Article Text

PDF

Affective behavioural disturbances in Alzheimer's disease and ischaemic vascular disease
  1. Rita Hargravea,
  2. Laurie C Geckc,
  3. Bruce Reedb,
  4. Dan Mungasb
  1. aDepartment of Psychiatry, University of California, Davis, CA, USA, bDepartment of Neurology, cNashua Clinic of New Hampshire, NH, USA
  1. Dr Rita Hargrave, 4338 Leach Street, Oakland, CA 94602, USA.

Abstract

OBJECTIVES To investigate affective change in Alzheimer's disease and ischaemic vascular disease and examine the contribution of white matter disease to psychopathology in these dementias. Based on earlier studies, it was predicted that: (1) depression would be more prevalent and severe in ischaemic vascular disease; (2) psychomotor slowing would be more prevalent in ischaemic vascular disease; (3) apathy would be more prevalent in ischaemic vascular disease; and (4) The degree of white matter disease would be positively correlated with the severity of psychomotor slowing.

METHODS Ratings of affective/behavioural states and white matter disease were compared in 256 patients with Alzheimer's disease and 36 patients with ischaemic vascular disease or mixed dementia with an ischaemic vascular component using analysis of variance (ANOVA) and linear regression models.

RESULTS The findings were: (1) decreased affect/withdrawal was more prevalent and severe in patients with ischaemic vascular disease and patients with white matter disease; (2) psychomotor slowing was more severe in patients with ischaemic vascular disease and patients with white matter disease; and (3) differences between Alzheimer's disease and ischaemic vascular dementia groups in the degree of psychomotor slowing were independent of the severity of white matter disease.

CONCLUSIONS Future studies using structural and functional neuroimaging techniques would be helpful for examining the relation between neurobiological factors and affective/behavioural disturbances in dementia.

  • dementia
  • behavioural disturbances
  • white matter disease

Statistics from Altmetric.com

Recent studies suggest that behavioural disturbances in dementia may vary according to the aetiology of the dementia. Although the types of behavioural disturbances in Alzheimer's disease and ischaemic vascular disease are similar, there are substantial differences in the symptom profiles of these two types of dementia.1-6 Behavioural disturbances in Alzheimer's disease commonly include anxiety, agitation, psychosis, and personality change.1 7-10Ischaemic vascular disease is often associated with irritability, apathy, and blunted affect.10-12 Patients with ischaemic vascular disease have greater severity of depression and anxiety than patients with Alzheimer's disease.13

A comparative analysis of behavioural disturbances in Alzheimer's disease and ischaemic vascular disease may provide clues about the underlying pathophysiology of affective/behavioural change in dementia. White matter abnormalities are associated with dementia, cognitive impairment, and functional impairment.14-19 Although white matter change may contribute to the development of affective/behavioural changes in dementia, few studies have specifically examined this association.

The purpose of our study was to investigate affective change in Alzheimer's disease and ischaemic vascular disease and to examine the contribution of white matter disease to psychopathology in these dementias. Previous studies have examined discrete affective/behavioural occurrences (for example, hallucinations). Our study focused on continuous, ongoing behaviour patterns that reflect more stable affective or arousal states (for example, depressed affect or agitation). We were interested in continuous affective changes for two reasons. Firstly, affective changes in dementia exert a significant impact on a patient's and caregiver's quality of life and social interactions.20 Secondly, ongoing affective/behaviour patterns may be related to measurable pathophysiological changes in the brain (for example, white matter disease). We developed a new behavioural rating instrument that focuses specifically on affective changes that occur in dementia and provides more comprehensive assessment of affect than is possible with other, existing instruments. Based on earlier studies, we predicted that: (1) depression would be more prevalent and severe in ischaemic vascular disease; (2) psychomotor slowing would be more prevalent in ischaemic vascular disease; (3) apathy would be more prevalent in ischaemic vascular disease; and (4) the degree of white matter disease would be positively correlated with the severity of psychomotor slowing.

Methods

SUBJECTS

Subjects were patients evaluated at the UC Davis Alzheimer's Disease Center (UCD-ADC) after a uniform clinical evaluation protocol. All patients are evaluated by a team of neurologists, geriatricians, neuropsychologists, nurses, and social workers. Routine dementia evaluation laboratory tests, other clinically indicated laboratory tests, and a CT or MRI study of the brain are obtained for each patient. Clinical findings, test results, and imaging films are reviewed in a multidisciplinary case conference and a consensus diagnosis is established. National Institute of Neurological and Communicative Disorders and Stroke/Alzheimer's Disease and Related Disorders Association (NINCDS/ADRDA) diagnostic criteria are used for possible and probable Alzheimer's disease and California ADDTC criteria are used for possible and probable ischaemic vascular disease.21 22

Patients evaluated between September 1992 and November 1995 were included in this study. All patients signed informed consent forms which had been approved by the local internal review boards. All patients (n=378) received initial clinical evaluations and ratings of affective states. Table 1 presents the demographic characteristics of the patients.

Table 1

Demographic characteristics of sample

Thirty six patients had possible or probable ischaemic vascular disease (n=15) or had a mixed dementia with an ischaemic vascular component (n=21). Among patients with ischaemic vascular disease, 20 had subcortical infarcts: 14 had basal ganglia infarcts, 10 had centrum semiovale infarcts, five had thalamic infarcts, and two had brainstem infarcts. Seven patients had cortical infarcts: four in the posterior cerebral artery distribution, two in the middle cerebral artery distribution and one in the anterior cerebral artery distribution. Two patients had both subcortical and cortical infarcts.

VARIABLES

Ratings of affect

The affect ratings utilised in this study constitute one component of the standard evaluation protocol of the UCD ADC. Ratings of 16 variables representing affect and relatively continuous, ongoing behavioural patterns were obtained as part of the routine evaluation. The variables were: (1) depressed affect,(2) anhedonia, (3) decreased energy, (4) anxious affect, (5) irritability, (6) agitation, (7) labile affect, (8) apathy, (9) social withdrawal, (10) decreased affective expression, (11) decreased humour, (12) hyperactivity, (13) slowed movement, (14) inappropriate humour, (15) sweet craving, and (16) slowed thinking. Quantitative methods for rating these variables were developed. Detailed definitions of each variable were created so that raters could objectively distinguish characteristics that at times have subtle differences. Variables were then rated according to a five point scale: 0=the affective/behaviour disturbance in question is not present; 1=the affective/behaviour disturbance is present but intensity is of a very mild degree; 2=present with intensity of mild degree; 3=present with intensity of moderate degree; 4=present with intensity of extreme degree. The intensity of a variable was rated based on the degree to which most people would regard the affective state/behaviour pattern as unusual or abnormal and the degree to which the behaviour is responsive to environmental change.

Interrater reliability of ratings was empirically tested by having tape recorded interviews with caregivers of 35 patients independently rated by two raters. Spearman r coefficients were calculated to assess interrater reliability for each variable. These reliability coefficients are presented in table 2. The mean interraterr was 0.74 (SD 0.08). Principal components analysis of the 16 ratings, based on the overall sample of 378 participants, yielded four components accounting for 66.4% of total variance. Rotated component loadings are shown in table 2. These components seem to measure decreased affect/withdrawal (component 1), agitation/irritability (component 2), disinhibition (component 3), and psychomotor speed (component 4). Interrater reliability components for coefficients component scores are presented in table 2. These results demonstrate very acceptable levels of interrater reliability, particularly for measuring complex phenomena.

Table 2

Interrater reliability and principal component loadings of affect ratings

Clinical ratings were based on semistructured caregiver interviews and information from detailed behaviour questionnaires completed by the caregivers. Ratings referred to a 1 month period preceding the interview. Ratings were made by doctoral level professionals with formal training in the assessment process.

Ratings of white matter disease

Neuroimaging films of each patient were reviewed by a UCD ADC neurologist and quantitative ratings were made for selected variables. We were particularly interested in the degree of white matter abnormality in this study. This variable was rated on a four point scale: 0=no white matter change beyond that expected on the basis of the patient's age; 1=mild white matter abnormality; 2=moderate abnormality; 3=severe abnormality. The same scale was used for rating CT images and MRI images. Interrater reliability was tested by having two neurologists independently rate films of 18 patients (10 MRI , 8 CT). The Spearman r coeffient comparing the two ratings of degree of white matter abnormality was 0.83, indicating good interrater agreement.

DATA ANALYSIS

JMP Statistical Discovery software was used for data analyses.23 Data analyses considered three primary questions: (1) the relation of diagnosis (Alzheimer's diseasev dementia with an ischaemic vascular component (ischaemic vascular disease)) to affect ratings; (2) the relation of white matter disease to affect; and (3) the interactive effects of diagnosis and white matter disease. For each analysis, a multivariate analysis of variance (MANOVA) was used in which component scores from each of the four principal components underlying the 16 affect ratings were dependent variables. These component scores are uncorrelated, linear combinations of the 16 affect ratings optimally weighted to provide a measure of each dimension, and the simple correlation of a given rating with a component score is equal to the loading of that rating on that component. The primary variable(s) of interest were included as independent variables. Sex, age, education, and mini mental state examination (MMSE) score were added as independent variables to control for effects of demographic variables and overall degree of dementia. t Tests were used to test for group differences in age, education, and MMSE scores. A χ2 test was used to test for group differences in sex. If the overall multivariate effect for all four dependent variables was significant, then individual univariate analyses of variance (ANOVAs) were performed for each of the four component scores. Secondary analyses were performed to examine effects on the 16 individual affect ratings. Bonferroni correction methods were used to adjust the significance level for multiple comparisons (p=0.05/16=0.0031).

Results

AFFECT RATING DIFFERENCES BETWEEN ALZHEIMER'S DISEASE AND ISCHAEMIC VASCULAR DEMENTIA

The first step of data analysis involved comparing patients with probable Alzheimer's disease (n=195) with patients in the ischaemic vascular dementia group (n=36). Table 3 shows demographic characteristics of patients in these groups. Groups did not significantly differ in age, education, MMSE scores, or sex.

Table 3

Demographic characteristics of probable Alzheimer's disease and ischaemic vascular dementia groups

A MANOVA was performed on the four affect rating components comparing the probable Alzheimer's disease group with the ischaemic vascular disease group, controlling for effects of sex, age, education, and MMSE. The Alzheimer's disease-ischaemic vascular disease effect in the overall model was significant (F (4, 219)=9.6, p<0.001, Wilk's lambda=0.98), indicating that there was a statistically significant group difference for at least one of the four dependent variables. Figure 1 presents adjusted group means and standard errors of the mean for each dependent variable.

Figure 1

Means (SEM) on affect rating components for Alzheimer's disease and ischaemic vascular dementia groups. Values are expressed in standard score units with a mean of “0” and SD if “1” based upon the overall subject sample.

ANOVAs were then performed for each dependent variable using the same five independent variables (Alzheimer's disease-ischaemic vascular disease, sex, age, education, MMSE). Significant Alzheimer's disease-ischaemic vascular disease group differences were noted for component 1 (decreased affect/withdrawal) (F=5.4; df=1, 222; p<0.03), with these characteristics present to a greater degree in in patients with ischaemic vascular dementia. Demographic variables and MMSE were not significantly related to this component. Component 4 (psychomotor speed) showed highly significant Alzheimer's disease-ischaemic vascular disease differences, (F=32.8; df=1, 222; p<0.001), with significantly slower psychomotor speed for the ischaemic vascular disease group. Sex (F(1, 222)=32.8 p<0.002; M female=-0.18, M male=-0.63) and MMSE (F=16.0; df=1, 222; p<0.001;r=0.26) effects were also significant. Alzheimer's disease-ischaemic vascular disease effects for the two other dependent variables were not significant.

Separate ANOVAs were performed using the 16 affect rating variables as dependent variables and including the five independent variables used in the above analyses. Bonferroni correction with a p value of 0.003 (0.05/16=0.0031) was used to control for the number of comparisons. Adjusted group means and standard errors for each variable, and significance levels of group differences are presented in table 4. Results are consistent with previous results. Several variables with loadings of 0.50 or greater on component 1 (decreased affect/withdrawal) and all variables defining component 4 (psychomotor speed) showed significant group differences. The ischaemic vascular disease group consistently showed higher scores on these ratings. None of the variables on the agitation/irritability or the disinhibition components differentiated the two groups. The depressed affect rating, which loaded on both component 1(decreased affect/withdrawal) and component 2 (agitation/irritability) components, clearly did not significantly differ across groups.

Table 4

Means and p values for Alzheimer's disease-ischaemic vascular dementia comparisons

RELATION OF AFFECT RATINGS AND WHITE MATTER CHANGES

The next phase of data analysis examined the relation of affect ratings to semiquantitative ratings from neuroimaging depicting varying degrees of white matter disease. All subjects with neuroimaging data were included in these analyses. Analyses were similar to those used to compare diagnostic groups. Firstly, a multivariate general linear model was used in which the four component scores were dependent variables. White matter disease was the primary independent variable, and sex, age, education, and MMSE were also included as secondary independent variables. The overall white matter disease effect was significant, (F(4,275)=3.6; p<0.008 Wilk's lambda=0.95), so individual ANOVAs were performed with each of the four component scores. Figure 2 shows the relation between white matter disease ratings and component scores. Component score means were calculated for patients with white matter disease ratings of 0, 1, 2, and 3 and these means and standard errors are presented in fig2.

Figure 2

Means and standard errors of means of affect rating principal components across categories of white matter disease (WMD). Values are expressed in standard score units with a mean of “0” and standard deviation if “1” based upon the overall subject sample.

Analyses were performed using each affect rating as a dependent variable and white matter disease, sex, age, education, and MMSE as independent variables. Significant white matter disease effects were found for decreased energy (p<0.001), slowed movement (p<0.002), and apathy (p<0.003). Consistent with Alzheimer's disease-ischaemic vascular disease analyses, individual variables from the decreased affect/ withdrawal and psychomotor speed components were significantly influenced by degree of white matter disease. Variables from agitation/irritability and disinhibition components were not significantly related.

INTERACTIVE EFFECTS OF DIAGNOSIS AND WHITE MATTER DISEASE ON AFFECT

A final group of analyses was performed to determine if Alzheimer's disease-ischaemic vascular disease and white matter disease have independent and interactive effects on affect. These analyses were performed using scores from component 1 (decreased affect/withdrawal) and component 4 (psychomotor speed) because these two components were both related to Alzheimer's disease-ischaemic vascular disease and white matter disease. Patients with neuroimaging data who had either the diagnosis of probable Alzheimer's disease (n=153) or an ischaemic vascular disease component to their dementia (n=30) were used for these analyses. Patients were divided into two groups based on white matter disease ratings: (1) those with a score of 0 or 1 (no to mild white matter disease; Alzheimer's disease n=137, ischaemic vascular disease n=14), and (2) those with a score of 2 or 3 (moderate to severe white matter disease; Alzheimer's disease n=16, ischaemic vascular disease n=16). None of the patients with Alzheimer's disease had severe white matter disease. Each of the two component scores was entered as a dependent variable. Independent variables were the Alzheimer's disease-ischaemic vascular disease main effect, the white matter disease group main effect, the Alzheimer's disease-ischaemic vascular disease by white matter disease group interaction effect, and covariates sex, age, education, and MMSE. Adjusted group means and standard errors of the decreased affect/withdrawal and psychomotor speed component scores are presented in figs 3 and 4.

Figure 3

Means and standard errors of means on the decreased affect/withdrawal component for groups defined by degree of white matter disease (no WMD or mild WMD versus moderate or severe WMD). Values are expressed in standard score units with a mean of “0” and standard deviation if “1” based upon the overall subject sample.

Figure 4

Means and standard errors of means on the psychomotor slowing component for groups defined by degree of white matter disease (no WMD or mild WMD versus moderate or severe WMD. Values are expressed in standard score units with a mean of “0” and standard deviation if “1” based upon the overall subject sample.

The Alzheimer's disease-ischaemic vascular disease (F(1, 172)=1.7, p<0.20) and white matter disease group (F(1, 172)=1.7, p<0.18) main effects and their interaction (F<1.0) were not significant for the decreased affect/withdrawal component. The lack of significance for either main effect is noteworthy because both Alzheimer's disease-ischaemic vascular disease and white matter disease change were related to this component in previous analyses. This difference in results is likely due to removal from each of the main effects of shared variance that is related to the dependent variable, thereby weakening the relation of both independent variables with the dependent variable. Results show the same general trend in which white matter disease corresponds to higher scores on this component (fig 3) but the degree of group differences failed to reach statistical significance.

For psychomotor speed, the Alzheimer's disease-ischaemic vascular disease main effect was significant (F(1,172)=28.6, p<0.001), but the white matter disease group main effect and the interaction effect were not significant (Fs<1.0). These results indicate that there are very striking Alzheimer's disease-ischaemic vascular disease differences in psychomotor speed that are independent of degree of white matter disease. Conversely, white matter disease effects previously found were likely related to confounding effects of Alzheimer's disease-ischaemic vascular disease on the white matter disease variable.

Discussion

Our study provides additional information on psychopathology in dementia and the influence of vascular disease on affective states/behavioural changes in Alzheimer's disease and ischaemic vascular disease. There were three main findings: (1) decreased affect/withdrawal was more prevalent and severe in patients with ischaemic vascular disease and in patients with white matter disease; (2) psychomotor slowing was more prevalent among patients with ischaemic vascular disease and patients with white matter disease; and (3) the degree of psychomotor slowing in ischaemic vascular disease was independent of the severity of white matter change. Secondary findings were: (1) psychomotor slowing was more severe in men than women and (2) psychomotor slowing was associated with lower MMSE scores in both Alzheimer's disease and ischaemic vascular disease.

The findings of higher scores on component 1 (decreased affect/withdrawal) among patients with ischaemic vascular disease supports our original hypotheses about depression and apathy and has been reported by other investigators.11 12 21 22Greater prevalence of decreased affect/withdrawal in patients with white matter disease is consistent with earlier studies of dementia and strokes.11 12 21 22 Our data show a quantitative relation between degree of white matter disease and severity of apathy, decreased affect, and social withdrawal. A direct quantitative relation of this nature has not been previously reported in the literature. Higher prevalence of psychomotor slowing among patients with ischaemic vascular disease and patients with white matter disease is consistent with earlier studies.11 21 24 Although the specific pathophysiology of apathy, depression, or psychomotor slowing is unknown, recent studies suggest that these symptoms may be due to pathological changes in the frontal-subcortical pathways and frontal lobe perfusion deficits and dysfunction of dopaminergic, serotonergic, and noradrenergic neurotransmission.11 21 24-32

Greater psychomotor slowing in ischaemic vascular disease than in Alzheimer's disease independent of the severity of white matter was an unexpected finding. The absence of a positive correlation between the severity of psychomotor slowing and degree of white matter disease has not been previously reported. Our results suggest that additional neuroanatomical and neurochemical factors beyond severity of white matter disease may contribute to greater psychomotor slowing in ischaemic vascular disease. For example, effects of discrete infarcts in specific subcortical structures may be more important than white matter changes in producing psychomotor slowing. Our sample of patients with ischaemic vascular disease was not large enough to examine the correlation of psychomotor slowing with infarcts in specific structures.

Greater psychomotor slowing among men than women has not been previously reported. Sex differences in psychomotor slowing persisted even after adjustments for age, education, and MMSE scores. Previous studies of sex differences in Alzheimer's disease have yielded conflicting results. Some investigators have reported sex differences in cognitive deficits, mood disorders, behavioural disorders, and brain metabolism.33-36 Other studies report no consistent sex differences in the clinical presentation of Alzheimer's disease.37 38

Future research is needed to validate our results and examine the contribution of additional demographic and neurobiological factors to the pathophysiology of affective states/behavioural disturbances in dementia. Future research could consider issues such as (1) the relation between infarct location and the progression of affective change in dementia and (2) the relation between infarct location and severity of psychomotor slowing in ischaemic vascular disease. Studies utilising functional imaging techniques such as PET might be valuable for relating affect patterns and changes to function of specific cortical and subcortical structures.

Acknowledgments

This work was performed at the University of California, Davis Alzheimer's Disease Center at 1771 Stockton Blvd, Suite 2005, Sacramento, California 95816.This study was supported in part by grants AG10220 and AG10129 and AG12435 from the National Institute on Aging, Bethesda, MD, and by the California Department of Health Services Alzheimer's Disease Program, contracts 91–13044 and 91–13045.

References

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.