ARTICLE OPEN ACCESS

Early predictors of mortality in parkinsonismand Parkinson diseaseA population-based study

David Backstrom MD Gabriel Granaringsen MSc Magdalena Eriksson Domellof PhD Jan Linder MD PhD

Susanna Jakobson Mo MD PhD Katrine Riklund MD PhD Henrik Zetterberg MD PhD

Kaj Blennow MD PhD and Lars Forsgren MD PhD

Neurologyreg 201891e2045-e2056 doi101212WNL0000000000006576

Correspondence

Dr Backstrom

davidbackstromumuse

AbstractObjectiveTo examine mortality and associated risk factors including possible effects of mild cognitiveimpairment imaging and CSF abnormalities in a community-based population with incidentparkinsonism and Parkinson disease

MethodsOne hundred eighty-two patients with new-onset idiopathic parkinsonism were diagnosedfrom January 2004 through April 2009 in a catchment area of 142000 inhabitants in SwedenPatients were comprehensively investigated according to a multimodal research protocol andfollowed prospectively for up to 135 years A total of 109 patients died Mortality rates in thegeneral Swedish population were used to calculate standardized mortality ratio and expectedsurvival and Cox proportional hazard models were used to investigate independent predictorsof mortality

ResultsThe standardized mortality ratio for all patients was 184 (95 confidence interval 150ndash222 plt 0001) Patients with atypical parkinsonism (multiple system atrophy or progressive supra-nuclear palsy) had the highest mortality In early Parkinson disease a mild cognitive impair-ment diagnosis freezing of gait hyposmia reduced dopamine transporter activity in thecaudate and elevated leukocytes in the CSF were significantly associated with shorter survival

ConclusionAlthough patients presenting with idiopathic parkinsonism have reduced survival the survival ishighly dependent on the type and characteristics of the parkinsonian disorder Patients withParkinson disease presenting with normal cognitive function seem to have a largely normal lifeexpectancy The finding of a subtle CSF leukocytosis in patients with Parkinson disease withshort survival may have clinical implications

RELATED ARTICLES

EditorialLife expectancy inParkinson disease

Page 991

Patient PageParkinson disease andmortality Understandinghow the two are connected

Page e2106

MORE ONLINE

PodcastDr Jeffery Ratliff interviewsDr David Baumlckstroumlm ona population-based studyinvestigating earlypredictors of mortality inparkinsonism andParkinson disease

NPuborgpatgbx

CME CourseNPuborgcmelist

From the Department of Pharmacology and Clinical Neuroscience (DB MED JL LF) Epidemiology and Global Health Unit Department of Public Health and Clinical Medicine(GG) Department of Psychology (MED) and Department of Radiation Sciences Diagnostic Radiology and Umearing Center for Functional Brain Imaging (SJM KR) UmearingUniversityInstitute of Neuroscience and Physiology (HZ KB) Department of Psychiatry and Neurochemistry Sahlgrenska Academy at University of Gothenburg Molndal Clinical Neuro-chemistry Laboratory (HZ KB) Sahlgrenska University Hospital Molndal Sweden Department of Molecular Neuroscience (HZ) University College London Institute of Neurologyand UK Dementia Research Institute at UCL (HZ) London UK

Go to NeurologyorgN for full disclosures Funding information and disclosures deemed relevant by the authors if any are provided at the end of the article

The Article Processing Charge was funded by Umearing University

This is an open access article distributed under the terms of the Creative Commons Attribution License 40 (CC BY) which permits unrestricted use distribution and reproduction in anymedium provided the original work is properly cited

Copyright copy 2018 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology e2045

In Parkinson disease (PD) the second most common neu-rodegenerative disorder life expectancy is reduced1 Thereare however conflicting data regarding the size of and spe-cific factors accounting for the reduced survival in comparisonto the general population Most studies of survival in PD havebeen hospital-based or have used register-based case-findingmethods These designs may produce biased results throughunderrepresentation of mild PD cases and lack of referral ofolder patients to hospital clinics2 There are also few studiesof the survival in unselected populations of patients withnew-onset idiopathic parkinsonism (including atypicalparkinsonism) rather than PD

In PD previous studies have found that a nontremor-dominant phenotype PD dementia (PDD) and early auto-nomic dysfunction are associated with a shorter survival134

The recently defined diagnosis of mild cognitive impairmentin PD (PD-MCI)5 has rarely been studied in this regardFurthermore the neurobiology of PD with short survival(in terms of factors such as striatal dopamine depletionpatterns CSF abnormalities or APOE genotype) is not wellknown

Against this background we assessed all-cause mortality andassociated risk factors in a population-based Swedish cohortof patients with incident carefully diagnosed idiopathic par-kinsonism including PD The patients underwent extensiveneurologic neuropsychological and laboratory testing as wellas multimodal neuroimaging and received standard or if in-dicated advanced treatments by movement disorder neurol-ogists during long-term follow-up

MethodsStudy populationAll participants were part of a population-based incidencestudy of unselected cases of new-onset idiopathic parkin-sonism from a defined geographic catchment area of 142000inhabitants in northern Sweden6 The local tradition is to referall patients with suspected PD to the Department of Neu-rology at Umearing University Hospital (our department) whichis the only neurologic department in the area To avoid se-lection bias and to make case identification as complete aspossible a population screening was performed throughmanysources including letters sent twice yearly to all health prac-titioners asking for referral of all suspected cases with incidentparkinsonism Eldercare institutions were surveyed by visits(the largest institution) with an examination by neurologists

of all cases that were reported to have signs of parkinsonismor by interview (remaining institutions)

All patients were recruited to the study in the early motor(drug-naive) phase between January 1 2004 and April 302009 After exclusion of patients with secondary parkinsonism(eg due to neuroleptic drugs or stroke) or dementia atbaseline (eg patients with dementia with Lewy bodies) 182participants were included and followed prospectively A di-agnosis of PD multiple system atrophy (MSA) or progressivesupranuclear palsy (PSP) required agreement among theexaminers (neurologists specialized in movement disorders)that the clinical criteria for the diagnosis were fulfilled basedon the UK Parkinsonrsquos Disease Society Brain Bank criteria7 orcriteria for MSA or PSP89

At the latest follow-up (up to death or 85ndash135 years followinginclusion in alive patients) 143 patients were diagnosed withPD 13 with MSA 18 with PSP 4 as having unclassifiableparkinsonism and 4 did not have idiopathic parkinsonism(figure 1) Autopsy confirmed the diagnosis in 3 cases of PDand 2 cases of PSP

Standard protocol approvals registrationsand patient consentsAll participants provided written informed consent The studywas approved by the Regional Medical Ethics Board in UmearingSweden and was performed in accord with the Declarationof Helsinki

Clinical assessmentsAll participants were investigated at study entry (baseline)and followed up prospectively with standardized clinicalexaminations including the Unified Parkinsonrsquos DiseaseRating Scale (UPDRS) the modified Hoehn and Yahr ScaleMini-Mental State Examination (MMSE) and the 39-itemParkinsonrsquos Disease Questionnaire at least yearly as de-scribed previously10 We divided the baseline UPDRS scoresinto subscores for tremor (sum of items 20 and 21) andpostural imbalance and gait disorder (PIGD sum of items 1314 15 29 and 30) and classified the phenotype of PD astremor predominant intermediate or PIGD in accordancewith the DATATOP (Deprenyl and Tocopherol Anti-oxidative Therapy of Parkinsonrsquos Disease) trial analyses11

Olfactory function was investigated by the 12-item Brief SmellIdentification Test12 global mobility by the Timed Up andGo (TUG) test which is the time it takes to rise up froma chair walk 3 m and sit down again13 and depression by theMontgomery-Aringsberg Depression Rating Scale14 Severe

GlossaryAβ42 = β-amyloid 1ndash42 CI = confidence interval DAT = dopamine active transporter MCI = mild cognitive impairmentMMSE =Mini-Mental State ExaminationMSA = multiple system atrophy PD = Parkinson disease PDD = Parkinson diseasedementia PIGD = postural imbalance and gait disorder PSP = progressive supranuclear palsy SMR = standardized mortalityratio TUG = Timed Up and Go UPDRS = Unified Parkinsonrsquos Disease Rating Scale

e2046 Neurology | Volume 91 Number 22 | November 27 2018 NeurologyorgN

hyposmia is defined by a score lt4 on the Brief Smell Identi-fication Test A possible REM sleep behavior disorder wasdefined by a history of dream enactment behavior ona screening questionnaire

CSF analysisAt study entry 99 of the patients with PD and 21 of thepatients with atypical parkinsonism agreed to CSF collectionby lumbar puncture using standard procedures CSF totaltau concentration was measured using a sandwich ELISA(INNOTEST hTAU Ag Fujirebio Ghent Belgium) con-structed to determine all tau isoforms irrespective ofphosphorylation status15 Tau phosphorylated at threonine181 was measured by a sandwich ELISA (INNOTESTPHOSPHO-TAU [181P] Fujirebio)16 CSF β-amyloid 1ndash42(Aβ42) levels were measured using a sandwich ELISA(INNOTEST β-AMYLOID [1ndash42] Fujirebio) constructedto measure Aβ containing the first and 42nd amino acids17

CSF levels of α-synuclein were analyzed using a commerciallyavailable α-synuclein human ELISA (KHB0061 InvitrogenCarlsbad CA) according to instructions by the manufacturerExperienced board-certified laboratory technicians per-formed the CSF analyses using procedures approved by theSwedish Board for Accreditation and Conformity AssessmentCells in the CSF were counted by automated flow cytometryA subtle pleocytosis of mononuclear leukocytes was opera-tionally defined by counts ranging between 2 and 20 cells permicroliter However no patient had more than 10 cells permicroliter

Dopamine active transporter imagingOf the 182 patients enrolled in the study 170 patients (934)underwent dopamine active transporter (DAT) imaging by123I-FP-CIT (DaTSCAN GE Healthcare BV Eindhoven TheNetherlands) SPECT DAT imaging was done 3 hours fol-lowing an IV bolus dose of 185MBq 123I-FP-CIT Imaging was

done prior to commencement of medication at baseline Theimaging protocol was done within the framework of a non-profit clinical trial (EU no 2009-011748-20) and constituteda substudy within the research project Semiquantitativeanalysis (based on regions of interest) and visual evaluation ofthe DAT SPECT were done unbiased by any clinical in-formation at all times Normal reference values were derivedfrom an age-matched group of healthy controls participatingin the study and reduction of DAT uptake in the patients withPD was measured in percent and SDs of the normal valuesThe most affected side (left or right) was defined by theputamen and caudate that showed the largest reduction of123I-FP-CIT uptake The putamen and caudate were in-vestigated separately The imaging protocol equipment andsemiquantitative evaluation methods that were used havebeen described earlier18 Two different SPECT cameras wereused during the course of the project one brain-dedicatedSPECT camera (the Neurocam) was later substituted bya multipurpose hybrid SPECTCT (both General ElectricMilwaukee WI) Normal reference values were establishedfor both equipments1819 All PD MSA and PSP patientsfulfilling diagnostic criteria and who participated in the DATimaging (n = 163) had a pathologic scan

Genetic testingOne hundred thirty-three of the patients with PD agreed toDNA analysis by peripheral blood sampling DNA was iso-lated from peripheral blood using standard procedures Thevariations of interest (the e2e3e4 polymorphisms) in theAPOE gene were genotyped using TaqMan Assays-by-Design(Applied Biosystems Foster City CA) The assay was per-formed according to manufacturerrsquos instructions and analyzedusing the allelic discrimination function of the TaqMan 7900HT Fast Real-Time PCR system (Applied Biosystems) Ge-notype success rates of 100 were obtained The patientswho declined all laboratory investigations (n = 10) were older

Figure 1 Flowchart of patients included in the study

Diagnosis was established according to clinical diagnosisat the latest follow-up and confirmed by autopsy in 5patients MSA = multiple system atrophy PSP = pro-gressive supranuclear palsy

NeurologyorgN Neurology | Volume 91 Number 22 | November 27 2018 e2047

than the others (791 vs 705 years) and had slightly higherUPDRS scores for motor dysfunction but were otherwisecomparable All laboratory analyses were performed blindedfrom clinical data

NeuropsychologyExtensive neuropsychological testing used for PD-MCI andPDD20 diagnostics was performed at baseline and after 1 3 5and 8 years At baseline the patients with PD could be dividedinto patients with MCI (PD-MCI n = 61) and patientswithout PD-MCI (PD with normal cognition n = 82) Be-cause all cognitive domains had been covered in the testbattery throughout the study period PD-MCI diagnoses wereapplied according to Movement Disorder Society guidelines5

using level 2 criteria21 Patients were classified as having MCIat baseline if (1) impaired in a minimum of 2 tests in onedomain (single-domain MCI) or in a minimum of one test in2 different domains (multiple-domain MCI) (2) impair-ments were ge15 SDs below mean of normative data (3) self-perceived cognitive decline was reported in a questionnaire(eg the 39-item Parkinsonrsquos Disease Questionnaire) andordirectly by patient andor family member and (4) the patienthad no functional impairment in basic activities of living(ie driving a car social or personal care medication man-agement) due to cognitive impairment The tests used forMCI classification were for episodic memory Free and CuedSelective Reminding Test Logical Memory and Paired As-sociative Learning from the Wechsler Memory Scale andBrief Visuospatial Memory Test (total recall) for workingmemory Digit Span Forward and Digit Span Backward fromWechsler Adult Intelligence Scale III for attention TrailMaking Test A and B for language Controlled Oral WordAssociation and Boston Naming Test for visuospatial func-tion the Benton Judgment of Line Orientation Test andPentagon Copying from MMSE and for executive functionWisconsin Card Sorting Testndashcomputer version 2 MentalControl fromWechsler Memory Scale and Category Fluency(animals in 60 seconds) In the multivariable analysis of fac-tors predicting mortality in PD baseline neuropsychologyresults were used However to investigate possible effects ofincident dementia (of which there were no cases at baseline)the neuropsychology results from the 3-year follow-up werealso analyzed in relation to mortality Incident PDD and PD-MCI were diagnosed at the 3-year follow-up on the basis ofprevious test results a documented decline cognitive declinereported by the patient andor family member and (forPDD) by the occurrence of functional impairment in basicactivities of living caused by cognitive impairment StructuralMRI and routine laboratory tests were performed to excludecognitive impairment due to other causes than PD-MCIor PDD

MortalityWe followed all surviving patients yearly for approximately 85to 135 years until August 31 2017 The average time sinceinclusion in all patients was 77 years Although a few olderpatients were followed by telephone rather than visits during

the last few years the survival data were complete A deathcertificate in which the cause of death was stated wasobtained for 98 (90) of all the fatalities in the cohort

StatisticsPotential group differences in clinical variables at baselinewere examined by 1-way analysis of variance Fisher exact testand Kruskal-Wallis test and correlation between variables byPerson r and Spearman ρ as appropriate Post hoc contrasts inanalysis of variance are shown corrected for multiple com-parisons using Tukey HSD (honestly significant difference)The age- and sex-specific standardized mortality ratio (SMR)was calculated stratified by clinical diagnosis and sex and inthe patients with PD by MCI status by dividing the observednumber of deaths counted from baseline in each group bythe expected number of deaths in each group The expectednumbers of deaths were calculated using the age- and sex-specific official Swedish National Statistics of 2004ndash2017mortalities multiplied by the person-time from each group inthe study Confidence intervals (CIs) for the SMRs werecalculated to the 95 level using the Poisson distribution andp values using the χ2 distribution Life expectancies (antici-pated mean remaining time to live) based on SMR werecalculated using a modified Gompertz function as donepreviously22 Cox proportional hazard analysis was used toinvestigate whether one or more covariates predicted mor-tality in PD A list of potential risk factors for mortality wasgenerated including factors previously found predictive andfactors of neurobiological interest (biomarkers) To in-vestigate dopaminergic denervation at baseline DAT-uptakeratios were normalized by average SDs above or below thenormal mean (ie z score) in order to equate the numericalvalues derived from the 2 different scanners that were usedThe number of mononuclear leukocytes in CSF was in-vestigated as a continuous predictor variable in relation tosurvival The presence of a CSF leukocytosis was also in-vestigated For univariate comparisons we controlled fora false discovery rate of 005 by using the Benjamini-Hochbergprocedure resulting in a significance level of p lt 0017 Be-cause age is strongly related to survival all results were ad-justed for age Multivariable models were then developedusing age with clinical variables and age with biomarker var-iables as predictors of survival Variables significantly associ-ated with survival at the p lt 02 level in the univariate modelswere included using a backward elimination procedure Ifvariables were highly correlated the variable with the lowerp value was included to avoid multicollinearity The only ex-ception was the PIGD score which was correlated to TUGresults but chosen over the TUG test because of perceivedhigher clinical usefulness In the final multivariable modelsonly variables with p lt 0005 were included to avoid over-fitting of the model and to restrict the number of predictorsfor ease of use in clinical practice To avoid confoundingrelated to the heterogeneity of the different parkinsoniandiseases detailed prognostic models were developed only forPD except for the CSF protein markers which were alsoinvestigated in atypical parkinsonism All statistical analyses

e2048 Neurology | Volume 91 Number 22 | November 27 2018 NeurologyorgN

were performed using SPSS 230 (IBM Corp Armonk NY)or R Statistical Software

Data availabilityAnonymized data can be obtained by request from anyqualified investigator for purposes of replicating proceduresand results

ResultsSurvival in incident idiopathic parkinsonismClinical characteristics at baseline for the patients with idio-pathic parkinsonism are shown in table 1 Survival data fromfirst evaluation to death or end of the study were obtained forall participants (figure 1) Of the 178 patients with idiopathicparkinsonism 109 (612) died during follow-up Seventy-seven (538 of 143) of the deaths occurred in the PD group12 (923 of 13) in the MSA group and 16 (889 of 18) inthe PSP group The 4 patients with unclassifiable parkinson-ism likely represent cases of late-onset PD but were excludedfrom further analyses as they did not fulfill specific diagnosticcriteria The overall mean age at death was 820 years Deepbrain stimulation or pumps for intestinal delivery of levodopa

were used or had been used by 12 (84) of the 143 patientswith PD

Survival was related to the age at the first visit (hazard ratio fordeath 303 for each decade older 95 CI 230ndash398 p lt0001) However when survival in patients with MSA or PSPwas analyzed separately from PD the age at the first visit wasnot a significant factor (p = 0766) As a measurement ofglobal cognitive function survival also correlated with thebaseline MMSE score after adjustment for age (119 timeshigher hazard for death for each lower point p = 0006) TheSMR for the whole parkinsonism cohort was 184 (p lt 0001)times higher than the comparable age and sex distribution-standardized mortality of the Swedish population during theyears 2004 to 2017 (table 2)

The SMR for the patients with PD was 158 and 332 for thepatients with atypical parkinsonism (table 2) The patientswith PD had an age-adjusted hazard ratio for death of 043(95 CI 027ndash067 p lt 0001) compared to the patients withatypical parkinsonism More specifically the age-adjustedhazard ratio for death in MSA was 276 (95 CI 148ndash515p = 0001) and 142 (95 CI 108ndash188 p = 0012) inPSP compared to the patients with PD (figure 2) As shown

Table 1 Characteristics of participants at the first (baseline) visit

VariableParkinsondisease MSA PSP

Unclassifiedparkinsonism p Value

No of participants 143 13 18 4

Age y 712 (98) 736 (96) 750 (71) 857 (75) 0012b

Disease duration y 18 (14) 19 (09) 21 (15) 26 (10) 0670

Sex MF ( M) 8558 (59) 85 (61) 810 (44) 22 (50) 0648

Hoehn and Yahr median (IQR) 20 (20ndash30) 25 (20ndash30) 25 (20ndash30) 30 (25ndash35) 0009

UPDRS total median (IQR) 350 (270ndash460) 370 (320ndash490) 350 (310ndash450) 505 (415ndash645) 0283

UPDRS III median (IQR) 260 (190ndash350) 230 (200ndash370) 265 (240ndash320) 350 (280ndash415) 0541

Possible RBD n () 15 (11) 2 (15) 0 (0) 0 (0) 0286

MADRS depression score 45 (37) 80 (74) 77 (94) 50 (00) 0025

Smell test no correct 66 (27) 70 (33) 71 (23) 0626

Freezing ever n () 95 (67) 9 (69) 13 (76) 4 (100) 0508

Orthostatic blood pressure dropa mm Hg 109 (188) 213 (222) 27 (150) 0048c

MMSE score 285 (17) 290 (11) 279 (16) 273 (21) 0181

MCI diagnosis n () 61 (43) 3 (23) 14 (78) 3 (75) 0007d

Abbreviations IQR = interquartile rangeMADRS =Montgomery-Aringsberg Depression Rating ScaleMCI =mild cognitive impairmentMMSE =Mini-Mental StateExaminationMSA=multiple systematrophy PSP = progressive supranuclear palsy RBD=REMsleep behavior disorder UPDRS=Unified ParkinsonrsquosDiseaseRating Scale Part III is the motor function subscaleThe participantswere naive to dopaminergic treatment at this time Values aremeans (SD) unless otherwise stated Disease duration is the time fromonset offirst motor symptom as recalled by the patient to baselinea Systolic pressure difference after 3 minutes of standing Post hoc comparisons using Tukey honestly significant difference with the significance level of p lt005 to control for multiple testing showed the following differencesb Parkinson disease younger than unclassified parkinsonismc MSA more orthostatic than PSPd MCI more common in PSP than in Parkinson disease and than in MSA

NeurologyorgN Neurology | Volume 91 Number 22 | November 27 2018 e2049

in table 3 the most common cause of death was pneumoniabut in many cases the exact cause of death could not bedetermined (eg patients who died alone in their homes)Assuming the mean age of 717 years at baseline in idiopathicparkinsonism the expected survival in PD was 96 years and61 years in atypical parkinsonism

Survival in PDAlthough there was no sex difference in the hazard ratio fordeath in the patients with PD the SMR was higher in female(209 ie 109 more cases of death in female patients withPD than in the general female Swedish population) than inmale (SMR 133) patients (table 2) The survival also cor-related with cognitive status Mortality in patients with PD(both male and female) with normal cognition at baseline wasnot significantly different from the general Swedish pop-ulation while it was increased in patients with PD-MCI(SMR 217 table 2) The patients with PD-MCI at baselinehad a 24 times higher age-adjusted hazard of death duringfollow-up compared to the patients without MCI at baseline(p lt 0001 table 4) Survival in PD also correlated with thebaseline MMSE score (117 times higher hazard for death foreach lower point after adjustment for age p = 0011) Feweryears of formal education was not associated with reducedsurvival Assuming the mean age of 712 years at baseline the

expected survival in PD was 116 years without MCI and 82years with MCI

A baseline PD phenotype comprising MCI freezing of gaita PIGD phenotype hyposmia high disease severity (mea-sured by the total and Part III UPDRS scores) slowness in theTUG test and onset of dementia during the first 3 yearspredicted shorter survival (table 4 and figure 3 AndashF) in theunivariate analysis Tremor was uncorrelated to survival Thepatients with PD had a 38 lower DAT uptake in the puta-men and a 24 lower DAT uptake in the caudate nucleus inthe most affected hemisphere compared to healthy controls Aneurobiological factor (biomarker) that predicted a shortersurvival in PD was reduced DAT uptake in the striatumparticularly in the caudate nucleus (148 times higher hazardfor death for each SD of lower uptake) while the baseline CSFprotein concentrations of neurodegenerative markers andAPOE genotype were not predictive (table 4) Higher Aβ42concentration in CSF was however moderately associatedwith increased survival when the whole cohort of patients withparkinsonian disorders was investigated (age-adjusted hazardratio 099 per pgmL 95 CI 098ndash100 p = 0048 in allpatients and 097 per pgmL 95 CI 095ndash100 p = 0036 inthe patients with atypical parkinsonism) Furthermorea proportion of all patients with PD (131) had a subtle

Table 2 Standardized mortality ratio in parkinsonism in relation to clinical diagnosis

Sex No No of events Expected SMR (95 CI) p Value

Idiopathic parkinsonism cohort all cases

Female 73 48 213 225 (166ndash298) lt0001

Male 101 57 358 159 (121ndash206) lt0001

Total 174 105 571 184 (150ndash222) lt0001

Atypical parkinsonism

Total 31 28 84 332 (221ndash480) lt0001

Parkinson disease

Female 58 34 163 209 (145ndash291) lt0001

Male 85 43 324 133 (096ndash179) 0077

Total 143 77 487 158 (125ndash198) lt0001

Parkinson disease with normal cognition

Female 37 17 108 157 (091ndash251) 0085

Male 45 18 185 097 (058ndash154) 0814

Total 82 35 293 119 (083ndash166) 0341

Parkinson disease with MCI

Female 21 17 55 312 (182ndash499) lt0001

Male 40 25 139 180 (1162ndash265) 0005

Total 61 42 194 217 (156ndash293) lt0001

Abbreviations CI = confidence interval MCI = mild cognitive impairment SMR = standardized mortality ratio

e2050 Neurology | Volume 91 Number 22 | November 27 2018 NeurologyorgN

pleocytosis of mononuclear leukocytes (2ndash10 cells) in theCSF These patients had a significantly worse prognosiscompared to the other patients (hazard ratio for death 631p lt 0001 table 4) independently of adjustment for eryth-rocytes Given this finding the CSF results were furtheranalyzed There was a trend that a higher number of leu-kocytes at baseline tended to correlate with higher CSF totaltau and phosphorylated tau concentrations (r = 024 p =0019 and r = 022 p = 0028 respectively) There was noevidence of CNS infection in any of the patients and there wasno difference in neurodegenerative marker or total proteinconcentrations in the CSF between PD patients with a leu-kocytosis and the other PD patients However the PDpatients with a CSF leukocytosis had a higher progression ofthe UPDRS III scores at the 1-year follow-up (median 4-pointincrease vs 3-point decrease p = 0007Mann-WhitneyU test)than the other patients with PD despite using equal medi-cation dosages (as measured by levodopa-equivalent dailydose)

After excluding factors strongly correlated to each other andincluding only strictly significant factors (at the p lt 0005level) in a multivariable model older age at baseline prevalentmild cognitive impairment higher PIGD score hyposmia

reduced DAT uptake in the caudate and an inflammatoryreaction (leukocytosis) in the CSF were independent clinicaland biomarker predictors of reduced survival in PD

DiscussionIn this prospective population-based cohort study involvingin-person examination of all participants we evaluated theprognosis of idiopathic parkinsonism and PD with respect tomortality stratified by factors related to MCI and clinical andneurobiological characteristics To avoid selection bias weincluded all patients with incident idiopathic parkinsonism inthe studied area rather than only PD and we explicitly di-agnosed atypical forms of parkinsonism Our results shouldtherefore provide information that is generalizable to theldquoreal-liferdquo experience of the overall population with idiopathicparkinsonism

In line with the findings from several previous studies2324

patients with MSA and PSP had a markedly worse prognosisthan patients with PD In the present cohort the patients withincident atypical parkinsonism (ie MSA or PSP) were olderthan in most previous hospital-based studies24ndash26 possibly

Figure 2 Survival in incident idiopathic parkinsonism

(A) Lexis diagram showing follow-up of patients throughout the study The ldquoatypical parkinsonismrdquo group comprises patientswith new-onsetMSA and PSP (B)Kaplan-Meier plot of survival in relation to diagnosis (for number at risk see supplemental table e-1 linkslwwcomWNLA762) MSA = multiple systematrophy PSP = progressive supranuclear palsy

NeurologyorgN Neurology | Volume 91 Number 22 | November 27 2018 e2051

because of the population-based design Our results confirmthe dire prognosis of these diseases in this cohort with a rel-atively high observed age at onset In fact age was not a sig-nificant predictor of survival in patients with MSA or PSPwhich is likely explained by the strong effect on mortalitycaused by these disorders themselves

Recent studies have also despite advances in modern treat-ment found a reduced life expectancy in PD1 Similarly wefound that the survival of patients with PD was reduced incomparison with the general Swedish population during thesame years (SMR 158 p lt 0001) The SMR in most modernPD mortality studies has been in the range of 15 to 2712

Hence the risk of death in the present study was in the lowerrange This may be a result of the population-based studyprotocol (including patients with mild parkinsonism and explic-itly diagnosing cases of atypical parkinsonism that have a worseprognosis) and the fact that all patients had access to compre-hensive health care services throughout the disease course

It is currently unclear why many patients with PD experienceshorter lifespans In the present PD cohort the excess mor-tality occurred most clearly in female patients and in patientswith MCI (with pneumonia being the most common cause ofdeath) Data on the general Swedish population show longersurvival in women than in men Therefore a similar death rate(ie a hazard ratio close to 1) between the sexes in the presentstudy as well as in previous studies2728 may be interpreted toindicate a worse than expected prognosis for female patientswith PD In keeping with this some previous PD studies haveexplicitly reported a highermortality among female patients29ndash31

The fact that life expectancy is reduced in PD compared to thegeneral population independently from comorbidities32 andthe finding of a correlation between mortality and severity of

PD symptoms as measured by clinical scales3334 suggests thatdisease-specific features (such as α-synuclein pathology) atleast partly account for the increased mortality We found thatthe increased mortality in PD correlated with core parkinso-nian symptoms (with the notable exception of tremor) andolfactory dysfunction independently of age The elevatedmortality in PD also correlated with the severity of striatalDAT imaging deficits both in the putamen and in the caudateOur findings indicate that the neurodegenerative process inPD which is linked to nigrostriatal denervation is in itselfassociated with increased mortality Of note resting tremor inPD has been found to only weakly correlate with striatal do-paminergic denervation35

In contrast mortality was not increased in patients with PDwho did not have cognitive impairment at study entry (malepatients without MCI actually had an SMR below 1) Thefindings underscore the importance of an early PD-MCI di-agnosis as defined by the Movement Disorders SocietyThere is evidence that PD is associated with lower risk ofsome diseases such as many cancers and a lower risk factorburden from tobacco smoking and arterial hypertension36ndash39

In PD patients with normal cognition who have a milderdisease phenotype such differences could counterbalancemortality increases caused by neurodegeneration Further-more a lower level of education was not associated with in-creased mortality This may indicate that the higher mortalityin patients with poorer cognitive function was related to thepathologic process leading to cognitive dysfunction ratherthan being explained by socioeconomic factors In keepingwith this interpretation a few previous studies have founddementia in PD to be an independent risk factor for death340

Of note we found higher mortality in PD patients with moresevere caudal DAT-uptake deficits Imaging studies (includingan fMRI study of the same patients with PD as those in the

Table 3 Causes of death in parkinsonism in relation to clinical diagnosis

Variable PD MSA PSP p Value

Deathsno of participants 77143 1213 1618 lt0001

Mean age at death (SD) 830 (63) 783 (110) 800 (71) 0073

Mean survival years from baseline to death (SD) 65 (30) 47 (28) 48 (19) 0031

Pneumonia 15 (195) 3 (250) 7 (438) 0116

Dementia 2 (156) 0 0 0085

Unknown 9 (117) 0 2 (125) 0451

Heart infarction 7 (91) 1 (83) 1 (63) 0934

Cancer all types 6 (78) 2 (167) 1 (63) 0556

Heart failure 4 (52) 2 (167) 0 0159

Other causes 24 (312) 4 (333) 5 (313) 0989

Abbreviations MSA = multiple system atrophy PD = Parkinson disease PSP = progressive supranuclear palsyData are n () unless otherwise stated In 2 further patients with PD pneumoniawas determined to have contributed to death in combinationwith sepsis andrenal failure respectively Difference in cause of death was calculated by χ2 analysis

e2052 Neurology | Volume 91 Number 22 | November 27 2018 NeurologyorgN

Table 4 Early predictors of mortality in Parkinson disease

Variable at baseline Observed range

Adjusted only for age Multivariable clinical model

HR (95 CI) p Value HR (95 CI) p Value

Age 397 to 900 y 111 (107ndash116) lt0001

Duration since first symptom 02 to 75 y 093 (081ndash108) 0366 mdash mdash

Sex F vs M 103 (065ndash163) 0909 mdash mdash

UPDRS total score 8 to 81 103 (101ndash105) 0001c mdash mdash

UPDRS III subscore 5 to 62 103 (101ndash105) 0003c mdash mdash

PIGD score 0 to 12 261 (164ndash417) lt0001c 325 (175ndash605) lt0001

Tremor score 0 to 12 066 (032ndash135) 0255 mdash mdash

Freezinga 0 to 3 217 (140ndash336) lt0001c mdash mdash

Fallinga 0 to 2 090 (045ndash179) 0753 mdash mdash

Dysphagiaa 0 to 2 180 (102ndash321) 0044 mdash mdash

Body mass index 192 to 365 kgm2 101 (094ndash109) 0800 mdash mdash

Severity of hyposmia (correctanswers on B-SIT)

0 to 12 083 (075ndash092) lt0001c 084 (075ndash093) lt0001

Timed Up and Go test 40 to 410 s 111 (107ndash115) lt0001c mdash mdash

Previous smoker ever vs never 154 (094ndash251) 0081 mdash mdash

MADRS depression score 0 to 18 102 (095ndash109) 0529 mdash mdash

Years of education 6 to 30 101 (093ndash110) 0780 mdash mdash

Mild cognitive impairment yesno 238 (147ndash384) lt0001c 181 (104ndash314) 0035

Dementia within first 3 y yesno 194 (112ndash334) 0017c mdash mdash

Orthostatic blood pressure dropb minus26 to 77 mm Hg 101 (099ndash102) 0425 mdash mdash

BiomarkersBiomarkermodel

Age 436 to 868 y 117 (111ndash123) lt0001

APOE laquo4 genotype 0 to 2 e4 alleles 114 (071ndash183) 0593 mdash mdash

CSF Aβ42 pgmL 249 to 1373 ngL 099 (098ndash101) 0377 mdash mdash

CSF t-tau pgmL 81 to 1053 ngL 101 (099ndash102) 0240 mdash mdash

CSF p-tau pgmL 17 to 129 ngL 108 (095ndash123) 0255 mdash mdash

CSF α-synuclein pgmL 323 to 2179 ngL 102 (095ndash109) 0594 mdash mdash

Inflammatory reaction in CSF yesno 631(304ndash1312)

lt0001c 559 (267ndash1171) lt0001

CSF leukocytes mononuclear cells 0 to 10 (countμL) 128 (112ndash146) lt0001c mdash mdash

DAT uptake most affected putamen minus516 to minus038 060 (044ndash082) 0002c mdash mdash

DAT uptake most affected caudate minus472 to 014 067 (053ndash085) 0001c 081 (061ndash106) 0132

Abbreviations Aβ42 = β-amyloid 1ndash42 B-SIT = Brief Smell Identification Test CI = confidence interval DAT = dopamine active transporter HR = hazard ratioMADRS = Montgomery-Aringsberg Depression Rating Scale PIGD = postural imbalance and gait difficulty p-tau = phosphorylated tau t-tau = total tau UPDRS =Unified Parkinsonrsquos Disease Rating ScaleHRs formortality in patientswith Parkinson disease (n = 143) within 85 to 135 years Themiddle 2 columns showunivariate association but for clarity valuesare shown after adjustment for agea Assessed by UPDRSb The systolic pressure difference after 3 minutes of standing Inflammatory reaction in CSF is defined by a subtle pleocytosis (2ndash10 mononuclear cellsμL)c Significant prior to inclusion in the multivariable model

NeurologyorgN Neurology | Volume 91 Number 22 | November 27 2018 e2053

Figure 3 Survival in Parkinson disease in relation to phenotype

Kaplan-Meier plots of survival in patients with Parkinson disease (n = 143) in relation to clinical and neurobiological phenotype at baseline (except panel Bwhich is related to phenotype at 3 years) Severe hyposmia is defined by a B-SIT score lt4 All variables (AndashF) were significantly related to survival at the p lt0001 level (log rank) except the tremor or PIGDintermediate variable (C) which was significant at the p = 0004 level (for number at risk see supplementaltable e-2 linkslwwcomWNLA762) B-SIT = Brief Smell Identification Test PIGD = postural imbalance and gait disorder

e2054 Neurology | Volume 91 Number 22 | November 27 2018 NeurologyorgN

present study) have shown that dopamine denervation in thecaudate nucleus in PD correlates with cognitive deficits4142

A positive APOE e4 carrier status or α-synuclein Aβ42 andtotal tau or phosphorylated tau concentrations in CSF wereuncorrelated with PDmortality However the number of CSFsamples may have made our study underpowered fordetecting such differences The difference compared to thefindings in a recent neuropathologic study43 could also relateto the fact that the present investigation used a more homo-geneous disease group (PD) When CSF samples from allpatients with idiopathic parkinsonism were included a lowerAβ42 concentration was associated with shorter survival Thismay indicate that Alzheimer diseasendashtype pathology is a pre-dictor of reduced survival in parkinsonian disorders in gen-eral43 mainly in MSA andor PSP

The finding of a low-grade inflammatory reaction in the CSFof 131 of the patients with PD was strongly related toa reduced survival (with a 631 times increased hazard fordeath) An increase of proinflammatory cytokines in the CNSand an inflammatory hyperreactive state in circulating mon-ocytes has previously been shown in PD4445 Two largeobservational studies have also found a lower risk of PD as-sociated with the use of nonsteroidal anti-inflammatory drugsin the general population4647 Taken together our findingsmight suggest a triggered immune system responding to thepresence of abnormal misfolded proteins in PD patients withshort lifespans possibly contributing to disease progressionThe shorter lifespans in patients with a CSF leukocytosisindicate a rationale for further investigating immunomodu-lation to reduce PD mortality

The possibility of uncontrolled confounding factors and thefact that neuropathologic diagnosis at autopsy of the nervoussystem was obtained in only 5 of 109 deaths are limitations ofthe study Several studies have reported that neuropathologicconfirmation of a clinical diagnosis of idiopathic PD rangesfrom 65 to 9348 although the accuracy is higher in expertcenters49 In addition no autopsy examinations were per-formed in any of the MSA cases While acknowledging thelimitations of a clinical diagnosis the risk of incorrect di-agnosis was nonetheless minimized by the long follow-upperiods and the finding of pathologic uptake on DAT imagingin all of the examined patients Even so the results especiallythe CSF findings need to be confirmed by future studies Ourstudy also has several strengths including a population-baseddesign a high proportion of patients who were examinedusing a multimodal research protocol and prospective follow-up for up to 135 years

The present study shows that patients with incident parkin-sonism have reduced survival but that the survival is highlydependent on the type and characteristics of the parkinsoniandisorder Early MCI in PD is an important predictor of theprognosis The finding of a low-grade immune reaction in theCSF of patients with PD who have short survival may have

important clinical implications and therefore merits furtherinvestigation

Author contributionsDB JL and LF designed the study DB and GG conductedthe statistical analysis with the help of LF DB wrote the firstversion of the report and contributed to the data collection SJMand KR contributed to the data collection statistical analysisandwriting of the reportMED JL HZ andKB contributedto the data collection and writing of the report

AcknowledgmentMona Edstrom RN and Jorgen Andersson laboratorytechnician (Department of Pharmacology and Clinical Neuro-science Umearing University) provided valuable assistance Theauthors are grateful to all patients for their participation

Study fundingThe study was supported by grants from the Swedish MedicalResearch Council Erling-Persson Foundation Umearing Uni-versity Vasterbotten County Council King Gustaf V andQueen Victoria Freemason Foundation Swedish ParkinsonFoundation Kempe Foundation Swedish Parkinsonrsquos Dis-ease Association the Torsten Soderberg Foundation theSwedish Brain Foundation the European Research Counciland the Knut and Alice Wallenberg Foundation

DisclosureD Backstrom G Granaringsen and M Domellof report no dis-closures relevant to the manuscript J Linder reports receivinghonoraria for lectures from GSK Lundbeck Boehringer Ingel-heim Abbott AbbVie Solvay Orion Pharma UCB NordicInfuCare Medtronic and IPSEN and serving on advisoryboards for Boehringer Ingelheim Lundbeck and GSKS Jakobson Mo and K Riklund report no disclosures relevantto the manuscript H Zetterberg has served at advisory boardsfor Roche Diagnostics and Eli Lilly has received travel sup-port from Teva and is a cofounder of Brain Biomarker Sol-utions in Gothenburg AB a GU Ventures-based platformcompany at the University of Gothenburg K Blennow hasserved as a consultant or at advisory boards for Alzheon Bio-Arctic Biogen Eli Lilly Fujirebio Europe IBL InternationalMerck Novartis Pfizer and Roche Diagnostics and is a co-founder of Brain Biomarker Solutions inGothenburg AB a GUVenture-based platform company at the University of Goth-enburg L Forsgren reports no disclosures relevant to themanuscript Go to NeurologyorgN for full disclosures

Publication historyReceived by Neurology April 16 2018 Accepted in final formAugust 15 2018

References1 Macleod AD Taylor KS Counsell CE Mortality in Parkinsonrsquos disease a systematic

review and meta-analysis Mov Disord 2014291615ndash16222 de Lau LM Breteler MM Epidemiology of Parkinsonrsquos disease Lancet Neurol 2006

5525ndash5353 Levy G Tang MX Louis ED et al The association of incident dementia with mor-

tality in PD Neurology 2002591708ndash1713

NeurologyorgN Neurology | Volume 91 Number 22 | November 27 2018 e2055

4 De Pablo-Fernandez E Tur C Revesz T Lees AJ Holton JL Warner TT Associationof autonomic dysfunction with disease progression and survival in Parkinson diseaseJAMA Neurol 201774970ndash976

5 Litvan I Goldman JG Troster AI et al Diagnostic criteria for mild cognitive im-pairment in Parkinsonrsquos disease Movement Disorder Society Task Force guidelinesMov Disord 201227349ndash356

6 Linder J Stenlund H Forsgren L Incidence of Parkinsonrsquos disease and parkinsonismin northern Sweden a population-based study Mov Disord 201025341ndash348

7 Gibb WR Lees AJ The significance of the Lewy body in the diagnosis of idiopathicParkinsonrsquos disease Neuropathol Appl Neurobiol 19891527ndash44

8 Gilman S Low PA Quinn N et al Consensus statement on the diagnosis of multiplesystem atrophy J Neurol Sci 199916394ndash98

9 Litvan I Agid Y Calne D et al Clinical research criteria for the diagnosis of pro-gressive supranuclear palsy (Steele-Richardson-Olszewski syndrome) report of theNINDS-SPSP international workshop Neurology 1996471ndash9

10 Domellof ME Elgh E Forsgren L The relation between cognition and motor dys-function in drug-naive newly diagnosed patients with Parkinsonrsquos disease Mov Disord2011262183ndash2189

11 Jankovic J McDermott M Carter J et al Variable expression of Parkinsonrsquos diseasea base-line analysis of the DATATOP cohort The Parkinson Study Group Neurology1990401529ndash1534

12 Doty RL Marcus A Lee WW Development of the 12-item cross-cultural SmellIdentification Test (CC-SIT) Laryngoscope 1996106353ndash356

13 Podsiadlo D Richardson S The timed ldquoUp amp Gordquo a test of basic functional mobilityfor frail elderly persons J Am Geriatr Soc 199139142ndash148

14 Montgomery SA Asberg M A new depression scale designed to be sensitive tochange Br J Psychiatry 1979134382ndash389

15 Blennow K Wallin A Agren H et al Tau protein in cerebrospinal fluid a biochemicalmarker for axonal degeneration in Alzheimer disease Mol Chem Neuropathol 199526231ndash245

16 Vanmechelen E Vanderstichele H Davidsson P et al Quantification of tau phos-phorylated at threonine 181 in human cerebrospinal fluid a sandwich ELISA witha synthetic phosphopeptide for standardization Neurosci Lett 200028549ndash52

17 Andreasen N Hesse C Davidsson P et al Cerebrospinal fluid β-amyloid (1ndash42) inAlzheimer disease differences between early- and late-onset Alzheimer disease andstability during the course of disease Arch Neurol 199956673ndash680

18 Mo SJ Linder J Forsgren L et al Pre- and postsynaptic dopamine SPECT in the earlyphase of idiopathic parkinsonism a population-based study Eur J Nucl Med MolImaging 2010372154ndash2164

19 JakobsonMo S Larsson A Linder J et al 123I-FP-CIT and 123I-IBZM SPECT uptakein a prospective normal material analysed with two different semiquantitative imageevaluation tools Nucl Med Commun 201334978ndash989

20 Emre M Aarsland D Brown R et al Clinical diagnostic criteria for dementia asso-ciated with Parkinsonrsquos disease Mov Disord 2007221689ndash1707 quiz 1837

21 Yarnall AJ Breen DP Duncan GW et al Characterizing mild cognitive impairment inincident Parkinson disease the ICICLE-PD study Neurology 201482308ndash316

22 Hobson P Meara J Ishihara-Paul L The estimated life expectancy in a communitycohort of Parkinsonrsquos disease patients with and without dementia compared with theUK population J Neurol Neurosurg Psychiatry 2010811093ndash1098

23 Savica R Grossardt BR Bower JH et al Survival and causes of death among peoplewith clinically diagnosed synucleinopathies with parkinsonism a population-basedstudy JAMA Neurol 201774839ndash846

24 Nath U Ben-Shlomo Y Thomson RG et al Clinical features and natural history ofprogressive supranuclear palsy a clinical cohort study Neurology 200360910ndash916

25 Wenning GK Geser F Krismer F et al The natural history of multiple systematrophy a prospective European cohort study Lancet Neurol 201312264ndash274

26 Low PA Reich SG Jankovic J et al Natural history of multiple system atrophy in theUSA a prospective cohort study Lancet Neurol 201514710ndash719

27 Ben-Shlomo Y Marmot MG Survival and cause of death in a cohort of patients withparkinsonism possible clues to aetiology J Neurol Neurosurg Psychiatry 199558293ndash299

28 Roos RA Jongen JC van der Velde EA Clinical course of patients with idiopathicParkinsonrsquos disease Mov Disord 199611236ndash242

29 Hoehn MM Yahr MD Parkinsonism onset progression and mortality Neurology196717427ndash442

30 Diamond SG Markham CH Hoehn MM et al An examination of male-femaledifferences in progression and mortality of Parkinsonrsquos disease Neurology 199040763ndash766

31 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathicParkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184

32 Driver JA Kurth T Buring JE et al Parkinson disease and risk of mortality a pro-spective comorbidity-matched cohort study Neurology 2008701423ndash1430

33 Marras C McDermott MP Rochon PA et al Survival in Parkinson disease thirteen-year follow-up of the DATATOP cohort Neurology 20056487ndash93

34 Forsaa EB Larsen JP Wentzel-Larsen T Alves G What predicts mortality in Par-kinson disease A prospective population-based long-term study Neurology 2010751270ndash1276

35 Otsuka M Ichiya Y Kuwabara Y et al Differences in the reduced 18F-Dopa uptakesof the caudate and the putamen in Parkinsonrsquos disease correlations with the threemain symptoms J Neurol Sci 1996136169ndash173

36 Vanacore N Spila-Alegiani S Raschetti R Meco G Mortality cancer risk in parkin-sonian patients a population-based study Neurology 199952395ndash398

37 Inzelberg R Jankovic J Are Parkinson disease patients protected from some but notall cancers Neurology 2007691542ndash1550

38 Ritz B Ascherio A Checkoway H et al Pooled analysis of tobacco use and risk ofParkinson disease Arch Neurol 200764990ndash997

39 Noyce AJ Bestwick JP Silveira-Moriyama L et al Meta-analysis of early nonmotorfeatures and risk factors for Parkinson disease Ann Neurol 201272893ndash901

40 Willis AW Schootman M Kung N et al Predictors of survival in patients withParkinson disease Arch Neurol 201269601ndash607

41 Marie RM Barre L Dupuy B et al Relationships between striatal dopamine denervationand frontal executive tests in Parkinsonrsquos disease Neurosci Lett 199926077ndash80

42 Ekman U Eriksson J Forsgren L et al Functional brain activity and presynapticdopamine uptake in patients with Parkinsonrsquos disease and mild cognitive impairmenta cross-sectional study Lancet Neurol 201211679ndash687

43 Irwin DJ Grossman M Weintraub D et al Neuropathological and genetic correlatesof survival and dementia onset in synucleinopathies a retrospective analysis LancetNeurol 20171655ndash65

44 Baufeld C OrsquoLoughlin E Calcagno N et al Differential contribution of microglia andmonocytes in neurodegenerative diseases J Neural Transm 2018125809ndash826

45 Grozdanov V Bliederhaeuser C Ruf WP et al Inflammatory dysregulation of bloodmonocytes in Parkinsonrsquos disease patients Acta Neuropathol 2014128651ndash663

46 Chen H Zhang SM Hernan MA et al Nonsteroidal anti-inflammatory drugs and therisk of Parkinson disease Arch Neurol 2003601059ndash1064

47 Chen H Jacobs E Schwarzschild MA et al Nonsteroidal antiinflammatory drug useand the risk for Parkinsonrsquos disease Ann Neurol 200558963ndash967

48 Adler CH Beach TG Hentz JG et al Low clinical diagnostic accuracy of early vsadvanced Parkinson disease clinicopathologic study Neurology 201483406ndash412

49 Hughes AJ Daniel SE Ben-Shlomo Y Lees AJ The accuracy of diagnosis of par-kinsonian syndromes in a specialist movement disorder service Brain 2002125861ndash870

e2056 Neurology | Volume 91 Number 22 | November 27 2018 NeurologyorgN

DOI 101212WNL0000000000006576201891e2045-e2056 Published Online before print October 31 2018Neurology

David Baumlckstroumlm Gabriel Granaringsen Magdalena Eriksson Domelloumlf et al population-based study

Early predictors of mortality in parkinsonism and Parkinson disease A

This information is current as of October 31 2018

ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology

ServicesUpdated Information amp

httpnneurologyorgcontent9122e2045fullincluding high resolution figures can be found at

Supplementary Material

576DC1httpnneurologyorgcontentsuppl20200304WNL0000000000006Supplementary material can be found at

References httpnneurologyorgcontent9122e2045fullref-list-1

This article cites 49 articles 17 of which you can access for free at

Citations httpnneurologyorgcontent9122e2045fullotherarticles

This article has been cited by 5 HighWire-hosted articles

Subspecialty Collections

httpnneurologyorgcgicollectionprogressive_supranuclear_palsyProgressive supranuclear palsy

httpnneurologyorgcgicollectionparkinsons_disease_parkinsonismParkinsons diseaseParkinsonism

httpnneurologyorgcgicollectionnatural_history_studies_prognosisNatural history studies (prognosis)

httpnneurologyorgcgicollectionmultiple_system_atrophyMultiple system atrophy

httpnneurologyorgcgicollectionall_imagingAll Imagingfollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online

ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology

ServicesUpdated Information amp

httpnneurologyorgcontent9122e2045fullincluding high resolution figures can be found at

Supplementary Material

576DC1httpnneurologyorgcontentsuppl20200304WNL0000000000006Supplementary material can be found at

References httpnneurologyorgcontent9122e2045fullref-list-1

This article cites 49 articles 17 of which you can access for free at

Citations httpnneurologyorgcontent9122e2045fullotherarticles

This article has been cited by 5 HighWire-hosted articles

Subspecialty Collections

httpnneurologyorgcgicollectionprogressive_supranuclear_palsyProgressive supranuclear palsy

httpnneurologyorgcgicollectionparkinsons_disease_parkinsonismParkinsons diseaseParkinsonism

httpnneurologyorgcgicollectionnatural_history_studies_prognosisNatural history studies (prognosis)

httpnneurologyorgcgicollectionmultiple_system_atrophyMultiple system atrophy

httpnneurologyorgcgicollectionall_imagingAll Imagingfollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online

ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2018 The Author(s) Published by

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology