Neuropsychiatric symptoms in Alzheimer’s disease : measurement, follow-up and associations with activities of daily living

(1)

Dissertations in Education, Humanities, and Theology

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

TONI SAARI

Neuropsychiatric symptoms in Alzheimer’s disease:

measurement, follow-up and

associations with activities of

daily living

(2)

(3)

Neuropsychiatric symptoms in Alzheimer’s

disease: measurement, follow-up and

associations with activities of daily living

(4)

(5)

Toni Saari

Neuropsychiatric symptoms in Alzheimer’s disease: measurement, follow-up and associations with activities of daily living

Publications of the University of Eastern Finland Dissertations in Education, Humanities, and Theology

No 173

University of Eastern Finland Kuopio

2021

(6)

Grano Oy Jyväskylä, 2021 Editor: Päivi Atjonen

Sales: University of Eastern Finland Library ISBN: 978-952-61-3800-8 (print)

ISBN: 978-952-61-3801-5 (PDF) ISSNL: 1798-5625

ISSN: 1798-5625 ISSN: 1798-5633 (PDF)

(7)

Saari, Toni

Neuropsychiatric symptoms in Alzheimer’s disease: measurement, follow-up and associations with activities of daily living

Kuopio: Itä-Suomen yliopisto, 2021

Publications of the University of Eastern Finland

Dissertations in Education, Humanities, and Theology; 173 ISBN: 978-952-61-3800-8 (print)

ISSNL: 1798-5625 ISSN: 1798-5625

ISBN: 978-952-61-3801-5 (PDF) ISSN: 1798-5633 (PDF)

ABSTRACT

Alzheimer’s disease (AD) is the most common cause of dementia. The core features of the disease are decline in cognitive functions and activities of daily living (ADLs). However, neuropsychiatric symptoms (NPS), such as depression, apathy and psychosis are also common in AD. NPS associate with worse quality of life and ADLs for the individuals with AD, but also with increasing costs of treatment and burden for caregivers. NPS, however, have received less attention in research than cognitive symptoms.

Conflicting results have been reported for associations between NPS and ADLs. The psychometric evidence for commonly used NPS rating scales is limited, and both self- and informant-reports of NPS are used although they provide discrepant information. These factors together complicate accurate detection of NPS and understanding their implications for functional impairment.

This study had three aims. First, the associations between NPS and ADLs were examined in a five-year follow-up of people with AD. Next, network structures of affective symptoms in AD, as rated by the person with AD and an informant, were investigated. Finally, the psychometric properties of the most widely used NPS rating scale, the Neuropsychiatric Inventory (NPI), was reviewed. The study used data of people with AD from the University of

(8)

Eastern Finland’s ALSOVA study and the US-based NACC data set. Database searches were carried out for the review.

Apathy, aberrant motor behavior and appetite disturbances associated with worse basic ADLs. In addition to these symptoms, delusions also associated with worse instrumental ADLs. Informant-rated depression, apathy and anxiety were weakly associated with the ratings of the person with AD regarding these symptoms. Instead, self-rated feelings of helplessness and worthlessness were among the most central symptoms in the network.

The psychometric review of the NPI established that facets of reliability are the measure’s major strengths. However, validity analyses of the measure were limited, leading to uncertainty about which constructs are captured by the NPI.

The results of this study highlight that different NPS relate to ADL impairment. Informant ratings may not capture the most crucial affective symptoms of the person with AD. In the future, more comprehensive understanding of NPS is needed. Advances in theory and adherence to good practices in psychometrics lay the groundwork for developing and evaluating NPS rating scales. In turn, more refined measurement could benefit planning treatment for NPS.

Keywords: Alzheimer’s disease, neuropsychiatric symptoms, activities of daily living, measurement, psychometrics, network analysis, longitudinal study

(9)

Saari, Toni

Alzheimerin taudin neuropsykiatriset oireet: mittaus, seuranta ja yhteydet päivittäisiin toimintoihin

Kuopio: Itä-Suomen yliopisto, 2021

Publications of the University of Eastern Finland

Dissertations in Education, Humanities, and Theology; 173 ISBN: 978-952-61-3800-8 (nid.)

ISSNL: 1798-5625 ISSN: 1798-5625

ISBN: 978-952-61-3801-5 (PDF) ISSN: 1798-5633 (PDF)

TIIVISTELMÄ

Alzheimerin tauti (AT) on yleisin dementiaan johtava muistisairaus. Taudin keskeisiä piirteitä ovat tiedonkäsittelyn ja päivittäisten toimintojen heikkene- minen. Kuitenkin neuropsykiatriset oireet, kuten masennus, apaattisuus, ja psykoottiset oireet ovat yleisiä AT:ssa. Näillä oireilla on negatiivinen yhteys elämänlaatuun ja päivittäisiin toimintoihin AT:a sairastavilla henkilöillä. Lisäk- si ne ovat yhteydessä omaisten suurempaan kuormitukseen sekä AT:n hoito- kustannuksiin. Neuropsykiatrista oireistoa on kuitenkin tutkittu vähemmän kuin tiedonkäsittelyn ongelmia.

Tutkimustieto neuropsykiatristen oireiden ja päivittäisten toimintojen vä- lisistä yhteyksistä on ristiriitaista. Neuropsykiatristen oireiden kyselyistä on usein myös rajallisesti psykometrista tietoa, ja niitä käytetään sekä potilaan että omaisen raportoimina vaikka eri menetelmillä saadaan eriävää tietoa.

Nämä tekijät yhdessä hankaloittavat neuropsykiatristen oireiden tarkkaa tun- nistamista sekä niiden toimintakykyä rajoittavien piirteiden ymmärtämistä.

Tutkimuksella oli kolme tavoitetta. Ensimmäisenä tavoitteena oli tarkas- tella neuropsykiatristen oireiden yhteyttä päivittäiseen toimintakykyyn AT:a sairastavilla henkilöillä viiden vuoden seurannassa. Lisäksi tutkimuksessa tar- kasteltiin AT:hen liittyviä mielialaoireita kartoittavien itsearviointi- ja omais- kyselyjen yhteistä verkostorakennetta. Lopuksi tehtiin katsaus käytetyimmän

(10)

muistisairauksia koskevan neuropsykiatrisen oirekyselyn (NPI) psykometri- sista ominaisuuksista. Aineisto koostui Itä-Suomen yliopiston ALSOVA-tutki- mushankkeeseen sekä yhdysvaltalaiseen NACC-tutkimukseen osallistuneista henkilöistä, jotka sairastivat AT:a. Katsausta varten tehtiin tietokantahakuja.

Neuropsykiatrisista oireista apatia, poikkeava motorinen käyttäytyminen sekä syömisen häiriöt olivat yhteydessä itsestä huolehtimisen haasteisiin.

Näiden oireiden lisäksi harhaluulot olivat yhteydessä haasteisiin välineellisis- sä toiminnoissa, jotka sisältävät vaativampaa asioiden hoitamista. Omaisen arvioimana potilaan masennus, apatia tai ahdistuneisuus olivat heikosti yh- teydessä AT:a sairastavan henkilön omiin arvioihin näistä oireista. Sen sijaan itsearvioidut avuttomuuden ja arvottomuuden tunteet olivat oireverkoston keskeisimpiä oireita. NPI:n psykometrisessa katsauksessa havaittiin, että re- liabiliteetin eri ulottuvuudet ovat kyselyn suurimpia vahvuuksia. Validiteettia koskevia analyyseja NPI:stä on kuitenkin tehty vähän, minkä vuoksi on epä- selvää, tavoittaako kyseinen kysely ne oireet, joita sen on tarkoitus mitata.

Tulokset korostavat, että moninaiset neuropsykiatriset oireet ovat yh- teydessä haasteisiin päivittäisissä toiminnoissa. AT:n varhaisissa vaiheissa pelkällä omaiskyselyllä saatetaan tavoittaa vain osa potilaan kokemista mie- lialaoireista. Tulevaisuudessa tarvitaan lisää ymmärrystä neuropsykiatristen oireiden luonteesta. Teorian edistäminen sekä yleisesti hyväksytyt psyko- metriset käytännöt luovat kestävän pohjan oirekyselyjen kehittämiseen sekä niiden hyödyllisyyden arviointiin. Neuropsykiatrisen oirekuvan tarkempi ar- viointi voi hyödyttää myös hoidon suunnittelua.

(11)

Acknowledgements

The present study was conducted at the Brain Research Unit of the University of Eastern Finland, Kuopio during 2018–2021. I am deeply grateful for everyone who contributed to these studies and supported me along the way. First, I wish to thank my supervisors. This project would not have been possible without Professor Anne Koivisto, who has been a key figure in introducing me to the world of Alzheimer’s disease research from 2016 onwards. She has always kept the clinical importance of research in mind, and enthusiastically supported my, often changing, interests in advancing the knowledge in this field. Professor Taina Hintsa has been a steady source of support, and she has offered me clarity at times when I needed it the most. I have been lucky enough to benefit from her focused approach to crafting manuscripts and refining my scientific thought in the process. She has encouraged me to find psychological insight in a domain where these concerns are often marginalized. During the manuscript writing process, Doctor Ilona Hallikainen was the first person I turned towards when something was unclear or something unfortunate came up, which means I was in her office quite often. Her expertise and active research in the topics of this dissertation benefited this project greatly, as did her unrivalled hands-on experience with the ALSOVA data set.

I wish to thank the expert reviewers of this dissertation. Docent Eero Vuoksimaa and Docent Marja Hietanen were pivotal in honing the arguments presented here and strengthening the presentation. Their kind, collegial and encouraging words left a lasting impression after having done my best to keep this project coherent and topical. Additionally, I wish to thank Docent Eero Vuoksimaa for agreeing to the double-role of being both the opponent and a reviewer for this dissertation.

I am also indebted to my co-author, Docent Tuomo Hänninen, who has supported me since my master’s thesis. He introduced me to neuropsychology both as a clinical and research enterprise during my master’s thesis and later during my internship at the Department of Neurology in Kuopio University Hospital. For this dissertation, he strengthened one of the manuscripts by providing crucial comments. I am also thankful for the ALSOVA Study

(12)

Group, who had collected these valuable data and paved the way for my contribution to the project. I also wish to thank Tuomas Selander for his invaluable guidance on statistical matters regarding the ALSOVA manuscript.

I am grateful for the countless individuals who have maintained and collected data for the NACC database.

Moreover, I want to thank Marja Äikiä and Maria Raatikainen for their insightful supervision, and Sonja Paavola for her collegiality during my internship at the Kuopio University Hospital. I am indebted to Professor Hannu Räty, who provided ample support and encouragement during my first steps in academic research. Doctor Merja Hallikainen and Professor Reetta Kälviäinen have made me feel very welcome in the University of Eastern Finland’s Department of Neurology, for which I am very grateful. I wish to thank all my current and previous co-workers at the Brain Research Unit for their support and company throughout the years. My colleagues at Verve have supported me through the last stages of this project, for which I am very grateful. Additionally, I wish to thank Mari Tikkanen for her work in administrative matters and Esa Koivisto for making this ride as smooth as possible by helping with practical issues. I want to thank Kaisu Kortelainen, who has warmly and patiently helped me through the hoops of doctoral studies. Zack Miller has helped me with numerous queries regarding the NACC dataset, for which I am deeply grateful.

I wish to thank all my friends for encouraging me on this path. Anton, Ilja, Nicholas, Heikki and Aaro have been there for me since the day I started my studies in psychology. Juha-Matti has been my n=1 clinical Alzheimer’s disease research peer support group, and with him I have had many fruitful discussions about the perplexities in psychology, neurology, and psychiatry.

Juho has been my closest psychologist-colleague during this project, and discussions with him have one way or another influenced the manuscripts.

For my mother, Arja, and my older brother, Jani, I am deeply thankful for you supporting my studies from early on. Of all people, I am most indebted to Tiia, for her love and support through this journey. She has been there through the highs and lows and stood patiently with me in times of uncertainty.

Finally, I want to thank everyone who participated in these studies. Without their invaluable effort, this dissertation would not have been possible. The

(13)

ALSOVA study was supported by Kuopio University Hospital (VTR grant 1V255/5772728) and the Päivikki and Sakari Sohlberg Foundation. The NACC database is funded by NIA/NIH Grant U01 AG016976. NACC data are contributed by the NIA-funded ADCs. This study was supported by personal grants from the Finnish Cultural Foundation and the Finnish Brain Foundation.

Kuopio, May 2021 Toni Saari

(14)

(15)

Koivisto et al. (2018). Methods are in regular typeface, whereas relevant findings are italicized. AD = Alzheimer’s disease, Aβ = beta amyloid, CERAD = Consortium to Establish a Registry for Alzheimer’s Disease, CSF = cerebrospinal fluid, MRI = magnetic resonance imaging, PET = positron emission tomography, p-tau = phosphorylated tau, t-tau = total tau. ...34 Figure 3. Clinical and research diagnostic criteria used in AD. The

spectrum indicates the degree to which clinical or biological features of AD are emphasized in the criteria. AD = Alzheimer’s disease, DSM-5 = Diagnostic and Statistical Manual of Mental Disorders – Fifth Edition, IWG-2 = International Working Group revised criteria, NIA-AA = National Institute of Aging – Alzheimer’s Association. ...41 Figure 4. Pathways between neuropsychiatric symptoms and

Alzheimer’s disease. A = etiologic pathway, B = shared risk/

confounding pathway, C = reverse causality, D = interaction (adapted from Geda et al., 2013). AD = Alzheimer’s disease, MCI = mild cognitive impairment. ...43 Figure 5. A simplistic network model. Conditional dependence

relationships between symptoms (nodes 1–10) are depicted by black edges. Node 3 is central based on the several connections it has in the network. Nodes 4 and 8, however, are poorly connected and may not contribute much to the phenomenon depicted by the network. All edge weights are equal for illustrative purposes. ...75 Figure 6. Flow-chart of the ALSOVA study. ...80 Figure 7. Flow-chart of the NACC study sample selection process. ...83

(20)

Figure 8. Correlation networks of NPI domain scores and IADL and BADL subscores, where figures A–F represent data from study visits from baseline through the fifth follow-up visit in chronological order. Edges are Spearman correlations (r ≥ .15), red edges denoting negative and blue positive correlations.

Thicker edges correspond to larger correlations. ABE = Aberrant motor behavior, AGI = Agitation, ANX = Anxiety, APA = Apathy, APP = Appetite disturbances, BA = Basic activities of daily living, DEL = Delusions, DEP = Depression, DIS = Disinhibition, EUF = Euphoria, HAL = Hallucinations, IA = Instrumental activities of daily living, IRR = Irritability, SLE = Sleep disturbances. ...94 Figure 9. Network structures of the GDS-15 and NPI-Q items of

depression, anxiety, and apathy at baseline (A) and at follow-up (B). Blue lines are edges between nodes, depicting conditional dependence relationships between nodes.

Thicker edges correspond to stronger associations. Color of the nodes is analogous to their community membership, i.e. nodes with the same color indicate the communities detected with the walktrap algorithm. The blue ring around a node estimates how well other nodes predict the state of the node (symptom present or not); a full blue ring would indicate the node being fully predicted by other nodes. Layout of the two networks was averaged to aid visual comparison. Items indicated by >< in the legend are reverse scored. ...99 Figure 10. Standardized strength estimates of nodes at baseline (red)

and follow-up (teal), indicating direct connectedness of

each node. ...101

(21)

ABBREVIATIONS

AD Alzheimer’s disease

AD+P Psychosis of AD

ADC Alzheimer’s Disease Center

ADCS-ADL Alzheimer’s Disease Cooperative Study – Activities of Daily Living inventory

ADL Activity of daily living

APA American Psychiatric Association

APA, AERA, NCME American Educational Research Association, American Psychological Association, National Council on

Measurement in Education, & Joint Committee on Standards for Educational and Psychological Testing

APOE Apolipoprotein E

AT(N) Amyloid-tau-(neurodegeneration)

Aβ Amyloid beta

BADL Basic activity of daily living

bvFTD Behavioral variant frontotemporal dementia CDR Clinical Dementia Rating scale

CSF Cerebrospinal fluid

CTT Classical test theory

DICE Describe, Investigate, Create, Evaluate

DLB Dementia with Lewy bodies

DSM-5 Diagnostic and Statistical Manual of Mental Disorders – Fifth Edition

DSM-IV Diagnostic and Statistical Manual of Mental Disorders – Fourth Edition

FTD Frontotemporal dementia

GDS-15 Geriatric Depression Scale—15 GEE Generalized Estimating Equation IADL Instrumental activity of daily living

ICD-10 International Statistical Classification of Diseases and Related Health Problems – Tenth Edition

IRT Item response theory

(22)

LVM Latent variable model

MBI Mild behavioral impairment

MCI Mild cognitive impairment

MRI Magnetic resonance imaging

NACC National Alzheimer’s Coordinating Center

NIA-AA National Institute of Aging – Alzheimer’s Association NIMH-dAD National Institute of Mental Health provisional

diagnostic criteria for depression in AD

NINCDS-ADRDA National Institute of Neurological and Communicative Disorders and Stroke and Alzheimer´s Disease and Related Disorders Association

NPI Neuropsychiatric Inventory

NPI-C Neuropsychiatric Inventory Clinician Version NPI-Q Neuropsychiatric Inventory Questionnaire

NPS Neuropsychiatric symptoms

PET Positron emission tomography

VaD Vascular dementia

(23)

List of original publications

This dissertation is based on the following original publications:

I Saari, T., Hallikainen, I., Hintsa, T., & Koivisto, A. (2020). Neuropsychiatric symptoms and activities of daily living in Alzheimer’s disease: ALSOVA 5-year follow-up study. International Psychogeriatrics, 32(6), 741–751

II Saari, T.T., Hallikainen, I., Hintsa, T., & Koivisto, A.M. (2020). Network structures and temporal stability of self- and informant-rated affective symptoms in Alzheimer’s disease. Journal of Affective Disorders, 276, 1084–1092.

III Saari, T., Koivisto, A., Hintsa, T., Hänninen, T., & Hallikainen, I. (2020).

Psychometric properties of the Neuropsychiatric Inventory: A Review.

Journal of Alzheimer’s Disease. Advance online publication.

The publications were adapted with the permission of the copyright owners.

(24)

(25)

1 Introduction

Alzheimer’s disease (AD) is a neurodegenerative disease characterized by gradual cognitive decline. AD is the most common memory disorder, accounting for some 70 % of all dementia cases (Reitz, Brayne, & Mayeux, 2011). Dementia refers to all-cause cognitive or neuropsychiatric symptoms which disturb the person’s functional independence, exhibit decline from previous levels of functioning and are not explained by a major psychiatric illness (McKhann et al., 2011). AD and other dementias affect over 40 million persons globally and constitute the fifth commonest cause of death (Nichols et al., 2019). The number of persons with dementia has more than doubled in the past decade, and is expected to increase in a similar fashion (Nichols et al., 2019; Prince et al., 2013). Disease modifying treatment is not available and only 1 % of clinical trials have favored the drug over a placebo (Cummings et al., 2019).

The majority of AD research has focused on cognitive symptoms, such as the hallmark episodic memory deficits. This is reflected in the 2011 National Institute on Aging-Alzheimer Association (NIA-AA) diagnostic criteria for probable AD, which require amnestic or non-amnestic cognitive decline (McKhann et al., 2011). Neuropsychiatric symptoms (NPS), on the other hand, have received less attention historically (Berrios, 1989), but have surfaced as a major topic of interest in the 21^st century (Lyketsos et al., 2011, 2002). NPS is an umbrella term referring to changes in behavior, mood and thought processes related to dementia. These symptoms can manifest in different ways, ranging from depressive and apathetic states to psychotic episodes and aggressiveness.

Neuropsychiatric symptoms are common in AD (Zhao et al., 2016), cause the most distress to caregivers out of any symptoms (Terum et al., 2017) and compromise functional independence of the patients (e.g., Wadsworth et al., 2012; Okura et al., 2010). The latter finding relates to shared neurobiology underlying behavioral and emotional regulation as well as activities of daily living (ADLs) (McKeith & Cummings, 2005). Recent work has clarified the definitions of different neuropsychiatric syndromes (Cummings et al., 2020;

(26)

Robert et al., 2018; Cummings et al., 2015), the sequence in which symptoms appear during the disease (Vik-Mo, Giil, Ballard, & Aarsland, 2018) and possible etiological pathways of NPS (Geda et al., 2013).

Despite research progress, two factors continue to obscure the relationships between NPS and clinical outcomes of interest, such as ADLs.

On one hand, NPS are often assessed with questionnaires with limited validity and reliability evidence (Perrault, Oremus, Demers, Vida, & Wolfson, 2000; Seignourel, Kunik, Snow, Wilson, & Stanley, 2008; Clarke et al., 2011).

Psychometric properties of NPS rating scales affect the validity of inferences made based on them (Shaw, Cloos, Luong, Elbaz, & Flake, 2020; Flake, Pek, &

Hehman, 2017; Loken & Gelman, 2017; Carlson & Herdman, 2012; Bollen &

Lennox, 1991), yet these concerns cannot be solely attributed to caregiver bias (McKeith & Cummings, 2005). On the other, associations between NPS and functional outcomes have limited generalizability as the studies have used varying samples, study designs and strategies of addressing confounders.

Mixed results of associations between NPS and ADLs have been reported, where some studies have suggested nearly all NPS to be related to ADLs (Okura et al., 2010; Tekin, Fairbanks, O’Connor, Rosenberg, & Cummings, 2001), while others have reported few associations between NPS and ADLs (You et al., 2015; Green et al., 1999).

This study has three objectives. First, longitudinal relationships between NPS domains and functional impairment are investigated. Second, a network analysis of both self- and informant-rated affective symptoms in AD is conducted and finally, validity and reliability evidence for the most common NPS measure, the Neuropsychiatric Inventory (NPI), is reviewed.

(27)

2 Review of the literature

2.1 Alzheimer’s disease

AD is a neurodegenerative disease characterized by progressive cognitive decline, particularly in episodic memory. The ultimate cause of AD remains unknown (Scheltens et al., 2016), but abnormal accumulation of two molecules, amyloid beta (Aβ) and tau, are the major pathological findings of the disease (Bloom, 2014; Polanco et al., 2018). These pathologies contribute to gradually spreading neuronal and synaptic damage in the brain, typically beginning from the entorhinal cortex, a region associated with memory and orientation (Braak & Braak, 1991). In line with the changes in the brain, episodic memory deficits are usually the first cognitive symptoms (Dubois et al., 2014). ADLs ultimately become compromised by the changes in brain structure and metabolism. At this point, the person with AD requires assistance to carry out his or her daily routines, such as managing finances, taking medication and, later, attending to self-care. This loss of ability to manage one’s daily affairs constitutes the syndrome of dementia. For all-cause dementia, the median times between diagnosis and entry into long-term care and death are estimated to be 3.9 and 5 years, respectively (Joling et al., 2020).

2.1.1 Epidemiology

2.1.1.1 Prevalence

Alzheimer’s disease is the most common neurocognitive disorder and cause of dementia, accounting for 70–80% of all dementias. Globally, over 40 million persons are estimated to have AD or other dementias (Nichols et al., 2019).

As societies age, this figure is expected to nearly double every two decades at least until 2050, with a disproportionately large contribution from low or middle income countries (Prince et al., 2013; Nichols et al., 2019). Based on US data, some 3 % of individuals aged 65–74 have AD, whereas 17.6 % of individuals aged 75–84 and 32.3 % aged over 85 have AD (Hebert, Weuve,

(28)

Scherr, & Evans, 2013). Of community-dwelling individuals over 60, some 4 % are estimated to have AD based on a meta-analysis (Fiest et al., 2016).

Some 190 000 persons have a memory disorder in Finland, and 86 000 persons bought prescription medicine for memory problems in 2019 (Finnish Institute of Health and Welfare, 2020). From the available dementia incidence data it can be estimated that there are some 10 000 new AD cases annually in Finland, which corresponds to a 10 % increase from 2010 (Finnish Institute of Health and Welfare, 2020; Finnish Ministry of Social Affairs and Health, 2013;

Memory Disorders: Current Care Guidelines, 2017).

2.1.1.2 Risk factors and prevention

There are several non-modifiable and potentially modifiable risk factors for dementia (World Health Organization [WHO], 2019). These risk factors have a marked temporal range, ranging from genetics, to education, midlife hypertension and obesity, to later life medical, psychosocial and lifestyle factors (Livingston et al., 2020; WHO, 2019). In a recent Lancet Commission report, the most significant modifiable risk factors for dementia were lesser education, hearing loss, and smoking for years < 45, 45–65 and > 65 years, respectively (Livingston et al., 2020). Together, modifiable risk factors could account for up to 40% of the risk for dementia (Livingston et al., 2020). WHO recommendations for interventions addressing known modifiable risk factors are compiled in Table 1. Of note, air pollution and traumatic brain injury were included as potentially modifiable risk factors in the more recent Lancet Commission report, but not in the WHO recommendations.

Age is recognized as the most important non-modifiable risk factor (Alzheimer’s Association, 2019). There are also genetic risk factors for AD, most importantly the apolipoprotein E allele 4 (APOE ε4, Saunders et al., 1993).

Age and APOE ε4 interact, so that APOE ε44 carriers have 5.6 and 15.8 times the risk for AD at age under 60 and age 70 to 79, respectively (Genin et al., 2011). For APOE ε34, the corresponding risks are 2.1 and 3.1. In addition to interactions with age, the deleterious effects of APOE ε4 are more apparent for women (Mielke, Vemuri, & Rocca, 2014). Overall, AD is more prevalent in women than in men, with nearly two-thirds of AD diagnoses being in women (Alzheimer’s Association, 2019).

(29)

Recent genome-wide association studies have revealed several new susceptibility loci for AD (Alzheimer Disease Genetics Consortium et al., 2019), while rare mutations in presenilin-1, presenilin-2 and amyloid protein precursor genes are known to cause dominantly inherited familial AD (Lanoiselée et al., 2017). In addition to APOE and the genes associated with familial AD, several common variants have been found to affect key processes tied to AD pathology (Alzheimer Disease Genetics Consortium et al., 2019).

Furthermore, genes associated with intelligence are inversely associated with AD, highlighting the partially biological origin of some protective factors (Savage et al., 2018; The Brainstorm Consortium et al., 2018).

Table 1. Modifiable risk factors and the intervention evidence for risk re- duction of cognitive decline (adapted from WHO, 2019)

Modifiable

risk factor Intervention Quality of

evidence Strength of recommendation Physical

activity Increasing physical activity Low to

moderate Conditional to strong

Smoking Tobacco cessation Low Strong

Nutritional

interventions Mediterranean diet Moderate Conditional

Balanced diet Low to high Strong

Alcohol use Reduction or cessation of

harmful drinking Moderate Conditional Cognitive

activity Cognitive training Very low to

low Conditional

Social activity - - -

Bodyweight Reduction of midlife obesity Low to

moderate Conditional Hypertension Management of

hypertension Very low to

high Conditional to strong

Diabetes Management of diabetes Very low to

moderate Conditional to strong

Dyslipidemia Management of dyslipidemia Low Conditional

Depression - - -

Hearing loss - - -

Blank rows denote risk factors for which the WHO guideline development group found insufficient evidence to make intervention recommendations.

(30)

2.1.2 Cognitive changes and neuropathology

AD is primarily characterized by progressive cognitive deterioration. Deficits in episodic memory, or remembering what has happened or been said recently, are often the prevalent and earliest cognitive symptoms in AD (McKhann et al., 2011), with other cognitive difficulties emerging later. This amnestic variant accounts for approximately 60 % of the cases, and is more prevalent in the older age strata (Scheltens et al., 2017). Atypical variants, in which the early cognitive impairment is in a non-memory domain, are less frequent.

The major atypical variants are: posterior variant of AD, where visuospatial functioning is impaired; logopenic variant of AD, where naming and sentence repetition deficits prevail; and frontal variant of AD, characterized by frontal neuropsychiatric symptoms and impairment in executive functions (Dubois et al., 2014). Different neuropathological and genetic profiles support categorization based on the primary cognitive difficulty (Crane et al., 2017).

The amnestic variant of AD is typically associated with older age, APOE ε4 positivity and more severe atrophy in hippocampal areas relative to posterior cortical areas (Scheltens et al., 2017). While cognitive changes are considered the hallmark features of AD, disturbances in physiological and motor functions are also acknowledged (Ahmed et al., 2018).

Several neuropathological mechanisms have been associated with typical AD, although the ultimate trigger(s) for the cascade of neurodegenerative events remain unclear (Scheltens et al., 2016). Of these, two related neuropathological pathways are the best understood: Aβ and tau (Bloom, 2014; Polanco et al., 2018). Both groups of molecules have important supporting functions in healthy brains, but chronic disturbance in their physiology leads to dysfunction and destruction of neurons and synapses in AD and other neurodegenerative diseases (Polanco et al., 2018). A sustained imbalance between Aβ production and clearance can form extracellular plaques, triggering processes leading to neuronal and synaptic insults (Hardy, 2002). Aβ accumulation in turn can lead to hyperphosphorylation of tau, which results in formation of intracellular neurofibrillary tangles (Medeiros, Baglietto-Vargas, & LaFerla, 2011). Aβ and tau also interact to produce neurotoxic effects not reduceable to effects of either molecule (Bloom,

(31)

2014). Recent studies also implicate the role of inflammation and microglial activation in the pathogenesis of AD (Polanco et al., 2018).

The accumulation of tangles has been characterized by what is known as Braak staging, where six stages of tangle accumulation are distinguished (Braak

& Braak, 1991). In stage I, tangles appear exclusively in the transentorhinal cortex, gradually spreading and increasing in density in already afflicted regions in further stages until nearly all cortical areas are affected in stage VI. These features are consistent with the clinical changes of AD (Dubois et al., 2010): tangles in the medial temporal lobes in stages III–VI correspond to memory impairment, and deficits in executive function, language and visuospatial skills become more pronounced in stages VI–VII as tangles develop in the brain regions associated with these functions (Nelson et al., 2012).

2.1.3 Clinical spectrum of Alzheimer’s disease

The clinical symptoms of AD develop over time. Years or even decades may lie between the first AD brain changes and the onset of clinical symptoms (Nelson et al., 2012). This intermediate stage is called preclinical AD, and can only be detected through biomarkers. The first clinical symptoms denote prodromal or mild cognitive impairment stages, whereas progression of cognitive and functional symptoms is used to stage the severity of AD dementia.

2.1.3.1 Preclinical states, mild cognitive impairment and prodromal Alzheimer’s disease

The preclinical states of AD comprise an asymptomatic at-risk state for AD, where evidence of abnormal amyloid accumulation can be detected either in positron emission tomography (PET) imaging or cerebrospinal fluid (CSF), and presymptomatic AD, where the person has a rare monogenic mutation that causes AD later (Dubois et al., 2010). Persons in these states do not have symptoms, and they differ in the likelihood of progression to the clinical syndrome of AD. For example, a positive amyloid PET scan alone in an asymptomatic person might not be a good predictor of future clinical AD (Iaccarino, Sala, Perani, & for the Alzheimer’s Disease Neuroimaging Initiative,

(32)

2019), whereas rare causal mutations inevitably lead to AD pathology and symptoms.

Early cognitive changes without functional impairment can be characterized by mild cognitive impairment (MCI) or by prodromal AD. Of these, MCI is a broader term including non-AD etiologies as well (Dubois & Albert, 2004), and refers to cognitive functioning below age and education expectations. In MCI, the cognitive symptoms are mild enough to not significantly compromise ADLs (Petersen, 2004; Winblad et al., 2004). The concept allows for highly heterogenous etiologies, and can be considered to refer to a clinical syndrome rather than a disease process (Dubois et al., 2010). The concept of mild neurocognitive disorder in the Diagnostic and Statistical Manual for Mental Disorders – Fifth Edition (DSM-5) captures similar clinical features (American Psychiatric Association [APA], 2013). MCI is a risk for future dementia (Petersen et al., 2009), although the reversal rate from MCI to normal cognition can be up to 30 % (Malek-Ahmadi, 2016), depending on the patient group and used criteria for MCI (Thomas et al., 2019). MCI can be further divided into subtypes based on the primarily affected cognitive domain(s): single domain amnestic MCI, multiple domain amnestic MCI, single domain non-amnestic MCI and multiple domain non-amnestic MCI (Petersen et al., 2001; Winblad et al., 2004). These subtypes correspond to an increased risk for specific neurodegenerative diseases (Figure 1).

The brevity of the all-cause MCI concept is problematic when more specific prodromal disease states are investigated. To that end, many researchers and clinicians prefer to use the concept prodromal AD to capture the combination of early amnestic cognitive impairment with biomarker and neuroimaging evidence that support AD pathology (Dubois & Albert, 2004; Dubois et al., 2014). Within the amnestic MCI subdomain, the evidence of AD pathology can increase the accuracy of progression to AD by reflecting a disease rather than merely a clinical syndrome (Dubois et al., 2010).

(33)

Figure 1. Flowchart for MCI diagnostic procedures, adapted from Winblad et al. (2004) and Petersen (2009). AD = Alzheimer’s disease, D = depression, DLB = dementia with Lewy bodies, FTD = frontotemporal dementia, MCI = mild cognitive impairment, VaD = vascular dementia.

2.1.3.2 Mild behavioral impairment

In recent years, early NPS have become a major focus of interest (Ismail et al., 2016). Spearheaded by research in non-AD neurocognitive disorders, such as behavioral variant frontotemporal dementia (bvFTD; Neary et al., 1998; Rascovsky et al., 2011), criteria for early neuropsychiatric syndromes have emerged (Taragano et al., 2009). The syndrome of mild behavioral impairment (MBI) was proposed to systematically define early NPS preceding dementia with the following criteria: a major change in patient behavior, symptom onset after age 60 and persistence for at least 6 months, absence of cognitive impairment, normal functioning in social and occupational settings with preserved activities of daily living (Taragano et al., 2009).

The Alzheimer’s Association International Society to Advance Alzheimer’s Research And Treatment – Neuropsychiatric Symptoms Professional Interest Area revised the Taragano criteria in 2016, where the most important change was the removal of MCI as an exclusion criterion for MBI (Ismail et al., 2016).

(34)

The revised criteria also outline five behavioral and personality changes, of which at least one must be present: decreased motivation, affective dysregulation, impulse dyscontrol, social inappropriateness, and abnormal perception or thought content. In a population-based study of individuals over 50 years of age, the prevalence of MBI was 10 % based on a previously established cut-off on a rating scale (Creese, Brooker, et al., 2019). Earlier studies have shown that 27 % of older individuals with normal cognition have at least one NPS, commonly in the affective domain (Geda et al., 2008).

While some patients may present with predominantly behavioral changes, cognitive and behavioral symptoms often coexist. Persons with MCI or subjective cognitive decline are at an increased risk of future cognitive decline if NPS are also present (Creese et al., 2019; Pink et al., 2015). If persons with comorbid cognitive symptoms and NPS in these early stages progress to dementia, the diagnoses are mainly AD or frontotemporal dementia (FTD), with a ratio of approximately 1:2 (Taragano et al., 2009).

2.1.3.3 Dementia due to Alzheimer’s disease

Based on the severity of cognitive and functional decline, dementia due to Alzheimer’s disease can be separated to mild, moderate and severe stages (Table 2; Hughes, Berg, Danziger, Coben, & Martin, 1982; Memory Disorders:

Current Care Guidelines, 2017). Ultimately, the patient will require assistance in carrying out their activities of daily living. The most complex ADLs are impaired first and the most basic functions of self-care are preserved longest (Lindbergh, Dishman, & Miller, 2016; Perneczky et al., 2006; Nygård, 2003;

Njegovan, Man-Son-Hing, Mitchell, & Molnar, 2001). NPS may be present even before AD diagnosis (Taragano et al., 2009; Pink et al., 2015), but they become increasingly disturbing as the disease progresses (Piccininni, Di Carlo, Baldereschi, Zaccara, & Inzitari, 2005; Hashimoto et al., 2015).

(35)

Table 2. Characteristic symptoms in different stages of AD dementia (adapt- ed from National Current Care Guidelines for Memory Disorders, 2017)

Mild stage (MMSE

18–26, CDR 0.5–1) Moderate stage (MMSE 10–22, CDR 1–2)

Severe stage (MMSE 0–12, CDR 2–3) Cognitive

symptoms Forgetfulness, impaired learning and executive functions, deteriorated reasoning, word finding difficulties, problems in calculations

Episodic memory deficits, difficulties in speech production, disorientation, lack of awareness and/or insight, visuospatial difficulties, apraxia

Limited speech production, marked difficulties in understanding speech, inattention, severe apraxia

Changes in activities of daily living

Difficulties in following a conversation, decreased reading, giving up complicated hobbies, problems planning and attending to finances, problems with medication, decreased ability to work and drive, increased use of memory aids

Impaired IADLs, difficulties in prepar- ing a meal, dressing inappropriately, misplacing personal belongings, getting lost, need for re- minders in BADLs

Need for assistance in BADLs, inconti- nence

Neuro- psychiatric symptoms

Apathy, detachment, irritability, anxiety, depression, delusions

Delusions, hallucinations, agitation, wandering, disturbed sleep-wake cycle, basic social skills intact

Agitation, aggressiv- ity, aberrant motor behavior, wandering, disturbed sleep-wake cycle, depression, or apathy

Somatic

symptoms Weight loss Weight loss Gait apraxia, primary reflexes, extrapy- ramidal symptoms, secondary frailty BADLs = basic activities of daily living, CDR = Clinical Dementia Rating,

IADLs = instrumental activities of daily living, MMSE = Mini-Mental State Examination.

(36)

2.1.4 Diagnostics and assessment

Diagnosis of AD rests on clinical judgment combining neuropsychological testing, taking history from the patient and informant, neuroimaging and laboratory tests (Figure 2; Koivisto et al., 2018). All components serve inclusionary and exclusionary roles by strengthening or weakening the confidence in the diagnosis of AD. Diagnostic certainty is achieved only by examining the patient’s brain post-mortem (Nelson et al., 2012).

Figure 2. Flow diagram for AD diagnostics in Finland, adapted from Koivis- to et al. (2018). Methods are in regular typeface, whereas relevant findings are italicized. AD = Alzheimer’s disease, Aβ = beta amyloid, CERAD = Con- sortium to Establish a Registry for Alzheimer’s Disease, CSF = cerebrospinal fluid, MRI = magnetic resonance imaging, PET = positron emission tomography, p-tau = phosphorylated tau, t-tau = total tau.

(37)

2.1.4.1 Clinical assessment

A careful history from both the patient and an informant is crucial for understanding the nature and development of symptoms, as neuropsychological testing results are rarely available before the diagnostic procedures for AD. Anosognosia, or lack of awareness about one’s symptoms or severity of them, may underlie differences in patient and informant reports (Robert et al., 2018). The information regarding onset of symptoms can be used to exclude etiologies that have a sudden onset, a rapid decline or fluctuating course not typical for AD.

Psychiatric symptoms should be screened, as they could be the primary explanation for cognitive symptoms, present comorbidly with a memory disorder, or precede cognitive symptoms related to a memory disorder (Byers

& Yaffe, 2011). Several screening tools are available for screening psychiatric symptoms in suspected cognitive impairment, such as the Geriatric Depression Scale (GDS-15; Yesavage & Sheikh, 1986) and the Neuropsychiatric Inventory (Cummings et al., 1994; Cummings, 1997a). Quality of the symptoms is also useful for differential diagnosis, for example, the content of psychotic symptoms differ in AD versus schizophrenia (Jeste & Finkel, 2000; Lyketsos et al., 2011). Neuropsychiatric symptoms currently do not play a key role in diagnosing typical AD.

Structured clinical interviews complement free-form history taking and offer quantitative evidence for longitudinal monitoring. ADL inventories can be used to evaluate the patient’s functional capabilities, or the practical problems arising from cognitive difficulties. The Alzheimer’s Disease Cooperative Study – Activities of Daily Living inventory (ADCS-ADL; Galasko et al., 1997) distinguishes between basic ADLs and instrumental ADLs, and has been used in both clinical trials and observational studies to characterize the longitudinal decline in AD. For younger persons, information about workplace performance is more relevant (Kuikka, Akila, Pulliainen, & Salo, 2018).

The neuropsychological profile of typical AD can be characterized as “an amnestic presentation of the hippocampal type” (Dubois & Albert, 2004). This pattern of difficulties in memory tasks is characterized by poor delayed recall of encoded stimuli that is not significantly improved by cueing. Persons with atypical variants of AD may show disproportionate impairment in visuospatial,

(38)

language or executive functions (Dubois et al., 2014). Several screening methods, such as the Consortium to Establish a Registry for Alzheimer’s Disease – Neuropsychological Battery (Morris et al., 1989; Hänninen et al., 1999), are available for detecting the cognitive symptoms, although more comprehensive neuropsychological testing is often desirable, particularly for younger persons.

2.1.4.2 Biomarkers

Currently, there are five biomarkers for AD that have sufficient validity evidence for their use in research, but also in clinical settings to a varying extent (Table 3, Frisoni et al., 2017). Of these, magnetic resonance imaging (MRI) is the first line of neuroimaging in AD diagnostics. Typical structural MRI findings for AD are atrophy in the hippocampus and entorhinal cortex compared to cognitively healthy controls. Several rating scales exist for grading the severity of atrophy (Harper, Barkhof, Fox, & Schott, 2015). Hippocampal (Scheltens et al., 1992), posterior (Koedam et al., 2011) and frontal (Pasquier, Leys, Mounier- Vehier, Barkhof, & Scheltens, 1996; Scheltens, Pasquier, Weerts, Barkhof, &

Leys, 1997; Ferreira et al., 2016) rating scales have been used in AD research and clinical practice. Parietal cortical atrophy may be more pronounced in atypical AD (Scheltens et al., 2017).

(39)

Table 3. The core biomarkers for diagnosis of Alzheimer’s disease (adapted from Frisoni et al., 2017)

Findings Underlying pathology

MRI

Regional anatomy Reduced volume of hippocampi and other regions in the temporal

lobes

Atrophy and neurodegeneration

PET

FDG-PET Reduced uptake in tem-

poroparietal cortex and posterior cingulate-pre- cuneus

Glucose hypometabolism and neurodegeneration

Amyloid PET Increased cortical reten-

tion Cortical deposition of Aβ

CSF

Aβ42 or Aβ42/Aβ40 Reduced concentration

or ratio Abnormal Aβ metabolism

Total tau and hyperphos-

phorylated tau Increased concentration Neuronal damage and accumulation of tau Aβ = beta amyloid, FDG = fluorodeoxyglucose, CSF = cerebrospinal fluid, MRI = magnetic resonance imaging, PET = positron emission tomography.

If the previously mentioned factors are inconclusive, cerebrospinal fluid (CSF) biomarkers can be examined for β amyloid (Aβ), total tau (t-tau) and phosphorylated tau (p-tau). In these CSF markers, a pattern of low Aβ with high t- and p-tau have been associated with AD (Dubois et al., 2007; Blennow

& Zetterberg, 2018). Recently, examining the levels of a 42-amino acid-long peptide, Aβ42, has been improved by more sensitive and specific ratios of Aβ42/Aβ40 and Aβ42/Aβ38 (Janelidze et al., 2016).

PET imaging may offer additional insight if diagnosis is not reached with clinical, MRI and CSF data. Fluorodeoxyglucose PET is an imaging modality based on glucose metabolism, and it is commonly used in differential diagnosis based on distinct uptake patterns in different memory disorders.

Recent fluorodeoxyglucose PET protocols may predict progression from MCI to AD better than other markers, but more research is needed to find

(40)

the optimal uses (Smailagic, Lafortune, Kelly, Hyde, & Brayne, 2018). In an exclusionary sense, a negative fluorodeoxyglucose PET may be the most reliable biomarker for predicting clinical stability or reversion to normal cognition in MCI (Caminiti et al., 2018).

Imaging amyloid directly in the brain is possible with amyloid PET imaging.

Amyloid imaging correlates highly with CSF Aβ42 but has the additional benefit of showing ligand retention distribution in the brain (Blennow & Zetterberg, 2018), not just a quantitative estimate of the amyloid levels. Amyloid PET is usually reserved for situations where objectively verified cognitive decline has not received an explanation from previous diagnostic procedures, but the knowledge of brain Aβ pathology is considered to improve diagnostic certainty (Johnson et al., 2013). Limitations of amyloid PET imaging include its cost, limited availability, and detection of AD pathology later than CSF (Palmqvist, Mattsson, Hansson, & for the Alzheimer’s Disease Neuroimaging Initiative, 2016; Lewczuk & Kornhuber, 2016; Rabinovici, 2016), thus, it is mostly used in research (Frisoni et al., 2017).

2.1.5 Clinical and research diagnostic criteria

2.1.5.1 Clinical criteria

AD has long been considered a clinical syndrome characterized by gradual cognitive decline (McKhann et al., 1984). Currently, the diagnosis is set following NIA-AA (McKhann et al., 2011) or DSM-5 (APA, 2013) criteria. The NIA-AA clinical diagnostic criteria for probable AD require gradual onset, history of cognitive decline as reported or demonstrated in neuropsychological assessment and dominantly either amnestic or non-amnestic profile (Table 4, McKhann et al., 2011). Probable AD cannot be diagnosed if temporally related and clinically significant cerebrovascular problems exist, if there are clinical signs and symptoms of another neurological disease, and if cognitive decline could be a result of medication or a comorbid illness affecting cognition (McKhann et al., 2011). In the NIA-AA criteria, confidence in the probable AD diagnosis is increased by the presence of AD biomarkers, but not required.

(41)

Table 4. NIA-AA diagnostic criteria for probable Alzheimer’s disease (adapt- ed from McKhann et al., 2011)

1. Dementia as evidenced by cognitive or neuropsychiatric symptoms that a. Interfere with the ability to function at work or usual activities

b. Indicate a decline from previous level of functioning

c. Cannot be explained by delirium or major psychiatric disorder

d. Manifest in at least two of the following domains — acquiring and remembering new information; reasoning and judgment; visuospatial abilities;

language functions; personality, behavior or comportment 2. Gradual onset of symptoms

3. History of cognitive decline by report or observation

4. The initial and most prominent cognitive symptoms appear as either

a. Amnestic presentation — deficits in learning and recall of recently learned information, as well as deficits in some other cognitive domain per criterion 1d.

b. Nonamnestic presentation — deficits in either language, visuospatial or executive functions.

5. The patient does not have a history of stroke with a temporal relation to onset or worsening of cognitive symptoms, prominent features of other neurodegenerative diseases, other neurological or non-neurological comorbidities that could account for the symptoms, or medication that could substantially influence cognitive performance.

NIA-AA = National Institute of Aging – Alzheimer’s Association.

The DSM-5 diagnosis of major neurocognitive disorder is not specific to AD, but rather includes all cognitive disorders with significant decline in at least one cognitive domain, these cognitive problems interfere with ADLs, and the problems are not better explained by mental disorders or delirium (APA, 2013). Evidence of cognitive decline in at least two domains, one of which is learning and memory, is required to attribute AD as the cause of major neurocognitive disorder (Sachdev et al., 2014). The NIA-AA and DSM-5 criteria largely overlap for AD, but NIA-AA criteria better account for atypical variants of AD, where non-amnestic symptoms are prominent.

2.1.5.2 Research criteria

Clinical criteria are useful in observational studies, where heterogeneity of the study population is preferred over homogeneity commonly required in interventional trials (Jack et al., 2018), as a complete biomarker profile of

(42)

amyloid, tau and neurodegeneration is neither feasible nor necessary for many observational studies. The International Working Group research diagnostic criteria for AD strike a balance between the more clinically oriented NIA-AA and DSM-5 criteria, and purely biological criteria by requiring presence of both clinical and biomarker evidence (Dubois et al., 2014).

Clinical-biological criteria, however, may not be sensitive or specific enough, nor can they advance understanding of the preclinical phase where biomarker findings indicate AD pathology in the absence of clinical symptoms (Jack et al., 2018). To address this gap, research criteria for biological definition of AD were recently proposed (Jack et al., 2018). These biological criteria are based on burgeoning evidence for the three biomarkers, Aβ, tau, and neurodegeneration [AT(N)], and the criteria are designed to standardize use of biomarkers in research settings (Jack et al., 2018). Classification based on AT(N) does not assume the neuropathological events to occur in any specific temporal order (Jack et al., 2018, 2016). This lack of temporal commitment allows the AT(N) scheme to describe biomarker profiles outside the popular amyloid cascade framework, where amyloid beta accumulation precedes tau deposits and neurodegeneration (Jack et al., 2013). While the relationships between AT(N) profiles in preclinical AD and future cognitive difficulties are extensively documented, the findings between these biomarkers and NPS are so far somewhat consistent for amyloid only (Ng et al., 2021).

The different diagnostic criteria for AD and their uses are summarized in Figure 3.

(43)

Figure 3. Clinical and research diagnostic criteria used in AD. The spec- trum indicates the degree to which clinical or biological features of AD are emphasized in the criteria. AD = Alzheimer’s disease, DSM-5 = Diagnostic and Statistical Manual of Mental Disorders – Fifth Edition, IWG-2 = Interna- tional Working Group revised criteria, NIA-AA = National Institute of Aging – Alzheimer’s Association.

2.2 Neuropsychiatric symptoms and syndromes in Alzheimer’s disease

NPS, sometimes called behavioral and psychological symptoms of dementia, non-cognitive symptoms or simply psychiatric symptoms, is an umbrella term for symptoms manifesting in abnormal behavior, affective regulation and thought processes in neurodegenerative disease (Lyketsos et al., 2011).

The lexicon of the 12 domains in the revised Neuropsychiatric Inventory (Cummings, 1997a) is used here, as it is used most commonly in the literature.

The 12 domains are: delusions, hallucinations, agitation, depression, anxiety, euphoria, apathy, disinhibition, irritability, aberrant motor behavior, sleep disturbances and appetite disorders. Some of these symptoms can be considered to form reliable clinical entities called syndromes (Ismail et al., 2016; Geda et al., 2013; Jeste, Meeks, Kim, & Zubenko, 2006).

(44)

2.2.1 Etiology, prevalence and burden

2.2.1.1 Etiology

In line with advances in neuroimaging and genetics, recent inquiries into etiology of NPS emphasize biological factors, such as neural correlates and genetic factors (Krell-Roesch et al., 2019; Pink et al., 2015; Rosenberg, Nowrangi, & Lyketsos, 2015). However, psychological risk factors, such as personality traits and coping, are increasingly recognized as well (Arenaza- Urquijo et al., 2020; Sutin, Stephan, Luchetti, & Terracciano, 2018; Archer et al., 2007; Terracciano, An, Sutin, Thambisetty, & Resnick, 2017). While it is commonly appreciated that NPS are multifactorial in their etiology (e.g., Lanctôt, Amatniek, et al., 2017; Lyketsos, 2015), the explicit causal mechanisms are not well understood (Geda, Krell-Roesch, Sambuchi, & Michel, 2017).

Hypotheses regarding possible pathways of NPS and incipient cognitive decline are illustrated in Figure 4 (Geda et al., 2013). Both the factors within the pathways and the pathways themselves may interact with one another (Geda et al., 2013; Kendler, 2012): for instance, one pathway may explain how a NPS could emerge, whereas another could explain how the same NPS contributes to cognitive decline. The pathways should also be interpreted with the nature of individual symptoms in mind as a pathway could be compatible with one NPS but not another (e.g., affective reactions to declining cognition).

(45)

Figure 4. Pathways between neuropsychiatric symptoms and Alzheimer’s disease. A = etiologic pathway, B = shared risk/confounding pathway, C = reverse causality, D = interaction (adapted from Geda et al., 2013). AD = Alzheimer’s disease, MCI = mild cognitive impairment.

The A pathway corresponds to an etiologic model where neuropsychiatric symptoms are causally related to subsequent brain damage leading to cognitive symptoms of MCI or AD. This pathway does not attempt to define the mechanisms giving rise to the NPS, rather, the focus is on the consequences of them (Geda et al., 2013). While genetic correlations between psychiatric disorders are notable, AD and other neurological disorders are genetically mostly independent of psychiatric disorders (The Brainstorm Consortium et al., 2018). Thus, the mechanisms through which (neuro)psychiatric symptoms increase the risk of AD seem to be distinct compared to those that relate to the comorbidity of psychiatric disorders. For example, post- traumatic stress disorder could lead to disturbances in the regulation of the

(46)

hypothalamic–pituitary–adrenal axis, which could increase the accumulation of AD pathology (Justice, 2018).

The shared risk or confounding pathway (B) supposes that NPS as such are not causally related to cognitive decline, rather, both groups of symptoms reflect a combination of risk factors. For example, genes associated with general intelligence are inversely associated with AD and psychiatric symptoms (Savage et al., 2018). Additionally, amyloid status based on PET imaging has been suggested to associate with NPS and cognitive decline, while cognitive decline alone would not be sufficient to account for the NPS (Krell-Roesch et al., 2019).

Reverse causality or psychological reaction (C) suggests that the patient is aware of his or her cognitive symptoms, to which NPS are a result of adaptation to the decline (Geda et al., 2013). It has been reported that patients with mild AD are more often depressed than patients with moderate AD (Fitz & Teri, 1994; Piccininni et al., 2005). Affective reactions to cognitive decline could also be mediated by personality traits (Archer et al., 2007). Reverse causality could provide an explanation for depressive symptoms: a recent 28-year follow-up study reported that instead of increasing the risk of AD from midlife onwards, depressive symptoms were associated with increased risk of dementia in the decade prior to diagnosis (Singh-Manoux et al., 2017).

Finally, the interaction pathway (D) suggests that a pre-existing vulnerability in combination with NPS increase the risk of AD (Geda et al., 2013). In this vein, a study by Pink et al. (2015) found that all NPS were risk factors for future AD, but only the interactions of depression and apathy with APOE ε4 portended a higher risk. Burke et al. (2016) replicated the finding that all NPS are risk factors for AD, but found that NPS interactions with APOE ε4 did not increase the risk – surprisingly, APOE ε4 carriers with delusions and aberrant motor behavior were even at a decreased risk for AD (Burke, Maramaldi, Cadet, & Kukull, 2016).

2.2.1.2 Prevalence and burden

NPS are prevalent in MCI and AD (Zhao et al., 2016; Apostolova & Cummings, 2008; Lyketsos et al., 2002). In AD, apathy and depression are the most common early symptoms and psychotic and aggressive symptoms emerge later (Lyketsos et al., 2011). Table 5 presents meta-analytic prevalence

(47)

estimates of symptom domains (Zhao et al., 2016) with their incidence and persistence in a five-year follow-up study of people with AD (Vik-Mo et al., 2018). In addition to being common, NPS become more severe as AD progresses (e.g., Piccininni et al., 2005; Hashimoto et al., 2015), associate with higher costs of care (Herrmann et al., 2006; Schnaider Beeri, Werner, Davidson, & Noy, 2002), caregiver burden (Terum et al., 2017), and worse quality of life (Hongisto et al., 2018; Conde-Sala et al., 2016). Depressive, psychotic and agitative symptoms may also increase the risk of placement in 24-hour care and mortality (Peters et al., 2015; Okura et al., 2011).

Table 5. Meta-analytic and 6-year prevalence of NPS in AD (adapted from Zhao et al., 2016; Vik-Mo, Giil, Ballard & Aarsland, 2018)

Meta-analytic prevalence estimate

Potential explanations for vari-

ability of prevalence estimates 6-year preva- lence Delusions 31% (27-35%) Ethnicity, disease duration, age 59%

Hallucinations 16% (13-18%) - 40%

Agitation 40% (33-46%) Disease duration, age 68%

Depression 42% (37-46%) Age 88%

Anxiety 39 (32-46%) Assessment method, age 74%

Euphoria 7% (5-9%) Study setting, age 25%

Apathy 49% (41-57%) Assessment method, education,

MMSE, disease duration 91%

Disinhibition 17% (12-21%) Disease duration, age 59%

Irritability 36 (31-41%) Study setting, disease duration,

age 76%

Aberrant Motor

Behavior 32% (25-38%) Disease duration, study setting 69%

Sleep Disturbances 39 % (30-47%) - 58%

Appetite Disorders 34% (27-41%) Study setting, disease duration 70%

Parentheses indicate 95% confidence intervals for meta-analytic prevalence estimates. Sample size range was 5488–10526 in the meta-analysis (Zhao et al. 2016), and the five-year follow-up study was conducted in 106 AD patients (Vik-Mo, Giil, Ballard & Aarsland, 2018). 6-year prevalence refers to the proportion of individuals who had the NPS at least once (Neuropsychiatric Inventory score ≥ 1) over six visits covering baseline and five-years of follow-up. AD = Alzheimer’s disease, MMSE = Mini-Mental State Examination, NPS = Neuropsychiatric symptoms.

Neuropsychiatric symptoms in Alzheimer’s disease : measurement, follow-up and associations with activities of daily living

Dissertations in Education, Humanities, and Theology

TONI SAARI

Neuropsychiatric symptoms in Alzheimer’s disease:

measurement, follow-up and

associations with activities of

daily living

Neuropsychiatric symptoms in Alzheimer’s

disease: measurement, follow-up and

associations with activities of daily living

Neuropsychiatric symptoms in Alzheimer’s disease: measurement, follow-up and associations with activities of daily living

Acknowledgements

Table of contents

List of original publications

1 Introduction

2 Review of the literature