Neurofilament light in serum: Reference values and effect of risk factors for multiple sclerosis

BACKGROUND: The measurement of neurofilament light (NFL) in blood samples has been established as a sensitive measure of neuroaxonal damage in a wide range of diseases in the peripheral and central nervous system, including multiple sclerosis (MS). Previous studies have identified confounding factors that may influence the serum concentration of NFL.

AIM: We aimed at investigating the relationship between known confounders (age, body mass index, blood volume) and risk factors for MS (smoking and human leukocyte antigen (HLA)) on serum concentrations of NFL in control subjects. In addition, we compared different methods for correction for confounders when applied to newly diagnosed patients with MS.

METHODS: We measured serum concentrations of NFL by single molecule array analysis in 1.101 control subjects without neurological disease from 4 different cohorts (including 906 healthy blood donors) and 72 patients with newly diagnosed relapsing-remitting MS. A questionnaire on smoking habits was distributed to the 906 healthy blood donors, and the HLA risk alleles HLA-DRB1*15:01 and HLA-A*02:01 were genotyped by TaqMan allelic-discrimination PCR analysis in these subjects.

RESULTS: We confirmed that serum concentrations of NFL increase with age, but we also found that sample storage conditions for the different cohorts of control subjects had a substantial effect. Prolonged storage time and storage at -20° were independently associated with lower serum concentrations of NFL than shorter storage time and storage at -80° In samples from the large cohort of blood donors, we confirmed an association between high BMI and high blood volume with lower serum concentrations of NFL and found that this association was marginally stronger for BMI than for blood volume. We found no association between smoking and HLA risk factors for MS with serum concentrations of NFL in the blood donor cohort. Finally, we found that a simple method for correcting for the effect of age on NFL performed as well as Z-scores, which consider the effect of both age and BMI. This was shown when discriminating between patients with MS and control subjects and between MS patients with and without Gd-enhancing MRI lesions.

CONCLUSIONS: We confirm an association between serum concentrations of NFL, age, and BMI, but we also find that it may often be sufficient to correct for the effect of age alone. The effect of BMI should, however, be considered along with the effect of other confounding factors, including various comorbidities.