Blood pressure and heart rate responses to orthostatic challenge and Valsalva manoeuvre in mild cognitive impairment with Lewy bodies

Abstract Objectives Orthostatic hypotension is a common feature of normal ageing, and age‐related neurodegenerative diseases, in particular the synucleinopathies including dementia with Lewy bodies. Orthostatic hypotension and other abnormal cardiovascular responses may be early markers of Lewy body disease. We aimed to assess whether abnormal blood pressure and heart rate responses to orthostatic challenge and Valsalva manoeuvre would be more common in mild cognitive impairment with Lewy bodies (MCI‐LB) than MCI due to Alzheimer's disease (MCI‐AD). Methods MCI patients (n = 89) underwent longitudinal clinical assessment with differential classification of probable MCI‐LB, possible MCI‐LB, or MCI‐AD, with objective autonomic function testing at baseline. Blood pressure and heart rate responses to active stand and Valsalva manoeuvre were calculated from beat‐to‐beat cardiovascular data, with abnormalities defined by current criteria, and age‐adjusted group differences estimated with logistic models. Results Orthostatic hypotension and abnormal heart rate response to orthostatic challenge were not more common in probable MCI‐LB than MCI‐AD. Heart rate abnormalities were likewise not more common in response to Valsalva manoeuvre in probable MCI‐LB. An abnormal blood pressure response to Valsalva (delayed return to baseline/absence of overshoot after release of strain) was more common in probable MCI‐LB than MCI‐AD. In secondary analyses, magnitude of blood pressure drop after active stand and 10‐s after release of Valsalva strain were weakly correlated with cardiac sympathetic denervation. Conclusions Probable MCI‐LB may feature abnormal blood pressure response to Valsalva, but orthostatic hypotension is not a clear distinguishing feature from MCI‐AD.


| INTRODUCTION
Autonomic symptoms are common features of the synucleinopathies, including Parkinson's disease and dementia with Lewy bodies (DLB), and may include changes in heart rate (HR), blood pressure, temperature regulation, skin response, urination and digestion. 1 Orthostatic hypotension (OH) in particular may be more common in DLB compared to other synucleinopathies. 2 OH is characterised by a significant drop in blood pressure (BP) when moving from the supine position to standing upright. 3 A reduction in orthostatic blood pressure without a compensatory increase in heart rate indicates a neurogenic cause of the OH. 4 OH may cause symptoms such as light-headedness, loss of balance, fatigue, or nausea, though a symptomatic response to OH is not required for diagnosis. 5 Aside from orthostatic challenge, abnormal BP and HR responses may also be elicited by the Valsalva manoeuvre (VM); as with the orthostatic challenge, abnormal responses may present as an exaggerated change in BP or HR in response to VM, or a slowed return to baseline/lack of overshoot after release of strain. 6 OH and its related symptoms may be of relevance to the screening and diagnosis of DLB, and contribute to the worse typical prognosis of this compared to Alzheimer's disease (AD), being linked to shorter survival time 7 ; shorter survival being a common feature of DLB. 8 However, OH is also reasonably common in AD and the normal ageing population, and so is not specific to DLB 9 : AD, which often cooccurs in DLB, is itself associated with pathological changes to autonomic nuclei of the reticular formation, 10 and so may also contribute to autonomic symptoms through a mixed pathology.
Autonomic symptoms in DLB such as OH may reflect the widespread cholinergic dysfunction common of Lewy body disease, 11 or the presence of Lewy pathology (Lewy bodies and Lewy neurites) in the peripheral nervous system. Degeneration of the sympathetic nervous system is an early feature of DLB demonstrated by abnormalities in cardiac 123I-metaiodobenzylguanidine (MIBG) scintigraphy, 12 a finding also observed in the cognitive prodrome of DLB: mild cognitive impairment (MCI) with Lewy bodies (MCI-LB). 13 In both the dementia and prodromal stages of DLB, sympathetic denervation may be quantified by reduced uptake of MIBG to the heart relative to the mediastinum (heart:mediastinum ratio; HMR).
As a reduced HMR has been demonstrated to often be present in the MCI stage of disease, 13 MCI-LB may also feature early autonomic dysfunction such as OH. However, this has not been examined in a prospective sample of MCI-LB with objective measurement of OH, or cardiovascular response to VM.
We therefore aimed to investigate whether objectively measured cardiovascular dysfunctions would be an early feature of MCI-LB, which might help distinguish this from normal ageing and MCI due to AD (MCI-AD). We hypothesised that probable MCI-LB would be more likely than MCI-AD to feature abnormal blood pressure and heart rate responses to orthostatic challenge and Valsalva manoeuvre.

| Participants
As previously described, 14 patients were recruited from older persons' health, psychiatry and neurology services in North East England with a health service diagnosis of MCI, and reported presence of either any core clinical features of LB disease (parkinsonism, REM sleep behaviour disorder, complex visual hallucinations, and fluctuating attention and cognition), or any other supportive features sensitive to LB disease but also found in AD (e.g. changes in mood, anxiety, autonomic symptoms, or sleep disturbance). All were aged ≥60 years and required to be medically stable. All provided written informed consent to undergo further screening, and were excluded from further assessment if they had dementia, no objective cognitive impairment, or evidence of any frontotemporal or vascular aetiology.

| Assessment and imaging
Participants underwent repeated cognitive and clinical assessment at All participants were similarly offered MIBG cardiac scintigraphy at baseline, as previously described. 13 HMR was quantified from delayed images, and classified as abnormal given a HMR <1.85, with this cut-off value derived from locally recruited cognitively healthy comparators. 16 Of this sample, three did not complete MIBG imaging (declined or were ineligible), and two declined FP-CIT imaging. Imaging results were then incorporated into baseline diagnosis and annual diagnostic reviews.
Structural and functional MRI were conducted at baseline, as previously reported. 17,18 Any significant cerebrovascular disease on MRI was cause for exclusion as a possible vascular cognitive impairment.

| Cognitive impairment diagnosis and differential classification
As previously described, 14

| Procedure
At baseline, all patients were offered autonomic function testing, measured using the CNSystems Task Force ® Monitor. This was administered by a medically qualified researcher (RD, SL) in a single session consisting of a 10-min supine rest to obtain baseline mea-

| Valsalva
To select which Valsalva manoeuvre (VM) to use for analysis, the data was visually inspected and the VM which achieved the greatest degree of hypotension in phase 2 of the manoeuvre was selected.
Participants who were unable to successfully perform the VM were excluded from this analysis, but retained for active stand analyses.

| Statistical analysis
Binary logistic models estimated differences between MCI-AD and probable MCI-LB in presence of each of four outcomes: orthostatic hypotension, abnormal HR response to orthostasis, abnormal BP response to VM, and abnormal HR response to VM. These were adjusted for age (mean centred), and the use of cholinesterase inhibitors, since these may be independently associated with OH. 26 The possible MCI-LB were included for additional context, but were not interpreted in relation to the primary hypothesis. There was no statistical adjustment for multiple comparisons.
Secondary analyses were undertaken with linear models to assess any associations between the magnitude of any BP abnormalities, and either MCI subtype or degree of MIBG abnormality.

| RESULTS
Of 103 participants, 89 provided useable cardiovascular data for at least one of the assessments. Baseline demographic and autonomic characteristics are described in Table 1; data completeness for the orthostatic challenge and Valsalva procedures are also included, as  Due to the gender imbalance across diagnostic groups, we conducted an additional sensitivity analysis to assess whether there was any gender-associated effect. This did not meaningfully change any of the observed findings, and there was no main effect of gender itself on any autonomic outcome.
A secondary analysis explored the associations between MIBG HMR, a measure of sympathetic denervation, and cardiovascular responses to orthostasis and VM.
There was a marginally significant association (see Figure 1) between higher HMR and an attenuated SBP drop at standing

| DISCUSSION
We aimed to assess whether probable MCI-LB would have a greater probability of presenting with early abnormalities in autonomic response to orthostasis and Valsalva manoeuvre than MCI-AD, as is seen in DLB compared with AD. An abnormal BP response to Valsalva was significantly more common in probable MCI-LB than in  participants than that of the wider clinical population they represent, as poor health is itself a barrier to research participation.
MCI-LB patients were more likely to be in receipt of calcium channel blockers and have previously been shown to be more likely to be in receipt of cholinesterase inhibitors 14 ; the administration of cholinesterase inhibitors in MCI-LB is consistent with local recommendations for managing other clinical features of LB disease, 28 even in the absence of dementia. While efforts were made to adjust for this, as with any observational study, differences in medication use might contribute to obscuring true group differences through either therapeutic effects or exclusion of participants with specific characteristics.
These data may also be limited more generally by the methods of collection. While the blood pressure measurements calibrate finger cuff BP against arm cuff oscillatory BP, neither reflects the gold standard: the Task Force ® Monitor has an accuracy of +/− 5 mmHg in the range of interest (50-250 mmHg). Artefacts were common and many participants were unable to provide data either through excessive artefacts or signal loss, or due to their medical history: such missing data may well reflect processes of interest (e.g. invalid BP measures due to dysautonomia related to neurodegeneration). Correct performance of the Valsalva manoeuvre is dependent on participant technique, and cognitive impairment could impose an additional barrier to this.

| CONCLUSIONS
Probable MCI-LB may feature a delayed return to baseline BP after Valsalva, but there was no clear evidence that MCI-LB are more likely to feature OH, neurogenic OH, or heart rate abnormalities after active stand or Valsalva.