Insomnia and autonomic dysfunction may identify patients at risk for developing early Parkinson's disease dementia

Background: The prospective study aims at identifying features predictive of early onset of dementia in patients with Parkinson's disease (PD). Methods: 89 non-demented PD patients underwent a complex evaluation (demographic data, UPDRS, Unified Multiple System Atrophy Rating Scale - UMSAR, Insomnia Severity Index - ISI, Neuro-Psychiatric Inventory - NPI, Hamilton Depression Rating Scale - HDRS, Mini Mental State Examination) at baseline and at 3-year follow-up. Results: At 3-year follow-up 43.8% of patients developed dementia. An ordinal regression of MMSE at follow-up showed that dementia developed in patients with autonomic dysfunctions (odds ratio 16.18, 95% CI 3.16 to 82.77, p = 0.001), old age (odds ratio 1.24, 95% CI 1.11 to 1.39, p < 0.001), and insomnia (odds ratio 1.23, 95% CI 1.09 to 1.38, p = 0.001). Conclusion: Patients with signs of autonomic dysfunction and insomnia are at higher risk for developing dementia and deserve closer monitoring of cognitive symptoms.

3 Advanced age is the most commonly established risk factor for dementia. Non-motor features such as REM sleep behavior disorder (RBD), baseline cognitive impairment, olfactory disturbances, and autonomic dysfunction have been identified as possible dementia predictors [3][4][5][6]. However, in many studies the tests were designed for clinical trial conditions, tests which are more difficult to apply as screening methods in real-life conditions and add to the costs of health care.
In our study, we assessed clinically, under conditions applicable in an outpatient care setting, a number of potential clinical risk factors for dementia in a 3-year prospective cohort study.

Methods
The present prospective study was conducted in accordance with the Declaration of Helsinki (1964) after approval of the Ethical Committee of the hospital.

Patient selection
Patients were recruited from the Neurology ward and outpatient clinic of our hospital. The period of enrollment extended from November 2012 to June 2014. A total of 130 patients were asked to undergo the neurological and psychological evaluation described below.
Exclusion criteria were: · Secondary parkinsonisms or Pakinson-plus syndromes · Dementia diagnosed at screening · Significant comorbidities (e.g., cancer) which could impede the patient to complete the follow-up period Parkinson's disease was diagnosed in accordance with the UK Brain Bank criteria [7].
Dementia was diagnosed by a certified psychiatrist according to the DSM-IV criteria taking into account also the particularities of Parkinson's disease dementia (impairment of at least two of the following domains: attention, executive function, visuo-constructive 4 ability, memory) [8].
On the basis of the presence of dementia at enrollment 18 patients were excluded, and 9 denied enrollment for various reasons. Written consent to participate was obtained from all the 103 remaining patients, which were allowed to withdraw at any time. A number of 14 patients were lost to follow-up. Therefore we analyzed the data from 89 patients who completed the 3-year study period.

Patient evaluation
Patients underwent a complete neurological evaluation performed by a certified neurologist and a psychological evaluation at enrollment and after 3 years (36 +/-3 months) performed by a psychologist certified in clinical psychology. Meanwhile they had regular follow-up visits, dose adjustments of antiparkinsonian medication, and psychological and/or psychiatric evaluation as needed. Information on demographics, disease duration, medication was recorded. PD evaluation included:  [14]. Screening questions for each of the 10 neuropsychiatric symptoms were posed first and if a positive response was obtained the symptom was further explored. Ratings from 1 through 4 were obtained for frequency and from 1 through 3 for severity of each symptom. A composite score for each symptom was calculated as the product of the frequency and severity (from 0 through 12) and was used for analysis. 5. Mood. The Hamilton Rating Scale for Depression, a 17 item questionnaire measuring psychological and physiological symptoms of depression [15] was administered to each patient at baseline, at 3-year follow-up visit, and between these 2 visits as needed.

Statistical analysis
Statistical analysis was performed using IBM-SPSS version 19.0. Analyses consisted of description of the various scores, and calculation of proportions. Differences between groups were tested for statistical significance with the student's t test. Factors associated with dementia were identified with a partial correlation analysis controlling for age and disease duration, after which they were included into an ordinal regression analysis. 95% 6 confidence intervals and odds-ratios were determined. Statistical significance was set at p < 0.05.

Results
The demographic and clinical description of the 89 patients included in the study is shown in table I. The gender distribution was 59.6 % male and 40.4 % female patients. In order to see which clinical characteristics may predict a rapid evolution towards cognitive decline, we ran a partial correlation analysis between the dose of levodopa at baseline, the UPDRS score, the presence of autonomic dysfunctions (as indicated by 1 or more symptoms of autonomic dysfunction scored on the UMSAR scale), the Insomnia Severity Index, the score on the Hamilton Depression Rating Scale, the score of the Neuropsychiatric Inventory, and the MMSE at baseline (at the moment of inclusion into this study). Since with age and duration of disease the likelihood of developing dementia increases, we controlled for these 2 factors.
In patients who went on to developing dementia, the MMSE score at 3-year follow-up was significantly influenced only by the Insomnia Severity index (Pearson coefficient -0.604, p <0.001) and the presence of autonomic symptoms and signs (Pearson coefficient -0.640, p < 0.001). Neither dose of levodopa, UPDRS score at baseline, nor Hamilton Depression Rating Scale score could predict which patients would develop dementia, while the correlation with the MMSE score at baseline failed to reach statistical significance (p = 0.065). However, patients more depressed were more likely to be on a higher dose of levodopa (Pearson coefficient 0.504, p = 0.001).
In patients who did not develop dementia the MMSE score at 3-year follow-up was

Discussion
Of our 89 patients included in the study, 39 developed dementia in the 3 year follow-up period, at higher risk being those patients who presented signs of autonomic dysfunction and insomnia.
Almost half of the patients complained about their sleep patterns, while depression 8 affected 37% of patients at inclusion in the study. After 3 years, sleep was disturbed in more than two third of patients, and depression rates exceeded 50%. The figures provided for the depression rates by other studies resemble our findings [17,18].
Focusing on the development of dementia, our study shows that a 3-year follow up was sufficient to show a significant decline in cognitive functions. It is true that the baseline MMSE was significantly lower in patients who were diagnosed with dementia after 3 years as compared with patients who were not demented after 36 months (21.39 vs. 26.38; p < 0.001). It may be that a simple MMSE test has too low specificity for evaluating cognition in Parkinson's disease since it focuses mainly on cortical functions while Parkinson's disease dementia is a subcortical type of dementia. It is possible that a part of these patients might have been diagnosed with dementia upon a more detailed testing. In addition, demented patients at follow up were older (69.67 years vs. 63.51 years; p < 0.001) and had a longer duration of the disease (7.62 vs. 5.59 years; p < 0.001) when included in the study. However, even after controlling for these variables, the partial correlation analysis shows that the cognitive functions of patients diagnosed with dementia at follow-up were influenced by the UMSAR score and the Insomnia Severity Index at baseline.
The UMSAR score reflects signs of autonomic dysfunction. Previous cross-sectional studies have found strong correlations between cognition and orthostatic hypotension [19]. The cause for this strong relationship is unclear. It is possible that autonomic dysfunction is a marker of a more "diffuse" disease subtype, in which autonomic areas and cortical regions involved in dementia degenerate as well [20]. However, an alternative mechanism may also be discussed. In the general population low diastolic blood pressure has been linked to brain atrophy, particularly when associated with reduced cerebral blood flow [21].
Although cerebrovascular autoregulation mechanisms prevent orthostatic changes from reducing cerebral blood flow, moderate changes in PD which may associate multisystem and cardiovascular dysautonomia might cause significant cerebral hypoperfusion. Hypoxia and hypotension have been shown in animal models and human studies to cause neuronal damage and increases in accumulation of beta-amyloid and other pathologic proteins [22,23]. Therefore, one may reasonably hypothesize that repeated episodes of hypotension in PD over the years can augment damage to cortical neurons already made vulnerable by the synucleopathy. The findings of a prospective observational study showing that treatment of orthostatic hypotension improved cognition in dementia supports this hypothesis. [24].
Another predictor of dementia was the presence of insomnia. The sleep rhythm is driven by a central circadian pacemaker located in the suprachiasmatic nucleus of the anterior hypothalamus [25]. Studies of sleep deprivation and sleep restriction in healthy subjects revealed the importance of sleep in memory function and mental health. Experimentally induced sleep loss in healthy subjects is associated with impairments in a broad range of cognitive functions, and cognitive impairments have been identified in people with sleep disorders such as insomnia and sleep apnea. Sleep appears to be necessary for both memory encoding and consolidation. Sleep deprivation also affects emotional memory, with preservation of mainly negative compared to positive and neutral memories [25].
Disruptions in the sleep-wake cycle and sleep complaints are common complaints found in community-based studies of older people. Over 50% of adults aged over 65 have at least one chronic sleep-related problem [26]. Advancing age with the associated decline in the melatonin and cortisol rhythms which entrain day-night activity patterns contribute to the changes in circadian rhythms which have been demonstrated in the elderly [27]. However, the link between sleep disturbances and Parkinson's disease is stronger. It has been estimated that a patient presenting typical Rapid Eye Movement sleep behavior disorder (REMBD) will have a 50% chance of developing a parkinsonian syndrome within 5 years [28].
We did not find a significant correlation between the Hamilton Depression Rating Scale score and MMSE score, neither in patients converted to dementia nor in patients who did not develop dementia, although in the latter group of patients the link was stronger.
Depression did not correlate with cognitive variables although it was more often present and more severe in patients with cognitive impairment. Previous studies have shown that mood changes in Parkinson's disease rather reflect disordered emotional functions [25]. It has been suggested that patients with no depression and no dementia have a normally functioning dopaminergic mesocortico-limbic system, while the opposite were true for PD patients with depression and/or cognitive impairment [25]. Supporting this hypothesis is the finding of a lower metabolic activity in the head of the caudate nucleus and in the inferior frontal cortex in the depressed PD patients as well as that of a significant inverse correlation between glucose metabolism in the inferior frontal cortex and depression scores [29].
The UPDRS score at baseline showed no significant correlation with the development of dementia or depression, highlighting the emphasis put by this rating scale on motor symptoms. Of a total of 179 possible points, 108 may be achieved from the motor items of the UPDRS scale.
An interesting finding is the link between dose of levodopa and decline of the MMSE at 3 year follow up. Levodopa is a precursor of dopamine. After crossing the blood-brain barrier it is decarboxylated to dopamine by aromatic amino acid decarboxylase present in neurons and glia, thereby contributing to its clinical efficacy in the treatment of Parkinson's disease. However, the presence of dopamine in the brain may be damaging as well. Un-sequestered dopamine is enzymatically degraded by monoamine oxidase B, a reaction during which hydrogen peroxide results as a by-product. Hydrogen peroxide is a powerful oxidant. In addition, auto-oxidation of dopamine results in highly reactive quinones which bind to proteins and interfere with the normal function of the latter.
Increased oxidative stress also causes mitochondrial dysfunction and DNA damage, adding to the levodopa induced neurotoxicity [30]. A relatively recent study concluded that by a dysregulation of brain 5-hydroxytryptamine chronic levodopa treatment may contribute to non-motor symptoms related to spatial memory and fear [31]. The results presented provide further reasons for investigating adjunctive therapies to L-dopa for PD, such as antioxidants, as well as to delay treatment with levodopa for as long as possible, especially in young patients, although it is known that only mildly impaired PD patients may obtain symptom relief with anticholinergics or dopamine agonists [17].
We are aware of the limitations of our study, namely: the lack of recordings of family history of psychiatric disturbances, of personal history of head trauma, of nonparkinsonian medication use, the lack of detailed cognitive testing both at baseline and at follow up, using a complex battery of tests and use only of the MMSE and psychiatric evaluation, assessment of the presence of sleep disturbances only through the ISI, and a relatively small sample size, observed for a rather short period of time (36 months).
However, we desired to remain in the frame of an evaluation applicable as part of a usual outpatient visit. Those patients who have symptoms and/or signs which may indicate an early onset of dementia should undergo a more detailed testing and have visits scheduled more often.

Conclusions
In our opinion, our findings show that: -Cognitive evaluation in initial and follow up assessment of Parkinson's disease patients may select those at risk for developing dementia raising the possibility of targeting 12 coming neuroprotective agents at patients with this condition.
-Identifying and analyzing non-motor symptoms, especially symptoms of autonomic dysfunction and sleep disturbances, may identify patients at risk for PDD and enable correction of these factors as well as more frequent assessments to diagnose dementia at early stages.
It would be interesting to investigate whether correction of these dysfunctions would delay the development of dementia.

Declarations
Ethics approval and consent to participate All patients gave written informed consent, and the ethical committee of the Clinical Municipal Hospital Oradea, Romania approved the study. Since patients with dementia at screening were not included in the study we did not ask for a family member or legal guardian's consent.

Not applicable
Availability of data and materials The datasets used during the current study are available from the corresponding author on reasonable request.