Article Text


Behavioural and psychological symptoms in vascular dementia; differences between small- and large-vessel disease
  1. Salka S Staekenborg1,
  2. Tanja Su1,
  3. Elisabeth C W van Straaten1,
  4. Roger Lane2,
  5. Philip Scheltens1,
  6. Frederik Barkhof3,
  7. Wiesje M van der Flier1,4
  1. 1Department of Neurology, Alzheimer Center, Vrije Universiteit Medical Center, Amsterdam, The Netherlands
  2. 2Novartis Neuroscience, Novartis Pharmaceuticals Corporation, East Hanover, New Jersey, USA
  3. 3Department of Radiology, Alzheimer Center, Vrije Universiteit Medical Center, Amsterdam, The Netherlands
  4. 4Department of Epidemiology & Biostatistics, Vrije Universiteit Medical Center, Amsterdam, The Netherlands
  1. Correspondence to Dr S S Staekenborg, Department of Neurology and Alzheimer Centre, Vrije Universiteit Medical Centre, PO Box 7057, 1007 MB, Amsterdam, The Netherlands; s.staekenborg{at}


Aim The authors investigated the prevalence of behavioural and psychological symptoms in vascular dementia (VaD) from baseline data of the VantagE study and compared the severity and relative frequency of symptoms between small-vessel VaD and large-vessel VaD.

Methods Behavioural and psychological symptoms of 484 VaD patients included in a large multicentre clinical trial (registration number NCT00099216) were determined using the 12-item Neuropsychiatric Inventory (NPI). Symptoms were considered present when the score was ≥1. Based on MRI, patients were classified as having small-vessel VaD (83%) or large-vessel VaD (17%).

Results Behavioural and psychological symptoms were reported in 92% of the VaD patients. The median NPI score of the total study population was 9 (0–76), with a median number of three symptoms per patient. Apathy (65%) was most prevalent, followed by depressive symptoms (45%), irritability (42%) and agitation/aggression (40%). Patients with small-vessel VaD reported more apathy, aberrant motor behaviour and hallucinations than patients with large-vessel VaD (p<0.05). In contrast, patients with large-vessel VaD reported a higher severity of agitation/aggression and euphoria (p<0.05).

Conclusion Behavioural and psychological symptoms are common in VaD. Patients with small-vessel and large-vessel VaD demonstrate different profiles of symptoms, with especially more apathy in small-vessel VaD and more agitation/agression in large-vessel VaD.

  • Vascular dementia
  • behavioural and psychological symptoms
  • MRI
  • behavioural disorder

Statistics from


Behavioural and psychological symptoms are increasingly recognised as important clinical features of the dementia syndrome. These symptoms impact the quality of life of patients with dementia, and have been associated with increased care giver burden, more rapid progression of cognitive and functional decline, earlier institutionalisation and mortality.1 In Alzheimer's disease (AD), a prevalence of about 90% for behavioural and psychological symptoms has been reported.2 3 Vascular dementia (VaD) is the second most common type of dementia worldwide, but in contrast to AD, only a couple of small studies on behavioural and psychological symptoms have been conducted in VaD patients.

In spite of the modest body of literature, diagnostic criteria for VaD such as the criteria of the National Institute of Neurological Disorders and Stroke Association Internationale pour la Recherche et l'Enseignement en Neurosciences (NINDS-AIREN) specify that clinical features consistent with the diagnosis include personality and mood changes, abulia, depression and emotional incontinence.4 Results of studies on behavioural and psychological symptoms in VaD differ slightly, but apathy, depressive symptoms and agitation/aggression have been reported as symptoms of highest prevalence and severity.5–10 These former studies are hampered by methodological shortcomings. First, VaD was mostly not the main diagnosis under study, as studies investigated a large sample of AD patients with a small additional group of VaD patients. Furthermore, due to differences in research populations, diagnostic criteria and methods to assess presence of symptoms, results are difficult to compare.

According to the NINDS-AIREN criteria for VaD, neuroimaging is required to demonstrate cerebrovascular disease. VaD can be based on small-vessel disease (lacunes, white matter hyperintensities or bilateral thalamic lesions) or large-vessel disease (large territorial or strategic infarcts). We hypothesised that the behavioural and psychological profile of VaD patients differs according to type of underlying vascular disease. In small-vessel VaD, symptoms such as apathy would be expected due to disruption of corticosubcortical circuits, while in large-vessel VaD, a wide range of symptoms would be plausible, related to size, location and cortical involvement of the large-vessel infarct. For example, after ischaemic stroke, a broad variety of psychological and behavioural symptoms have been demonstrated with depressive symptoms and irritability being of highest frequency.11 In VaD, only two previous reports made the differentiation between small-vessel and large-vessel disease.6 9 Both studies used computed tomography with different criteria to define small-vessel and large-vessel VaD, and inconsistent results were reported.

The aim of this study was to determine the presence of behavioural and psychological symptoms in a large cohort of 484 VaD patients enrolled in a clinical trial, and to compare the severity and prevalence of symptoms according to type of underlying vascular disease (small-vessel or large-vessel) as assessed on MRI using well-defined radiological criteria.12


Study design and patients

We examined the baseline data of patients enrolled in the VantagE study (registration number NCT00099216). The VantagE study was a multicentre, phase III, prospective, randomised, double-blind, placebo-controlled clinical trial of the effects of rivastigmine in patients with mild to moderate VaD.13 Trial inclusion criteria included both fulfilment of the DSM-IV diagnostic criteria for VaD and fulfilment of the NINDS-AIREN criteria for VaD according to central assessment of the neuroimaging criteria.4 The NINDS-AIREN criteria for probable VaD were slightly modified: if neuroimaging criteria for subcortical VaD were met as assessed by the central neuroradiologist, patients were not required to have evidence of a temporal relationship between the dementia syndrome and the evidence of cerebrovascular disease. Accordingly, patients with cortical VaD entered the study with a clinical diagnosis of probable VaD, but patients with subcortical VaD were permitted to enter the study with a clinical diagnosis of possible VaD by NINDS-AIREN criteria. Excluded from entry in the study were patients with a current diagnosis of any primary neurodegenerative disorder, a history of stroke within the 3 months before baseline unless the patient was considered to have fully stabilised in function, a current DSM IV diagnosis of major depression, any other DSM-IV Axis 1 diagnosis that might have interfered with the response of the patient to study medication, initiated treatment with newly prescribed dopaminergic agents during the 4 weeks prior to baseline or ingested lithium during the 4 weeks prior to screening. Patients with space occupying lesions or lobar haemorrhages were excluded. For the current study, patients were required to have data on psychological and behavioural symptoms (n=525). Furthermore, MRI data had to fulfil either criteria of small-vessel VaD or criteria of large-vessel VaD (see below; n=484). Patients who fulfilled criteria for both small-vessel and large-vessel disease were excluded (n=39). All patients gave written informed consent. The study was approved by the local Ethics Committees.

Baseline clinical assessment

Diagnostic evaluation included complete medical history, physical and neurological examination, laboratory tests, neuropsychological testing including Mini-Mental State Examination (MMSE) and MRI of the brain. The use of sleep medication (yes/no), antidepressants (yes/no) and other behavioural regulating medication (including neuroleptics and anxiolytics; yes/no) was recorded. Antidepressants and anxiolytics were permitted if a constant dose had been maintained for 3 months prior to baseline evaluations, and patients who were being treated with low doses of neuroleptics and in whom discontinuation was not considered feasible were allowed to remain on their current dose during the study, with the recommendation that maintenance doses should be kept as low as needed to manage the patient. Global staging of dementia severity was performed with the Global Deterioration Scale (GDS). To evaluate behavioural and psychological symptoms (including delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy, disinhibition, irritability, aberrant motor behaviour, night-time behaviour disturbances and appetite changes), the Neuropsychiatric Inventory (NPI, 12-item), an informant rated instrument, was used.14 Using the NPI, the 12 behavioural domains are evaluated by sample questions including (but not limited to): delusions (‘Does the patient believe that others are stealing from him?’); hallucinations (‘Does the patient talk to people who are not there?’); agitation/aggression (‘Is the patient uncooperative, resistive to help from others?’); depression/dysphoria (‘Does the patient say or act that he is sad or in low spirits?’) anxiety (does the patient say that he is worried about planned events?); euphoria (‘Does the patient find humour in and laugh at things that others do not find funny?’); apathy (‘Does the patient seem less spontaneous and less active than usual?’); disinhibition (‘Does the patient say crude things or make sexual remarks that he would not usually say?’); irritability (‘Does the patient have sudden flashes of anger?’); aberrant motor behaviour (‘Does the patient pace around the house without apparent purpose?’); night-time behaviour disturbances (‘Does the patient have problems with sleeping and are they often awake during the night?’); appetite changes (‘Has the patient changed in appetite, eating habits or in physical weight?’). For each neuropsychiatric domain, severity (0=absent, 1=mild, 2=moderate, 3=severe) and frequency (0=absent, 1=occasionally, less than once per week, 2=often, about once per week, 3=frequently, several times per week but less than every day, 4=very frequently, once or more per day or continuously) are rated. Subsequently, the domain score is calculated as the product of the frequency and severity. The sum of the 12 domain scores provides the total NPI score (maximum score=144). Presence of behavioural and psychological symptoms was defined as a score of ≥1.

MRI protocol

All patients underwent MRI examination before randomisation. Scanners operating between 0.5 and 1.5 T were used. Axial spin-echo T2-weighted images (T2-WI; echo time (TE): 80–120 ms; repetition time (TR): 3000–4000 ms; slice thickness=5 mm); axial fluid-attenuated inversion recovery (FLAIR) images (TE: 110–150 ms; TR: 9000–10 000 ms; inversion time: 2000–2200 ms; slice thickness=5 mm); and axial, sagittal, and coronal spin-echo T1-weighted images (T1-WI; TE: 11–20 ms; TR: 500–700 ms; slice thickness=5 mm) were acquired.

Image assessment

Image assessment was performed centrally at the Image Analysis Center (VU Medical Center, Amsterdam, The Netherlands) by agreement of two experienced readers blinded to clinical information, with the use of digital image files. The assessment of vascular abnormalities included the items of the radiological NINDS-AIREN criteria for VaD, according to operational definitions earlier proposed.12 Based on these criteria, patients were classified as having large-vessel VaD (strategic large-vessel infarct of the dominant hemisphere or bilateral hemispheric strokes) or small-vessel VaD (white-matter hyperintensities involving at least a quarter of the white matter, multiple lacunes or bilateral thalamic lesions).


A statistical analysis was performed by means of SPSS 14.0 (SPSS, Chicago, Illinois). To compare baseline characteristics between small-vessel and large-vessel VaD, χ2 tests were used for dichotomous variables, Mann–Whitney U tests for non-parametric data and independent sample t tests for continuous data. Comparison of the total NPI score, the total number of symptoms and the NPI score per symptom between small-vessel VaD and large-vessel VaD was performed using Mann–Whitney U tests. The prevalence of symptoms (present/not present) was compared between large-vessel and small-vessel VaD using the χ2 test. Subsequently, to control for age, sex, MMSE, dementia duration and the use of sleep medication, antidepressants and other behavioural regulating medication, logistic regression analysis was performed with the individual NPI symptoms as dependent variable, and the different types of VaD (small-vessel or large-vessel) as independent variable.


Baseline demographics are shown in table 1. The total population included n=484 VaD patients with a mean (SD) age of 72(8) years and included 182 (38%) women. On average, patients were mildly to moderately demented with a mean MMSE score of 19 (4) and a median (range) GDS score of 4 (2–6). Based on the operational definitions for the radiological part of the NINDS-AIREN criteria, n=401 (83%) had small-vessel VaD, and n=83 (17%) had large-vessel VaD. Patients with small-vessel VaD were slightly older and less educated, but had a shorter duration of dementia compared with patients with large-vessel VaD. The use of psychotropic medication was comparable in both groups.

Table 1

Patient demographics and characteristics

In the total cohort, the prevalence of any behavioural or psychological symptom was 92%, with a median NPI score of 9 (0–76). A median number of three (0–11) symptoms per patient were reported. Thirty-seven patients did not report any symptom, and two patients reported 11 of the 12 NPI symptoms. Apathy was the symptom that was reported most often (65% of the total population), followed by depressive symptoms (45%), irritability (42%) and agitation/aggression (40%). Hallucinations (6%) and euphoria (6%) were the symptoms that were reported most infrequently.

No differences were found between small- and large-vessel VaD comparing the total NPI score or the total number of reported symptoms per patient. However, there appeared to be differences in the individual NPI symptoms between small-vessel VaD and large-vessel VaD (figure 1). Patients with small-vessel VaD had a higher severity of apathy, aberrant motor behaviour and hallucinations than patients with large-vessel VaD. In contrast, symptoms of agitation/aggression and euphoria were more severe in large-vessel VaD compared with small-vessel VaD. The prevalence of individual NPI symptoms in small-vessel and large-vessel VaD is shown in figure 2. Patients with small-vessel VaD had a higher prevalence of apathy, aberrant motor behaviour and hallucinations than patients with large-vessel VaD. Patients with large-vessel VaD were reported as having more frequent euphoria compared with patients with small-vessel VaD. Furthermore, the prevalence of agitation/aggression (48% compared with 38%) and irritability (49% compared with 41%) was higher in large-vessel than small-vessel VaD, but these differences did not reach significance. Logistic regression analysis with adjustment for age, sex, MMSE, dementia duration and the use of sleep-, antidepressant- and behavioural-medication yielded comparable results, with a higher risk of apathy (OR (95% CI) 1.8 (1.1;3.0)) and aberrant motor behaviour (OR (95% CI) 3.2 (1.1;9.4)) for small-vessel VaD than large-vessel VaD, and more euphoria (OR (95% CI): 2.4 (1.1; 5.5)) in large-vessel VaD than in small-vessel VaD.

Figure 1

Severity of Neuropsychiatric Inventory symptoms in small-vessel vascular dementia (VaD) and large-vessel VaD: mean scores of individual behavioural and psychological symptoms in patients with small-vessel VaD (n=401) and large-vessel VaD (n=83). Please note that, although means (with SD) are represented in the figure, statistics were performed using Mann–Whitney U tests. *Higher in small-vessel VaD, p<0.05; §higher in large-vessel VaD, p<0.05.

Figure 2

Prevalence of Neuropsychiatric Inventory symptoms in small-vessel vascular dementia (VaD) and large-vessel VaD. Prevalence of individual psychological and behavioural symptoms in patients with small-vessel VaD (n=401) and large-vessel VaD (n=83). A symptom was considered present with a Neuropsychiatric Inventory score ≥1. Data were compared using χ2 tests.


The main finding of the present study is that the profile of neuropsychiatric symptoms differs between small-vessel and large-vessel VaD. Apathy, aberrant motor behaviour and hallucinations are more severe and more prevalent in patients with small-vessel VaD compared with large-vessel VaD. Conversely, agitation/aggression and euphoria are more severe in patients with large-vessel VaD.

The observed prevalence of 92% of any behavioural and psychological symptom in the total study population is in agreement with earlier reports.5 8 10 15 Furthermore, the highest prevalence for apathy, followed by depressive symptoms, irritability and agitation/aggression confirms previous findings in populations of VaD patients.5 6 Other studies have demonstrated a slightly different order of these symptoms, for example with the highest frequency of depressive symptoms, followed by agitation/aggression and apathy.8 10 In our study, apathy had the greatest severity and prevalence in both small-vessel and large-vessel VaD. Both severity and prevalence were even higher in small-vessel VaD than in large-vessel VaD. Apathy may be induced by changes in the neural networks that generate and control goal-directed actions, which are mostly represented within the prefrontal cortex connections to basal ganglia, thalamus and limbic system structures.16 It seems plausible that disruption of the white-matter tracts between frontal cortex and basal ganglia by severe WMH may result in apathy. Furthermore, other vascular damage according to the definition of small-vessel disease includes multiple lacunes (at least two lacunes in the frontal lobe and two lacunes in the basal ganglia) or bilateral thalamic lesions, also structures that are thought to play a role in the origin of apathy.12 16 In line with our results, a recent study in patients with cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy (CADASIL) demonstrated that CADASIL patients with apathy had a higher WMH burden and more lacunes than patients without apathy.17 Furthermore, apathy has been recognised as an important clinical feature in many neurodegenerative diseases, such as AD, Parkinson's disease and supranuclear palsy.18–20 We have no neuropathological confirmation of the diagnosis of VaD, and the possibility of concomitant pathology or even misdiagnoses cannot be excluded. Previously, we reported in the same cohort VaD patients moderate to severe medial temporal lobe atrophy (MTA) and cortical atrophy, with the atrophy scores being higher in patients with small-vessel than large-vessel VaD. This MTA and cortical atrophy may be caused by secondary neurodegeneration due to ischaemia but also indicate the presence of a concomitant neurodegenerative disease such as AD. However, all our patients were carefully diagnosed by the clinical and radiological NINDS-AIREN criteria, which are known to have a high specificity.21

Patients with large-vessel VaD showed a different profile of neuropsychiatric symptoms, with a higher severity of agitation/aggression and euphoria than small-vessel VaD. We expected a wide variety of symptoms in large-vessel VaD, but the total NPI score and the total number of symptoms were comparable with small-vessel VaD. In contrast to our results, a previous study in patients with VaD found a higher score for each symptom in large-vessel than with small-vessel VaD, but without any significant differences between both groups.6 However, to differentiate between small-vessel and large-vessel disease, computed tomography was used instead of MRI, which may have influenced the subdivision of VaD patients. A high severity of agitation/aggression in VaD and specifically to large-vessel VaD has been previously described.6 8 10 Moreover, aggressive behaviour has been described as a possible and, if present, important clinical feature after hemispheric infarction in territories of the anterior, middle or posterior cerebral arteries.22 23

Other symptoms were equally distributed between small-vessel and large-vessel VaD, such as depressive symptoms. In accordance with the ‘vascular depression hypothesis,’ subcortical vascular lesions frequently have been related to depression at late life, and also after ischaemic stroke the risk of depression has been estimated at 33%.24–26 We have to think about how we can place our findings in a hypothetical framework. We found a high prevalence of depressive symptoms in both small-vessel (43%) and large-vessel VaD (45% of the patients), suggesting a possible vascular role in the pathogenesis of depression. However, which part can be explained by white-matter changes, the location of a cerebral lesion (especially disruption of frontal- and frontostriatal pathways) or other factors such as disability of the patient caused by vascular lesions, has to be clarified in future research.

Strengths of the current study include the large study population, as it is one of the largest clinical series of patients affected by VaD to date, and the fact that a wide spectrum of psychiatric domains was investigated using the NPI. All patients were carefully screened for fulfilment of the clinical and radiological NINDS-AIREN criteria for VaD, generally considered accurate and specific criteria. We chose to use a low cut-off for symptoms being classed as present/absent comparable with most other studies in the field. However, some other reports used a NPI score >4 per symptom as a cut-off to define ‘clinically relevant symptoms,’ which makes the results of different studies not always interchangeable. Limitations include setting, entry criteria and other design features of a randomised clinical trial, which may have introduced a selection bias on the inclusion of patients. Although this might have led to an underestimation of symptoms, we do not believe it has played a role in the differences between small-vessel and large-vessel disease. Another limitation includes the study design of a cross-sectional cohort study, which precludes the assessment of causality or the evolution of symptomatology. Furthermore, the informant-based structure of the NPI, which uses information of the care giver to grade the occurrence and severity of neuropsychiatric symptoms in patients, may partially reflect the coping style of the care giver.

In conclusion, our study shows a high prevalence of behavioural and psychological symptoms among VaD patients, in both small- and large-vessel VaD. Furthermore, our results demonstrate that small-vessel VaD demonstrates a different profile of behavioural and psychological symptoms compared with large-vessel VaD. In particular, apathy, aberrant motor behaviour and hallucinations were more severe and more common in small-vessel VaD than in large-vessel VaD, possibly reflecting disruption of white-matter tracts between frontal cortex and basal ganglia, while euphoria and agitation/aggression were more severe among patients with large-vessel VaD. The treatment and prognostic implications of the different neuropsychiatric profiles between patients with small-vessel and large-vessel VaD remain to be elucidated in further research. However, recognition of these symptoms may be of importance with regard to optimising care and determining prognosis.


View Abstract


  • Competing interests RL is an employee of Novartis. Please note that this was a post-hoc analysis of baseline data and does not discuss treatment.

  • Ethics approval Ethics approval was provided by METc VUmc.

  • Provenance and peer review Not commissioned; externally peer reviewed.

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.