With global dementia prevalence rising steeply [1], the need for precise diagnosis and effective disease-modifying therapies has never been greater. The research framework of the Amyloid/Tau/Neurodegeneration (ATN) system for confirming the presence of Alzheimer’s disease (AD) pathology incorporates specific cerebrospinal fluid (CSF) biomarkers with PET neuroimaging to classify AD into stages of severity. The particular CSF biomarkers of interest are amyloid β [Aβ] (A)-, tau (T)- and neurodegeneration (N)-related biomarkers which indicate distinctive aspects of AD pathophysiology [2]. This system has been increasingly applied to drug design and intervention studies. However, the widespread use of CSF biomarkers and neuroimaging is restricted due to high costs as well as the invasive nature of lumbar puncture and its more challenging nature in older people compared to younger people.

Previously, routine blood biomarkers for dementia seemed a remote possibility. However, recent studies show that biomarker concentrations in blood correlate closely to those in CSF [2]. There is also growing evidence that typical AD pathological characteristics such as extracellular amyloid plaque deposition, intra-neuronal neurofibrillary tangle formation resulting from hyperphosphorylation of tau protein and related hallmarks of neurodegeneration are present several years before overt clinical symptoms occur. Consequently, blood-based brain biomarker indicators of these changes in the pre-symptomatic phase of AD will very likely be transformative in dementia diagnosis and management.

Amyloid Markers

Physiological amyloid beta 1–40 (Aβ40), pathological amyloid beta 1–42 (Aβ42) and the Aβ42/Aβ40 ratio are already well-established amyloid biomarkers in CSF. Aβ seemed to be a feasible blood-based biomarker as it readily crosses the blood–brain barrier. However, initial studies found inconsistent results, likely due to less advanced measurement techniques and the use of clinical criteria for diagnosis rather than the proven post hoc demonstration of Aβ pathology [3]. Since then, with development of ultrasensitive immunoassays, more recent studies show that plasma Aβ42/Aβ40 ratio correlates closely with both PET-defined cerebral amyloid deposition and CSF Aβ42/Aβ40 ratios [4]. Research evidence indicates that plasma Aβ42/Aβ40 ratio accurately reflects cerebral Aβ deposition in all stages of AD to a high level of accuracy. When combined with APOE genotype analysis in one study, diagnostic accuracy for AD was further increased (AUC = 0.85 [95% CI, 0.82–0.88]) [5] i.e. clinically applicable accuracy. More recent research indicates that plasma Aβ changes (as well as those in CSF) precede amyloid PET positivity [6] and that plasma Aβ42/Aβ40 ratio measurement achieved high accuracy in predicting amyloid PET status (AUC = 0.941 in a discovery study, AUC = 0.868 in a validation study) [7].

Tau Markers

Hyperphosphorylation of tau protein mediates neurofibrillary tangle formation, an essential element of AD pathology. Three subtypes of phosphorylated tau (p-tau) appear particularly important: p-tau181 (tau phosphorylated at threonine 181), p-tau217 and p-tau231. Initial studies showed that plasma total p‐tau was significantly elevated in AD compared with age‐matched cognitively unimpaired controls and patients with non‐AD dementias [3]. At first, measuring p-tau in blood samples was technically difficult given the ultra-low p-tau concentrations involved. However, recent advances in ultrasensitive immunoassay techniques have made accurate plasma fractional tau measurements feasible. Milà-Alomà et al. have recently shown that plasma p-tau181, p-tau217 and p-tau231 are all significantly elevated in preclinical AD and that plasma p-tau231 and p-tau217 reflect the earliest cerebral Aβ changes before overt Aβ plaque pathology is detectable by amyloid PET [8]. Further recent work shows that plasma p-tau181 concentration starts to increase approximately 15 years before clinical disease onset in familial AD [9] and also predicts disease neuropathology at least 8 years before autopsy confirmation in sporadic AD [10]. The predictive accuracy of P-tau181 and P-tau217 when combined with APOE genotype and psychometric testing is impressive i.e. > 90% for mild cognitive impairment (MCI) to progress to overt AD [11]. Wilson et al. have recently described a high-throughput and fully automated Lumipulse® plasma p-tau181 assay for the detection of AD [12]. The assay has robust performance in differentiating AD from control participants (AUC 0.959, CI: 0.912 to 0.990), and was strongly correlated with CSF p-tau181, CSF Aβ42/Aβ40 ratio and amyloid-PET global standardised uptake value ratios. Mass spectrometry-based measures of p-tau217 perform best when identifying MCI patients with abnormal brain Aβ imaging who will subsequently progress to overt AD [13]. Several other p-tau assays (p-tau217Lilly, p-tau217Janss, p-tau181ADx and p-tau181WashU) show high and consistent accuracy comparable to the gold standards of Aβ-PET and CSF Aβ42/40 in detecting abnormal brain Aβ status and prediction of future progression to overt AD [13].

Total-tau (t-tau) concentration in CSF or plasma reflects the severity of Aβ‐induced neuronal or axonal injury, and can be elevated in both AD and non-AD dementias, as well as stroke, traumatic brain injury and Creutzfeldt-Jakob disease [14]. Current evidence indicates that plasma t-tau may not be useful as a specific diagnostic biomarker for AD, but is potentially useful as a general neurodegenerative disease prognostic indicator [15].

Neurofilament Light-Chain (NfL)

Plasma NfL is a non-specific marker of neuronal injury and is increased in several neurodegenerative and non-neurodegenerative brain diseases [11]. Plasma NfL concentrations, similar to CSF NfL, correlate with brain white matter pathology [16] and with severity of neurofibrillary tangle pathology in AD assessed at autopsy [17]. However, like t-tau, plasma NfL lacks diagnostic specificity but has potential as a neurodegenerative disease severity biomarker.

Glial Fibrillary Acidic Protein (GFAP)

Plasma GFAP is another blood biomarker of neuroinflammatory cascade activation. Increased CSF and plasma GFAP concentrations indicate abnormal activation and proliferation of astrocytes resulting from neuronal cell damage [11]. Significant increases in plasma GFAP concentration occur in AD and a significant association between baseline GFAP and grey matter volume diminution has been observed over time [18]. Elevated plasma GFAP is also predictive of cognitive decline in MCI [19]. A summary of brain biomarkers measurable in blood is provided in Table 1.

Table 1 Summary of principal specific and non-specific blood biomarkers of Alzheimer’s disease
Full size table

Role of blood biomarkers in diagnosis, treatment, and screening for AD

The conditional FDA approval of aducanumab in 2021 and the more recent FDA approval of lecanemab signal the arrival of disease-modifying therapies (DMTs) for early overt AD. It seems possible that DMTs for preclinical AD could be approved for use in the current decade. Although the primary measures of drug efficacy will necessarily be based on clinical psychometric assessment, it will also be important to know whether any DMT improves AD-specific or non-specific biomarkers of neuronal injury and associated neuroinflammatory activity. Amyloid and tau-specific PET neuroimaging, whilst available in highly specialized centres, is expensive and likely to remain inaccessible for assessment and monitoring of most AD patients. Semiquantitative neuroimaging to detect reduction of cerebral amyloid load is desirable when treating AD patients with amyloid-reducing therapies. However, simple and much less expensive blood biomarker tests of amyloid and tau-specific pathology will likely be more practical as a treatment monitoring test in AD, alongside the all-important psychometric measurements. Plasma concentrations of p-tau181, p-tau-217, p-tau231, GFAP, NfL and Aβ42/Aβ40 ratio change significantly in preclinical AD patients compared to age-matched cognitively normal individuals. This introduces the intriguing possibility of treating preclinical AD to prevent or delay overt cognitive impairment if anti-amyloid or anti-p-tau DMTs prove effective.

Recently, Mielke et al. have shown that both plasma p-tau217 and p-tau181 are excellent predictors of elevated brain total amyloid (AUC > 0.80) and entorhinal cortical tau on PET imaging (AUC > 0.80), but less so for temporal lobe tau (AUC < 0.70) [20]. However, there are some limitations in patients with certain comorbidities in relation to plasma brain biomarker interpretation. For example, patients with chronic kidney disease or previous myocardial infarction or stroke may cause elevated brain biomarker concentrations. Also, patient sex affects some brain biomarkers, women having higher average baseline plasma p-tau181 concentrations than men [21]. Nevertheless, brain biomarkers in blood have been used to monitor treatment effects in a number of recent clinical trials of DMTs in some neuroinflammatory and neurodegenerative diseases [22] (Table 2).

Table 2 Application of blood-based brain biomarkers in current and recent clinical trials of disease-modifying therapies in Alzheimer’s disease
Full size table

The FDA approval of lecanemab [23] indicates that several more anti-amyloid DMTs will likely become available soon for treatment of early-stage AD. Valiltramiprosate (ALZ-801), a prodrug of homotaurine which works by inhibiting Aβ42 aggregation into toxic oligomers is currently being assessed APOLLOE4 clinical trial in early-stage AD [24]. In this trial, plasma p-tau181 and p-tau181/Aβ42 ratio are being used to evaluate valiltramiprosate efficacy versus placebo in early-stage AD patients who are homozygous for Apo-E4. It is expected that APOLLOE4 will be completed in June 2024 [25].

The recently published TRAILBLAZER trial data showing efficacy of donanemab in early overt AD appears to confirm the arrival of DMTs in dementia [26]. Personalized precision pharmacotherapy for dementia depends critically on precision diagnostics. The imminent arrival of highly disease-sensitive and specific brain biomarkers in blood [27] signals a new and exciting phase for precision diagnosis and effective pharmacotherapy of dementing illness.