 since census); these have been detailed previously.(11) Therefore, we identified 1260 eligible people, 62 (4.9%) of whom refused to participate. The CSI-D was administered to 1198 people in Phase
SC
I; 673 (56.2%) were female. In Phase II, we fully assessed 168 of the 184 people (91.3%) with poor performance on the CSI-D, 56 of 104 people (53.8%) with
M AN U
intermediate performance and 72 of 910 people (7.9%) with good performance on the CSI-D; giving a Phase II cohort of 296, 78 of whom were diagnosed with dementia.
Prevalence was calculated by weighting each case according to their initial CSI-D
TE D
diagnosis category from phase I (poor, intermediate or good performance). The weightings used were based on the percentage of the Phase I population within that category who were assessed during phase II. The entire data collection was
EP
completed in 6 months and the number of cases of probable dementia who could not
AC C
be followed-up was small. These cases are assumed to be missing at random and no adjustment for non-response during phase II has been made.(25) Age-standardisation, using the direct method, was to the WHO world standard.(26) Confidence intervals (CI) for the prevalence estimates were calculated using jackknife methods and adjusted for clustering of subjects by village. The 1-jackknife method involves the repeated calculation of estimates with a single observation removed for each estimate. The estimates are then combined to give a measure of the variability within the data. Six separate 1-jackknife samples, with weighting for 10
ACCEPTED MANUSCRIPT the population of each sample, were taken, with one entire village deleted from the analysis each time. Weightings were based on the percentage of the study population represented by the village. This allowed the calculation of 95% CIs based on the t-distribution.
RI PT
After phase II assessment, 78 cases of dementia were identified, giving a crude dementia prevalence of 7.5% (95% CI 6.0 to 9.0) and an age-adjusted prevalence of
SC
6.4% (95% CI 4.9 to 7.9). Statistical methods
M AN U
Data were analyzed using standard statistical software, PASW-18 for windows (PASW, Chicago, IL, USA). Only those aged 70 years and over were included in this study. As a result, ages were non-normally distributed, as were data related to CSID discriminant function score and cholesterol levels. For these data, non-parametric
TE D
descriptive statistics and significance tests (Mann-Whitney U test) were used. Risk factors for VAD were investigated in univariate and multivariable analysis using logistic regression. There were a number of missing values for cholesterol and BMI
EP
data. In univariate analysis, no attempt has been made to impute missing values. In
Results
AC C
multivariable analysis, multiple imputation methods were used.
Prevalence of sub-types From the 78 cases of dementia identified, there were 38 (48.7%) cases of ADD and 32 (41.0%) cases of VAD. Of the remaining 8 cases, 3 cases had Parkinson’s disease dementia (PDD), 1 had probable Lewy body dementia (LBD), 1 had possible LBD, 2 had mixed alcohol induced dementia and ADD and 1 case of normal pressure hydrocephalus on clinical examination (who refused CT scanning). The 11
ACCEPTED MANUSCRIPT crude prevalence of ADD was 3.7% (95% CI 2.5 to 4.9) and of VAD was 2.9% (95% CI 1.9 to 3.9). The age-adjusted prevalence was 3.0% (95% CI 1.8 to 4.2) for ADD and 2.6% (95% CI 1.6 to 3.6) for VAD. The demographic and clinical characteristics of those with different dementia sub-types are presented in Table 2. Those with
RI PT
VAD had a significantly higher CSI-D discriminant function score than those with ADD (U = 275.0, z = -3.814, p < 0.001). A previous stroke was identified clinically in 24 cases, 23 with VAD and one with LBD. Of 29 cases who underwent CT scan, 15
SC
showed signs of small vessel disease and six revealed a previous infarct.
M AN U
Risk factors for VAD
The VAD and ADD groups were compared with regard to vascular and demographic risk factors, see Table 3. In univariate analysis, high total cholesterol and a previous diagnosis of diabetes mellitus were significantly more common in VAD than ADD. In multivariable analysis, only diabetes mellitus was associated with a significantly
TE D
greater odds of having VAD, rather than ADD.
EP
Discussion
AC C
Prevalence of dementia sub-types This study was part of one of the largest dementia prevalence studies conducted to date in SSA, and the first to be conducted in East Africa. The proportion of dementia cases with VAD was noticeably higher in our study population than seen in previous studies of dementia from SSA, all of which were conducted in West and Central Africa, see Table 1. Previous studies have tended to report rates of ADD three to four times higher than those of VAD. Ratios in black people of African origin from other world regions are similar. In our population, rates of VAD and ADD were 12
ACCEPTED MANUSCRIPT almost equal. There may be many reasons for the apparently higher contribution of VAD to the overall burden of dementia. It is thought that ADD and VAD exist on a continuum with a greater or lesser contribution from a range of factors influencing the final diagnosis.(27) Most previous studies of dementia have been in those aged 65
RI PT
years and over, whilst our population was aged 70 years and over. It is well
established that vascular risk factors are more common in older age groups and this is likely to translate into a more vascular picture when causes of dementia are
SC
investigated. However, data from European populations suggest that it is ADD, rather than VAD, that is largely responsible for the increase in prevalence of
M AN U
dementia with age. Rates of VAD increase much more slowly than ADD with increasing age.(9, 28) This trend is also thought to be apparent in SSA.(29) With continuing demographic transition, rates of ADD, relative to VAD, are likely to rise in Hai.
TE D
One possible reason for the higher prevalence of VAD seen in this study may be under-reporting of VAD in previous studies from SSA. Routine use of CT brain for diagnosis of dementia is rare in this setting. Indeed, of the studies listed in Table 1,
EP
only that of Hendrie et al in 1995 used CT brain to help In diagnosis and this was
AC C
only performed in around half of all cases from Ibadan.(3) Given the reliance on clinical signs and symptoms it is possible that VAD has been under-diagnosed in some previous studies from SSA. The population in Hai have relatively high rates of hypertension, when compared to other African communities.(30) Hypertension was notably more common in our population across all sub-types than reported in other studies of dementia from SSA, and this may be one of the major contributory factors to the high rates of VAD seen in Hai.(4-5, 31-32) This is supported by previous studies of stroke in this 13
ACCEPTED MANUSCRIPT population.(33) Age-standardised stroke incidence in two demographic surveillance sites in Tanzania (Hai and Dar-es-Salaam) were found to be similar to that reported in developed countries in Hai, and higher in Dar es Salaam. A previous stroke was recorded in almost a third of all dementia cases in the current study. Treating
RI PT
hypertension may therefore represent an important future public health intervention in dementia. In a resource poor setting, prevention of chronic diseases is likely to be much more cost effective than treatment. However, many health services in SSA
SC
have been designed primarily to treat acute medical conditions and a re-modelling of services would be required.(34) However, this may be the only way that the growing
M AN U
burden of chronic non-communicable diseases can be reduced.(35) Comparison with studies from other world regions
Rates of dementia and the contribution of ADD and VAD to the overall burden, in black populations living outwith Africa have been investigated. The Ibadan-
TE D
Indianapolis study found overall rates of dementia in African-Americans to be higher than in Nigeria, suggesting that lifestyle factors played a major role in the excess
EP
burden.(36) However, the contribution of VAD and ADD at both sites appeared similar.(3, 37) A study that investigated dementia rates in people living in London,
AC C
found an excess burden in people of Afro-Caribbean origin compared to Caucasians and concluded that this was due to an excess of vascular risk factors.(38) However, drawing comparisons between our East African population and black people living in the USA and UK, who are likely to be of predominantly West African origin, is difficult given differences in genetic make-up, lifestyle and diet across Africa.(39) VAD was thought to be more common in some Asian populations than in Europe and the USA.(40) A study comparing the prevalence of dementia sub-types in Japan across two decades reported VAD to be more common that ADD in the 1980s, but 14
ACCEPTED MANUSCRIPT that ADD was becoming comparatively more prevalent by the 1990s.(41) A systematic review of dementia prevalence in Japan published in 2012 suggested that this trend appears to be continuing in more recent decades.(42) Indeed, the ratio of ADD to VAD is now thought to be similar to that seen in the USA, with ADD
RI PT
becoming increasingly common in the older age bands. The review authors note that changes in diagnostic criteria, as well as demographic ageing and lifestyle
changes, may partly explain the apparent shift in the profile of dementia. It may be
SC
that, with demographic transition, our rural Tanzanian population will see a similar
M AN U
shift in the relative contribution of dementia sub-types to the overall disease burden. Risk factors for VAD
We were able to report on a number of vascular and demographic risk factors which may be thought to be more common in VAD than ADD and may have accounted for the high prevalence of VAD seen in this cohort. However, only a previous diagnosis
TE D
diabetes mellitus and high total cholesterol emerged as significant modifiable risk factors for VAD, when compared to ADD. Being overweight, hypertension and low
EP
HDL cholesterol levels were not significantly associated with a diagnosis of VAD compared to ADD. Obesity, hypertension, smoking and diabetes mellitus have been
AC C
noted as risk factors for dementia,(43) with hypertension noted as a specific risk factor for VAD.(44) In a prospective study of 1753 elderly Yoruba Nigerians, 120 developed dementia over the six year follow-up period, with hypertension at baseline significantly more common in those with dementia.(45) However, sub-type specific risk factor data from SSA are lacking. The low number of cases studied is likely to have limited the power of our analyses and this area requires further study.(46) Limitations
15
ACCEPTED MANUSCRIPT The main limitation of our study is that CT brain was not available for all people found to have dementia. Although this would have helped to provided information on dementia sub-types, due to distances involved and frailty of some participants this was felt to be inappropriate in some cases. However, where the diagnosis was in
RI PT
doubt diagnoses were reviewed by two independent old age psychiatrists. Another difficulty was the lack of hospital or medical documentation of stroke in many cases, and the resultant difficulty in obtaining clear evidence of dementia occurring within
SC
three months of stroke for a diagnosis of VAD. In many cases we were reliant on the accuracy of an informant history. In a number of subjects, the absence of small
M AN U
vessel disease or evidence of a previous stroke by CT scan helped rule out VAD, despite a history which was suggestive of VAD. However, relatives in most cases resided with the patient they were caring for and were asked about onset of dementia and previous strokes in different sections of the interview as open-ended
TE D
answers. As a result, we do not feel any systematic bias will have resulted. We were only able to obtain blood samples from 50 of the 78 people with dementia. Refusal to provide blood samples is common in rural SSA and we feel this would have been a
EP
limitation of any study conducted in this setting. Our data relating to diabetes
AC C
mellitus are based on self-reported previous diagnosis and are therefore likely to represent an under estimate. Finally, we acknowledge that the numbers identified with each dementia subtype were relatively small. Nevertheless, our study is one of the largest and most comprehensive studies of dementia sub-types conducted in SSA to date. The lack of previous data highlights the unique challenges in conducting epidemiological research in this low resource setting. Conclusions
16
ACCEPTED MANUSCRIPT Our data are unique in that they were collected from an East African population. The contribution of VAD to the overall dementia burden was high. This is likely to be due in part to high rates of vascular risk factors within the background population, most notably hypertension. As the population undergoes demographic transition the
RI PT
prevalence and profile of dementia is likely to change. Health services in SSA need
SC
to adapt to the changing health needs of their people.
Acknowledgements: We wish to acknowledge the help of all health care workers,
M AN U
village enumerators and medical supervisors who assisted in identification of cases, clinical assessment and completion of the census. We are especially grateful to patients, carers and family members for taking part in the study. Conflicts of Interest: There were no conflicts of interest.
TE D
Funding: This study was part funded by a British Geriatric Society SpR start up grant and an Academy of Medical Sciences (UK) Clinical Lecturer start up grant. Role of the Funding Source: The sponsors of this study had no role in designing
EP
the study; in the collection, analysis, and interpretation of data; in the writing of the
AC C
report; or in the decision to submit the paper for publication.
Contributions
Design/conception – Richard Walker, Catherine Dotchin, Stella-Maria Paddick, Anna Longdon, Paul Chaote, Ahmed Jusabani. Literature search - Richard Walker, Catherine Dotchin, Stella-Maria Paddick. Data collection – Stella-Maria Paddick, Anna Longdon, Aloyce Kisoli, Ahmed Iqbal. Data analysis - Richard Walker, William K. Gray, Catherine Dotchin, Stella-Maria Paddick. Interpretation of results - Richard 17
ACCEPTED MANUSCRIPT Walker, Catherine Dotchin, William Gray, Stella-Maria Paddick, Anna Longdon, Andrew Teodorczuk, Julian Hughes. Writing of paper - Richard Walker, Catherine Dotchin, William Gray, Stella-Maria Paddick, Anna Longdon, Andrew Teodorczuk,
AC C
EP
TE D
M AN U
SC
RI PT
Julian Hughes, Aloyce Kisoli, Paul Chaote, Ahmed Jusabani and Ahmed Iqbal.
18
ACCEPTED MANUSCRIPT Figure 1. Diagram of the recruitment and randomisation process Census population: n = 161,119
Died: n = 77
Too young: n = 16
Duplicates: n=4
M AN U
Moved away: n = 64
SC
Census population aged 70 and over in the 6 randomly selected villages: n = 1277
RI PT
Census population of 6 randomly selected villages: n = 34,078
Not found: n =8
TE D
Actual population from census: n = 1108
Final phase I study population: n = 1198
EP
Refused: n = 62
Identified by enumerators postcensus: n = 152
AC C
Probable dementia by CSI-D: n = 184
Follow-up in phase II: n = 168 (91.3%)
Possible dementia by CSI-D: n = 104
No dementia by CSI-D: n = 910
Randomly selected for phase II: n = 56 (53.8%)
Randomly selected for phase II: n = 72 (7.9%)
19
ACCEPTED MANUSCRIPT
Table 1. Dementia sub-types reported in previous sub-Saharan Africa studies Setting
Diagnostic Criteria
Imaging Available
Reported sub-types
Prevalence
Ibadan, Nigeria (urban)
DSM-III-R and ICD-10 for dementia
Yes. Only 10 dementia patients due to strikes/logistical issues at time of study.
ADD, 18 (64.3%)
RI PT
Hendrie et al, 1995(3)
Design
NINCDS-ADRDA for ADD
SC
ICD-10 for VAD.
Hendrie et al, 2001(37)
Incidence
Ibadan, Nigeria (urban)
DSM-IV and ICD-10 for dementia
NINCDS-ADRDA for ADD.
Guerchet et al, 2010(5)
Prevalence
Bangui, Central African Republic (urban)
NINCDS-ADRDA for ADD
DSM-IV for dementia
Brazzaville, Congo (urban)
Lewy body dementia, 1 (3.6%) Depression related dementia, 1 (3.6%)
Yes.
ADD, 62 (88.6%)
28 cognitive impairment,
VAD, 4 (5.7%)
15 dementia
PDD, 1(1.4%)
19 controls
Dementia with severe chronic illness, 2 (2.9%)
No
Other subtypes based on clinical history
EP
Prevalence
Djidja, Benin (rural)
Unspecified dementia, 1 (1.4%) ADD, 11 (84.6%) VAD, 1 (7.7%) Dementia with progressive aphasia, 1 (7.7%)
No
AC C
Guerchet et al, 2010(5)
Prevalence
TE D
ICD-10 for VAD.
Guerchet et al, 2009(31)
M AN U
ICD-10 for secondary dementia
VAD, 8 (28.6%)
NINCDS-ADRDA for ADD
ADD, 33 (82.5%) VAD, 7 (17.5%)
Hachinski score and clinical history for VAD DSM-IV for dementia
No
NINCDS-ADRDA for ADD
ADD, 24 (68.6%) VAD, 11 (31.4%)
20
ACCEPTED MANUSCRIPT
Hachinski score and clinical history for VAD Prevalence
Cotonou, Benin (urban)
DSM-IV for dementia
No
RI PT
Paraiso et al, 2011(32)
NINCDS-ADRDA for ADD.
Yusuf et al, 2011(47)
Prevalence
Northern Nigeria (rural)
ICD and DSM-IV for dementia
No
DSM-IV for ADD
M AN U
ICD10 for VAD. McKeith for dementia with Lewy bodies Lund for dementia Ibadan, Nigeria (urban)
ICD-10 and DSM-IV for dementia
No
TE D
Incidence
VAD, 9 (21.4%) Parkinson’s disease dementia, 1 (2.4%) ADD, 6 (66.7%) VAD, 2 (22.2%) Frontotemporal dementia, 1 (11.1%)
frontotemporal
EP
Ogunniyi et al, 2011(45)
SC
NINDS-AIREN for VAD
ADD, 32 (76.2%)
ADD, 99 (82.5%) VAD, 11 (9.2%) Dementia with depression, 5 (4.2%) Dementia with Parkinson’s disease, 2 (1.7%) Other causes, 3 (2.5%)
AC C
ADD Alzheimer’s disease dementia, VAD Vascular dementia, PDD Parkinson’s disease dementia, DSM Diagnostic and statistical manual of mental disorders, ICD International Classification of Disease, NINCDS-ADRDA National Institute of Neurological and Communicative Disorders and Stroke - Alzheimer's Disease and Related Disorders Association, NINDS-AREN National Institute of Neurological Disorders and Stroke Association Internationale pour la Recherche et l'Enseignement en Neurosciences
21
ACCEPTED MANUSCRIPT
Table 2. Demographic and clinical characteristics of dementia cases by sub-type Other (n = 8)
No dementia (n = 218)
Median age in years (IQR)
85.0 (80.0 to 91.0)
84 (76.25 to 90.0)
84.5 (75.5 to 94.25)
78.0 (74.0 to 86.0)
Gender
30 female (78.9%)
21 female (65.6%)
5 female (62.5%)
145 female (66.5%)
Median CSI-D discriminant function score (IQR)
0.26 (0.21 to 0.49)
0.47 (0.34 to 0.75)
0.44 (0.22 to 0.68)
0.17 (0.08 to 0.26)
Diabetes previously diagnosed
3 (7.9%)
9 (28.1%)
1 (12.5%)
39 (17.9%)
Median total cholesterol in mmol/L (IQR)*
4.5 (4.3 to 4.8)
4.9 (4.0 to 5.6)
4.8 (4.5 to 6.5)
-
Median HDL cholesterol in mmol/L (IQR)*
1.2 (0.9 to 1.4)
1.2 (1.0 to 1.5)
1.1 (1.1 to 1.3)
-
Median total cholesterol/HDL cholesterol ratio (IQR)*
3.8 (3.3 to 4.2)
4.1 (3.3 to 4.7)
4.4 (3.9 to 6.3)
-
Median body mass index (IQR)**
19.9 (17.4 to 25.1)
21.5 (18.7 to 23.5)
21.5 (18.5 to 28.9)
20.8 (18.4 to 23.5)
Mean systolic blood pressure in mmHg (SD)
162.1 (36.631)
175.8 (30.450)
167.0 (17.696)
161.5 (33.831)
Mean diastolic blood pressure in mmHg (SD)
84.9 (17.212)
94.3 (19.157)
81.5 (11.796)
87.5 (16.508)
M AN U
TE D
RI PT
VAD (n = 32)
SC
ADD (n = 38)
EP
IQR = inter-quartile range, SD standard deviation, HDL high density lipoprotein * Based on blood samples taken from 23 subjects with ADD, 22 with VAD and 5 with other causes of dementia
AC C
** Based on 34 subjects with ADD, 22 with VAD, 7 with other causes of dementia and 201 subjects without dementia
22
ACCEPTED MANUSCRIPT
Table 3. Univariate analysis of VAD risk factors Variable
Univariate odds ratio (95% CI, significance)
Age band
1
75-79 years
0.28 (0.04 to 2.04, p = 0.208)
80-84 years
0.25 (0.04 to 1.70, p = 0.157)
85 years and over
0.23 (0.04 to 1.28, p = 0.093)
Male gender
1.96 (0.68 to 5.72, p = 0.215)
Diabetes previously diagnosed
4.57 (1.12 to 18.67, p = 0.035)
Hypertension
2.49 (0.77 to 8.06, p = 0.127)
M AN U
High total cholesterol*
SC
RI PT
70-74 years
3.29 (0.83 to 12.98, p = 0.089)
Low HDL cholesterol*
1.32 (0.36 to 4.82, p = 0.672)
High total cholesterol to HDL cholesterol
ratio* High body mass index**
1.05 (0.19 to 5.87, p = 0.953)
TE D
0.51 (0.12 to 2.19, p = 0.368)
CI confidence interval, HDL high density lipoprotein VAD coded 1, ADD coded 0
* Based on blood samples taken from 23 subjects with ADD and 22 with VAD
AC C
EP
** Based on 34 subjects with ADD and 22 with VAD
23
ACCEPTED MANUSCRIPT
References 1.
Ferri CP, Prince M, Brayne C, et al: Global prevalence of dementia: a
2.
RI PT
Delphi consensus study. Lancet 2005; 366:2112-2117 Llibre Rodriguez JJ, Ferri CP, Acosta D, et al: Prevalence of dementia in
Latin America, India, and China: a population-based cross-sectional survey.
3.
SC
Lancet 2008; 372:464-474
Hendrie HC, Osuntokun BO, Hall KS, et al: Prevalence of Alzheimer's
M AN U
disease and dementia in two communities: Nigerian Africans and African Americans. Am J Psychiatry 1995; 152:1485-1492 4.
Ochayi B,Thacher TD: Risk factors for dementia in central Nigeria. Aging
Ment Health 2006; 10:616-620
Guerchet M, M'Belesso P, Mouanga AM, et al: Prevalence of dementia in
TE D
5.
elderly living in two cities of Central Africa: the EDAC survey. Dement Geriatr Cogn Disord 2010; 30:261-268
Gureje O, Ogunniyi A,Kola L: The profile and impact of probable dementia
EP
6.
in a sub-Saharan African community: Results from the Ibadan Study of Aging. J
7.
AC C
Psychosom Res 2006; 61:327-333 Wimo A,Prince M: World Alzheimer Report 2010: The Global Economic
Impact of Dementia, London, UK, Alzheimer Disease International, 2010 8.
Prince M, Bryce R, Albanese E, et al: The global prevalence of dementia:
a systematic review and metaanalysis. Alzheimers Dement 2013; 9:63-75 e62
24
ACCEPTED MANUSCRIPT
9.
Lobo A, Launer LJ, Fratiglioni L, et al: Prevalence of dementia and major
subtypes in Europe: A collaborative study of population-based cohorts.
10.
RI PT
Neurologic Diseases in the Elderly Research Group. Neurology 2000; 54:S4-9 Kalaria RN, Maestre GE, Arizaga R, et al: Alzheimer's disease and
vascular dementia in developing countries: prevalence, management, and risk
11.
SC
factors. Lancet Neurol 2008; 7:812-826
Longdon AR, Paddick SM, Kisoli A, et al: The prevalence of dementia in
M AN U
rural Tanzania: a cross-sectional community-based study. Int J Geriatr Psychiatry 2013; 28:728-737 12.
Adult Morbidity and Mortality Project (AMMP): Policy Implications of Adult
Morbidity and Mortality; final report, Dar-es-Salaam, Tanzanian Ministry of
13.
TE D
Health, 2004
Hall KS, Hendrie HC, Brittain HM, et al: The Development of a Dementia
Screening Interview in 2 Distinct Languages. International Journal of Methods in
14.
EP
Psychiatric Research 1993; 3:1-28
Chen CH, Mizuno T, Elston R, et al: A comparative study to screen
AC C
dementia and APOE genotypes in an ageing East African population. Neurobiol Aging 2010; 31:732-740 15.
Prince M, Acosta D, Chiu H, et al: Dementia diagnosis in developing
countries: a cross-cultural validation study. Lancet 2003; 361:909-917
25
ACCEPTED MANUSCRIPT
16.
American Psychiatric Association: Diagnostic and statistical manual of
mental disorders, 4th edition, Washington DC, American Medical Association,
17.
RI PT
1994 Hachinski V, Iadecola C, Petersen RC, et al: National Institute of
Neurological Disorders and Stroke-Canadian Stroke Network vascular cognitive
18.
SC
impairment harmonization standards. Stroke 2006; 37:2220-2241
McKeith IG, Galasko D, Kosaka K, et al: Consensus guidelines for the
M AN U
clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international workshop. Neurology 1996; 47:1113-1124 19.
Petersen RC, Smith GE, Waring SC, et al: Mild cognitive impairment:
clinical characterization and outcome. Arch Neurol 1999; 56:303-308 Rogoza AN, Pavlova TS,Sergeeva MV: Validation of A&D UA-767 device
TE D
20.
for the self-measurement of blood pressure. Blood Press Monit 2000; 5:227-231 21.
Williams B, Poulter NR, Brown MJ, et al: Guidelines for management of
EP
hypertension: report of the fourth working party of the British Hypertension
22.
AC C
Society, 2004-BHS IV. J Hum Hypertens 2004; 18:139-185 Third Report of the National Cholesterol Education Program (NCEP)
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 2002; 106:3143-3421 23.
WHO MONICA Project: The World Health Organization MONICA Project
(monitoring trends and determinants in cardiovascular disease): a major
26
ACCEPTED MANUSCRIPT
international collaboration. WHO MONICA Project Principal Investigators. J Clin Epidemiol 1988; 41:105-114 Paraiso MN, Houinato D, Guerchet M, et al: Validation of the use of
RI PT
24.
historical events to estimate the age of subjects aged 65 years and over in Cotonou (Benin). Neuroepidemiology 2010; 35:12-16
Gao S, Hui SL, Hall KS, et al: Estimating disease prevalence from two-
SC
25.
phase surveys with nonresponse at the second phase. Stat Med 2000; 19:2101-
26.
M AN U
2114
Ahmad OB, Boschi-Pinto C, Lopez AD, et al: Age standardization of rates:
A new WHO standard (GPE discussion paper Series: No. 31), Geneva, World Health Organization, 2001
Prince M,Jackson J: World Alzheimer Report 2009, London, UK,
TE D
27.
Alzheimer Disease International, 2009 28.
Fratiglioni L, Launer LJ, Andersen K, et al: Incidence of dementia and
EP
major subtypes in Europe: A collaborative study of population-based cohorts.
29.
AC C
Neurologic Diseases in the Elderly Research Group. Neurology 2000; 54:S10-15 George-Carey R, Adeloye D, Chan KY, et al: An estimate of the
prevalence of dementia in Africa: A systematic analysis. J Glob Health 2012; 2:20401 30.
Dewhurst MJ, Dewhurst F, Gray WK, et al: The high prevalence of
hypertension in rural-dwelling Tanzanian older adults and the disparity between
27
ACCEPTED MANUSCRIPT
detection, treatment and control: a rule of sixths? J Hum Hypertens 2013; 27:374-380 Guerchet M, Houinato D, Paraiso MN, et al: Cognitive impairment and
RI PT
31.
dementia in elderly people living in rural Benin, west Africa. Dement Geriatr Cogn Disord 2009; 27:34-41
Paraiso MN, Guerchet M, Saizonou J, et al: Prevalence of Dementia
SC
32.
among Elderly People Living in Cotonou, an Urban Area of Benin (West Africa).
33.
M AN U
Neuroepidemiology 2011; 36:245-251
Walker R, Whiting D, Unwin N, et al: Stroke incidence in rural and urban
Tanzania: a prospective, community-based study. Lancet Neurol 2010; 9:786792
Reidpath DD,Allotey P: The burden is great and the money little: Changing
TE D
34.
chronic disease management in low- and middle-income countries. J Glob Health 2012; 2:20301
Di Cesare M, Khang YH, Asaria P, et al: Inequalities in non-communicable
EP
35.
36.
AC C
diseases and effective responses. Lancet 2013; 381:585-597 Hendrie HC, Murrell J, Gao S, et al: International studies in dementia with
particular emphasis on populations of African origin. Alzheimer Dis Assoc Disord 2006; 20:S42-46 37.
Hendrie HC, Ogunniyi A, Hall KS, et al: Incidence of dementia and
Alzheimer disease in 2 communities: Yoruba residing in Ibadan, Nigeria, and African Americans residing in Indianapolis, Indiana. JAMA 2001; 285:739-747 28
ACCEPTED MANUSCRIPT
38.
Livingston G, Leavey G, Kitchen G, et al: Mental health of migrant elders--
the Islington study. Br J Psychiatry 2001; 179:361-366 Tishkoff SA, Reed FA, Friedlaender FR, et al: The genetic structure and
RI PT
39.
history of Africans and African Americans. Science 2009; 324:1035-1044 40.
Suh GH,Shah A: A review of the epidemiological transition in dementia--
SC
cross-national comparisons of the indices related to Alzheimer's disease and vascular dementia. Acta Psychiatr Scand 2001; 104:4-11
Kiyohara Y, Yoshitake T, Kato I, et al: Changing patterns in the prevalence
M AN U
41.
of dementia in a Japanese community: the Hisayama study. Gerontology 1994; 40 Suppl 2:29-35 42.
Dodge HH, Buracchio TJ, Fisher GG, et al: Trends in the prevalence of
43.
TE D
dementia in Japan. Int J Alzheimers Dis 2012; 2012:956354 Whitmer RA, Sidney S, Selby J, et al: Midlife cardiovascular risk factors
and risk of dementia in late life. Neurology 2005; 64:277-281 Sharp SI, Aarsland D, Day S, et al: Hypertension is a potential risk factor
EP
44.
669 45.
AC C
for vascular dementia: systematic review. Int J Geriatr Psychiatry 2011; 26:661-
Ogunniyi A, Lane KA, Baiyewu O, et al: Hypertension and incident
dementia in community-dwelling elderly Yoruba Nigerians. Acta Neurol Scand 2011; 124:396-402
29
ACCEPTED MANUSCRIPT
46.
Davis DH, Kreisel SH, Muniz Terrera G, et al: The epidemiology of
delirium: challenges and opportunities for population studies. Am J Geriatr
47.
RI PT
Psychiatry 2013; 21:1173-1189 Yusuf AJ, Baiyewu O, Sheikh TL, et al: Prevalence of dementia and
dementia subtypes among community-dwelling elderly people in northern
AC C
EP
TE D
M AN U
SC
Nigeria. Int Psychogeriatr 2011; 23:379-386
30
S1064-7481(14)00074-8
DOI:
10.1016/j.jagp.2014.02.004
Reference:
AMGP 343
To appear in:
The American Journal of Geriatric Psychiatry
Received Date: 2 October 2013 Revised Date:
5 February 2014
Accepted Date: 6 February 2014
Please cite this article as: S.-M. Paddick, A. Longdon , A. Kisoli , W.K. Gray , C.L. Dotchin , A. Jusabani , A. Iqbal , J. Hughes , A. Teordorczuk , P. Chaote , R.W. Walker, The prevalence of dementia sub-types in rural Tanzania, The American Journal of Geriatric Psychiatry (2014), doi: 10.1016/ j.jagp.2014.02.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
The prevalence of dementia sub-types in rural Tanzania Stella-Maria Paddick MRCPsych,1,2 Anna Longdon MRCP,3 Aloyce Kisoli MSc,4 William K. Gray PhD,1 Catherine L. Dotchin MD1,4 Ahmed Jusabani MD5, Ahmed Iqbal FRCR,6 Julian Hughes PhD,1 Andrew Teordorczuk MD,1 Paul Chaote MD,4 Richard W. Walker, MD1,7 Institute of Ageing and Health, Newcastle University, Newcastle upon Tyne, UK
2.
Northumbria Healthcare NHS Foundation Trust, North Tyneside General Hospital, North Shields,
UK
RI PT
1.
South Devon Healthcare NHS Foundation Trust, Torquay, UK
4.
District Medical Office, Hai District Hospital, Boma’ngombe, Kilimanjaro Region, Tanzania.
5.
Kilimanjaro Christian Medical Centre, Moshi, Tanzania
6.
Institute of Neurosciences, Southern General Hospital, Glasgow, UK.
7.
Institute of Health and Society, Newcastle University, Newcastle upon Tyne, UK.
M AN U
SC
3.
Correspondence to: Catherine Dotchin, Department of Medicine, North Tyneside General Hospital, Rake Lane, North Shields, Tyne and Wear, NE29 8NH, UK; telephone and fax: 0191 293 2709. [email protected]
TE D
Word count: Text 3777, Abstract 250, References 47, Tables 3. Key words: Dementia. Alzheimer’s disease. Vascular dementia. Prevalence. Tanzania. Africa
AC C
EP
Running Title: Dementia sub-types in Tanzania
1
ACCEPTED MANUSCRIPT Abstract Objectives: The prevalence of dementia is predicted to increase rapidly in developing countries .Vascular risk factors may contribute to this rise. Our aim was to estimate the proportions of Alzheimer’s disease (ADD) and vascular dementia
Design: A two-stage door-to-door dementia prevalence study.
SC
Setting: Hai district, Tanzania
RI PT
(VAD) in a prevalent cohort of dementia cases in rural Tanzania.
Participants: In Phase I, the Community Screening Instrument for Dementia (CSI-D)
M AN U
was used to screen 1198 community-dwelling people for dementia. In Phase II, 168/184 (91.3%) of those with poor performance, 56/104 (53.8%) of those with intermediate performance and 72/910 (7.9%) of those with good performance on CSI-D were interviewed and diagnoses were made using the DSM-IV criteria.
TE D
Measurements: For subtype diagnosis, DSM-IV dementia criteria plus NINCDSADRDA criteria were used for ADD and NINDS-AIREN criteria for VAD. Other dementias were diagnosed by international consensus criteria. Diagnoses were
EP
confirmed or excluded by computerised tomography where clinically appropriate.
AC C
Results: Of 78 dementia cases, 38 (48.7%) were ADD and 32 (41.0%) were VAD. The crude prevalence of ADD was 3.7% (95% CI 2.5 to 4.9) and of VAD was 2.9% (95% CI 1.9 to 3.9). The age-adjusted prevalence was 3.0% (95% CI 1.8 to 4.2) for ADD and 2.6% (95% CI 1.6 to 3.6) for VAD. A previous diagnosis of diabetes mellitus was independently associated with greater odds of having VAD than ADD. Conclusions: VAD accounted for a greater proportion of dementia cases than expected. Further investigation and treatment of risk factors is required in this setting. 2
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
3
ACCEPTED MANUSCRIPT Introduction The prevalence of dementia in sub-Saharan Africa (SSA) was, until recently, thought to be remarkably low compared to that in high-income countries.(1-3) This view has been challenged by a number of recent studies suggesting a prevalence of dementia
RI PT
similar to that seen in the USA and Western Europe.(4-6) Indeed, the majority of people with dementia worldwide are thought to live in low- and middle-income
countries.(7-8) Furthermore, with demographic transition it is projected that the
SC
prevalence of dementia in these countries will increase at three times the rate of that
and middle-income countries.(7)
M AN U
in the developed world. By 2040 over 70% of people with dementia will live in low-
Worldwide, Alzheimer’s disease dementia (ADD) is the most prevalent form of dementia, with vascular dementia (VAD) being the second most prevalent form.(1-2) In Europe it is estimated that 16% of dementias are VAD in those aged 65 years and
TE D
over.(9) However, considerable geographical and inter-ethnic variation in the relative contribution of VAD and ADD to the overall burden of dementia may exist. A
EP
review of worldwide data suggested that in developing countries, ADD represented 60% and VAD represented 26% of all dementias.(9-10) As populations in these
AC C
regions urbanise, their risk factor profile is likely to change. This may influence overall dementia prevalence and the relative importance of the various dementia sub-types. This is particularly true in the case of cardiovascular risk factors such as hypertension, smoking, physical activity, cholesterol and obesity. There is little current reliable data on the relative proportions of dementia sub-types in SSA. Dementia incidence and prevalence studies that used either the International Classification of Disease (ICD) or Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria for diagnosis are listed in Table 1. Across all of the 4
ACCEPTED MANUSCRIPT prevalence studies, ADD accounted for 124 of 167 dementia cases (74.3%) and VAD for 38 cases (22.8%). Accurate characterisation of dementia subtypes may be impeded by lack of access to radiology services. Nevertheless, given the projected increases in dementia prevalence in SSA, it follows that it is of paramount
RI PT
importance to characterise not only the prevalence of dementia, but also its subtypes in this population. This will assist in healthcare planning in this resource poor setting.
SC
We have previously reported the prevalence of dementia in a rural site in Tanzania,
M AN U
East Africa.(11) The aim of the current study was to report the relative proportion of dementia sub-types within the cases identified and to consider the role of risk factors for each sub-type.
TE D
Methods
This study was approved by the National Institute of Medical Research, Dar-esSalaam, Tanzania and by the Newcastle and North Tyneside Joint Ethics Committee
EP
in the UK. Signed informed consent was obtained from each participant. We obtained a thumbprint for those that could not read and write. The purpose and
AC C
implications of the study were verbally explained. In cases where patients were unable to give valid consent, written assent was obtained from a close relative. The study methods have been described previously and only a brief description is given here.(11) Setting The Hai district of Northern Tanzania is a rural area on the slopes of Mount Kilimanjaro. The majority of the population in the Hai district are subsistence farmers 5
ACCEPTED MANUSCRIPT although some families are able to sell produce such as coffee and tomatoes. For financial, diagnostic and supply reasons, few older adults are on regular medication.(12).
RI PT
Case identification The Hai district of Northern Tanzania contains a demographic surveillance site
(DSS) with regular population censuses. The most recent census finished collecting
SC
data at the end of May 2009. On the 1st of June 2009 the population of the Hai DSS was 161,119. The data presented here were collected as part of a larger dementia
M AN U
prevalence study; initial findings from which have already been published.(11) We aimed to interview and assess all those aged 70 years and over living in each of six villages, selected by a random number generator from the 52 villages within the DSS.
TE D
Dementia screening and assessment
Initial phase I screening for dementia was carried out using the Community Screening Instrument for Dementia (CSI-D).(13) The CSI-D includes an interview of
EP
the person suspected of having dementia and of an informant, usually a carer or close relative. The scores for the two interviews are entered into a model which
AC C
produces an overall discriminant function score, with a higher score indicating a greater likelihood of dementia being present. The CSI-D has been validated for use in Swahili.(14) We also carried out further assessments using the 10/66 research group’s protocol for assessment of cognitive disorders in developing countries.(2, 15) As part of the protocol, the Geriatric Mental State (GMS) examination and the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) 10 word task were administered in addition to the CSI-D.
6
ACCEPTED MANUSCRIPT The local enumerators for each village administered the CSI-D after attending two training workshops delivered by our team. The CSI-D ranks individuals into ‘good performance’, ‘intermediate performance’ and ‘poor performance’ categories. During phase II we aimed to clinically assess all those in the ‘poor performance’ category,
RI PT
50% of those in the ‘intermediate performance’ category and 5% of those in the
‘good performance’ category, in accordance with the 10/66 dementia research group protocol.(15)
SC
Interviews were carried out by two UK research doctors with an interest in dementia
M AN U
accompanied by a trained study nurse fluent in both English, and Swahili, who was familiar with the local area and culture. Dementia diagnoses were based on DSM-IV criteria.(16) Dementia sub-type diagnosis used the NINDS-AIREN (National Institute of Neurological Disorders and Stroke - Association Internationale pour la Recherche et l'Enseignement en Neurosciences) criteria and Hachinski score for VAD, and
TE D
NINCDS-ADRDA (National Institute of Neurological and Communicative Disorders and Stroke - Alzheimer's Disease and Related Disorders Association) criteria for ADD.(17) Dementia with Lewy bodies (DLB) was diagnosed using the modified
EP
McKeith criteria(18) and Parkinson’s Disease Dementia (PDD) using the DSM-IV
AC C
criteria.(16) Mild cognitive impairment (MCI) was defined using the Petersen international consensus criteria.(19) Although Magnetic Resonance Imaging (MRI) was not available locally, computerised tomography (CT) scanning was available to support diagnoses where clinically indicated. Twenty-nine patients underwent CT scan. We aimed to offer CT scanning to all those with a history suggestive of VAD, moderate to severe ADD and those where there were symptoms suggestive of another cause of dementia such as normal pressure hydrocephalus or chronic subdural haematoma. Due to the 7
ACCEPTED MANUSCRIPT distances involved and the frailty of some of the participants, a full series of CT scans was not feasible in practical terms or clinically appropriate. Standard scans were obtained using a Philips Tomoscan SR 4000 5mm-10mm 37 slices. These were reported locally in Tanzania and then reported and reviewed by a UK based
RI PT
neuro-radiologist blind to clinical diagnoses (A.I.). In cases of doubt regarding
dementia sub-type, a UK-based old age psychiatrist (A.T.) reviewed case notes. In addition, a random selection of 10 suspected VAD cases were reviewed by a second
SC
UK-based old age psychiatrist (J.H.). Cases of probable and possible ADD/VAD were categorized as ADD/VAD. In cases where the cause was unclear, ‘mixed
and alcohol induced dementia. Other assessment
M AN U
aetiology’ was recorded. This included two cases where there were features of ADD
Blood pressure was recorded on all participants in the seated position (three
TE D
measurements taken one minute apart after five minutes resting quietly, with an average taken of the last two readings) using an A&D UA-767TM (A&D Instruments
EP
Ltd, UK) monitor.(20) A cut-off of ≥ 140mmHg systolic blood pressure or ≥ 90mmHg diastolic blood pressure was used to define hypertension.(21) Height and weight
AC C
were recorded where possible, although for a small number of cases weight could not be recorded due to disability. Height was recorded without footwear using a steel tape measure, with demispan recorded where the subject could not stand. Weight was recorded on a set of digital scales with subject wearing light clothes, but without footwear. Body mass index (BMI) was calculated using standard methods and a cut-off of ≥ 25 used to indicate being overweight. Patients were asked if they had been previously diagnosed with diabetes mellitus, though no formal testing was conducted as part of the current study. Non-fasting blood samples were taken from 8
ACCEPTED MANUSCRIPT those who consented. The collected blood samples were centrifuged, frozen and transported back to the UK where they were analysed at North Tyneside General Hospital using an automated biochemical analyser. In accordance with the recommendations of the US National Cholesterol Education Program,(22) high
RI PT
serum total cholesterol was defined as ≥ 200mg/dL (5.18 mmol/L) and low levels of high density lipoprotein (HDL) cholesterol as < 40 mg/dL (1.04mmol/L). Although no universally accepted cut-off exists, high total cholesterol/HDL cholesterol ratio was
SC
defined as ≥ 5.0.
M AN U
Patients and relatives were asked if they had ever had a stroke that was diagnosed by a medical doctor or if they had experienced symptoms suggestive of stroke for those unfamiliar with the term. Neurological deficits were assessed in all patients using the standard 10/66 protocol examination including tone, deep tendon reflexes, visual fields, coordination and gait. Since this was not a study of stroke, a more
TE D
detailed examination and formal assessment for stroke, according to World Health Organization criteria (23) was not undertaken.
EP
Estimation of prevalence rates
A point prevalence estimate was used. The denominator population was defined as
AC C
those aged 70 years and over who were resident in the six villages on the prevalence date, 12th April 2010. The numerator for the study was calculated based on cases identified by the study team between 12th April and 30th September 2010. Collecting accurate information on patient age can be difficult in SSA (12). Age was calculated from birth year and confirmed using memory prompts where the date of birth was in doubt. These methods have been validated in SSA (24).
9
ACCEPTED MANUSCRIPT Figure 1 depicts the study recruitment process. The six villages selected had a census population of 34,078 (21.1% of the Hai DSS population), of whom 1277 (3.7%) were aged 70 years and over. There were 169 exclusions (death since census, moved away, incorrect age recorded in census, duplicates etc.) and 152
RI PT
additions (moved to area since census, or turned 70 since census); these have been detailed previously.(11) Therefore, we identified 1260 eligible people, 62 (4.9%) of whom refused to participate. The CSI-D was administered to 1198 people in Phase
SC
I; 673 (56.2%) were female. In Phase II, we fully assessed 168 of the 184 people (91.3%) with poor performance on the CSI-D, 56 of 104 people (53.8%) with
M AN U
intermediate performance and 72 of 910 people (7.9%) with good performance on the CSI-D; giving a Phase II cohort of 296, 78 of whom were diagnosed with dementia.
Prevalence was calculated by weighting each case according to their initial CSI-D
TE D
diagnosis category from phase I (poor, intermediate or good performance). The weightings used were based on the percentage of the Phase I population within that category who were assessed during phase II. The entire data collection was
EP
completed in 6 months and the number of cases of probable dementia who could not
AC C
be followed-up was small. These cases are assumed to be missing at random and no adjustment for non-response during phase II has been made.(25) Age-standardisation, using the direct method, was to the WHO world standard.(26) Confidence intervals (CI) for the prevalence estimates were calculated using jackknife methods and adjusted for clustering of subjects by village. The 1-jackknife method involves the repeated calculation of estimates with a single observation removed for each estimate. The estimates are then combined to give a measure of the variability within the data. Six separate 1-jackknife samples, with weighting for 10
ACCEPTED MANUSCRIPT the population of each sample, were taken, with one entire village deleted from the analysis each time. Weightings were based on the percentage of the study population represented by the village. This allowed the calculation of 95% CIs based on the t-distribution.
RI PT
After phase II assessment, 78 cases of dementia were identified, giving a crude dementia prevalence of 7.5% (95% CI 6.0 to 9.0) and an age-adjusted prevalence of
SC
6.4% (95% CI 4.9 to 7.9). Statistical methods
M AN U
Data were analyzed using standard statistical software, PASW-18 for windows (PASW, Chicago, IL, USA). Only those aged 70 years and over were included in this study. As a result, ages were non-normally distributed, as were data related to CSID discriminant function score and cholesterol levels. For these data, non-parametric
TE D
descriptive statistics and significance tests (Mann-Whitney U test) were used. Risk factors for VAD were investigated in univariate and multivariable analysis using logistic regression. There were a number of missing values for cholesterol and BMI
EP
data. In univariate analysis, no attempt has been made to impute missing values. In
Results
AC C
multivariable analysis, multiple imputation methods were used.
Prevalence of sub-types From the 78 cases of dementia identified, there were 38 (48.7%) cases of ADD and 32 (41.0%) cases of VAD. Of the remaining 8 cases, 3 cases had Parkinson’s disease dementia (PDD), 1 had probable Lewy body dementia (LBD), 1 had possible LBD, 2 had mixed alcohol induced dementia and ADD and 1 case of normal pressure hydrocephalus on clinical examination (who refused CT scanning). The 11
ACCEPTED MANUSCRIPT crude prevalence of ADD was 3.7% (95% CI 2.5 to 4.9) and of VAD was 2.9% (95% CI 1.9 to 3.9). The age-adjusted prevalence was 3.0% (95% CI 1.8 to 4.2) for ADD and 2.6% (95% CI 1.6 to 3.6) for VAD. The demographic and clinical characteristics of those with different dementia sub-types are presented in Table 2. Those with
RI PT
VAD had a significantly higher CSI-D discriminant function score than those with ADD (U = 275.0, z = -3.814, p < 0.001). A previous stroke was identified clinically in 24 cases, 23 with VAD and one with LBD. Of 29 cases who underwent CT scan, 15
SC
showed signs of small vessel disease and six revealed a previous infarct.
M AN U
Risk factors for VAD
The VAD and ADD groups were compared with regard to vascular and demographic risk factors, see Table 3. In univariate analysis, high total cholesterol and a previous diagnosis of diabetes mellitus were significantly more common in VAD than ADD. In multivariable analysis, only diabetes mellitus was associated with a significantly
TE D
greater odds of having VAD, rather than ADD.
EP
Discussion
AC C
Prevalence of dementia sub-types This study was part of one of the largest dementia prevalence studies conducted to date in SSA, and the first to be conducted in East Africa. The proportion of dementia cases with VAD was noticeably higher in our study population than seen in previous studies of dementia from SSA, all of which were conducted in West and Central Africa, see Table 1. Previous studies have tended to report rates of ADD three to four times higher than those of VAD. Ratios in black people of African origin from other world regions are similar. In our population, rates of VAD and ADD were 12
ACCEPTED MANUSCRIPT almost equal. There may be many reasons for the apparently higher contribution of VAD to the overall burden of dementia. It is thought that ADD and VAD exist on a continuum with a greater or lesser contribution from a range of factors influencing the final diagnosis.(27) Most previous studies of dementia have been in those aged 65
RI PT
years and over, whilst our population was aged 70 years and over. It is well
established that vascular risk factors are more common in older age groups and this is likely to translate into a more vascular picture when causes of dementia are
SC
investigated. However, data from European populations suggest that it is ADD, rather than VAD, that is largely responsible for the increase in prevalence of
M AN U
dementia with age. Rates of VAD increase much more slowly than ADD with increasing age.(9, 28) This trend is also thought to be apparent in SSA.(29) With continuing demographic transition, rates of ADD, relative to VAD, are likely to rise in Hai.
TE D
One possible reason for the higher prevalence of VAD seen in this study may be under-reporting of VAD in previous studies from SSA. Routine use of CT brain for diagnosis of dementia is rare in this setting. Indeed, of the studies listed in Table 1,
EP
only that of Hendrie et al in 1995 used CT brain to help In diagnosis and this was
AC C
only performed in around half of all cases from Ibadan.(3) Given the reliance on clinical signs and symptoms it is possible that VAD has been under-diagnosed in some previous studies from SSA. The population in Hai have relatively high rates of hypertension, when compared to other African communities.(30) Hypertension was notably more common in our population across all sub-types than reported in other studies of dementia from SSA, and this may be one of the major contributory factors to the high rates of VAD seen in Hai.(4-5, 31-32) This is supported by previous studies of stroke in this 13
ACCEPTED MANUSCRIPT population.(33) Age-standardised stroke incidence in two demographic surveillance sites in Tanzania (Hai and Dar-es-Salaam) were found to be similar to that reported in developed countries in Hai, and higher in Dar es Salaam. A previous stroke was recorded in almost a third of all dementia cases in the current study. Treating
RI PT
hypertension may therefore represent an important future public health intervention in dementia. In a resource poor setting, prevention of chronic diseases is likely to be much more cost effective than treatment. However, many health services in SSA
SC
have been designed primarily to treat acute medical conditions and a re-modelling of services would be required.(34) However, this may be the only way that the growing
M AN U
burden of chronic non-communicable diseases can be reduced.(35) Comparison with studies from other world regions
Rates of dementia and the contribution of ADD and VAD to the overall burden, in black populations living outwith Africa have been investigated. The Ibadan-
TE D
Indianapolis study found overall rates of dementia in African-Americans to be higher than in Nigeria, suggesting that lifestyle factors played a major role in the excess
EP
burden.(36) However, the contribution of VAD and ADD at both sites appeared similar.(3, 37) A study that investigated dementia rates in people living in London,
AC C
found an excess burden in people of Afro-Caribbean origin compared to Caucasians and concluded that this was due to an excess of vascular risk factors.(38) However, drawing comparisons between our East African population and black people living in the USA and UK, who are likely to be of predominantly West African origin, is difficult given differences in genetic make-up, lifestyle and diet across Africa.(39) VAD was thought to be more common in some Asian populations than in Europe and the USA.(40) A study comparing the prevalence of dementia sub-types in Japan across two decades reported VAD to be more common that ADD in the 1980s, but 14
ACCEPTED MANUSCRIPT that ADD was becoming comparatively more prevalent by the 1990s.(41) A systematic review of dementia prevalence in Japan published in 2012 suggested that this trend appears to be continuing in more recent decades.(42) Indeed, the ratio of ADD to VAD is now thought to be similar to that seen in the USA, with ADD
RI PT
becoming increasingly common in the older age bands. The review authors note that changes in diagnostic criteria, as well as demographic ageing and lifestyle
changes, may partly explain the apparent shift in the profile of dementia. It may be
SC
that, with demographic transition, our rural Tanzanian population will see a similar
M AN U
shift in the relative contribution of dementia sub-types to the overall disease burden. Risk factors for VAD
We were able to report on a number of vascular and demographic risk factors which may be thought to be more common in VAD than ADD and may have accounted for the high prevalence of VAD seen in this cohort. However, only a previous diagnosis
TE D
diabetes mellitus and high total cholesterol emerged as significant modifiable risk factors for VAD, when compared to ADD. Being overweight, hypertension and low
EP
HDL cholesterol levels were not significantly associated with a diagnosis of VAD compared to ADD. Obesity, hypertension, smoking and diabetes mellitus have been
AC C
noted as risk factors for dementia,(43) with hypertension noted as a specific risk factor for VAD.(44) In a prospective study of 1753 elderly Yoruba Nigerians, 120 developed dementia over the six year follow-up period, with hypertension at baseline significantly more common in those with dementia.(45) However, sub-type specific risk factor data from SSA are lacking. The low number of cases studied is likely to have limited the power of our analyses and this area requires further study.(46) Limitations
15
ACCEPTED MANUSCRIPT The main limitation of our study is that CT brain was not available for all people found to have dementia. Although this would have helped to provided information on dementia sub-types, due to distances involved and frailty of some participants this was felt to be inappropriate in some cases. However, where the diagnosis was in
RI PT
doubt diagnoses were reviewed by two independent old age psychiatrists. Another difficulty was the lack of hospital or medical documentation of stroke in many cases, and the resultant difficulty in obtaining clear evidence of dementia occurring within
SC
three months of stroke for a diagnosis of VAD. In many cases we were reliant on the accuracy of an informant history. In a number of subjects, the absence of small
M AN U
vessel disease or evidence of a previous stroke by CT scan helped rule out VAD, despite a history which was suggestive of VAD. However, relatives in most cases resided with the patient they were caring for and were asked about onset of dementia and previous strokes in different sections of the interview as open-ended
TE D
answers. As a result, we do not feel any systematic bias will have resulted. We were only able to obtain blood samples from 50 of the 78 people with dementia. Refusal to provide blood samples is common in rural SSA and we feel this would have been a
EP
limitation of any study conducted in this setting. Our data relating to diabetes
AC C
mellitus are based on self-reported previous diagnosis and are therefore likely to represent an under estimate. Finally, we acknowledge that the numbers identified with each dementia subtype were relatively small. Nevertheless, our study is one of the largest and most comprehensive studies of dementia sub-types conducted in SSA to date. The lack of previous data highlights the unique challenges in conducting epidemiological research in this low resource setting. Conclusions
16
ACCEPTED MANUSCRIPT Our data are unique in that they were collected from an East African population. The contribution of VAD to the overall dementia burden was high. This is likely to be due in part to high rates of vascular risk factors within the background population, most notably hypertension. As the population undergoes demographic transition the
RI PT
prevalence and profile of dementia is likely to change. Health services in SSA need
SC
to adapt to the changing health needs of their people.
Acknowledgements: We wish to acknowledge the help of all health care workers,
M AN U
village enumerators and medical supervisors who assisted in identification of cases, clinical assessment and completion of the census. We are especially grateful to patients, carers and family members for taking part in the study. Conflicts of Interest: There were no conflicts of interest.
TE D
Funding: This study was part funded by a British Geriatric Society SpR start up grant and an Academy of Medical Sciences (UK) Clinical Lecturer start up grant. Role of the Funding Source: The sponsors of this study had no role in designing
EP
the study; in the collection, analysis, and interpretation of data; in the writing of the
AC C
report; or in the decision to submit the paper for publication.
Contributions
Design/conception – Richard Walker, Catherine Dotchin, Stella-Maria Paddick, Anna Longdon, Paul Chaote, Ahmed Jusabani. Literature search - Richard Walker, Catherine Dotchin, Stella-Maria Paddick. Data collection – Stella-Maria Paddick, Anna Longdon, Aloyce Kisoli, Ahmed Iqbal. Data analysis - Richard Walker, William K. Gray, Catherine Dotchin, Stella-Maria Paddick. Interpretation of results - Richard 17
ACCEPTED MANUSCRIPT Walker, Catherine Dotchin, William Gray, Stella-Maria Paddick, Anna Longdon, Andrew Teodorczuk, Julian Hughes. Writing of paper - Richard Walker, Catherine Dotchin, William Gray, Stella-Maria Paddick, Anna Longdon, Andrew Teodorczuk,
AC C
EP
TE D
M AN U
SC
RI PT
Julian Hughes, Aloyce Kisoli, Paul Chaote, Ahmed Jusabani and Ahmed Iqbal.
18
ACCEPTED MANUSCRIPT Figure 1. Diagram of the recruitment and randomisation process Census population: n = 161,119
Died: n = 77
Too young: n = 16
Duplicates: n=4
M AN U
Moved away: n = 64
SC
Census population aged 70 and over in the 6 randomly selected villages: n = 1277
RI PT
Census population of 6 randomly selected villages: n = 34,078
Not found: n =8
TE D
Actual population from census: n = 1108
Final phase I study population: n = 1198
EP
Refused: n = 62
Identified by enumerators postcensus: n = 152
AC C
Probable dementia by CSI-D: n = 184
Follow-up in phase II: n = 168 (91.3%)
Possible dementia by CSI-D: n = 104
No dementia by CSI-D: n = 910
Randomly selected for phase II: n = 56 (53.8%)
Randomly selected for phase II: n = 72 (7.9%)
19
ACCEPTED MANUSCRIPT
Table 1. Dementia sub-types reported in previous sub-Saharan Africa studies Setting
Diagnostic Criteria
Imaging Available
Reported sub-types
Prevalence
Ibadan, Nigeria (urban)
DSM-III-R and ICD-10 for dementia
Yes. Only 10 dementia patients due to strikes/logistical issues at time of study.
ADD, 18 (64.3%)
RI PT
Hendrie et al, 1995(3)
Design
NINCDS-ADRDA for ADD
SC
ICD-10 for VAD.
Hendrie et al, 2001(37)
Incidence
Ibadan, Nigeria (urban)
DSM-IV and ICD-10 for dementia
NINCDS-ADRDA for ADD.
Guerchet et al, 2010(5)
Prevalence
Bangui, Central African Republic (urban)
NINCDS-ADRDA for ADD
DSM-IV for dementia
Brazzaville, Congo (urban)
Lewy body dementia, 1 (3.6%) Depression related dementia, 1 (3.6%)
Yes.
ADD, 62 (88.6%)
28 cognitive impairment,
VAD, 4 (5.7%)
15 dementia
PDD, 1(1.4%)
19 controls
Dementia with severe chronic illness, 2 (2.9%)
No
Other subtypes based on clinical history
EP
Prevalence
Djidja, Benin (rural)
Unspecified dementia, 1 (1.4%) ADD, 11 (84.6%) VAD, 1 (7.7%) Dementia with progressive aphasia, 1 (7.7%)
No
AC C
Guerchet et al, 2010(5)
Prevalence
TE D
ICD-10 for VAD.
Guerchet et al, 2009(31)
M AN U
ICD-10 for secondary dementia
VAD, 8 (28.6%)
NINCDS-ADRDA for ADD
ADD, 33 (82.5%) VAD, 7 (17.5%)
Hachinski score and clinical history for VAD DSM-IV for dementia
No
NINCDS-ADRDA for ADD
ADD, 24 (68.6%) VAD, 11 (31.4%)
20
ACCEPTED MANUSCRIPT
Hachinski score and clinical history for VAD Prevalence
Cotonou, Benin (urban)
DSM-IV for dementia
No
RI PT
Paraiso et al, 2011(32)
NINCDS-ADRDA for ADD.
Yusuf et al, 2011(47)
Prevalence
Northern Nigeria (rural)
ICD and DSM-IV for dementia
No
DSM-IV for ADD
M AN U
ICD10 for VAD. McKeith for dementia with Lewy bodies Lund for dementia Ibadan, Nigeria (urban)
ICD-10 and DSM-IV for dementia
No
TE D
Incidence
VAD, 9 (21.4%) Parkinson’s disease dementia, 1 (2.4%) ADD, 6 (66.7%) VAD, 2 (22.2%) Frontotemporal dementia, 1 (11.1%)
frontotemporal
EP
Ogunniyi et al, 2011(45)
SC
NINDS-AIREN for VAD
ADD, 32 (76.2%)
ADD, 99 (82.5%) VAD, 11 (9.2%) Dementia with depression, 5 (4.2%) Dementia with Parkinson’s disease, 2 (1.7%) Other causes, 3 (2.5%)
AC C
ADD Alzheimer’s disease dementia, VAD Vascular dementia, PDD Parkinson’s disease dementia, DSM Diagnostic and statistical manual of mental disorders, ICD International Classification of Disease, NINCDS-ADRDA National Institute of Neurological and Communicative Disorders and Stroke - Alzheimer's Disease and Related Disorders Association, NINDS-AREN National Institute of Neurological Disorders and Stroke Association Internationale pour la Recherche et l'Enseignement en Neurosciences
21
ACCEPTED MANUSCRIPT
Table 2. Demographic and clinical characteristics of dementia cases by sub-type Other (n = 8)
No dementia (n = 218)
Median age in years (IQR)
85.0 (80.0 to 91.0)
84 (76.25 to 90.0)
84.5 (75.5 to 94.25)
78.0 (74.0 to 86.0)
Gender
30 female (78.9%)
21 female (65.6%)
5 female (62.5%)
145 female (66.5%)
Median CSI-D discriminant function score (IQR)
0.26 (0.21 to 0.49)
0.47 (0.34 to 0.75)
0.44 (0.22 to 0.68)
0.17 (0.08 to 0.26)
Diabetes previously diagnosed
3 (7.9%)
9 (28.1%)
1 (12.5%)
39 (17.9%)
Median total cholesterol in mmol/L (IQR)*
4.5 (4.3 to 4.8)
4.9 (4.0 to 5.6)
4.8 (4.5 to 6.5)
-
Median HDL cholesterol in mmol/L (IQR)*
1.2 (0.9 to 1.4)
1.2 (1.0 to 1.5)
1.1 (1.1 to 1.3)
-
Median total cholesterol/HDL cholesterol ratio (IQR)*
3.8 (3.3 to 4.2)
4.1 (3.3 to 4.7)
4.4 (3.9 to 6.3)
-
Median body mass index (IQR)**
19.9 (17.4 to 25.1)
21.5 (18.7 to 23.5)
21.5 (18.5 to 28.9)
20.8 (18.4 to 23.5)
Mean systolic blood pressure in mmHg (SD)
162.1 (36.631)
175.8 (30.450)
167.0 (17.696)
161.5 (33.831)
Mean diastolic blood pressure in mmHg (SD)
84.9 (17.212)
94.3 (19.157)
81.5 (11.796)
87.5 (16.508)
M AN U
TE D
RI PT
VAD (n = 32)
SC
ADD (n = 38)
EP
IQR = inter-quartile range, SD standard deviation, HDL high density lipoprotein * Based on blood samples taken from 23 subjects with ADD, 22 with VAD and 5 with other causes of dementia
AC C
** Based on 34 subjects with ADD, 22 with VAD, 7 with other causes of dementia and 201 subjects without dementia
22
ACCEPTED MANUSCRIPT
Table 3. Univariate analysis of VAD risk factors Variable
Univariate odds ratio (95% CI, significance)
Age band
1
75-79 years
0.28 (0.04 to 2.04, p = 0.208)
80-84 years
0.25 (0.04 to 1.70, p = 0.157)
85 years and over
0.23 (0.04 to 1.28, p = 0.093)
Male gender
1.96 (0.68 to 5.72, p = 0.215)
Diabetes previously diagnosed
4.57 (1.12 to 18.67, p = 0.035)
Hypertension
2.49 (0.77 to 8.06, p = 0.127)
M AN U
High total cholesterol*
SC
RI PT
70-74 years
3.29 (0.83 to 12.98, p = 0.089)
Low HDL cholesterol*
1.32 (0.36 to 4.82, p = 0.672)
High total cholesterol to HDL cholesterol
ratio* High body mass index**
1.05 (0.19 to 5.87, p = 0.953)
TE D
0.51 (0.12 to 2.19, p = 0.368)
CI confidence interval, HDL high density lipoprotein VAD coded 1, ADD coded 0
* Based on blood samples taken from 23 subjects with ADD and 22 with VAD
AC C
EP
** Based on 34 subjects with ADD and 22 with VAD
23
ACCEPTED MANUSCRIPT
References 1.
Ferri CP, Prince M, Brayne C, et al: Global prevalence of dementia: a
2.
RI PT
Delphi consensus study. Lancet 2005; 366:2112-2117 Llibre Rodriguez JJ, Ferri CP, Acosta D, et al: Prevalence of dementia in
Latin America, India, and China: a population-based cross-sectional survey.
3.
SC
Lancet 2008; 372:464-474
Hendrie HC, Osuntokun BO, Hall KS, et al: Prevalence of Alzheimer's
M AN U
disease and dementia in two communities: Nigerian Africans and African Americans. Am J Psychiatry 1995; 152:1485-1492 4.
Ochayi B,Thacher TD: Risk factors for dementia in central Nigeria. Aging
Ment Health 2006; 10:616-620
Guerchet M, M'Belesso P, Mouanga AM, et al: Prevalence of dementia in
TE D
5.
elderly living in two cities of Central Africa: the EDAC survey. Dement Geriatr Cogn Disord 2010; 30:261-268
Gureje O, Ogunniyi A,Kola L: The profile and impact of probable dementia
EP
6.
in a sub-Saharan African community: Results from the Ibadan Study of Aging. J
7.
AC C
Psychosom Res 2006; 61:327-333 Wimo A,Prince M: World Alzheimer Report 2010: The Global Economic
Impact of Dementia, London, UK, Alzheimer Disease International, 2010 8.
Prince M, Bryce R, Albanese E, et al: The global prevalence of dementia:
a systematic review and metaanalysis. Alzheimers Dement 2013; 9:63-75 e62
24
ACCEPTED MANUSCRIPT
9.
Lobo A, Launer LJ, Fratiglioni L, et al: Prevalence of dementia and major
subtypes in Europe: A collaborative study of population-based cohorts.
10.
RI PT
Neurologic Diseases in the Elderly Research Group. Neurology 2000; 54:S4-9 Kalaria RN, Maestre GE, Arizaga R, et al: Alzheimer's disease and
vascular dementia in developing countries: prevalence, management, and risk
11.
SC
factors. Lancet Neurol 2008; 7:812-826
Longdon AR, Paddick SM, Kisoli A, et al: The prevalence of dementia in
M AN U
rural Tanzania: a cross-sectional community-based study. Int J Geriatr Psychiatry 2013; 28:728-737 12.
Adult Morbidity and Mortality Project (AMMP): Policy Implications of Adult
Morbidity and Mortality; final report, Dar-es-Salaam, Tanzanian Ministry of
13.
TE D
Health, 2004
Hall KS, Hendrie HC, Brittain HM, et al: The Development of a Dementia
Screening Interview in 2 Distinct Languages. International Journal of Methods in
14.
EP
Psychiatric Research 1993; 3:1-28
Chen CH, Mizuno T, Elston R, et al: A comparative study to screen
AC C
dementia and APOE genotypes in an ageing East African population. Neurobiol Aging 2010; 31:732-740 15.
Prince M, Acosta D, Chiu H, et al: Dementia diagnosis in developing
countries: a cross-cultural validation study. Lancet 2003; 361:909-917
25
ACCEPTED MANUSCRIPT
16.
American Psychiatric Association: Diagnostic and statistical manual of
mental disorders, 4th edition, Washington DC, American Medical Association,
17.
RI PT
1994 Hachinski V, Iadecola C, Petersen RC, et al: National Institute of
Neurological Disorders and Stroke-Canadian Stroke Network vascular cognitive
18.
SC
impairment harmonization standards. Stroke 2006; 37:2220-2241
McKeith IG, Galasko D, Kosaka K, et al: Consensus guidelines for the
M AN U
clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international workshop. Neurology 1996; 47:1113-1124 19.
Petersen RC, Smith GE, Waring SC, et al: Mild cognitive impairment:
clinical characterization and outcome. Arch Neurol 1999; 56:303-308 Rogoza AN, Pavlova TS,Sergeeva MV: Validation of A&D UA-767 device
TE D
20.
for the self-measurement of blood pressure. Blood Press Monit 2000; 5:227-231 21.
Williams B, Poulter NR, Brown MJ, et al: Guidelines for management of
EP
hypertension: report of the fourth working party of the British Hypertension
22.
AC C
Society, 2004-BHS IV. J Hum Hypertens 2004; 18:139-185 Third Report of the National Cholesterol Education Program (NCEP)
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 2002; 106:3143-3421 23.
WHO MONICA Project: The World Health Organization MONICA Project
(monitoring trends and determinants in cardiovascular disease): a major
26
ACCEPTED MANUSCRIPT
international collaboration. WHO MONICA Project Principal Investigators. J Clin Epidemiol 1988; 41:105-114 Paraiso MN, Houinato D, Guerchet M, et al: Validation of the use of
RI PT
24.
historical events to estimate the age of subjects aged 65 years and over in Cotonou (Benin). Neuroepidemiology 2010; 35:12-16
Gao S, Hui SL, Hall KS, et al: Estimating disease prevalence from two-
SC
25.
phase surveys with nonresponse at the second phase. Stat Med 2000; 19:2101-
26.
M AN U
2114
Ahmad OB, Boschi-Pinto C, Lopez AD, et al: Age standardization of rates:
A new WHO standard (GPE discussion paper Series: No. 31), Geneva, World Health Organization, 2001
Prince M,Jackson J: World Alzheimer Report 2009, London, UK,
TE D
27.
Alzheimer Disease International, 2009 28.
Fratiglioni L, Launer LJ, Andersen K, et al: Incidence of dementia and
EP
major subtypes in Europe: A collaborative study of population-based cohorts.
29.
AC C
Neurologic Diseases in the Elderly Research Group. Neurology 2000; 54:S10-15 George-Carey R, Adeloye D, Chan KY, et al: An estimate of the
prevalence of dementia in Africa: A systematic analysis. J Glob Health 2012; 2:20401 30.
Dewhurst MJ, Dewhurst F, Gray WK, et al: The high prevalence of
hypertension in rural-dwelling Tanzanian older adults and the disparity between
27
ACCEPTED MANUSCRIPT
detection, treatment and control: a rule of sixths? J Hum Hypertens 2013; 27:374-380 Guerchet M, Houinato D, Paraiso MN, et al: Cognitive impairment and
RI PT
31.
dementia in elderly people living in rural Benin, west Africa. Dement Geriatr Cogn Disord 2009; 27:34-41
Paraiso MN, Guerchet M, Saizonou J, et al: Prevalence of Dementia
SC
32.
among Elderly People Living in Cotonou, an Urban Area of Benin (West Africa).
33.
M AN U
Neuroepidemiology 2011; 36:245-251
Walker R, Whiting D, Unwin N, et al: Stroke incidence in rural and urban
Tanzania: a prospective, community-based study. Lancet Neurol 2010; 9:786792
Reidpath DD,Allotey P: The burden is great and the money little: Changing
TE D
34.
chronic disease management in low- and middle-income countries. J Glob Health 2012; 2:20301
Di Cesare M, Khang YH, Asaria P, et al: Inequalities in non-communicable
EP
35.
36.
AC C
diseases and effective responses. Lancet 2013; 381:585-597 Hendrie HC, Murrell J, Gao S, et al: International studies in dementia with
particular emphasis on populations of African origin. Alzheimer Dis Assoc Disord 2006; 20:S42-46 37.
Hendrie HC, Ogunniyi A, Hall KS, et al: Incidence of dementia and
Alzheimer disease in 2 communities: Yoruba residing in Ibadan, Nigeria, and African Americans residing in Indianapolis, Indiana. JAMA 2001; 285:739-747 28
ACCEPTED MANUSCRIPT
38.
Livingston G, Leavey G, Kitchen G, et al: Mental health of migrant elders--
the Islington study. Br J Psychiatry 2001; 179:361-366 Tishkoff SA, Reed FA, Friedlaender FR, et al: The genetic structure and
RI PT
39.
history of Africans and African Americans. Science 2009; 324:1035-1044 40.
Suh GH,Shah A: A review of the epidemiological transition in dementia--
SC
cross-national comparisons of the indices related to Alzheimer's disease and vascular dementia. Acta Psychiatr Scand 2001; 104:4-11
Kiyohara Y, Yoshitake T, Kato I, et al: Changing patterns in the prevalence
M AN U
41.
of dementia in a Japanese community: the Hisayama study. Gerontology 1994; 40 Suppl 2:29-35 42.
Dodge HH, Buracchio TJ, Fisher GG, et al: Trends in the prevalence of
43.
TE D
dementia in Japan. Int J Alzheimers Dis 2012; 2012:956354 Whitmer RA, Sidney S, Selby J, et al: Midlife cardiovascular risk factors
and risk of dementia in late life. Neurology 2005; 64:277-281 Sharp SI, Aarsland D, Day S, et al: Hypertension is a potential risk factor
EP
44.
669 45.
AC C
for vascular dementia: systematic review. Int J Geriatr Psychiatry 2011; 26:661-
Ogunniyi A, Lane KA, Baiyewu O, et al: Hypertension and incident
dementia in community-dwelling elderly Yoruba Nigerians. Acta Neurol Scand 2011; 124:396-402
29
ACCEPTED MANUSCRIPT
46.
Davis DH, Kreisel SH, Muniz Terrera G, et al: The epidemiology of
delirium: challenges and opportunities for population studies. Am J Geriatr
47.
RI PT
Psychiatry 2013; 21:1173-1189 Yusuf AJ, Baiyewu O, Sheikh TL, et al: Prevalence of dementia and
dementia subtypes among community-dwelling elderly people in northern
AC C
EP
TE D
M AN U
SC
Nigeria. Int Psychogeriatr 2011; 23:379-386
30
