For decades, diagnosing Alzheimer’s disease (AD) has required access to specialist memory clinics, neuroimaging or invasive cerebrospinal fluid (CSF) testing. That paradigm is starting to shift. In 2025, the US Food and Drug Administration (FDA) cleared the first blood-based biomarker tests intended to aid the assessment of AD.
“Blood-based biomarkers for Alzheimer’s disease are probably the most important breakthrough in 20 years,” says Craig Ritchie, honorary professor of the psychiatry of ageing at the University of Edinburgh. Yet alongside their promise, questions arise about how, where and when these tools should be used.
From specialist diagnostics to blood draws
In October 2025, the FDA cleared Roche’s Elecsys pTau181 assay, after having approved Fujirebio’s Lumipulse G pTau217/β-amyloid 1-42 plasma ratio test earlier that year in May. Both assays detect proteins associated with AD pathology in plasma, offering a far less invasive and more accessible alternative to conventional diagnostics such as positron emission tomography (PET) imaging or lumbar puncture.
The ability to diagnose AD accurately and at an earlier stage has taken on renewed urgency1 with the arrival of disease-modifying therapies, including monoclonal antibodies that target amyloid, such as lecanemab and donanemab. These treatments can slow, but not reverse, neurodegeneration, which makes timely intervention critical. They are most effective when administered before substantial neuronal loss has occurred and are indicated only for patients with biologically confirmed AD pathology.
“Those therapeutics are quite precisely targeting a particular biology,” Ritchie says. “But with disease-modifying therapies, you have to know you have the disease first.”
Although they are often discussed together, the two FDA-cleared assays serve distinct clinical purposes and target different proteins.
Roche’s Elecsys measures tau protein phosphorylated at amino acid 181 (pTau181) in plasma. In a multi-center, non-interventional study of 312 participants, the assay ruled out AD with a negative predictive value of 97.9%. The FDA clearance specifies its use as an aid to assessment in adults 55 years of age and older who present with signs or symptoms of cognitive decline, with results interpreted alongside other clinical information. It is the first blood test for AD to be approved for use in primary care.
Fujirebio’s Lumipulse assay, by contrast, measures both pTau217 and a β-amyloid peptide of amino acids 1–42 and calculates their ratio, which correlates with the presence of amyloid plaques in the brain. The test is intended for patients who present at a specialized care setting with signs and symptoms of cognitive decline. In a clinical study of 499 patients, 91.7% of patients with positive results had amyloid pathology confirmed by PET or CSF testing, while 97.3% of those with negative results were amyloid negative. Around 20% of patients, however, received an indeterminate result.
Alicia Algeciras-Schimnich, a consultant at the Mayo Clinic in Rochester, Minnesota, emphasizes that this distinction matters. “The Roche test is best considered a rule-out or screening test,” she explains. “Its reported positive predictive value is around 22%, meaning most positive results require confirmatory testing with PET or CSF biomarkers.” By contrast, she says, the Fujirebio assay functions more as a confirmatory test, although indeterminate results still require follow-up.
Matching biomarkers to clinical questions
Whether FDA clearance translates into widespread clinical use is uncertain. Joshua Grill, director of the Institute for Memory Impairments and Neurological Disorders at the University of California, Irvine, calls the approvals “a milestone,” but cautions against assuming immediate impact.
“These tests have been available, if not authorized, for some time,” Grill says. “What the uptake will be in primary care is unknown. Physician comfort with the tests is a key variable, as will be professional guidelines instructing their appropriate use.”
In primary care, most people who present with memory or cognitive complaints do not have AD pathology, says Ritchie. Blood-based biomarkers could therefore serve an important role as early triage tools, helping clinicians rule out AD before pursuing more invasive or costly investigations. “Biomarkers like pTau181, which have a high negative predictive value, are very good at identifying people who are unlikely to have Alzheimer’s,” Ritchie says. “The doctor can then focus on other causes like sleep problems, mood disorders or metabolic issues.”
He compares the shift to other chronic diseases already managed in primary care. “You don’t wait to see a cardiologist to get a lipid profile,” he says. “The direction of travel should be using these tools safely and confidently in primary care, even if not everyone agrees on that yet.”
“These tests should never be used as standalone measures of amyloid pathology or to make an Alzheimer’s diagnosis,” cautions Algeciras-Schimnich. “They must be interpreted within the context of their intended application, whether for triage or confirmation, and with an understanding of potential confounding factors that may affect results.”
Who should be tested?
Research into whether blood biomarkers actually improve care when deployed outside specialist settings is ongoing. The BioFINDER-Primary Care study, for example, is enrolling patients who present to primary care with cognitive symptoms across multiple sites in Sweden. The study is prospectively evaluating whether adding plasma biomarkers of AD, alongside brief cognitive assessments, improves diagnostic accuracy, confidence of the general practitioner and referral decisions, compared with usual care.
Another unresolved question is whether blood biomarkers should be used in asymptomatic people. A 2025 study in Nature Medicine exploring the predictive performance of six AD blood biomarkers in 2148 dementia-free older adults from Sweden suggested that their greatest clinical value lies not in population screening but in stratifying risk among people who already have cognitive symptoms2. Two international expert panels have also recommended against biomarker testing in cognitively unimpaired populations until evidence-based symptom-delaying treatments become available3,4.
However, a meta-analysis of 17 biomarker studies in JAMA Neurology concluded that plasma pTau217 can reliably predict AD pathology in the preclinical stages of the disease, which raises the possibility that blood-based biomarkers could eventually support earlier identification, even before symptoms emerge5.
“I do think these tests hold incredible promise for an eventual practice that includes screening. And that would mean that older adults with no cognitive problems, when they reach a certain age, would undergo testing to identify whether they may be good candidates for symptom-delaying treatments,” says Grill. “I think that's where we're headed, but I don't think we're there yet.”
Accelerating research
Blood biomarkers are already reshaping clinical trials of AD. “They’ve turned a shot in the dark into a shot on goal,” says neurologist Jeffrey Cummings from the University of Nevada, Las Vegas. “Before them, we often didn’t know which trial patients truly had Alzheimer’s.”
Ian McDonough, an associate professor of psychology at Binghamton University, says FDA-cleared blood tests could dramatically expand trial access. “They expand the reach of biomarker testing to any facility that does blood draws,” he says. “That includes people in harder-to-reach or more impoverished regions who often can’t access PET or CSF testing.”
Larger, more representative trials could help address gaps in AD research, by including populations who have long been under-represented. Blood tests may also enable recruitment earlier in the disease course, when interventions are more likely to succeed.
However, recent real-world data have underscored the fragility of biomarker performance. At the 2025 Clinical Trials on Alzheimer’s Disease meeting in Boston, Algeciras-Schimnich presented data suggesting higher-than-expected false positive rates for the Lumipulse assay in certain cohorts.
“Our findings stress the need to independently validate test performance in real-world situations, even for FDA-cleared tests,” she says.
