An siRNA eased myasthenia gravis in the phase 3 NIMBLE trial

Cemdisiran, a small-interfering-RNA drug that switches off production of the immune protein complement component 5, reduced the symptoms of generalised myasthenia gravis more than placebo in the phase 3 NIMBLE trial [s1]. The finding matters because cemdisiran is given as a subcutaneous injection just once every three months, a far lighter treatment schedule than the existing complement-blocking antibodies for this rare neuromuscular disease [s1][s2].

Generalised myasthenia gravis is an autoimmune disorder in which antibodies attack the connection between nerve and muscle, causing fluctuating weakness of the eyes, face, limbs and, in severe cases, the muscles of breathing and swallowing. In the most common form, antibodies against the acetylcholine receptor recruit the complement system — a cascade of blood proteins ending in component 5 (C5) — to damage that junction [s1]. Blocking C5 works: the antibody eculizumab was shown in the phase 3 REGAIN trial to help patients with refractory antibody-positive disease and is an approved therapy, though it must be infused every two weeks [s2]. NIMBLE tested a different way to shut down the same target — not by mopping up C5 in the blood but by instructing the liver to make less of it.

What the trial did

NIMBLE was a randomised, double-blind, placebo-controlled trial run at 86 centres in 13 countries and funded by Regeneron Pharmaceuticals, the drug's developer [s1]. It enrolled adults with generalised myasthenia gravis, antibodies against the acetylcholine receptor or the related protein LRP4, and a Myasthenia Gravis Activities of Daily Living (MG-ADL) score of at least 6 — a questionnaire measure of how much the disease interferes with tasks such as chewing, speaking and rising from a chair [s1]. Patients were assigned to cemdisiran alone (600 mg every 12 weeks), the C5 antibody pozelimab alone, a combination of the two at lower doses, or placebo, all given under the skin over a 24-week double-blind period [s1]. The pozelimab-only arm was included to gauge each drug's contribution and was not itself tested against placebo [s1].

The primary endpoint was the change in MG-ADL score from baseline to week 24, assessed in the first 245 patients randomised [s1]. Between January 2022 and July 2025, 284 patients were randomly assigned — 79 to cemdisiran, 50 to pozelimab, 80 to the combination and 75 to placebo — and 263 of the 277 who received any treatment (95%) completed the double-blind period [s1].

What it found

At week 24, the least-squares mean MG-ADL score fell by 4.5 points on cemdisiran and 4.0 points on the combination, against 2.2 points on placebo [s1]. That translated to a placebo-adjusted improvement of 2.3 points for cemdisiran (95% confidence interval −3.6 to −1.0; P=0.0005) and 1.7 points for the combination (95% CI −3.0 to −0.4; P=0.0086) — both statistically significant, with cemdisiran monotherapy showing the larger effect [s1]. A roughly two-point movement on MG-ADL is the kind of change generally considered clinically meaningful for the scale.

Safety over the double-blind period looked manageable. Adverse events affected 69% of the cemdisiran group, 81% of the combination group and 77% of the placebo group; the most common event on cemdisiran, upper respiratory infection, occurred at 12% versus 11% on placebo [s1]. No deaths occurred during the double-blind phase; two occurred afterward, one judged treatment-related by the investigator but not by the sponsor [s1]. Notably for a complement drug, the trial reported no serious or meningococcal infections in the cemdisiran group — meningococcal disease is the signature danger of C5 blockade and the reason such drugs carry vaccination requirements [s1][s2].

How to read it

Two design points bound the result. This is 24-week data on a symptom questionnaire, so it speaks to short-term relief of daily-living impairment, not yet to years of disease control or to harder outcomes such as myasthenic crises [s1]. And because cemdisiran silences C5 production rather than neutralising the protein directly, its onset and offset differ from an antibody — a property that makes quarterly dosing possible but that also warrants watching the infection risk carefully as longer data accrue [s1]. The single most useful comparison for a reader is the schedule: an injection every three months versus infusions every fortnight for the established C5 antibody, against a broadly similar mechanism [s1][s2].

The myasthenia field has crowded quickly with targeted drugs — complement blockers and antibodies that strip pathogenic antibodies from the blood through the neonatal Fc receptor. Related coverage has examined the FDA approval of efgartigimod, an FcRn blocker, in a related nerve disorder and the approval of that class in seronegative myasthenia gravis.

What to watch

The open-label extension and any longer-term reporting will show whether the week-24 benefit holds and whether the reassuring infection profile survives extended dosing [s1]. Regulators will also weigh how a quarterly siRNA fits alongside the antibody-based complement drugs already available, and whether the convenience translates into real-world adherence and outcomes [s1][s2].

This article describes research and is not medical advice. Treatment choices in myasthenia gravis are decisions for treating clinicians.

Sources

• Efficacy and safety of cemdisiran siRNA in myasthenia gravis (NIMBLE): a double-blind, randomised, placebo-controlled, phase 3 trial — The Lancet, 21 April 2026
• Safety and efficacy of eculizumab in anti-acetylcholine receptor antibody-positive refractory generalised myasthenia gravis (REGAIN) — The Lancet Neurology, 2017