Understanding the progression of Parkinson’s disease: a review

Abstract

Over the long term, Parkinson’s disease (PD) appears to progress, in a linear fashion, at an annual rate of about 2% of the maximum motor disability score. This figure aligns quite well with pathological research on the rate that substantia nigra dopaminergic neurons are lost. An unexpected finding from cohort studies and clinical trials is that progression is twice as fast in prodromal PD, leading up to clinical diagnosis, and in recently diagnosed PD prior to the commencement of dopaminergic therapy. Levodopa initiation reduces motor disability by 40% of the pretreatment level. This benefit is composed of the short duration response, which is easily measured as the difference between on and off states, and the long duration response, which is comparable in size though not directly observable. Despite clinical impressions to the contrary, there is little evidence that the response to levodopa wanes over time or that axial motor deficits affecting speech, gait and balance become increasingly resistant to treatment. While not revealed by prospective longitudinal studies, the advanced PD phase, accompanied by visual hallucinations and cognitive decline, may show an exponential rate of change. Serial motor scale assessment, informed by a knowledge of symptomatic dopaminergic treatment effects, is probably still the best way to measure the underlying rate of progression of PD in clinical trials.

Introduction

In degenerative disorders such as Parkinson’s disease (PD), the best clinical severity measurements are indicators of pathological change—neuronal damage and loss. Motor disability scales are the longest established method for tracking the progression of PD. Better understanding of PD’s multifaceted neuropathology has led to greater emphasis on non-motor symptoms and their different patterns with respect to disease staging. The Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) is the most widely used clinical rating framework.¹ While its Part III gives an objective motor score, non-motor symptoms are captured by Part I and functional effects of motor disability by Part II.

Motor scales are well suited to clinical trials of symptomatic dopaminergic drugs. With an appropriate placebo-controlled study design, a symptomatic motor benefit can be proven or disproven over several months. Clinical research on disease-modifying therapies in PD faces much bigger challenges. Here, the aim is to show that a trial intervention alters the speed of the pathological process. Changes in the slow underlying progression rate need to be detected in the face of strong and variable symptomatic treatment influences on motor scoring. The possibility that a disease-modifying agent also possesses symptomatic effects creates further difficulty. Many investigators feel that new approaches to design and outcome measurement in disease-modifying trials are needed.^{2 3} Symptom-independent laboratory biomarkers would be ideal, though none as yet are sufficiently sensitive. It may be possible to create reliable gauges of clinical progression by combining sensor-based recordings of motor and non-motor deficits.⁴ At the present time, however, conventional motor disability scales are, despite limitations, the most sensitive way to track progression of PD.

This article will attempt to review measurement of the clinical progression of PD. It will rely heavily, though not exclusively, on longitudinal motor observations. There is a wealth of published material containing serial motor scoring, from clinical trial and from cohort research. A particular focus will be the ways that motor assessments reflect the complex influences of dopaminergic treatments. The trajectory that is determined by pathological progression can best be perceived by understanding the effect of symptomatic drugs on the measurement tool.

Pathological considerations

Substantia nigra

Loss of pigmented neurons in association with Lewy body deposition in the compacta region of the substantia nigra is the pathological lesion for dopamine deficiency in PD. Four studies have employed cell counting methods to correlate nigral cell degeneration with control brains and PD duration.5,8 About 50% of nigral cells are lost at the point of PD diagnosis. The annual rate of nigral cell loss ranges between 1.4% and 1.9% with reference to control neuronal populations. Inevitable differences in rates of cell loss and causes of death restrict the accuracy of these figures. There is always a small population of residual compacta cells. These studies estimate that 12–25% have survived at the endpoint of a long disease course. There are greater and lesser degrees of cell loss when counting is performed by anatomical subregion of the substantia nigra compacta.⁵ Nevertheless, there does remain a source of intrinsic dopaminergic neurotransmission for the whole duration of the disease.

Lewy pathology and the Braak staging system

Two things about Lewy bodies had been known for a long time—that they are not confined to the substantia nigra in PD; and that they are not specific to PD, being present in a number of other neurodegenerative disorders. In 1997, alpha-synuclein was found to be a major component of these intracellular inclusions.⁹ Immunohistochemical stains for alpha-synuclein then disclosed a range of additional Lewy pathologies, with aggregations in nerve cell bodies and processes. Braak and colleagues, from observations on brains with incidental Lewy pathology and clinically diagnosed PD, determined a hierarchical, caudal to rostral six-stage scheme for the pathological progression of the disorder.¹⁰ It rests on several assumptions—that Lewy body deposition is not simply a consequence of ageing, but is part of a pathological continuum that leads on to disease states; and that Lewy body deposition is integral to the processes causing nerve cell death. As a unifying account of Lewy pathology, the Braak staging system has shortcomings—a minority of PD brains diverge from its pattern, and many cases of dementia with Lewy bodies do not conform because alpha-synuclein inclusions tend, initially, to bypass the brainstem. But the strength of the model is its ability to explain why olfactory or sleep-related non-motor symptoms can be premonitory PD features (stages 1–3), and why cognitive and neuropsychiatric deficits dominate its advanced phase (stages 5–6).

Many neuronal populations show alpha-synuclein deposition. They are anatomically and neurochemically diverse. Braak remarked on properties common to susceptible neurons—projecting axons that are long in relation to their cell body and that tend to be unmyelinated or thinly myelinated.¹¹ While the involvement of other regions may contribute to non-motor symptoms, cell loss in relation to Lewy pathology is generally greatest in the substantia nigra, explaining why PD is defined by its motor deficits.¹²

Other pathologies

Upward spread of Lewy pathology to limbic and neocortical brain areas coincides with dementia in PD,¹³ though this is not the sole pathological substrate. Deposition of abnormal proteins associated with Alzheimer’s disease, both amyloid-beta and tau, is also present in PD dementia.^{13 14} Simply additive neurodegenerative effects, shared pathways to protein misfolding and synergistic properties of the misfolded proteins themselves are all possible interpretations. Composite indicators that embrace all three pathological proteins appear best to predict the presence of dementia.¹⁵

Tracking motor disability in PD

Clinimetric tools for motor rating in PD are accurate, sensitive and have been well validated. The MDS-UPDRS Part III motor scale (2008, maximum score=132) is now used for most research purposes. A considerable amount of data about motor progression is expressed in the predecessor scale, the UPDRS-III (1987, maximum=108).¹⁶ These two scoring systems show strong correlation and share many attributes.¹⁷ About 55% of scoring is for bradykinesia, as opposed to rigidity or tremor. Two-thirds of items assess limb motor function, with less weighting for axial and gait deficits. Nigral cell density shows an inverse linear relationship with UPDRS-III scoring.⁶ The UPDRS-III was modelled on older scales, particularly the Columbia University Rating Scale (maximum=80).¹⁸ Some very long-term progression research employed the compact Webster scale,¹⁹ which, with modification (maximum=36), has a more even balance of lateralised and axial/gait motor scoring.

The ‘rate what you see’ instruction for the MDS-UPDRS-III²⁰ means that other factors interfering with repetitive movement or gait—non-parkinsonian tremor or musculoskeletal problems—can generate spurious scores. A floor effect may exist for some items of this scale, not surprising for ‘midline’ motor disabilities that develop later on.²¹ As will be shown, the UPDRS-III and MDS-UPDRS-III perform well enough when measuring mild parkinsonism. The scales accurately register the motor benefit of dopaminergic medications, though the strength of symptomatic treatment effects creates difficulties with tracking underlying disease progression. There is a relative paucity of longitudinal research that extends into the second decade of the disease course and its more advanced stages.

This review will follow a convention of representing motor progression as an annual percentage increase of the maximum score, irrespective of the actual objective motor scale. Initial response to dopaminergic treatment refers to the percentage decrease from pretreatment disability score, again irrespective of scale. In many studies, prevailing motor state has been rated without regard to medication cycle. Some researchers try to classify prevailing motor states as either off or on. The most rigorous methods use practically defined off states (overnight withholding of medication, preferably while fasting) followed by levodopa test doses with ascertainment of maximum benefit. Reference will be made to a recent systematic review and meta-analysis of motor progression when discussing pooled results and selected groups of studies.²²

The disease course

Pre-motor, early motor PD

Non-motor symptoms in established PD have long been recognised. James Parkinson seems to have been aware of some of them when he wrote An Essay on the Shaking Palsy in 1817.²³ An implication of the Braak model of PD stages was that certain non-motor symptoms could precede motor involvement, allowing a prodromal disease phase to be identified. Logically, neuroprotective or disease-modifying treatments should be more effective when cell loss in the substantia nigra and more rostral brain areas is relatively mild. As a result of this interest, there is now research data on the earliest motor manifestations of PD.

Some older individuals develop ‘mild parkinsonian signs’. These can be classified in various ways, but most previous publications refer to UPDRS-III scores of <5, including one or more domains of parkinsonian disability.²⁴ This entity captures some physical effects of normal ageing as well as minor motor changes from other degenerative processes like Alzheimer’s disease. 25–40% of older persons in a population sample may comply with a minimal parkinsonian signs definition, depending on its strictness.²⁴ On those figures, only a small minority will convert to diagnosable PD. In community-dwelling older adults with minimal parkinsonian signs by the criteria of Louis et al,²⁵ motor progression occurred only at 0.13% per annum (p.a.).²⁶

The presence of non-motor symptoms as risk factors for later development of PD does increase MDS-UPDRS-III mild parkinsonism scoring in community-based research.²⁷ Prodromal disease indicators have also been used to assemble at-risk cohorts that could be followed while some progressed to diagnosable PD. This approach has enabled valuable observations about the earliest parkinsonian motor features. One study used olfaction and dopamine transporter scans, back-analysing disability scores from the time point of phenoconversion.²⁸ The UPDRS score worsened at a rate of 7.5% p.a. in the 2 years before diagnosis, whereas progression had been 0.9% p.a. 3 and more years before diagnosis. A cohort identified by rapid eye movement sleep disorder progressed at 5.2% p.a. over 2 years before phenoconversion compared with 0.6% p.a. 3 and more years prior to diagnosis.²⁹

Diagnosis and pretreatment phase

At the point of clinical diagnosis of PD, patients have accrued roughly 20% of a maximum disability score. The average of UPDRS-III scores at phenoconversion in two prodromal disease studies just mentioned was 18.4.^{28 29} In eight clinical trials that recruited untreated subjects with early PD and collected serial motor scores, the median baseline disability score was 19% of maximum.30,37 Individuals in placebo arms of these trials progressed at almost 5% p.a. This accords with the observations on prodromal PD in the couple of years leading up to diagnosis. But it is twice the long-term progression rate from trial or cohort studies that contain individuals on dopaminergic treatment.²²

There seems to be no reason to suspect this difference is artefactual. All these clinical trials were conducted with rigorous methodology, and multiple blinded motor assessments were published. Clinical trial placebo effects would work in the opposite direction. Distorted motor scaling properties at low scores is not the explanation. Patients actually traverse the 10–20% range of the measurement scale twice—once before treatment is started; again, more slowly, during long-term progression after initial dopaminergic benefit. This can be appreciated from figure 1.

Figure 1. Schematic representation of motor progression based on pretreatment disability and serial defined off state measurements of levodopa response. Top of solid bars, off state; bottom of solid bars, on state. Red arrow, total initial levodopa response; green arrow, short duration response; blue arrow, long duration response. The grey shaded area at the top of the figure imagines the trajectory of untreated disability (as if drugs had never been given) and its margin of uncertainty to estimate the magnitude of the long duration response over time.

Dopaminergic treatment effect

From 16 published longitudinal studies in which there was concurrent commencement of levodopa therapy, the mean initial motor response was 40.3%±15.2% of pretreatment disability.²² Mean time to maximum improvement was 7 months in 10 of those studies where multiple early time points permitted this to be estimated.

The Movement Disorder Society Clinical Diagnostic Criteria for Parkinson’s Disease (2015) define the supportive criterion of ‘clear and dramatic beneficial response to dopaminergic therapy’ both qualitatively and quantitatively.³⁸ The objective benchmark is >30% improvement in MDS-UPDRS III score. The origin of this figure is not given, though based on the SD statistic for the initial motor response, 75–80% of initial responses might comply if recorded scores during the first treatment year are all considered.

The two levodopa responses

The short duration response (SDR) to levodopa follows the pharmacokinetic profile of the drug and is easily measured as the difference between baseline (off state) and peak post-dose effect (on state).³⁹ It is present from the very first administration of the drug and can be accurately tracked over time through off and on scoring.

The long duration response (LDR) is an additional motor benefit that does not relate to dosing and does not even rely on the presence of the drug in the brain since it can outlast withholding of medication by days or weeks.^{40 41} Right at the beginning of the era of dopaminergic therapy, George Cotzias and his colleagues were aware of it. They noted that after prolonged use of d,l-dopa, it took 4–14 days for motor states to return to baseline when the drug was ceased.⁴² Drugs other than levodopa, dopamine receptor agonists, for instance, also have LDRs.^{43 44} It can be measured with difficulty by prolonged levodopa withdrawal. The LDR has been estimated to compose 30–50% of the overall motor response.^{40 41} It is not present when the first levodopa tablets are given, and it seems to take many months to become fully established.⁴⁵ Later, its presence can be inferred by comparing off scores with pretreatment disability and with the likely trajectory of the untreated motor deficit (see figure 1).⁴⁶

Long-term motor progression

Based on an analysis of a large number of published longitudinal studies, motor progression in PD is about 2% p.a.²² Serial motor measurements ranging from intervals as short as 6 months to as long as 20 years, and including patients before and during standard pharmacological therapy, were used. Simple averaging of all available longitudinal scoring and meta-analysis adjusted for sample size and statistical spread converged around the 2% p.a. figure. The source data contained an estimated 5.7×10⁴ patient-years of PD progression.

Two factors present in extended longitudinal research—attrition of sample size from loss to follow-up or death, and sequential commencement of dopaminergic therapy as part of normal management of PD—had a minimal effect on progression, though both might have been expected to depress the rate. Similarly, baseline age, observation period, sample size and motor disability scale made little difference to calculated rates of deterioration.

Virtually without exception, studies with sufficient serial observations to be plotted showed a linear relation between motor disability and time. A caveat to this conclusion, however, is that only a small number of studies captured data extending into the advanced stage of PD.

Nine studies that conducted defined off state and levodopa test dose assessments show a narrow but significant difference in progression rates, with average off disability increasing faster than on by 0.6% p.a.47,55 In five of these studies with scoring into the second decade of the disease course, this divergence was still present.51,55 The one with the longest set of continuous observations, spanning more than 20 years, suggested that eventually, on and off disabilities progress in parallel.⁵¹

Motor fluctuations and dyskinesia

A ‘rule of thumb’ that motor fluctuations develop at a rate of 10% per levodopa treatment year dates to the early levodopa era.^{56 57} While some subsequent surveys suggest a lower incidence, it is still broadly true to say that about half of patients will have developed fluctuations and dyskinesia after 5 years of treatment,⁵⁸ and that these treatment complications are often the dominant management issue during the middle stage of PD. Motor fluctuations do not develop without a good levodopa response in the first place. Duration of PD and the severity of the intrinsic dopaminergic deficit are more important than duration of levodopa treatment. Cotzias had documented this when he reported the first effective use of levodopa.⁴² The observation is occasionally still made—fluctuations and dyskinesia quite quickly developed in a high percentage of sub-Saharan African patients who were disabled by many years of PD before commencing levodopa monotherapy.⁴⁸

When motor fluctuations begin, patients are aware of the onset and ending of their motor responses and will often say that the duration of benefit seems to be contracting. Earlier research concentrated on shortening of the SDR as a major factor for the instability. Some shortening can be shown in longitudinal measurements, though the change in response duration is not great.⁵⁹ Furthermore, it is hard accurately to define the start and finish of an SDR in early PD when the SDR amplitude is relatively small. Cross-sectional studies confirm that fluctuating patients have greater response amplitudes.^{60 61} This implies that off scores must increase faster than on scores to account for the widening SDR and is consistent with the small difference registered by serial defined off state and test dose measurements.47,55 One longitudinal observational study showed that a significantly greater SDR amplitude in fluctuators developed because of better on phase scoring during the first treatment decade, with little difference in the rate of off deterioration.⁶²

Advanced PD

The appearance of clinical milestones such as frequent falling, visual hallucinations, cognitive disability and need for high-level care identifies an advanced stage of PD. These are somewhat inter-related—visual hallucinations often accompany cognitive decline;⁶³ cognitive impairment is a risk factor for falling⁶⁴ and capacity for independent living is undermined by each of the other milestones. When plotted against the entire disease course for clinico-pathological correlation, the milestones delineate an advanced PD phase of approximately 5 years.⁶⁵ They do have pathological significance—scoring for Lewy body burden rises in proportion to the number of milestones recorded, and multiple clinical milestones correspond to neocortical Lewy body disease. Alignment of the disease course with time of death reveals an unexpected pattern. The length of the time section that is heralded by one or more clinical milestones is about the same irrespective of the duration of the preceding disease course or the age at diagnosis. Patients with young onset PD may take decades to reach the advanced disease state. Some patients diagnosed with PD in their 80s develop cognitive decline and other advanced PD milestones within a few years. This observation suggests that disease progression might not wholly be explained by a linear model.

Controversies and uncertainties about motor progression in PD

What is happening in emergent PD before the commencement of dopaminergic treatment?

The suggestion that motor scores increase twice as fast in early, untreated PD comes from placebo arms of clinical trials designed for quite different purposes. While further confirmation is needed, the differences appear real. The pretreatment phase has been thought of as a useful research window to evaluate putative disease-modifying treatments without confounding symptomatic treatment effects. Participants are temporarily denied the benefit of dopaminergic drugs, restricting such trials to short time scales. These unanticipated differences in progression rate call into question the assumptions for such clinical trial design.

There seems to be no biological reason for nigral cell loss to surge temporarily around the time that PD is diagnosed. Some other explanation, possibly involving mechanisms by which surviving nigral cells compensate for reduced dopaminergic innervation of the striatum, is more likely. These compensatory mechanisms, which may operate at presynaptic and postsynaptic levels, are not well understood.⁶⁶ But they must exist to allow a 50% deficit of dopaminergic substantia nigra neurons to be borne with minimal clinical parkinsonism. Eventually, their capacity is exceeded and the motor symptoms appear. The way this tolerance is lost may explain faster accrual of motor disability score in pretreatment PD.

An observation that is sometimes made in the clinic could be relevant here. A patient is diagnosed with PD. There are prominent motor signs, yet symptoms have been present only for a number of months, suggesting that progression must be rather rapid. Dopaminergic treatment is started, a satisfactory response ensues and several years later the patient seems to be following a standard PD course.

What causes the LDR, and how might it affect measurement of motor progression of PD?

The LDR is the ‘dark matter’ of anti-parkinsonian motor benefit, substantive but not directly observable. Its presence means that from the day a patient commences levodopa tablets, it may be as long as 5 years before their treating clinician again observes as much parkinsonism, irrespective of medication cycle (see figure 1).^{22 51}

Muenter and Tyce first suggested that separate SDR and LDR to levodopa implied that these treatment effects occurred in different compartments.⁶⁷ There are a number of theories for the LDR.^{43 68} One straightforward explanation does involve two compartments for dopaminergic action. The SDR disappears with the pharmacokinetic decay curve of pharmaceutical levodopa. The LDR persists because surviving substantia nigra dopaminergic neurons continue to synthesise and release dopamine. Some of this capacity runs down over days and weeks, possibly because compensatory metabolic and synaptic activities of these neurons cannot be sustained without the presence of an exogenous source of dopamine.

Though largely invisible, the LDR should be considered when mounting clinical trials for PD. There is uncertainty about the duration of withdrawal needed fully to reveal it. Some trial designs have employed washout phases to account for symptomatic drug effects,^{31 69} but a residual LDR will be present if these are too short. Neuroprotection trials may employ defined off state measurements to minimise interference from symptomatic levodopa effects. Unobservable variations in the LDR could still affect estimation of underlying disease progression, especially over shorter trial durations.

Does levodopa stop working over time?

The concept of a ‘honeymoon’ phase for levodopa, which can be traced back to the early era of the drug in the 1970s,⁷⁰ cast a long shadow over the treatment choices for PD. Authors of these publications were often referring to the appearance of motor complications (fluctuation and dyskinesia) after an initial period of stable motor benefit.

Development of an unstable response to a drug is not the same as a general reduction in its pharmacological action. Particularly in literature prepared to inform patients, though, these two things were often conflated. Some clinicians do have the impression that the levodopa response wanes in advanced PD.⁷¹ It has been suggested that as many as 20% of patients who initially enjoyed a beneficial response develop resistance to levodopa and that peripheral pharmacokinetic factors could be responsible.⁷²

If the levodopa motor effect decays with time, this should be apparent in longitudinal measurements of its SDR. Yet studies covering various portions of the first 15 years of the disease course that conducted defined off state and levodopa test dose measurements show no sign of this.47,55 Best motor function on treatment continues to worsen but the SDR magnitude is maintained. It should be noted that there are few serial measurements extending into the advanced disease state, particularly when the progression milestone of cognitive impairment is well established. It is also possible that increasing instability leads to a larger gap between objectively measured levodopa responses and the perceived benefits of the drug in daily living. Nonetheless, there is ample evidence to reassure patients that levodopa does not stop working and that they will continue to respond to it indefinitely.

What is the evidence about levodopa-resistant ‘midline’ motor deficits during disease progression?

The Hoehn and Yahr clinical staging scheme, first published in 1967,⁷³ affirms a truism about PD—axial motor deficits affecting gait, postural stability and speech develop once limb involvement is bilateral and cause increasing disability over time. It is often stated that these ‘midline’ motor deficits show poor responsiveness to levodopa. Cross-sectional studies of PD⁷⁴ and broad clinical impressions about its progression are the main sources of this conclusion—there is less relevant longitudinal clinical research. In such statements, it can be hard to tell whether it is being implied that axial motor deficits are intrinsically less treatment responsive throughout the disease course, as opposed to becoming more levodopa resistant over time.⁷⁵

The UPDRS-III and MDS-UPDRS-III scores for axial deficits can be summed to give a general idea of the behaviour of this group of motor signs. A composite measure for postural instability/gait difficulty motor subtype contains three motor items plus two subjective experiences of daily living subscores in the MDS-UPDRS.⁷⁶ But objective rating for speech, truncal mobility, overall gait function and postural stability each rely on single four-point scales, allowing only coarse sensitivity to change. Cross-sectional studies have evaluated levodopa responsiveness of these deficits using more discriminating measurement. Perceptual and acoustic analysis of speech characteristics shows levodopa responsiveness in some but not other studies.77,79 Gait and balance assessment using inertial sensors confirms significant improvement in gait parameters after levodopa doses. By contrast, static balance did not respond and may be worse where levodopa-induced dyskinesia increases postural sway.⁸⁰ Gait freezing, which eventually affects more than half of patients with PD,⁸¹ shows pharmacological heterogeneity. It is usually worse in off phases, denoting levodopa benefit.⁸² But it may be resistant to levodopa or paradoxically worsened by it (on freezing).⁸³ Apparently resistant gait freezing sometimes responds to supranormal doses of levodopa.⁸⁴

Longitudinal measurement of the levodopa responsiveness of axial motor deficits by defined off and test dose methodology is available from just two studies. Merola et al obtained an axial motor score based on six UPDRS-III items in a patient group with mean 14 years duration of PD.⁵⁴ A second assessment was done 5 years later. Axial disability had worsened, but its SDR actually increased slightly. Ganga et al separated subscores of a modified Webster motor scale (maximum score=36) to ‘midline’ and ‘lateralised’ categories.⁸⁵ Multiple observations, beginning at the commencement of levodopa treatment and extending more than 15 years into the disease course, were published. Axial motor disability progressed faster than lateralised disability, and the proportion of off phase midline deficits increased with time. However, the magnitude of midline SDR was preserved. While evidence that axial deficits become progressively resistant to levodopa is weak, it is important to acknowledge that a given axial SDR may deliver much less functional benefit at higher disability score levels.

Does the worsening of parkinsonian disability accelerate when the advanced disease phase is reached?

The impression of PD as a disorder of linear progression appears to be supported by serial motor assessments. Yet biological processes of growth or decay are usually governed by exponential laws. The observation from clinico-pathological research that a set of clinical milestones delineates an advanced disease stage of fixed duration and not dependent on length of prior disease course implies exponentiality.⁶⁵ This may be difficult to perceive in prospective cohort research, in which most participants do not have advanced PD and survivors dominate the motor scoring statistics. But one attempt to align serial motor scores looking backwards from the end of the disease course suggested a late trend of acceleration.⁵¹

Braak proposed that PD progresses by spread of Lewy pathology across contiguous brain regions.¹⁰ It did not take long to realise that this could be explained by a prion-like disease mechanism,⁸⁶ especially when autopsy studies found Lewy bodies in fetal dopaminergic neurons that had been grafted into the brains of PD patients.^{87 88} It is now known that alpha-synuclein aggregates can propagate aggregation of monomers of alpha-synuclein, though only in in vitro conditions with supraphysiological concentrations of the native protein.⁸⁹ A transmissible molecular basis for Braak staging, though unproven, would likely result in some degree of exponentiality. The topographic scheme alone predicts involvement of exponentially higher numbers of neurons once the large neocortical cell populations are affected in stages 5–6.

Some conclusions

For much of the PD course, the rate of motor progression is about 2% p.a. This figure correlates quite well with estimated rates of nigral cell loss from clinico-pathological research. But around the time of PD diagnosis and before the commencement of dopaminergic drugs, motor scores seem to be increasing twice as quickly. The explanation is not clear. While study designs involving prodromal or untreated PD are free of symptomatic treatment effects, this period may not reflect the long-term progression rate.

Pathological studies show that there is always a population of surviving nigral cells. Their importance throughout the disease course is hard to assess. They determine pretreatment motor function. They may play a role in setting the level of off scores through the LDR to levodopa.

Motor scales do not represent all consequences of neurodegeneration in PD, particularly its non-motor symptomatology. But the scales are reliable and sensitive markers of the disorder’s temporal behaviour. All things considered, defined off state measurements are probably the best guide to the rate of motor progression for clinical trial purposes.

Based on an average disease duration of 15 years, PD can be considered to have early, middle and advanced segments, each lasting roughly 5 years. The usual supposition of progression in a linear manner may not be accurate. There is some evidence for exponential increase in disability in advanced PD.