Clinical trials—An overview

Clinical trials generally evaluate therapeutics in four phases before a new drug can be approved for use in the clinic. Each of these phases has a different goal, with results from the prior phase being used to design the next phase. Phase I trials evaluate a drug for the first time in people and are often short in duration with a small sample size (5–50 patients). The goal is to establish a potential treatment’s safety profile and identify a safe dosage range. Phase II trials usually run 6–24 months in duration and have a larger sample size (50–400 patients), while seeking to evaluate a treatment’s safety as well as any impact on biomarkers. Phase III trials are typically performed with a large sample size (1000–4000 patients) and attempt to compare a treatment’s efficacy against standard treatments, if any exist. Phase III studies are usually a multicenter effort and can take, particularly for slowly progressive disease such as Alzheimer’s disease, many years to enroll and complete. Typically, two phase III trials must be conducted for a new drug to win approval with the Food and Drug Administration (FDA). Finally, phase IV trials are performed after a new treatment has received FDA approval. These postmarketing studies help identify and evaluate long-term side effects and further establish effectiveness in clinical practice. The progression of a drug through these phases is meant to establish its clinical effectiveness in a safe, methodical, and reproducible way.

For a clinical trial to evaluate a new therapeutic, there are a number of key elements that must be considered. This includes (1) clear identification of the trial’s main objective, (2) selection of the appropriate primary and secondary endpoint(s), (3) use of adequate methods of data collection, (4) choice of sample to be included, (5) sample size with scientific justification, (6) method of handling data (including missing data), (7) statistical methods and assumptions for accurate interpretation of results. This comprehensive plan is termed the “clinical trial protocol.”

For a clinical trial to be unbiased in evaluating the efficacy of a drug, there should be a clearly defined treatment group or “active” arm and a single control or “placebo” arm. However, often times, phase I trials are open label, that is, they are unblinded, with all participants receiving treatment, and both the participants and study teams being aware that treatment is given. However, phase II and III studies are almost always placebo controlled. The placebo arm is essential to allow for blinding of participants and study teams to the assigned arm to avoid biases. The placebo, a pharmaceutically inert substance identical in appearance to the treatment being evaluated, is the clinical researcher’s analog to the bench scientist’s control experiment. To prove that a new treatment is effective, the study team compares the results of the experimental treatment for an illness with those obtained from the placebo. The placebo-controlled trial is therefore regarded as the “gold standard” for testing the efficacy of new treatments. Many studies will include an open-label (unblinded) extension after the blinded phase. This allows participants to get access to the treatment under evaluation (after the double-blind phase) and provides additional safety and efficacy data on the treatment to the study team.

A power analysis is performed prior to starting a clinical trial and is often reported with the published results. This analysis determines the sample size needed to detect the expected treatment difference between the placebo and treatment groups on specific variables or outcome measures. Determining the optimal sample size for a study assures an adequate power to detect statistical significance. The sample size estimate takes into account the natural rate of change in a selected outcome in a specific population as well as the expected effect size of the drug. The variance in both rate of change and effect size is also critical in estimating sample size.

Well-designed clinical trials often specify how the treatment group is allocated. Most trials utilize randomization, the process by which participants are randomly allocated to either the intervention or control group, hence the term “Randomized Clinical Trial (RCT).” Both the method of randomization and the ratio of randomization (usually 1:1 treatment to placebo ratio) are specified in advance of study start. Participants are usually randomized using a randomized block scheme with blocks of varying size. The block scheme refers to the number of enrolled participants who are randomized at a particular time. By using blocks of varying size, there is a reduction in selection bias while helping ensure that each arm of the study will contain an equal number of participants. Stratified randomization refers to the situation in which strata are constructed based on values of prognostic variables and a randomization scheme is performed separately within each stratum. Strata are constructed based on the sample size and randomization is performed separately within each stratum. This ensures balance between the treatment and placebo arms with respect to these variables. Stratification should be considered when performing small trials (sample size < 300 patients), planning interim analyses for larger trials using small numbers of patients, or within set limits for type I (alpha) and II (beta) errors. A type I error is the chance of a trial erroneously concluding that the difference in an outcome is due to the treatment when it was really due to chance. A type II error is the chance of a trial erroneously concluding that there is no effect of the treatment on outcome when there actually is a differential effect. In most clinical trials, the type I error is set at 5% and type II error at 20%. Power is the chance that a trial will correctly conclude that the difference in outcomes is due to the treatment (i.e., a difference in outcomes is detected and this difference is truly due to the drug). Power is calculated as 1 minus type II error; therefore, most clinical trials are designed to have 80% power.

Finally, a statistical analysis plan is created prior to beginning any clinical trial. This plan includes how the primary and secondary outcomes will be analyzed, any planned statistical modeling, how missing data is handled or imputed, and any proposed subgroup analyses. Analyses in trials can be classified as either an intention-to-treat analysis (ITT), where a comparison of the treatment groups that includes all patients as originally allocated after randomization or a per-protocol (PP) analysis, where there is comparison of treatment groups that includes only those patients who completed the treatment originally allocated. The use of ITT analysis has become the de facto standard for analysis as it ensures maintenance of comparability between the treatment arms as obtained through randomization, maintains sample size, and eliminates bias. Guidance for these activities has been under the Consolidated Standards of Reporting Trials (CONSORT) guidelines which is an evidence-based minimum set of recommendations for reporting randomized clinical trials. It offers a standard way for authors to prepare reports of trial findings, facilitating their complete and transparent reporting, reducing the influence of bias on their results, and aiding their critical appraisal and interpretation. It consists of a 25-item checklist and a participant flow diagram, along with some brief descriptive text. The checklist items focus on reporting how the trial was designed, analyzed, and interpreted; the flow diagram displays the progress of all participants through the trial. The CONSORT guidelines for reporting of “parallel group randomized controlled trials” recommend that both ITT and PP analyses should be reported for all outcomes to allow readers to interpret the effect of an intervention [1].

Clinical trials for Alzheimer’s disease in Down syndrome

In considering the design of clinical trials for Alzheimer’s disease (AD) in Down syndrome (DS), it is essential to clearly understand the characteristics of the ideal sample population (e.g., age range, baseline intellectual functioning), the behavior of the intervention (e.g., pharmacologic mechanism of action, compound’s safety profile), and to select the correct outcome measures (e.g., clinically meaningful measure of efficacy) as described in the earlier section. In addition, the US Code of Federal Regulations 21 CFR Parts 50 and 56 outline specific requirements to enhance protections for “vulnerable populations” including individuals DS, which further raises issues on the need to utilize the most appropriate informed consent and assent processes for persons with developmental disabilities.

In DS, where intellectually disability (ID) is lifelong and precedes dementia, the question arises as to what degree the changes in cognition and functioning can be related to the emergence of AD. Baseline ID in DS is a relatively static process related to inherent cerebral dysfunction and not associated with progressive neurodegeneration and cerebral atrophy, whereas declines in memory and other cognitive domains are progressive. Results from recent biomarker studies suggest these techniques may be able to accurately discriminate the progressive brain changes due to AD from the baseline ID in DS [2]. Aging adults with DS (over 40) demonstrate cognitive decline in discrete phases. Early in the disease process isolated memory deficits emerge and slowly progress [3] in a manner that is congruent with the mild cognitive impairment stage of AD in the sporadic population. Following this, affected individuals experience a decline in cognitive function that occurs alongside function decline, and dementia ensues [4]. Thus older adults with DS constitute a population in which the neuropsychological profile of cognitive change due to AD can be observed in conjunction with biomarker changes decades before dementia occurs.

The DS population will require a unique set of considerations in terms of sensitive and valid assessments. Baseline assessments will be critical as will be the need for carefully planned and consistent testing sessions. Both direct neuropsychological measures as well as observer-rated scales will likely be required for accurate assessment of AD in DS. The key will be to identify measures with the greatest dynamic range that overlaps with changes in biomarkers of AD. Here, there are a number of possible neuropsychological measures that are being validated in longitudinal natural history and biomarker studies of AD in DS. Additional considerations include floor and ceiling effects, test-retest reliability, as well as the clinical meaningfulness of scores on such neuropsychological tests. The strategy of a trial ready cohort and the utilization of run-in data will be essential to conducting clinical trials in this population. A similar approach has been utilized in the Dominantly Inherited Alzheimer Network (DIAN-TU) and its associated Trials Unit (DIAN-TU) for autosomal dominant AD and in the Trial Ready Cohort for Preclinical and Prodromal Alzheimer’s Disease (TRC-PAD). Both of these efforts have focused on identifying participants who are more likely to be eligible for clinical trials based on inclusion/exclusion criteria. This will be all the more important for the DS population given the need to ensure stable preexisting medical conditions and ability to participate in various clinical trial procedures which often include MRI, and PET scans as well as blood collection and cognitive testing.

Amyloid as a drug target

The “amyloid hypothesis” posits that soluble Aβ fragments effectuate plaque accumulation, intracellular neurofibrillary tangles (NFTs), neurodegeneration, and subsequent dementia [5]. Supporting this notion is evidence that amyloid precursor protein (APP), whose gene is located on chromosome 21 and triplicated in DS, contributes to an increased prevalence of dementia in affected individuals (see Chapter 1). Moreover, it has been shown that inherited forms of AD that are associated with mutations within the APP gene lead to autosomal-dominant forms of this disease [6–8]. Further supporting the link between triplication of the APP gene and overproduction of Aβ and dementia in DS are two published cases of individuals with partial trisomy of chromosome 21 who were disomic for the APP gene and did not develop dementia or AD pathology [9, 10]. Finally, evidence suggests that a mutation that leads to a reduction in Aβ formation [Icelandic mutation (A673T)] reduces the risk for AD dementia [11]. Given the pivotal role of Aβ in neuropathology, it is not surprising that several candidate therapeutics have been deployed to mitigate production of Aβ, inhibit enzymes that process APP (e.g., β-secretase, β-amyloid cleavage enzyme (BACE)), and facilitate amyloid plaque clearance (e.g., anti-Aβ immunotherapy) [12].

One of the guiding aims of current clinical trials for the prevention of AD dementia is earlier detection and intervention. AD is known to present on a continuum of preclinical, prodromal, and dementia stages. Persons who are asymptomatic yet exhibit AD neuropathology, as evidenced by amyloid PET or CSF Aβ levels, represent the preclinical stage [13]. Individuals who exhibit significant impairment in a single cognitive domain represent the prodromal stage. Persons experiencing cognitive deficits across multiple domains along with functional deficits represent the dementia stage. Identification of persons with AD would ideally occur during the preclinical stage of AD so that disease-modifying treatment could be administered earlier in the disease process. Recognizing this, the Food and Drug Administration (FDA) has indicated that an effect of a candidate therapeutic on a single primary efficacy measure may be considered for approval in the context of positive biomarkers [14].

In DS, the triplication of an extra copy of APP on chromosome 21 results in an increase in APP mRNA and protein expression along with increased Aβ production [15]. The excess production of Aβ in persons with DS effectuates AD-like neuropathological changes by the age of 40 [16] (see Chapter 2). Cholinergic losses present in the brains of persons with DS are identical to those seen in AD [17]. Yet in contrast to sporadic AD, amyloid plaques and NFTs start accumulating in people as early as 12 years of age in DS [18]. Recent data indicate that changes in AD biomarker in persons with DS are similar to those observed in familial and sporadic AD. There is a sixfold increase in plasma Aβ in individuals with DS as compared to age-matched non-DS individuals [19]. The results of amyloid positron emission tomography (PET) imaging from individuals with DS are consistent with those seen in non-DS individuals with AD [20–23]. The presence of the APOE ɛ4 allele results in a greater risk of early onset of dementia in adults with DS [24] in a manner similar to that seen in familial and sporadic AD [25–27], probably as a result of excess Aβ accumulation. Postmortem studies indicate that adults with DS exhibit prominent patterns of cerebral atrophy of the medial temporal lobe structures congruent with the presentation observed in early stages of familial and sporadic AD [28–30]. Volumetric MRI studies of age-related brain changes in DS demonstrate patterns of hippocampal-specific atrophy similar to that observed in persons with AD, a change that correlates with memory measures [31].

Clinical trials targeting amyloid can focus on reduction in the production of Aβ, accelerating removal of Aβ, or blocking the expression of APP (which is triplicated in DS and thought to drive AD pathology). An important consideration when planning clinical trials for AD in DS targeting amyloid will involve an assessment of cerebrovascular disease, in particular cerebral amyloid angiopathy.

Dyrk1A as a drug target

The DYRK1A gene, located on the 21st chromosome, is thought to play an important role in the pathogenesis of the intellectual disability associated with DS by virtue of its involvement in neuronal differentiation [32]. Its overexpression as studied in transgenic mice (TgDyrk1A) is thought to lead to cognitive impairment and less complex dendritic trees [33]. Epigallocatechin gallate EGCG is a flavonoid and natural antioxidant found in green tea and has been shown to be a potent DYRK1A inhibitor [34, 35]. A phase II clinical trial involving 31 DS participants ages 14–29 years enrolled in a 6 month long, placebo-controlled study of oral EGCG (9 mg/kg/day) found significant effects on episodic memory function after 3 months of treatment [36]. The primary outcome measures were as follows: Pattern Recognition Memory (PRM), Fuld Object Memory Evaluation (FULD), and Paired Associates Learning (PAL) as well as Plasma homocysteine, NAD(P)H:quinone oxidoreductase activity and dyrk1a gene expression in lymphocytes. Overall, EGCG was well tolerated but no significant impact was seen on efficacy. An additional study was conducted with a combination of both EGCG and cognitive training for 12 months and was found to be significantly more effective than either placebo or cognitive training at improving visual recognition memory, inhibitory control, and adaptive behavior. Phase 3 trials with a larger population of individuals with Down’s syndrome are needed to confirm the potential efficacy of EGCG and cognitive training [37].

The Alzheimer’s Biomarker Consortium for Down syndrome

As the largest AD biomarker study in DS to date, the landmark Alzheimer’s Biomarker Consortium for Down syndrome (ABC-DS) is setting the stage for powering clinical trials for AD in DS. Similar to the three largest ongoing longitudinal studies of AD biomarkers in the general population: the Alzheimer’s disease Neuroimaging Initiative (ADNI), the Dominantly Inherited Alzheimer Network (DIAN), and the Alzheimer Prevention Initiative (API), it is looking at deep phenotyping of the natural history of AD in DS. This 5-year longitudinal study launched in 2015, and now beginning its second 5-year cycle, is examining the progression of AD-related biomarkers (Aβ-, tau-, and fluorodeoxyglucose-PET, MRI, cerebrospinal fluid plasma biomarkers, and neuropathology) as well as cognitive and functional measures in over 500 adults with DS [38]. The data from ABC-DS will enable secondary prevention trials for AD in DS by anchoring standard AD biomarkers with cognitive and clinical outcomes. It will also allow sample size estimates and a clear understanding of the natural history of rates of change of AD biomarkers in adults with DS.

Treatment trials in DS

The conduct of clinical trials in the DS population raises many methodological challenges. Given the wide variations in baseline intellectual capabilities and cognitive functioning, these differences must be taken into account for accurate assessment in testing situations. The main defining feature of dementia in the typical population is a decline from the baseline level of function and performance of daily skills. Consensus guidelines for the evaluation and management in DS have been proposed [39]. However, the earliest signs of dementia in adults with DS may be subtle and will often require an astute observer. In addition, concomitant medications, recent medical illnesses (including laboratory testing such as TSH and Vitamin B12), or recent life events that can impact psychosocial functioning must also be considered when interpreting cognitive performance (see Chapters 11, 13, and 14). In individuals with DS, these issues are even further magnified. Although some validated measures exist [40, 41] there is currently no single, standard clinical instrument for the assessment of dementia in adults with DS. Nonetheless, a few small multicenter clinical trials have been successfully conducted in the DS population and have provided important insights into trial design and implementation in this group.

Vitamin E

Oxidative stress as a mechanism of pathology in both AD and the DS brain is supported by studies demonstrating oxidative damage in Alzheimer brain. A randomized, double-blind controlled clinical trial recruited adults with DS older than 50 years to participate. Participants were randomly assigned to receive 1000 IU of vitamin E orally twice daily for 3 years or identical placebo. The primary outcome was change on the Brief Praxis Test (BPT). Secondary outcomes included incident dementia and measures of clinical global change, cognition, function, and behavior. A total of 337 individuals were randomized, 168 to vitamin E and 169 to placebo with no difference between drug and placebo [42].

Memantine

Memantine is approved for the treatment of moderate-to-severe AD by the United States Food and Drug Administration (FDA). Memantine is believed to exert its therapeutic effect by acting as a low-to-moderate affinity, noncompetitive N-methyl-d-aspartate (NMDA) receptor antagonist that binds preferentially to open NMDA receptor-operated calcium channels [43, 44]. This activity-dependent binding blocks NMDA-mediated ion flux and ameliorates the deleterious effects of sustained, pathologically elevated levels of glutamate (excitotoxicity) that may lead to neuronal dysfunction [45]. The efficacy of memantine in the management of patients with AD, vascular dementia, and mixed dementia was assessed in a Cochrane metaanalysis including 12 randomized controlled trials (RCTs) [46]. The metaanalysis showed that memantine was superior to placebo in benefiting cognitive function for mild-to-moderate AD and moderate-to-severe AD [mild-to-moderate AD, using Alzheimer’s Disease Assessment Scale cognitive subscale (ADAS-cog)] [47].

A prospective randomized double-blind trial enrolled adults (> 40 years) with DS, with and without dementia, to receive memantine or placebo for 52 weeks. Eighty-eight patients received memantine and 85 to received placebo. Although the drug was well tolerated, both groups declined in cognition and functioning and rates did not differ between the two groups for any outcomes [48].

Scyllo-inositol

Elevated levels of myo-inositol are thought to contribute to the cognitive impairment in DS [49]. ELND005 (scyllo-inositol), an endogenous myo-inositol isomer, is thought to reduce amyloid toxicity and regulate myo-inositol levels to improve cognitive function in patients with DS. ELND005 (scyllo-inositol; cyclohexane-1,2,3,4,5,6-hexol) has also been evaluated as a potential disease-modifying treatment of AD. In preclinical studies, ELND005 has shown amyloid antiaggregation effects in vitro, protective effects on oligomer-induced neuronal toxicity, and positive effects on learning in animal models of AD [50, 51]. ELND005 has shown both amyloid- and myo-inositol-lowering effects in cerebrospinal fluid (CSF) and brain, respectively, in patients with AD [52]. At a dose with acceptable long-term safety, ELND005 showed beneficial trends on cognition in mild AD [52]. A phase II randomized, double-blind, placebo-controlled study of oral ELND005 in 26 individuals with DS without dementia has been completed [53]. In addition to demonstrating the safety and tolerability of the antiamyloid compound in DS, it also demonstrated the feasibility of conducting a PK/PD study in adults with DS.

ACI-24

ACI-24 is a liposomal vaccine that is designed to elicit an antibody response only against aggregated Aβ peptides without concomitant pro-inflammatory T-cell activation [54]. ACI-24 is based on the truncated Aβ1–15 sequence, which is devoid of T-cell epitopes located closer to the peptide’s C-terminus. The peptide sequence is anchored into the surface of liposomes in such a way that peptides adopt an aggregated β-sheet structure, forming a conformational epitope. Previous active vaccines (e.g., AN-1792) elicited a T-cell response, which led to an increased risk of meningoencephalitis [55]. However, this vaccine’s epitopes for T-cell activation have been removed, such that there is only a B-cell-mediated or humoral response, with generation of Aβ-specific antibodies only.

A phase I/II trial of ACI-24 is ongoing in Europe and aims to address safety, immunogenicity, and efficacy in mild to moderate Alzheimer’s dementia in the typical population. A phase Ib placebo-controlled, multicenter study with ACI-24 for the treatment of AD in individuals with DS was recently completed (NCT02738450). The study enrolled 16 adults with DS aged 25–45 years and treated them with ACI-24 or placebo for 1 year with an additional year of follow-up. Primary endpoints included measures of safety, tolerability, and immunogenicity. Secondary endpoints of this clinical trial are effects on biomarkers of AD pathology as well as cognitive and clinical function. Top-line results are anticipated in 2021. A phase II trial of ACI-24 in DS is being planned.

Horizon21

The Horizon21 DS Consortium consists of various existing DS cohorts from the United Kingdom (the London Down Syndrome Consortium [LonDownS] and the Cambridge Dementia in Down’s Syndrome [DiDS] cohort), Netherlands (the Rotterdam Down syndrome study), Germany (AD21 study group, Munich), France (TriAL21 for Lejeune Institute, Paris), and Spain (the Down Alzheimer Barcelona Neuroimaging Initiative [DABNI]). The consortium will be developing a DS trial-ready cohort that builds on existing studies to establish a large multinational cohort of individuals with DS with run-in data, which will also serve as a registry of potential trial participants [56].

The Alzheimer’s Clinical Trial Consortium for Down syndrome

The ABC-DS project is setting the stage for conducting secondary prevention trials for AD in DS. The NIH-funded Alzheimer’s Clinical Trial Consortium for Down Syndrome (ACTC-DS) recently launched the Trial Ready-Cohort for Down syndrome (TRC-DS). TRC-DS will enroll 120 nondemented participants with DS into a longitudinal safety-run in study with MRI, amyloid PET, tau PET, cognitive testing, and biofluid biomarker analysis in advance of upcoming randomized, placebo-controlled clinical trials for AD in DS. The clinical protocols of both the ABC-DS and TRC-DS are closely harmonized and permit co-enrollment of participants. Data from the ABC-DS project is therefore providing key insights into study design of the clinical trial including sample size selection and duration of treatment period of the planned antiamyloid therapeutic agent.

Conclusions

Over the past decade, great progress has been made in understanding AD in DS utilizing brain MRI, amyloid and tau PET, and biofluid biomarkers. Indeed, several research groups from around the world have shown that there exist remarkable similarities between AD in DS and in other populations. A few clinical trials for AD have been conducted in DS but are now poised to ask whether therapies that are currently being tested for the sporadic and autosomal-dominant forms of AD can be evaluated in people with DS using the latest research tools available.