Correlates of neuropsychiatric and motor tests with language assessment in patients with Lewy body dementia

Background: Lewy body dementia (LBD) impairs performance in daily activities and affects motor, language and visuospatial tasks. Objective: We aimed to correlate neuropsychiatric and motor assessments with language and visual organization tests in LBD. Methods: Twenty-two patients with dementia with Lewy bodies and ten patients with Parkinson’s disease dementia participated on a cross-sectional study that assessed cognition, functionality, caregiver burden, verbal fluency, the primer-level dictation section of the Boston Diagnostic Aphasia Examination (PLD-BDAE), the Hooper Visual Organization Test, the Neuropsychiatric Inventory and the Movement Disorder Society – Unified Parkinson’s Disease Rating Scale. Results: Language and visuospatial test results followed motor impairment and general cognitive performance. Whereas visual organization did not predict performance in the PLD-BDAE, visuospatial abilities and verbal fluency were concurrently associated, suggesting that linguistic impairment in LBD may be attributed to neuropsychological components of cognition and language. Only visual organization was associated with behaviour, suggesting that neuropsychiatric symptoms associate with differential impairment of visual organization in comparison with language in LBD. Schooling did not affect visual organization or language test performance, while the length of dementia was negatively associated with visual organization and verbal fluency. Discussion: Though visual organization tests follow behaviour and motor performance in LBD, there is differential impairment regarding language skills.


Introduction
Deficits in visuospatial abilities, memory, executive functions and language are the most evident neuropsychological symptoms in patients with Lewy body dementia (LBD) syndromes, corresponding to frontal-striatal dopaminergic dysmodulation associated with diffuse cholinergic cortical dysfunction 1 . Essentially, the spectrum of these syndromes consists of dementia with Lewy bodies and Parkinson's disease dementia 2 , comprising the second leading cause of degenerative dementia in older people after Alzheimer's dementia (AD) 3,4 . Nevertheless, sensitivity of clinical diagnosis of LBD is not always good, particularly in severe dementia, although specificity tends to be high 5 .
The main etiological hypothesis for LBD entails the histopathological presence of Lewy bodies in the brainstem, subcortical nuclei, limbic cortex (cingulate cortex and amygdala), and in the neocortex 6,7 . In Parkinson's disease dementia, Lewy body formation and neuron loss usually start in brainstem nuclei and in the substantia nigra, whereas in dementia with Lewy bodies they occur in paralimbic and neocortical structures from disease onset 8 . In addition, amyloid pathology helps predict the onset of dementia in parkinsonian syndromes 9 . Nevertheless, cholinergic denervation is the main source of linguistic impairments in patients with LBD, particularly when involving cortical or subcortical language networks 10 .
Approximately 80% of all patients with Parkinson's disease develop dementia, more frequently when they are male and have more severe motor signs at examination 11 . The burden of motor and neuropsychiatric manifestations of LBD considerably affects functional independence and social activities, impacting quality of life 2,8 . Visuospatial skills and behavioural symptoms are helpful for differential diagnosis between LBD and AD, whereas cerebrovascular risk might be more important for pathogenesis of AD 8,12 , but cholinesterase inhibitors are usually valuable for treatment of neuropsychiatric symptoms of both AD and LBD 13,14 .
Despite the fact that some associations are well established for dementia syndromes, such as functional decline following cognitive decline in severe dementia 15 , impairment of language has not been deeply studied in LBD. We hypothesized that language domains could be primarily affected in LBD, whereas motor signs, behavioural symptoms, linguistic and cognitive features would be concurrently impaired; therefore, we aimed to analyse associations of neuropsychiatric and motor assessments with language and visual organization test results in patients with LBD.

Participants
In this cross-sectional study, consecutive outpatients with LBD in different levels of clinical evolution were recruited from the Department of Neurology and Neurosurgery at Hospital São Paulo, Federal University of São Paulo (Unifesp). All patients with LBD who were followed at the outpatient clinic were assessed from January 2014 to April 2015. Diagnosis of Parkinson's disease followed traditional clinical criteria 16 . Patients had to be diagnosed with either probable or possible Parkinson's disease dementia according to Movement Disorder Society Task Force clinical diagnostic criteria 17 , or either probable or  possible dementia with Lewy bodies 18 . Basically, Parkinson's disease  dementia developed within the context of established Parkinson's  disease, requiring a combination of typical cognitive and behavioural  features for diagnosis, while dementia with Lewy bodies preceded  motor manifestations by at least one year, with a combination of core features (fluctuating cognition with varied attention or alertness, recurrent well-formed visual hallucinations, or spontaneous features of parkinsonism) and suggestive features (REM sleep behaviour disorder, severe neuroleptic sensitivity, low dopamine transporter uptake in basal nuclei). None of the patients had neuroimaging evidence of focal cerebrovascular diseases or any other structural brain diseases that could account for the cognitive or language deficits.

Clinical assessment
After diagnostic confirmation, patients and caregivers were assessed for: patient age, gender, schooling, estimated age at dementia onset, sleep satisfaction and estimated daily length of sleep 19 28 .
In a second evaluation, patients with LBD were also assessed with the Hooper Visual Organization Test (HVOT) 29 , verbal fluency (VF) 30 , and the primer-level dictation section of the Boston Diagnostic Aphasia Examination (PLD-BDAE) 31,32 , including primer words, regular phonics, and common irregular words. All cognitive assessments were conducted on weekdays at morning time, by two examiners (FCM and FFO).
The Schwab & England scale 24 was employed for overall performance in activities of daily living. A trichotomous version (1 = unable; 2 = able with help; 3 = able without help) of Lawton's Scale for Instrumental Activities of Daily Living 25 was employed, with scores for using the telephone, getting to places beyond walking distance, grocery shopping, meal preparation, housekeeping, doing handyman work, doing laundry, taking own medications, and handling finances; caregivers provided all information, with a total score of 9 to 27.
For the HVOT, each participant was presented with 30 figures of fragmented objects in ascending order of difficulty 29 . For the VF tasks, the patient should generate the largest possible number of words in one minute for each category, including words beginning with F, A and S (F-A-S), as well as all animals, fruits and grocery items that might be known 33 . In the PLD-BDAE, the patient writes regular and irregular words to dictation 31 .

Statistical analyses
Fisher's exact test and the Mann-Whitney test were used for comparisons of neurological features between dementia syndromes. Simple linear regressions were employed for comparisons between test results. A multiple linear regression model was employed for associations between each visual organization or language test (HVOT, VF, and the PLD-BDAE) and the following independent variables: schooling and length of the dementia syndrome; p-values were corrected with the Bonferroni test. The threshold of significance was set at p < 0.05.

Ethical aspects
This study is part of the research project 064990/2013 approved by the Ethics Committee of Hospital São Paulo, Unifesp, in October 2013. All invited patients and their legal representatives agreed to participate on the research and signed the Informed Consent Form before the evaluation.

Results
Overall, 39 participants were recruited; between the first and the second assessments, three patients passed away (7.7%), and four patients did not complete the second evaluation (10.3%), resulting in a final sample of 32 patients -19 women (59.4%) and 13 men (40.6%). Twenty-one patients were diagnosed with probable dementia with Lewy bodies, one patient was diagnosed with possible dementia with Lewy bodies, nine patients were diagnosed with probable Parkinson's disease dementia, and one patient was diagnosed with possible Parkinson's disease dementia. Nineteen patients with dementia with Lewy bodies (86.4%) had visual hallucinations, versus six patients with Parkinson's disease dementia (60.0%), p = 0.165. Sixteen patients with dementia with Lewy bodies (72.7%) had parkinsonism, versus ten patients with Parkinson's disease dementia (100.0%), p = 0.142. Moreover, fifteen patients with dementia with Lewy bodies had fluctuations (68.2%). Demographic data and test results for all patients with LBD are summarized in Table 1.
There was no statistically significant difference between patients with dementia with Lewy bodies and patients with Parkinson's disease  Table 2 summarizes the results from simple linear regressions between visual organization and language test results for all patients with LBD. Visual organization was associated with category VF, whereas all language tests were correlated with each other, except for the association between VF (animals) and the PLD-BDAE. Tables 3 and 4 list results from simple linear regressions regarding visual organization and language tests for predictions of associations with other neuropsychiatric features. Visual organization was associated with basic (but not instrumental) functionality, general cognitive tests, motor examination and the Neuropsychiatric Inventory total scores. All categories of VF were associated with general cognitive tests, and inversely associated with global dementia rating; however, only VF for F-A-S and for fruits was associated with the clock drawing test, only VF for F-A-S and for grocery items was associated with non-motor experiences of daily living, and only VF for fruits was associated with motor experiences of daily living and motor examination. All categories of VF were associated with forward digit span, except for fruits, the only category associated with reverse digit span. Sleep satisfaction was negatively associated with all categories of VF, except for fruits. The PLD-BDAE was associated with basic (but not instrumental) functionality, general cognitive tests, motor experiences of daily living and motor examination. Length of sleep, instrumental functionality, and caregiver distress regarding behavioural symptoms had no significant associations with visual organization or language tests. Table 5 lists results from multiple linear regressions involving language and visual organization tests. Schooling did not affect performance in any test, while the length of the dementia syndrome was negatively associated with performance in the HVOT and VF (animals).

Discussion
In this study, associations among neuropsychiatric features of patients with LBD could be more accurately evaluated by specific tests. Knowledge of less studied clinical features, such as language disorders in LBD, can be useful to promote deinstitutionalized care and caregiver education.
Visual organization and PLD-BDAE test results were not significantly correlated, suggesting that the HVOT does not predict performance in the PLD-BDAE; in other words, this finding confirms that impairment of language may be a primary feature of LBD, and not necessarily secondary to cognitive deficits 34 . Nonetheless, VF and visuospatial abilities were concurrently associated in LBD. Visual organization is related to the frontal-subcortical circuitry that is  affected early in the course of LBD 35 . Moreover, poor VF is associated with incident dementia in Parkinson's disease, and could be due to impaired self-generated search 13 . Nevertheless, other studies have found that naming tests can be the best predictor of performance in the HVOT 35 . Patients with dementia with Lewy bodies have disproportionate deficits in visuospatial skills, attention and letter fluency 36 . Visuospatial processing, attention and executive functions in Parkinson's disease dementia have also been described to be similar to dementia with Lewy bodies 37 . Impaired connectivity with the frontal cortex leads to severely impaired grammatical expression in patients with LBD: failing to complete sentences, omitting the verb phrase, perseveration, requiring additional time to plan sentences 34 . We observed similar errors in the narrative discourse of our patients, but it should be noted that cortical involvement may occur earlier in dementia with Lewy bodies compared to Parkinson's disease dementia 38 .
Cognitive and functional tests are usually correlated with one another in all stages of AD 23 . In our study with patients with LBD, we found that the higher the Clinical Dementia Rating sum-of-boxes scores, the lower were language and visuospatial test results. Likewise, language and visuospatial test results followed Mini-Mental State Examination scores. This could be helpful for differential diagnoses, since visuospatial abilities are more impaired in LBD than in other dementia syndromes 39 . In comparison with AD, patients with LBD have better contextual verbal delayed recall and recognition, and less short-term memory deficits, but worse letter fluency deficits and qualitative measures of executive functioning, and worsening visuoperception following overall cognitive decline 37,40 .
The Hoehn and Yahr stages grade severity of parkinsonism 28 , while the clock drawing test is a measure of visuospatial dysfunction also useful for screening cognitive impairment 23 . When performances in the HVOT and in the PLD-BDAE were worse, patients also had more severe parkinsonism and lower scores in the clock drawing test. It has been shown that patients with LBD who lose more motor function also have the greatest visuospatial impairment 13 .
Only VF had a negative association with sleep satisfaction, but not with length of sleep, suggesting that sleep satisfaction may be inversely correlated with the stage of LBD. Sleep disorders occur in three quarters of autopsy-confirmed cases of dementia with Lewy bodies, but may not necessarily correlate with sleep satisfaction; in addition to attentional, executive functioning, and visuospatial impairments, the presence of impaired verbal learning helps identify prodromal dementia with Lewy bodies in patients with sleep disorders 41 . Moreover, it has been shown that the severity of psychotic symptoms in patients with Parkinson's disease is directly associated with the severity of cognitive impairment and sleep disturbances 42 .
In our analyses, only the HVOT was associated with total scores of the Neuropsychiatric Inventory. Behavioural symptoms may affect sustained attention and, therefore, cognitive functioning 43 . Despite the increased frequency of visual hallucinations in LBD when compared to other dementia syndromes, they also lead to worse prognosis 8 . Still regarding neuropsychiatric symptoms, caregiver distress had no significant associations with visual organization or language tests, possibly representing low sensitivity to score variations in these tests.
All categories of VF were associated with forward digit span, except for fruits, the only category associated with reverse digit span. These findings suggest that attention and executive functions are important for most forms of category VF, but working memory might not decline concurrently.
Visual organization and language performance were not affected by education. This could be due to the cross-sectional nature of our study, but also to the fact that mechanisms of neurodegeneration supersede protective factors in these patients.
The length of the dementia syndrome was negatively associated with visual organization and VF for animals, an important finding to be correlated with the rapid cognitive decline usually found in patients with LBD 12 . On the other hand, instrumental functionality had no significant associations with visual organization or language tests, possibly due to the fact that instrumental functional decline occurs earlier, while visuospatial and language decline happen throughout the course of LBD.
The most important limitations of our study comprise its small sample size, its cross-sectional nature, and the fact that all patients were recruited from a single centre, thus limiting generalizability. Also, the size of our sample did not allow stratification into patient groups according to diagnoses (dementia with Lewy bodies or Parkinson's disease dementia), but pathophysiology is similar for these two diseases 16,37 , and our results were mostly unaffected by this choice. Furthermore, the wide age range of the patients (spanning 45 years) could have affected our results due to the fact that young and older adults use different strategies to accommodate to impairments in executive function 44 but, considering that all patients were over 50 years-old and had at least one year of dementia diagnosis, we believe this to be unlikely.
We conclude that language and visual organization tend to follow motor skills and general cognitive performance in patients with LBD. Whereas visual organization did not predict performance in the PLD-BDAE, visuospatial abilities and VF were concurrently associated, suggesting that features of linguistic impairment in LBD may be attributed to components of cognition and language. Moreover, only visual organization was associated with behavioural performance, suggesting that neuropsychiatric symptoms are differentially associated with visual organization in comparison with linguistic features in LBD. Future studies should address neuropsychiatric correlations in prospective assessments.