Cardiac Stress Test Is Normal in Pre-motor Parkinson’s Disease

Gilad Yahalom, MD,1,8*† Elad Maor, MD, PhD,2,7† Sharon Hassin-Baer, MD,1,8 Shlomo Segev, MD,3 Yechezkel Sidi, MD,4,8 andShaye Kivity, MD4,5,6,7,8

1The Parkinson Disease and Movement Disorders Clinic, Department of Neurology and Sagol Neuroscience Center2The Olga and Lev Leviev Heart Center

3Institute for Medical Screening4Department of Medicine C5Department of Medicine A

6Center for Autoimmune Diseases, Sheba Medical Center7The Dr. Pinchas Borenstein Talpiot Medical Leadership Program, Tel Hashomer, Israel

8Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel

ABSTRACT: Cardiac sympathetic denervation isan early nonmotor feature of Parkinson’s disease (PD).The aim of the current study was to trace evidence forcardiac dysfunction abnormalities in the premotor phaseof PD. We retrospectively reviewed treadmill ergometrictests of a large cohort (n 5 16,841) between 2000 and2012, that attended the Executive Screening Survey (ESS)at Sheba Medical Center. Heart rate and blood pressureprofiles as well as exercise capacity were comparedbetween subjects who later developed PD and age- andsex-matched subjects (ratio 1:2) who did not. We identi-fied 28 subjects (24 males) who developed PD at follow-up. The PD group was older than the group of subjectswho did not develop PD on first ergometric test(64.82 6 8.82 vs. 48.91 6 10.60 years, P < 0.001). The time

between the first ergometric test and motor symptomsonset was 4.64 6 2.86 years. Patients who later devel-oped PD had lower maximal heart rate (P < 0.001) andlower heart rate reserve than healthy controls (P < 0.001);however, compared with age- and sex-matched subjects,subjects who developed PD had similar exercise capacityand heart rate profile during rest, exercise, and recovery,even 1 year before diagnosis. In this study, we did notdetect significant signs of sympathetic dysfunction duringthe premotor phase of PD. VC 2014 International Parkinsonand Movement Disorder Society

Key Words: ergometric; exercise stress test; Parkin-son’s disease; premotor; preclinical

Parkinson’s disease (PD) is a common neurodegener-ative disease, mainly of the elderly. Some nonmotorsymptoms, predating the motor phase and diagnosis ofPD by many years, commonly occur, including sensoryand autonomic phenomena.1 Over the last few years,some clinical, laboratory, and imaging procedures

have been developed for diagnostic purposes, some-times enabling early detection of PD in the premotorphase. These procedures include the detection of sleepabnormalities, neurobehavioral symptoms, and olfac-tory dysfunction.2 This will hopefully pave the wayfor major advances in disease-modifying therapies.One nonmotor feature that was found to be associatedwith PD is cardiac sympathetic denervation,3,4 as dem-onstrated by the metaiodobenzyl-guanidine (MIBG)cardiac scan. The role of the MIBG-cardiac scan inpreclinical PD has not yet been determined. A singlecase report has been published of decreased sympa-thetic innervation demonstrated by a 6-[18F]fluorodop-amine cardiac positron emission tomography scanperformed 4 years before the appearance of initialmotor symptoms of PD.5 In another study, positiveimmunostaining for alpha-synuclein aggregates, inbiopsy specimens of autonomic plexuses obtained

------------------------------------------------------------*Correspondence to: Gilad Yahalom, The Parkinson Disease and Move-ment Disorders Clinic, Department of Neurology, the Chaim Sheba Medi-cal Center, Ramat-Gan, Israel, E-mail: gyahalom@gmail.com

Funding agencies: None.

Relevant conflicts of interest/financial disclosures: Nothing to report.Full financial disclosures and author roles may be found in the online ver-sion of this article.

†Both authors contributed equally to this article.

Received: 2 September 2013; Revised: 9 May 2014; Accepted: 28May 2014

Published online 20 June 2014 in Wiley Online Library(wileyonlinelibrary.com). DOI: 10.1002/mds.25943

R E S E A R C H A R T I C L E

1158 Movement Disorders, Vol. 29, No. 9, 2014

from subjects without known neurodegenerative dis-ease, correlated with a lower uptake of MIBG-cardiacscan 16 months after biopsy (P 5 0.03), perhaps sug-gesting early pathological involvement of the periph-eral autonomic neurons in Lewy body disorders,including PD.6

Because heart rate (HR) at any moment reflects abalance between sympathetic and parasympatheiceffects on the heart, previous studies investigated exer-cise stress testing (EST) of PD patients.7-10 However, asingle publication has recently presented a study thatevaluated the HR profile at the premotor phase of PD.Palma and colleagues10 retrospectively studied acohort of subjects who underwent EST and were fol-lowed until some developed PD. They compared 18such detected PD patients with healthy controls andfound differences in prediagnosis EST, concluding thatchronotropic insufficiency may constitute an early signof PD during the premotor phase.

The aim of the current study was to try to confirmthe results of Palma et al. in a large Israeli cohort ofapparently healthy men and women, some of whomlater developed PD; EST was performed at severaltimes before the development of PD motor symptoms,and the analysis included comparison of results at twotimes during the premotor period for most patients.

With sympathetic denervation, we might assumethat HR during effort as well as the HR reserve willbe lower than average because of a reduced output ofthe sympathetic system. These early signs, if found,may serve as a biomarker for PD at the premotorphase.

Methods

Patients

This cohort was composed of subjects attendingthe Executive Screening Survey (ESS) at the ShebaMedical Center (SMC), Ramat Gan, Israel, forannual medical checkups. The population consistsmainly of apparently healthy men and women whoundergo annual examinations for health screening.The study population was described previously.11,12

At the time of each annual examination, every partic-ipant in the ESS is interviewed, using standard healthquestionnaires that gather information regardingdemographic characteristics, medical history, andhealth-related habits. Thereafter, blood samples aredrawn after a 12-hour fast and analyzed immediately.A physician at the center performs a complete physi-cal examination, including blood pressure measure-ment. The subject undergoes a standard EST eachyear. A computerized database of all of the annualvisits in this center was established in the year 2000and is the source of data for this study. The follow-up of our cohort ended in December 2012.

Study Protocol

In this retrospective study, subjects who attendedthe ESS and were diagnosed at some point with PD,according to the UK PD Society brain bank clinicaldiagnostic criteria, were included.13 Patients wereexcluded from the study if they were diagnosed withPD before their first EST, or if they had atypical par-kinsonism, medication-induced parkinsonism, severecardiovascular comorbidity, or diabetes mellitus. Thisstudy was part of the Cardiovascular Risk and Meta-bolic Assessment study.11

The diagnosis of PD was reviewed and confirmed byreview of the patients’ files (in patients who were fol-lowed up in the Movement Disorders Clinic at SMC),by correspondence with their neurologists, or by avisit and personal examination by a movement disor-der specialist (G.Y.) at the PD and Movement Disor-ders Clinic at SMC. The age at PD motor symptomsappearance and PD diagnosis was retrieved.

The Institutional Review Board of the SMCapproved this study contingent on strict maintenanceof participants’ anonymity during the database analy-ses. Data from patients were recorded anonymously.Informed consent was signed by each PD patient whowas approached. For PD subjects who alreadyattended the PD and Movement Disorders Clinic atSMC, and for the healthy controls, the requirementfor informed consent contingent on a de-identifiedanalysis of the study database was waived.

Subjects from the ESS who did not develop PD orparkinsonism at the time of their last visit at the ESSwere matched to the PD patients for age at the time ofthe EST and for sex and served as controls. A sub-

analysis of the PD group was performed by ESTmatching at 3 different time points: first visit, second(subsequent) visit, and at the last annual visit beforethe PD diagnosis. The propensity score matching wasrepeated each time, and the control subjects were notnecessarily identical at the different matchings.

All subjects who completed treadmill exerciseaccording to the BRUCE protocol,14 under the super-vision of a board-certified cardiologist, were includedin the analysis. Subjects on anti-hypertensive agentswere requested not to take their anti-hypertensivemedications on the morning of the EST. Heart ratewas measured and documented before the beginningof the exercise test, at the end of the test, and duringrecovery (4 minutes after the end of the test). The HRreserve was defined as the difference between HR atrest and during maximal exercise. Additional docu-mented variables included exercise duration (in sec-onds), systolic and diastolic blood pressure at rest, atthe end of the test, and during recovery, the presenceof atrial or ventricular arrhythmias during the exer-cise, and metabolic equivalents (METs). A MET is ameasure to estimate maximal oxygen uptake for a

E X E R C I S E S T R E S S T E S T S I N P R E M O T O R P D

Movement Disorders, Vol. 29, No. 9, 2014 1159

given workload. It is calculated by the speed andgrade of the treadmill.15

Statistical Analysis

Baseline clinical characteristics and all ergometricparameters were compared between the PD patientsand the control group, either by using a t test for con-tinuous variables or by using a chi-squared test forcategorical variables.

To balance covariates between the two studygroups, initial propensity score matching was used ona 1:10 ratio (PD: controls, respectively), as previouslydescribed by Thoemmes.16 The propensity score analy-sis included matching for sex and age. All controlpatients’ files were carefully reviewed. Control subjectswith cardiovascular disease, diabetes mellitus, or con-gestive heart failure were manually excluded. Theremaining control subjects were then matched again,using propensity score matching with respect to age,sex, and hypertension, with a ratio of 1:2. Pvalue<0.05 was defined as significant. All analyseswere performed using SPSS software version 20.

Results

The entire ESS cohort consisted of 16,841 partici-pants who completed full assessment, including EST.After a screening of the entire cohort, we identified101 patients who were given the diagnosis of PD atsome time point during their follow-up at the ESS.After including only patients who fulfilled inclusioncriteria and performed their ergometic testing beforePD diagnosis, 28 PD patients (24 males) wereenrolled. The numbers of individuals at the differentstages of the enrollment to the study are presented inthe flow diagram in Figure 1. At their first visit to the

ESS, the average age of the subjects who eventuallydeveloped PD was higher as compared with the con-trol group (mean ages at first EST were 64.826 8.82years and 48.91 6 10.60 years, respectively,P< 0.001). Parkinson’s disease manifested 4.64 6 2.86years after the first EST (range, 1-10 years).

Initial unadjusted analysis showed that PD patients hada significantly lower HR at maximal effort (P< 0.001), alower HR reserve (P<0.001), and a higher systolic bloodpressure (SBP) at rest (P<0.002) and during recovery(P 5 0.03). METS was significantly lower in PD patients(P 5 0.008). In addition, exercise duration was signifi-cantly shorter in PD patients (P< 0.001).

However, after propensity score matching and second-ary matching of the PD group to 56 age- and sex-matchedcontrols, no significant differences were found betweenthe PD and the control group in all of the parameters men-tioned (Table 1). The median follow-up time of the con-trol group was 7 years (interquartile range of 3-11 years).There was a trend, approaching significance, of reducedSBP at maximal effort (170.54 20.65 6 mmHg vs.179.09 6 18.86 mmHg, P 5 0.07), and a reduced HRreserve (62.21 6 19.45 beats per minute vs. 69.85 6 16.19beats per minute, P 5 0.08) in PD patients.

Looking at the change in ergometric performancebetween first and subsequent ESTs, no significant dif-ference was seen between patients with PD andhealthy controls in all parameters. The median follow-up time of the control group was 10 years (interquar-tile range, 4-11 years). The median interval betweenthe first and second ESTs was 365 days; interquantilerange, 350-374 days; Table 2). There was a trend ofreduction in HR reserve, far from reaching signifi-cance, in the PD group (25.64 6 16.28 beats perminute vs. 20.80 6 18.99 beats per minute, P 5 0.26).

Looking at the EST 1 year before the appearance ofmotor symptoms, no significant differences were foundin any parameters in the EST between the PD and thecontrol groups. The median follow-up time of the con-trol group was 9 years (interquartile range, 6-12years). Further details are given in Table 3.

Discussion

The current analysis failed to demonstrate significantEST abnormalities suggestive of cardiac autonomicdysfunction in patients who later developed PD. Ourresults suggest that ergometric findings do not serve aspremotor markers for PD.

In a recently published study by Palma et al.,10

results of treadmill exercise tests were comparedbetween patients with PD in their premotor phase(average 4.27 years before diagnosis) and apparentlyhealthy subjects. The authors found that the maximalHR at peak exercise was significantly lower in PDpatients as compared with the controls, and

FIG. 1. This flow diagram represents the pathway through whichpatients were selected to be enrolled in the study. Controls werematched using a propensity score matching method with respect toage, sex, and hypertension. The control sampling was repeated ateach time point and was not similar at the different matchings.

Y A H A L O M E T A L

1160 Movement Disorders, Vol. 29, No. 9, 2014

commented that this might be a reflection of reducedsympathetic outflow in PD patients.10

The reason for the discrepancy between our resultsand those of Palma et al. is unclear. Our cohort ofpremotor PD patients did not differ substantially fromthe cohort of Palma et al. in terms of age(64.82 6 8.82 years vs. 65.22 6 6.83 years, respec-tively), inclusion criteria, or EST protocols. The rela-tively small sample (n 5 18) in the study by Palmaet al. should be noted. Furthermore, our cohort (bothpatients with PD and healthy controls) showedimprovement in EST performance over the years(Fig. 2). This paradoxical improvement despite theaging of the cohort throughout the years may repre-sent a selection bias of people who turned to the ESSas a first step to change their lifestyle because of grow-ing awareness of their general health.

Another study7 assessing cycling EST in 16 patientswith existing PD treated with dopaminergic agents, atan early stage, showed no differences between patientswith PD and controls, in changes of HR or of SBPacross all ranges of effort skills. A significant differ-ence was found in changes in cardiac contractility atthe highest degrees of effort only in patients with evi-dent sympathetic denervation as per MIBG scan,whereas patients with an intact sympathetic system asper MIBG scan did not differ from controls, despitethe possible deleterious effects of the dopaminergicagents. Whether the low cardiovascular response seenin PD patients in the motor phase at peak exercise isattributable to the sympathetic denervation or second-ary to the use of dopaminergic agents is undetermined.One study examined the effect of dopaminergic agents

on EST in 14 PD patients and found that autonomicabnormalities during exercise were related to the dis-ease and not to dopaminergic agents.17

Another study8 assessing cycling EST in 15 PDpatients (11 at Hoeh and Yahr [H&Y] stage 3 and 4 atH&Y stage 2) did not find striking differences in cardi-ovascular adaptation to physical work in PD patients.Werner et al.9 assessed treadmill EST in 16 PD patientsat stage 2 of H&Y, all treated with levodopa, andfound abnormal cardiovascular responses at higherexercise intensities. At peak exercise, blood pressure,but not HR, was significantly higher in the controlgroup. This was also a trend in our study 4.64 6 2.86years before motor symptoms appearance, but thistrend was minimized at 1 year before motor symptoms’

TABLE 1. Earliest exercise stress test in PD and controlsubjects after adjustment for age, sex, and hypertension

PD Controls P-Value

N 28 56Age, years6SD 64.8268.82 65.5068.82 0.74Males 24 48HR at rest, bpm6SD 78.14612.08 75.52612.57 0.36Peak HR, bpm6SD 140.36621.99 145.38617.62 0.30HR at recovery, bpm6SD 89.086 21.44 88.53616.11 0.91DBP at rest, mmHg6SD 78.2169.05 79.0267.65 0.69SBP at rest, mmHg6SD 133.21619.01 128.57614.98 0.27Peak DBP, mmHg6SD 79.4667.50 81.36608.69 0.31Peak SBP, mmHg6SD 170.54620.65 179.09 618.86 0.07DBP at recovery,mmHg6SD

78.4667.97 77.7766.46 0.70

SBP at recovery,mmHg6SD

135.00619.18 134.73619.67 0.95

HRR, bpm6SD 62.21619.45 69.85616.19 0.08METS6SD 10.2062.95 9.9862.48 0.77Stress duration(seconds6SD)

475.716173.84 467.436164.19 0.84

Abbreviations: PD, Parkinson’s disease; N, number of subjects; BMI, bodymass index; HR, heart rate; DBP, diastolic blood pressure; SBP, systolicblood pressure; METS, metabolic equivalents; HRR, heart rate reserve.

TABLE 2. The change in performance in 2 subsequentexercise stress tests in PD and control subjects, adjusted

for age, sex, and hypertension

PD Controls P-value

N 25 50Age, years6SD 65.2468.96 65.2468.88 1.00Males 21 42dHR at rest, bpm6SD 3.04612.11 0.44619.44 0.48Peak dHR, bpm6SD 22.60614.15 20.3667.28 0.46dHR at recovery, bpm6SD 21.08629.40 2.52615.62 0.57dHRR, bpm6SD 25.64616.28 20.80618.99 0.26Change in stressduration (seconds6SD)

21.726165.13 4.926181.56 0.69

Abbreviations: PD, Parkinson’s disease; N, number of subjects; dHR, thechange (5delta) in heart rate between the first and second exercise stresstests; dHRR, the change (5delta) in heart rate response between the firstand second exercise stress tests.

TABLE 3. Exercise stress test in PD subjects 1 year beforemotor symptoms appearance as compared with age, gen-

der, and hypertension matched control subjects

PD Controls P-value

N 20 40Age, years6SD 69.0569.90 69.0369.74 0.99Males 18 36HR at rest, bpm6SD 73.906 13.56 75.77613.71 0.62Peak HR, bpm6SD 133.30621.79 135.40621.97 0.73HR at recovery, bpm6SD 86.35624.26 85.37616.30 0.87DBP at rest, mmHg6SD 76.2567.76 77.8967.77 0.45SBP at rest, mmHg6SD 123.25614.17 127.44622.12 0.38Peak DBP, mmHg6SD 79.2567.12 79.7568.55 0.81Peak SBP, mmHg6SD 166.50615.65 172.70621.68 0.21DBP at recovery,mmHg6SD

76.5066.09 76.7165.96 0.90

SBP at recovery,mmHg6SD

132.25611.06 134.21617.18 0.60

HRR, bpm6SD 59.40623.04 59.54621.60 0.98METS6SD 9.7363.10 9.9662.99 0.80Stress duration(seconds6SD)

503.156186.62 452.036194.61 0.33

Abbreviations: PD, Parkinson’s disease; N, number of subjects; BMI, bodymass index; HR, heart rate; DBP, diastolic blood pressure; SBP, systolicblood pressure; METS, metabolic equivalents; HRR, heart rate reserve.

E X E R C I S E S T R E S S T E S T S I N P R E M O T O R P D

Movement Disorders, Vol. 29, No. 9, 2014 1161

appearance. In the Werner et al. study, half of the PDpatients failed to attain 85% of target HR.

Based on these results, several differences are seen inEST in PD patients; however, in most studies, patientswere already diagnosed and under treatment withdopaminergic agents. Our study did not show markeddifferences in EST between PD patients at the premo-tor phase and relatively healthy controls, even at 1year before the appearance of PD motor symptoms.The trend to lower HR reserve, which signifies areduction in sympathetic outflow, found in our PDgroup in the first EST, an average of 4.64 years beforethe appearance of motor symptoms, did not reach sig-nificance at a further stage up to 1 year before motorsymptoms’ appearance. This trend was even annulled.

Several limitations were identified in the presentstudy. First, despite being to the best of our knowl-edge the largest study assessing EST in premotor PDpatients, the analysis is still underpowered withrespect to minor changes in HR profile and exercisecapacity. Second, this is a retrospective observationalstudy, and therefore the risk of selection bias and

residual confounding still exists, despite validationanalyses using propensity-score matching. Third, thestudy was limited to a primarily Jewish, relativelyyoung, healthy, and high-socioeconomic-backgroundpopulation. This might limit the validity of our con-clusions to other populations. We might expect thatpeople of high socio-economic status have moreawareness of the importance of physical activity, andincreasing evidence indicates that physical activity isassociated with reduced risk of developing PD.18

Finally,we conclude that EST seems to have no diag-nostic role in subjects at the premotor phase of PD.

References1. Adler CH. Premotor symptoms and early diagnosis of Parkinson’s

disease. Int J Neurosci 2011;121(Suppl 2):3-8.

2. Chahine LM, Stern MB. Diagnostic markers for Parkinson’s dis-ease. Curr Opin Neurol 2011;24:309-317.

3. Spiegel J. Diagnostic and pathophysiological impact of myocardialMIBG scintigraphy in Parkinson’s disease. Parkinsons Dis 2010;2010: 295346.

4. Orimo S, Suzuki M, Inaba A, Mizusawa H. 123I-MIBG myocar-dial scintigraphy for differentiating Parkinson’s disease from otherneurodegenerative parkinsonism: a systematic review and meta-analysis. Parkinsonism Relat Disord. 2012;18:494-500.

5. Goldstein DS, Sharabi Y, Karp BI, et al. Cardiac sympatheticdenervation preceding motor signs in Parkinson disease. Cleve ClinJ Med 2009;76(Suppl 2):S47-50.

6. Minguez-Castellanos A, Chamorro CE, Escamilla-Sevilla F, et al.Do alpha-synuclein aggregates in autonomic plexuses predate Lewybody disorders?: a cohort study. Neurology 2007;68:2012-2018.

7. Nakamura T, Hirayama M, Yamashita F, et al. Lowered cardiacsympathetic nerve performance in response to exercise in Parkin-son’s disease. Mov Disord 2010;25:1183-1189.

8. Reuter I, Engelhardt M, Freiwaldt J, Baas H. Exercise test in Par-kinson’s disease. Clin Auton Res 1999;9:129-134.

9. Werner WG, DiFrancisco-Donoghue J, Lamberg EM. Cardiovascu-lar response to treadmill testing in Parkinson disease. J NeurolPhys Ther 2006;30:68-73.

10. Palma JA, Carmona-Abellan MM, Barriobero N, et al. Is cardiacfunction impaired in premotor Parkinson’s disease? A retrospectivecohort study. Mov Disord 2013;28:591-596.

11. Shaye K, Amir T, Shlomo S, Yechezkel S. Fasting glucose levelswithin the high normal range predict cardiovascular outcome. AmHeart J 2012;164:111-116.

12. Maor E, Kivity S, Kopel E, et al. Differences in heart rate profileduring exercise among subjects with subclinical thyroid disease.Thyroid 2013;23:1226-1232.

13. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diag-nosis of idiopathic Parkinson’s disease: a clinico-pathological studyof 100 cases. J Neurol Neurosurg Psychiatry 1992;55:181-184.

14. Bruce RA. Exercise testing of patients with coronary heart disease:principles and normal standards for evaluation. Ann Clin Res1971;3:323-332.

15. Bruce RA, Kusumi F, Hosmer D. Maximal oxygen intake andnomographic assessment of functional aerobic impairment in cardi-ovascular disease. Am Heart J 1973;85:546-562.

16. Thoemmes F. Propensity score matching in SPSS. arXiv:120163852012.

17. DiFrancisco-Donoghue J, Elokda A, Lamberg EM, Bono N,Werner WG. Norepinephrine and cardiovascular responses to max-imal exercise in Parkinson’s disease on and off medication. MovDisord 2009;24:1773-1778.

18. Xu Q, Park Y, Huang X, et al. Physical activities and future riskof Parkinson disease. Neurology 2010;75:341-348.

FIG. 2. The overall fitness of the Parkinson’s disease cohort (a) and of thehealthy population (b) represented by exercise duration at different visits.

Y A H A L O M E T A L

1162 Movement Disorders, Vol. 29, No. 9, 2014