Personality, Alzheimer's disease and behavioural and cognitive symptoms of dementia: the PACO prospective cohort study protocol

Alzheimer’s disease (AD) is characterised by both a progressive cognitive impairment and Behavioural and Psychological Symptoms of Dementia (BPSD) [1]. BPSD are frequent in AD. They may occur at any stage of the disease and progressively worsen as cognitive function declines. Besides the apathy and depression frequently seen in patients with mild to moderate dementia, some of the other BPSD such as agitation, irritability, delusions, aberrant motor behaviour and hallucinations may be extremely upsetting for both the patient and his caregivers [2‐4]. When they occur early in the evolution of the disease, their frequency and the intensity of these BPSD can lead to increased difficulties in carrying out everyday tasks [5], an increasingly rapid decline in cognitive function [6], a reduction in the quality of life of both the caregiver and patient [7], an increased likelihood of depression in the caregiver, an increased risk of hospitalisation, premature institution placement [8] and increased health costs [9].

The difficulty in caring for patients with AD is more closely linked to the existence and intensity of the BPSD than the actual cognitive decline. Current treatment for these BPSD is often difficult but is generally approached by medications and/or psycho-social support therapies that still need to be validated.

The prevalence of BPSD has been shown to be increased by a certain number of factors: they may exist at any stage of the disease but become progressively worse as cognitive abilities decline. Evolution of the disease and the level of cognitive impairment have been identified by several authors [4]. A better understanding of the risk factors for the onset of BPSD would allow us to better target those patients at risk, thus favouring earlier treatment and so limiting their consequences on patients, caregivers and the community. The BPSD are the combined result of neurodegenerative lesions [10], somatic comorbidities, iatrogenic effects and also psychological vulnerability. Clinical experience suggests that premorbid personality could play a role in the onset of BPSD with a possible link between personality traits and specific BPSD.

Nevertheless, very few studies have evaluated the relationship between premorbid personality, life events and BPSD. Even childhood trauma could have consequences for the psychosocial adaptation of the elderly [11] and so favour the onset of BPSD [12]. A correlation between a previous history of neuroticism and depression [13, 14] or delusions [15] has been observed in patients with AD. If a high level of social interaction appears linked to an increased risk of agitation, the degree of extraversion does not influence the onset of BPSD [16].

In a previous study, we assessed the influence of previous psychiatric history, premorbid personality and family context on the emergence and typology of BPSD in 99 patients in crisis situations [17]. In this retrospective study, the prevalence of delusions, anxiety, irritability and eating disorders were higher in the patient group with premorbid personality disorders. Those with a history of paranoiac, schizoid or schizotypal problems were more irritable and presented more delusional tendencies. Depression was present in one out of three patients with obsessive-compulsive, avoidant or dependent personality disorders. Finally, eating disorders were more frequent in patients with antisocial, histrionic or narcissistic personalities.

Most of studies assessing the relationship between personality and BPSD are based on the five factors model that defines personality using five dimensions: openness, conscientiousness, extraversion, agreeability and neuroticism. A cross sectional study in 208 patients with moderate to severe Alzheimer’s disease showed an association between a previous history of neuroticism or agreeableness and BPSD, especially for previous neuroticism and anxiety and for previous agreeableness, agitation and irritability [18]. Moreover, a systematic review found a positive relationship between neuroticism and challenging behaviour in 72% of the studies, in particular, mood, aggression and overall challenging behaviour [19]. However, these retrospective studies lacked representative samples and were impacted by confounding factors and flaws in statistical analyses. Finally, a recent study including 52 MCI subjects found a relationship between premorbid neuroticism and openness and affective behavioural symptoms and apathy [20].

Overall, the methodology used in the previous studies makes interpreting results difficult, particularly as most studies are retrospective and include only a small number of patients, often with late stage AD. Such parameters can lead to bias in a patient’s self-evaluation since cognitive decline invariably leads to modification in the ability to express one’s own personality. A second area of bias comes from collecting data retrospectively from relatives. This implies subjective judgements, in an effort to preserve the image of the loved one, it is easy for family to minimise their BPSD or on the contrary to overestimate these behavioural problems and their effects on the family.

Social cognition is a cognitive domain involved in the detection and the recognition of social and emotional messages, as well as integration in a personal emotional context and a social complex context. It concerns one of the major aspects of our cognitive functioning allowing us to interact with our environment. A key component of social cognition is the ability to recognize facial emotion expressions, to identify and to judge mental state and the intentions of the others. It has been shown that impairment in the ability to recognise facial expressions is correlated to social behaviour in various psychiatric and neurological pathologies [21, 22]. Recent work has already shown that patients with AD also had impaired ability to recognise facial expressions [23]. Thus, impairment of social cognition ability in AD could influence the onset and the course of BPSD. Moreover, performance in social cognition tests like facial recognition and perspective taking could be linked to certain personality traits and the subsequent risk of BPSD.

Furthermore, imaging markers could help to understand the link between brain structures changes and the presence of BPSD. The neuropsychiatric symptoms observed in AD patients could be mainly due to frontal structural abnormalities [24]. More specific regional associations with a range of behavioural symptoms have been identified: apathy was associated with frontal structures [25]; delusions were correlated with frontal, parietal, and temporal structures [26]; depressive symptoms with thalamus, lentiform nucleus, and medial temporal cortex [27]; agitation was associated with temporal and frontal structures [28] and amygdale atrophy was related to aberrant motor behaviour, with potential relationships to anxiety and irritability [29]. These same regions participate in the recognition of social messages, particularly those of facial expression, perspective taking and empathy [30]. Consequently, measurement of MRI markers may allow us to better understand the link between personality, social cognition and BDSP and the implication of the anatomic structures in these different domains.

We propose to assess personality in patients with prodromal or mild AD, and then to collect BPSD data prospectively with the objective to evaluate the causality link between personality and BPSD occurrence. Social cognition performance and cortical atrophy will also be measured at baseline.

The PACO study aims to better understand the factors underlying neuropsychiatric symptoms in patient with baseline prodromal or mild AD. This prospective multidisciplinary project brings together specialists in old age psychiatry, geriatric and neurology. Its main objective is to evaluate the effect of premorbid personality on the phenomenology and the frequency of BPSD. Premorbid personality will be assessed with the Five Factors Model, which is considered as the most widely used dimensional personality approach [53].

Several studies suggest that personality may influence the behavioral course of AD [19]. It has been yielded that persons with high neuroticism experience have more behavioral disturbances when measured with NPI [18]. The association between personality and specific neuropsychiatric symptoms is so far not fully established and should be better specified. For instance, high levels of neuroticism have been sometimes linked to depression [54] or not [18]. Most studies in this field are retrospective and include few patients, often in advanced stages of the disease. It can induce a substantial risk of recall bias, especially when personality is assessed by informants. The longitudinal design of PACO project, including more than 250 AD patients at prodromal or mild stage, can control some of these biases. Self-report of personality has been preferred to observer rating. In healthy people, this methodology is considered to be more accurate and trustworthy because observers do not have access to a person’s internal motives [55]. In the context of cognitive decline, self-assessment of personality could be criticized. However, Archer et al. showed that the NEO-FFI, a personality questionnaire based on the Big Five Inventory, can be used reliably to measure premorbid personality in AD patients [56]. Moreover, it has been shown that when asked to evaluate their current personality, persons with mild AD tended to describe their former personality [57], maybe due to an inability to update their self-image [58].

In PACO study, we are expecting to find correlations between personality dimensions and the risk of being affected by BPSD. We hypothesize that high level of neuroticism and low level of conscientiousness will be more specifically associated with BPSD, in particular affective domains such as depression, anxiety and apathy. As cognitive decline could be positively or negatively influenced by personality, this potential association will be studied [59, 60]. Moreover, as personality development is considered to be influenced by environmental factors, childhood and adult life events will be collected to assess their possible influence on the course of AD. This would allow us, for example, to highlight the possible effects of previous exposure to a trauma or childhood deficiencies.

Subjects included in PACO study will perform a comprehensive social cognition battery including facial emotion recognition tasks. Several publications have highlighted that AD patients experience an increasing deficit in facial emotional expression recognition as the disease progresses [23], but little is known concerning the very early stage of AD [61]. Consequences of impaired facial emotion recognition on behavioral troubles in AD are questioned and should be clarified [62]. Thus, a secondary objective of the PACO study is to compare social cognition ability between MCI and mild AD, and to identify potential relationships between social cognition abilities and BPSD occurrence during the patient’s follow-up.

PACO project includes MRI 3D sequence at baseline. Different studies have linked impaired emotion recognition in psychiatric or neurologic disease to disturbances or damages of the fronto-temporal brain network [63‐65]. Social cognition parameters could be correlated to initial regional cerebral volume and/or the atrophy observed one year later, which could itself be correlated to a higher risk of developing BPSD. For instance, some orbito-frontal atrophy may predict the occurrence of BPSD. Recent evidence suggests that personality traits may be related to neuroanatomical structures. Jackson et al. [66] found that higher neuroticism was associated with specific brain regions atrophy, whereas higher conscientiousness was related to larger regional volumes in healthy aging. It could be hypothesised that an interaction between personality dimension, regional brain atrophy, and BPSD may exist in AD.

The PACO study may have several limitations. Firstly, personality should ideally have been assessed before AD onset. However, it would have been necessary to include a very large number of patients with a long follow-up. We have chosen to include patients at early stage of AD: at this stage we can suppose that the personality assessment reflects premorbid personality. Secondly, a longer follow-up would allow us to detect a larger number of BDSP. The absence of long-term follow-up could underestimate the relationship between personality and BDSP. Indeed, a certain number of SCPD increase with the evolution of the disease, but we hypothesized that some of them occur early in patients with specific personality traits. Thirdly, at baseline only patients with no or mild behavioural disturbances are selected. This inclusion methodology excludes subjects with personality dimensions associated to early behaviour disorder. Finally, the AD diagnosis was based on clinical and MRI criteria. Dubois et al. criteria refer to both imaging and to CSF markers [67]. This would allow better patient selection. However, in our centres, CSF markers are available for a small percentage of the patients, in particular when the diagnosis remains unclear. On the opposite, patients included in the PACO study had a typical clinical presentation, and CSF markers were not available for the majority of them. A systematic lumbar puncture should have limited the feasibility of our study.

The present study also has strengths. In most former studies personality assessment was performed by relatives and only a few included patients at prodromal or mild stage of AD. Moreover, to our knowledge, PACO is the first study to evaluate personality and social cognition performance and their link with subsequent BDSP.

The results of this research may help clinicians to better indentify, from a combined assessment of personality, performance in social cognition tests and the degree of cerebral atrophy, patients with AD at enhanced risk of developing BPSD and to improve patient and family care. Demonstrating a link between premorbid personality and behavioural problems will also allow us to argue for a veritable cooperation between neurologists, geriatricians and psychiatrists in the patients care throughout the course of the disease.

This study also aims to better understand the repercussions of early stage AD on social behaviour, identify the neuroanatomical correlates of BPSD. It should allow us to set up a regional data base to include clinical, neuropsychological, psychiatric, neuroradiological and biological data with longitudinal follow-up of the cohort, and then further inter-regional and inter-speciality cooperation in studying and dealing with Alzheimer’s disease.

Isabelle Rouch, MD, PhD, senior epidemiologist, Jean-Michel Dorey, MD, old age psychiatrist, Nawèle Boublay, MSc, epidemiologist, Marie-Anne Henaff, MSc, researcher, Florence Dibie-Racoupeau, MD, old age psychiatrist, Zaza Makaroff, MD, geriatrician, Sandrine Harston, MD, neurologist, Michel Benoit, MD, PhD, old age psychiatrist, Marie-Odile Barrellon, MD, geriatrician, Denis Fédérico, MD, geriatrician, Bernard Laurent, MD, PhD, Head of university neurology unit, Catherine Padovan PhD, neuropsychologist, Pierre Krolak-Salmon, MD, PhD, Head of university geriatric unit.