Cerebrospinal fluid levels of chitinase 3-like 1 and interleukin-6 can predict response to platform therapies in relapsing multiple sclerosis.

Background: Platform therapies, such as dimethylfumarate, teriflunomide, and interferons beta, are insufficient to control relapsing-remitting multiple sclerosis in many patients. Having biomarkers available to stratify patients as good or poor responders before starting treatment would represent a considerable advance.

Methods: We tested serum and cerebrospinal fluid (CSF) levels of chitinase 3-like 1 (c and s-CHI3L1), soluble trigger receptor expressed on myeloid cells (sTREM2) and neuronal pentraxin 2 (NPTX2), and CSF levels of neurofilament light chains and of interleukin-6 (c-IL6) collected before treatment start in 70 patients in Eastern France. We explored associations with the status responder or non-responder after 12 months of efficient treatment. Non-responders were defined as patients with sign of activity, or confirmed increase of disability, or treatment cessation due to « inefficacy» according to the treating physician.

Results: 39 patients were non-responders (55.7%). c-CHI3L1 taken as a binary variable (under or above 164.6 ng/mL) and c-IL6 either as a binary variable (detectable or not) were associated with being non-responders. The positive predictive value of being non-responder in case of c-CHI3L1 higher than 164.6 ng/mL was 0.63 (95 confidence interval 95CI 0.43-0.82). No responder patients had detectable c-IL6, 12 non-responders (46.2%) had detectable c-IL6. Having high values of c-CHI3L1 and/or detectable c-IL6 was associated with an 8.33 higher risk of being non-responder after 1 year of effective platform treatment (95CI 1.44-33.33; p = 0.001).

Conclusions: CHI3L1 and IL6 in CSF could predict the first-year response to platform therapies.