Aging & Mental Health, 2015 Vol. 19, No. 3, 247 257, http://dx.doi.org/10.1080/13607863.2014.924900
Agitation-associated behavioral symptoms in mild cognitive impairment and Alzheimer’s dementia Stefan Van der Musselea,b, Nathalie Le Bastarda, Jos Saerensc, Nore Somersc, Peter Mari€enc,d, Johan Goemanc, Peter P. De Deyna,c,e,f and Sebastiaan Engelborghsa,c* a
Laboratory of Neurochemistry and Behavior, Reference Centre for Biological Markers of Dementia (BIODEM), Institute Born-Bunge, University of Antwerp (UA), Antwerp, Belgium; bDepartment of Nursing and Midwifery Sciences, Faculty of Medicine and Health Sciences, University of Antwerp (UA), Antwerp, Belgium; cDepartment of Neurology and Memory Clinic, Hospital Network Antwerp (ZNA), Middelheim and Hoge Beuken, Antwerp, Belgium; dDepartment of Clinical and Experimental Neurolinguistics (CLIN), Vrije Universiteit Brussel, Brussels, Belgium; eDepartment of Health Care Sciences, Artesis University College Antwerp, Antwerp, Belgium; f Department of Neurology and Alzheimer Research Center, University Medical Center Groningen, Groningen, The Netherlands (Received 3 November 2013; accepted 13 April 2014) Objectives: The aim of this study is to determine the prevalence of agitation in mild cognitive impairment (MCI, Petersen’s criteria) and patients with Alzheimer’s dementia (AD), and to characterize the associated behavioral symptoms. Method: A cross-sectional analysis of baseline data from a prospective, longitudinal study on behavioral symptoms was performed, including 268 MCI and 393 AD patients. Behavioral assessment was performed through Middelheim Frontality Score (MFS), Behavioral Pathology in Alzheimer’s Disease Rating Scale (Behave-AD) and Cornell Scale for Depression in Dementia (CSDD). Agitated behavior was considered to be clinically relevant when one or more items of the CohenMansfield Agitation Inventory (CMAI) occurred at least once a week. Results: The prevalence of agitation in AD (76%) was higher than in MCI (60%; p < 0.001). Patients with agitation showed more severe frontal lobe, behavioral and depressive symptoms (MFS, Behave-AD and CSDD total scores). In agitated AD patients, all behavioral symptoms and types of agitation were more severe compared to non-agitated AD patients, but in agitated MCI patients only for diurnal rhythm disturbances. This resulted in more severe Behave-AD global scores in patients with agitation as compared to patients without agitation. Comparing MCI and AD patients, MCI patients with agitation showed more severe behavioral and depressive symptoms than AD patients without agitation. The structure of agitation in AD consisted of more aggressive and physically non-aggressive behavior than in MCI. Conclusion: Frontal lobe, behavioral and depressive symptoms are more severe in MCI and AD patients with clinically relevant agitation as compared to patients without agitation. However, this association is less pronounced in MCI. Keywords: neuropsychiatric symptoms; Cohen-Mansfield Agitation Inventory (CMAI); aggressiveness; aggression; MCI; AD
Introduction Behavioral and psychological dementia-related problems increase caregiver burden in mild cognitive impairment (MCI) (Bruce, McQuiggan, Williams, Westervelt, & Tremont, 2008; Frank et al., 2006; Garand, Dew, Eazor, Dekosky, & Reynolds, 2005; Garand et al., 2007) and dementia due to Alzheimer’s disease (AD) (Levy, Lanctot, Farber, Li, & Herrmann, 2012) and this might affect the quality of care for patients and the quality of life of relatives and caregivers (Hazzan, Ploeg, Shannon, Raina, & Oremus, 2013). Moreover, these neuropsychiatric manifestations, such as aggression, depression and hallucinations, are consistent predictors for institutionalization (Gaugler, Yu, Krichbaum, & Wyman, 2009; Luppa, Luck, Brahler, Konig, & Riedel-Heller, 2008). The thorough understanding of behavior in MCI and AD is necessary, as behavioral features may represent risk factors for MCI or predictors for the progression of MCI to AD (Monastero, Mangialasche, Camarda, Ercolani, & Camarda, 2009). In addition, better behavioral understanding may be important for (non-)pharmacological behavioral research and management, aiming to reduce
institutionalization and to improve the quality of life for patients, relatives and caregivers. Agitation belongs to the behavioral and psychological signs and symptoms of dementia (Finkel, Costa e Silva, Cohen, Miller, & Sartorius, 1996) and is defined as inappropriate verbal, vocal or motor activity that is not explained by apparent needs or confusion (Cohen-Mansfield & Billig, 1986). The inappropriateness of the activity refers to it being repetitive, abusive/aggressive and/or not according to social standards. Furthermore, four categories of agitation can be distinguished: aggressive physical, non-aggressive physical, aggressive verbal and nonaggressive verbal (Cohen-Mansfield & Billig, 1986). Agitated behavior is common in dementia with a prevalence of 24% 45% in non-institutionalized older adults with dementia (Chan, Kasper, Black, & Rabins, 2003; Lyketsos et al., 2000, 2002; Margallo-Lana et al., 2001), and up to 85% in institutionalized dementia patients (Margallo-Lana et al., 2001; Pitkala, Laurila, Strandberg, & Tilvis, 2004; Selbaek, Kirkevold, & Engedal, 2007; Zuidema, Derksen, Verhey, & Koopmans, 2007). However, agitation prevalence percentages seem to vary
*Corresponding author. Email: [email protected] Ó 2014 Taylor & Francis
248
S. Van der Mussele et al.
greatly, depending on patient selection and assessment methods. For the same reasons, the prevalence of agitation in MCI varies even more with 5% 25% in populationbased and 4% 45% in hospital-based studies (Apostolova & Cummings, 2008; Monastero et al., 2009). An agitation syndrome in AD is repeatedly found as a principal behavioral component beside a mood and a psychosis factor (Cummings, McRae, & Zhang, 2006; Frisoni et al., 1999; Hollingworth et al., 2006; Kang, Ahn, Kim, & Kim, 2010; Mirakhur, Craig, Hart, McLlroy, & Passmore, 2004; Spalletta, Baldinetti et al., 2004; Spalletta, Musicco et al., 2010; VilaltaFranch et al., 2010). Furthermore, aggressive behavior in patients with dementia is associated with internal factors such as depression, psychosis and pain (Cipriani, Vedovello, Nuti, & Di, 2011). To our knowledge, the relationship, between clinically relevant agitation and associated behavioral symptoms in MCI and as compared to AD, has not been investigated. This study aims to test the hypotheses that agitated patients display more frontal lobe, depressive and other neuropsychiatric symptoms, such as psychosis, activity, sleep and affective disturbances and anxiety/phobias, as compared to non-agitated patients; and that agitated AD patients have more associated behavioral symptoms than agitated MCI patients, as behavioral symptoms are more present in AD (Van der Mussele et al., 2012). Methods Study population and diagnostic criteria The study population consisted of MCI patients (n D 268) and AD patients (n D 393). Patients were consecutively included in our memory clinic at the moment of their diagnostic work-up, consisting of a general physical and neurological examination, blood screening, structural neuroimaging, standard electroencephalogram and an extensive time-linked (§3 months) neuropsychological examination with adjustment for age and education, comprising amongst others the Mini-Mental State Examination (MMSE) (Folstein, Folstein, & McHugh, 1975) and the Wechsler Memory Scale III (The Psychological Corporation, 1998), Hierarchic Dementia Scale (Cole & Dastoor, 1987) and/or Repeatable Battery for the Assessment of Neuropsychological Status (Randolph, Tierney, Mohr, & Chase, 1998). To diagnose MCI, Petersen’s diagnostic criteria (2004) were applied. Objective cognitive impairment for age and education was quantified as a performance of more than 1.5 standard deviation below the appropriate mean on the neuropsychological subtests. As all cognitive domains of subjects were tested in an extensive timelinked (§3 months) neuropsychological examination, all MCI patients were categorized as amnestic/non-amnestic, single/multiple domain. The MCI group (n D 268) consisted of 53 amnestic single domain, 149 amnestic multiple domain, 29 non-amnestic single domain and 37 nonamnestic multiple domain MCI patients. Probable AD was diagnosed according to NINCDS/ ADRDA criteria (National Institute of Neurological and
Communicative Disorders and Stroke Alzheimer’s Disease and Related Disorders Association) (McKhann et al., 1984), though all patients as well fulfilled the DSM-IV (dementia criteria of the Diagnostic and Statistical Manual of Mental Disorders, fourth edition) criteria (American Psychiatric Association, 1994). Clinical and neuropsychological follow-up of included patients and autopsy in deceased consented patients during follow-up period, did contribute to the diagnostic accuracy of the subjects in this study. The AD group consisted of 35 definite and 358 probable AD patients. Staging of cognitive deterioration was assessed by means of the Global Deterioration Scale (GDetS) (Reisberg, Ferris, de Leon, & Crook, 1982). To get a cursory idea about how psychotropic drug intake might have affected our study results, we investigated the difference in behavior between treated and untreated patients with agitation by comparing the total scores of the behavioral assessment scales. MCI and AD patients were categorized as treated when they took medication from at least one of the psychotropic drug categories mentioned in Table 1. The local ethics committee approved this study. All patients and/or patients’ caregivers gave written informed consent. All patients were of Caucasian origin. Behavioral assessment All subjects underwent in-depth behavioral assessment at inclusion (baseline) consisting of an interview of both patient and caregiver, covering a period of two weeks prior to inclusion. In case a non-professional caregiver was lacking, the patient’s main professional caregiver was contacted and interviewed. The interview was performed by the clinician or researcher who was not blinded for the subject’s cognitive diagnosis. The battery of behavioral assessment scales comprised: Cohen-Mansfield Agitation Inventory (CMAI), Middelheim Frontality Score (MFS), Behavioral Pathology in Alzheimer’s Disease Rating Scale (Behave-AD) and Cornell Scale for Depression in Dementia (CSDD). The MFS is a validated clinical and behavioral assessment scale that measures frontal lobe features and reliably discriminates frontotemporal dementia (FTD) from AD patients with a sensitivity and specificity of almost 90% and with good inter- and intra-rater reliability (Aries et al., 2010; De Deyn et al., 2005). According to the Instructions for Administration and Scoring, the MFS was obtained by summating scores in a standardized fashion on 10 items. Each item was scored either 0 (absent) or 1 (present) yielding a total maximal score of 10. The 10 items scored are noted in Figure 1. The presence of frontal lobe symptoms was considered to be clinically relevant in case of a total MFS score of 5 (De Deyn et al., 2005). The Behave-AD is a 25-item scale that measures behavioral symptoms in seven clusters (Figure 2), scored on a four-point scale of increasing severity (Reisberg et al., 1987). A psychosis cluster score is calculated by summating the scores of the paranoid/delusional ideation (delusions) and hallucinations clusters. Besides a total score, a
Aging & Mental Health
249
Table 1. Demographic, clinical and neuropsychological data. MCI
AD Agitation
Total n D 268 % Male/Female
Agitation
No n D 108
Yes n D 160
40.3%
59.7%
Total n D 393
No n D 96
Yes n D 297
24.4%
75.6%
26/70
103/194
119/149
46/62
73/87
129/264
Age at inclusion (years)
75.5 § 8.1 (50 94)
76.3 § 8.6 (50 94)
75.0 § 7.7 (53 90)
80.1 § 7.6 (51 97)
81.8 § 8.1 (51 97)
79.6 § 7.3 (54 97)
Age at onset (years)
73.2 § 8.4 (48 90)
73.9 § 8.6 (48 90)
72.7 § 8.2 (51 87)
76.9 § 8.4 (48 96)
79.2 § 8.7 (48 96)
76.1 § 8.2 (50 95)
Disease duration (years)
2.2 § 1.7 (0 11)
2.2 § 1.8 (0 11)
2.2 § 1.7 (0 10)
3.2 § 2.4 (0 14)
2.6 § 1.8 (1 10)
3.4 § 2.6 (0 14)
GDetS (1 7)
3.0 § 0.7 (1 5)
3.1 § 0.8 (1 5)
3.3 § 0.7 (1 5)
5.1 § 1.0 (2 7)
4.9 § 0.9 (2 7)
5.2 § 1.0 (3 7)
MMSE score (/30)
25.6 § 2.9 (16 30)
25.7 § 2.9 (16 30)
25.6 § 2.9 (16 30)
15.1 § 6.3 (0 30)
16.5 § 5.9 (0 30)
14.7 § 6.4 (0 28)
52.4
50.0
54.1
15.6
16.7
15.2
37.5
40.4
33.3
50.5
Free of psychotropic medication (%) Antidepressants (%) Antipsychotics (%) Hypnotics, sedatives, anxiolytics (%) Cholinesterase inhibitors (%) Antiparkinsonian agents (%) Antiepileptics (%)
24.8
20.2
28.0
39.7
4.7
6.6
3.3
46.3
28.5
29.5
27.8
23.4
20.8
24.2
3.4
5.7
1.9
37.7
44.2
35.6
2.7
2.8
2.6
1.5
3.1
1.0
2.3
2.8
1.9
1.5
1.0
1.7
Note: Data are given as ratio, percentage or mean § SD with ranges represented between brackets. Significant differences comparing the MCI group with the AD group and the groups with and without agitation within the MCI group and within the AD group are mentioned as follows: p < 0.01, p < 0.001. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. For comparison of male female ratios and percentages, chi-square statistics were used. For other comparisons, Mann Witney U test was applied. The level of significance was set at p < 0.01. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease; GDetS D Global Deterioration Scale.
global score on a four-point scale of increasing severity is provided, reflecting how troubling to the caregiver or dangerous to the patient the behavioral symptoms are, from not troubling or dangerous (score 0) to severe (score 3). Depressive symptoms were assessed by means of the CSDD, a 19-item depression scale (Alexopoulos, Abrams, Young, & Shamoian, 1988a). Item scores range from 0 (absent) to 2 (severe), with a maximum total score of 38 points. The presence of significant depressive symptoms was defined as a total CSDD score >7 (Burns, Lawlor, & Craig, 2004). Studies have shown the CSDD to be valid for screening depression in non-demented patients too (Alexopoulos, Abrams, Young, & Shamoian, 1988b). The CMAI assesses 29 agitated behaviors on a seven-point scale of increasing frequency/severity (1 D never; 7 D several times an hour) (Cohen-Mansfield, 1996). CMAI cluster scores include aggressive behavior (10 items), physically non-aggressive behavior (11
items) and verbally agitated behavior (8 items); a total score is provided as well. Agitation was considered to be clinically relevant when one or more items occurred at least once a week (any individual item score 3); and aggressive, physically non-aggressive and verbally agitated behavior was considered to be clinically relevant when one or more items within the respective cluster occurred at least once a week (Choy, Lam, Chan, Li, & Chiu, 2001; Cohen-Mansfield, Marx, & Rosenthal, 1989; Gruber-Baldini, Boustani, Sloane, & Zimmerman, 2004; Suh, 2004; Testad, Aasland, & Aarsland, 2007; Zuidema et al., 2007; Zuidema, de Jonghe, Verhey, & Koopmans, 2010). Statistical analyses The (behavioral) assessment scales provide semicontinuous variables. Therefore, non-parametric statistics were used:
250
S. Van der Mussele et al.
Figure 1. Prevalence (%) of frontal lobe symptoms in MCI and AD patients. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. Chi-square statistics were used, comparing the groups with and without agitation within the MCI and AD groups. Note: Significant differences are mentioned as follows: p < 0.01, p < 0.001. The level of significance was set at p < 0.01. Abbreviations: MFS D Middelheim Frontality Score; MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease.
Aging & Mental Health
251
Figure 2. Prevalence (%) of behavioral symptoms in MCI and AD patients. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. Chi-square statistics were used, comparing the groups with and without agitation within the MCI and AD groups. Note: Significant differences are mentioned as follows: p < 0.01, p < 0.001. The level of significance was set at p < 0.01. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease.
252
S. Van der Mussele et al.
Mann Whitney U test was applied to compare (semi)continuous variables, chi-square statistics for categorical data and Spearman rank to correlate (semi)continuous data. To reduce the type I error rate, probability levels of 0.01 were considered significant. Statistical analyses were carried out using Statistical Package for the Social Sciences (SPSS) Statistics 17.0. Results The AD group comprised a higher percentage of women than the MCI group. MCI patients were younger at inclusion and disease onset and their disease duration was shorter. In AD patients, GDetS scores were higher and MMSE scores were lower as compared to the MCI group. Table 1 shows that there is no difference in gender, age, disease duration, GDetS, MMSE and psychotropic drug intake between MCI patients with and without agitation. In contrast, agitated AD patients had a younger age at onset, a longer disease duration yet, higher GDetS scores and used more antipsychotics than non-agitated AD patients. The prevalence of agitation in AD patients (76%) was higher than the prevalence of agitation in MCI patients (60%; p < 0.001). Our study also subcategorized agitation in aggressive, non-aggressive and verbally agitated behavior. The prevalence of aggressive behavior, as measured by the CMAI, was 14% in AD patients and 1% in MCI patients (p < 0.001). The prevalence of physically non-aggressive behavior was 60% in AD patients and 24% in MCI patients (p < 0.001). The prevalence of verbally agitated behavior was 58% in AD patients and 50% in MCI patients (p D 0.039). Agitation, as defined in this study, had a different structure in MCI and AD patients. Agitated MCI patients displayed less aggressive behavior (1% versus 19%; p < 0.001) and less physically non-aggressive behavior (41% versus 80%; p < 0.001) than agitated AD patients. Verbally agitated behavior was present in 83% of the agitated MCI patients and in 76% of the agitated AD patients (p D 0.095). In MCI patients, severity of agitation (CMAI total score) did not correlate with the MMSE (rs D 0.001; p D 0.985) or GDetS (rs D 0.076; p D 0.213) scores. In AD patients, the severity of agitation correlated positively with the GDetS (rs D 0.284; p < 0.001) and negatively with the MMSE (rs D 0.211; p < 0.001). The prevalence of MFS items and MFS total scores in MCI and AD patients with and without agitation are displayed in Figure 1 and Table 2. Frontal lobe symptoms were present in 8% of the MCI patients with agitation and in 3% of the MCI patients without agitation (p D 0.099). Frontal lobe symptoms were present in 33% of the AD patients with agitation and in 7% of the AD patients without agitation (p < 0.001). AD patients with agitation had higher MFS total scores as compared to MCI patients with agitation (Figure 3). The prevalence of Behave-AD clusters and the severity of Behave-AD clusters, total and global scores in MCI and AD patients with and without agitation are
displayed in Figure 2 and Table 2. Only diurnal rhythm disturbances were more prevalent and severe in MCI patients with agitation. The Behave-AD total scores in MCI patients remained statistically significant higher in patients with agitation, even after removing the cluster scores activity disturbances and aggressiveness from the Behave-AD total score (p D 0.009). In addition to the prevalence and severity of behavioral symptoms, Behave-AD global scores, indicating how troubling to the caregiver and/or dangerous to the patient the behavioral symptoms are, were higher in MCI patients with agitation. From the MCI patients with agitation, 45% was rated as at least mildly troubling or dangerous, as compared to 16% in MCI patients without agitation (p < 0.001). All behavioral symptoms were more severe and more prevalent in AD patients with agitation. The Behave-AD total scores in AD patients remained statistically significant higher in patients with agitation, even after removing the cluster scores activity disturbances and aggressiveness from the Behave-AD total score (p < 0.001). The Behave-AD global scores were higher in AD patients with agitation. From the AD patients with agitation, 85% was rated as at least mildly troubling or dangerous, as compared to 31% in AD patients without agitation (p < 0.001). AD patients with agitation displayed more severe behavioral symptoms as compared to MCI patients with agitation (Figure 3). Furthermore, MCI patients with agitation showed more severe behavioral symptoms than AD patients without agitation (Figure 3). These last two findings remained statistically significant even after removing cluster scores from the Behave-AD total score that also measure agitated behavior such as activity disturbances and aggressiveness (p < 0.001). MCI patients with agitation displayed more depressive symptoms than patients without agitation (Table 2). The prevalence of significant depressive symptoms in MCI patients with agitation was 19% and in patients without agitation 8% (p D 0.013). AD patients with agitation displayed more depressive symptoms than patients without agitation (Table 2). The prevalence of significant depressive symptoms in AD patients with agitation (31%) was higher than the prevalence of depressive symptoms in patients without agitation (8%; p < 0.001). AD patients with agitation displayed more depressive symptoms as compared to MCI patients with agitation (Figure 3). Furthermore, MCI patients with agitation showed more severe depressive symptoms than AD patients without agitation (Figure 3). In MCI, there was no difference (p D 0.433) in prevalence of agitation between the patients treated with psychotropic medication (n D 121) and the untreated (n D 131) patients with respectively 56% and 61%. Treated (n D 68) and untreated (n D 80) MCI patients with agitation did neither differ regarding MFS total scores (p D 0.169) and behavioral symptoms (Behave-AD total score: p D 0.050), nor with regard to being troubling to the caregiver and/or dangerous to themselves (Behave-AD global score: p D 0.731). However, treated MCI patients with agitation
Aging & Mental Health
253
Table 2. Severity of behavioral symptoms.
Behavior/disturbances (score range)
MCI (n D 268)
AD (n D 393)
Agitation
Agitation
No n D 108
Yes n D 160
No n D 96
Yes n D 297
MFS (0 10) Total score
1.5 § 1.3 (0 6)
2.4 § 1.5 (0 8)
2.0 § 1.6 (0 7)
3.8 § 1.7 (1 9)
Behave-AD (0 21) Delusions
0.3 § 1.1 (0 9)
0.2 § 0.6 (0 4)
0.4 § 1.1 (0 5)
1.4 § 2.5 (0 20)
Behave-AD (0 15) Hallucinations
0.2 § 1.0 (0 6)
0.1 § 0.6 (0 5)
0.2 § 0.6 (0 4)
0.4 § 1.1 (0 7)
Behave-AD (0 36) Psychosis
0.5 § 1.9 (0 15)
0.3 § 1.0 (0 9)
0.6 § 1.5 (0 9)
1.9 § 3.1 (0 26)
Behave-AD (0 9) Activity
0.1 § 0.5 (0 3)
0.3 § 0.7 (0 3)
0.2 § 0.6 (0 4)
2.0 § 2.0 (0 9)
Behave-AD (0 9) Aggressiveness
0.3 § 0.8 (0 5)
1.9 § 1.9 (0 9)
0.4 § 1.2 (0 9)
2.8 § 2.7 (0 9)
Behave-AD (0 3) Diurnal rhythm
0.3 § 0.6 (0 3)
0.5 § 0.7 (0 3)
0.2 § 0.5 (0 3)
0.6 § 0.8 (0 3)
Behave-AD (0 6) Affective
0.7 § 1.1 (0 5)
0.9 § 1.1 (0 5)
0.5 § 1.0 (0 4)
0.9 § 1.3 (0 6)
Behave-AD (0 12) Anxiety/phobias
0.6 § 1.1 (0 8)
0.8 § 1.2 (0 6)
0.3 § 0.7 (0 5)
0.6 § 1.1 (0 5)
Behave-AD (0 9) Total score
2.5 § 3.0 (0 17)
4.7 § 3.6 (0 20)
2.1 § 3.4 (0 17)
8.8 § 6.5 (0 47)
Behave-AD (0 3) Global score
0.2 § 0.4 (0 2)
0.6 § 0.8 (0 3)
0.4 § 0.7 (0 2)
1.4 § 0.8 (0 3)
CMAI (10 70) Aggressive
10.0 § 0.0 (10 10)
10.1 § 0.6 (10 16)
10.1 § 0.6 (10 15)
12.0 § 5.4 (10 50)
CMAI (11 77) Physically non-aggressive
11.0 § 0.1 (11 12)
13.0 § 3.0 (11 24)
11.1 § 0.3 (11 13)
18.7 § 7.0 (11 48)
CMAI (8 56) Verbally agitated
8.1 § 0.5 (8 11)
13.0 § 4.6 (8 36)
8.1 § 0.4 (8 10)
14.4 § 6.2 (8 40)
CMAI (29 203) Total score
29.2 § 0.5 (29 32)
36.2 § 6.4 (31 69)
29.3 § 0.9 (29 34)
45.1 § 13.3 (31 105)
Cornell Scale for Depression (0 38) Total score
3.3 § 2.9 (0 14)
5.2 § 4.9 (0 31)
2.7 § 3.3 (0 16)
6.1 § 4.1 (0 25)
Note: Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. Data are given as ratio, percentage or mean § SD with ranges represented between brackets. Psychosis D delusions C hallucinations. Significant differences comparing the groups with and without agitation within the MCI group and within the AD group are mentioned as follows: p < 0.01, p < 0.001. For all comparisons, Mann Witney U test was applied and the level of significance was set at p < 0.01. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease; MFS D Middelheim Frontality Score; Behave-AD D Behavioral Pathology in Alzheimer’s Disease Rating Scale; CMAI D Cohen-Manfield Agitation Inventory; CSDD D Cornell Scale for Depression in Dementia.
displayed more depressive symptoms as compared to untreated MCI patients with agitation (CSDD total score: p D 0.002). In AD, there was no difference (p D 0.793) in prevalence of agitation between the treated (n D 330) and untreated (n D 62) patients with respectively 76% and 74%. Treated (n D 250) and untreated (n D 46) AD patients with agitation did neither differ regarding MFS total scores (p D 0.061) and behavioral (p D 0.315) and
depressive symptoms (p D 0.177), nor with regard to being troubling to the caregiver and/or dangerous to themselves (p D 0.205).
Discussion Our finding on the prevalence of agitation in AD (76%) is in line with previous research, stating that up to 45% 85% of dementia patients are suffering from
254
S. Van der Mussele et al.
Figure 3. Severity of frontal lobe, behavioral and depressive symptoms in MCI and AD. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. For all comparisons, Mann Whitney U test was applied. The level of significance was set at p < 0.01. (a) MCI patients without agitation; (b) MCI patients with agitation; (c) AD patients without agitation and (d) AD patients with agitation. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease; MFS D Middelheim Frontality Score; Behave-AD D Behavioral Pathology in Alzheimer’s Disease Rating Scale; CSDD D Cornell Scale for Depression in Dementia.
agitation (Chan et al., 2003; Lyketsos et al., 2000, 2002; Margallo-Lana et al., 2001; Pitkala et al., 2004; Selbaek et al., 2007; Zuidema et al., 2007). However, our finding on the prevalence of agitation in MCI (60%) is somewhat
higher than the prevalence found in hospital-based studies in previous research, where the maximum observed prevalence is 45% (Apostolova & Cummings, 2008; Monastero et al., 2009). Nevertheless, none of the hospital-based studies, referred to in this two reviews (Apostolova & Cummings, 2008; Monastero et al., 2009), used the CMAI as agitation assessment scale. Therefore, we could assume, apart from the two discussed limitations (recruitment in a memory clinic and not blinded raters), that the CMAI might be a quite sensitive approach of assessing agitation in MCI, including three clusters: aggressive, physically non-aggressive and verbally agitated behavior. We hypothesize that a large part of MCI patients have MCI due to AD (Albert et al., 2011) or prodromal AD (Dubois, 2014). In addition, neuroanatomic studies about agitation in MCI and AD are sparse, but point toward the involvement of frontal and temporal areas (Trzepacz et al., 2013). Recently Trzepacz et al. (2013) and Tsai et al. (2013) added evidence to this theory and confirmed the relation between agitation and AD pathology in these specific brain areas through neurodegeneration and neurochemical changes. Also, worse cognitive performance on the MMSE correlated with these AD brain changes in these two studies. Then, our study found similar evidence as previous research that agitation/aggressiveness in AD is associated with internal factors such as depression and psychosis (Cipriani et al., 2011): agitation in MCI and AD is associated with depressive symptoms and vice versa and depressive symptoms are prevalent and more severe in MCI (19%) and AD (31%) patients with agitation (Van der Mussele et al., 2013). Therefore, the following hypothesis might also explain the prevalence and severity of agitation and frontal lobe symptoms in MCI and AD patients: decreased serotonergic activity has been related to depression (Coppen, 1967) and dopaminergic neurons are modulated by serotonergic innervation (Di Giovanni, Esposito, & Di Matteo, 2010). Indeed, the serotonergic system appears to have an inhibitory effect on the dopaminergic function (Di Giovanni et al., 2010; Engelborghs et al., 2008). Furthermore, ascending dopaminergic pathways are part of the frontal-subcortical circuitry (Bonelli & Cummings, 2007) and the dopaminergic system is one of the important modulators of frontal lobe function (Goldman-Rakic, Lidow, & Gallager, 1990). Consequently, frontal-subcortical circuit dysfunction leads to impaired executive functions, apathy and impulsivity (Bonelli & Cummings, 2007) and might thus as well explain the prevalence and severity of agitation and frontal lobe symptoms in MCI and AD patients. In brief, MCI can be an early stage of AD, agitation and frontal lobe symptoms can be provoked by the frontal and temporal AD brain pathology and therefore be of prognostic value in MCI for the progression to AD; and different types of behavioral disturbances are interrelated in AD, thus occurring sometimes together or associated. The clinical implications of this study are that in MCI patients with agitation, attention should also be given to the (non-)pharmacological treatment of associated depressive symptoms and diurnal rhythm disturbances, whereas in AD, agitation can be associated with all types of behavioral
Aging & Mental Health disturbances. Moreover, the management of agitation might decrease associated behavioral symptoms, as well as the opposite approach, that the treatment of associated behavioral symptoms might reduce agitation in patients. The quality of care for patients with cognitive deterioration could be increased by the implementation of a systematic behavioral screening policy in the geriatric in- and outpatient services. Specific tools and training could support the implementation of this behavioral screening policy. Further research would assist in understanding the underlying pathophysiology of associated behavioral symptoms in agitation and in understanding the possible interrelation of behavioral symptoms, the epidemiology of MCI and the diagnostic and prognostic value of behavioral symptoms and syndromes. Follow-up of the included MCI patients is ongoing to test the hypothesis that behavioral disturbances in MCI predict progression to dementia. The occurrence of agitation in MCI might be of prognostic value for progression to AD. To our knowledge, this study is the first to investigate associated behavioral symptoms in and between agitated AD and MCI patients. This study has additional strengths. First, the study included large and well-characterized MCI and AD patient groups that have been diagnosed following the strict application of clinical diagnostic criteria. Second, we broke up our study groups in subgroups with and without agitation in a reliable way, as it has been proven trustworthy in other studies (Choy et al., 2001; CohenMansfield et al., 1989; Gruber-Baldini et al., 2004; Suh, 2004; Testad et al., 2007; Zuidema et al., 2007; Zuidema et al., 2010). Third, as the data came from a prospective, longitudinal study, the follow-up of included subjects and AD autopsy confirmation (n D 35) contributed to increased AD diagnostic certainty. Lastly, all patients were diagnosed by clinicians in the same center in order to preserve homogeneity in the diagnostic groups. Three limitations to the study are known to the authors. First, our study population was recruited in a memory clinic, which might have introduced a selection bias as behavioral symptoms might have contributed to referral. Second, behavioral assessment raters were not blinded for the subjects’ clinical MCI or AD diagnosis. On the other hand, due to strict application of clinical diagnostic criteria, which do not include behavioral changes, behavior did not impact the clinical diagnosis. Third, included subjects were not free of psychotropic medication at the moment of their behavioral observation. However, there is no difference in the prevalence of agitation between the treated and untreated MCI or AD patients and intra-diagnostic group analysis did not reveal important differences in behavior between treated and untreated patients. We did not assess in detail the effects of psychotropic drugs on behavior, as this is not a pharmacological study. Furthermore, psychotropic drug intake rates vary among countries and Belgium is known to have a high prevalence of psychotropic drug utilization in community-dwelling elderly and nursing home residents (Azermai, Elseviers, Petrovic, Van Bortel, & Vander Stichele, 2011). This also partially explains the high
255
prevalence of psychotropic drug intake in MCI and AD patients in this study. A matter of debate could be the wide range of MMSE scores in our MCI population. Some studies use MMSE scores as part of their key eligibility criteria for MCI, e.g. the Alzheimer’s Disease Neuroimaging Initiative (ADNI) with MMSE scores 24 30 (inclusive) for MCI subjects. This might raise questions about our MCI study population with the broad MMSE range of 16 30. First, from the MCI patients with an MMSE score of
Agitation-associated behavioral symptoms in mild cognitive impairment and Alzheimer’s dementia Stefan Van der Musselea,b, Nathalie Le Bastarda, Jos Saerensc, Nore Somersc, Peter Mari€enc,d, Johan Goemanc, Peter P. De Deyna,c,e,f and Sebastiaan Engelborghsa,c* a
Laboratory of Neurochemistry and Behavior, Reference Centre for Biological Markers of Dementia (BIODEM), Institute Born-Bunge, University of Antwerp (UA), Antwerp, Belgium; bDepartment of Nursing and Midwifery Sciences, Faculty of Medicine and Health Sciences, University of Antwerp (UA), Antwerp, Belgium; cDepartment of Neurology and Memory Clinic, Hospital Network Antwerp (ZNA), Middelheim and Hoge Beuken, Antwerp, Belgium; dDepartment of Clinical and Experimental Neurolinguistics (CLIN), Vrije Universiteit Brussel, Brussels, Belgium; eDepartment of Health Care Sciences, Artesis University College Antwerp, Antwerp, Belgium; f Department of Neurology and Alzheimer Research Center, University Medical Center Groningen, Groningen, The Netherlands (Received 3 November 2013; accepted 13 April 2014) Objectives: The aim of this study is to determine the prevalence of agitation in mild cognitive impairment (MCI, Petersen’s criteria) and patients with Alzheimer’s dementia (AD), and to characterize the associated behavioral symptoms. Method: A cross-sectional analysis of baseline data from a prospective, longitudinal study on behavioral symptoms was performed, including 268 MCI and 393 AD patients. Behavioral assessment was performed through Middelheim Frontality Score (MFS), Behavioral Pathology in Alzheimer’s Disease Rating Scale (Behave-AD) and Cornell Scale for Depression in Dementia (CSDD). Agitated behavior was considered to be clinically relevant when one or more items of the CohenMansfield Agitation Inventory (CMAI) occurred at least once a week. Results: The prevalence of agitation in AD (76%) was higher than in MCI (60%; p < 0.001). Patients with agitation showed more severe frontal lobe, behavioral and depressive symptoms (MFS, Behave-AD and CSDD total scores). In agitated AD patients, all behavioral symptoms and types of agitation were more severe compared to non-agitated AD patients, but in agitated MCI patients only for diurnal rhythm disturbances. This resulted in more severe Behave-AD global scores in patients with agitation as compared to patients without agitation. Comparing MCI and AD patients, MCI patients with agitation showed more severe behavioral and depressive symptoms than AD patients without agitation. The structure of agitation in AD consisted of more aggressive and physically non-aggressive behavior than in MCI. Conclusion: Frontal lobe, behavioral and depressive symptoms are more severe in MCI and AD patients with clinically relevant agitation as compared to patients without agitation. However, this association is less pronounced in MCI. Keywords: neuropsychiatric symptoms; Cohen-Mansfield Agitation Inventory (CMAI); aggressiveness; aggression; MCI; AD
Introduction Behavioral and psychological dementia-related problems increase caregiver burden in mild cognitive impairment (MCI) (Bruce, McQuiggan, Williams, Westervelt, & Tremont, 2008; Frank et al., 2006; Garand, Dew, Eazor, Dekosky, & Reynolds, 2005; Garand et al., 2007) and dementia due to Alzheimer’s disease (AD) (Levy, Lanctot, Farber, Li, & Herrmann, 2012) and this might affect the quality of care for patients and the quality of life of relatives and caregivers (Hazzan, Ploeg, Shannon, Raina, & Oremus, 2013). Moreover, these neuropsychiatric manifestations, such as aggression, depression and hallucinations, are consistent predictors for institutionalization (Gaugler, Yu, Krichbaum, & Wyman, 2009; Luppa, Luck, Brahler, Konig, & Riedel-Heller, 2008). The thorough understanding of behavior in MCI and AD is necessary, as behavioral features may represent risk factors for MCI or predictors for the progression of MCI to AD (Monastero, Mangialasche, Camarda, Ercolani, & Camarda, 2009). In addition, better behavioral understanding may be important for (non-)pharmacological behavioral research and management, aiming to reduce
institutionalization and to improve the quality of life for patients, relatives and caregivers. Agitation belongs to the behavioral and psychological signs and symptoms of dementia (Finkel, Costa e Silva, Cohen, Miller, & Sartorius, 1996) and is defined as inappropriate verbal, vocal or motor activity that is not explained by apparent needs or confusion (Cohen-Mansfield & Billig, 1986). The inappropriateness of the activity refers to it being repetitive, abusive/aggressive and/or not according to social standards. Furthermore, four categories of agitation can be distinguished: aggressive physical, non-aggressive physical, aggressive verbal and nonaggressive verbal (Cohen-Mansfield & Billig, 1986). Agitated behavior is common in dementia with a prevalence of 24% 45% in non-institutionalized older adults with dementia (Chan, Kasper, Black, & Rabins, 2003; Lyketsos et al., 2000, 2002; Margallo-Lana et al., 2001), and up to 85% in institutionalized dementia patients (Margallo-Lana et al., 2001; Pitkala, Laurila, Strandberg, & Tilvis, 2004; Selbaek, Kirkevold, & Engedal, 2007; Zuidema, Derksen, Verhey, & Koopmans, 2007). However, agitation prevalence percentages seem to vary
*Corresponding author. Email: [email protected] Ó 2014 Taylor & Francis
248
S. Van der Mussele et al.
greatly, depending on patient selection and assessment methods. For the same reasons, the prevalence of agitation in MCI varies even more with 5% 25% in populationbased and 4% 45% in hospital-based studies (Apostolova & Cummings, 2008; Monastero et al., 2009). An agitation syndrome in AD is repeatedly found as a principal behavioral component beside a mood and a psychosis factor (Cummings, McRae, & Zhang, 2006; Frisoni et al., 1999; Hollingworth et al., 2006; Kang, Ahn, Kim, & Kim, 2010; Mirakhur, Craig, Hart, McLlroy, & Passmore, 2004; Spalletta, Baldinetti et al., 2004; Spalletta, Musicco et al., 2010; VilaltaFranch et al., 2010). Furthermore, aggressive behavior in patients with dementia is associated with internal factors such as depression, psychosis and pain (Cipriani, Vedovello, Nuti, & Di, 2011). To our knowledge, the relationship, between clinically relevant agitation and associated behavioral symptoms in MCI and as compared to AD, has not been investigated. This study aims to test the hypotheses that agitated patients display more frontal lobe, depressive and other neuropsychiatric symptoms, such as psychosis, activity, sleep and affective disturbances and anxiety/phobias, as compared to non-agitated patients; and that agitated AD patients have more associated behavioral symptoms than agitated MCI patients, as behavioral symptoms are more present in AD (Van der Mussele et al., 2012). Methods Study population and diagnostic criteria The study population consisted of MCI patients (n D 268) and AD patients (n D 393). Patients were consecutively included in our memory clinic at the moment of their diagnostic work-up, consisting of a general physical and neurological examination, blood screening, structural neuroimaging, standard electroencephalogram and an extensive time-linked (§3 months) neuropsychological examination with adjustment for age and education, comprising amongst others the Mini-Mental State Examination (MMSE) (Folstein, Folstein, & McHugh, 1975) and the Wechsler Memory Scale III (The Psychological Corporation, 1998), Hierarchic Dementia Scale (Cole & Dastoor, 1987) and/or Repeatable Battery for the Assessment of Neuropsychological Status (Randolph, Tierney, Mohr, & Chase, 1998). To diagnose MCI, Petersen’s diagnostic criteria (2004) were applied. Objective cognitive impairment for age and education was quantified as a performance of more than 1.5 standard deviation below the appropriate mean on the neuropsychological subtests. As all cognitive domains of subjects were tested in an extensive timelinked (§3 months) neuropsychological examination, all MCI patients were categorized as amnestic/non-amnestic, single/multiple domain. The MCI group (n D 268) consisted of 53 amnestic single domain, 149 amnestic multiple domain, 29 non-amnestic single domain and 37 nonamnestic multiple domain MCI patients. Probable AD was diagnosed according to NINCDS/ ADRDA criteria (National Institute of Neurological and
Communicative Disorders and Stroke Alzheimer’s Disease and Related Disorders Association) (McKhann et al., 1984), though all patients as well fulfilled the DSM-IV (dementia criteria of the Diagnostic and Statistical Manual of Mental Disorders, fourth edition) criteria (American Psychiatric Association, 1994). Clinical and neuropsychological follow-up of included patients and autopsy in deceased consented patients during follow-up period, did contribute to the diagnostic accuracy of the subjects in this study. The AD group consisted of 35 definite and 358 probable AD patients. Staging of cognitive deterioration was assessed by means of the Global Deterioration Scale (GDetS) (Reisberg, Ferris, de Leon, & Crook, 1982). To get a cursory idea about how psychotropic drug intake might have affected our study results, we investigated the difference in behavior between treated and untreated patients with agitation by comparing the total scores of the behavioral assessment scales. MCI and AD patients were categorized as treated when they took medication from at least one of the psychotropic drug categories mentioned in Table 1. The local ethics committee approved this study. All patients and/or patients’ caregivers gave written informed consent. All patients were of Caucasian origin. Behavioral assessment All subjects underwent in-depth behavioral assessment at inclusion (baseline) consisting of an interview of both patient and caregiver, covering a period of two weeks prior to inclusion. In case a non-professional caregiver was lacking, the patient’s main professional caregiver was contacted and interviewed. The interview was performed by the clinician or researcher who was not blinded for the subject’s cognitive diagnosis. The battery of behavioral assessment scales comprised: Cohen-Mansfield Agitation Inventory (CMAI), Middelheim Frontality Score (MFS), Behavioral Pathology in Alzheimer’s Disease Rating Scale (Behave-AD) and Cornell Scale for Depression in Dementia (CSDD). The MFS is a validated clinical and behavioral assessment scale that measures frontal lobe features and reliably discriminates frontotemporal dementia (FTD) from AD patients with a sensitivity and specificity of almost 90% and with good inter- and intra-rater reliability (Aries et al., 2010; De Deyn et al., 2005). According to the Instructions for Administration and Scoring, the MFS was obtained by summating scores in a standardized fashion on 10 items. Each item was scored either 0 (absent) or 1 (present) yielding a total maximal score of 10. The 10 items scored are noted in Figure 1. The presence of frontal lobe symptoms was considered to be clinically relevant in case of a total MFS score of 5 (De Deyn et al., 2005). The Behave-AD is a 25-item scale that measures behavioral symptoms in seven clusters (Figure 2), scored on a four-point scale of increasing severity (Reisberg et al., 1987). A psychosis cluster score is calculated by summating the scores of the paranoid/delusional ideation (delusions) and hallucinations clusters. Besides a total score, a
Aging & Mental Health
249
Table 1. Demographic, clinical and neuropsychological data. MCI
AD Agitation
Total n D 268 % Male/Female
Agitation
No n D 108
Yes n D 160
40.3%
59.7%
Total n D 393
No n D 96
Yes n D 297
24.4%
75.6%
26/70
103/194
119/149
46/62
73/87
129/264
Age at inclusion (years)
75.5 § 8.1 (50 94)
76.3 § 8.6 (50 94)
75.0 § 7.7 (53 90)
80.1 § 7.6 (51 97)
81.8 § 8.1 (51 97)
79.6 § 7.3 (54 97)
Age at onset (years)
73.2 § 8.4 (48 90)
73.9 § 8.6 (48 90)
72.7 § 8.2 (51 87)
76.9 § 8.4 (48 96)
79.2 § 8.7 (48 96)
76.1 § 8.2 (50 95)
Disease duration (years)
2.2 § 1.7 (0 11)
2.2 § 1.8 (0 11)
2.2 § 1.7 (0 10)
3.2 § 2.4 (0 14)
2.6 § 1.8 (1 10)
3.4 § 2.6 (0 14)
GDetS (1 7)
3.0 § 0.7 (1 5)
3.1 § 0.8 (1 5)
3.3 § 0.7 (1 5)
5.1 § 1.0 (2 7)
4.9 § 0.9 (2 7)
5.2 § 1.0 (3 7)
MMSE score (/30)
25.6 § 2.9 (16 30)
25.7 § 2.9 (16 30)
25.6 § 2.9 (16 30)
15.1 § 6.3 (0 30)
16.5 § 5.9 (0 30)
14.7 § 6.4 (0 28)
52.4
50.0
54.1
15.6
16.7
15.2
37.5
40.4
33.3
50.5
Free of psychotropic medication (%) Antidepressants (%) Antipsychotics (%) Hypnotics, sedatives, anxiolytics (%) Cholinesterase inhibitors (%) Antiparkinsonian agents (%) Antiepileptics (%)
24.8
20.2
28.0
39.7
4.7
6.6
3.3
46.3
28.5
29.5
27.8
23.4
20.8
24.2
3.4
5.7
1.9
37.7
44.2
35.6
2.7
2.8
2.6
1.5
3.1
1.0
2.3
2.8
1.9
1.5
1.0
1.7
Note: Data are given as ratio, percentage or mean § SD with ranges represented between brackets. Significant differences comparing the MCI group with the AD group and the groups with and without agitation within the MCI group and within the AD group are mentioned as follows: p < 0.01, p < 0.001. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. For comparison of male female ratios and percentages, chi-square statistics were used. For other comparisons, Mann Witney U test was applied. The level of significance was set at p < 0.01. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease; GDetS D Global Deterioration Scale.
global score on a four-point scale of increasing severity is provided, reflecting how troubling to the caregiver or dangerous to the patient the behavioral symptoms are, from not troubling or dangerous (score 0) to severe (score 3). Depressive symptoms were assessed by means of the CSDD, a 19-item depression scale (Alexopoulos, Abrams, Young, & Shamoian, 1988a). Item scores range from 0 (absent) to 2 (severe), with a maximum total score of 38 points. The presence of significant depressive symptoms was defined as a total CSDD score >7 (Burns, Lawlor, & Craig, 2004). Studies have shown the CSDD to be valid for screening depression in non-demented patients too (Alexopoulos, Abrams, Young, & Shamoian, 1988b). The CMAI assesses 29 agitated behaviors on a seven-point scale of increasing frequency/severity (1 D never; 7 D several times an hour) (Cohen-Mansfield, 1996). CMAI cluster scores include aggressive behavior (10 items), physically non-aggressive behavior (11
items) and verbally agitated behavior (8 items); a total score is provided as well. Agitation was considered to be clinically relevant when one or more items occurred at least once a week (any individual item score 3); and aggressive, physically non-aggressive and verbally agitated behavior was considered to be clinically relevant when one or more items within the respective cluster occurred at least once a week (Choy, Lam, Chan, Li, & Chiu, 2001; Cohen-Mansfield, Marx, & Rosenthal, 1989; Gruber-Baldini, Boustani, Sloane, & Zimmerman, 2004; Suh, 2004; Testad, Aasland, & Aarsland, 2007; Zuidema et al., 2007; Zuidema, de Jonghe, Verhey, & Koopmans, 2010). Statistical analyses The (behavioral) assessment scales provide semicontinuous variables. Therefore, non-parametric statistics were used:
250
S. Van der Mussele et al.
Figure 1. Prevalence (%) of frontal lobe symptoms in MCI and AD patients. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. Chi-square statistics were used, comparing the groups with and without agitation within the MCI and AD groups. Note: Significant differences are mentioned as follows: p < 0.01, p < 0.001. The level of significance was set at p < 0.01. Abbreviations: MFS D Middelheim Frontality Score; MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease.
Aging & Mental Health
251
Figure 2. Prevalence (%) of behavioral symptoms in MCI and AD patients. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. Chi-square statistics were used, comparing the groups with and without agitation within the MCI and AD groups. Note: Significant differences are mentioned as follows: p < 0.01, p < 0.001. The level of significance was set at p < 0.01. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease.
252
S. Van der Mussele et al.
Mann Whitney U test was applied to compare (semi)continuous variables, chi-square statistics for categorical data and Spearman rank to correlate (semi)continuous data. To reduce the type I error rate, probability levels of 0.01 were considered significant. Statistical analyses were carried out using Statistical Package for the Social Sciences (SPSS) Statistics 17.0. Results The AD group comprised a higher percentage of women than the MCI group. MCI patients were younger at inclusion and disease onset and their disease duration was shorter. In AD patients, GDetS scores were higher and MMSE scores were lower as compared to the MCI group. Table 1 shows that there is no difference in gender, age, disease duration, GDetS, MMSE and psychotropic drug intake between MCI patients with and without agitation. In contrast, agitated AD patients had a younger age at onset, a longer disease duration yet, higher GDetS scores and used more antipsychotics than non-agitated AD patients. The prevalence of agitation in AD patients (76%) was higher than the prevalence of agitation in MCI patients (60%; p < 0.001). Our study also subcategorized agitation in aggressive, non-aggressive and verbally agitated behavior. The prevalence of aggressive behavior, as measured by the CMAI, was 14% in AD patients and 1% in MCI patients (p < 0.001). The prevalence of physically non-aggressive behavior was 60% in AD patients and 24% in MCI patients (p < 0.001). The prevalence of verbally agitated behavior was 58% in AD patients and 50% in MCI patients (p D 0.039). Agitation, as defined in this study, had a different structure in MCI and AD patients. Agitated MCI patients displayed less aggressive behavior (1% versus 19%; p < 0.001) and less physically non-aggressive behavior (41% versus 80%; p < 0.001) than agitated AD patients. Verbally agitated behavior was present in 83% of the agitated MCI patients and in 76% of the agitated AD patients (p D 0.095). In MCI patients, severity of agitation (CMAI total score) did not correlate with the MMSE (rs D 0.001; p D 0.985) or GDetS (rs D 0.076; p D 0.213) scores. In AD patients, the severity of agitation correlated positively with the GDetS (rs D 0.284; p < 0.001) and negatively with the MMSE (rs D 0.211; p < 0.001). The prevalence of MFS items and MFS total scores in MCI and AD patients with and without agitation are displayed in Figure 1 and Table 2. Frontal lobe symptoms were present in 8% of the MCI patients with agitation and in 3% of the MCI patients without agitation (p D 0.099). Frontal lobe symptoms were present in 33% of the AD patients with agitation and in 7% of the AD patients without agitation (p < 0.001). AD patients with agitation had higher MFS total scores as compared to MCI patients with agitation (Figure 3). The prevalence of Behave-AD clusters and the severity of Behave-AD clusters, total and global scores in MCI and AD patients with and without agitation are
displayed in Figure 2 and Table 2. Only diurnal rhythm disturbances were more prevalent and severe in MCI patients with agitation. The Behave-AD total scores in MCI patients remained statistically significant higher in patients with agitation, even after removing the cluster scores activity disturbances and aggressiveness from the Behave-AD total score (p D 0.009). In addition to the prevalence and severity of behavioral symptoms, Behave-AD global scores, indicating how troubling to the caregiver and/or dangerous to the patient the behavioral symptoms are, were higher in MCI patients with agitation. From the MCI patients with agitation, 45% was rated as at least mildly troubling or dangerous, as compared to 16% in MCI patients without agitation (p < 0.001). All behavioral symptoms were more severe and more prevalent in AD patients with agitation. The Behave-AD total scores in AD patients remained statistically significant higher in patients with agitation, even after removing the cluster scores activity disturbances and aggressiveness from the Behave-AD total score (p < 0.001). The Behave-AD global scores were higher in AD patients with agitation. From the AD patients with agitation, 85% was rated as at least mildly troubling or dangerous, as compared to 31% in AD patients without agitation (p < 0.001). AD patients with agitation displayed more severe behavioral symptoms as compared to MCI patients with agitation (Figure 3). Furthermore, MCI patients with agitation showed more severe behavioral symptoms than AD patients without agitation (Figure 3). These last two findings remained statistically significant even after removing cluster scores from the Behave-AD total score that also measure agitated behavior such as activity disturbances and aggressiveness (p < 0.001). MCI patients with agitation displayed more depressive symptoms than patients without agitation (Table 2). The prevalence of significant depressive symptoms in MCI patients with agitation was 19% and in patients without agitation 8% (p D 0.013). AD patients with agitation displayed more depressive symptoms than patients without agitation (Table 2). The prevalence of significant depressive symptoms in AD patients with agitation (31%) was higher than the prevalence of depressive symptoms in patients without agitation (8%; p < 0.001). AD patients with agitation displayed more depressive symptoms as compared to MCI patients with agitation (Figure 3). Furthermore, MCI patients with agitation showed more severe depressive symptoms than AD patients without agitation (Figure 3). In MCI, there was no difference (p D 0.433) in prevalence of agitation between the patients treated with psychotropic medication (n D 121) and the untreated (n D 131) patients with respectively 56% and 61%. Treated (n D 68) and untreated (n D 80) MCI patients with agitation did neither differ regarding MFS total scores (p D 0.169) and behavioral symptoms (Behave-AD total score: p D 0.050), nor with regard to being troubling to the caregiver and/or dangerous to themselves (Behave-AD global score: p D 0.731). However, treated MCI patients with agitation
Aging & Mental Health
253
Table 2. Severity of behavioral symptoms.
Behavior/disturbances (score range)
MCI (n D 268)
AD (n D 393)
Agitation
Agitation
No n D 108
Yes n D 160
No n D 96
Yes n D 297
MFS (0 10) Total score
1.5 § 1.3 (0 6)
2.4 § 1.5 (0 8)
2.0 § 1.6 (0 7)
3.8 § 1.7 (1 9)
Behave-AD (0 21) Delusions
0.3 § 1.1 (0 9)
0.2 § 0.6 (0 4)
0.4 § 1.1 (0 5)
1.4 § 2.5 (0 20)
Behave-AD (0 15) Hallucinations
0.2 § 1.0 (0 6)
0.1 § 0.6 (0 5)
0.2 § 0.6 (0 4)
0.4 § 1.1 (0 7)
Behave-AD (0 36) Psychosis
0.5 § 1.9 (0 15)
0.3 § 1.0 (0 9)
0.6 § 1.5 (0 9)
1.9 § 3.1 (0 26)
Behave-AD (0 9) Activity
0.1 § 0.5 (0 3)
0.3 § 0.7 (0 3)
0.2 § 0.6 (0 4)
2.0 § 2.0 (0 9)
Behave-AD (0 9) Aggressiveness
0.3 § 0.8 (0 5)
1.9 § 1.9 (0 9)
0.4 § 1.2 (0 9)
2.8 § 2.7 (0 9)
Behave-AD (0 3) Diurnal rhythm
0.3 § 0.6 (0 3)
0.5 § 0.7 (0 3)
0.2 § 0.5 (0 3)
0.6 § 0.8 (0 3)
Behave-AD (0 6) Affective
0.7 § 1.1 (0 5)
0.9 § 1.1 (0 5)
0.5 § 1.0 (0 4)
0.9 § 1.3 (0 6)
Behave-AD (0 12) Anxiety/phobias
0.6 § 1.1 (0 8)
0.8 § 1.2 (0 6)
0.3 § 0.7 (0 5)
0.6 § 1.1 (0 5)
Behave-AD (0 9) Total score
2.5 § 3.0 (0 17)
4.7 § 3.6 (0 20)
2.1 § 3.4 (0 17)
8.8 § 6.5 (0 47)
Behave-AD (0 3) Global score
0.2 § 0.4 (0 2)
0.6 § 0.8 (0 3)
0.4 § 0.7 (0 2)
1.4 § 0.8 (0 3)
CMAI (10 70) Aggressive
10.0 § 0.0 (10 10)
10.1 § 0.6 (10 16)
10.1 § 0.6 (10 15)
12.0 § 5.4 (10 50)
CMAI (11 77) Physically non-aggressive
11.0 § 0.1 (11 12)
13.0 § 3.0 (11 24)
11.1 § 0.3 (11 13)
18.7 § 7.0 (11 48)
CMAI (8 56) Verbally agitated
8.1 § 0.5 (8 11)
13.0 § 4.6 (8 36)
8.1 § 0.4 (8 10)
14.4 § 6.2 (8 40)
CMAI (29 203) Total score
29.2 § 0.5 (29 32)
36.2 § 6.4 (31 69)
29.3 § 0.9 (29 34)
45.1 § 13.3 (31 105)
Cornell Scale for Depression (0 38) Total score
3.3 § 2.9 (0 14)
5.2 § 4.9 (0 31)
2.7 § 3.3 (0 16)
6.1 § 4.1 (0 25)
Note: Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. Data are given as ratio, percentage or mean § SD with ranges represented between brackets. Psychosis D delusions C hallucinations. Significant differences comparing the groups with and without agitation within the MCI group and within the AD group are mentioned as follows: p < 0.01, p < 0.001. For all comparisons, Mann Witney U test was applied and the level of significance was set at p < 0.01. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease; MFS D Middelheim Frontality Score; Behave-AD D Behavioral Pathology in Alzheimer’s Disease Rating Scale; CMAI D Cohen-Manfield Agitation Inventory; CSDD D Cornell Scale for Depression in Dementia.
displayed more depressive symptoms as compared to untreated MCI patients with agitation (CSDD total score: p D 0.002). In AD, there was no difference (p D 0.793) in prevalence of agitation between the treated (n D 330) and untreated (n D 62) patients with respectively 76% and 74%. Treated (n D 250) and untreated (n D 46) AD patients with agitation did neither differ regarding MFS total scores (p D 0.061) and behavioral (p D 0.315) and
depressive symptoms (p D 0.177), nor with regard to being troubling to the caregiver and/or dangerous to themselves (p D 0.205).
Discussion Our finding on the prevalence of agitation in AD (76%) is in line with previous research, stating that up to 45% 85% of dementia patients are suffering from
254
S. Van der Mussele et al.
Figure 3. Severity of frontal lobe, behavioral and depressive symptoms in MCI and AD. Agitation was considered to be clinically relevant (Yes) when one or more items occurred at least once a week. For all comparisons, Mann Whitney U test was applied. The level of significance was set at p < 0.01. (a) MCI patients without agitation; (b) MCI patients with agitation; (c) AD patients without agitation and (d) AD patients with agitation. Abbreviations: MCI D mild cognitive impairment; AD D dementia due to Alzheimer’s disease; MFS D Middelheim Frontality Score; Behave-AD D Behavioral Pathology in Alzheimer’s Disease Rating Scale; CSDD D Cornell Scale for Depression in Dementia.
agitation (Chan et al., 2003; Lyketsos et al., 2000, 2002; Margallo-Lana et al., 2001; Pitkala et al., 2004; Selbaek et al., 2007; Zuidema et al., 2007). However, our finding on the prevalence of agitation in MCI (60%) is somewhat
higher than the prevalence found in hospital-based studies in previous research, where the maximum observed prevalence is 45% (Apostolova & Cummings, 2008; Monastero et al., 2009). Nevertheless, none of the hospital-based studies, referred to in this two reviews (Apostolova & Cummings, 2008; Monastero et al., 2009), used the CMAI as agitation assessment scale. Therefore, we could assume, apart from the two discussed limitations (recruitment in a memory clinic and not blinded raters), that the CMAI might be a quite sensitive approach of assessing agitation in MCI, including three clusters: aggressive, physically non-aggressive and verbally agitated behavior. We hypothesize that a large part of MCI patients have MCI due to AD (Albert et al., 2011) or prodromal AD (Dubois, 2014). In addition, neuroanatomic studies about agitation in MCI and AD are sparse, but point toward the involvement of frontal and temporal areas (Trzepacz et al., 2013). Recently Trzepacz et al. (2013) and Tsai et al. (2013) added evidence to this theory and confirmed the relation between agitation and AD pathology in these specific brain areas through neurodegeneration and neurochemical changes. Also, worse cognitive performance on the MMSE correlated with these AD brain changes in these two studies. Then, our study found similar evidence as previous research that agitation/aggressiveness in AD is associated with internal factors such as depression and psychosis (Cipriani et al., 2011): agitation in MCI and AD is associated with depressive symptoms and vice versa and depressive symptoms are prevalent and more severe in MCI (19%) and AD (31%) patients with agitation (Van der Mussele et al., 2013). Therefore, the following hypothesis might also explain the prevalence and severity of agitation and frontal lobe symptoms in MCI and AD patients: decreased serotonergic activity has been related to depression (Coppen, 1967) and dopaminergic neurons are modulated by serotonergic innervation (Di Giovanni, Esposito, & Di Matteo, 2010). Indeed, the serotonergic system appears to have an inhibitory effect on the dopaminergic function (Di Giovanni et al., 2010; Engelborghs et al., 2008). Furthermore, ascending dopaminergic pathways are part of the frontal-subcortical circuitry (Bonelli & Cummings, 2007) and the dopaminergic system is one of the important modulators of frontal lobe function (Goldman-Rakic, Lidow, & Gallager, 1990). Consequently, frontal-subcortical circuit dysfunction leads to impaired executive functions, apathy and impulsivity (Bonelli & Cummings, 2007) and might thus as well explain the prevalence and severity of agitation and frontal lobe symptoms in MCI and AD patients. In brief, MCI can be an early stage of AD, agitation and frontal lobe symptoms can be provoked by the frontal and temporal AD brain pathology and therefore be of prognostic value in MCI for the progression to AD; and different types of behavioral disturbances are interrelated in AD, thus occurring sometimes together or associated. The clinical implications of this study are that in MCI patients with agitation, attention should also be given to the (non-)pharmacological treatment of associated depressive symptoms and diurnal rhythm disturbances, whereas in AD, agitation can be associated with all types of behavioral
Aging & Mental Health disturbances. Moreover, the management of agitation might decrease associated behavioral symptoms, as well as the opposite approach, that the treatment of associated behavioral symptoms might reduce agitation in patients. The quality of care for patients with cognitive deterioration could be increased by the implementation of a systematic behavioral screening policy in the geriatric in- and outpatient services. Specific tools and training could support the implementation of this behavioral screening policy. Further research would assist in understanding the underlying pathophysiology of associated behavioral symptoms in agitation and in understanding the possible interrelation of behavioral symptoms, the epidemiology of MCI and the diagnostic and prognostic value of behavioral symptoms and syndromes. Follow-up of the included MCI patients is ongoing to test the hypothesis that behavioral disturbances in MCI predict progression to dementia. The occurrence of agitation in MCI might be of prognostic value for progression to AD. To our knowledge, this study is the first to investigate associated behavioral symptoms in and between agitated AD and MCI patients. This study has additional strengths. First, the study included large and well-characterized MCI and AD patient groups that have been diagnosed following the strict application of clinical diagnostic criteria. Second, we broke up our study groups in subgroups with and without agitation in a reliable way, as it has been proven trustworthy in other studies (Choy et al., 2001; CohenMansfield et al., 1989; Gruber-Baldini et al., 2004; Suh, 2004; Testad et al., 2007; Zuidema et al., 2007; Zuidema et al., 2010). Third, as the data came from a prospective, longitudinal study, the follow-up of included subjects and AD autopsy confirmation (n D 35) contributed to increased AD diagnostic certainty. Lastly, all patients were diagnosed by clinicians in the same center in order to preserve homogeneity in the diagnostic groups. Three limitations to the study are known to the authors. First, our study population was recruited in a memory clinic, which might have introduced a selection bias as behavioral symptoms might have contributed to referral. Second, behavioral assessment raters were not blinded for the subjects’ clinical MCI or AD diagnosis. On the other hand, due to strict application of clinical diagnostic criteria, which do not include behavioral changes, behavior did not impact the clinical diagnosis. Third, included subjects were not free of psychotropic medication at the moment of their behavioral observation. However, there is no difference in the prevalence of agitation between the treated and untreated MCI or AD patients and intra-diagnostic group analysis did not reveal important differences in behavior between treated and untreated patients. We did not assess in detail the effects of psychotropic drugs on behavior, as this is not a pharmacological study. Furthermore, psychotropic drug intake rates vary among countries and Belgium is known to have a high prevalence of psychotropic drug utilization in community-dwelling elderly and nursing home residents (Azermai, Elseviers, Petrovic, Van Bortel, & Vander Stichele, 2011). This also partially explains the high
255
prevalence of psychotropic drug intake in MCI and AD patients in this study. A matter of debate could be the wide range of MMSE scores in our MCI population. Some studies use MMSE scores as part of their key eligibility criteria for MCI, e.g. the Alzheimer’s Disease Neuroimaging Initiative (ADNI) with MMSE scores 24 30 (inclusive) for MCI subjects. This might raise questions about our MCI study population with the broad MMSE range of 16 30. First, from the MCI patients with an MMSE score of
