INTRODUCTION

PROBLEM STATEMENT

The course of Parkinson’s disease (PD) after typical motor symptoms are apparent is one of gradual worsening over a decade or more, corresponding to ongoing neuronal loss affecting cells in the pigmented nuclei of the brainstem, particularly in the substantia nigra. There is a preclinical phase of uncertain duration (1), during which loss of dopaminergic neurons and other neuropathologic changes progress until the threshold for clinical motor symptoms is reached. Recently, the diagnostic criteria of PD have been challenged and the Movement Disorders Society has settled a Task Force to establish a new set of diagnostic criteria (2). The estimated neuron loss by the time of diagnosis is about 60% (3). PD progression ultimately leads to important disability, handicap, and death (4). Despite current best standard of care, mortality among PD patients is still higher than the general population one (5). Given this course, there is a growing interest in developing interventions that can change it for the better. It is believed that to obtain an important clinical impact on the natural course of disease, the intervention should have a long-lasting effect that goes beyond the immediate control of signs and symptoms. This is the essence of the disease-modifying concept.

DEFINITIONS

There is no consensus on the definition of the term disease-modifying. A disease-modifying drug for relapsing–remitting multiple scleroses is one that can postpone disability in contrast to those that only reduce the frequency of relapses (6). For a neurodegenerative disorder, a disease-modifying drug is usually considered to be one that can reduce the progression rate. To some extent, this intuitively implies an effect in the physiopathologic mechanism of the disease (7). This last, narrow perspective equates with the concept of neuroprotection, which describes a mechanism of action rather than the consequence of an intervention. From the patient perspective, what is relevant, however, is the occurrence of long-lasting changes in disability, regardless and independently of the mechanism. Thus, we believe that a disease-modifying intervention should be defined as one that is able to have a “long-lasting” effect on disability, although we recognize the difficulties in defining the qualifier “long-lasting.”

The Food and Drug Administration (FDA) links disease-modifying effects in a neurodegenerative disorder with an effect in the mechanism of the neurodegenerative process, meaning that it prefers the narrow, mechanistic approach to the concept. The European Medicines Evaluation Agency (CHMP/EMA) has adopted a two-step approach. Like FDA, the disease-modification qualifier is linked to the ability to prove that an intervention is able to intervene in the physiopathology of the disease. However, for CHMP it is important to demonstrate a clinical relevant effect, and if this is achieved without the proof of the effect in physiopathology, claims on delay on disability are possible (8).

Overall, two stances can be adopted. One is a disease-centered stance where the dichotomy is treatments that interfere with the mechanisms of the disease versus treatments that only have an effect on the expression of signs and symptoms. The other stance is patient-centered. It distinguishes treatments that produce a clinically relevant long-lasting benefit from the ones that produce only transitory effects. In this chapter, we adopt the patient-centered approach in which a disease-modifying intervention is one that prevents/postpones disability. We favor this approach because it is all encompassing and does not overemphasize the importance of preserving neurons to get a relevant benefit. There is a widespread belief that only by protecting or rescuing threatened neurons is it possible to obtain a long-lasting benefit, but this is not necessarily so. The concept of neuroprotection is undoubtedly attractive, but a long-lasting benefit will only be obtained, by this mechanism, if the proportion of neurons that is saved or rescued is significant. Estimates of how large this fraction should be are lacking. It is unlikely that small neuroprotective effects will be clinically relevant. Moreover, most of the preclinical experiments to test potential effects upon neuronal death in PD focus on dopaminergic cells. A number of the most debilitating parkinsonian signs—including dementia, depression, psychosis, falls, orthostatic hypotension, urinary incontinence, constipation, and impotence—are not dependent on dopaminergic denervation. Therefore, it is not expected that they should be sensitive to interventions that would spare dopaminergic neurons alone.

Assuming the broad concept of disease modification, we will argue that levodopa (L-dopa) is the best-documented disease-modifying intervention for PD given that it changes the course of the disease by inducing motor fluctuations and dyskinesias (undesirable) and that it has impact on an important milestone: death (desirable) (7).

OUTCOMES OF A DISEASE-MODIFYING STRATEGY IN PD

To establish if an intervention has or has not had an effect on prevention/postponement of disability in controlled clinical trials, it is necessary to evaluate clinically relevant end points. Outcomes definition is a critical step in the process of establishing if an intervention is disease-modifying. For example, the series of trials comparing L-dopa to various dopamine agonists (9–12), which used time to occurrence of dyskinesias or time to motor complications as a primary outcome, has been criticized for not having used an outcome clearly related with disability (13), which would have made the translation of the results to clinical practice less contentious. These debates are unavoidable because there is a trade-off between the clinical relevance of an outcome and the time frame in which it occurs, with the more-relevant outcomes happening later in the disease than the less-relevant outcomes, which makes the options for an outcome a matter of feasibility (14,15). A parameter for evaluating the feasibility of an outcome is what we designated by the outcome T30 or T50, which is the time it takes, in months or years, for 30% (T30) or 50% (T50) of a theoretical cohort of early PD patients to reach the defined outcome. Table 9.1 provides examples of values for T30 and T50.

CLINICALLY RELEVANT OUTCOMES

Clinically relevant outcomes usually are established by means of consensus expert opinion, although other, more systematic approaches that can reflect patients concerns are available, such as qualitative research methodologies including focus groups (16). Among those outcomes that are consensual, those in Table 9.1 are the most frequently mentioned. However, it is doubtful that postponing death is a goal if disability and handicap are not positively modified. This discussion has never been particularly heated among researchers of PD because the remoteness of this outcome from the moment of diagnosis makes it an unlikely choice. So far, only one trial used death as a prespecified outcome (17,18). Nevertheless, contrary to other opinions (14,15) we consider death inappropriate as an outcome to evaluate a potential disease-modifying intervention in PD because it is not obligatory that delaying death is necessarily associated with delaying disability.

The binary outcomes presented in Table 9.2 should be considered stand-alone outcomes, with a clinical importance of their own, which makes them worth pursuing and distinguishing from surrogate end points. Nonetheless, those end points are not equal in their clinical relevance. In particular, “need for dopaminergic therapy” calls for the following explanation: Postponing dopaminergic therapy is not a goal in itself. In fact, a review of the scientific evidence released most of the fears that L-dopa would be neurotoxic (19). Hitherto, “need for dopaminergic therapy” represents a level of disability that is clinically relevant and has been operationalized in different protocols (20,21). It is arguable that this end point is too early and any benefit related to it might not be maintained later. The counterargument is that if a patient is kept, long enough, in a level of disability below the threshold for needing dopaminergic therapy, it might be worthwhile, even if after reaching that level the disease follows its usual course without change.

In this context and from the patient’s perspective, it is important to note that it is not relevant how the listed outcomes are delayed, whether by preserving neuron health or by providing symptomatic control or both. What is relevant is for how long those outcomes are delayed. However, the view that disease modification implies preservation or rescue of neurons pushed investigators to attempt to disentangle what is a strictly symptomatic effect (short-lasting) from a neuroprotective effect (long-lasting) by means of innovative trial designs, like randomized start trials or randomized withdrawal trials (22). There is no room in this context to engage in a detailed discussion of those designs. Nevertheless, it is important for the sake of information to note that none of those designs has been validated for the purpose. Designs that include washout periods are confounded by dropouts (patients do not stay for the duration of the washout), nocebo effects, unknown biologic half-life for the drugs that have been washed out, etc. The randomized start trial, has been used in the ADAGIO (23) study, where rasagiline was compared with placebo and in the PROUD study (24) where pramipexole was compared to placebo, is far from having been confirmed as useful for disentangling symptomatic from protective effects.

It has been recently proposed that potential disease-modifying agents in PD could be studied by a sequence of trials beginning by a safety phase I study, then by a futility study, followed by a delayed-start study and finally by a long-term classical randomized controlled trial (RCT) (25). These recommendations are based on the facts that, the delayed-start study provides an opportunity to define therapies that provide benefit that cannot be explained by an early symptomatic effect alone, but does not necessarily provide a meaningful measure of the effect of the intervention on cumulative disability. In contrast, the long-term simple study provides a measure of the effect of the drug on cumulative disability but does not address mechanism of action. Therefore, combining results from both kinds of trials might give a complete account of drugs’ mechanism and potency as a disease-modifying therapy.

Futility trials have been purposed as a means to detect “signals” of drug inefficacy (25). Such design has been used in other areas of clinical pharmacology, including the development of drugs for the treatment of cancer. Candidate agents are screened for nonsuperiority in comparison to a placebo or a natural history comparator with regard to their capacity to slow UPDRS (United Parkinson’s Disease Rating Scale) deterioration or some other measure of PD progression. The null hypothesis in these trials is that the drug is superior to the comparator; therefore, if the study is negative, the null hypothesis is rejected and the treatment is deemed futile and is discarded. In theory, the use of futility trials helps in reducing the risk of embarking into “negative” long, large, and expensive parallel group–randomized trials. However, using historical comparative groups raises many methodologic issues (see below). Moreover, futility trials in PD cannot be shorter than trials using a more classical design, because of the intrinsic slow progression of the disease; the advantage is the sample size be smaller than those classical trials. However, timewise there is no advantage except the one acquired in shorter recruitment time. It is also becoming apparent that drugs declared nonfutile by this method end up being considered nonefficacious after a larger trial has been conducted (refers to the NIH PDCoQ10 trial). It will be important to have a better understanding of the diagnostic accuracy of futility trials as a screening tool before using such a resource-consuming tool more generally. If the rate of false positives is high alternatives as a more widespread use of adaptive designs is probably advisable.

BIOMARKERS AND SURROGATE END POINTS

As defined by Temple (26), a surrogate end point of a clinical trial is a laboratory measurement or a physical sign used as a substitute for a clinically meaningful end point that measures directly how a patient feels, functions, or survives. Changes induced on a surrogate end point by a therapy are expected to reflect changes in a clinically meaningful end point.

A correlate does not a surrogate make. It is a common misconception that if an outcome is a correlate (i.e., correlated with the true clinical outcome), it can be used as a valid surrogate end point (that is, a replacement for the true clinical outcome). However, proper justification for such replacement requires that the effect of the intervention on the surrogate end point predicts the effect on the clinical outcome—a much stronger condition than correlation (27). There are no validated surrogate end points in PD.

Biomarker is a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention (28). Several biomarkers have been proposed or used in the context of PD. They were comprehensively reviewed by Mitchell et al. (29). For the discussion of disease-modifying interventions, functional brain imaging is one type of biomarker of special interest.

In ¹⁸F-fluorodopa (L-3,4-dihydroxyphenylalanine) positron emission tomography (PET), fluorodopa is injected intravenously and taken up by nigrostriatal neurons in a manner similar to L-dopa. It is then decarboxylated to fluorodopamine, which is released and metabolized like dopamine. The level of radioactivity during this process is measured. It is dependent on the number of active neurons.

Another approach to functional imaging is based on visualizing the dopamine transporter located on dopaminergic nerve terminals where it actively pumps dopamine back into neurons. Two radio ligands—123I-β-CIT (2β-carboxymethoxy-3β-[4-iodophenyl]tropane) and 123I-FPCIT (N-ω-fluoropropyl-2β-carboxymethoxy-3β-[4-iodophenyl]tropane)—are used to label the presynaptic dopamine transporter for SPECT imaging. The imaging reflects the uptake of the tracer into the striatum, its binding to the dopamine transporter, and its release and metabolism.

Although several trials have used these biomarkers either as primary end points—REAL-PET (30)—or secondary end points—CALM-PD (10), PELMOPET (12), ELLDOPA (31)—the interpretation of the results is full of hurdles: Some patients classified as PD patients by clinicians have normal scans (these can be 5%–15% of a given cohort), the treatment used might impact on the image results, and the existence of longitudinal correlation between imaging and clinical status is not yet demonstrated. There also are very important challenges to such correlation. The most striking example comes from the transplantation experiments where a clear survival of the transplant does not translate into any obvious clinical benefit (32,33). Beyond this example, because it can be argued that transplanted cells are exogenous and do not reflect the paradigm of the analysis in cohorts of early PD patients, the fact is that there is no strong evidence: The few studies available are small and show a large intersubject variability, the decrease in signal in imaging techniques does not correlate with the clinical status, and this would be just the first criteria for imaging to achieve the status of a surrogate end point (see above). Several factors are to be considered when interpreting imaging data: the effect of the treatment being assessed on the marker—for example, dopamine agonist may downregulate dopamine transporters; this effect being wrongly interpreted as reflecting missing neurons; and the sensitivity and reproducibility of the technique being used.

At present, there are no validated surrogate end points for disability in PD. In certain circumstances, the available imaging techniques can be used as biomarkers, but the data generated should be interpreted with caution. Nevertheless, it is desirable to use them as secondary end points in clinical trials, which will increase the knowledge database and ultimately lead to a better-informed use.

The community of scientists (academic and industry) interested in developing new treatments for neurodegenerative disorders is struggling not only with the paradigms to prove a clinical relevant effect, as described, but also with paradigms that will allow a quick evaluation of the potential of candidate drugs to avoid the need to start long-term, large-size sample trials in the dark, not knowing which is the optimal dose or if there is a good basis to believe the drug in question is worth testing. To address this issue, the National Institutes of Health (NIH) in the United States provided funds to a dedicated program. In the context of this program, four candidate drugs (creatine, coenzyme Q10 [CoQ10], minocycline, and neuroimmunophilin) were tested in clinical trials. This program also stimulated creative thinking around the topic of how to explore the potential of candidate drugs. This led to importation from the field of oncology of the concept of futility studies, which have been described before (34).

DISEASE-MODIFYING STRATEGIES IN PD

A summary of the current status of the most relevant agents evaluated for a possible disease-modifying effect during the last decade is offered in Table 9.3.

PROVEN STRATEGIES: TREATMENT WITH L-DOPA

L-dopa has an established role as one of the most efficacious antiparkinsonian agents documented by decades of clinical use. The effect size of L-dopa in PD is large and robust and argues against possible biases that usually affect uncontrolled studies. Therefore, it now seems irrelevant to discuss the evidence basis of L-dopa’s well-established efficacy. Nevertheless, this unquestionable efficacy was confirmed in a history-making randomized clinical trial known as ELLDOPA (31). The discussion of the remarkable achievement that ELLDOPA represents is beyond our discussion. Yet it is relevant to highlight that ELLDOPA results show a clear dose–response for L-dopa efficacy on PD disability as measured by the UPDRS. It also suggests, at a clinical level, that L-dopa may have a “protective” effect, while the imaging data suggest the opposite. This is another blow for the attempt to establish imaging as a surrogate end point in PD. To the end of supporting that L-dopa is disease-modifying, the most relevant argument is that the introduction of L-dopa in the therapeutic armamentarium changed the natural history of the disease by reducing the disability drastically and by having an impact on the time of death, at least in the first 5 years after initiation of treatment, as documented by observational studies (7,35). The most likely mechanism by which this result is achieved is the large symptomatic effect of size on motor symptoms, yet the contribution from other mechanisms cannot be excluded. It is also well documented that the chronic use of L-dopa does produce long-lasting molecular changes at the level of the basal ganglia, which are associated and probably are an important contributor to the occurrence of motor fluctuations and dyskinesias that were unknown in the pre–L-dopa era (36). Such changes do not correspond to a favorable or desirable modification of the disease, but they indisputably change the clinical phenomenology of disease progression. It must be mentioned that in the recent Movement Disorders Society–sponsored Evidence-Based Medicine (EBM) review, L-dopa was considered “Investigational” for the delay of clinical progression, whereas its efficacy for PD symptoms is proven (37).

Related posts:

Epidemiology of Parkinson’s Disease Corticobasal Degeneration Sydenham Chorea and Other Choreas Disorders of Gait Ballism Botulinum Toxin in Movement Disorders

Stay updated, free articles. Join our Telegram channel

Join

Full access? Get Clinical Tree