Schematic representation of the theoretical effect of neuroprotective, symptomatic, or mixed neuroprotective and symptomatic treatments compared with placebo in different study designs. The x-axis represents time and the y-axis represents the measure of progression, for instance Unified Parkinson’s Disease Rating Scale (UPDRS) score. The parallel-group design (A–C), washout design (D–F) and delayed-start design (G–I) are illustrated. The placebo group is depicted as a blue line. The green lines (A, D, G) represent a pure neuroprotective agent, the red lines (B, E, H) represent a pure symptomatic agent, and the purple lines (C, F, I) represent a mixed symptomatic and neuroprotective agent (see text for details). Of note, the curves are artificially simplified; in particular, a beneficial clinical effect of the placebo (i.e. a placebo effect) is not reflected.
Washout period studies
To minimize the symptomatic effects, a washout period design (Figure 25.1D–F) has been proposed and used in some studies. Assuming complete washout of the agent, this theoretically has a significant advantage over parallel-group studies. However, such trials often have limitations in the duration of the washout period, considering the availability of symptomatic treatments and the gradual progression of the disease. Moreover, patients receiving active treatment during the study may not tolerate the washout as well as those who received placebo, and this may lead to differential dropout rates in the two treatment groups. Another related problem is that it is easier to retain patients in an active-treatment arm (due to the symptomatic benefit) than in a placebo arm, and this can also lead to differential dropout during the trial (this is a problem with all studies that have an active versus placebo control period). Furthermore, long-term symptomatic effects persisting after the end of the washout period (as proposed in the ELLDOPA trial, see below) cannot be excluded.
Delayed-start trials
The delayed-start design (Figure 25.1G–I) was conceived with the objective of dissociating neuroprotective effects from confounding symptomatic benefits [5]. This type of study consists of two phases: in phase 1, patients are divided into placebo and active-treatment groups. The primary outcome, usually the UPDRS, is measured at the end of phase 1 to evaluate symptomatic effects, and the slope of progression is also calculated. In phase 2, patients in the placebo group are converted to active treatment, while patients in the early-treatment group maintain the same intervention. Outcomes and slopes are then analyzed the same way as in the initial phase.
The ADAGIO (Attenuation of Disease Progression with Azilect GIven Once daily) trial was the first to formally use the delayed-start design in a neuroprotective trial in PD [34]. In this study, it was determined from the outset that, to establish neuroprotective effects, the intervention had to fulfill all three of its predefined endpoints:
1. A significant difference in the primary outcome at the end of phase 2, favoring the early-treatment group, i.e. an advantage remains in early-treated patients when both groups are receiving the active intervention and the symptomatic effect is presumed to be comparable in both groups.
2. Divergent slopes of progression in phase 1 (slower progression in the early-treatment group than in the placebo group), i.e. the disease-modifying effects are increasing over time.
3. Nonconvergent slopes of progression in phase 2 (parallel slopes or slower progression in the early-treatment group), i.e. the disease-modifying effects are not lost over time.
Although this design seemed promising, many criticisms have emerged [5, 33, 35, 36]. First, the duration of the initial placebo phase is limited ethically by the availability of symptomatic treatments. If the placebo period is too short, an agent that has its maximal effect over a longer period than the duration of phase 1 might be falsely rejected and neuroprotective effects may be missed. Likewise, if the symptomatic effect takes longer to peak than the duration of phase 2, a difference at the end might be misinterpreted as neuroprotection, when in fact the delayed-treatment group has not yet reached maximal benefit. It has also been argued that inclusion criteria in these types of studies might favor slow progressors, as patients are recruited with the expectation that they will not require treatment in the next 6–9 months. This might diminish the power of the study to demonstrate neuroprotection because small changes in endpoints may not be considered statistically significant. A differential placebo response in early-treatment and delayed-treatment groups with an agent that has a symptomatic benefit may also be confused with a neuroprotective effect. Despite these shortcomings, currently this design is probably the best available method to assess neuroprotective effects in PD.
Futility studies
Because of the high number of candidates for neuroprotection, futility studies have been introduced to facilitate selection of agents that have a higher chance of obtaining positive results in phase 3 studies [37]. Futility studies use small groups of patients, which are often compared with a historic control group instead of a placebo group, in order to minimize the sample size. In these studies, the null hypothesis states that the intervention is superior to placebo. If results are positive, the null hypothesis cannot be rejected (the agent does not meet the futility threshold), and the agent can be considered for further studies. To minimize the risk of falsely considering an agent futile (and rejecting a possibly effective neuroprotective agent), the P value is usually set higher, at 0.10 or 0.15. This method minimizes costs and enables the study of more agents with minimal sample sizes. One important challenge in using this approach in PD has been the finding that historical controls may not adequately represent modern patient outcomes and so concurrent placebo groups are deemed necessary (i.e. the sample size advantage to the design no longer applies).
Large simple clinical trials
The progressive nature of PD and the availability of highly effective symptomatic treatments limit the duration of studies in untreated PD. Furthermore, the heterogeneity of clinical manifestations makes the choice of a valid endpoint for neuroprotection difficult. An intervention given over a long time period could make determination of disease modification easier, by permitting symptomatic treatment according to standard practice [33]. Endpoints such as quality of life, necessity of mechanical aids for ambulation, cognitive dysfunction or independence in activities of daily living could reflect disease modification in a more global manner.
Animal models
This topic, which is critical to developing effective neuroprotective treatment in PD, is largely beyond the scope of this chapter. We will only touch briefly on some of the more important issues.
Different types of animal models have been developed; however, none of them has accurately reproduced the complete pathophysiology of PD. The ideal animal model would have a chronic and progressive evolution, and would involve not only dopamine but also multiple other neurotransmitter systems [38]. The variable properties of neurons in different animal species are an issue, and it has been suggested that a neuroprotective agent should be demonstrated to be effective in primate models, which might be closer to human PD. Age is possibly the most important risk factor for PD, and it should be emphasized that this is largely not considered or accounted for in current animal models.
Toxin-induced models
The most widely used models are toxin-induced models, such the as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mice and nonhuman primates, as well as the 6-hydroxydopamine (6-OHDA) rodent. These models, while useful to evaluate symptomatic treatment and induction of motor complications, are inaccurate in terms of pathophysiology for PD, and neuroprotection studies are therefore difficult to translate into human clinical trials. These models consist of an acute lesion induced by the toxin, followed by a cascade of neurotoxicity and cell death that persists and may progress despite cessation of exposure [39].
The MPTP model was thought to involve primarily the dopaminergic system and therefore to represent only dopaminergic manifestations of PD; however, other monoamines are affected to a lesser extent. The toxin 6-OHDA has less specificity for dopamine neurons, and this model requires pretreatment with selective serotonin reuptake inhibitors, tricyclic antidepressants or monoamine oxidase B (MAO-B) inhibitors to be more similar to PD. Even with these precautions, PD pathophysiology is very complex and cannot be reproduced perfectly in these models. Most notably, we now recognize a critical role for α-synuclein in the pathogenesis of most types of PD, and this is largely lacking in toxin models in current use.
Other criticisms relate to the acute nature of the lesion, which is not believed to be present in idiopathic PD. The introduction of the neuroprotective candidate agent is often made before the lesion is induced; ideally, the agent should be started once the neurodegenerative process is already underway, to more accurately reproduce clinical practice. Furthermore, the cascade of neurotoxic events might be different in toxin-exposed animals compared with idiopathic PD, and the pathological findings are also different, which again raises questions on the reliability of these animal models. One recent model that has been generating considerable interest involves the inoculation of “preformed fibrils” of α-synuclein into the brain of mice [40]. This has been shown to result in cell-to-cell spread (in a prion-like fashion) of α-synuclein pathology similar to that seen in PD [41].
Transgenic models
New animal models are based on genetic forms of PD, most notably (with respect to idiopathic PD) α-synuclein and leucine-rich repeat kinase 2 (LRRK2) [42]. These models might better represent the chronic evolution of PD compared with toxic models. However, they still do not reproduce all of the pathological features of PD, with some showing α-synuclein aggregation but no cell death, and some showing neurodegeneration but no Lewy bodies. Genetic-based models also do not accurately reproduce the presumed contribution from environmental exposures (infections, inflammation, toxins) that may be a determining factor in “sporadic” PD. Vector-based genetic models are currently in development and may be more representative of PD. This method allows the use of older animals, which decreases contributions from adaptive changes, and the chronic evolution is also more similar to PD. In addition, the possibility of injecting the vector unilaterally permits usage of the unaffected side as a control.
Results from major clinical studies
As mentioned above, currently no agent has been unequivocally proven to slow the progression of PD. Multiple studies have been conducted, some with conflicting results, but none with convincing evidence of benefit. Here, we review data from selected major clinical studies, emphasizing the important lessons learned (Table 25.2). A large number of agents have been studied in smaller trials or are currently under evaluation; the reader is referred to a recent review paper in which many more treatments are summarized [43].
DATATOP, Deprenyl And Tocopherol Antioxidant Therapy Of Parkinson’s disease; SINDEPAR, Sinemet-Deprenyl-Parlodel; TEMPO, TVP-1012 in Early Monotherapy for Parkinson’s Disease Outpatients; ADAGIO, Attenuation of Disease Progression with Azilect Given Once-daily; ELLDOPA, Early versus Later Levodopa Therapy in Parkinson Disease; REAL-PET, Requip as Early Therapy versus l-DOPA-PET; CALM-PD, Comparison of the Agonist pramipexole with Levodopa on Motor complications of Parkinson’s Disease; PROUD, PRamipexole On Underlying Disease; NET-PD, Neuroprotection Exploratory Trials in Parkinson’s Disease; LS1, Long-term Study 1; PRECEPT, Parkinson Research Examination of CEP-1347 Trial; QE3, Effects of Coenzyme Q10 (CoQ) in Parkinson Disease; MAO-B, monoamine oxidase B.
Monoamine oxidase B inhibitors
DATATOP
The DATATOP (Deprenyl and Tocopherol Antioxidant Therapy of Parkinson’s disease) study [44] was conducted in 800 patients with early PD. This trial involved a 2 × 2 factorial design consisting of four parallel groups: selegiline (10 mg/day) and placebo, tocopherol (2000 IU/day) and placebo, selegiline and tocopherol, or placebo and placebo. The endpoint was the time to progression of disability requiring treatment with levodopa. Selegiline treatment, but not tocopherol, delayed the time to start of levodopa by almost 9 months. Washin and washout effects were observed, consistent with symptomatic benefit from treatment. Symptomatic effects may have overshadowed possible disease modification, and generally the results have not been interpreted as supporting true neuroprotection.
SINDEPAR
The SINDEPAR (Sinemet-Deprenyl-Parlodel) study [45] compared selegiline versus placebo in patients receiving other treatments for PD. Overall, 101 patients with early untreated PD were included and were divided into four groups after randomization: selegiline and levodopa/carbidopa, placebo and levodopa/carbidopa, selegiline and bromocriptine, or placebo and bromocriptine. The UPDRS scores were analyzed at 14 months, after a washout period of 2 months for the study drug (selegiline) and 7 days for the treatment drug (levodopa/carbidopa or bromocriptine). The effect of selegiline was noted, with a difference in UPDRS scores of 5.4 points compared with the placebo group. The authors concluded that a neuroprotective effect was demonstrated for selegiline, because the washout period was believed to be long enough to exclude a purely symptomatic effect. However, others have suggested that long-term effects of selegiline might explain the difference in motor scores [46].
TEMPO
The TEMPO (TVP-1012 in Early Monotherapy for Parkinson’s Disease Outpatients) study [47] examined the effects of rasagiline in early PD. It was designed as a parallel-group, placebo-controlled study. Two doses of rasagiline, 1 and 2 mg/day, were studied. A total of 404 patients with early PD were included. At the end of the study, the UPDRS score was improved with rasagiline compared with placebo, for both 1 mg (−4.20 points) and 2 mg (−3.56 points) groups. An extension study revealed worse scores in patients with deferred treatment (originally receiving placebo), which could be consistent with neuroprotection, although the extension study was not specifically designed for this purpose.
ADAGIO
The ADAGIO (Attenuation of Disease Progression with Azilect Given Once-daily) trial [33] was the first study that used the delayed-start design. Two different doses of rasagiline were studied (1 and 2 mg/day). Early treatment with rasagiline 1 mg resulted in a positive outcome for all three predefined endpoints (as outlined above), with a difference of 1.7 points in the UPDRS score at the end of phase 2. The 2 mg dose group, however, failed to fulfill the predefined criteria. Post-hoc analysis of the 2 mg group data showed that the quartile with the highest UPDRS scores had positive results for all three outcomes. This has been interpreted as possible disease-modifying effects, masked by symptomatic effects. However, other authors have criticized the study and argued that the difference of 1.7 UPDRS points was not clinically significant [36, 48]. The shortcomings of the delayed-start trial design, as described above, have also been cited to claim false-positive results for the 1 mg dose, and the absence of a neuroprotective effect [49, 50].
Levodopa
ELLDOPA
In the past, concern had been raised about the potential toxic effects of levodopa through increased dopamine turnover and resultant increased oxygen reactive species [51, 52]. The ELLDOPA (Earlier versus Later Levodopa Therapy in Parkinson Disease) study [8] was designed to analyze the effects of levodopa on the progression of PD. Patients were randomized to three doses of levodopa/carbidopa (50/12.5 mg three times daily, 100/25 mg three times daily, or 200/50 mg three times daily) or placebo. Assessment of motor scores was done at 42 weeks, after a 2-week washout period from the study drug. A significant dose–response therapeutic benefit was observed, with worsening of UPDRS scores during the washout period, which was more pronounced for the higher dose. After the 2-week washout, the difference in UPDRS scores between active-treatment and placebo groups was sustained, which was interpreted as possible disease modification. However, these results have been contested, because the washout period necessary for complete removal of therapeutic effects might be as long as 1–2 months [53]. An imaging substudy with [123I]β-CIT SPECT was conducted in a subgroup of 142 patients. When patients with a normal baseline scan were excluded, the decrease in [123I]β-CIT was faster in levodopa-treated patients (−1.4% with placebo compared with −6.0, −4.0 and −7.2% for the 150, 300 and 600 mg/day levodopa doses, respectively; P = 0.036). The discordance between clinical and imaging findings has created problems in interpreting the findings of the ELLDOPA study. Nevertheless, this study reassured patients and physicians about the risk of toxicity induced by levodopa, as a 2-year follow-up did not demonstrate worsening of motor scores.
Related posts:
Oral dopamine agonists in the management of Parkinson’s disease
Management of disease-related behavioral disturbances in Parkinson’s disease
A neurobehavioralist approach to the management of cognitive impairment in Parkinson’s disease
Lessons learned: symptomatic trials in early Parkinson’s disease
Lessons and challenges of trials for other nonmotor complications of Parkinson’s disease
Lessons and challenges of trials involving ancillary therapies for the management of Parkinson’s disease
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
