Introduction
It is now considered that the pathophysiological process of Alzheimer's disease (AD) begins years, if not decades, before the development of dementia. Thus the clinical course of AD starts with a presymptomatic or preclinical phase.Reference Sperling, Aisen and Beckett1 Affected subjects will develop symptomatic AD if they live long enough, the timing being influenced by cognitive reserve and other factors that are currently unknown. This review is based upon a review of published papers from the past 5 years identified from PubMed. Search terms used were: Alzheimer's disease, preclinical stages, amyloidosis, neurodegeneration and cognition.
Pathology of Alzheimer's disease
A definitive diagnosis of AD requires histopathological evidence from brain biopsy or autopsy. Neurofibrillary tangles (NFTs) and senile plaques (SPs) are the main neuropathological hallmark. However, lesions such as NFTs and SPs can also be observed in cognitively normal older people. The senile neuritic plaques are spherical microscopic lesions with a core of extracellular Aβ surrounded by degenerating axonal endings. The Aβ peptide is derived from the amyloid precursor protein (APP), which undergoes proteolysis by α, β and γ secretase enzymes. When cleaved by α secretase a soluble peptide derivative of amyloid is formed, but when cut by β secretase first γ secretase generates two peptides, Aβ40 and Aβ42. The Aβ peptide monomer binds other Aβ peptides forming oligomers, ultimately leading to the accumulation of amyloid in the brain.
Neurofibrillary tangles are fibrillary, cytoplasmatic structures within the neurons. The electron microscope shows paired helical filaments of a protein called tau. In AD, tau aggregates due to hyperphosphorylation, possibly secondary to the aggregation of extracellular Aβ. Once tau aggregates, it forms twisted paired helical filaments, recognized as neurofibrillary tangles. Most of the cortical areas with a high density of tangles also display significant loss of synapses and neurons.
Cell loss, senile plaques and tangles appear regularly in the neocortex, especially in the association areas, in the hippocampus including the entorhinal cortex, in the amygdala and in the basal nucleus of Meynert. Other neurological features of AD include granulovacuolar degeneration and amyloid angiopathy.
In preclinical AD, the neuropathological changes of the disease are developing, but have not yet caused any noticeable clinical symptoms. The onset of very mild dementia is related to neuronal and synaptic loss arising from the accumulation of the NFTs and SPs. A large body of evidence is pointing to the fact that the pathology of AD represents an insidious process starting 15–20 years before any clinical manifestations can be seen in the patient. The delay between the pathological and the clinical manifestations of AD emphasizes the challenge of making an early diagnosis and the need for biological markers of the disease.
Biomarkers of preclinical stages of AD
The preclinical phase of AD is primarily defined by changes in biomarkers. At present the diagnosis is relevant only to research as these biomarkers have not been shown to have adequate sensitivity or specificity for use in clinical practice, and no treatments for preclinical AD are yet available.
The biomarker model (Figure 1) of the preclinical stage of AD was adapted from the original graph proposed by Jack et al.Reference Jack, Knopman and Jagust2 It has the following features:
(1) Firstly, Aβ biomarkers become abnormal, with a substantial Aβ load accumulating before the appearance of clinical symptoms. The brain Aβ accumulation is necessary but not sufficient to produce clinical symptoms of mild cognitive impairment (MCI) and dementia.
(2) Biomarkers of synaptic dysfunction become evident, including decrease of fluorodeoxyglucose 18F (FDG) reuptake in temporo-parietal regions on positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). The severity and changes over time of these synaptic markers correlate with clinical symptoms in MCI and AD dementia.
(3) Brain atrophy on structural magnetic resonance (MRI) has a characteristic pattern involving the medial temporal lobes, paralimbic and temporoparietal cortices. It is a marker of neuronal loss and MRI retains a close relationship with cognitive performance through the clinical phases of MCI and dementia.
(4) None of the biomarkers is static: rates of change in each biomarker vary over time and follow a non-linear time course.
Model of preclinical Alzheimer's disease
Stages of preclinical AD
If a continuum in the preclinical stages of AD is accepted, it is possible to suggest the following schema, although this will probably require further modifications as new findings emerge. The feasibility of defining these stages for recruitment of clinical research cohorts is also still unclear.
Stage 1: the stage of asymptomatic cerebral amyloidosis
These individuals have elevated tracer retention on PET amyloid imaging and/or low Aβ1–42 in cerebrospinal fluid (CSF) assay, but no detectable evidence of additional brain changes suggestive of neurodegeneration. The standard for determining ‘amyloid positivity’ is still evolving. It remains unclear whether there is a biologically relevant continuum of Aβ accumulation, or if there is a clear threshold or ‘cut-off’ value that can be defined on the basis of predictive value for subsequent clinical decline.Reference Shaw, Vanderstichele and Knapik-Czajla3 It also remains unknown whether this threshold should be adjusted for age or genotype.
Although limited, current data suggest that soluble or oligomeric forms of Aβ are in balance with plaques, which may serve as reservoirs. It remains unknown if there is any identifiable pre-plaque stage in which only soluble forms of Aβ would be present. Thus it may be possible, in the future, to detect a stage of the disease preceding stage 1.
PET imaging of the amyloid-binding agent Pittsburgh compound B (PET-PiB) allows a semi-quantitative, in vivo analysis of brain Aβ load and its spatial distribution. This is closely correlated with amyloid plaque burden at autopsy and is inversely related to CSF Aβ42 levels.Reference Ikonomovic, Klunk and Abrahamson4
Recent results from clinical trials of the F-labelled tracer F-AV-45 (Flobetapir®) are promising for the next generation of amyloid imaging. It has several characteristics that make it suitable for Aβ plaque imaging: excellent binding affinity, high selectivity, excellent brain penetration and rapid kinetics.Reference Mintun, Saha and Fleisher5
Stage 2: amyloid positivity plus evidence of synaptic dysfunction and/or early degeneration
The current markers of neuronal injury with the greatest validation are:
(1) Increased levels of CSF tau or phosphor tau can be defined as markers of neuronal injury from multiple causes and can be seen in neurodegenerative disorders other than AD. Studies have consistently shown that phosphorylated tau offers equivalent (or possibly better) diagnostic utility for AD than total tau. Enzyme-linked immunosorbent assays have been developed to recognize various phosphorylated epitopes (p-tau 231, p-tau 181, p-tau199), all of them equally effective in discriminating AD patients from healthy subjects.Reference Hample, Buerger and Zinkowsky6
(2) Hypometabolism in an AD-like pattern on FDG PET. A reduction of glucose metabolism has been seen on PET imaging in the bilateral temporal regions, precuneus and posterior cingulate cortex. Hypometabolism on FDG PET is also correlated with CSF Aβ42 levels and cognitive function.Reference Jagust, Landau and Shaw7
(3) Cortical thinning/grey matter loss in a specific anatomical distribution and/or hippocampal atrophy on volumetric MRI. The hierarchical patterns of cortical atrophy vary greatly over time, but there is consensus that the first MRI signs of emerging AD are related to atrophy of the medial temporal lobe (particularly the hippocampus) and cortical thinning in vulnerable regions.Reference Dickerson and Wolk8 Shrinkage in the entorhinal cortex may precede significant cognitive decline by about 9 years and, in conjunction with the hippocampal volume, can predict cognitive decline in normal subjects with accuracy of about 80%. Decreased grey matter volume in the parietal lobe, especially in the angular gyrus, has also been described in normal subjects in advance of development of MCI. Atrophy in these areas is also correlated with the Aβ42/tau ratio in CSF and amyloid load measured by PiB PET binding.Reference Becker, Hedden and Carmasin9
The functional connectivity between different brain regions is disrupted, possibly reflecting the deleterious effect of Aβ on synapses and glucose metabolism. Using fMRI, it has been shown that in individuals who are APOε4 carriers, the magnitude of brain activation in the parietal and prefrontal regions during memory tasks is higher than in controls and the extent of brain activation correlates with subsequent memory decline.Reference Seidenberg, Guidotti and Nielson10 Future markers may also include fMRI measures of default network connectivity. Sensitive imaging biomarkers that can detect early synaptic dysfunction and functional and structural disconnection, such as diffusion tensor imaging (DTI) and magnetoencephalography (MEG), may one day prove to be useful to track early response to amyloid-lowering therapies.Reference Fernandez, Turrerro, Zuluaga, Gil Gregorio, del Pozo, Maestu and Moratti11
Stage 3: amyloid positivity, plus evidence of neurodegeneration, plus subtle cognitive decline
Individuals with subtle cognitive decline are in the last stage of preclinical AD and are approaching the stage with the proposed clinical criteria for MCI. The research criteria for the preclinical stage of AD have always included cognitive impairment as a relevant variable. Dubois and colleagues, in their first review in 2007,Reference Dubois, Feldman and Jacova12 introduced episodic memory impairment as an independent factor in the diagnosis, whilst also outlining the abnormal biomarkers and neuronal neurodegeneration that accompany the preclinical stage of the disease. The memory impairment also needed to be progressive for at least the last 6 months, as perceived by the patient and even by the family. Subsequently a change in other cognitive functions has been described, with detailed neuropsychological assessment identifying differences between normal healthy ageing subjects and people at risk of developing AD up to 8–10 years before the diagnosis. Whilst most of the studies have focused on verbal episodic memory as an initial cognitive symptom of preclinical AD,Reference Sperling, Aisen and Beckett1, Reference Dubois, Feldman and Jacova12, Reference Ewers, Insel, Stern and Weiner13 impairment in functions like semantic memory, executive dysfunction, processing speed and visuospatial perception have also been shown.Reference Knopman and Caselli14, Reference Schmid, Taylor, Foldi, Berres and Monsch15
Most of these studies have had a longitudinal design, with a diagnostic battery including repeated neuropsychological assessment. This presents a challenge in deciding appropriate cut-off points in cognitive performance indicative of preclinical AD.Reference Knopman and Caselli14 Some authors theorize that neuropsychological assessment should focus on changes or inflection points in the task presented and in the speed of the decline in cognitive functions over time.Reference Schmid, Taylor, Foldi, Berres and Monsch15 The best way to know if a subject is declining is by making repeated measures of cognition over several years, using parallel forms to avoid practise effects.
Related to the problem of finding an appropriate cut-off point, recent studies are pointing to the importance of qualitative changes in cognitive assessment, based on the idea that these may precede the development of quantitative changes evident in the neuropsychological measures. Thus factors such as subjective memory complaints, or intrusion errors in the assessment of episodic memory, or number of errors and repetitions in tasks of verbal fluency must be recorded as part of cognitive assessment, as these could predict the future development of AD.Reference Schmid, Taylor, Foldi, Berres and Monsch15 Some researchers studying preclinical AD are focusing on memory complaints and self-rating of cognitive abilities, based on the thinking that the subject is likely to be most aware of minor changes in cognition over time.
Neuropsychological assessment needs to consider a number of possible confounding influences that may impact on performance: visual and hearing impairments, physical limitations and neurological status, educational level and effect of pharmacological treatments are just some.Reference Yubero, Salmon and Gil Gregorio16 Ageing itself is associated with cognitive changes, mostly with frontal lobe function and episodic memory, leading to a decline in reasoning, memory and processing speed beginning in the forties.Reference Knopman and Caselli14 There are also studies comparing early versus late onset of AD, demonstrating worse performance on visuospatial functioning in younger patients and better performance on memory tasks.Reference Mandal, Joshi and Saharan17
The cognitive reserve hypothesis presents another challenge to detecting preclinical AD through use of neuropsychological assessment. Many studies indicate successful compensation for cognitive impairment in people with cognitive reserve, not only in cognitive test performance, but also on FDG-PET imaging.Reference Ewers, Insel, Stern and Weiner13, Reference Schmid, Taylor, Foldi, Berres and Monsch15 Studies also show poorer performance on tasks such as digit symbol, phonetic fluency and delayed recall in people with lower education and so less cognitive reserve.Reference Knopman and Caselli14
Based on these findings, it may be concluded that brief cognitive screening tests are not useful instruments in the evaluation of preclinical AD. A more wide-ranging cognitive battery, including assessment of verbal and visual episodic memory, processing speed, executive tasks and visuospatial functions, may be more discriminatory. Classical tests used for the evaluation of episodic memory in older people and with some evidence of utility in identifying people at risk of developing AD are the Auditory Verbal Learning Test, the California Verbal Learning Test, the Free and Cued Selective Reminding Test and the Logical Memory or Visual Reproduction tests of the Wechsler Memory Scale. Other tasks that may have utility in the assessment of preclinical AD are the Trail Making Test, Block Design from the Wechsler Adult Intelligence Scale (WAIS), Verbal Fluency, and classical tests used for the evaluation of executive function. In addition, measures such as the Functional Activities Questionnaire (FAQ) or the Everyday Cognition Scale may be useful for documenting and quantifying any self-perception of cognitive impairment.
Conflicts of interest
None.
