Cerebrospinal fluid neurofilament light chain in multiple sclerosis and its subtypes: a meta-analysis of case-control studies

Abstract

Objective: Neurofilament is a biomarker of axonal injury proposed as a useful adjunct in the monitoring of patients with multiple sclerosis (MS). We conducted a systematic review and meta-analysis of case-control studies that have measured neurofilament light chain (NfL) levels in cerebrospinal fluid (CSF) of people with MS (pwMS), in order to determine whether, and to what degree, CSF NfL levels differentiate MS from controls, or the subtypes or stages of MS from each other.

Methods: Guidelines on Preferred Reporting Items for Systematic Reviews and Meta-Analyses were followed. Electronic databases were searched for published and 'grey' literature, with 151 hits. Of 51 full articles screened, 20 were included in qualitative analysis, and 14 in meta-analysis.

Results: CSF NfL was higher in 746 pwMS than 435 (healthy and disease) controls, with a moderate effect size of 0.61 (p < 0.00001). Mean CSF NfL levels were significantly higher in 176 pwMS with relapsing disease than 92 with progressive disease (2124.8 ng/L, SD 3348.9 vs 1121.4 ng/L, SD 947.7, p = 0.0108). CSF NfL in 138 pwMS in relapse (irrespective of MS subtype) was double that seen in 268 pwMS in remission (3080.6 ng/L, SD 4715.9 vs 1541.7 ng/L, SD 2406.5, p < 0.0001).

Conclusions: CSF NfL correlates with MS activity throughout the course of MS, reflecting the axonal damage in pwMS. Relapse is more strongly associated with elevated CSF NfL levels than the development of progression, and NfL may be most useful as a marker of disease 'activity' rather than as a marker of disability or disease stage.

Keywords: CSF; multiple sclerosis; neuroimmunology.

Figure 1

PRISMA 2009 flow diagram. PRISMA, Preferred Reporting Items for Systematic Review and Meta-analysis.

Figure 2

Clinically definite MS versus controls (HCs and NINDCs), subgroup and combined meta-analysis. Four studies compared CSF NfL levels in patients with MS with HCs, and nine studies used NINDCs. One group (Novakova et al) used the same control cohort in two papers (Novakova a and b). (Novakova a) was therefore excluded from the overall analysis to avoid duplication. This study is used in later subanalyses, and a sensitivity analysis including it did not alter results. CSF NfL levels are higher in MS than healthy and disease controls. CDMS, clinically definite MS; CSF, cerebrospinal fluid; HCs, healthy controls; MS, multiple sclerosis; NfL, neurofilament light chain; NINDCs, non-inflammatory neurological disease controls; SMD, standard mean deviation.

Figure 3

(A) RRMS in remission versus controls (NINDCs) and (b) RRMS in relapse versus controls (HCs and NINDCs), subgroup and combined meta-analysis. (a) CSF NfL levels were higher in patients with RRMS in remission than disease controls, with a moderate effect size of 0.67; (b) CSF NfL levels were higher in patients with RRMS in relapse than both healthy and disease controls, with a large effect size of 1.13; however, heterogeneity between studies was also significant. CSF, cerebrospinal fluid; HCs, healthy controls; NfL, neurofilament light chain; MS, multiple sclerosis; NINDCs, non-inflammatory neurological disease controls; RRMS, relapsing remitting MS; SMD, standard mean deviation.

Figure 4

(A) RRMS in relapse versus RRMS in remission and (b) all patients with MS in relapse versus all MS. (a) CSF NfL levels are higher in RRMS in relapse than in remission; (b) when progressive and patients with RRMS are combined as ‘all patients’, CSF NfL levels remain higher in relapse patients than remission patients. CSF, cerebrospinal fluid; MS, multiple sclerosis; NfL, neurofilament light chain; RRMS, relapsing remitting MS; SMD, standard mean deviation.

Figure 5

Progressive MS (combined SPMS and PPMS) versus controls (HC and NINDCs), subgroup and combined meta-analysis. CSF NfL levels were higher in patients with progressive disease than healthy and disease controls, but heterogeneity between studies was significant. HC=healthy controls; MS, multiple sclerosis; NINDCs, non-inflammatory neurological disease controls; PPMS, primary progressive MS; SMD, standard mean deviation; SPMS, secondary progressive MS.

Figure 6

RRMS versus progressive MS (combined SPMS and PPMS) meta-analysis. CSF NfL levels were higher in RRMS than patients with progressive MS, with a small, but statistically significant, effect size of 0.34. PPMS, primary progressive MS; SMD, standard mean deviation; RRMS, relapsing remitting MS; SPMS, secondary progressive MS.