Does Personality Predict Behavioral and Psychological Symptoms of Dementia? Results from PACO Prospective Study

INTRODUCTION

Alzheimer’s disease (AD) is characterized by a progressive neurocognitive decline frequently accompanied by behavioral and psychological symptoms (BPS). BPS emerge at early stages of the disease and worsen with cognitive decline [1]. Among factors implicated in BPS occurrence, personality could play an important role. Clinical physicians have suggested that BPS may be linked to an exaggerated expression of prior personality [2]. The majority of studies dealing with the relationship between personality and dementia relies on the 5-factor model, which defines personality across five dimensions: openness, conscientiousness, extraversion, agreeability, and neuroticism. A large systematic review showed that neuroticism was the personality trait the most strongly associated with BPS [3]. Cross-sectional studies uncovered relationship between premorbid neuroticism and BPS with respect to anxiety, delusion, and hallucinations [3–5]. A longitudinal study assessing predictors of depressive symptoms in AD patients identified high neuroticism as a risk factor for subsequent depressive symptoms [6]. Links between BPS and other personality dimensions have also been suggested but are less conclusive. A low level of agreeableness has been correlated to higher levels of BPS in dimensions such as apathy and irritability [3, 7]. This relationship, however, is debated, others studies having shown that higher agreeableness can lead to hallucinations and increased apathy, irritability, aggressiveness, affective disturbances, and troublesome behavior in nursing home residents [5]. Many studies have pointed out that a high level of conscientiousness can protect against the risk of dementia [8, 9]. Surprisingly, only one study highlighted a link between a high level of conscientiousness and a decreased risk of overall behavioral disorders [9].

It is noteworthy that most of studies included patients with moderate to severe dementia. Less attention has been paid to prodromal or mild stages of the disease. A possible association between higher neuroticism, lower openness, and higher affective symptoms and apathy has been suggested in patients with mild cognitive disorder [10]. However, another study led on mild AD did not confirm the hypothesis of a personality influence on BPS [11]. This confusing array of results points to the need to better understand the influence of personality on BPS at early dementia stages. Literature on this important topic is weak since most of the studies were retrospective, with small sample, and focusing on the moderate and severe stages of AD. At the advanced stage, participants become unable to describe their prior personality and reports from their close relatives are biased.

To deal with these potential weaknesses, the present prospective study aimed at assessing the influence of premorbid personality traits on subsequent occurrence of BPS in a population of patients with prodromal or mild stages of AD, before any significant BPS. We conjectured that personality factors would have an impact on behavioral symptoms at the early stage of the disease. Given that neuroticism is known to affect the risk of developing psychiatric symptoms, we predict that this dimension will be associated with BPS, especially depression and anxiety. Of other dimensions, high conscientiousness, which has been associated with better cognitive functioning and lower risk of dementia outcome, could be a protective factor with respect to BPS [8, 10].

METHODS

RESULTS

A total of 237 patients were included in the PACO study. Among them, 198 patients were seen at 6-month, 181 at 12-month, and 176 at 18-month follow-up. Table 1 shows baseline demographic and clinical characteristics of the patients. The hierarchical organization of the personality dimensions found in our study follows the same order as in the majority of other studies.

The mean values of NPI scores at each visit are summarized in Table 2. The global NPI and the different NPI factors increased between baseline and 18-month follow-up, and the increase was significant for all BPS except for depression and sleep disorders. Principal components analysis identified a five-factor model with eigenvalues superior or near to 1 (eigenvalue respectively 2.73, 1.37, 1.26, 1.1, and 0.92) which together accounted for 61.3% of the variances in NPI scores. The pattern matrix is presented in Table 3. Factor 1 was identified as ‘psychotic’ symptoms (with factor loading >0.6 for delusions, hallucinations, aberrant motor behavior, and >0.4 for agitation), factor 2 as ‘affective’ symptoms (with factor loading >0.6 for anxiety and depression, and >0.4 for agitation and irritability), factor 3 as ‘behavioral dyscontrol’ (with factor loading >0.6 for elation and disinhibition), factor 4 as ‘apathy/appetite’ symptoms (with factor loading >0.6 for appetite disorders and near to 0.6 for apathy), and factor 5 as ‘sleep disorders’ (with factor loading >0.6 for sleepdisorders).

The association between each personality domain and the evolution of global NPI score and PCA factors is presented in Table 4. The analyses were adjusted on age, sex, educational level, and baseline MMSE score. A higher score of neuroticism was associated with a higher score on ‘affective’, ‘apathy/appetite’, ‘sleep disorders’, and global NPI score. A higher score of conscientiousness was associated with lower ‘psychotic’, ‘affective’, ‘apathy/appetite’, and global NPI scores. A significant relationship was shown between higher extraversion and lower NPI ‘affective’ and higher ‘behavioral dyscontrol’ scores. A higher level of openness was associated with a lower ‘affective’ score.

DISCUSSION

To our knowledge, the PACO cohort study is the first prospective and longitudinal study to specifically investigate the relationship between personality and future BPS in prodromal or mild AD. In connection with our expectations, our findings demonstrated an influence of personality on BPS progression independently of potential demographic and cognition confounders (age, sex, education, and baseline MMSE). We observed that neuroticism and conscientiousness showed the strongest and most reliable link to behavior manifestations.

Our analysis highlighted that high neuroticism was associated during the follow-up with a significant raise of NPI global score, affective and apathy symptoms, and sleep disorders. The relationship between neuroticism and affective disorder revealed in the present study confirms, for the first time with a prospective approach, the results of prior publications where neuroticism was identified as a determinant factor to predict overall BPS and affective disorders [3, 20]. Individuals with high level of neuroticism tend to be pessimistic, to negatively appraise life events, and to experience more negative feelings and emotions [19]. It is accepted that neuroticism is a risk factor for developing a range of psychiatric conditions, especially depression or anxiety, and that this personality trait is associated with personality disorders [20–22]. Furthermore, individuals scoring high in neuroticism have decreased coping ability with inadequate responses to stressors. Cognitive impairment could be considered as a stressor itself with bidirectional effect. On one hand, affective troubles observed in neurotic individuals could be the expression of maladaptive thoughts patterns and inadequate emotional reactions, when experiencing cognitive deterioration. On the other hand, cognitive decline could contribute to major the coping deficit. Highly neurotic individuals have also been described as needing greater efforts to regulate negative feelings and as failing to control their emotions [23]. Emotional dyscontrol could worsen with cognitive decline, leading to a higher risk of depression and anxiety.

A speculative possibility is that the impact of neuroticism on BPS could be explained by a premorbid latent psychiatric vulnerability, sometimes never expressed across lifespan, and revealed by the neurodegenerative process. Evidences from neuroimaging bring support to this hypothesis. An emerging body of neuroimaging studies revealed that personality dimensions may impact brain structure and connectivity [24–26]. The most robust findings implicated neuroticism which has been found to be negatively correlated with grey matter volumes in the prefrontal cortex, amygdala, and anterior cingulate cortex [24, 26]. High neuroticism is also suspected to enhance the aging effects on frontal areas, with an accelerated decrease of grey matter in these regions [26]. Recent findings using surface-based morphometry and cortical thickness measures reported an association between high neuroticism, thicker cortex, and decreased surface areas in prefrontal temporal regions [23, 25]. These anatomical variations are considered as the structural markers of specific neurodevelopmental trajectories associated with neuroticism [27]. Moreover, some studies suggested that a high level of neuroticism was linked to an altered connectivity between cortical and subcortical structures, including prefrontal cortex and amygdala [28, 29]. These connectivity abnormalities may compromise the top down regulation of emotion [23]. Thus, in neurotic patients suffering from AD, neurodegenerative lesions could potentiate pre-existing neurodevelopmental vulnerability, and then favor the early onset of affective disorder and emotion dysregulation.

Another interesting result of the PACO study is the association between apathy evolution and neuroticism. Archer [17], using stepwise regression analysis adjusted for potential confounders, found ‘agitation/apathy’ syndrome to be negatively associated with agreeableness, while Mendez-Rubio [10] identified openness as a predictor of apathy. No prior study identified a link between apathy and neuroticism. Lack of motivation is considered as a core clinical dimension of apathy [30]. Above-mentioned findings concerning neuroimaging pointed a neurodevelopmental vulnerability of neuroticism in many cerebral areas, including anterior cingulate cortex, known to be involved in motivation. Apathy in AD has been found to be correlated with grey matter atrophy in bilateral anterior cingulate cortex and frontal areas [31]. As discussed with affective symptoms, apathy could also be interpreted as a manifestation of a deleterious association between neurodegenerative lesions and neurodevelopmental vulnerability in strategic brain regions implicated in motivation.

The PACO study also provided a strong support for a positive mediating of conscientiousness on BPS, as well as extraversion to a lesser extent. The present data showed that high conscientiousness is related to lower evolution of global NPI score, psychotic symptoms, affective symptoms, and apathy. High extraversion was significantly linked to lower affective symptoms and higher behavioral dyscontrol. Individuals with high conscientiousness are efficient and organized, aimed for achievement, act dutifully, and show self-discipline. Conscientiousness is regarded as an important determinant of health-related outcomes, such as lower risks of affective diseases, cognitive decline, and risk of dementia [8, 32], but little evidence from previous researches have emerged concerning its link with BPS. High conscientiousness has been reported to be negatively related to overall behavior troubles [9], verbally aggressive behavior or depression [33]. Hallucinations or delusions have been previously found to be associated with high neuroticism [5], high openness [4], and high or low agreeableness [5, 17]. A novel finding from the present study is the negative association between conscientiousness and psychotic symptoms emergence in prodromal or mild AD. With regard of the discrepancy of previous results, our conclusions have to be interpreted with caution. However, they would be consistent with the cognitive profiles of subjects with high levels of conscientiousness. It is commonly assumed that conscientiousness is underpinned by executive function that promote thought monitoring, inhibitory control, and self-control [34]. It might be supposed that individuals with high conscientiousness monitor and inhibit more easily their inadequate thoughts, avoiding the risk of misinterpretations and delusions.

Unlike neuroticism, conscientiousness and extraversion seem to have a protective effect on affective symptoms. Chatterjee et al. [4] have already reported that patients with depressive features were described by their proxy as less extraverted, but no association between conscientiousness and affective symptoms has been described before. The negative association between high conscientiousness and affective symptoms occurrence could reflect greater efficacy to handle daily stressors [35].

In neuroimaging studies, higher conscientiousness appeared to be associated with larger orbitofrontal volume [26], and to be a resilience factor with regard to AD pathological lesions [36]. Besides, delusions were shown to be linked to orbitofrontal, anterior cingulate, temporal and subcortical cortex atrophy [31]. Given these findings, it could be speculated that the protective influence of high conscientiousness on BPS could be in part underpinned by neurological features that would contribute to curb the behavioral manifestations of AD at earlystage.

Compared to previous studies, PACO study has several strengths including the large sample of participants recruited, the prospective design, the detailed evaluation of personality, and AD diagnosis relying on standardized criteria. To our knowledge, our findings are the first to show a prospective link between personality dimensions and BPS evolution in prodromal and mild AD patients. They reinforce the hypothesis of a link between neuroticism and BPS occurrence, and show the potentially protective effect of conscientiousness.

This study has also several limitations. Self-assessment of personality in AD patients may be questionable for two reasons. On one hand, it has been shown that AD is accompanied by early personality changes, and on the other hand cognitive disorders such as anosognosia could skew the patients’ ability to metacognition especially at the severe stage of the disease [15]. It also has been argued that patients with AD suffering from mild AD, tend to describe their personality as it was before the clinical onset because of self-image deficits updating [15, 37]. Therefore, we consider that personality evaluation in PACO study to be reliable. Another limitation concerns the low prevalence and incidence of BPS. This low prevalence is explained by the prospective design of the study which required not including patients, with meaningful BPS, in particular productive disorders. Thus, the impact of personality in patients with severe BPS at early stages of AD could not have been evaluated. Moreover, our results are observed in AD patients, and they should not be generalized to a premorbid sample in the general population. The diagnosis was based on clinical criteria, including clinical visit with a neurologist, geriatrist or psychiatrist, neuropsychological tests, and MRI biomarker. Cerebrospinal fluid biomarkers were available for a limited number of patients only. The systematic assay of cerebrospinal biomarkers could have ensured a better diagnosis certainty but probably have hindered the feasibility of the study. Concerning analyzes, some statistical links between personality traits and BPS evolution were close to 0.05. This would deserve reproducing the study in a larger cohort with a longer follow-up to confirm the results with lower significance thresholds. Lastly, the participants’ follow-up was relatively short in PACO study. A longer follow-up would have possibly provided more links between personality traits and BPS evolution.

In summary, the present study based on a prospective design showed that high neuroticism in patients suffering from prodromal or mild AD was associated with a higher risk of developing behavioral and psychological manifestations. Conversely, high conscientiousness, extraversion and agreeableness seemed to be associated with a lower risk of subsequent BPS. This study raises evidences that in prodromal and mild stage of AD, not only neurological factors but also psychological mechanisms may contribute to the arising of BPSD. Additional studies on a larger cohort of patients followed-up over a longer duration, would be needed to confirm these results. Our findings have clinical implication, by highlighting the need to better take into consideration the premorbid personality in people with prodromal or mild AD to improve the management of BPS and held to set up preventive care.