Reference: Neurology. 2026 Apr 14;106(7):e214779
Practice Point: Dementia biomarkers can aid in the evaluation of patients with mild cognitive impairment (MCI) or early dementia, but they add little actionable value for patients without cognitive symptoms.
EBM Pearl: A multivariable analysis can help elucidate how different factors independently influence disease risk and can also identify when multiple factors have overlapping contributions.
As the global population continues to age and dementia becomes more common, the need for early detection and diagnosis grows. While cognitive tests are the standard screening tools used for early detection, plasma biomarkers have become a major focus of clinical research. Certain biomarkers such as amyloid beta (AB) and phosphorylated tau at threonine 181 (p-tau 181) have been used to help confirm the diagnosis of early Alzheimer disease, while other biomarkers such as neurofilament light (NfL) and glial fibrillary acidic protein (GFAP) may be general indicators of dementia risk. What is still unclear, however, is the relative predictive power of these plasma biomarkers compared to clinical cognitive status. How much can these biomarkers tell us about dementia risk in people with a normal cognitive status? Could a blood test meaningfully predict who will develop dementia years before clinical symptoms appear?
A new study published in Neurology analyzed data from 1,577 older adults (mean age 76 years) from the Atherosclerosis Risk in Communities (ARIC) study in order to address this question. At baseline in 2011-2013, all included individuals were free of dementia; were tested for AB, p-tau 181, NfL, and GFAP; and were stratified by cognitive function (with or without mild cognitive impairment [MCI]). They were then followed for an additional 9 to 11 years, during which time they were monitored for incident dementia. A multivariable analysis was performed to evaluate to what extent different factors, including age, plasma biomarker levels, and cognitive status, predicted the incidence of dementia in the ensuing follow-up period.
As one might expect, older age, baseline MCI, and elevated levels of plasma biomarkers were each independently associated with a higher risk of incident dementia. However, baseline cognitive status was a stronger predictor than the biomarkers. Individuals without MCI at baseline had a much lower risk of dementia compared to patients with MCI, regardless of whether they had high or low levels of plasma biomarkers. For example, the 10-year cumulative incidence of dementia in individuals with MCI and low AB was about 40%, compared to almost 60% with high AB. Conversely, in patients without MCI, the cumulative incidence with high AB was only about 30% (and under 20% for those with low AB).
Furthermore, regression models incorporating age alone, cognitive status alone, plasma biomarkers alone, or various combinations each showed broadly similar discriminatory accuracy (as measured by area under the curve) for the risk of incident dementia. This suggests that these measures capture overlapping (but not necessarily identical) aspects of disease risk. Therefore, adding plasma biomarkers to models that already include age and cognitive status may not substantially improve the model’s ability to predict who will develop dementia.
So, what does all this mean for clinicians? The utility of any given biomarker level is highly dependent on cognitive status. In cognitively-intact patients, abnormal biomarker levels carry a low positive predictive value, whereas the same levels in patients with MCI are far more predictive of progression to dementia. A cognitively unimpaired patient with elevated plasma biomarkers does carry an elevated relative risk compared to someone with low plasma biomarker levels, but their absolute risk over the next decade remains substantially lower than that of a patient already demonstrating clinical impairment. Therefore, cognitive status, particularly the distinction between normal cognition and MCI, remains one of the strongest determinants of short-term dementia risk, and the clinical significance of biomarker test results in cognitively intact individuals is questionable.
In current clinical practice, biomarker tests for dementia are usually performed only to clarify etiology in patients who are already showing signs of cognitive decline. This study suggests that routinely checking plasma biomarkers in asymptomatic adults probably isn’t very useful because the results may not add meaningful prognostic value beyond what can be gleaned from their current, unimpaired cognitive status.
As plasma biomarker testing becomes more widely available, it is important to keep in mind that more testing is not necessarily indicated, especially in those without MCI. Moreover, the interpretation of abnormal biomarker levels should always take into consideration the full context of the individual’s clinical presentation.
For more information, see the topic Dementia Evaluation in DynaMed.
DynaMed EBM Focus Editorial Team
This EBM Focus was written by Claire Symanski, PhD, Medical Editor and Team Lead for ENT at DynaMed. Edited by Katharine DeGeorge, MD, MS, Executive Editor at DynaMed and Professor of Family Medicine at the University of Virginia; Alan Ehrlich, MD, FAAFP, Executive Editor at DynaMed and Associate Professor in Family Medicine at the University of Massachusetts Medical School; McKenzie Ferguson, PharmD, BCPS, Senior Clinical Writer at DynaMed; Rich Lamkin, MPH, MPAS, PA-C, Senior Clinical Writer at DynaMed; Michael Butler, PhD, Medical Writer at DynaMed; Matthew Lavoie, BA, Senior Medical Copyeditor at DynaMed; Hannah Ekeh, MA, Senior Associate Editor II at DynaMed; and Jennifer Wallace, BA, Senior Associate Editor at DynaMed.