Differentiation of vascular dementia from AD on neuropsychological tests

Intelligence.

Five global measures of intelligence were used in five studies. In two assessments in two studies (40%), VaD patients performed significantly better on the verbal subtests of the Wechsler Adult Intelligence Scale (WAIS-R) than did AD patients.33,34

Three assessments in three studies (60%) failed to show any significant differences between AD and VaD groups on the WAIS information subtest,35-36 WAIS-R, and Wechsler Memory Scale (WMS-R).37

Language.

There were 32 assessments of language function in 20 studies. Nineteen assessments (59%) of language function in 16 studies showed no significant difference. Using the Controlled Oral Word Association Test (COWAT) and the Boston Naming test (BNT), 16 studies demonstrated no significant differences between AD and VaD patients.35-49 None of the studies using the Token test reported a significant difference between the groups.44-46

Nine assessments in nine studies (28%) demonstrated the AD group was superior to the VaD group in language function. Five recent studies with good matching and utilizing newer diagnostic criteria found that AD patients performed superiorly on the COWAT compared with VaD patients.35,36,46,47,50 Three qualitative studies investigating reading, writing, comprehension, and grammar found that AD patients performed better than VaD patients.51-53 In telling a story about a picture, AD patients generated more words per minute than did VaD patients.41

Four assessments in four studies (13%) found VaD patients to perform superiorly to AD patients on the BNT42,44 and phrase repetition.40,53

Attention/immediate memory.

Ten assessments in nine studies investigated attention/immediate memory. No significant difference was found in eight assessments in seven studies (80%). No significant difference was found between patients in six studies on digit span.34,35,43,44,46,49 No difference was noted on the Corsi Block Tapping Test.44 No significant result difference was noted on memory registration.41

Two assessments in two studies (20%) found that AD patients performed better than VaD patients. VaD patients were superior to AD patients in word span.52 In the Corsi test, AD patients were found to have a significantly higher rate of forgetting.54

Verbal learning and memory.

Eighteen studies investigated verbal learning and memory. VaD patients demonstrated clearly superior performance on verbal learning and memory compared with AD patients in 11 studies (61%). Three studies35,44,52 found that story recall in VaD patients was superior to that of AD patients. On the California Verbal Learning Test (CVLT), VaD patients had lower rates of false positives and intrusions than did AD patients50 as well as general superiority.35,41 In the Rey Auditory Verbal Learning Test (RAVLT), VaD patients performed better,43,54 the latter study showing that AD patients had an increased forgetting rate. Superiority in the Hopkins Verbal Learning Test and Fuld Retrieval Test was found for VaD patients versus AD patients.40,42 On a memory assessment scale, VaD patients performed better and AD patients had higher rates of intrusions.47 VaD patients had superior free recall (Consortium to Establish a Registry for Alzheimer’s Disease neuropsychological test battery).55

In only one study (6%) did AD patients perform better than VaD patients on the CVLT.56

Six out of 18 studies (33%) showed no significant difference between the VaD and AD groups. Studies showing no difference in verbal long-term memory function used selective reminding and mnestic subtests of the Luria D-battery33,46,51,57 and cued recall in the CVLT.36 The above studies are summarized in table 1.

Nonverbal memory.

Seven assessments in six studies investigated nonverbal memory. Five assessments in five studies (71%) demonstrated no significant difference between VaD and AD patients on tests of nonverbal memory. These tests included: visuospatial memory,43 WMS-R visual reproduction subtest35,44; the Benton Visual Retention test46; and Famous Faces and Recognition Memory for Faces.58

Two assessments in two studies (29%) showed VaD function to be superior. Superiority of VaD patients’ performance on the 7/24 Spatial Recall and Rey-Osterrieth Complex Figure Recall has also been demonstrated.35,49

Conceptual function.

Ten assessments in eight studies investigated conceptual function. Six assessments (55%) in five studies investigating conceptual function found no significant differences. Three studies found no significant differences for WAIS-R similarities.34,40,46 Raven’s Coloured Progressive Matrices in two studies,44,46 and Picture Arrangement and Picture Completion in one,34 were also negative.

Three assessments in three studies (27%) showed VaD patients performed better than AD patients on Picture Completion33 and Raven’s Coloured Progressive Matrices.43,52

In one study (18%), AD subjects performed better than VaD subjects. AD patients have been found to perform better on Picture Arrangement tests than do VaD patients.53

Arithmetic function.

No significant differences were noted for the two studies (100%) that tested arithmetic function.34,51

Frontal executive function.

Ten assessments in nine studies investigated executive function. Nine assessments in eight studies (90%) that examined frontal executive function demonstrated superior performance by AD patients when compared with VaD patients. The executive function subtests of the Luria D-test battery57 and the Mattis Dementia Rating Scale53 were completed in a superior manner by AD patients. On the Lezak Tinker Toy Task, AD patients produced significantly more completed designs.41 AD patients were superior on planning (Porteus maze), self-regulation (visual pattern completion), and the Gibson spiral maze.44 AD patients also performed at a higher level on the Graphical Sequence Test than did VaD patients,48,50 VaD patients showing increased perseveration in the former study. AD patients have been shown to develop more categories on the Wisconsin Card Sorting Test49 and demonstrate less perseveration,35,49 in contrast to a single report (10%) of no difference on the Wisconsin Card Sorting Test.46 The above studies are summarized in table 2.

Constructional abilities.

Sixteen assessments in 11 studies tested constructional abilities. Thirteen assessments (81%) in nine studies showed no difference in constructional abilities between AD and VaD groups as evidenced by performance on the following tests: Block Design test in four studies,34,35,40,44 Object Assembly test in two studies,34,40 Clock Drawing in two studies,42,48 and Rey-Osterrieth Complex Figure in three studies.35,44,49 Tests of simple copying, copying with landmarks,43 and the Developmental Test of Motor Integration42 yielded no differences.

Three assessments (19%) in three studies found the AD group superior to the VaD group. Two studies found that clock drawing by AD patients was superior to that of VaD patients,50,56 and Block Design and Object Assembly tests were performed significantly better by AD patients in one study.53

Working memory and concentration.

Fifteen assessments in nine studies investigated working memory and concentration. Ten assessments in six studies (66%) showed no significant difference between AD and VaD groups on tests of attention and tracking, which included Trail-Making34,35,41,42; Symbol Digit Modalities Test35,49; Digit Cancellation, Hidden Cues, Continuous Performance and Sentence Repetition33; Knox Cube and Figure Cancellation41; Paced Auditory Serial Addition Test and Stroop test35; and Letter Pick-Out test.49

In five assessments in three studies (33%), the AD group performed better than the VaD group. In two studies, AD patients performed better than VaD patients on the Continuous Performance Test.36,44 Superior performance for AD patients was also reported on Computerized Reaction Time36 and Trail-Making Tests.36,46

Motor speed.

No significant differences were reported for the single investigation of motor speed in one study, the Grooved Pegboard test.37

Orientation.

Three studies investigated orientation. VaD patients were significantly better orientated than AD patients in one study (33%).44 Two studies51,57 (66%) showed no difference in orientation.

Visual perception.

Five out of five studies (100%) failed to demonstrate any significant difference in visual perception between the dementia groups, including: three-dimensional geometric recognition, embedded figures, recurring figures33; visual perception subtest of the Luria D-test battery55,57; overlapping figures, Hooper Visual Organization Test44; and Judgement of Line Orientation.58

Tactile perception.

The Sensory-Perceptual examination was applied in one study without detecting significant differences.37

Structural–functional correlations.

The following imaging studies discussed comprise a subset of five studies, which included imaging data, selected from the neuropsychological studies included in the review.

In a study38 that performed CT scanning on all patients, rating the scans for ventricular dilatation and the severity of leukoaraiosis (LA),59 it was noted that 97% of the VaD patients displayed LA as opposed to 55.5% of those with AD. Overall, 87% of VaD patients had both infarctions and LA on CT scanning. The degree of ventricular dilatation correlated more strongly with intellectual impairment in VaD than AD patients.

MRI was performed on all patients in a study44 and 16 of the 30 VaD patients had frontal vascular lesions on MRI. These 16 patients were compared with 14 VaD patients with nonfrontal lesions, 48 AD and 48 normal controls.44 Both AD and frontal VaD patients performed better than nonfrontal VaD patients on the Rey Complex Figure Test. It was also noted that frontal VaD patients were more compromised than nonfrontal VaD and AD patients on the Porteus maze and visual pattern completion. MRI and SPECT were performed on a subset of patients of 10 VaD and 20 AD patients in a study.46 The SPECT scans demonstrated significantly lower blood flow in the frontal regions of those with VaD, in particular in the frontal cortices and basal ganglia. Group differences were demonstrated in the form of increased levels of both white matter and infratentorial hyperintensities on MRI in VaD patients. There was a correlation between lesions in frontal subcortical circuits and deficits in planning, set alteration and verbal fluency, increased anosognosia, and pathological crying in VaD.

Investigating the number of infarcts, white matter lucency, ventricular enlargement, and cortical atrophy on CT scans in a subset of 20 VaD patients in a study,47 more severe leukoaraiosis was associated with the production of fewer words on the COWAT and poorer delayed recall on the Memory Assessment Scale (MAS). An increasing number of infarcts was related to poorer discriminability on the MAS, and ventriculomegaly was associated with poorer performance on MAS delayed cued recall. Greater cortical atrophy was associated with poorer performance on all neuropsychological tests except the COWAT.

A further study investigated deep white matter alterations that were measured as a leukoaraiosis score. They found that an increased or more severe level of LA was associated with poorer performance on tests of executive function (WMS mental control) but superior performance on memory tests (CVLT-D). The differences in cognitive function between AD and subcortical VaD patients were hypothesized to be analogous to those of other subcortical dementias.

Reasons for differences in neuropsychological test performance.

To understand the pathogenesis of these differences in the two dementia syndromes, one must examine their neuropathologic basis. Memory is supported by multiple neural systems, with particular involvement of the medial temporal and diencephalic structures.60-62 In AD, the neuropathologic lesions impact directly and early on structures closely associated with memory.63 It has been suggested that the initial episodic memory impairment in AD is due to trans-entorhinal neuropathology.64 This is followed by semantic memory deficits reflecting the spread of pathology to the temporal neocortex.64 In contrast, verbal memory may be relatively spared in VaD due to the heterogeneity of the neuropathology, resulting in a lesser impact upon widely distributed memory systems and the frequent sparing of medial temporal lobe structures.

The relative excess of deficits in faculties ascribed to prefrontal lobe function in VaD may be explained by the frequent presence of lesions in structures that comprise the frontal subcortical circuits.1,65-68 Frontal subcortical circuits have been an area of research interest over the last decade, based on research in primates into the oculomotor and motor circuits.69 These circuits typically consist of neuronal connections from frontal cortex to basal ganglia down to the thalamus, with feedback from the thalamus to the frontal cortex.66 Arguably, dysfunction can occur due to damage to any part of the circuit. The circuits disrupted in VaD include: the dorsolateral prefrontal neuronal circuit mediating executive function, orbito-frontal circuit mediating emotional lability, and the anterior cingulate circuit responsible for motivation and initiation,13,67 thereby explaining the excess frontal executive dysfunction in VaD. Correlated imaging studies lend support to the conceptualization of VaD as being characterized by a greater degree of frontal subcortical dysfunction than in AD of the same severity.38,44,46-48

As VaD is currently conceptualized, the presence of frontal executive dysfunction is not a necessary criterion for the diagnosis. The fact that it is a common and important aspect of VaD, and may help distinguish it from AD, raises the question whether greater importance should be assigned to this feature. Subcortical pathology due to vascular factors leading to a dementia syndrome has been recognized for over a century,70 but whether all VaD patients have significant subcortical pathology has not been specifically examined. Moreover, the concept of white matter pathology is now extended to frontal subcortical lesions that include the grey matter as well as the connections in these circuits.

Language as a possible differentiating function.

The majority of studies suggested that there was no evidence for differentiation of VaD and AD based on tests of verbal fluency and naming. However, in five recent studies, the COWAT, a test of verbal fluency, was performed better by AD than VaD patients. Because the COWAT is influenced by frontal lesions and mental inflexibility,71 this finding may again reflect the more severe executive dysfunction in VaD. Therefore, the published studies do not suggest a differential language dysfunction in the two disorders. One possible explanation for the discrepancy is that most investigators of VaD exclude patients with dysphasia because of the limitations it imposes on neuropsychological assessments. The result is a sample biased toward less language impairment. This methodologic limitation should receive greater attention in future studies so that the contribution of language impairment on dysfunction in dementia can be adequately assessed.

Limitations of the review.

The first limitation of this review stems from the considerable heterogeneity in the concept of VaD that is prevalent in the extant literature. Dementia may arise from a number of different vascular lesions, each of which may be associated with a particular neuropsychological profile.68 McPherson and Cummings68 have proposed that frontal executive dysfunction may be more prominent in small vessel disease, especially lacunar states. In contrast, they note that in single strategic infarctions such as the angular gyrus syndrome (comprising mainly fluent aphasia, alexia, agraphia, and Gerstmann’s syndrome), the picture may closely resemble AD.68 Conversely, caudate, pallidal, and thalamic strategic infarctions would be expected to disrupt frontal executive function. It is possible that the failure to demonstrate differences in many cognitive functions between AD and VaD arises from the heterogeneity within VaD, thereby obscuring group differences. It is also possible that different diagnostic criteria capture different thresholds or even subtypes of VaD.72 Additionally, there is evidence suggesting that pure VaD is a rarer phenomenon than first conceptualized and recent neuropathologic findings suggest that mixed AD–VaD is quite common.73 Therefore, the cases of VaD identified in the original articles reviewed may result from mixed AD–VaD pathology and thus prove difficult to distinguish from AD on neuropsychological testing.

Secondly, there is the problem of circularity bias in diagnosis as the patients in the studies reviewed received clinical as opposed to neuropathologic diagnoses of VaD. The diagnostic biases of the original authors of the articles will therefore impact upon the neuropsychological findings for subjects diagnosed in each group, resulting in inclusion bias. For example, if the original authors allocated more severely memory impaired patients a diagnosis of AD, then neuropsychological testing would reflect the same characteristic. Studies of large numbers of patients with dementia from vascular causes recruited from community samples may be necessary to address these issues. Furthermore, the use of clinical samples, especially when patients are drawn from memory clinics largely catering to AD patients, may bias the sample to a subgroup of VaD that closely resembles AD in its clinical presentation.10

Thirdly, the diversity of diagnostic criteria for VaD used in the studies under review cautions against uncritical acceptance of the findings. The relatively poor validity of diagnosis and the low agreement between diagnostic systems for VaD is now well recognized. There is an up to 10-fold variance in rates of classification of dementia when different diagnostic criteria for VaD were compared on the same community population.72 Internationally acceptable criteria for VaD have only recently been published, and future studies are likely to demonstrate greater consistency. Additionally, the diagnosis of AD involves some possible misclassification and mixed AD–VaD cases may occur.

Fourth, the studies reviewed had many methodologic shortcomings. We had to exclude 18 of the 45 studies from this analysis because they did not meet our strict inclusion criteria. Matching of cases is relevant in that to demonstrate the differences in cognitive function between AD and VaD subjects, factors known to confound neuropsychological assessment must be minimized. It is well known that age, sex, and education are codeterminants of neuropsychological performance.71 It is therefore surprising that some studies failed to match subjects according to these factors. Additionally, in order that the comparison is valid, the severity of the dementia must be controlled for, or the more severely impaired group will necessarily perform more poorly. Very few studies appeared to match adequately for the severity of the dementia. The utilization of mental status scales for the matching of the severity of the dementia may have biased findings by more readily detecting some types of dementia. A more sound approach would be to use indices of function and thus not be confounded by cognitive characteristics of the dementia. Even the studies that were included had many deficiencies. Because the neuropsychological tests used were diverse, our focus was necessarily on cognitive functions rather than scores on particular tests.

Despite these limitations, we consider the literature consistent on some aspects of neuropsychological function in what is generally regarded as VaD, and it provides us with a map for future neuropsychological assessments of this syndrome.

Comparison with conclusions reached by other authors.

The review by Almkvist13 did not specify inclusion criteria and included many studies that were methodologically unacceptable by our criteria. He also concluded that there was a slight excess of deficits in VaD compared to AD in executive functions, verbal fluency, attention, and motor functions.13 Additionally, relative advantage for VaD compared to AD in naming and intrusion errors was noted in our review. He concluded that frontal subcortical dysfunction may be more pronounced in VaD than AD. McPherson and Cummings68 reviewed the neuropsychological aspects of VaD. They observed that most studies focusing on the cognitive deficits in VaD had combined all patients with cerebrovascular pathology into a single group, and suggested that the “patchy” deficits observed may be an artifact of studying patients with varying neuropathologic changes. Multi-infarct, strategic infarct and small vessel disease were described as three major subgroupings of VaD. The authors noted VaD may result from a summation of pathologic processes. They described executive dysfunction as prominent in thalamic, lacunar, and Binswanger subtypes of VaD. In contrast, they concluded that memory differences between AD and VaD were not significant. The authors also speculated on the possible etiologic role of frontal subcortical dysfunction in VaD.68

Reviews of the VaD construct have characterized the neuropsychological features of VaD in comparison to AD as follows1,5,74: greater impairment of executive functions; greater attentional deficits; slower mental processing; lesser impairment of memory; lower frequency of intrusion errors; and lower level of language dysfunction. The conclusions regarding attention and slower mental processing were not supported in our review. Other reviewers75 quoted a number of studies as inconclusive regarding the neuropsychological profile of VaD, concluding that no specific pattern of frontal deficit had been consistently demonstrated in VaD.

General textbook reviews of VaD have emphasized a “patchy” pattern of cognitive deficits comprising slowing of cognitive processing, visuospatial abnormalities, and impairment of executive function.76,77 Findings of cognitive slowing and visuospatial abnormalities were not supported by this review. The methodologic difficulties in comparing AD to VaD, in particular the failure to match for the severity of dementia, have been noted, as has the observation that frontal subcortical dysfunction is the most common feature in VaD.10

The findings of this review concur with an emphasis on executive and memory dysfunction in discrimination between AD and VaD.

Clinical and research recommendations.

Even while the concept of VaD is being refined, and amidst calls for its abandonment,4 the clinician is frequently required to distinguish VaD from AD. Our review suggests that the neuropsychological profile, although not prototypical for either disorder, does differ significantly between the two dementias to be useful in the clinical assessment. In addition to the positive discriminators, it is useful to recognize that the following do not permit the differentiation: memory registration (digit span); conceptual functions (arithmetic, similarities, and Raven’s Coloured Progressive Matrices); attention and tracking (trail-making, symbol digit modality); and construction (block design, object assembly, and the Rey-Osterrieth Complex Figure). The clinician must also apply standard criteria for diagnosis. The review suggests that the emphasis on memory impairment in the definition may not be appropriate. The clinician should be sensitive to frontal executive dysfunction, which can sometimes be overlooked if one has a restricted concept of dementia.

Our review suggests that future research into dementia from vascular causes should not be restricted to patients who meet extant criteria. The use of broad criteria, and resorting to community samples as well as individuals who are at high risk, may help encompass the range of vascular dementias and help determine if there are indeed any core features. A hypothesis that could be examined is that frontal executive dysfunction is the core disturbance in VaD and that variability is added by lesions in other brain regions. Any investigation of the neuropsychology of VaD should be informed by contemporaneous neuroimaging data. Both structural and functional neuroimaging are necessary to help understand the mechanisms involved. The limitations of the application of a predetermined concept of dementia to what is dementia of vascular etiology are quite apparent.