Statistical Models of Parkinson’s Disease Progression: Predictive Validity in a 3-Year Follow-up

Abstract

Background: The rate of Parkinson’s disease (PD) progression varies widely between patients. Current knowledge does not allow to accurately predict the evolution of symptoms in a given individual over time.

Objectives: To develop regression-based models of PD progression and to explore its predictive value in a three-year follow-up.

Methods: At baseline, 300 consecutive PD patients were assessed using the Unified Parkinson’s Disease Rating Scale (UPDRS) - subscales II and III, Hoehn & Yahr (H&Y) and Schwab and England Independence Scale (S&E); and the Freezing of Gait Questionnaire (FOG-Q). UPDRS-III and H&Y were applied in OFF and ON medication conditions. An axial index was derived from the UPDRS-III. Based on multiple linear regression coefficients, algorithms were developed to adjust test scores to the characteristics of each individual. Sixty-eight patients were reevaluated three years later.

Results: In the construction of the models, disease duration, age ≥70, age at disease onset ≥55, tremor as the first symptom alone, and medication description explained between 35% (UPDRS-III in ON) and 57% (axial index in ON) of the variance of test scores. The predictive r² of the models in a 10-fold cross-validation ranged between 33% (UPDRS-III in ON) and 55% (axial index in ON and S&E in OFF). All measures, except UPDRS-III OFF, H&Y ON, and S&E ON, had moderate/good absolute agreement (intraclass correlation coefficient between 0.60 and 0.72) between baseline and follow-up.

Conclusions: A cross-sectional assessment of a PD population allowed the development of models of disease progression, whose predictive value was validated on a three-year longitudinal study.

Keywords

Parkinson disease disease progression statistical models validity of results

INTRODUCTION

Parkinson’s disease (PD) is a heterogeneous disease [1 –3], with between-patients’ variability in progression of motor and non-motor symptoms. A better understanding of how PD motor symptoms progress at the individual level may help patients make informed decisions about their personal and professional lives and would support clinicians in treatment decisions and in the design of clinical trials [4]. One of the issues when proposing Subthalamic Nucleus – Deep Brain Stimulation (STN-DBS) in the early stages of PD is the current inability to predict the rate of disease progression on an individual basis, which may lead to selection inaccuracies [5].

The current knowledge on the progression of PD without the effects of medication is limited. Before the emergence of levodopa in the late sixties [6], no validated clinical scales were available [7]. Since the widespread use of levodopa, the long follow-up of drug naïve patients has become impossible. In the last decades, drug against placebo studies [8 –22] have explored disease progression in non-treated PD patients, but only in early stages of the disease and with limited time of follow-up. A recent study analysed motor complications associated with dopaminergic treatment in a series of previously untreated Ghanaian patients with disease duration of 4 years [23]. This interesting study did not explore however in detail the progression of the disease. Important methodological issues have also been identified in the literature of treated patients, such as, the wide variety of outcome measures and the diversity of sampling procedures [4 , 25]. Longitudinal studies in PD are afflicted by attrition and mortality. Furthermore, interpretation of data in the very long follow up studies is difficult confronted with the simple impact of ageing and above all the high rates of comorbidities interfering with evaluation of symptoms, function, and autonomy [1 , 26–36].

It is currently accepted that progression of motor symptoms in PD on dopaminergic treatment does not follow a linear path. At least for some symptoms, progression appears to be faster in the first years of the disease [37 –39] and in the late stages of the disease motor and non-motor symptoms resistant to treatment become prevalent [27, 32]. There is evidence that some factors are prognostic to the progression of the disease, such as, age at PD onset [2 , 40–58], age [37 , 59–62], phenotype with tremor dominant motor symptoms [3 , 63], phenotype with predominance of postural instability and gait disturbance (PIGD) [2 , 64] and tremor as the first symptom [7 , 65]. The predictive effect of sex [35 , 64], depressed mood [54, 66], and cognitive decline [24, 35] in disease progression is less clear.

A key issue in the study of PD progression is: how to reliably measure severity and progression of PD. The assessment instruments ought to have a strong association with disease duration [67] and the assessment conditions need to be well defined, especially in treated patients with motor fluctuations. The non-linearity of symptoms progression should also be taken into account. Furthermore, it is important to measure not only motor disability but also those non-motor symptoms, known to be the main drivers of poorer quality of life, such as anxiety and depression [38 , 69].

The main aim of this study was to modulate the progression of PD, based on cross-sectional associations between patients’ demographic and clinical characteristics and motor and non-motor symptoms of PD. We hypothesized that these statistical models for a series of clinical scales and indexes would show good predictive value in a 10-fold cross-validation analysis and in a three-year follow-up. We also hypothesized that the relative position of a patient in relation to his/her peers with similar demographic and clinical features is somewhat stable throughout the disease, even though motor and non-motor symptoms may differ from patient to patient and from time to time.

MATERIAL AND METHODS

Participants

A comprehensive observational study was proposed, between November 2012 and September 2014, to 322 consecutive patients with diagnosis of PD from Centro Hospitalar do Porto’s Movement Disorders outpatient clinic. PD was defined according to the United Kingdom Brain Bank criteria [70]. Exposure to neuroleptics, vascular parkinsonism, and possible or probable atypical parkinsonian syndromes were considered a priori exclusion criteria. From the 322 possible subjects, only one refused to participate in the study. From those included (n = 321), 6 died before assessment (2 from respiratory infection, 1 from heart disease, 2 were bedridden with advanced PD, and 1 patient from unknown cause), 2 were excluded because they developed other debilitating conditions, and 13 moved geographically to a region not dependent from our center or could not be reached between inclusion and assessment. Three hundred PD patients participated in the cross-sectional part of the study (Table 1).

A three-year follow-up was attempted to the first 88 patients observed in the cross-sectional study. However, 20 were lost to follow-up: 5 died (2 fromrespiratory infection, 1 from pulmonary cancer, 1 from heart disease, and 1 from unknown cause), 2 were submitted to STN-DBS surgery, 1 had an active psychosis, 1 was not off medication for 12 h on the day of the evaluation, 5 could not be reached, and 6 abandoned the Movement Disorders’ outpatient clinic. The 68 patients with follow-up assessment (Table 1) and the 20 patients lost to follow-up had similar (p > 0.05) demographic (sex and age) and clinical characteristics (age at disease onset, disease duration, and medication description) at the cross-sectional assessment.

The local ethical committee approved the study and all participants (or legal representatives) provided their informed consent.

Procedures

PD patients were evaluated by movement disorders specialists in the morning without anti-parkinsonian medication for 12 hours, the practically defined OFF [71], using the Unified Parkinson’s Disease Rating Scale [72] subscale III (UPDRS-III), and the modified Hoehn & Yahr scale (H&Y) [7]. After the assessment in OFF, patients took their usual first dose of anti-parkinsonian medication and were re-evaluated one hour later, using the same instruments. Then, UPDRS subscale for activities of daily living (UPDRS-II), the Schwab and England Independence Scale (S&E) [73] and the freezing of gait questionnaire (FOG-Q) [74], were applied. Thirty patients (10%) from the cross-sectional study and four (6%) from the follow-up study were evaluated at home due to the severity of their motor symptoms.

Based on UPDRS-III OFF and ON, an axial index (i.e., sum of items 18, 27, 28, 29, and 30) was calculated.

Based on interview and patient records, the following data were collected: gender, age, age at disease onset (i.e., age at the first subjective motor symptom), first motor symptom (bradykinesia, tremor alone or in conjunction with other symptoms, gait disorder or other), family history of PD or essential tremor, current anti-parkinsonian medication, and ability of the patient to describe all medication schedules [75, 76]. Patients who reported tremor as the first subjective symptom alone, also had bradykinesia when observed, fulfilling the United Kingdom Brain Bank criteria for PD diagnosis. Current anti-parkinsonian medication was converted to levodopa equivalent dose (LED) [77]. Age and age at disease onset were dichotomized (respectively, <70 vs. ≥70 and <55 vs. ≥55). The relationship between age and disability, and between age and cognitive decline, may be particularly important over 70 years [61 , 78]. Guided by recent references [54 , 79–81], 55 years was used as the cut-off age for late versus young-onsetPD.

Statistical analysis

Patients’ demographic and clinical characteristics were summarized as percentages for categorical variables and means (standard deviations) for continuous ones. Based on the longitudinal subgroup data, an annual rate of clinical progression was calculated for each scale and index. Baseline and follow-up raw scores were compared using paired t-tests.

The association between certain demographic (sex and age) and clinical (duration, age at disease onset, tremor as first symptom, tremor as first symptom alone and medication description) variables with the cross-sectional test scores (UPDRS-II, UPDRS-III, axial index, H&Y, and S&E in OFF and ON conditions, and FOG-Q) was analyzed using simple linear regressions. The same test scores were then modeled using multiple linear regressions. Covariates significantly associated with the test score in the univariate analyses were initially considered in the multivariable models (age >70, age at disease onset >55, disease duration, tremor as first symptom alone, and patients’ capacity to describe their medication, i.e., medication description). Then backward selection was used to construct the final model for each test-score. Independent variables with p-values higher than 0.1 were removed from the models.

All multiple linear regression analyses, except for H&Y and S&E, were conducted with the scores square-root transformed to stabilize the variance and improve the homoscedasticity. Also, the association of disease duration with the scores was found to be non-linear, therefore a quadratic and cubic effect were allowed in the model for this variable.

Based on the regression coefficients (Tables 3a and 3b), algorithms were developed to adjust test scores (Supplementary Table 2). These algorithms convert “raw” into standardized Z scores, with mean 0 and standard deviation 1. For all adjusted test scores, lower Z scores correspond to increased severity of symptoms. The Z scores represent the difference between the score of an individual and the mean score of individuals with the same demographic and clinical characteristics (i.e., variables included in the regression model). The percentiles associated with the adjusted score are shown in Supplementary Table 3. These common metrics were applied to facilitate intra-individual and inter-individual comparisons.

To study the predictive value of the models, we computed the predictive r² using a 10-fold cross-validation for each model. The regression-based algorithms developed in the cross-sectional study (Supplementary Table 2) were then applied to adjust test scores at baseline and three-years follow-up of the longitudinal sample. The intraclass correlation coefficient (ICC) was used to explore the absolute agreement between adjusted scores at baseline and at follow-up.

For all statistical tests we used 0.05 as the significance level. The statistical analysis was conducted using SPSS and R.

RESULTS

Cross-sectional study

The mean test scores are presented in Table 2. Simple linear regression analyses showed significant associations (p < 0.05) with age ≥70, age at disease onset ≥55, disease duration, tremor as first symptom alone, and medication description (Supplementary Tables 1a and b). Disease duration, age ≥70, and medication description were significantly related with all test scores. No significant associations were found with sex and tremor as first symptom (p > 0.05).

The multiple linear regression model (age≥70, age at onset ≥55, disease duration, tremor as first symptom alone, and medication description) partly explained the variance of all test scores (Tables 3a and 3b). The adjusted r² for OFF and ON conditions was respectively: 50% and 42% for UPDRS-II, 41% and 35% for UPDRS-III, 55% and 57% for the axial index, 52% and 54% for H&Y, and 56% and 51% for S&E. The variance of FOG-Q explained by the model was 49%.

The multiple linear regression analyses revealed that disease duration and medication description were associated with all assessment measures, even when taking into account covariates. Non-linear associations were found between disease duration and all measures except H&Y in ON condition. Patients that were unable to describe their medication had worse test scores and indexes in all assessment measures. Patients with older age at examination also had worse test scores in all assessment measures except for the H&Y in OFF condition. Older age at disease onset was associated with poorer functioning in daily life activities (UPDRS-II in ON condition) and with increased motor symptom severity (axial index and H&Y in OFF and ON conditions). Patients with tremor as first symptom alone had less severe motor symptoms (axial index in OFF and ON) and better functioning in activities of daily living (UPDRS-II and S&E in OFF condition).

Cross-sectional validation

The cross-validation r² ranged between 0.33 to 0.55. The regression models with the best predictive ability were: axial index ON (55%), S&E OFF (55%), axial index OFF (54%), H&Y OFF (51%), H&Y ON (50%), UPDRS-II OFF (49%), S&E ON (48%), and FOG-Q (48%). The regression models with the poorest predictive ability were: UPDRS-III ON (33%), UPDRS-II ON (39%), and UPDRS-III OFF (40%).

Longitudinal study

The interval between baseline and follow-up was 35.9 months (sd = 2.1). The mean raw scores at baseline and follow-up and the mean annual rate of clinical progression of each measure are presented in Table 2. All test scores were worse at three years compared to baseline.

The longitudinal reliability of the adjusted scores is presented in Table 4. The ICC of the adjusted scores for measures UPDRS-II OFF and ON, UPDRS-III ON, axial index OFF and ON, H&Y OFF, S&E OFF, and FOG-Q ranged between 0.60 and 0.72. For measures UPDRS-III OFF, H&Y ON, and S&E ON, the ICC varied between 0.28 and 0.52.

DISCUSSION

PD progression was investigated in a series of eleven clinical measures, covering motor symptoms and functional disability. Statistical models of PD progression were developed based on a cross-sectional evaluation of 300 consecutive PD patients without relevant selection bias. The predictive value of the models was demonstrated in a three-year follow-up study of 68 patients.

As expected, disease duration was strongly associated with all assessment scores and indexes. The observed quadratic and cubic effects of disease duration in most measures may reflect a nonlinear progression of motor symptoms and functional disability [35 , 82].

Patients with age at disease onset below 55 had less severe motor symptoms (axial index and H&Y in OFF and ON conditions) and better functioning in activities of daily living (UPDRS-II in ON condition), even when disease duration and other covariates were taken into account. These findings are in agreement with clinical studies that have encountered a slow progression of motor symptoms when PD begins at younger age [28 , 54]. Older age at disease onset has been consistently associated with greater postural instability [3 , 51] and cognitive decline [3 , 83].

Patients 70 or older had poorer scores in almost all measures, even when disease duration and other variables were taken into account, which suggests an independent effect of age in symptom severity and functional disability. These results support the notion put forward by other clinical and pathological studies that age and disease duration have additive and independent effects in the clinical presentation of the disease [37 , 84].

Patients with tremor as first subjective symptom alone appear to have a more benign progression of the disease. These patients had less severe motor symptoms (axial index in OFF and ON conditions), better functioning in activities of daily living (UPDRS-II in OFF condition) and slower progression of dependence (S&E in OFF condition), when disease duration and other covariates were taken into account. This association between tremor as first symptom alone and less severity of the disease is consistent with other reports [7, 65]. Though, no clear association was found between tremor as the first symptom in conjunction with other symptoms and the assessment measures.

The inability to describe medication schedules is considered a sensitive indicator of cognitive impairment [75]. In our study, incorrect medication description was associated with poorer scores in all measures. These findings are consistent with the known association between cognitive decline and severity of motor symptoms in PD [3 , 85].

No significant associations were found between sex and test scores. Our study did not confirm reports in the literature that suggest that male patients may have greater progression of symptoms [58, 64] or different progression of motor scores during different phases of the disease [35].

In general, the statistical models of PD progression were stable in a three-year follow-up. The measures with greatest temporal stability (i.e., measures with highest absolute agreement in the relative position of a patient in relation to his/her peers with similar demographic and clinical features between baseline and follow-up) were: UPDRS-II OFF, axial index OFF and ON, S&E OFF, and FOG-Q. These measures also had good predictive validity in a cross-validation analysis. The good predictability of the model for axial symptoms is remarkable given its great variability between subjects and progression with disease duration [86]. Another longitudinal study [87], with a 4-year follow-up, also suggests a continuous progression of the axial symptoms throughout the disease, whereas other symptoms (e.g., rigidity, tremor, and bradykinesia) appear to have a variable progression. These findings point to the importance of exploring the axial symptoms when studying PD progression. Noteworthy, most longitudinal reports to date did not explore axial symptoms on the UPDRS-III scale[1 , 88].

The UPDRS-III had the lowest and the axial index had the highest variance explained by the model in the cross-sectional assessment in ON condition. In the cross-validation analysis, UPDRS-III in ON also had the poorest predictive value, even though the axial index in ON had the best predictive value. In the longitudinal analyses, UPDRS-III in the OFF condition had the poorest stability of the model, whereas the axial index in OFF had the highest frequency of patients scoring as expected at follow-up. Another study [67] also found a poor association between disease duration and UPDRS-III scores in ON medication condition. Though the authors did not isolate the axial symptoms from the motor subscale of UPDRS. This set of findings point to a complex association between disease duration and UPDRS-III scores, especially those that reflect non-axial symptoms. Also, UPDRS-III non-axial symptoms may be more vulnerable than axial symptoms to random variation, to measurement error, or to systematic variation associated with other independent variables not considered in this study.

The axial index represents disturbances in speech, rising from siting, posture, postural stability, and gait. These symptoms are known to gradually appear with disease progression and to develop with decreasing response to levodopa over time, representing progressive pathology in extranigral sites in the brain [39 , 89–91]. Freezing of gait is another closely related symptom that affects patients in more advanced stages of the disease [92]. These symptoms are usually associated with disability and poor quality of life [27 , 93]; and ought to be taken into account when studying disease progression and when defining endpoints in neuroprotective studies.

Based on the longitudinal data, the annual rate of clinical progression in a three-year follow up was calculated for each measure. The mean rates of progression observed in our cohort varied between 1.2% and 3.8% depending on the measure. These results fall within the wide range of rates of clinical progression per year reported by other longitudinal studies with comparable disease durations [28 , 94]. The diversity of findings in the literature may be partly explained by methodological and sample differences. The considerable inter-individual variability in rates of progression (as demonstrated by the large standard deviations of the percentage of progression) limits the clinical usefulness of the mean rates.

The consecutive nature of the recruitment and the extremely high acceptance rate of the study by patients followed in our center decreased potential selection biases. A hospital-based sample may not be perfectly representative of the whole population of PD. However, our Movement Disorders Outpatient Clinic, as part of the National Health Service with universal coverage, receives PD patients in all stages of the disease and provides long-term follow-up of these patients. Moreover, in this study, to have a representative sample of the whole population and to avoid losses at baseline and at follow-up as much as possible, patients with severe motor symptoms in the OFF-drug condition, rendering difficult an evaluation at the hospital, were evaluated at home. An important limitation associated with cross-sectional assessment of patients with PD is the possibility of misdiagnosis in the initial phases of the disease, namely atypical parkinsonian syndromes. However, this caveat is likely to have had only minor impact in the present study, given that all patients were evaluated by neurologists in a single center specialized in PD, the large population size, and the mean disease duration of 9 years. The wide range of age at disease onset, age at examination, and disease duration is a strength of the study. The inclusion of patients in different stages of the disease may facilitate the translation of the study results into clinical practice.

To control for the confounder effect of motor fluctuations on motor evaluations, patients were assessed in a well-defined motor condition in relation to the anti-parkinsonian medication. Both UPDRS-III and the H&Y were applied in OFF and ON conditions to each patient in the same day. Even though 12 hours without anti-parkinsonian medication may not represent a true untreated condition, due to the long-duration effect of levodopa and dopamine agonists [95 –97], the OFF medication interval of 12 hours was uniformly applied to all subjects. Longer intervals are deemed impracticable for the more severe patients. We used the habitual morning dose of medication for the ON state, which may not represent the best ON state for some patients. However, it represents the habitual ON in the morning for each patient. This dual medication condition protocol is an important advantage in relation to previous studies, that explored only one condition [1 , 90] or failed to differentiate between conditions [28 , 94]. Another strength of this prospective study is the patients’ assessment by movement disorders specialists.

Ideally the longitudinal study would have included all patients observed in the cross-sectional study. However, due to logistic limitations, we could only target the first ∼30% of participants at three-year follow-up. From these 88 patients, 20 (23%) were lost to follow up. This percentage of attrition at three years is similar to other longitudinal studies with comparable time intervals [28 , 93]. We acknowledge a survivor bias in our longitudinal study. This limitation could have been minimized if patients at baseline were in the early stages of PD. Though, such approach would have sacrificed the representativeness of the sample and would have overlooked the progression of symptoms in later stages of the disease. Noteworthy, in our study, patients lost to follow-up had similar demographic and clinical characteristics in the cross-sectional assessment to those evaluated at three years.

In sum, we proposed algorithms to modulate the progression of PD in a series of clinical scales and indexes, based on a cross-sectional evaluation of 300 PD patients in OFF and ON conditions. We then estimated how well the regression-based models generalize to an independent dataset (cross-sectional validation) and confirmed the stability of the models over time (longitudinal study). The study results show that the proposed methods can be employed in longitudinal modeling of PD and may be used as a blueprint for exploring the impact of existing and new therapies. This approach can be advantageous in the identification of significant deviations from expected with disease progression, for instance those resultant from STN-DBS. The estimation of the effects of STN-DBS in advanced PD is quite challenging, due to the progressive nature of the disease and the ethical constraints associated with a long-term placebo condition. Data from this and from other disease progression studies also provide an important support for clinical practice, namely, when counselling individual patients and when making therapeutic decisions [5, 98].

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

Footnotes

ACKNOWLEDGMENTS

The study received financial support from: Centro Hospitalar do Porto’s Department of Teaching, Education, and Research; the Portuguese National Funding Agency for Science, Research and Technology (FCT) PEST-OE/SAU/UI0215/2011.

Appendix

References

Hely

, Morris

JGL

, Traficante

, Reid

WGJ

, O’Sullivan

, & Williamson

(1999) The Sydney multicentre study of Parkinson’s disease: Progression and mortality at 10 years. J Neurol Neurosurg Psychiatry, 67, 300–307.

Auyeung

, Tsoi

, Mok

, Cheung

, Lee

, Li

, & Yeung

(2012) Ten year survival and outcomes in a prospective cohort of new onset Chinese Parkinson’s disease patients. J Neurol Neurosurg Psychiatry, 83, 607–611.

Williams-Gray

, Mason

, Evans

, Foltynie

, Brayne

, Robbins

, & Barker

(2013) The CamPaIGN study of Parkinson’s disease: 10-year outlook in an incident population-based cohort. J Neurol Neurosurg Psychiatry, 84, 1258–1264.

Post

, Merkus

, Haan

, & Speelman

On behalf of the CARPA Study Group (2007) Prognostic factors for the progression of Parkinson’s disease: A systematic review. Mov Disord, 22, 1839–1851.

Mestre

, Espay

, Marras

, Eckman

, Pollak

, & Lang

(2014) Subthalamic nucleus-deep brain stimulation for early motor complications in Parkinson’s disease - the EARLYSTIM trial: Early is not always better. Mov Disord, 29, 1751–1756.

Cotzias

, Van Woert

, & Schiffer

(1967) Aromatic amino acids and modification of parkinsonism. N Engl J Med, 276, 374–379.

Hoehn

, & Yahr

(1967) Parkinsonism: Onset, progression, and mortality. Neurology, 17, S11–S26.

The Parkinson Study, Group (1993) Effects of tocopherol and deprenyl on the progression of disability in earlyParkinson’s disease. N Engl J Med, 328, 176–183.

Olanow

, Hauser

, Gauger

BSC

, Malapira

, Koller

, Hubble

, Bushenbark

, Lilienfeld

, & Esterlitz

(1995) The effect of deprenyl and levodopa on the progression of Parkinson’s disease. Ann Neurol, 38, 771–777.

10.

The Parkinson Study Group (1996) Effect of lazabemide on the progression of disability in early Parkinson’s disease. Ann Neurol, 40, 99–107.

11.

Shults

, Oakes

, Kieburtz

, Beal

, Haas

, Plumb

, Juncos

, Nutt

, Shoulson

, Carter

, Kompoliti

, Perlmutter

, Reich

, Stern

, Watts

, Kurlan

, Molho

, Harrison

, & Lew

, Parkinson Study Group (2002) Effects of coenzyme Q10 in early Parkinson disease: Evidence of slowing of the functional decline. Arch Neurol, 59, 1541–1550.

12.

The Parkinson Study Group (2004) Levodopa and the progression of Parkinson’s disease. N Engl J Med, 351, 2498–2508.

13.

The Parkinson Study Group (2004) A controlled, randomized, delayed-start study of rasagiline in early Parkinson disease. Arch Neurol, 61, 561–566.

14.

Olanow

, Schapira

, LeWitt

, Kieburtz

, Sauer

, Olivieri

, Pohlmann

, & Hubble

(2006) TCH346 as a neuroprotective drug in Parkinson’s disease: A double-blind, randomised, controlled trial. Lancet Neurol, 5, 1013–1020.

15.

Giladi

, Boroojerdi

, Korczyn

, Burn

, Clarke

, & Schapira

(2007) SP513 investigators, Rotigotine transdermal patch in early Parkinson’s disease: A randomized, double-blind, controlled study versus placebo and ropinirole. Mov Disord, 22, 2398–2404.

16.

Jankovic

, Watts

, Martin

, & Boroojerdi

(2007) Transdermal rotigotine: Double-blind, placebo-controlled trial in Parkinson disease. Arch Neurol, 64, 676–682.

17.

Olanow

, Rascol

, Hauser

, Feigin

, Jankovic

, Lang

, Langston

, Melamed

, Poewe

, Stocchi

, & Tolosa

(2009) ADAGIO Study Investigators, A double-blind, delayed-start trial of rasagiline in Parkinson’s disease. N Engl J Med, 361, 1268–1278.

18.

Snow

, Rolfe

, Lockhart

, Frampton

, O’Sullivan

, Fung

, Smith

, Murphy

, & Taylor

(2010) Protect Study Group, A double-blind, placebo-controlled study to assess the mithochondria-targeted antioxidant MitoQ as a disease-modifying therapy in Parkinson’s disease. Mov Disord, 25, 1670–1674.

19.

Hauser

, Schapira

, Rascol

, Barone

, Mizuno

, Salin

, Haaksma

, Juhel

, & Poewe

(2010) Randomized, double-blind, multicentre evaluation of pramipexole extended release once daily in early Parkinson’s disease. Mov Disord, 25, 2542–2549.

20.

Poewe

, Rascol

, Barone

, Hauser

, Mizuno

, Haaksma

, Salin

, Juhel

, & Schapira

Pramipexole ER Studies Group (2011) Extended-release pramipexole in early Parkinson Disease: A 33-week randomized controlled trial. Neurology, 77, 759–766.

21.

Schapira

, McDermott

, Barone

, Comella

, Albrecht

, Hsu

, Massey

, Mizuno

, Poewe

, Rascol

, & Marek

(2013) Pramipexole in patients with early Parkinson’s disease (PROUD): A randomized delayed-start trial. Lancet Neurol, 12, 747–755.

22.

Beal

, Oakes

, Shoulson

, Henchcliffe

, Galpern

, Haas

, Juncos

, Nutt

, Voss

, Ravina

, Shults

, Helles

, Snively

, Lew

, Griebner

, Watts

, Gao

, Pourcher

, Bond

, Kompoliti

, Agarwal

, Sia

, Jog

, Cole

, Sultana

, Kurlan

, Richard

, Deeley

, Waters

, Figueroa

, Arkun

, Brodsky

, Ondo

, Hunter

, Jimenez-Shahed

, Palao

, Miyasaki

, So

, Tetrud

, Reys

, Smith

, Singer

, Blenke

, Russell

, Cotto

, Friedman

, Lannon

, Zhang

, Drasby

, Kumar

, Subramanian

, Ford

, Grimes

, Cote

, Conway

, Siderowf

, Evatt

, Sommerfeld

, Lieberman

, Okun

, Rodriguez

, Merritt

, Swartz

, Martin

, King

, Stover

, Guthrie

, Watts

, Ahmed

, Fernandez

, Winters

, Mari

, Dawson

, Dunlop

, Feigin

, Shannon

, Nirenberg

, Ogg

, Ellias

, Thomas

, Frei

, Bodis-Wollner

, Glazman

, Mayer

, Hauser

, Pahwa

, Langhammer

, Ranawaya

, Derwent

, Sethi

, Farrow

, Prakash

, Litvan

, Robinson

, Sahay

, Gartner

, Hinson

, Markind

, Pelikan

, Perlmutter

, Hartlein

, Molho

, Evans

, Adler

, Duffy

, Lind

, Elmer

, Davis

, Spears

, Wilson

, Leehey

, Hermanowicz

, Niswonger

, Shill

, Obradov

, Rajput

, Cowper

, Lessig

, Song

, Fontaine

, Zadikoff

, Williams

, Blindauer

, Bergholte

, Propsom

, Stacy

, Field

, Mihaila

, Chilton

, Uc

, Sieren

, Simon

, Kraics

, Silver

, Boyd

, Hamill

, Ingvoldstad

, Young

, Thomas

, Kostyk

, Wojcieszek

, Pfeiffer

, Panisset

, Beland

, Reich

, Cines

, Zappala

, Rivest

, Zweig

, Lumina

, Hilliard

, Grill

, Kellermann

, Tuite

, Rolandelli

, Kang

, Young

, Rao

, Cook

, Severt

, & Boyar

(2014) Parkinson Study Group QE3 investigators, A randomized clinical trial of high-dosage coenzyme Q10 in early Parkinson disease: No evidence of benefit. JAMA Neurol, 71, 543–552.

23.

Cilia

, Akpalu

, Sarfo

, Cham

, Amboni

, Cereda

, Fabbri

, Adjei

, Akassi

, Bonetti

, & Pezzoli

(2014) The modern pre-levodopa era of Parkinson’s disease: Insights into motor complications from sub-Saharan Africa. Brain, 137, 2731–2742.

24.

Marras

, Rochon

, & Lang

(2002) Predicting motor decline and disability in Parkinson disease. Arch Neurol, 59, 1724–1728.

25.

Lerche

, Liepelt-Scarfone

, Alves

, Barone

, Behnke

, Ben-Shlomo

, Berendse

, Burn

, Dodel

, Grosset

, Heinzel

, Hu

, Kasten

, Krüger

, Maetzler

, Moccia

, Mollenhauer

, Oertel

, Roeben

, Sünkel

, Walter

, Wirdefeldt

, & Berg

(2015) Methods in neuroepidemiology characterization of european longitudinal cohort studies in Parkinson’ disease – report of the JPND working group BioLoC-PD. Neuroepidemiology, 45, 282–297.

26.

Aarsland

, Andersen

, Larsen

, Lolk

, Kragh-Sørensen

(2003) Prevalence and characteristics of dementia in Parkinson disease. An 8-year prospective study. Arch Neurol, 60, 387–392.

27.

Hely

, Morris

JGL

, Reid

WGJ

, & Trafficante

(2005) Sydney multicenter study of Parkinson’s disease: Non-L-Dopa-respondive problems dominate at 15 years. Mov Disord, 20, 190–199.

28.

Alves

, Wentzel-Larsen

, Aarsland

, & Larsen

(2005) Progression of motor impairment and disability in Parkinson disease. A population-based study. Neurology, 65, 1436–1441.

29.

Alves

, Larsen

, Emre

, Wentzel-Larsen

, & Aarsland

(2006) Changes in motor subtype and risk for incident dementia in Parkinson’s disease. Mov Disord, 21, 1123–1130.

30.

Burn

, Rowan

, Allan

, Molloy

, O’Brien

, & McKeith

(2006) Motor subtype and cognitive decline in Parkinson’s disease, Parkinson’s disease with dementia, and dementia with Lewy bodies. J Neurol Neurosurg Psychiatry, 77, 585–589.

31.

Hauser

, Rascol

, Korczyn

, Jon Stoessl

, Watts

, Poewe

, De Deyn

, & Lang

(2007) Ten-year follow-up of Parkinson’s disease patients randomized to initial therapy with ropinirole or levodopa. Mov Disord, 22, 2409–2417.

32.

Hely

, Reid

WGJ

, Adena

, Halliday

, & Morris

JGL

(2008) The Sydney multicenter study of Parkinson’s disease: The inevitability of dementia at 20 years. MovDisord, 23, 837–844.

33.

Forsaa

, Larsen

, Wentzel-Larsen

, Herlofson

, & Alves

(2008) Predictors and course of health-related quality of life in Parkinson’s disease. Mov Disord, 23, 1420–1427.

34.

Nandhagopal

, Kuramoto

, Schulzer

, Mak

, Cragg

, Lee

, McKenzie

, McCormick

, Samii

, Troiano

, Ruth

, Sossi

, de la Fuente-Fernandez

, Calne

, & Stoessl

(2009) Longitudinal progression of sporadic Parkinson’s disease: A multi-tracer positron emission tomography study. Brain, 132, 2970–2979.

35.

Reinoso

, Allen

, Au

W-L

, Seah

S-H

, Tay

K-Y

, & Tan

LCS

(2015) Clinical evolution of Parkinson’s disease and prognostic factors affecting motor progression: 9-year follow-up study. Eur J Neurol, 22, 457–463.

36.

Hariz

, & Bergenheim

(2001) A 10-year follow-up review of patients who underwent Leksell’s posteroventral pallidotomy for Parkinson disease. J Neurosurg, 94, 552–558.

37.

Lee

, Schulzer

, Mak

, Snow

, Tsui

, Calne

, Hammerstad

, & Calne

(1994) Clinical observations on the rate of progression of idiopathic parkinsonism. Brain, 117, 501–507.

38.

Schrag

, Dodel

, Spottke

, Bornschein

, Siebert

, & Quinn

(2007) Rate of clinical progression in Parkinson’s disease. A prospective study. Mov Disord, 22, 938–945.

39.

Maetzler

, Liepelt

, & Berg

(2009) Progression of Parkinson’s disease in the clinical phase: Potential markers. Lancet Neurol, 8, 1158–1171.

40.

Blin

, Dubois

, Bonnet

, Vidailhet

, Brandabur

, & Agid

(1991) Does ageing aggravate Parkinsonian disability?J Neurol Neurosurg Psychiatry, 54, 780–782.

41.

Caparros-Lefebvre

, Pécheux

, Petit

, Duhamel

, & Petit

(1995) Which factors predict cognitive decline in Parkinson’s disease?J Neurol Neurosurg Psychiatry, 58, 51–55.

42.

Arevalo

, Jorge

, Garcia

, Scipioni

, & Gershanik

(1997) Clinical and pharmacological diferences in early- versus late-onset Parkinson’s disease. Mov Disord, 12, 277–284.

43.

Schrag

, Ben-Shlomo

, Brown

, Marsden

, & Quinn

(1998) Young-onset Parkinson’s disease revisited - clinical features, natural history, and mortality. Mov Disord, 13, 885–894.

44.

Aarsland

, Andersen

, Larsen

, Lolk

, Nielsen

, Kragh-Sørensen

(2001) Risk of dementia in Parkinson’s disease. A community-based study. Neurology, 56, 730–736.

45.

Foltynie

, Brayne

, & Barker

(2002) The heterogeneity of idiopathic Parkinson’s disease. J Neurol, 249, 138–145.

46.

Diederich

, Moore

, Leurgans

, Chmura

, & Goetz

(2003) Parkinson disease with old-age onset. A comparative study with subjects with middle-age onset. Arch Neurol, 60, 529–533.

47.

Foltynie

, Brayne

CEG

, Robbins

, & Barker

(2004) The cognitive ability of an incident cohort of Parkinson’s patients in the UK. The CamPaIGN study. Brain, 127, 550–560.

48.

Hobson

, & Meara

(2004) Risk and incidence of dementia in a cohort of older subjects with Pakinson’s disease in the United Kingdom. Mov Disord, 19, 1043–1049.

49.

Lewis

SJG

, Foltynie

, Blackwell

, Robbins

, Owen

, & Barker

(2005) Heterogeneity of Parkinson’s disease in the early clinical stages using a data driven approach. J Neurol Neurosurg Psychiatry, 76, 343–348.

50.

Suchowersky

, Reich

, Perlmutter

, Zesiewicz

, Gronseth

, & Einer

(2006) Practice parameter: Diagnosis and prognosis of new onset Parkinson disease (an evidence-based review). Neurology, 66, 968–975.

51.

Levy

(2007) The relationship of Parkinson disease with aging. Arch Neurol, 64, 1242–1246.

52.

Williams-Gray

, Evans

, Goris

, Foltynie

, Ban

, Robbins

, Brayne

, Kolachana

, Weinberger

, Sawcer

, & Barker

(2009) The distinct cognitive syndromes of Parkinson’s disease: 5 year follow-up of the CamPaIGN cohort. Brain, 132, 2958–2969.

53.

Wickremaratchi

, Ben-Shlomo

, & Morris

(2009) The effect of onset age on the clinical features of Parkinson’s disease. Eur J Neurol, 16, 450–456.

54.

Selikhova

, Williams

, Kempster

, Holton

, Revesz

, & Lees

(2009) A clinico-pathological study of subtypes in Parkinson’s disease. Brain, 132, 2947–2957.

55.

Zhao

, Wee

, Chan

Y-H

, Seah

, Au

, Lau

, Pica

, Li

, Luo

, & Tan

(2010) Progression of Parkinson’s disease as evaluated by Hoehn and Yahr stage transition times. Mov Disord, 25, 710–716.

56.

Forsaa

, Larsen

, Wentzel-Larsen

, & Alves

(2010) What predicts mortality in Parkinson’s disease? A prospective population-based long-term study. Neurology, 75, 1270–1276.

57.

Aarsland

, & Kurz

(2010) The epidemiology of dementia associated with Parkinson disease. J Neurol Sci, 289, 18–22.

58.

Post

, Muslimovic

, Geloven

, Speelman

, Schmand

, & Han

(2011) On behalf of the CARPA-study group, Progression and prognostic factors of motor impairment, disability and quality of life in newly diagnosed Parkinson’s disease. Mov Disord, 26, 449–456.

59.

Hughes

, Ross

, Musa

, Bhattacherjee

, Nathan

, Mindham

, & Spokes

(2000) A 10-year study of the incidence of and factors predicting dementia in Parkinson’s disease. Neurology, 54, 1596–1602.

60.

Aarsland

, Kvaløy

, Andersen

, Larsen

, Tang

, Lolk

, Kragh-Sørensen

, & Marder

(2007) The effect of age of onset of PD on risk of dementia. J Neurol, 254, 38–45.

61.

Poewe

(2009) Clinical measures of progression inParkinson’s disease. Mov Disord, 24(Suppl 2), S671–S676.

62.

Kempster

, Williams

, Selikhova

, Holton

, Revesz

, & Lees

(2007) Patterns of levodopa response in Parkinson’s disease: A clinico-pathological study. Brain, 130, 2123–2128.

63.

Rajput

, Voll

, Rajput

, Robinson

, & Rajput

(2009) Course in Parkinson disease subtypes. A 39-year clinicopathologic study. Neurology, 73, 206–212.

64.

Jankovic

, & Kapadia

(2001) Functional decline in Parkinson disease. Arch Neurol, 58, 1611–1615.

65.

Roos

RAC

, Jongen

JCF

, & van der Velde

(1996) Clinical course of patients with idiopathic Parkinson’s disease. Mov Disord, 11, 236–242.

66.

Starkstein

, Mayberg

, Leiguarda

, Preziosi

, & Robinson

(1992) A prospective longitudinal study of depression, cognitive decline, and physical impairments in patients with Parkinson’s disease. J Neurol NeurosurgPsychiatry, 55, 377–382.

67.

Harrison

, Wylie

, Frysinger

, Patrie

, Huss

, Currie

, & Wooten

(2009) UPDRS activity of daily living score as a marker of Parkinson’s disease progression. Mov Disord, 24, 224–230.

68.

Poewe

(2006) The natural history of Parkinson’s disease. J Neurol, 253(Suppl 7), VII/2–VII/6.

69.

Schrag

, Jahanshahi

, & Quinn

(2000) What contributes to quality of life in patients with Parkinson’s disease?J Neurol Neurosurg Psychiatry, 69, 308–312.

70.

Hughes

, Daniel

, Kilford

, & Lees

(1992) Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: A clinic-pathological study of 100 cases. J Neurol Neurosurg Psychiatry, 55, 181–184.

71.

CAPIT Committee: Langston

, Widner

, Goetz

, Brooks

, Fahn

, Freeman

, & Watts

(1992) Core assessment program for intracerebral transplantations. Mov Disord, 7, 2–13.

72.

Fahn

, & Elton

, Members of the UPDRS Development Committee (1987) Unified Parkinson’s Disease Rating Scale. In: Recent develoments in Parkinson’s disease, Fah

, Marsde

, Caln

, Lieberma

. Eds. Macmillan Health Care Information, Florham Park, NJ, pp. 153–163.

73.

Schwab

, & England

ACJ

(1969) Projection technique for evaluating surgery in Parkinson’s disease. In: Third Symosium on Parkinson’s Disease, held at Royal College of Surgeons of on 20, 21 and 22 May 1968, Gillingha

, Donaldso

IML

, eds. E & S Livingstone, Edinburgh, Scotland, pp. 152–157.

74.

Giladi

, Shabtai

, Simon

, Biran

, Tal

, & Korczyn

(2000) Construction of freezing of gait questionnaire for patients with parkinsonism. Parkinsonism Relat Disord, 6, 165–170.

75.

Dubois

, Burn

, Goetz

, Aarsland

, Brown

, Broe

, Dickson

, Duyckaerts

, Cummings

, Gauthier

, Korczyn

, Lees

, Levy

, Litvan

, Mizuno

, McKeith

, Olanow

, Poewe

, Sampaio

, Tolosa

, & Emre

(2007) Diagnostic procedures for Parkinson’s disease dementia: Recommendations from the Movement Disorder Society Task Force. Mov Disord, 22, 2314–2324.

76.

Dujardin

, Dubois

, Tison

, Durif

, Bourdeix

, Péré

, & Duhamel

, EXECUTIVE study group (2010) Parkinson’s disease dementia can be easily detected in routine clinical practice. Mov Disord, 25, 2769–2776.

77.

Tomlinson

, Stowe

, Patel

, Rick

, Gray

, & Clarke

(2010) Systematic review of levodopa dose equivalency reporting in Parkinson’s disease. Mov Disord, 25, 2649–2653.

78.

Williams-Gray

, Foltynie

, Brayne

CEG

, Robbins

, & Barker

(2007) Evolution of cognitive dysfunction in an incident Parkinson’s disease cohort. Brain, 130, 1787–1798.

79.

Wickremaratchi

, Knipe

, Sastry

, Morgan

, Jones

, Salmon

, Weiser

, Moran

, Davies

, Ebenezer

, Raha

, Robertson

, Butler

, Ben-Shlomo

, & Morris

(2011) The motor phenotype of Parkinson’s disease in relation to age at onset. Mov Disord, 26, 457–463.

80.

Spica

, Pekmezovic

, Svetel

, & Kostic

(2013) Prevalence of nonmotor symptoms in young-onset versus late-onset Parkinson’s disease. J Neurol, 260, 131–137.

81.

Mendes

, Gonçalves

, Vila-Chã

, Moreira

, Fernandes

, Damásio

, Teixeira-Pinto

, Taipa

, Lima

, & Cavaco

(2015) Appendectomy may delay Parkinson’s disease onset. Mov Disord, 30, 1404–1407.

82.

Kempster

, O’Sullivan

, Holton

, Revesz

, & Lees

(2010) Relationships between age and late progression of Parkinson’s disease: A clinico-pathological study. Brain, 133, 1755–1762.

83.

Muslimović

, Schmand

, Speelman

, & De Haan

(2007) Course of cognitive decline in Parkinson’s disease: A meta-analysis. J Int Neuropsychol Soc, 13, 920–932.

84.

Levy

, Louis

, Cote

, Perez

, Mejia-Santana

, Andrews

, Harris

, Waters

, Ford

, Frucht

, Fahn

, & Marder

(2005) Contribution of aging to the severity of different motor signs in Parkinson disease. Arch Neurol, 62, 467–472.

85.

Riggeal

, Crucian

, Seignourel

, Jacobson

, Okun

, Rodriguez

, & Fernandez

(2007) Cognitive decline tracks motor progression and not disease duration in Parkinsonian patients. Neuropsychiatr Dise Treat, 3, 955–958.

86.

Evans

, Mason

, Williams-Gray

, Foltynie

, Trotter

, & Barker

(2011) The factor structure of the UPDRS as an index of disease progression in Parkinson’s disease. J Parkinsons Dis, 1, 75–82.

87.

Goetz

, Stebbins

, & Blasucci

(2000) Differential progression of motor impairment in levodopa-treated Parkinson’s disease. Mov Disord, 15, 479–484.

88.

Hely

, Morris

, Reid

, ÓSullivan

, Williamson

, Rail

, Broe

, & Margrie

(1994) The Sydney multicentre study of Parkinson’s disease: A randomised, prospective five year study comparing low dose bromocriptine with low dose levodopa-carbidopa. J Neurol Neurosurg Psychiatry, 57, 903–910.

89.

Klawans

(1986) Individual manifestations of Parkinson’s disease after ten or more years of levodopa. Mov Disord, 1, 187–192.

90.

Evans

, Mason

, Williams-Gray

, Foltynie

, Brayne

, Robbins

, & Barker

(2011) The natural history of treated Parkinson’s disease in an incident, community based cohort. J Neurol Neurosurg Psychiatry, 82, 1112–1118.

91.

Bonnet

, Loria

, Saint-hilaire

, Lhermitte

, & Agid

(1987) Does long-term aggravation of Parkinson’s disease result from nondopaminergic lesions?Neurology, 37, 1539–1542.

92.

Giladi

, Tal

, Azulay

, Rascol

, Brooks

, Melamed

, Oertel

, Poewe

, Stocchi

, & Tolosa

(2009) Validation of the freezing of gait questionnaire in patients with Parkinson’s disease. Mov Disord, 24, 655–661.

93.

Ellis

, Cavanaugh

, Earhart

, Ford

, Foreman

, Thackeray

, Thiese

, & Dibble

(2016) Identifying clinical measures that most accurately reflect the progression of disability in Parkinson disease. Parkinsonism Relat Disord, 25, 65–71.

94.

Louis

, Tang

, Cote

, Alfaro

, Mejia

, Marder

(1999) Progression of parkinsonian signs in Parkinson disease. Arch Neurol, 56, 334–337.

95.

Nutt

, Carter

, & Woodward

(1995) Long-duration response to levodopa. Neurology, 45, 1613–1616.

96.

Barbato

, Stocchi

, Monge

, Vacca

, Ruggieri

, Nordera

, & Marsden

(1997) The long-duration action of levodopa may be due to a postsynaptic effect. Clin Neuropharmacol, 20, 394–401.

97.

Clissold

, McColl

, Reardon

, Shiff

, & Kempster

(2006) Longitudinal study of the motor response to levodopa in Parkinson’s disease. Mov Disord, 21, 2116–2121.

98.

Martinez-Martin

, Reddy

, Katzenschlager

, Antonini

, Todorova

, Odin

, Henriksen

, Martin

, Calandrella

, Rizos

, Bryndum

, Glad

, Dafsari

, Timmermann

, Ebersbach

, Kramberger

, Samuel

, Wenzel

, Tomantschger

, Storch

, Reichmann

, Pirtosek

, Trost

, Svenningsson

, Palhagen

, Volkmann

, & Chaudhuri

(2015) EuroInf: A multicenter comparative observational study of apomorphine and levodopa infusion in Parkinson’s disease. Mov Disord, 30, 510–516.