GFAP and NfL Are Robust Dementia Markers When Considering Hypercholesterolemia

Serum glial fibrillary acidic protein (GFAP) and neurofilament light chain (NfL) may be more effective dementia risk predictors than phosphorylated tau181 (p-tau181), according to a study published in Alzheimer’s & Dementia.

There are blood biomarkers for dementia risk that have been established and high cholesterol has been considered to be a modifiable risk factor for dementia risk. For the study, researchers from centers in Germany and the United Kingdom evaluated whether blood biomarkers of dementia differ in the setting of hypercholesterolemia.

Researchers sourced data for this analysis from the nested case-control study ESTHER which recruited study participants between 2000 and 2002 in Germany. During 14- and 17-year follow-ups, participants (N=768) were evaluated for plasma biomarkers. Hypercholesterolemia was defined as total cholesterol ≥240 mg/dL.

The cohort was stratified into dementia cases (n=261) and control individuals without dementia (n=507). The cases and control individuals comprised 57.9% and 54.8% women, 58.2% and 50.9% were aged 60-70 years, 39.5% and 25.8% were apolipoprotein E (APOE) ε4 carriers, and 6.9% and 1.8% had a history of stroke, respectively.

Among the entire study population, a GFAP value of 123.50 pg/ml, an NfL value of 21.50 pg/ml, and a p-tau181 of 2.07 pg/ml separated the third and fourth quartiles for each biomarker. In general, dementia risk was increased among individuals with biomarkers in the fourth quartile comapred with the lower 3 quartiles for GFAP (odds ratio [OR], 3.08) and NfL (OR, 1.57) but not for p-tau181 (OR, 1.29; 95% CI, 0.86-1.93).

Stratified by APOE ε4 status, risk was elevated among individuals with high GFAP and NfL regardless of APOE ε4 status whereas only individuals who were not APOE ε4 carriers with high p-tau181 were at elevated dementia risk (OR, 1.67). In general, the interactions between biomarkers and APOE ε4 status were not significant.

Stratified by hypercholesterolemia, dementia risk tended to be higher among the subset of patients with high total cholesterol compared with lower total cholesterol, in which the risk for every one-standard deviation (SD) increase in GFAP was an OR of 2.54 compared with 1.41, in NfL was 1.56 compared with 1.40, and in p-tau181 was 1.23 compared with 1.13, respectively.

When additionally stratifying by APOE ε4 status, significant associations were observed per one-SD increase in GFAP among APOE ε4 carriers (OR, 2.15) and noncarriers (OR, 3.15) in the hypercholesterolemia cohort and in GFAP among APOE ε4 carriers in the non-hypercholesterolemia cohort (OR, 1.72) and in NfL among noncarriers in the non-hypercholesterolemia cohort (OR, 1.71).

The best-fit model for predicting dementia risk included age, gender, educational level, APOE ε4 status, total cholesterol, GFAP, NfL, and p-tau181 (area under the curve [AUC], 0.8452).

The results of this study may not be generalizable for a more diverse population.

“In this cohort the association between biomarkers, especially GFAP and NfL, and risk of a dementia diagnosis seems to depend also on cholesterol levels,” the researchers noted.

They concluded that their “[S]tudy showed that the informative and diagnostic value of biomarkers developed in research cohorts with highly selected participants might be different in community-based cohorts and that the interplay with cerebrovascular injuries, and vascular and genetic risk factors shall also be considered when such biomarkers are used in community settings.”