Each of these stages can be targeted for specific drug treatments, requiring different trial designs and outcomes:
• a healthy elderly population can be tested with a safe antioxidant such as Ginkgo biloba versus placebo over five1 to seven years using incident dementia as the primary outcome;2
• an aMCI population could be tested with a symptomatic or a potential DMD versus placebo over three years using time to diagnose dementia as the primary outcome;3
• patients with mild to moderate AD must be offered the usual standard of care to which a potential DMD or placebo is added over 12–18 months, with a slower decline on cognitive, global, or functional autonomy as primary end-points;4 and
• patients with severe AD are usually in a nursing home setting requiring shorter six-month studies with appropriate cognitive, functional, and behavioral outcomes.5
Disease milestones have also been defined in AD6 and can be used as endpoints in a survival analysis. For instance, α-tocopherol delayed loss of residual autonomy and progression to severe dementia or death in one study by the Alzheimer Disease Co-operative Study group (ADCS),7 thus influencing clinical practice to use vitamin E in all stages of AD (at least in the US until the negative results in the Memory Impairment Study),8 and a meta-analysis of clinical studies using high-dosage vitamin E suggesting a higher mortality rate.9
Symptomatic domains in dementia include cognition, activities of daily living (ADL), and behavior. In many patients, early changes in mood and anxiety precede the formal diagnosis of AD. Cognitive, functional, and global decline are relatively linear over time, whereas neuro-psychiatric symptoms peak midway into the disease course and improve spontaneously through the severe stage as mobility becomes impaired with emerging Parkinson-like physical signs.10 These natural changes in the symptomatic domains through the stages of AD have an impact on trial design and outcomes.11
Symptomatic Clinical Trials using Cholinesterase Inhibitors and Memantine
The modern treatment for AD was initiated by the report that tacrine improved some aspects of cognition and daily life.12 The follow-up confirmatory studies used cross-over and parallel-group designs. The latter design group offers the possibility of short-term (minimum of three-month) studies comparing the efficacy of different doses of the drug with placebo.
The primary analysis is carried out on outcomes at end-point, using the last observation carried forward (LOCF) or intent to treat (ITT) to compensate for missing values in case of drop-outs. Although LOCF/ITT has been favored for symptomatic studies, there is a trend for using observed cases (OC), e.g. completers, for studies of 12 months or longer. For practical purposes, both types of analysis are performed. Although ‘cognitive enhancement’ was the main hope for cholinesterase inhibitors (CIs) as a therapeutic class, the reality that has emerged from six-month studies with open-label extensions and the one-year placebo-controlled Nordic study13 is that although there is a small but statistically significant improvement in cognition peaking at three months with CIs, the most clinically relevant finding has been the stabilization of cognitive decline with ‘return to baseline’ at nine to 12 months for the actively treated groups at the higher therapeutic doses compared with placebo-treated groups, who decline steadily. These results were considered sufficient for CIs to become ‘standard treatment’ for AD in the mild to moderate stage,14–16 notwithstanding the challenges to the effectiveness of CI17 using the arguments that the trial designs were flawed;17 a pragmatic study over three years comparing donepezil with placebo was negative for the primary end-point of delaying entry to institutional care and progression of disability;18 and there was insufficient cost–benefit, at least in mild stages.19 Memantine acting though N-methyl-d-aspartate (NMDA) receptor inhibition has been found to be effective in studies using parallel group, in moderate-to-severe AD.20 Guidelines appropriate for this stage of disease have been used and accepted by regulatory agencies. The novel design of adding memantine or placebo to a stable dose of a CI has been used successfully,21 and is of great importance as it has paved the way for a number of studies in which novel drugs or placebo are added on to ‘standard treatment.’ Post hoc analysis demonstrated an antiagitation/ aggression effect,22 which is being studied prospectively in Canada. These non-cognitive effects may turn around the current negative appraisal by the National Institute for Clinical Excellence (NICE) regarding the effectiveness of memantine.19
Although no drug has yet been determined to delay disease modification, attempts are under way using parallel groups over one year or longer with the novel agent or a placebo added on to standard treatment. Outcomes known to have relatively linear changes over time are used, such as the Clinical Dementia Rating Sum of Boxes (CDR-SB),23 the Alzheimer’s Disease Assessment Scale–Cognitive Component (ADAS-cog), the Alzheimer’s Disease Co-operative Study (ADCS)-ADL,24 or the Disability Assessment in Dementia (DAD),25 supplemented by volumetric brain measurements using magnetic resonance imaging (MRI) at the beginning and the end of treatment. As an example, tramiprosate acting as gag-mimetic26 has been tested in mild AD over 78 weeks with changes from baseline to week 78 in ADAS-cog, CDR-SB, and DAD as the main clinical outcomes, with the rate of brain atrophy being calculated using MRI. It is possible that delaying progression from CDR total score of 1 (mild dementia) to 2 (moderate dementia) will be the most convincing argument of effectiveness. The next study using flurbiprofen will be completed in June 2008.
Patterns of responses based on phenotype (age at onset of disease, severity at onset of treatment, rapid decliners) and genotype apolipoprotein E (ApoE) will be of great interest, since it is likely that subgroups of patients with AD will respond well to a given class of drugs or a drug within a class, allowing for a pharmacogenomic approach to the use of DMD. It is encouraging that European regulators have initiated discussions on these matters with investigators and sponsors.27
We have gained a better understanding of the natural history of AD and have developed appropriate trial designs and outcomes for the various stages of this condition. There is clear benefit for the treatment of symptoms in mild to severe AD using CIs and memantine. There is cautious optimism for successful disease modification using a number of agents currently under study. Treatment guidelines must be constantly updated to take into account new evidence for the ultimate benefit of patients and care-givers.