Neurology · Medical Literature
Alzheimer's blood-based biomarkers, incident dementia, and interactions with age, APOE status, and hormone therapy.
Abstract
Cognitive impairment among older adults is often due to multiple pathologies and heterogenous risk factors. We assessed whether Alzheimer's blood-based biomarkers (BBMs) were associated with incident mild cognitive impairment (MCI)/probable dementia, and whether associations were modified by age, apolipoprotein E (APOE), and hormone therapy (HT). Analyses included 2467 Women's Health Initiative Memory Study women (≥65 years of age) randomized between 1995 and 1998 to 3-5-years of HT or placebo. Cox regression (mean 18-year follow-up) assessed associations between the z-scored BBMs and MCI/dementia. Lower baseline amyloid beta (Aβ)42/40 ratio and higher phosphorylated tau 181 (p-tau181), glial fibrillary acidic protein (GFAP) and neurofilament light chain (NfL) were associated with an increased risk of MCI and dementia; GFAP was most strongly associated. The p-tau181 and NfL associations were stronger among APOE ε4 carriers; BBMs varied non-linearly by age. The associations of BBMs with the cognitive outcomes also varied inconsistently between HT groups. BBMs for AD and related dementias (ADRD) are associated with incident MCI/dementia in older women. Interactions between the BBMs and HT were inconsistent and require further investigation.
Clinical relevance
This study addresses a critical gap in the early identification of Alzheimer's disease and related dementias (ADRD) by evaluating the long-term predictive utility of blood-based biomarkers (BBMs) for incident MCI and probable dementia in older women over a mean 18-year follow-up — one of the longest such follow-up periods reported. The findings are directly relevant to the growing postmenopausal female population, who bear a disproportionate burden of AD, and to clinicians now facing real-world decisions about incorporating plasma biomarkers into risk stratification. The additional interrogation of APOE ε4 status and hormone therapy (HT) as effect modifiers adds actionable nuance for individualized patient counseling.
Key findings
1. In 2467 Women's Health Initiative Memory Study (WHIMS) participants (≥65 years), all four BBMs — lower Aβ42/40 ratio, higher p-tau181, higher GFAP, and higher NfL — were independently associated with increased risk of incident MCI/probable dementia on Cox regression over a mean 18-year follow-up. 2. GFAP demonstrated the strongest association with cognitive outcomes among all BBMs evaluated, suggesting superior long-term predictive power for incident MCI/dementia in this cohort. 3. The associations of p-tau181 and NfL with MCI/dementia were significantly stronger among APOE ε4 carriers compared to non-carriers, indicating a gene–biomarker interaction that amplifies risk in this genetically predisposed subgroup. 4. BBM associations with cognitive outcomes varied non-linearly by age, indicating that the predictive value of these plasma markers is age-dependent and cannot be uniformly applied across the older adult age spectrum. 5. Associations between BBMs and MCI/dementia outcomes varied inconsistently across HT (active vs. placebo) groups, with no definitive or reproducible interaction pattern identified between hormone therapy and any individual biomarker. 6. All BBMs were analyzed as z-scored continuous variables, and the study design was a randomized controlled trial (HT vs. placebo, 1995–1998, 3–5 years of treatment) with long-term observational follow-up, lending methodological rigor to the longitudinal biomarker associations.
Practice implications
Clinicians evaluating cognitive risk in women aged 65 and older should consider plasma GFAP as the highest-yield single BBM for long-term MCI/dementia risk stratification, while a multi-biomarker panel (Aβ42/40, p-tau181, GFAP, NfL) provides the most comprehensive assessment. APOE ε4 genotyping should be integrated into biomarker interpretation, as elevated p-tau181 and NfL confer meaningfully greater risk in ε4 carriers, warranting earlier and more intensive surveillance or intervention discussions in this subgroup. Hormone therapy should not currently be recommended or withheld based on anticipated modulation of AD biomarker risk, as HT–BBM interactions were inconsistent across the cohort and require further prospective investigation before informing clinical decision-making.