Multiple Sclerosis Research iconMultiple Sclerosis ResearchSep 23, 2026 ~4 min source read

Detecting the “crap gap”: wearables pick up fatigue-linked wearing-off under natalizumab

A small prospective study used wearables and biomarkers to test whether wearing-off symptoms between natalizumab infusions reflect disease activity or fatigue-related motor changes. Wearables registered subtle afternoon declines in steps that tracked reported fatigue, while serum and MRI markers did not point to ongoing inflammation or degeneration in this cohort.

Detecting the Crap Gap

Share this story

Send the public story page.

Useful takeaways from this story.

Wearing-off symptoms (WoS) were common: 19 of 28 participants (67.9%) reported WoS at least once during follow-up.

WoS correlated with self-reported fatigue but not with dosing interval, diffusion MRI changes, or serum biomarkers assessed at baseline and month 6.

Passive smartwatch data showed a modest end-of-cycle drop in afternoon steps and dynamic metrics (morning-to-afternoon step ratios, reduced peak step variability) that correlated with fatigue scores and WoS rate.

What this study looked for

The study set out to test whether wearing-off symptoms (WoS) reported by people with multiple sclerosis (PwMS) on natalizumab represent returning disease activity or are better explained by fatigue-related motor changes. The investigators combined clinical questionnaires, blood biomarkers, diffusion MRI, and passive smartwatch monitoring in a prospective, single-center cohort (SATURATE, NCT05701423).

Who and how many

Core findings

WoS were frequent: 19 of 28 participants (67.9%) reported WoS at least once and 63 of 177 questionnaires (35.6%) flagged WoS. The rate of WoS associated with higher fatigue scores but did not associate with dosing interval, diffusion MRI changes, or the serum biomarkers measured.

Wearable results were the key mechanistic signal. Passive smartwatch data revealed a subtle decline in afternoon steps toward the end of the dosing cycle. Derived dynamic features — notably the morning-to-afternoon step ratio and reduced variability in peak step counts — correlated with both fatigue scores and WoS frequency. Those patterns suggest measurable motor changes linked to fatigue that manifest as reduced afternoon activity.

What the authors conclude

In this clinically stable cohort, WoS under natalizumab did not show evidence of ongoing inflammatory or neurodegenerative disease activity based on the biomarkers and diffusion MRI used. Instead, WoS matched fatigue-related motor signatures that passive wearable devices can detect. The authors propose that dynamic wearable metrics could serve as digital endpoints for fatigue-focused interventions and clinical care.

Practical implications for clinicians and patients

  • Wearable devices can capture subtle activity changes tied to fatigue and wearing-off that standard clinic visits may miss. They offer an objective complement to symptom reports.
  • In similar clinically stable patients, WoS may reflect fluctuating fatigue and motor performance rather than subclinical inflammation detectable by the specific serum and diffusion MRI measures used here.
  • If validated, wearable-derived metrics could help track treatment-related fatigue, guide timing of clinical assessments, and inform trials targeting fatigue or wearing-off.

Next steps suggested by the study

The study is exploratory and limited by small sample size and single-center design. The next studies should test whether the wearable signatures replicate in larger, multicenter cohorts, and whether more sensitive or different biomarkers can link the observed activity changes to underlying biology when present.

More context around this story.

When Your Benchmark Leaks the Answer
Dzone iconDzoneSep 18, 2026

When Your Benchmark Leaks the Answer

A detector I built was scoring 0.067 recall on temporal errors, meaning it caught about one in fifteen of the wrong dates it was supposed to find. Wrong dates are supposed to be the easy category: extract the years from the claim, extract the years from the source, compare. There is no semantics to get wrong. I assumed

Loading more related stories...

Keep reading in the app

Open the app view to save this story, compare related coverage, and continue from the same source.

Open in app