Abstract
Background:
Presence of behavioral and psychological symptoms of dementia (BPSD) is very common in subjects with cognitive impairment, representing an important determinant of disease progression, institutionalization, and worse prognosis. Knowledge of the prevalence and correlates of BPSD in community-living old subjects with cognitive impairment is limited so far, but it is essential for establishing specifically tailored care and cure in such a vulnerable population.
Objective:
With this study, we aimed at investigating, in a large sample of old age subjects with cognitive impairment, BPSD prevalence and correlates including the main demographic, clinical, and socio-environmental characteristics.
Methods:
Data were gathered from the ReGAl project (Rete Geriatrica Alzheimer; Geriatric Network on Alzheimer’s disease), a large longitudinal Italian multicentric clinical-based study, promoted by the Italian Society of Gerontology and Geriatrics (SIGG).
Results:
We evaluated data from 4,157 old-age subjects affected by mild cognitive impairment (MCI) (541; 13%) or dementia (3616; 87%). 85.2% of all the population presented with at least one BPSD. Using a factor analysis, we identified four factors of BPSD: psychotic, affective, maniac, and impulse control behaviors. Logistic regression analyses revealed that among the main demographic, clinical, and socio-environmental aspects considered, only comorbidity was associated with all factors, independently of multiple covariates.
Conclusion:
Identification of BPSD is crucial in everyday clinical practice and necessary to develop specific interventions and to define appropriate outcomes in their management. BPSD occur in a complex psychopathological context, influenced by several demographic and environmental factors that must be taken into account for a correct diagnosis and treatment.
INTRODUCTION
Behavioral and psychological symptoms of dementia (BPSD)— defined as signs and symptoms of disturbed perception, thought content, mood, or behavior [1]— are very common in subjects with cognitive impairment. They can affect up to 90% of subjects with dementia over the course of their illness [2], with a high interindividual variability [3], which strongly impacts on caregivers’ stress as well [4, 5]. Also in the early stages of cognitive impairment, neuropsychiatric symptoms are frequent in old age subjects, with an estimated rate of 35–85%, even in those with mild cognitive impairment (MCI) [6, 7]. Moreover, in this vulnerable population, the presence of BPSD represents the main distressing outcome and a leading cause for a more rapid cognitive decline, earlier institutionalization, and increased morbidity and mortality [8 –10]. Taken together, all these factors significantly contribute to the social and economic burden of dementia worldwide. For their prevalence, the severe negative impact on individual health status, and use of healthcare resources, BPSD have emerged as an important focus of research and intervention. Recent literature shows that BPSD are not independent, rather a group or cluster of related symptoms, providing evidence for potential clinical benefits in treating groups of BPSD over individual symptoms [11]. However, the presence of BPSD varies widely among populations and different settings, which imposes limitations in adopting a specific algorithm for their assessment andtreatment.
Identification of correlates associated with BPSD also play an important role and several studies [12] have demonstrated that both severity of dementia and male gender [13, 14] are the most associated factors. Other correlates such as socio-demographic factors, including education, previous job activity, marital and socio-environmental status, have been less explored. Moreover, these studies have been often conducted in young old subjects, in relatively small populations and, mainly, in nursing home setting. For this reason, evidence among very old subjects in community-based studies is limited.
In light of such evidence, with this study we aim to investigate, in depth and in a large population of old age subjects affected by cognitive impairment at different severity stages, the prevalence of BPSD, the factor structure, and their correlates among demographic and socio-environmental aspects.
MATERIALS AND METHODS
Study population
Data were gathered from the ReGAl project (Rete Geriatrica Alzheimer-Geriatric Network on Alzheimer’s disease), a large longitudinal Italian multicentric clinical-based study, promoted by the Italian Society of Gerontology and Geriatrics (SIGG), and focused on cognitive impairment and dementia in old age subjects. The project has been described extensively elsewhere [15]. From 5,165 subjects, with the exclusion of those cognitively healthy or with subjective memory complaints, we analyzed data of 4,157 subjects with cognitive impairment, older than 65 years. The diagnosis of dementia was based on DSM-IV [16] and on Petersen’s criteria [17] for mild cognitive impairment (MCI). Alzheimer’s disease (AD) was diagnosed according to the NINDS-ADRDA criteria [18], vascular dementia (VaD) on the NINDS-AIREN criteria [19], frontotemporal dementia (FTD) on the criteria described by Neary et al. [20], dementia with Lewy bodies (LBD) on the criteria by McKeith et al. [21], and Parkinson dementia (PDD) according to Emre et al. [22]. The severity of dementia was rated by the Clinical Dementia Rating scale (CDR) [23]. All subjects underwent a complete physical and neurological examination, neuropsychological evaluation, and an interview about medical conditions and social and familial status. Subjects with clinically severe psychiatric or systemic disease, intellectual disability, and severe sensory impairment (blindness, deafness) were excluded.
Socio-demographic evaluation
All subjects underwent a complete evaluation of the general socio-demographic status. In particular information about marital status (single, widow/er, married, divorced, or religious), social status (living alone, with one family member, with more than one family member, with a private assistance, or in a residential facility), previous job activity (worker, housekeeper, farmer, merchant, craftsman, employee, teacher, professional, or other), have been collected. Educational level was measured as the highest grade achieved at school (in years) and subsequently categorized into five groups, according to the division of the Italian educational system: (1) 0–4 years; (2) 5 years; (3) 6–8 years; (4) 9–13 years; and (5) more than 13 years. We considered “psychoactive drug user” a subject who was taking at least one among anxiolytics, antidepressants, hypnotics, antipsychotics, and anti-dementia drugs.
Comorbidity
Comorbidity was assessed using the Cumulative Illness Rating Scale (CIRS) [24, 25]. This rating scale consists of 14 items covering heart, hypertension, vascular and respiratory disorders, a combined eye-ear nose-throat item, upper and lower gastrointestinal systems, hepatobiliary system, kidney, genitourinary diseases, musculoskeletal diseases, endocrine/metabolic disorders, neurological system, and behavioral-psychiatric disorders. Severity in each single item is rated according to the following algorithm: 1 = no, 2 = mild, 3 = moderate, 4 = severe, 5 = life-threatening. No subject obtained a score of 1 in our sample. After completion of CIRS, two summary measures can be constructed: the illness Severity Index (CIRS-SI), which reflects the overall severity of diseases and the average rating of the 13 CIRS items, with the exclusion of behavioral-psychiatric disorders; and the Comorbidity Index (CIRS-CI), computed by counting the number of items with a score≥3 (moderate to severe pathology). As a result, the CIRS-CI can be considered as the number of clinically relevant concomitant diseases.
Cognitive, functional, and neuropsychological assessment
Cognitive performances were evaluated with the Mini-Mental State Examination (MMSE), as a test of general cognition [26], and then with a large battery of specific tests evaluating different cognitive areas [15]. To avoid the underestimation of a self-rated level of functional capacity, an informant-based rating of functional status was carried out using the BADL [27, 28] and the IADL scales [29]. BPSD were evaluated by the Neuropsychiatric Inventory questionnaire (NPI) [30], a fully structured caregiver interview measuring the following symptoms: delusions, hallucinations, agitation, depression, anxiety, elation, apathy, disinhibition, irritability, aberrant motor behavior, sleeping and eating abnormalities. The composite score (frequency×severity, FxG) of each subscale ranges between 0 and 12, and the total composite score between 0 and 144. An NPI score≥1 for any symptom was considered entry criteria. Current depressive symptoms were assessed by the 15-item version of the Geriatric Depression Scale (GDS) [31].
Statistical analyses
The observed data were normally distributed (Shapiro-Wilk W-Test) and are presented as means±standard deviation (SD). To assess differences among groups, unpaired t test, ANOVA, or Pearson’s Chi squared test were used, as appropriate. Principal component factor analysis with varimax rotation was used to examine the joint variation and interdependencies among NPI variables in all sample. Four factors were identified and the association between each factor and the main demographic, clinical, and socio-environmental characteristics was examined by logistic regression analyses. The odds ratio (OR) was determined to estimate the strength of the association. All p values are 2-tailed and level of significance was set at p≤0.05. Statistical analyses were performed using the SPSS 20 software package (SPSS, Inc., Chicago, IL).
RESULTS
Sample characteristics
All subjects (n = 4,157) had a mean age of 78.9±5.9 years (age range: 65–101) with a prevalence of women (65.8%). 541 (13%) subjects were affected by MCI (CDR = 0.5) and 3,616 (87%) by dementia (specifically 1683 with mild dementia CDR 1, 1,250 with moderate dementia CDR 2, and 683 with severe dementia CDR≥3). Among subjects with dementia, 2,398 (66.3%) were diagnosed as AD, 638 (17.6%) as mixed dementia, 331 (9.2%) as VaD, 130 (3.6%) as FTD, 52 (1.4%) as DLB, 36 (1%) as PDD, and 31 (0.9%) as secondary dementia. Clinical and demographic characteristics of all sample population and subjects stratified by severity of cognitive impairment are displayed in Table 1. In all population, 2396 subjects (57.6%) used at least one psychoactive drug, mostly antidepressants, and anxiolytics.
Demographic and clinical characteristics of sample population stratified by severity of cognitive impairment
Data are shown as mean±standard deviation (SD) or frequencies where appropriated. * χ 2 = 43.435; ** χ 2 = 64.177. MMSE, Mini-Mental State Examination; GDS, Geriatric Depression Scale; CIRS-CI, Cumulative Illness Rating-Comorbidity Index; CIRS-SI, Cumulative Illness Rating-Severity Index; BADL, Activities of Daily Living; IADL, Instrumental Activities ofDaily Living; CDR 0.5, mild cognitive impairment; CDR 1, mild dementia; CDR 2, moderate dementia; CDR ≥3, severe dementia.
Prevalence of BPSD and correlates
3,543 (85.2%) of all population presented at least one BPSD, namely 356 (65.8%) of MCI, 1,442 (85.7%) of mild dementia, 1,134 (90.7%) of moderate dementia, and 611 (89.5%) of severe dementia. The mean NPI score (expressed as FxG) was 21.26±16.51 (range 1–110). In all sample population, the most frequent symptoms were depression (59.6%), anxiety (54.2%), and apathy (51.9%). All frequencies increased with increasing of dementia severity, except for anxiety and depression (Supplementary Table 1). Subjects with BPSD had significantly lower education level, lower MMSE, GDS, BADL, and IADL scores and higher comorbidities, as compared with subjects without disturbances (data not shown). Among subjects with BPSD, women (21.6±16.6 compared with 20.3±16.2 in men, p = 0.027), and psychoactive drug users (23.2±17.0 compared with not users 18.1±15.1, p < 0.0001) had a higher NPI score (FxG). The majority lived with a family member (53.3%), mostly widower (41.9%), with a lower education level (38.2%; 0–4 years), housekeeper if woman (48.6%) and worker (24.2%) if man.
Factor analysis
Using a factor analysis with varimax rotation on NPI sub-items, we identified four factors of BPSD (Table 2): Psychotic behaviors (delusions, hallucinations, agitation, and irritability) loaded on factor 1; Affective behaviors (depression, anxiety, and apathy) had the strongest loading on factor 2; Maniac behaviors (elation and disinhibition) loaded on factor 3; and Impulse control behaviors (aberrant motor behavior, sleeping and eating disorders) loaded on factor 4. On this basis, in our population we found that 2,375 subjects suffered from psychotic, 3,057 affective, 649 maniac, and 1,767 impulse control behaviors. No statistically significant differences in age and gender distribution were found among groups (p≥0.05). In all population, subjects with psychotic behaviors were mostly housekeepers (χ 2 = 24.991, p = 0.002) and living with a family member (χ 2 = 19.013, p = 0.001). Subjects with affective behaviors were mostly married or widow/ers (χ 2 = 11.968, p = 0.018). Subjects with maniac behaviors lived mostly with a family member (χ 2 = 11.128, p = 0.025). Subjects with impulse control behaviors were more often housekeepers (χ 2 = 19.435, p = 0.013). Lower education was found among all factors of BPSD, psychotic (χ 2 = 13.991, p = 0.007), affective (χ 2 = 12.079, p = 0.017), maniac (χ 2 = 10.307, p = 0.036), and impulse control behaviors (χ 2 = 34.902, p < 0.0001).
Factor loadings based on a principle component analysis with varimax rotation for NeuropsychiatricInventory (NPI) variables
Correlates of BPSD factors
Logistic regression analyses (Table 3) were constructed with each factor of identified BPSD as dependent variable and age, gender, psychoactive drug use, type of dementia, cognitive impairment severity, CIRS score, education, and socio-environmental characteristics (including previous job activity, marital as social status) as covariates. They revealed that: 1) advanced age, male gender, psychoactive drug use, higher severity of cognitive impairment, and higher comorbidity index were independently associated with the presence of psychotic behaviors; 2) younger age, psychotic drug use, and higher comorbidity index were independently associated with the presence of affective behaviors; 3) male gender, higher severity of dementia, higher comorbidity index, and marital status (being never married) were associated with the presence of maniac behaviors; 3) advanced age, psychoactive drug use, higher severity of cognitive impairment, higher comorbidity index, lower education, and social status (living with a private assistance) were independently associated with the presence of impulse control behaviors.
Correlates of identified BPSD factors by logistic regression analyses in all population sample
Gender indicated as F = 0 and M = 1; psychoactive drug use indicated as users = 1 or not users = 0; type of dementia indicated as AD = 1; FLD = 2; LBD = 3; PDD = 4; VaD = 5; Mixed dementia = 6; Secondary dementia = 7; dementia severity indicated as CDR 0.5 = 0; CDR 1 = 1; CDR 2 = 2; CDR ≥ 3 = 3. Education indicated as 0–4 years = 1; 5 years = 2; 6–8 years = 3; 9–13 years = 4 and more than 13 years = 5; previous job activity indicated as worker = 1; housemaker = 2; farmer = 3; merchant = 4; craftsman = 5; employee = 6; teacher = 7, professional = 8; other = 9; Marital status indicates as single = 1, married = 2, widow/er=3, separated = 4, religious = 5; Social status indicated as living alone = 1, with one family’s member = 2, with more than one family’s member = 3, with a private assistance = 4; in a residential facility = 5.
DISCUSSION
There is an overall agreement that BPSD are very common regardless of the type of dementia and present during the course of the disease. Even in the early stages of cognitive impairment, neuropsychiatric symptoms are frequent and associated with a more rapid progression, from mild to severe dementia [32]. BPSD are a source of significant distress and poor quality of life for both patients and their caregivers [33]. Moreover, increased number of BPSD negatively correlates with survival rates, as examined in an over three-year study [34], showing that the presence of psychosis in AD is associated with increased mortality and acceleration of cognitive decline [35, 36].
Cross-sectional studies report a prevalence ranging from 50% to 100% [37] largely depending on the considered type of sample and setting. In our community-based population, we found that 85% of subjects with cognitive impairment present at least one neuropsychiatric symptom and 58% is under treatment with one or more psychoactive drugs. In accordance with the previous literature [2], depression, anxiety, and apathy were found to be the most frequent symptoms that are often difficult to cure. In fact, drug selection requires careful consideration for possible adverse effects and pharmacological interactions. Many antidepressants are effective for depressive and anxiety disorders, and choice should be based on their safety profile, even if there is little evidence of the effectiveness of individual agents in people with dementia. Apathy also is a common disorder but often overlooked or misdiagnosed. It is primarily a deficit in motivation and often presents as a comorbid feature of other clinical conditions [38]. Thus, if, on the one hand, BPSD are often difficult to treat pharmacologically, on the other hand, use of psychotropic medications represents an important challenge for the clinician because of an inappropriate prescription, sometimes cause of adverse events. Current guidelines recommend non-pharmacological interventions as the first-line treatment, followed by the less harmful medications for the shortest time.
Recent evidence supports the observation that BPSD onset or presence are not independent factors, rather they can be considered as a group or cluster of related symptoms [39]. Many correlates including age, gender, type, and severity of dementia have been shown to affect their expression that clinicians should consider before starting a specific treatment. After grouping BPSD in four distinctive factors (psychotic, affective, maniac, and impulse control behaviors), we found that among the main demographic, clinical, and socio-environmental correlates considered, only comorbidity index and age are constantly associated with all behavioral factors. It is known that patients with dementia have many additional chronic diseases (comorbidities) [40, 41] that can accelerate the progression of dementia and worsen its prognosis [42]. In fact, the presence of comorbidities lead to extended hospital accesses and stays, increased health care costs and mortality rates [43, 44]. Moreover, although comorbidities are equally common in older people with or without dementia, people with dementia are less likely to be diagnosed for other multiple pathologies. Accordingly, we found a lower comorbidity index in subjects with more severe dementia. Thus, the presence of comorbidities in patients with dementia requires a careful evaluation, prioritizing some health aspects and outcomes over others, and taking into account possible conflicts between multiple treatments and recommendations. In fact, in treating comorbid conditions in people with dementia, drug interactions, adverse drug reactions, and worsening of underlying conditions are very critical in clinical practice.
Despite previous studies failed to demonstrate a significant effect of age on prevalence of BPSD, we found that advanced age is more significantly associated with psychotic and impulse control behaviors, while younger age with affective and maniac behaviors. These data denote the importance of a different intervention approach in these groups (younger and older) in the community. Some studies on the relationship between BPSD and severity of dementia showed that they were more pronounced with worse cognitive performances [45]. A significant association was found between dementia severity and presence of agitation and physical or verbal aggression while only depression was associated with less severe stages of dementia. Also in our cohort severity of dementia was significantly associated with psychotic, maniac, and impulse behavioral symptoms while a higher prevalence of affective behaviors was found in subjects with MCI.
Gender also affects clinical presentation of behavioral disorders, which suggests a better management with a personalized program. In fact, aggressiveness and aberrant motor behavior have been more frequently reported in men whereas in women depressive/anxious symptoms and verbally agitated help-seeking behaviors are observed more often [44, 46]. Accordingly, we found that men suffered more frequently from psychotic behaviors.
Education plays certainly a role in BPSD expression, even if such an association is still unclear and under debate. Previous studies reported a negative relationship between education and prevalence of BPSD in AD [45, 32] or a lack of association [45], while we observed that lower education is associated with impulse control behaviors, independently of multiple covariates.
Marital status has been poorly investigated in relation to BPSD. Previous studies conducted in nursing homes found that patients who lived alone before admission showed a higher prevalence of psychosis than those who were married [47]. We were unable to confirm such an association, but we found that being “never married” is associated with maniac control behaviors. Only a previous finding on marital status has been published [48] showing that being never married is associated with “psychosis and behavior dysregulation” in a community based older adults with dementia.
We extended previous findings in this field including into the analyses social-environmental aspects and found that living with a private assistance leads to a higher frequency of impulse control behaviors. Previous studies performed in nursing home settings showed higher levels of agitation in cognitively impaired subjects [45]. In our free-living setting, we also found this strong association with aberrant motor behavior but also a high frequency of sleep and eating disorders. These results underline the importance of social environment that represents a modifiable factor with important practical consequences. Living with paid assistance is associated with the presence of aberrant motor behavior, sleep disorder, and, most importantly, with eating disorders. A skilled and trained paid caregiver is hard to find; he/she is often a person with no experience in assisting a subject with dementia and this may cause a stressful relationship between caregiver and patient. An improvement in social interaction and family contact may reduce these symptoms and further research is needed to investigate such arelationship.
This study has several strengths. Due to its large sample size, it reports important observations on BPSD prevalence. The large and representative sample allowed us to perform logistic regression analyses on reasonably large subgroups. This type of analysis is more appropriate than linear regression models or correlation analyses used in previous studies since logistic regression analysis has the power to correct for multiple confounding variables, and most importantly, it is not based on linear assumptions but can give an odds ratio adjusted for all the covariates included in the model. Moreover, we also included the use of psychoactive medications as a covariate, to control association analyses for its confounding effect. However, caution is needed in extending these findings to the general population because the study sample was recruited in specialized clinical settings, what might have introduced a selection bias. Finally, the cross-sectional design of the study may not be the optimal one, since behavioral disturbances can fluctuate.
In conclusion, this community-based study shows a high prevalence of BPSD mainly associated with age, comorbidity burden, and dementia severity. Identification of BPSD is crucial in clinical practice for the development of interventions and for setting clearly expected outcomes. Our results confirm that BPSD are not independent phenomena but they occur in a complex psychopathological context, what further supports a syndromic approach to their diagnosis and management. The psychopathological profile of each syndrome is highly variable across patients and the co-occurrence of sub-syndromes is common, reflecting a complex and multi-level interaction between BPSD. Management strategies, in favor of non-pharmacological treatments, together with care-focused training and education approaches, can be effective in BPSD modulation but require a concerted multidisciplinary team input.
