Amyloid beta misfolding measured in blood is the strongest early warning signal for Alzheimer's disease yet identified in asymptomatic people, according to a study led by Professor Klaus Gerwert of PRODI at Ruhr University Bochum. The work, published in EMBO Molecular Medicine on July 16, 2026, drew on 17 years of blood samples from 779 participants in the ESTHER cohort, a long-term population study of older adults run by the German Cancer Research Center (DKFZ). That kind of archive is unusual: it let the team test how well competing biomarkers performed across the entire arc of the disease rather than at a single snapshot.
The comparison produced a clear split by disease stage. P-tau 217, currently the most widely used blood marker, performed excellently once Alzheimer's had reached the clinical stage, matching established measures such as the amyloid beta 42/40 ratio in cerebrospinal fluid and amyloid PET scans. In the preclinical, symptom-free phase, however, its predictive power was only moderate, with an AUC of 0.67. Researchers generally treat values around 0.8 or above as strongly predictive, so P-tau 217 falls well short of that bar before symptoms emerge. Amyloid beta misfolding, by contrast, signaled elevated risk years earlier.
The practical implication is that no single blood test is likely to serve every purpose. A marker that confirms Alzheimer's in someone already showing memory problems is not necessarily the marker that identifies who will develop the disease a decade from now. As prevention trials increasingly try to recruit people before damage accumulates, having a test that works in the silent phase matters for deciding whom to enroll and when.
These findings come from a single population cohort, and the study is a comparison of biomarker performance rather than a test of any treatment. Confirmation in other cohorts will determine whether misfolding-based testing is ready to move toward wider use.