The Autoimmune Hypothesis, Tested Directly

Rozanolixizumab: The Autoimmune Hypothesis Put to the Test

For years, one of the most debated questions in fibromyalgia research has been whether the condition involves autoantibodies — immune proteins that mistakenly target the body's own tissue. In September 2026, the Lancet Rheumatology published the first trial designed to answer that question directly by removing those antibodies from the blood.

The result is genuinely informative, and it is not a simple yes or no.

The Mechanism Being Tested

Rozanolixizumab is a monoclonal antibody that blocks the neonatal Fc receptor (FcRn). This receptor recycles IgG antibodies and keeps them circulating. Blocking it accelerates their breakdown, lowering total IgG levels.

The rationale came from earlier laboratory work. When researchers injected IgG taken from people with fibromyalgia into mice, the animals developed pain-like behaviour. That IgG also bound to satellite glial cells in the dorsal root ganglia — the clusters of nerve cells just outside the spinal cord. People with more widespread pain tended to have higher levels of these antibodies.

If autoantibodies were driving the pain, removing them should reduce it. That is the hypothesis this trial tested.

Trial Design

This was a Phase 2A randomized, double-blind, placebo-controlled multicentre study (NCT05643794). It screened 165 people between December 2022 and July 2024 and randomized 63 adults with severe fibromyalgia, diagnosed by the 2016 American College of Rheumatology criteria. Of those, 55 completed all phases.

Participants received a fixed 560 mg dose of rozanolixizumab once weekly by subcutaneous infusion, or a matching placebo. The study ran two 12-week double-blind treatment periods, followed by a two-week placebo run-out and a five-week safety follow-up.

What Happened: A Split Result

The drug did what it was designed to do pharmacologically. Total IgG fell by roughly 37% within one week and plateaued by four weeks, with median reductions of 62% and 59% in the two active treatment sequences. That confirms the mechanism worked.

The clinical picture was more complicated:

• The primary endpoint was missed. The change in Brief Pain Inventory interference score at 12 weeks was −0.5 points versus placebo (95% CI −1.2 to 0.2; p=0.13). The trial did meet its prespecified one-sided 10% threshold, but not the stricter two-sided 5% standard.
• A secondary measure did improve. The Revised Fibromyalgia Impact Questionnaire (FIQR) total score improved by 8.4 points at 10 weeks (95% CI −13.8 to −3.0; p=0.0026). All three domains improved: physical function, symptoms, and overall impact.
• Pain and fatigue ratings did not improve. Neither the Pain Numeric Rating Scale nor the Fatigue Numeric Rating Scale showed gains.

The pattern — improvement on a broad impact measure but not on pain intensity — is exactly the kind of result that resists a headline.

Safety

No serious treatment-emergent adverse events occurred in the rozanolixizumab group during treatment. Headache was the most common side effect, affecting 41% of participants in the first 12 weeks and 30% in the second. There were no cases of suspected aseptic meningitis.

What the Researchers Concluded

The authors were careful not to overstate the finding. Most people in the trial did not respond meaningfully to IgG reduction. But responses varied widely between individuals, and the researchers suggest a biologically defined subgroup — those carrying genuinely pathogenic IgG autoantibodies — might respond differently.

That subgroup cannot currently be identified. There are no validated assays to select patients on these antibodies, their exact targets are unknown, and their disease-causing role remains unproven. The study was also not powered to test whether the size of the IgG drop predicted the size of the clinical response.

The honest summary: lowering IgG across an unselected severe fibromyalgia population did not reliably reduce pain. Whether it helps a narrower, correctly identified group is still an open question.