The prevalence and the course of neuropsychiatric symptoms in patients with dementia

Dementia is prevalent in Norway and other countries and is hallmarked by a decline in memory and other cognitive abilities. In addition to cognitive decline, the vast majority of patients with dementia experience neuropsychiatric symptoms (NPS) such as depression, agitation, psychosis, apathy, or irritability. In this review, we describe the prevalence and the course of NPS in patients with dementia, referring to results of population-based studies, studies of outpatients, and studies of patients in long-term care. For a better comparison of the included studies and for clarification, we have included studies that have assessed NPS with the Neuropsychiatric Inventory (NPI). Overall, we found a high prevalence rate of NPS. At least one NPS (NPI > 0) was present in 56% to 98% of the patients and 62% to 84% of the patients had at least one clinically significant NPS (NPI > 3). The NPS with the highest prevalence rate were apathy, irritability, agitation, depression and anxiety; while delusion, hallucination, disinhibition, aberrant motor behavior and euphoria were the least frequent NPS.


INTRODUCTION
Dementia is a syndrome consisting of several symptoms and hallmarked by a decline in cognition that leads to functional impairment.In order to diagnose dementia by the ICD-10 criteria, a decline in memory, a decline in other cognitive abilities, a decline in emotional control or motivation, or a change in social behaviour should be present.The patient's awareness should be preserved, and the symptoms should be present for more than six months.Dementia is progressive and can be caused by several different diseases, such as Alzheimer's disease, Lewy Body dementia, Parkinson's disease, frontotemporal dementia, cerebrovascular diseases, and dementia due to excessive alcohol use.The impact of dementia on quality of life, on caregiver burden, and on the use of resources is substantial (1,2).
The hallmark of dementia is cognitive decline, but accompanying neuropsychiatric symptoms (NPS) of dementia, such as apathy, depression, agitation and delusion, are prevalent.NPS are often more devastating and create more discomfort for patients and their caregivers than do the cognitive deficits, and studies have shown an inverse relationship between NPS, especially depression, and quality of life (3,4).NPS are often treated with psychotropic drugs, in spite of limited evidence for efficacy and a considerable risk of adverse events (5).
In this article, we will refer to studies on the prevalence and the course of NPS in patients with dementia.We will include population-based studies, studies describing NPS in outpatients, and studies of patients living in long-term care.To be able to compare prevalence rates of NPS across studies, we have focused on studies assessing NPS with the 10-item version of the Neuropsychiatric Inventory (NPI -10).We present prevalence rates of individual neuropsychiatric symptoms according to four sub-syndromes; agitation, affective symptoms, psychosis, and apathy.

Search strategy
We have searched the electronic databases of MEDLINE from inception to June 2011.We used the following medical subject headings (MeSH-terms) and free-text words to identify articles on the prevalence and the course of neuropsychiatric symptoms in population-based studies, studies of outpatients, and patients in long-term care: "dementia" combined with "neuropsychiatric inventory" or "NPI" combined with "neuropsychiatric symptoms" or "behavioral and psychological symptoms of dementia" combined with "prevalence" or "course" combined with "populationbased studies" or "outpatients" or "nursing homes" or "long-term care".We included only articles in English.In total 108 citations was retrieved, and the abstracts were reviewed by the two authors.We also searched the reference lists of the retrieved articles for other relevant articles.The search resulted in four population based studies, 10 studies including out-patients and five studies including nursing home patients that are presented in this paper.

Assessment scales for NPS
Due to the cognitive decline in patients with dementia, loss of insight, and language and communication impairments, most of the patients, particularly at severe  (7).Individual symptoms are rated as present or not during the last four weeks; subsequently, the frequency (1-4) and the severity (1-3) of the symptoms are rated.The frequency score and the severity score are multiplied into a sub-score (0 to 12 points, with the higher score indicating more severe symptoms), and the 12 sub-scores are added (min 0, max 144 points).Sub-scores above 3 on an individual item are regarded as clinically significant symptoms (8).
Based on factor analyses of the NPI, four clusters have been identified; agitation (agitation/aggression, disinhibition, and irritability), affective symptoms (depression and anxiety), psychosis (delusion and hallucination), and apathy (9,10).In patients with Parkinson's disease dementia (PDD) four clusters have been identified (11); mood (depression, anxiety and apathy), psychosis (delusions and hallucinations), agitation (agitation) and apathy (apathy).In studies with longitudinal design, authors have described persistence and resolution of NPS."Persistence" is defined as the ratio of patients with Clinically Significant NPS (CS-NPS) at follow-up to patients with CS-NPS at the previous assessment, and "resolution" is the ratio of patients without CS-NPS at follow-up to the patients with CS-NPS at the previous assessment.

NPS in persons with dementia, population-based studies
We identified four cross-sectional studies reporting prevalence rates for NPS assessed by the NPI (12)(13)(14)(15) 3).To be able to compare the results between studies, we will refer to the prevalence rates for NPS defined as an NPI score > 0. The prevalence rates of at least one NPS ranged from 56% to 88%, (12)(13)(14)(15).Most of the studies reported an increase in prevalence rates as the dementia becomes more severe (14).

Psychosis Prevalence:
The prevalence rates for symptoms of psychosis differed among the four studies, and ranges of prevalence rates were 12% to 27% (median 18.5%) for delusion and 8% to 15% (median 12.5%) for hallucination.Course: Only one population-based study has investigated the course of NPS in patients with dementia.In the Cache County study by Lyketsos et al., the patients were assessed four times over five years (16).Results are presented as prevalence rates for the remaining patients in the cohort at each assessment, showing an increasing prevalence rate for both delusion (18% to 38%) and hallucination (10% to 24%) (16).
Affective symptoms Prevalence: Prevalence rates for depression ranged from 22% to 38% (median 28%).Prevalence rates for anxiety ranged from 17% to 25% (median 22.5%).Course: In the Cache County study, there was an increase in the prevalence rates of affective symptoms in the remaining patients during the five-year follow-up period: anxiety (14% to 32%) and depression (29% to 47%) (16).

Apathy Prevalence:
The prevalence rates of apathy in the five samples ranged from 27% to 57% (median 44.5%).
Course: There was an increase in the prevalence rate of apathy in the Cache County study from 20% at baseline to 51% after 5 years (16).

NPS in persons with dementia attending outpatient clinics
Results from studies assessing the prevalence rates of NPS in outpatient samples applying the NPI are summarized in table 2 (9,(17)(18)(19)(20)(21)(22)(23)(24)(25).Four studies (9,17,23,24) presented the proportion of patients who experienced at least one NPI symptom (> 0), which ranged from 88% to 98%.Correspondingly, Benoit et al. found that 88% of the patients experienced at least one clinically significant NPI symptom (> 3) (25).The study of Aalten et al., in which one of the inclusion criteria was having at least one clinically significant (> 3) NPI symptom, pooled several datasets, and some of the participants in other studies noted in the table may have been included in this study (18).Toyota et al. compared patients with early-onset dementia to patients with late-onset dementia and found that delusions, hallucinations, agitation/aggression, disinhibition, and aberrant motor behavior were more common in lateonset dementia (19).Longitudinal studies indicate that virtually all persons with dementia experienced NPS during the follow-up.One French study demonstrated that the prevalence of NPS overall increased over a four-year follow-up (26).Longitudinal studies using assessment scales other than the NPI demonstrate heterogeneity in the course of symptoms and that NPS may arise at any stage of dementia.However, agitation seems to be the most persistent symptom and shows an increasing severity with disease progression in a majo-rity of the studies, whereas the persistence of psychosis and depression shows greater variability (27)(28)(29)(30).It should be noted that apathy was investigated rarely in these studies.

Psychosis
Prevalence: The prevalence rates for symptoms of psychosis differed among the eight studies (NPI>0) and ranged between 22% and 51% (median 30.5%) for delusion and 10% and 24% (median 15.5%) for hallucination.Aalten (2007) and Benoit (2003) found lower prevalence applying NPI>3 as a threshold value.A higher prevalence of both delusions and hallucinations was associated with increasing severity of dementia (22,23).Course: Studies applying the NPI have reported minor changes in psychosis during up to two-year follow-ups (31).However, a French study reported decreased severity of delusions (32).A later French study, where 151 patients were followed for four years, showed that the prevalence of hallucinations, but not delusions, increased over the follow-up period (26).

Affective symptoms
Prevalence: For the affective symptoms of depression and anxiety, the prevalence rates (NPI>0) ranged between 30% and 68% (median 48%) and 36% and 54% (median 41%), respectively.In the studies applying NPI>3 as a threshold value Aalten (2007) found that prevalence rates of 37% for both depression and anxiety, whereas Benoit (2003) found that 40% experienced depression, whereas 37% had anxiety symptoms.Mega et al. found that increased severity of depression and anxiety was associated with increasing dementia severity (24).Course: Aalten et al. found a decrease in severity of depressive symptoms over a two-year follow-up (31).

Agitation
Prevalence: The prevalence rates of the individual symptoms in the agitation sub-syndrome (NPI>0), presented in decreasing order, were irritability (range 25% to 66%; median 42%), aberrant motor behavior (range 31% to 47%; median 40.5%), agitation/aggression (range 29% to 60%; median 37.5%), and disinhibition (range 13% to 36%; median 25.5%).Using NPI>3 as a threshold value Aalten (2007) and Benoit (2003) found that the figures of irritability and agitation/aggression were inside the range of the other studies, whereas the figures for aberrant motor behavior and disinhibition were lower than the range in the other studies.More severe agitation/aggression, disinhibition, and aberrant motor behavior were associated with increasing severity of the dementia (20,(22)(23)(24).Course: Agitation symptoms were among the most persistent and aberrant motor behavior, and disinhibition increased during follow-up (31,32).Gonfrier et al. reported that the prevalence of aberrant motor activity, agitation/aggression, and disinhibition increased over a four-year follow-up (26).

Apathy
Prevalence: Apathy showed the highest prevalence in six of the eight studies (NPI>0), ranging from 42% to 74% with a median of 61.5%.The studies that failed to find apathy as the most common symptom were conducted in Asia.Correspondingly, apathy was the most common symptom in the two studies using NPI>3 as a threshold value.The severity of apathy was associated with increasing severity of dementia (20,23,24).Course: Apathy was among the most persistent symptoms and showed increasing severity during the follow-up (26,31,32).

NPS in persons with dementia living in long-term care
The prevalence rates of clinically significant NPS, defined as a score > 3 on an individual NPI item, for patients in long-term care settings have been presented in several studies in recent years (see table 3) (33)(34)(35)(36)(37).The proportion of patients with at least one NPI item score > 3, was 68% to 84% (median 72%), depending on the severity of the dementia and the etiologic diagnosis.A large study in Norwegian nursing homes reported an overall prevalence of NPS of 55% in mild dementia (CDR = 1), 70% in moderate dementia (CDR = 2) and 84% in severe dementia (CDR = 3) (34).Several studies report the course of NPS in patients in nursing homes, but the studies show heterogeneity in the course of the symptoms as well as the symptoms showed intermittent courses.Psychotic symptoms remained constant over time, while the prevalence rates of aberrant motor behavior, depression, anxiety, and euphoria declined and apathy, agitation/aggression, irritability, and disinhibition increased over time (38).

Psychosis
Prevalence: The prevalence rates of delusion vary across the different studies, with a range of 9% to 26% (median 16%).Hallucination was approximately half as frequent as delusion, with prevalence rates ranging between 3% and 14%, and a median of 8%.Course: The course of the psychotic symptoms over one year was described in two studies (39,40).In one of the studies, a significant increase in the prevalence rates of delusion and hallucination was found over the one-year follow-up (39), while there was no change in psychotic symptoms in the other study (40).Wetzels et al. described a resolution rate over a six-month period of 50% for hallucination and 64% to 88% for delusion, indicating that psychotic symptoms have a fluctuating course (35).The same pattern was described by Bergh et al., who found a resolution rate close to 40% for delusion and hallucination over a four-month period (36).
Affective symptoms Prevalence: Prevalence rates of affective symptoms were comparable across the seven studies: depression 8% to 22% (median 20%) and anxiety 17% to 26% (median 22%).Course: The resolution rates for depression in the four studies reporting the course of NPS in nursing home patients were extremely divergent, ranging from 30% to 100% depending on the cohort, the severity of the dementia, and the time interval between the assessments, making a summary of the results difficult (35,36,39,40).The resolution rates of anxiety were slightly more consistent, ranging from 33% to 75%.
Course: One study reported that approximately half of the patients with symptoms of agitation at inclusion were symptom-free after one year (39).As the time interval between the assessments decreased, the resolution rates for symptoms of agitation/aggression decreased as well, being approximately 35% in a study with assessments every fourth month (36).

Apathy
Prevalence: Six of the seven studies reported prevalence rates for apathy, which is a frequent symptom in persons with dementia, ranging from 19% to 34% and a median of 26%.
Course: In the largest Norwegian nursing home study, there was a resolution rate for apathy of 47% over 1 year (40), while studies with assessments of apathy every fourth or sixth month reported a resolution rate of 30% to 47% and 45% to 64%, respectively (35,36).

DISCUSSION
In this review, we have described the prevalence and the course of neuropsychiatric symptoms (NPS) in patients with dementia, based on population-based studies, studies of outpatients, and studies of patients in long-term care.We found a high prevalence rate of NPS in patients with dementia; 56% to 88% of patients in population-based studies had at least one NPS (NPI > 0) and 74% to 95% of outpatients had at least one NPS (NPI > 0), while 68% to 84% of nursing home patients had at least one clinically significant NPS (NPI > 3).Overall, it has been estimated that up to 90% of patients with dementia will experience at least one clinically significant NPS through the course of the disease (41).The NPS with the highest prevalence rates at all three levels of care was apathy, while irritability, agitation/aggression, depression, and anxiety were slightly less frequent than apathy.Delusion, hallucination, disinhibition, aberrant motor behavior, and euphoria were the least frequent NPS in patients, independent of the level of care received by the patients who were recruited to the study.Delusions were considerably more frequent than hallucinations.Although we have reviewed several studies and summarized the prevalence rates in them as medians and ranges, there are differences between individual studies with the same inclusion criteria.Further, comparison of prevalence rates between different levels of care is difficult.Nevertheless there are some trends.Patients in nursing homes have higher prevalence rates of all NPS, except for apathy, than patients in population-based cohorts.One explanation for this could be due to better treatment of apathy in nursing homes.The differences in severity of dementia between different levels of care may also explain differences in NPS prevalence rates between the cohorts, as there are increasing prevalence rates of NPS with increased severity of the dementia (34).Nevertheless, a recent review focusing on depression in dementia concluded that the prevalence rates of depression did not change across different stages of dementia (42).For psychotic symptoms, the prevalence rates of the patient's symptoms increase during the first three years of the disease, but later they don't change (43).
Studies from nursing homes in Norway generally report higher prevalence rates of psychotic symptoms than studies from nursing homes in other European countries.There are several possible explanations for this: 1) there are differences between the prevalence rates of psychosis in Norway and the rest of Europe; 2) the concept of delusion and hallucination is interpreted differently in the Norwegian version of the NPI compared to other versions of the NPI; and 3) the treatment of psychosis is less effective in Norway, resulting in higher prevalence rates.When looking at the prescription rates of antipsychotics in Norway and in the Netherlands, they are comparable (33,35,36,44).On the other hand environmental aspects and how nursing homes are organized in the two countries differ, which may explain the differences in prevalence rates of psychotic symptoms.Also, patients in Dutch nursing homes have more severe dementia compared to patients in Norwegian nursing homes (33,35,36,44).
The studies included in our review have different inclusion criteria.Ten studies have included patients independently of etiology of dementia diagnosis, and nine studies have included only those patients with a diagnosis of Alzheimer's disease.Generally, studies including only patients with Alzheimer's disease report higher prevalence rates of depression, anxiety, and apathy than studies that include patients independently of etiological dementia diagnosis.Psychotic symptoms seem to be less prevalent in patients with Alzheimer's disease than in patients with other etiologies of dementia.Previous studies have shown that as many as 50 to 80% of patients with Lewy Body dementia (LBD) or Parkinson's disease with dementia have visual hallucinations (45,46).Disinhibition is also frequent in LBD, found in 65% of the patients (47).Depression, emotional lability and apathy is common in vascular dementia (VaD), especially depression, while psychosis is less prevalent in VaD than in other types of dementia (48).Depression and anxiety are found to be equally common in VaD as in AD (49).In patients with frontotemporal dementia (FTD), NPS such as apathy, disinhibition, irritability and aberrant social behaviour dominate their dementia while psychotic symptoms are very rare in patients with FTD (47).The differences in the prevalence of NPS in different types of dementia may be explained by underlying biological mechanisms.Low levels of serotonin in the CNS are known to be associated with depression or anxiety.Studies with functional neuroimaging demonstrate serotonergic dysfunction both in patients with AD and FTD, but the location of the serotonergic dysfunction in the brain varies between studies (50).In patients with Parkinson's disease and LBD, psychosis is thought to be a result from overstimulation of dopaminerigic receptors as well as an imbalance of serotonin and acetylcholine (51).
The presented review has some limitations.We have included only studies where the NPS is assessed by the NPI, which limits the number of included studies.The majority of the studies included in our review were from settings in North America and Europe, which is a paradox, as the majority of patients with dementia are living in countries in other parts of the world.There is a trend that studies from Asian countries report lower prevalence rates for symptoms of apathy and affective symptoms than studies from other countries (21,23).There are differences in the criteria of a presenting NPS, which contributes to differences in the prevalence rates between the studies.It may also preclude the possibility of following the disease trajectory, in terms of prevalence rates, in patients with dementia in population-based samples through outpatients with dementia to patients with dementia in nursing homes.
What kinds of studies are needed in order to further increase our understanding of NPS in patients with dementia?More knowledge on the course of NPS is needed, and large, longitudinal cohort studies following patients with dementia from the beginning of their dementia disease until end-stage dementia should be planned.As most previous studies are on patients with Alzheimer's disease, future studies should be large enough to allow analyses of subgroups of dementia diagnoses as vascular dementia, Parkinson's disease with dementia, Lewy Body dementia, and frontotemporal dementia.The prevalence and course of NPS are influenced by the treatment given to the patients, and data on pharmacological treatment should be collected to allow for adjustments in the statistical analyses.

CONCLUSION
Neuropsychiatric symptoms (NPS) are prevalent in patients with dementia, and the prevalence rates increases with increasing severity of the dementia.There is also evidence that the prevalence rates of NPS are higher in nursing-home populations than in population-based samples.The most prevalent NPS are apathy, irritability, agitation/aggression, depression, and anxiety.Although NPS are prevalent, the patient's individual symptoms show a fluctuating course.Policy-makers and clinicians should be aware of the high prevalence rates of NPS and the fluctuating course of individual symptoms to better plan for further treatment and care of the elderly patients with dementia.

Table 1 .
Prevalence rates (%) of neuropsychiatric symptoms in studies with population based samples assessed by the neuropsychiatric inventory.
(6)resent neuropsychiatric symptom = NPI > 0, Clinically significant neuropsychiatric symptoms = NPI > 3 1 DSM-IV 2 DSM-III-R stages of the disease are unable to rate their own symptoms.Hence, several assessment scales for NPS are based on observations of the patients by caregivers or relatives.The Neuropsychiatric Inventory (NPI) is a well-known global assessment scale widely used in research and clinical practice(6).The NPI originally consisted of 10 items (NPI-10): delusions, hallucinations, depression/dysphoria, anxiety, agitation/

Table 3 .
Prevalence rates (%) of clinically significant neuropsychiatric symptoms in nursing home patients assessed by the neuropsychiatric inventory.