Associations of plasma NfL, GFAP, and t-tau with cerebral small vessel disease and incident dementia: longitudinal data of the AGES-Reykjavik Study

We investigated the associations of plasma neurofilament light (NfL), glial fibrillary acidic protein (GFAP), and total tau (t-tau) with markers of cerebral small vessel disease (SVD) and with incident dementia. We also investigated whether associations of NfL, GFAP, and t-tau with incident dementia were explained by SVD. Data are from a random subsample (n = 1069) of the population-based AGES-Reykjavik Study who underwent brain MRI and in whom plasma NfL, GFAP, and t-tau were measured at baseline (76.1 ± 5.4 years/55.9% women/baseline 2002–2006/follow-up until 2015). A composite SVD burden score was calculated using white matter hyperintensity volume (WMHV), subcortical infarcts, cerebral microbleeds, and large perivascular spaces. Dementia was assessed in a 3-step process and adjudicated by specialists. Higher NfL was associated with a higher SVD burden score. Dementia occurred in 225 (21.0%) individuals. The SVD burden score significantly explained part of the association between NfL and incident dementia. WMHV mostly strongly contributed to the explained effect. GFAP was not associated with the SVD burden score, but was associated with WMHV, and WMHV significantly explained part of the association between GFAP and incident dementia. T-tau was associated with WMHV, but not with incident dementia. In conclusion, the marker most strongly related to SVD is plasma NfL, for which the association with WMHV appeared to explain part of its association with incident dementia. This study suggests that plasma NfL may reflect the contribution of co-morbid vascular disease to dementia. However, the magnitude of the explained effect was relatively small, and further research is required to investigate the clinical implications of this finding. Supplementary Information The online version contains supplementary material available at 10.1007/s11357-023-00888-1.

Data are means (standard deviation) or median (interquartile range).Abbreviations: HDL, high-density lipoprotein; TBV, total brain volume; SVD, cerebral small vessel disease; WMHV, white matter hyperintensity volume.a Data available in n=1191 in the biomarker substudy and n=5331 in the original cohort, respectively.b Data available in n=1199 in the biomarker substudy and n=5569 in the original cohort, respectively.c Data available in n=1200 in the biomarker substudy and in n=5717 in the original cohort, respectively.d Data available in the biomarker substudy in n=1199 and n=5696 in the original cohort, respectively.e Data available in 1200 in the biomarker substudy and n=5717 in the original cohort, respectively.f Data available in 1200 in the biomarker substudy and n=5760 in the original cohort, respectively.g Data available in 1200 in the biomarker substudy and n=5718 in the original cohort, respectively.h Data available in 1152 in the biomarker substudy and n=5452 in the original cohort, respectively.I TBV was expressed as percentage of intracranial volume.Data available in 1200 in the biomarker substudy and n=2588 in the original cohort, respectively.j SVD burden score was calculated by assigning one point per cerebral small vessel disease marker based on the following cut-offs (range 0-4): WMHV highest quartile vs lowest three quartiles, and for subcortical infarcts, cerebral microbleeds, and large perivascular spaces presence vs absence.Data available in 1192 in the biomarker substudy and n=2588 in the original cohort, respectively.k WMHV was expressed as percentage of intracranial volume.Data available in 1200 in the biomarker substudy and n=2588 in the original cohort, respectively.l Data available in 1200 in the biomarker substudy and n=2672 in the original cohort, respectively.m Data available in 1192 in the biomarker substudy and n=2588 in the original cohort, respectively.Hazard ratios for incident dementia are expressed per natural log-transformed pg/ml higher plasma NfL, GFAP or t-tau.Total effect, direct effect and explained effect are defined in Supplementary Figure 2. The explained effect quantifies the degree to which the SVD markers attenuated the association of plasma biomarkers with incident dementia.All analyses adjusted for age, sex, education level, diabetes status, smoking history, body mass index, total cholesterol-to-HDL cholesterol ratio, use of lipid-modifying medication, systolic blood pressure, and use of antihypertensive medication.Abbreviations: SVD, cerebral small vessel disease; NfL, neurofilament light; GFAP, glial fibrillary acidic protein; ttau, total tau; WMHV, white matter hyperintensity volume.a Highest quartile versus lowest three quartiles of WMHV.Hazard ratios for incident dementia are expressed per natural log-transformed pg/ml higher plasma NfL, GFAP or t-tau.Total effect, direct effect and explained effect are defined in Supplementary Figure 2. The explained effect quantifies the degree to which the SVD burden score a attenuated the association of plasma biomarkers with incident dementia.All analyses adjusted for age, sex, education level, diabetes status, smoking history, body mass index, total cholesterol-to-HDL cholesterol ratio, use of lipidmodifying medication, systolic blood pressure, and use of antihypertensive medication.Abbreviations: SVD, cerebral small vessel disease; TBV, total brain volume; NfL, neurofilament light; GFAP, glial fibrillary acidic protein; t-tau, total tau; a SVD burden score was calculated by assigning one point per cerebral small vessel disease marker based on the following cut-offs (range 0-4): WMHV highest quartile vs lowest three quartiles, and for subcortical infarcts, cerebral microbleeds, and large perivascular spaces presence vs absence.b TBV was expressed as percentage of intracranial volume.Hazard ratios for incident dementia are expressed per natural log-transformed pg/ml higher plasma NfL, GFAP or t-tau.Total effect, direct effect and explained effect are defined in Supplementary Figure 2. The explained effect quantifies the degree to which the SVD markers attenuated the association of plasma biomarkers with incident dementia.All analyses adjusted for age, sex, education level, diabetes status, smoking history, body mass index, total cholesterol-to-HDL cholesterol ratio, use of lipidmodifying medication, systolic blood pressure, and use of antihypertensive medication.Abbreviations: SVD, cerebral small vessel disease; WMHV, white matter hyperintensity volume; NfL, neurofilament light; GFAP, glial fibrillary acidic protein; t-tau, total tau; TBV, total brain volume.a TBV was expressed as percentage of intracranial volume.Hazard ratios for incident dementia are expressed per natural log-transformed pg/ml higher plasma NfL, GFAP or t-tau.Total effect, direct effect and explained effect are defined in Supplementary Figure 2. The explained effect quantifies the degree to which WMHV explained the association of plasma biomarkers with incident dementia.All analyses adjusted for age, sex, education level, diabetes status, smoking history, body mass index, total cholesterol-to-HDL cholesterol ratio, use of lipid-modifying medication, systolic blood pressure, and use of antihypertensive medication.Abbreviations: WMHV white matter hyperintensity volume; NfL, neurofilament light; GFAP, glial fibrillary acidic protein; t-tau, total tau.a WMHV expressed per SD percentage of total intracranial volume.

Table 2 . Analytical ranges and inter-assay coefficient of variation of the plasma biomarkers Plasma biomarker Analytical range (pg/ml) Inter-Assay Coefficient of Variance, %
a TBV and WMHV were expressed as percentage of intracranial volume.bSVD burden score was calculated by assigning one point per cerebral small vessel disease marker based on the following cut-offs (range 0-4): WMHV highest quartile vs lowest three quartiles, and for subcortical infarcts, cerebral microbleeds, and large perivascular spaces presence vs absence.Supplementary Table 4.

Characteristics of the total study population, and according to tertiles of plasma GFAP
: GFAP, glial fibrillary acidic protein; HDL, high-density lipoprotein; TBV, total brain volume; SVD, cerebral small vessel disease; WMHV, white matter hyperintensity volume; NfL, neurofilament light; t-tau, total tau.a TBV and WMHV were expressed as percentage of intracranial volume.b SVD burden score was calculated by assigning one point per cerebral small vessel disease marker based on the following cut-offs (range 0-4): WMHV highest quartile vs lowest three quartiles, and for subcortical infarcts, cerebral microbleeds, and large perivascular spaces presence vs absence.
Data are means (standard deviation) or median (interquartile range).Abbreviations

Table 5 . Characteristics of the total study population, and according to tertiles of plasma t-tau
: t-tau, total tau; HDL, high-density lipoprotein; TBV, total brain volume; SVD, cerebral small vessel disease; WMHV, white matter hyperintensity volume; NfL, neurofilament light; GFAP, glial fibrillary acidic protein.a TBV and WMHV were expressed as percentage of intracranial volume.b SVD burden score was calculated by assigning one point per cerebral small vessel disease marker based on the following cut-offs (range 0-4): WMHV highest quartile vs lowest three quartiles, and for subcortical infarcts, cerebral microbleeds, and large perivascular spaces presence vs absence.
Data are means (standard deviation) or median (interquartile range).Abbreviations

Table 7 . Association between plasma NfL, GFAP, and t-tau and the SVD burden score a on an ordinal scale
SVD burden score was calculated by assigning one point per cerebral small vessel disease marker based on the following cut-offs (range 0-4): WMHV highest quartile vs lowest three quartiles, and for subcortical infarcts, cerebral microbleeds, and large perivascular spaces presence vs absence. a

Supplementary Table 11. Total effects, direct effects, and explained effects by WMHV a on a continuous scale and expressed as higher versus lower than the median of the associations between plasma NfL, GFAP, and t-tau and incident dementia Plasma biomarker
b Highest quartile versus lowest three quartiles of WMHV.