Plasma neurofilament light(pNFL) levels account for 30-60% of the variance in CSF neurofilament light(cNFL) levels depending on the study. Age, disability, relapses, and the presence of contrast enhancing MRI lesions can increase both pNFL and cNFL. Additional nervous system biomarkers can now be studied in plasma. Understanding the factors that increase their variability in blood may be helpful in normalizing levels to better understand what levels are concerning for ongoing disease activity.
To evaluate factors contributing to blood and cerebrospinal fluid(CSF) discordance and determine if a correction of blood levels can better estimate what is happening in the CSF compartment.
Matched plasma and CSF samples were identified in the Rocky Mountain Multiple Sclerosis Center Biorepository at the University of Colorado. Neurofilament Light(NFL), Glial Fibrillary Acidic Protein(GFAP), tau, and Ubiquitin carboxy-terminal hydrolase L1(UCHL1) levels were assessed using Single Molecule Array(SIMOA) in a Quanterix SR-X machine. Analyses were done on log transformed NFL concentrations.
Fifty-seven patients had matched plasma and cerebrospinal fluid samples evaluated for neurofilament light which included 24 patients with multiple sclerosis(MS), 7 with neuromyelitis optica spectrum disorder(NMOSD), and 18 patients with headache whose opening pressures were <20cmH2O. These patients had a mean age of 46.5(+/-11.2) years, 75% female, mean albumin index of 6.3(+/-5.5), and BMI of 27.4(+/-5.8). The CSF and plasma concentrations in pg/ml were for NFL 1059.3(+/-3052.4) and 12.2(+/-32.4), GFAP 7621.5(+/-9713.4) and 52.9(+/-39.7), tau 41.5(+/-41.3) and 1.3(+/-0.8), UCHL1 1356.0(+/-1677.1) and 23.6(+/-32.8). Respectively the CSF vs plasma Spearman correlations (95% confidence intervals, p values) were: 0.79(0.67-0.87,<0.0001), 0.67(0.50-0.79,<0.0001), 0.75(0.61-0.85,<0.0001), and 0.70(0.54-0.81,<0.0001). Adjusting individually for age, BMI, or albumin index did not affect the correlation for NFL.
Blood and CSF levels of NFL, GFAP, tau, and UCHL1 correlated well. Models will be created that explore the relationship between Blood and CSF levels of these biomarkers.