For Doctors in a Hurry
- This study assessed the independent and combined value of cognitive status and plasma p-tau181 for predicting incident dementia.
- Researchers followed 1,577 dementia-free participants from the Atherosclerosis Risk in Communities study, with a mean age of 76.5 years.
- Mild cognitive impairment was associated with a nearly threefold increased dementia risk (HR 2.94), independent of plasma biomarker concentrations.
- The authors concluded that cognitive status and plasma p-tau181 provide independent but overlapping information about future dementia risk.
- These findings underscore the continued importance of an accurate clinical diagnosis, especially for patients who are cognitively unimpaired.
Integrating Plasma Biomarkers into Clinical Dementia Risk Stratification
The diagnostic framework for Alzheimer disease is shifting toward a biological definition centered on the detection of amyloid and tau pathologies [1]. Blood-based assays for phosphorylated tau 181 (p-tau181) provide a less invasive alternative to lumbar punctures, with data indicating these tests can distinguish Alzheimer dementia from cognitively unimpaired controls with an area under the curve (AUC) of 90.21% to 98.24% [2, 3]. Clinicians must integrate these molecular signals with bedside assessments of mild cognitive impairment and vascular risk factors, such as perivascular spaces (fluid-filled cavities surrounding brain blood vessels that serve as markers of small vessel disease) which correlate with elevated p-tau181 and neurofilament light levels (p < 0.01) [4, 5]. In large prospective cohorts, combining plasma p-tau with brief cognitive tests and apolipoprotein E (APOE) genotyping achieved a prognostic accuracy of 0.91 AUC for predicting progression to dementia within four years [2]. This multidimensional approach clarifies the independent and combined value of biomarkers for long-term risk stratification in primary and secondary care [6, 2].
To clarify the prognostic relationship between clinical assessments and molecular markers, researchers analyzed data from the Atherosclerosis Risk in Communities (ARIC) study. The analysis included 1,577 ARIC participants with a mean age of 76.5 years. The demographic breakdown of the cohort was 60% women, 73% White, and 27% Black. All individuals included in this sample were clinically adjudicated as free of dementia during baseline assessments conducted between 2011 and 2013. At that time, the participants had plasma samples analyzed for p-tau181 and other biomarkers to establish their baseline biological risk profiles. Following the baseline evaluations, the participants were surveilled for incident dementia through December 31, 2022, providing approximately a decade of longitudinal follow-up. To assess how well different factors predicted the onset of cognitive decline, the investigators utilized cumulative incidence curves, Cox models (a statistical technique used to determine the association between specific variables and the time it takes for an event to occur), and Fine-Gray models (a method used to estimate the probability of an event while accounting for competing risks, such as death). These statistical models were specifically designed to evaluate the independent and combined discriminatory accuracy of baseline cognitive status and plasma biomarkers for incident dementia. By comparing these variables, the researchers aimed to determine whether molecular tests offer additional predictive value beyond standard clinical evaluations of cognitive impairment.
Independent Prognostic Value of Clinical Status and Biomarkers
The longitudinal data revealed that a clinical diagnosis alone carries substantial prognostic weight. Specifically, the risk of incident dementia was higher in persons with a baseline status of mild cognitive impairment compared with those who were cognitively unimpaired. The researchers calculated a covariate-adjusted hazard ratio (HR) of 2.94 (95% CI 2.61-3.33) for incident dementia in individuals presenting with mild cognitive impairment. Notably, this increased dementia risk in persons with mild cognitive impairment was independent of biomarker status, indicating that bedside clinical assessments provide critical predictive information regardless of a patient's underlying molecular profile. Parallel to the clinical findings, the blood-based assays demonstrated their own distinct predictive value. The study showed that the risk of dementia was also higher in persons with more abnormal concentrations of p-tau181 and other biomarkers independent of cognitive status. This demonstrates that elevated plasma p-tau181 and related proteins signal an impending trajectory of cognitive decline even in patients who currently appear cognitively intact during standard office evaluations. For the practicing physician, this suggests that a high p-tau181 level in a cognitively normal patient is a significant biological red flag that warrants closer long-term monitoring.
The researchers observed that the predictive signals from clinical cognitive status and plasma p-tau181 are not purely additive, suggesting they capture some of the same underlying risk. When age, cognitive status, and p-tau181 were included in the same statistical models, the risk of incident dementia was attenuated (a statistical reduction in the strength of an association when multiple related variables are analyzed together) relative to models that analyzed cognitive status or plasma biomarkers in isolation. This reduction in predictive strength indicates that these two measures share common prognostic variance. Specifically, for continuous p-tau181 concentrations, the covariate-adjusted hazard ratio (HR) without accounting for cognitive status was 1.45 (95% CI 1.36-1.54). However, when the researchers included cognitive status in the same model with p-tau181, the HR for p-tau181 decreased to 1.37 (95% CI 1.29-1.46). These findings indicate that cognitive status and plasma biomarker concentrations convey independent but overlapping information about the risk of incident dementia. For the practicing clinician, this means that a patient's current cognitive performance already reflects some of the pathological burden indicated by p-tau181 levels. The study models demonstrated that when combined with age, p-tau181 alone, cognitive status alone, or the combination of both had similar discriminatory accuracy (the ability of a test to correctly distinguish between those who will develop a condition and those who will not).
Clinical Takeaway: Similar Accuracy, Different Absolute Risk
As plasma biomarkers such as phospho-tau species are increasingly used in clinical practice for the diagnosis of Alzheimer disease, these findings provide critical context for their interpretation. The study confirms that both a patient's cognitive status and their level of phosphorylated-tau at threonine 181 (p-tau181) convey prognostic information about the risk of incident dementia. However, the statistical models demonstrated that these two sources of information are not entirely distinct in their predictive power. The analysis showed that when combined with age, models using p-tau181 alone, cognitive status alone, or the combination of p-tau181 and cognitive status had similar discriminatory accuracy. This suggests that from a purely predictive standpoint, an elevated biomarker level and a diagnosis of mild cognitive impairment offer comparable signals of future risk. The clinical implication, however, is that similar statistical accuracy does not equate to equivalent absolute risk for the patient. The study's most salient conclusion for clinicians is that the incidence rate of dementia is far lower in persons who are initially cognitively unimpaired, even if their p-tau181 levels are elevated. This disparity in outcomes underscores the continued primacy of the clinical evaluation. The authors emphasize that this lower incidence in the cognitively normal group highlights the importance of an accurate clinical diagnosis. For counseling patients, an abnormal biomarker result in a cognitively intact individual represents a long-term risk factor, but it does not confer the same near-term prognosis as a formal diagnosis of mild cognitive impairment, which remains a more powerful indicator of imminent decline.
References
1. Jack CR, Bennett DA, Blennow K, et al. NIA‐AA Research Framework: Toward a biological definition of Alzheimer's disease. Alzheimer s & Dementia. 2018. doi:10.1016/j.jalz.2018.02.018
2. Palmqvist S, Tideman P, Cullen N, et al. Prediction of future Alzheimer’s disease dementia using plasma phospho-tau combined with other accessible measures. Nature Medicine. 2021. doi:10.1038/s41591-021-01348-z
3. Karikari T, Pascoal T, Ashton N, et al. Blood phosphorylated tau 181 as a biomarker for Alzheimer's disease: a diagnostic performance and prediction modelling study using data from four prospective cohorts.. Lancet Neurology. 2020. doi:10.1016/s1474-4422(20)30071-5
4. Ong JJH, Leow Y, Wang JDJ, Kandiah N. Perivascular Spaces as Early Indicators of Alzheimer's Pathology and Cognitive Impairment in a Southeast Asian Cohort: Insights from the BIOCIS Study. Alzheimer's & Dementia. 2025. doi:10.1002/alz70856_097658
5. O’Brien JT, Holmes C, Jones MS, et al. Clinical practice with anti-dementia drugs: A revised (third) consensus statement from the British Association for Psychopharmacology. Journal of Psychopharmacology. 2017. doi:10.1177/0269881116680924
6. Huang W, Liao L, Liu Q, et al. Blood biomarkers for vascular cognitive impairment based on neuronal function: a systematic review and meta-analysis.. Frontiers in neurology. 2025. doi:10.3389/fneur.2025.1496711
