Chair: Anne Marie Mork Rokstad
O3.2.1 Educational level and its Association with the domains of the Montreal Cognitive Assessment Test
Miguel German Borda Borda1, Carlos Reyes-Ortiz 2, Mario Ulises Pérez-Zepeda 3, Daniela Patiño-Hernandez1, Camilo Gomez1, Carlos Cano1
1Pontificia Universidad Javeriana, Colombia, 2 Texas University, USA, 3 Instituto Nacional De Geriatría, Mexico
To explore the association between educational level and the scores obtained of each of the domains of the Montreal Cognitive Assessment test(MoCA).
This is a secondary analysis of the SABE/2012 Bogotá survey, a cross-sectional study including 2000 subjects aged ≥60years. The MoCA score was the dependent variable and was stratified by cognitive domains. Educational level was assessed through years of formal education. Age and sex were also included to adjust the multivariate models. Bivariate analyses and fitted logistic regression models were employed for analyzing association between these variables.
The proportion of incorrect answers increased as schooling years decreased and as age increased. In the multivariate analysis, visuospatial and executive function were the most affected domains. Educational level did not influence the short memory-recall task as it did with other domains.
It seems convenient to consider that performance in most domains of the MoCA is influenced by education. Therefore, low scores on these tasks could lead to low MoCA scores and thus to bias and over diagnosis of cognitive impairment in patients with lower educational levels. Results in memory-recall are not affected much by education and applying it separately could be useful in patients with low educational level in whom we suspect memory impairment.
This is also important in countries as Norway with high rates of immigration from developing and underdeveloped countries.
O3.2.2 Mortality prediction tables for dementia based on 50,076 patients from the Swedish Dementia Registry
Miriam Haaksma1,2,3, René Melis2, Marcel Olde Rikkert2,4, Maria Eriksdotter5, Sara Garcia-Ptacek2, 4
1 Radboud University Medical center, The Netherlands, 2 Radboudumc Alzheimer Center, 3 Radboud Institute for Health Sciences, The Netherlands, 4 Donders Institute, The Netherlands, 5 Karolinska Institutet, Sweden
Dementia is accompanied by increased morbidity and mortality. The prognosis of dementia is highly heterogeneous, causing a lot of uncertainty among patients and caregivers. Current prognostic models for dementia often focus on disease-related characteristics, ignoring other patient traits. Moreover, these models rarely result in applicable tools for clinicians. This study aims to develop usable risk tables to predict mortality in dementia patients.
We used data from 50,076 incident dementia patients aged 65 and over, obtained through the Swedish Dementia Registry. We built Cox proportional hazards models to predict time-to-death. Demographics, disease characteristics, e.g. cognitive performance measured by Mini-Mental State Examination (MMSE), and patient traits such as the Charlson Comorbidity Index (CCI) and medication use, were included as predictors.
After a maximum follow-up time of 9.7 years, 20,828 patients had died. Median (IQR) time-to-death was 4.8 (2.6-7.5) years. The strongest predictors of time-to-death were CCI (HR: 1.131, 95%-CI: 1.123-1.139) and MMSE (HR: 0.959, 95%-CI: 0.957-0.961), next to age (HR: 1.067, 95%-CI: 1.065-1.070) and gender (HR for men versus women: 1.467, 95%-CI: 1.425-1.509). As survival probabilities differed for patients from primary care versus memory clinics, we created separate risk tables for each diagnostic setting. These models yielded c-indexes of 0.70-0.72.
Time-to-death in dementia patients can be predicted with good accuracy using five easily obtainable patient characteristics. We constructed tables with 3-year survival probabilities stratified by gender and diagnostic setting. These tables may aid Nordic clinicians and patients in shared decision making and advanced care planning.
O3.2.3 Vital exhaustion and incidence of dementia: Results from the Copenhagen City Heart Study
Sabrina Islamoska1, Åse Marie Hansen1, Denmark; Anne Helene Garde2, Matias Brødsgaard Grynderup1, Finn Gyntelberg2, Kazi Ishtiak-Ahmed1, Erik Lykke Mortensen1, Thien Kieu Thi Phung1, Eva Prescott4, Naja Hulvej Rod1, Eszter Török1, Gunhild Waldemar4,; Kirsten Nabe-Nielsen1,
1 University of Copenhagen, Denmark, 2 The National Research Centre for the Working Environment, Denmark, 3 Bispebjerg Hospital, Denmark, 4 Danish Dementia Research Centre, Rigshospitalet, Copenhagen, Denmark
Psychological distress has been linked to dementia through neurologic and cardiovascular mechanisms. Psychological distress can be operationalized as vital exhaustion (VE) which is a mental state in which individuals complain about unusual fatigue, increased irritability and feelings of demoralization. As an indicator of psychological distress, VE might be a risk factor for dementia. We investigated whether VE was associated with incidence of dementia in later life.
The study population consisted of participants from the Copenhagen City Heart Study (CCHS) in 1991-1994. The 6718 participants were followed from time of enrolment until death, emigration, dementia diagnosis, or end of follow-up in 2015. VE was measured by 17 different symptoms from the Maastricht Questionnaire (symptom score: 0-17). Information on dementia diagnoses was drawn from Danish national registers. We applied Poisson regression and adjusted for potential confounding. Age-stratified analyses were performed to compare participants reporting VE in midlife (<55 years) and late-life (≥55 years).
During an average follow-up of 17 years, we found that for every additional VE symptom reported, the dementia incidence increased by 3% (Incidence Rate Ratio = 1.03; 95% Confidence Interval: 1.01-1.05), when adjusting for age, sex, marital status and educational level. The age-stratified analyses showed approximately the same estimates for participants <55 years and ≥55 years.
The more VE symptoms reported, the higher the incidence of dementia. However, we cannot preclude that VE could be a consequence of preclinical dementia pathology among the oldest participants.
O3.2.4 Alzheimer related determinants, not co-morbidity, are associated with mortality in young AD patients
Hanneke Rhodius1, Hilkka Liedes2, Ted Koene1, Afina Lemstra1, Charlotte Teunissen1, Frederik Barkhof13, Philip Scheltens1, Mark van Gils2, Jyrki Lötjönen3, Wiesje van der Flier1
1VU University Medical Center, Amsterdam, The Netherlands, 2 VTT Technical Research Centre of Finland,
Average survival after dementia diagnosis varies considerably. We aimed to identify clinical measures that are associated with risk of mortality in patients with dementia due to Alzheimer’s disease (AD). To combine multiple determinants and to facilitate use in clinical practice, we used a clinical decision support system (CDSS) based on the disease state index (DSI) classifier.
We included 616 patients (50% female, mean age 67±8 years, mean MMSE 22±3 ) with dementia due to AD from the Amsterdam Dementia Cohort. Information on mortality was obtained from the Dutch Municipal Register. We assessed associations of baseline clinical data, including co-morbidity, neuropsychology, MRI and cerebrospinal fluid (CSF) biomarkers, with mortality. In addition, we built a multivariate model based on the DSI to provide a prognosis of mortality.
After a mean of 4.9±2.0 years, 213(35%) patients had died. Cox proportional hazards models showed that higher age and male gender, and after adjustment for age and sex, worse scores on neuropsychological tests for memory and executive functioning and more severe atrophy on MRI were associated with increased risk of mortality. Effect sizes ranged from HR (95%CI) 1.1(1.0-1.3) to 2.4(1.6-3.5). APOE e4, co-morbidity and CSF biomarkers were not associated with mortality. When we used the DSI classifier to combine these determinants, classification was modest. The DSI classifier did provide visualization of how each variable contributes to the prediction.
In this memory clinic cohort, AD-related determinants were associated with mortality. A CDSS was instrumental in translating these findings to clinical practice.
O3.2.5 Does physical activity moderate the association between ApoE and cognitive decline in older adults?
Najada Stringa, Natasja van Schoor, Yuri Milaneschi, Hannie Comijs, Martijn Huisman
VU University Medical Center, Amsterdam, The Netherlands,
Aim: To determine whether physical activity(PA) moderates the association between ApoE ɛ4 and cognitive decline in older individuals. Methods: A total of 1742 individuals 55 years and older, from the Longitudinal Aging Study of Amsterdam (LASA) were followed-up for 9 years. Apo E phenotyping was done at the Immunochemistry Laboratory of VUmc. PA as assessed by LASA PA Questionnaire was analyzed in tertiles. Cognitive decline was coded as a decrease of 3 or more points in Mini-Mental State Examination (MMSE) in the next follow up measurement. We fitted logistic regression models using Generalized Estimating Equations with exchangeable structure to take into account the within-subject correlation in repeated measurements.
Findings: The mean age at baseline was 69.8 (SD 7.86) years and 26.9% of the participants had at least one Apo E ɛ4 allele. After adjustment for age, gender, education, depression and presence of chronic disease, the presence of Apo E ɛ4 was associated with higher odds of cognitive decline (OR 1.28, 95%CI (1.05-1.55)). There was no significant effect of PA on cognitive decline and the interaction term between Apo E ɛ4 and PA was not statistically significant.
Conclusions: In the LASA sample, there is no indication for a moderation effect of physical activity in the association between Apo E ɛ4 and cognitive decline. Careful replication is needed before drawing reliable conclusions from gene-environment interaction studies. We are currently replicating these results in two similar population-based cohort studies (Rotterdam Study and InChianti). The pooled results will be presented during the congress.
O3.2.6 Acetylcholinesterase inhibitors and risk of stroke and death in persons with dementia
Edwin Tan1, Kristina Johnell1, Sara Garcia-Ptacek1, Miriam Haaksma1, Johan Fastbom1, J Simon Bell2, Maria Eriksdotter1
1Karolinska Institutet, Sweden, 2 Monash University, Australia
People with dementia have a higher incidence of stroke than people without dementia. Acetylcholinesterase inhibitors (AChEIs) are used in the symptomatic treatment of dementia. It has been hypothesized that AChEIs may have anti-inflammatory and endothelial protective effects which lowers stroke risk. We investigated the association between AChEI use and risk of ischemic stroke and death in people with dementia.
This was a cohort study based on 44288 people diagnosed with dementia who were registered in the Swedish Dementia Registry (SveDem) from 2007 – 2014. Data on AChEI use was linked to diagnosed ischemic strokes and death using nationwide registers. Propensity-score matched competing risk regression models were used to compute hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between time-dependent AChEI use and risk of stroke and death.
During a median follow-up period of 715 [quartiles 345 – 1300] days, 2084 people had a stroke and 11276 died. In comparison with matched controls, those who used AChEIs had a lower risk of stroke (HR: 0.87, 95%-CI: 0.77 – 0.98) and all-cause death (HR: 0.77, 95%-CI: 0.73 – 0.81). After considering death as a competing risk, high doses (≥1.33 defined daily doses) of AChEI remained significantly associated with reduced stroke risk (subdistribution HR: 0.78, 95%-CI: 0.66 – 0.93). Subgroup analyses in those with Alzheimer’s disease produced similar findings.
The use of AChEIs in people with dementia may be associated with reduced risk of ischemic stroke and death. These results call for a closer examination of the cardiovascular effects of AChEIs.
O3.2.7 How does exercise interact with fall-related drugs on falls among persons with Alzheimer disease?
Niko Perttila1, Hanna Öhman1, Timo Strandberg2, Hannu Kautiainen1, Minna Raivio1, Marja-Liisa Laakkonen1, Niina Savikko1, Reijo Tilvis1, Kaisu Pitkala1
1 University of Helsinki, Finland, 2 University of Oulu, Finland
Background: Certain drugs increase fall-risk. No study has investigated the interaction of exercise and fall-related drugs (FRDs) on falls among participants with Alzheimer’s Disease (AD). Aim is to investigate how exercise intervention and FRDs interact on fall-risk among participants with AD.
Methods: In FINALEX trial, community-dwelling persons with AD received either home-based or group-based exercise twice weekly for one year (n=129); the control group received normal care (n=65). Number of falls were based on spouses’ fall diaries. We examined Incidence Rate Ratios (IRRs) for falls among non-users and users of various FRDs in intervention and control groups separately. P-values were adjusted for age, sex, physical functioning.
Research results: Whereas there was no difference in the number of falls among those without antihypertensive drug(s) between the intervention and control groups, in intervention group with antihypertensive(s) IRR was 0.5 falls/person-year [95%CI 0.4-0.6] and among controls 1.5 falls/person-year [95%CI 1.2-1.8] (p<0.001 for group, p=0.067 for medication, p<0.001 for interaction). Neither there was difference in falls between the groups among participants without psychotropics. Among participants using psychotropic(s) the intervention group had IRR 0.7 falls/person-year [95%CI 0.6-0.9] whereas the respective figure for the control group was 2.0, [95%CI 1.6-2.5] (p<0.001 for group, p=0.071 for medication, p<0.001 for interaction). Similar interaction was not found in respect of drugs with anticholinergic properties (p<0.001 for group, p=0.014 for medication, p=0.97 for interaction)).
Conclusion: Exercise has a potential to modify the risk for falls among people with dementia using antihypertensive(s) and psychotropic(s).