On the path to 2025: understanding the Alzheimer’s disease continuum

In the century since Alois Alzheimer discovered Alzheimer’s disease (AD), scientists have made remarkable strides in understanding the illness [1], although it was not until the 1980s that two key molecular culprits in disease pathophysiology, amyloid beta (Aβ) and tau proteins, were identified [2, 3].

Historically, the clinical definition of AD was considered probable because it was based on the systematic exclusion of other potential etiologies in a patient with a dementia syndrome and not on positive proof of AD pathology. Thus, AD could be confirmed only through postmortem findings or, rarely, in life by brain biopsy. Later, AD was defined in a broader sense based on clinical manifestations. With basic research advances in recent years, it is now recognized that, like many chronic diseases, pathophysiological changes begin many years prior to clinical manifestations of disease such that the spectrum of AD spans from clinically asymptomatic to severely impaired. As a result, defining AD purely by its clinical presentation is artificial and efforts have been made to recognize the disease based on both clinical and biomarker findings.

Advances with biomarkers have prompted a shift in how the disease is considered as a clinico-pathophysiological entity, with an increasing appreciation that AD should not be viewed with discrete and defined clinical stages, but as a multifaceted process moving along a continuum. Appreciating this concept is critical to understanding the distinction between AD and AD dementia, in order to initiate effective diagnostic and disease management options and for the development of effective disease-modifying therapies (DMTs).

The aim of this article is to discuss how our understanding of AD pathophysiology and the currently available biomarkers and clinical tools can assist with creating a more uniform approach (in both research and clinical practice) to conceptualizing AD as a continuum. This will help set the stage for the future management of AD and help lay a foundation for the prevention or effective treatment of AD by 2025, the year set by global leaders as the target for finding an effective way to treat or prevent AD [4].

Based on currently available information, AD is best conceptualized as a biological and clinical continuum covering both the preclinical (clinically asymptomatic individuals with evidence of AD pathology) and clinical (symptomatic) phases of AD. In the broadest sense, a continuum is defined as a seamless sequence in which adjacent elements (severities) are not perceptibly different from each other, although the extremes are distinct. In AD, this equates to disease progression from an asymptomatic phase, through a long preclinical period during which pathophysiological changes are reflected by increasing biomarker evidence of disease, to the symptomatic phase, during which biomarker changes continue and symptoms of cognitive and then functional impairment become increasingly evident, with the eventual loss of independence and death. These changes in the individual components of the continuum occur in a sequential but overlapping manner.

The etiology of AD is complex and much remains to be fully elucidated. The close link between genetic mutations and disorders associated with AD (mutations of presenilin 1 (PS1), presenilin 2 (PS2), amyloid beta precursor protein (APP), and Trisomy 21) and the accumulation of Aβ strongly implicates this molecule as a pathological driver in AD, but there is controversy over whether Aβ accumulation alone indicates inevitable progression to AD. Furthermore, evidence indicates that Aβ accumulation alone is probably insufficient to produce symptoms [5–7]. At some point during the disease course, additional factors are involved in determining regional neurodegeneration [8]. Tau pathology has been suggested as a facilitator of the downstream effects of amyloid [9]. Other investigators have proposed that synaptic, mitochondrial, metabolic, inflammatory, neuronal, cytoskeletal, myelin, and other age-related alterations may also play a role in the pathogenesis of AD [10].

Based on our current understanding, histopathological characteristics of AD include 1) accumulation of amyloid plaques—extracellular deposition of Aβ protein, both diffuse plaques of amorphous, primarily nonfibrillar Aβ aggregates and neuritic plaques of fibrillar Aβ arranged in a β-pleated conformation; 2) formation of neurofibrillary tangles (NFTs)—intraneuronal bundles of aggregated tau protein, including hyperphosphorylated tau (p-tau), forming paired helical filaments that aggregate within the neurons to create NFTs, leading to disruption of microtubule function, impaired axonal transport, and synaptic and neuronal injury; and 3) neurodegeneration—progressive loss of neurons or their processes (axons and dendrites) with a corresponding progressive impairment in neuronal function and loss of neurons and synapses (atrophy) [11]. Of note, these features individually are not diagnostic of AD. Several lines of evidence currently suggest that in AD the interplay between Aβ and tau is such that Aβ can drive tau pathology and tau pathology may drive Aβ pathology [12, 13]. Our understanding of the pathophysiology of AD is further complicated by primary age-related tauopathy (PART) [14, 15], a neuropathological condition revealed by tau imaging. PART is characterized by medial–temporal neurofibrillary pathology; the pathology remains localized and there are few or no Aβ deposits.

The transition between healthy aging and preclinical AD is not well defined, at least with our current understanding. This shift is likely subtle and without discernible steps; one can suppose that a combination of genetic and environmental factors plays a role in the process [16]. Genetic factors that may contribute as disease modifiers include the apolipoprotein protein E epsilon 4 (APOE4) allele, which conveys an increased risk of disease, and more rapid cognitive decline in the setting of early AD pathology [17]. Other factors that may play a role include cardiovascular risk factors and lifestyle factors such as diet, physical exercise, and cognitive engagement. These lifestyle characteristics influence “cognitive reserve” and onset of objective cognitive decline [18]. The concept of brain or cognitive reserve was originally invoked to provide an explanation for the observation that the extent of AD histopathological changes at autopsy did not always align with the degree of clinical impairment. “Brain reserve” refers to the capacity of the brain to withstand pathological insult, perhaps because of greater synaptic density or a larger number of healthy neurons, such that sufficient neural substrate remains to support normal function. “Cognitive reserve” is thought to represent the ability to engage alternate brain networks or cognitive strategies to cope with the effects of encroaching pathology. While cognitive reserve may help delay the onset of clinical symptoms, once symptoms do emerge the rate of impairment may be greater [18] (i.e., there is a steeper trajectory for clinical impairment) because, while clinical symptoms are delayed, pathological changes progress.

The AD continuum is composed of multiple interconnected components (pathophysiological processes, biomarker findings, and clinical symptoms), each occurring on its own trajectory, with individual trajectories generally parallel to each other but with some temporal offsets. The trajectories are influenced by modulating factors and, for both biomarkers and clinical symptoms, are dependent upon the sensitivity of the measurement.

The preclinical phase of AD

Exploring the continuum through the preclinical phase requires the use of biomarkers because individuals are apparently clinically normal. A number of diagnostic and progression markers have been described (Table 1), with the distinction being that diagnostic biomarkers are direct surrogates of brain AD lesions (amyloidosis or tauopathy) and therefore indicative of the presence of the disease, irrespective of stage, while progression markers identify downstream changes (metabolic changes, neuronal loss with atrophy) indicative of disease progression but not necessarily specific to AD [23]. Biomarker trajectories paralleling the hypothetical pathophysiological sequence of AD and specific biomarker changes over time were described by Jack et al. [24, 25] (Fig. 1).

The model presents the temporal evolution of five well-established AD biomarkers that provide a measure of brain Aβ deposition, tau, or neurodegeneration both in relation to each other and to the onset and progression of clinical symptoms. The model assumes that each biomarker follows a nonlinear temporal course, which is hypothesized to be sigmoid shaped, and that the maximum rate of change moves sequentially from one biomarker to the next. While there is some individual variation, changes in markers of Aβ deposition generally precede those of tau and neurodegeneration. The model not only describes biomarker trajectories in the preclinical phase, but also in the clinical phase, when symptoms become evident. Support for this hypothetical model has been provided by cross-sectional biomarker data across the preclinical/clinical continuum [26–28].

For the clinician, applying the continuum concept of AD raises several points of importance for counseling and prognostic discussions. AD can be diagnosed without dementia; an understanding of the biomarker changes is equally important to understanding clinical manifestations; and an appreciation for the temporal course of AD from preclinical biomarker evidence of disease through the presence of clinical symptoms is critical for effective disease diagnosis and management.

As 2025 approaches, there are other considerations concerning the continuum concept. There is currently no cure available but if forthcoming new treatments include DMTs that may have more impact if initiated earlier in the disease process, then earlier detection of disease is necessary. For more effective disease management, we also need to predict the future clinical course of disease more accurately. To achieve these goals, more research is needed to define biomarker profiles that best predict progression from the preclinical to the clinical stage and biomarker and/or cognitive profiles that best predict progression and rate of progression along the clinical course.

While we know that pathological changes begin long before symptoms appear, determining whether biomarker evidence of pathophysiological changes in the preclinical stage implies definite progression to clinical disease during an individual’s lifetime is difficult. Individual biomarkers do not provide definitive prognostic information. Recently, there have been efforts to improve diagnostic accuracy and ability to predict those at risk for clinical symptoms by considering a combination of biomarker findings. Jack et al. [46] proposed that diagnosis should be based on both the presence and absence of seven biomarkers in three categories (amyloid, tau, and neurodegeneration (A/T/N)). Dubois et al. [47] proposed that diagnosis be based on both Aβ and tau pathology. Using these criteria, the authors went further to differentiate between a “state” and a “stage”. In simple terms, a state is considered asymptomatic at risk of AD (cognitively normal and amyloid or tau positive but not both) or AD (amyloid and tau positive), while a stage refers to the degree of disease progression within a given state (e.g., clinical AD, preclinical AD, MCI due to AD or prodromal AD, dementia due to AD). Conceptualizing AD as a continuum may favor describing the state as dichotomous and the stage as continuous.

Only a few years ago, diagnosis supported by biomarker findings (Table 1) was considered appropriate only for research-related purposes [41]; however, there is increasing effort to integrate biomarkers into clinical decision-making, with several PET tracers having received regulatory approval and CSF markers being measured with higher precision using advanced automated systems. CSF Aβ₄₂, total tau (t-tau), and phosphorylated tau (p-tau) biomarkers and amyloid PET have been the most widely studied [48, 49]. More recently, tau PET has shown promise as a fairly specific marker of tau deposits characteristic of AD [30, 50].

Biomarkers are now a key component of AD clinical trials, playing a central role in selecting individuals for whom the study treatment is most likely to be effective and providing objective evidence of target engagement and disease-modifying effects. Integration of biomarker assessment into more widespread clinical use would help with earlier and more accurate diagnosis. For example, approximately 25% of individuals referred for clinical trial participation with a clinical diagnosis of mild AD dementia have been shown to be Aβ-negative [51], an observation inconsistent with the clinico-biological concept of AD, and this may be even higher in MCI due to AD [52]. Biomarker assessment would also assist with creation of uniform disease staging criteria for use in clinical research and practice, to enhance communication between research and practice, and help with transition of AD treatments from research to regulatory review to clinical practice. However, the clinical care environment has not yet evolved adequately for this to occur. In part, this is due to challenges with cost, standardization, accessibility, and incorporation of biomarker findings into patient treatment plans, but also, as mentioned previously, currently available biomarkers have limited ability to predict the clinical disease course. For clinical practice, we need simpler, less invasive, and more affordable biomarkers, as well as biomarkers that relate to other aspects of the disease.

In recent years, the focus in tool development has been on more sensitive cognitive measures that are able to detect subtle cognitive and functional impairment, to lower the threshold at which clinical disease can be detected as well as to detect smaller changes in cognition/function, to monitor disease progression along the continuum or therapeutic responses to interventions more accurately. Computerized tests that can be readily applied as screening tools in the primary care setting are also being introduced.

One key purpose of clinical assessment is to be able to predict the temporal path of AD at the individual level; for example, if and when MCI due to AD or prodromal AD will progress to AD dementia. Impairment in episodic memory (i.e., the ability to learn and retain new information) is seen most commonly in individuals who subsequently progress to a diagnosis of AD dementia [42, 54]. Based on this, tools that assess episodic memory, both immediate and delayed recall, are valuable. Decline in executive functions may also flag incident AD dementia; by contrast, change in information processing speed/attention seems less informative [54].

In the development of therapeutic interventions, there has generally been a shift in focus from AD dementia to MCI due to AD/prodromal AD and earlier; there is a need for a parallel shift in the diagnostic domain, beyond the research environment, to encourage lifestyle modifications and participation in clinical trials. An appreciation of the disease continuum engenders an awareness of our need to consider both diagnosis and therapies in a similar manner, along a continuum. That is, an individual treatment or management option may be most appropriate at a defined stage along the continuum, but its use will likely extend beyond this stage, and there will be overlap among the various treatment and management options such that more than one may be appropriate at any point on the continuum. This will be particularly relevant as new DMTs with different mechanisms of action become available.

With DMTs, treatment earlier in the disease continuum will likely be required to achieve more disease modification and maximize the opportunity of having an effect through slowing decline. If and when DMTs are approved for use, biomarker testing is likely to be central for early and more accurate diagnosis as well as for monitoring treatment effect. We can envision that biomarker findings will influence both initiation and termination of a specific treatment and may help determine whether combination therapies are appropriate and effective at the individual patient level. In particular, amyloid biomarkers will likely be necessary to identify candidates for anti-amyloid interventions at the predementia stages of AD.

The goal of managing any disease is primary prevention and, ultimately, we anticipate blood-based assays of amyloid dysregulation or other changes that will enable primary prevention studies. We can envision eventual management of such indicators of risk of amyloid accumulation with secretase inhibitors, other drug therapies, and lifestyle management to minimize reversible contributors to disease.