A neurofilament that does not fall in year one predicts activity in year two

Design

prospective real-world cohort study with multivariable logistic regression

Population

329 adults with relapsing multiple sclerosis starting disease-modifying treatment; 182 with complete year-2 outcome data

Primary outcome

evidence of disease activity between years 1 and 2

Effect

no z-score reduction at year 1: OR 3.07 (95% CI 1.47–6.39) overall; 14.04 (2.47–79.93) treatment-naïve; 4.54 (1.26–16.30) among those with NEDA-3 at year 1

This prospective real-world cohort at the Multiple Sclerosis Centre of Catalonia followed 329 adults with relapsing multiple sclerosis starting a disease-modifying treatment, with serum neurofilament light measured at baseline and one year. The exposure was the change in z-score — age- and body-mass-adjusted — over that first year, and the outcome was evidence of disease activity between years one and two. Of 352 treatment courses, 342 ran at least a year and 182 had complete outcome data.

Mean change was −0.62 z-score points, with larger falls in treatment-naïve patients and those on monoclonal antibodies. Failure to fall at all over the first year was associated with roughly three times the odds of disease activity in year two (odds ratio 3.07, 95% CI 1.47 to 6.39). In treatment-naïve patients the association was much stronger (14.04, 2.47 to 79.93), and — the finding that changes practice — it held in patients who had achieved no evidence of disease activity at one year (4.54, 1.26 to 16.30). Adding the change to the established Rio, modified Rio and MAGNIMS response scores left it independently associated with later activity (Rio stratum 2.26, 1.12 to 4.59).

The practical content is about the reassuring patient. A person with no relapses and a clean MRI at one year is currently told their treatment is working. If their neurofilament has not come down, that reassurance is worth less than it appears, and the case for an earlier review or an earlier escalation conversation is stronger. The authors are careful: discriminative performance was modest, and this supports a complementary role rather than a standalone test.

• Measure serum neurofilament at treatment initiation as well as at one year — the change is the informative quantity, not the level
• Interpret a flat neurofilament in an otherwise stable patient as a reason to review sooner
• Use it alongside Rio or MAGNIMS scoring, not instead of it
• Do not escalate on neurofilament alone; discriminative performance was modest

Why it matters

It identifies risk in exactly the patients current monitoring calls stable.

Don't overread it

This is an observational association — no trial has shown that escalating treatment on a flat neurofilament improves outcomes.

The statistics, in plain English

The treatment-naïve odds ratio of 14.04 has an interval running from 2.47 to 79.93 — the association is almost certainly real and its magnitude is almost entirely unknown, which is what small subgroup numbers do. The overall estimate of 3.07 (1.47 to 6.39) is the usable one. Being observational, this shows that a flat neurofilament flags patients who go on to have activity, not that acting on it changes their course. 'Modest discriminative performance' means many patients flagged will do fine, and some not flagged will relapse.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.