Efficacy and safety of cemdisiran siRNA in myasthenia gravis (NIMBLE): a double-blind, randomised, placebo-controlled, phase 3 trial.

# Efficacy and safety of cemdisiran siRNA in myasthenia gravis (NIMBLE): a double-blind, randomised, placebo-controlled, phase 3 trial *The Lancet 2026* Articles Efficacy and safety of cemdisiran siRNA in myasthenia gravis (NIMBLE): a double-blind, randomised, placebo-controlled, phase 3 trial Tuan Vu, Ali A Habib, Saiju Jacob, Renato Mantegazza, Hiroyuki Murai, John Vissing, Aziz Shaibani, Todd Levine, Yessar Hussain, Andreas Meisel, Agnieszka Adamczak-Ratajczak, Jan Ilkowski, Paulo Sgobbi, Alexandre Guerreiro, Sushan Luo, Rodrigo Pavani, Umesh Chaudhari, Neda Jalali, Michelle DeVeaux, Warren Moore, Karoline A Meagher, Michael E Burczynski, Shital Kokate, Steven Sherman, Dipti Pawaskar, Nikhil Singh, Amal Souttou, Andres Sirulnik, Lorah Perlee, George D Yancopoulos, James F Howard Jr, for the NIMBLE Trial Investigators* Summary Lancet 2026; 407: 1712–25 Background Autoantibody-mediated complement activation drives pathology in acetylcholine receptor (AChR) Published Online antibody-positive generalised myasthenia gravis. Newer targeted therapies lower overall antibody concentrations or April 21, 2026 inhibit complement activity. We aimed to evaluate an siRNA (cemdisiran) targeting complement component 5 (C5) as https://doi.org/10.1016/ monotherapy and in combination with a C5 antibody (pozelimab) in generalised myasthenia gravis. S0140-6736(26)00690-2 See Comment page 1660 Methods NIMBLE is a randomised, double-blind, placebo-controlled, phase 3 trial conducted at 86 centres in *Investigators listed in 13 countries. We enrolled patients aged 18 years or older with a diagnosis of generalised myasthenia gravis, positive appendix 1 (pp 3–14) serology for anti-AChR or anti-LRP4 antibodies, and a Myasthenia Gravis-Activities of Daily Living (MG-ADL) University of South Florida score of 6 or greater. Patients were randomly allocated to receive cemdisiran monotherapy (600 mg every 12 weeks; Morsani College of Medicine, Tampa, FL, USA (Prof T Vu MD); cemdisiran group), pozelimab monotherapy (200 mg every 4 weeks; pozelimab group), combined cemdisiran (200 mg University of California Irvine, every 4 weeks) and pozelimab (200 mg every 4 weeks; combination group), or placebo, all administered subcutaneously, Irvine, CA, USA (A A Habib MD); during a 24-week double-blind treatment period. The primary endpoint was change from baseline in MG-ADL scores University Hospitals at week 24, assessed in the modified intention-to-treat (mITT) primary analysis set (the first 245 randomly allocated Birmingham NHS Foundation Trust, Birmingham, UK patients who received any dose of study treatment and had at least one post-baseline assessment). A prespecified (Prof S Jacob MD); Institute of hierarchal statistical testing strategy was used; multiplicity was controlled for comparisons of cemdisiran versus Immunology and placebo and combination versus placebo (pozelimab was not statistically tested against placebo and was used only to Immunotherapy, University of assess the contribution of components to the combination treatment). Safety was mainly assessed by recording of Birmingham, Birmingham, UK (Prof S Jacob MD); Fondazione treatment-emergent adverse events in all patients who received any dose of study treatment. This study is registered IRCCS Istituto Neurologico with ClinicalTrials.gov (NCT05070858); the current status is active, not recruiting. Carlo Besta, Milan, Italy (R Mantegazza MD); International University of Findings Between Jan 20, 2022 and July 18, 2025, 390 participants were screened and, as of the data cut-o presented in Health and Welfare, Narita, this report (July 8, 2025), 284 were randomly allocated to the cemdisiran group (79 [28%] patients), pozelimab group Japan (Prof H Murai MD); (50 [18%]), combination group (80 [28%]), and placebo group (75 [26%]). Of 277 patients who received any study Rigshospitalet University of treatment, 263 (95%) completed the double-blind treatment period. In the mITT primary analysis set, at week 24, the Copenhagen, Copenhagen, Denmark (Prof J Vissing MD); least-squares mean change from baseline in MG-ADL total score was –4·5 (SE 0·4) in the cemdisiran group (n=64), Nerve and Muscle Center of –4·0 (0·4) in the combination group (n=67), and –2·2 (0·5) in the placebo group (n=59). Placebo-adjusted least- Texas, Houston, TX, USA squares mean dierences in MG-ADL score at week 24 were –2·3 (SE 0·7; 95% CI –3·6 to –1·0; p=0·0005) in the (Prof A Shaibani MD); Baylor cemdisiran group and –1·7 (0·7; –3·0 to –0·4; p=0·0086) in the combination group. The proportion of participants College of Medicine, Houston, TX, USA (Prof A Shaibani MD); with at least one adverse event during the double-blind treatment period was 54 (69%) of 78 in the cemdisiran group, Bob Bové Neuroscience 65 (81%) of 80 in the combination group, 40 (82%) of 49 in the pozelimab group, and 54 (77%) of 70 in the placebo Institute, Scottsdale, AZ, USA group. The most common adverse event in the cemdisiran group was upper respiratory tract infection (nine [12%] of (T Levine MD); Austin 78 patients), which occurred at a similar incidence in the placebo group (eight [11%] of 70). No serious or meningococcal Neuromuscular Center, Austin, TX, USA (Y Hussain MD); Charité infections occurred in the cemdisiran group. Adverse events leading to treatment discontinuation occurred in Universitätsmedizin Berlin, two participants (3%) in the placebo group and one participant (2%) in the pozelimab group. No deaths occurred Berlin, Germany during the double-blind treatment period. Two deaths occurred after the double-blind treatment period, one of which (Prof A Meisel MD); Clinical was assessed as treatment-related by the investigator but not treatment-related by the sponsor. Research Center, Medic-R, Poznań, Poland (A Adamczak-Ratajczak PhD); Interpretation Cemdisiran monotherapy and combination therapy were eective in the treatment of generalised NZOZ Neuro-Kard, Ilkowski I myasthenia gravis, and were generally well tolerated. Subcutaneous dosing of cemdisiran, administered every Partnerzy Spółka Partnerska 3 months, could provide a convenient treatment approach for generalised myasthenia gravis. Lekarzy, Poznań, Poland (J Ilkowski MD); PSEG Centro de Pesquisa Clinica, São Paulo, Funding Regeneron Pharmaceuticals. Brazil (P Sgobbi PhD); Centro de Pesquisa Neurológica—Porto Copyright © 2026 Elsevier Ltd. All rights reserved, including those for text and data mining, AI training, and similar Alegre, Porto Alegre, Brazil (A Guerreiro MD); Huashan technologies. 1712 Articles Hospital Fudan University, Research in context Shanghai, China (S Luo MD); Regeneron Pharmaceuticals, Evidence before this study gravis, the concept that complete inhibition of the complement Tarrytown, NY, USA We searched PubMed for articles published until April, 2026, pathway is necessary to achieve optimal clinical response in (R Pavani MD, U Chaudhari MD, using various combinations of search terms “generalized patients is supported by complement activity data of a peptide N Jalali PhD, M DeVeaux PhD, W Moore PharmD, myasthenia gravis”, “C5 inhibition”, “FcRn inhibition”, inhibitor of C5, zilucoplan, from a phase 2 clinical trial; however, K A Meagher PhD, “complement cascade”, “myasthenic crisis”, “Myasthenia it has not been fully validated. While targeted therapies are M E Burczynski PhD, Gravis-Activities of Daily Living”, “Quantitative Myasthenia available, there remains an unmet need for treatments that S Kokate BPharm, Gravis”, and “immunosuppresive therapies”. No language or provide improved efficacy across dosing intervals, with a S Sherman MPH, D Pawaskar PhD, N Singh MD, year restrictions were applied. Complement activation by the favourable safety profile and reduced treatment burden. A Souttou MD, A Sirulnik MD, classical pathway is central to myasthenia gravis Added value of this study L Perlee PhD, manifestations, triggered by autoantibodies to acetylcholine G D Yancopoulos MD); The phase 3 NIMBLE study evaluated the efficacy and safety of receptor (AChR). Uncontrolled complement activation in University of North Carolina at cemdisiran monotherapy administered every 3 months by Chapel Hill, Chapel Hill, NC, patients with generalised myasthenia gravis causes membrane subcutaneous injection, and a lower dose of cemdisiran in USA (Prof J F Howard Jr MD) attack complex-dependent damage and C5a-mediated combination with low-dose pozelimab administered Correspondence to: inflammation at the neuromuscular junction, leading to failed subcutaneously every month. The study included patients with Dr Rodrigo Pavani, Regeneron neuromuscular transmission. Patients with generalised Pharmaceuticals, Tarrytown, AChR autoantibody-positive and LRP4 autoantibody-positive myasthenia gravis present with skeletal muscle weakness, NY 10591, USA generalised myasthenia gravis, who generally require advanced rodrigo.pavani@regeneron. including diplopia or ptosis, facial weakness, and swallowing therapies. The study design allowed measurement of the com difficulties. C5 inhibitors (eg, eculizumab, ravulizumab, and contribution of each component (cemdisiran and pozelimab) in See Online for appendix 1 zilucoplan) have been tested in randomised, double-blind, the treatment of generalised myasthenia gravis. Cemdisiran placebo-controlled, phase 3 trials in patients with generalised monotherapy showed robust efficacy with partial inhibition of myasthenia gravis, and approved for the treatment of complement activity, with no meningococcal infections and no generalised myasthenia gravis in several countries. Cemdisiran evidence of increased infective risk compared with placebo. is a novel triantennary N-acetylgalactosamine-conjugated siRNA targeting complement C5 mRNA that reduces plasma C5 Implications of all the available evidence concentrations. In phase 1/2 trials, cemdisiran showed robust This study showed that cemdisiran monotherapy had rapid, suppression of C5 and complement activity but, as sustained, and clinically meaningful effects compared with monotherapy, did not fully control intravascular haemolysis in placebo for the treatment of generalised myasthenia gravis, another complement dependent disease, paroxysmal nocturnal with no waning of efficacy near the end of respective 3-month haemoglobinuria (PNH). However, a low-dose combination of dosing intervals. Cemdisiran monotherapy could provide a new subcutaneous cemdisiran and pozelimab (a fully human approach targeting the complement C5 pathway, specifically C5 antibody targeting the C5 protein, which on its own would also mRNA in hepatocytes. Although this regimen only led to partial not be expected to be efficacious with a low-dose regimen) in complement inhibition, it could reduce treatment burden, with patients with PNH led to a more complete complement convenient subcutaneous dosing every 3 months. The blockade that improved efficacy compared with a standard-of- underlying mechanisms explaining the residual complement care C5 inhibitor, ravulizumab, from an initial exploratory activity while improving clinical outcomes in generalised cohort in an ongoing phase 3 study. In generalised myasthenia myasthenia gravis warrant future investigations. Introduction that lower complement activity (eg, therapies targeting Generalised myasthenia gravis is a rare, chronic, complement component 5 [C5]).7–11 These therapies both autoantibody-mediated disease aecting the neuro- appear to be immunosuppressive: FcRn antibodies muscular junction.1,2 Patients present with diplopia and/ (eg, nipocalimab) administered chronically can activate or ptosis, facial weakness, dysphagia, and skeletal muscle latent Epstein–Barr virus in rare instances,12 and C5 weakness.2 Life-threatening myasthenic crisis occurs in antibodies are associated with a risk of serious around 15–20% of patients, mostly triggered by meningococcal infections.13–15 Furthermore, due to the respiratory infection and often managed by mechanical high plasma concentrations of C5, antibodies blocking ventilation.3 this target require very high-dose intravenous In acetylcholine receptor (AChR) antibody-positive administrations or frequent subcutaneous injections. generalised myasthenia gravis, pathology at the Thus, there remains an unmet need for targeted neuromuscular junction is driven by activation of the therapies providing rapid, sustained, and improved classical complement pathway, leading to the endplate ecacy without waning of eect throughout dosing damage.4 Emerging biological therapies for generalised intervals, with a favourable safety profile and a reduced myasthenia gravis include those that generally lower treatment burden. overall antibody concentrations (eg, therapies targeting Cemdisiran is an investigational, subcutaneously neonatal Fc receptors [FcRn] or CD19+ B cells),5,6 or those administered small interfering RNA (siRNA) targeting Articles C5 mRNA, reducing plasma C5 concentrations without review board and/or ethics committee at each binding or directly inhibiting C5. It is covalently linked to participating site (appendix 1 pp 3–14). All participants a triantennary N-acetylgalactosamine (GalNAc) ligand, provided written informed consent. Participant safety targeting cemdisiran to hepatocytes (thereby reducing was monitored by an independent data monitoring C5 production; appendix 1 p 24). GalNAc–siRNA committee. conjugates generally have a more favourable safety This study is registered with ClinicalTrials.gov profile (in healthy humans and those with other (NCT05070858). complement-mediated diseases)16,17 and have a longer duration of eect than previously used lipid nanoparticle Participants encapsulated siRNAs, while also oering convenient Eligible participants were those aged 18 years or older subcutaneous dosing.18 Pozelimab is an investigational with a documented diagnosis of generalised myasthenia fully human monoclonal antibody19,20 that binds and gravis and positive serology for anti-AChR or anti- blocks C5, which is approved to treat CD55-deficient lipoprotein receptor-related protein 4 (LRP4) antibodies protein-losing enteropathy.21,22 (these participants might benefit from complement Cemdisiran monotherapy (which alone lowers C5 inhibition, as it is likely that complement is involved in concentrations by >90%) was not ecacious in the pathogenesis of patients with generalised myasthenia paroxysmal nocturnal haemoglobinuria (PNH), another gravis who have anti-LRP4 antibodies), symptoms of C5-dependent disease.16,23 However, a combination of generalised myasthenia gravis as defined by a low-dose subcutaneous cemdisiran and low-dose Myasthenia Gravis-Activities of Daily Living (MG-ADL) subcutaneous pozelimab (which on its own would also score of 6 or greater and Myasthenia Gravis Foundation not be expected to be ecacious), given in a monthly of America (MGFA) clinical class II to IVa, and dosing regimen to patients with PNH, allowed for deeper concurrent treatment with a stable-dose regimen of an complement blockade that improved ecacy compared acetylcholinesterase inhibitor and immunosuppressive with a standard-of-care C5 antibody from an initial therapy for myasthenia gravis (or a documented reason exploratory cohort in an ongoing pivotal phase 3 study.24 for not receiving therapy). The first dose of a We therefore reasoned that combining subcutaneous meningococcal vaccination (serotype ACWY and cemdisiran and pozelimab might also allow for deeper serotype B [if available]) was required before complement blockade that could improve ecacy in randomisation. Sex (male or female), race generalised myasthenia gravis and also allow for a (American Indian or Alaskan Native; Asian; Black or convenient monthly dosing regimen. African American; Native Hawaiian or other We aimed to evaluate the ecacy of a combination of Pacific Islander; White; or other), and ethnicity (Hispanic cemdisiran and pozelimab, as well as cemdisiran or Latino; or not Hispanic or Latino) were self-reported monotherapy, allowing us to measure the contribution of by the participant. Full inclusion and exclusion criteria each component in the treatment of generalised are provided in the protocol (appendix 2 pp 77–81). myasthenia gravis. Randomisation and masking Methods Patients were initially randomly allocated in a 1:1:1:1 ratio Study design to receive pozelimab monotherapy (pozelimab group), NIMBLE is a phase 3, multinational, randomised, cemdisiran monotherapy (cemdisiran group), pozelimab double-blind, double-dummy, placebo-controlled trial in and cemdisiran (combination group), or placebo (placebo participants with symptomatic generalised myasthenia group). When 50 participants had been allocated to the gravis, conducted at 86 clinical research centres and pozelimab group, further enrolment was limited to the hospitals in 13 countries (appendix 1 pp 3–14). The trial other three groups (1:1:1 randomisation). When the consists of a 24-week placebo-controlled double-blind placebo group reached approximately 65 participants, treatment period, a 28-week double-blind extension randomisation proceeded 2:2:1 to the cemdisiran, treatment period, a 68-week open-label treatment period, combination, and placebo groups. Randomisation was and a 52-week post-treatment follow-up period stratified by presence of MGFA clinical class II at (appendix 1 p 25). This report presents the prespecified screening (yes or no) and by treatment with primary ecacy and safety findings from the double- immunosuppressive therapies (none or discontinued; all blind treatment period. The trial protocol is available in immunosuppressive therapies <1 year duration; or any See Online for appendices 2 and 3 appendix 2 and statistical analysis plan in appendix 3. immunosuppressive therapy ≥1 year duration) on day 1. The trial was conducted in accordance with the Block randomisation was used to assign participants to International Council for Harmonisation E6 Guideline each treatment via Interactive Response Technology. for Good Clinical Practice, the principles of the All participants, the principal investigators, and study Declaration of Helsinki, and all applicable regulatory site personnel were masked to group allocation; masking requirements. The trial protocol and its amendments was maintained with use of dummy drugs (pozelimab (appendix 2 pp 3–17) were approved by an institutional placebo and cemdisiran placebo). 1714 Articles Procedures change from baseline in MGC total score at week 24, All study treatments were administered subcutaneously change from baseline in MG-QoL15r total score at by study site personnel. Throughout the 24-week double- week 24, consistent response on the MG-ADL (defined as blind treatment period, patients in the pozelimab group a ≥2-point reduction at two or more consecutive received pozelimab 200 mg and cemdisiran placebo at assessments spanning ≥4 weeks during the double-blind baseline (day 1) and every 4 weeks; those in the cemdisiran treatment period, based on previously reported group received cemdisiran 600 mg at baseline and thresholds of clinically meaningful improvement),31 and weeks 12, and 24, cemdisiran placebo at weeks 4, 8, 16, minimal symptom expression (defined as a score of 0 or 1 and 20, and pozelimab placebo at baseline and every on the MG-ADL, based on previously published 4 weeks; those in the combination group received assessments of therapeutic eectiveness in people with pozelimab 200 mg and cemdisiran 200 mg at baseline myasthenia gravis)11,32 at week 24. and every 4 weeks; and those in the placebo group We also assessed the following prespecified exploratory received pozelimab placebo and cemdisiran placebo at ecacy endpoints: number of myasthenic crises baseline and every 4 weeks. (including impending crises); number of any Rescue therapy was allowed in the event of myasthenic hospitalisations; total number of days of hospitalisation; crisis or impending crisis. and use of rescue therapy. Ecacy during the double-blind treatment period was Safety assessments included treatment-emergent assessed with validated outcome measures: MG-ADL,25,26 adverse events (graded on a scale of mild, moderate, or Quantitative Myasthenia Gravis (QMG),27 Myasthenia severe), clinical laboratory tests, electrocardiograms, Gravis Composite (MGC),28 and Myasthenia Gravis physical examination, bodyweight, vital signs, and Quality of Life 15-item (revised; MG-QOL15r).29,30 The suicidal ideation and behaviour. Immunogenicity was MG-ADL is a clinician-administered, patient-reported measured by the incidence of antidrug antibodies to outcome measure assessing daily functions impacted by cemdisiran and pozelimab over time. generalised myasthenia gravis; total score ranges from 0 Pharmacokinetic outcomes included concentrations of to 24 with higher scores indicating greater functional cemdisiran and its predominant active metabolite in impairment and disability. The QMG is a 13-item plasma and concentration of pozelimab in serum. performance test that assesses a person’s current signs Pharmacodynamic outcomes included percentage and symptoms severity of myasthenia gravis; QMG total change from baseline in total complement haemolysis scores range from 0 to 39 with higher scores representing activity (measured as CH50)33 in serum and greater impairment. The MGC scale consists of ten items concentrations of total C5 in plasma. A full list of that measure symptoms and signs of myasthenia gravis; outcomes is provided in the protocol (appendix 2 scores range from 0 to 50, with higher scores indicating pp 60–61). higher impairment. The MG-QOL15r is a self- administered, validated patient-reported outcome Statistical analysis instrument consisting of 15 items that measure a The prespecified hierarchical statistical analysis plan person’s quality of life as related to myasthenia gravis; (appendix 3) provided for multiplicity-corrected total score ranges from 0 to 30, with higher scores comparisons of combination versus placebo, and representing greater impairment. Additional details of cemdisiran monotherapy versus placebo; the formal these measures are described in appendix 1 (p 16). These analysis plan did not evaluate the pozelimab versus ecacy assessments were collected at baseline and at placebo, which was included solely to confirm the weeks 2, 4, 8, 12, 16, 20, and 24. The MG-ADL, QMG, and contribution of components (the low pozelimab dose used MGC were administered by a trained evaluator; the was expected to minimally inhibit C5 as a monotherapy, MG-QoL15r was completed by the participant. and to only show ecacy in the setting of the lower C5 Details of pharmacokinetic, pharmacodynamic, and concentrations in combination with cemdisiran). immunogenicity analysis methods are described in Power calculations based on the primary endpoint appendix 1 (pp 16–17). indicated that 65 participants per group (placebo, combination, and cemdisiran groups) from concurrently Outcomes randomised participants were required to achieve The primary ecacy endpoint was change from baseline 88% power to detect a treatment dierence in MG-ADL total score at week 24. The key secondary of –2·3 between combination and placebo at the overall ecacy endpoint was change from baseline in QMG type I error rate of α=0·05, assuming an SD of 4 and an total score at week 24. attrition rate of 7%.9 Using the same SD and attrition Other prespecified secondary ecacy endpoints rate, the power was 81% to detect a similar treatment included MG-ADL response (defined as a ≥3-point dierence between cemdisiran and placebo at α=0·025. reduction in MG-ADL total score from baseline at We defined a modified intention-to-treat population as week 24),9 QMG response (defined as a ≥5-point all randomly allocated participants who received any reduction in QMG total score from baseline at week 24),9 amount of study drug and had at least one post-baseline Articles ecacy assessment in the double-blind treatment period. The same analysis prespecified for the primary Ecacy analyses of primary and all secondary endpoints endpoint was used to assess the contribution of each included in the hierarchical testing procedure (appendix 1 monotherapy to the combination therapy. The ecacy of p 26) were conducted in the mITT primary analysis set, cemdisiran monotherapy versus combination therapy defined as all participants who met mITT criteria among (ie, the contribution of pozelimab to combination the first 245 participants to be randomly allocated in the therapy) was assessed using the mITT analysis set. The double-blind treatment period. ecacy of pozelimab monotherapy versus combination 390 patients assessed for eligibility 101 excluded 95 not meeting inclusion or exclusion criteria 5 participant decision 1 other 5 not randomised at time of database cutoff 4 completed screening but not randomised 1 in screening 284 randomly assigned 75 assigned to placebo group 80 assigned to combination group 79 assigned to cemdisiran group 50 allocated to pozelimab group 5 excluded 1 excluded due to GCP quality 1 excluded due to GCP quality 2 did not receive study treatment issues issues 3 excluded due to GCP quality issues 70 received at least one dose of 80 received at least one dose of 78 received at least one dose of 49 received at least one dose of assigned study treatment assigned study treatment assigned study treatment assigned study treatment (included in safety analysis set) (included in safety analysis set) (included in safety analysis set) (included in safety analysis set) 0 included in PK analysis set* 66 included in PK analysis set* 62 included in PK analysis set* 41 included in PK analysis set* 10 discontinued study during 3 discontinued study during 1 discontinued study during double-blind treatment period double-blind treatment period double-blind treatment period 1 adverse event 1 physician decision† (due to adverse event) 2 physician decision 2 participant decision‡ 1 lost to follow-up 6 participant decision 60 ongoing and completed 77 ongoing and completed 78 ongoing and completed 48 ongoing and completed double-blind treatment period double-blind treatment period double-blind treatment period double-blind treatment period 4 with no post-baseline 5 with no post-baseline 3 with no post-baseline assessments assessments assessments 66 with at least one post-baseline 75 with at least one post-baseline 75 with at least one post-baseline 49 with at least one post-baseline assessment (included in mITT assessment (included in mITT assessment (included in mITT assessment (included in mITT analysis set) analysis set) analysis set) analysis set) 59 in mITT primary analysis 67 in mITT primary analysis 64 in mITT primary analysis 49 in mITT primary analysis set§ set§ set§ set§ 48 in mITT-PC analysis set¶ 55 in mITT-PC analysis set¶ 52 in mITT-PC analysis set¶ 49 in mITT-PC analysis set¶ Figure 1: Flow diagram for the double-blind treatment period GCP=Good Clinical Practice. mITT=modified intention-to-treat. mITT-PC=mITT pozelimab-concurrent. PK=pharmacokinetic. *The PK analysis set consisted of all patients who received at least one dose of active study drug and had at least one drug concentration result after receiving the first dose. †One participant was discontinued by physician decision in conjunction with a psychiatric consultation of the participant’s suicidal ideation (expressed while hospitalised for myasthenic crisis). ‡Participants withdrew consent from the study (one due to an unspecified personal or family situation, and one due to feeling worse). §mITT primary analysis set consisted of those who received at least one dose of treatment and had at least one post-baseline assessment among the first 245 patients to be randomly allocated. ¶mITT-PC analysis set consisted of all patients who were randomly allocated on or before the date on which the 50th patient was allocated to the pozelimab group, received at least one dose of treatment, and had at least one post-baseline assessment. 1716 Articles therapy (ie, the contribution of cemdisiran to combination Placebo Combination Cemdisiran Pozelimab therapy) was assessed using the mITT pozelimab- (n=59) (n=67) (n=64) (n=49) concurrent (mITT-PC) analysis set, which included all Age, years participants within the mITT who were randomly Mean 49·5 (15·6) 48·7 (15·1) 52·7 (15·7) 51·0 (15·6) allocated on or before the date on which the <35 12 (20%) 13 (19%) 10 (16%) 11 (22%) 50th participant was randomly allocated in the pozelimab 35–64 33 (56%) 45 (67%) 34 (53%) 27 (55%) monotherapy group. ≥65 14 (24%) 9 (13%) 20 (31%) 11 (22%) The assessment of the contribution of components was Region conducted through 95% Cls comparing the combination Japan 3 (5%) 2 (3%) 3 (5%) 2 (4%) therapy to each of the monotherapies. For cemdisiran Rest of world 56 (95%) 65 (97%) 61 (95%) 47 (96%) monotherapy versus combination therapy, assuming an Ethnicity SD of 4·0 and a dropout rate of around 7%, 50 participants Hispanic or Latino 4 (7%) 6 (9%) 8 (13%) 3 (6%) per group would yield a 95% CI halfwidth of 1·65 points. Not Hispanic or Latino 53 (90%) 56 (84%) 54 (84%) 44 (90%) For pozelimab monotherapy versus combination therapy, Not reported 0 2 (3%) 0 0 assuming an SD of 4·0 and a dropout rate of around 7%, Missing* 2 (3%) 3 (4%) 2 (3%) 2 (4%) 50 participants per group would yield a 95% CI halfwidth Race of 1·65 points. White 31 (53%) 40 (60%) 35 (55%) 29 (59%) The ecacy of primary and continuous secondary endpoints was assessed by a linear mixed model for Black or African American 1 (2%) 5 (7%) 4 (6%) 0 repeated measures (MMRM) with treatment, visit, Asian 24 (41%) 16 (24%) 22 (34%) 17 (35%) baseline total score, randomisation stratification factors American Indian or Alaska 0 0 0 1 (2%) Native (MGFA clinical class II at screening and treatment with Multiple 1 (2%) 2 (3%) 1 (2%) 0 immunosuppressive therapy on day 1), treatment-by- Not reported 2 (3%) 4 (6%) 2 (3%) 2 (4%) visit, and baseline total score-by-visit as fixed eects; a Sex prespecified stratification factor, region at screening Female 35 (59%) 42 (63%) 36 (56%) 24 (49%) (Japan or rest of the world), was removed from the Male 24 (41%) 25 (37%) 28 (44%) 25 (51%) MMRM due to the small number of participants in Height, cm 166·25 (10·49) 167·00 (11·47) 168·02 (9·70) 167·65 (10·69) Japan. Values after intercurrent events, including study discontinuation due to death, lack of ecacy, adverse Weight, kg 80·11 (21·61) 80·41 (23·29) 80·86 (23·37) 82·01 (26·05) event, and use of rescue therapies, were imputed by the Duration of myasthenia gravis 8·16 (9·21) 6·10 (7·14) 8·10 (8·69) 8·50 (7·56) since diagnosis, years worst observation carried forward plus random noise. Number of previous myasthenic crises or impending crises The summary of intercurrent events during the double- 0 39 (66%) 47 (70%) 39 (61%) 30 (61%) blind treatment period is provided in appendix 1 (p 20). 1 12 (20%) 13 (19%) 14 (22%) 14 (29%) Imputation was done 50 times with results summarised 2 5 (8%) 4 (6%) 7 (11%) 3 (6%) using the Rubin formula.34 The same MMRM model was 3 1 (2%) 3 (4%) 2 (3%) 2 (4%) used to assess the contribution of each monotherapy to >3 2 (3%) 0 2 (3%) 0 the combination. The ecacy analyses for all binary Time since most recent crisis, 22·94 (23·14) 23·45 (30·92) 30·27 (51·28%) 49·94 (50·04) secondary endpoints, pharmacokinetic analyses, and key months exploratory endpoints are described in appendix 1 Myasthenia gravis-related hospitalisation in the 2 years before randomisation (pp 16–18). Pharmacokinetic analyses were performed in Yes 22 (37%) 22 (33%) 24 (38%) 19 (39%) the pharmacokinetic analysis set, which was defined for No 37 (63%) 45 (67%) 40 (63%) 30 (61%) each study drug separately and included all participants Mean 1·4 (1·0) 1·8 (1·1) 1·5 (0·7) 1·3 (0·7) who received any amount of active study treatment History of positive serological test for anti-AChR antibodies† (pozelimab, cemdisiran, or both) and who had at least Yes 55 (93%) 63 (94%) 60 (94%) 48 (98%) one non-missing drug concentration result following the No 4 (7%) 4 (6%) 4 (6%) 1 (2%) first dose of the respective study treatment. History of positive serological test for anti-LRP4 antibodies† All safety analyses were done in the safety analysis set, which included all randomly allocated participants who Yes 8 (14%) 6 (9%) 5 (8%) 7 (14%) received any amount of study drug. No 51 (86%) 61 (91%) 59 (92%) 42 (86%) The cuto date for data presented in this report was History of serological test for anti-AChR antibodies and anti-LRP4 antibodies‡ July 8, 2025. Analyses were conducted using SAS Both negative 1 (2%) 0 1 (2%) 0 version 9.4. Both positive 5 (8%) 2 (3%) 2 (3%) 6 (12%) Only positive for anti-AChR 50 (85%) 61 (91%) 58 (91%) 42 (86%) antibodies Role of the funding source Only positive for anti-LRP4 3 (5%) 4 (6%) 3 (5%) 1 (2%) The funder of the study (Regeneron Pharmaceuticals) antibodies was involved in the study design, data collection, data (Table 1 continues on next page) analysis, data interpretation, writing of the report, and Articles Baseline demographics and disease characteristics Placebo Combination Cemdisiran Pozelimab were generally balanced across treatment groups (table 1 (n=59) (n=67) (n=64) (n=49) [mITT primary analysis set], appendix 1 pp 21–22 [mITT (Continued from previous page) anlaysis set]). The mITT primary analysis set consisted of MGFA classification§ 137 (57%) women and 102 (43%) men, with a mean age of IIa 12 (20%) 18 (27%) 12 (19%) 12 (24%) 50·5 years (SD 15·5). 270 (98%) participants received at IIb 12 (20%) 10 (15%) 14 (22%) 5 (10%) least one dose of quadrivalent meningococcal vaccine IIIa 18 (31%) 23 (34%) 23 (36%) 16 (33%) within 5 years before randomisation and IIIb 14 (24%) 15 (22%) 13 (20%) 13 (27%) 160 (58%) participants received serotype B meningococcal IVa 3 (5%) 1 (1%) 2 (3%) 3 (6%) vaccine within 3 years before randomisation. Number of immunosuppressive therapies used At week 24, the least-squares mean change from 0 16 (27%) 14 (21%) 14 (22%) 10 (20%) baseline in MG-ADL total score (the primary endpoint) 1 18 (31%) 31 (46%) 25 (39%) 22 (45%) was –4·5 (SE 0·4) in the cemdisiran group, –4·0 (0·4) in 2 25 (42%) 21 (31%) 25 (39%) 17 (35%) the combination group, and –2·2 (0·5) in the placebo ≥3 0 1 (1%) 0 0 group among patients in the mITT primary analysis set History of complement inhibitor therapies (table 2). Placebo-adjusted dierences in least-squares Yes 4 (7%) 1 (1%) 4 (6%) 3 (6%) means were –2·3 (SE 0·7; 95% CI –3·6 to –1·0; No 55 (93%) 66 (99%) 60 (94%) 46 (94%) p=0·0005) in the cemdisiran group and –1·7 (0·7; History of FcRn inhibitor therapy –3·0 to –0·4; p=0·0086) in the combination group. Yes 3 (5%) 4 (6%) 3 (5%) 6 (12%) Clinically meaningful improvements in MG-ADL total No 56 (95%) 63 (94%) 61 (95%) 43 (88%) score (defined as ≥2-point reduction) occurred within MG-ADL total score 9·4 (2·6) 9·4 (2·9) 8·9 (2·4) 9·8 (3·4) 2 weeks in the cemdisiran and combination groups, QMG total score 15·5 (6·2) 15·6 (5·5) 15·5 (4·9) 16·1 (6·1) deepened over time, and were sustained throughout the MGC total score 17·5 (7·2) 17·3 (6·6) 17·2 (5·8) 18·4 (8·3) double-blind treatment period (figure 2A). Per the MG-QOL15r total score 16·4 (6·1) 16·3 (5·6) 15·9 (5·8) 16·0 (6·7) statistical analysis plan, pozelimab monotherapy was not statistically tested against placebo, but showed little eect Data are mean (SD) or n (%). AChR=acetylcholine receptor. FcRN=neonatal Fc receptor. LRP4=lipoprotein receptor- at the studied dose (table 2). Pozelimab results are related protein 4. MG-ADL=Myasthenia Gravis-Activities of Daily Living. MG-QOL15r=Myasthenia Gravis Quality of Life 15-item (revised). MGC=Myasthenia Gravis Composite. MGFA=Myasthenia Gravis Foundation of America. presented in the figures and tables but are not discussed QMG=Quantitative Myasthenia Gravis. *Some countries did not report ethnicity. †Refers to both documented historical in detail since pozelimab monotherapy was included positivity before screening and positive result in central laboratory during screening. ‡Generated post-hoc. §MGFA solely to confirm the contribution of components. classification refers to baseline classification at screening; participants were not reclassified after randomisation. Least-squares mean change from baseline at week 24 in Table 1: Demographics and baseline characteristics (modified intention-to-treat primary analysis set) QMG total score (the key secondary endpoint) was –4·2 (SE 0·6) in the cemdisiran group, –3·3 (0·6) in the combination group, and –1·5 (0·7) in the placebo group decision to submit the paper for publication. The trial (table 2). Placebo-adjusted dierences in least-squares was conducted in collaboration with the sponsor and the means were –2·8 (SE 0·9; 95% CI –4·5 to –1·1; p=0·0015) NIMBLE trial investigators. in the cemdisiran group and –1·9 (0·9; –3·6 to –0·1; p=0·035) in the combination group. Reduction in QMG Results total scores also occurred within 2 weeks in both treatment Patients were recruited into the study between groups, deepened over time, and were sustained Jan 20, 2022 and July 18, 2025. 390 participants were throughout the double-blind treatment period (figure 2B). screened, of whom 101 were excluded, five had not been Additional secondary endpoints are shown in table 2 randomly allocated at the time of database cuto, and and figure 2. The proportions of participants with 284 were randomly allocated: 75 (26%) to the placebo responses on the MG-ADL and QMG in the cemdisiran group, 80 (28%) to the combination group, 79 (28%) to and combination groups were significantly greater than the cemdisiran group, and 50 (18%) the pozelimab those in the placebo group. Change from baseline at group (figure 1). Of those randomly allocated, week 24 in MGC total score did not meet statistical 277 (98%) participants received at least one dose of significance for the comparison of combination versus assigned study treatment and comprised the safety placebo; p values are therefore not reported for the population; 265 (93%) also had at least one post-baseline subsequent endpoints in the statistical testing strategy, assessment within the double-blind treatment period including MG-QOL15r (figure 2; table 2). and comprised the mITT population; and 239 (84%) Among the exploratory endpoints, the number of were included in the mITT primary analysis set. Among participants with myasthenic crisis and the number the 277 patients who received study treatment, 14 (5%) administered rescue therapy were lower in the discontinued the study before the end of the double-blind cemdisiran group than in the combination and placebo treatment period and 263 (95%) completed this treatment groups (table 3). The cemdisiran group also had the period (figure 1). lowest proportion of participants hospitalised for any 1718 Articles Placebo (n=59) Combination (n=67) Cemdisiran (n=64) Pozelimab (n=49)* Primary endpoint Change from baseline at week 24 in MG-ADL total score Mean (SD) –2·6 (3·0) –4·3 (3·5) –4·3 (3·3) –2·9 (4·0) LS mean (SE) –2·2 (0·5) –4·0 (0·4) –4·5 (0·4) –2·8 (0·5) Placebo-adjusted difference LS mean (95% CI) ·· –1·7 (–3·0 to –0·4) –2·3 (–3·6 to –1·0) NA p value ·· 0·0086 0·0005 NA Key secondary endpoint Change from baseline at week 24 in QMG total score Mean (SD) –2·0 (4·6) –3·6 (4·3) –4·1 (4·8) –1·9 (4·2) LS mean (SE) –1·5 (0·7) –3·3 (0·6) –4·2 (0·6) –1·5 (0·7) Placebo-adjusted difference LS mean (95% CI) ·· –1·9 (–3·6 to –0·1) –2·8 (–4·5 to –1·1) NA p value ·· 0·035 0·0015 NA Other secondary endpoints Proportion with MG-ADL response† Patients, n (%) 26 (44%) 44 (66%) 49 (77%) 24 (49%) RR (95% CI) ·· 1·514 (1·071 to 2·138) 1·843 (1·331 to 2·552) NA p value ·· 0·014 0·0001 NA Proportion with QMG response‡ Patients, n (%) 11 (19%) 24 (36%) 31 (48%) 9 (18%) RR (95% CI) ·· 1·929 (1·017 to 3·657) 2·678 (1·447 to 4·958) NA p value ·· 0·037 0·0006 NA Change from baseline to week 24 in MGC total score Mean (SD) –4·6 (7·9) –6·4 (6·6) –7·4 (6·4) –3·9 (7·6) LS mean (SE) –3·4 (1·0) –6·0 (0·9) –7·8 (0·9) –3·2 (1·0) Placebo-adjusted difference LS mean (95% CI) ·· –2·6 (–5·2 to 0·0) –4·3 (–6·9 to –1·8) NA p value ·· 0·048§ NA¶ NA Change from baseline to week 24 in MG-QOL15r total score Mean (SD) –2·9 (5·4) –4·3 (6·2) –4·5 (6·6) –1·6 (5·8) LS mean (SE) –2·3 (0·9) –4·3 (0·8) –4·7 (0·8) –1·6 (0·9) Placebo-adjusted difference LS mean (95% CI) ·· –2·0 (–4·2 to 0·3) –2·4 (–4·7 to –0·2) NA p value ·· NA¶ NA¶ NA Proportion with minimal symptom expression on MG-ADL|| Patients, n (%) 4 (7%) 15 (22%) 13 (20%) 7 (14%) RR (95% CI) ·· 3·676 (1·194 to 11·318) 3·390 (1·047 to 10·969) NA p value ·· NA¶ NA¶ NA Proportion with consistent response on MG-ADL** Patients, n (%) 37 (63%) 55 (82%) 52 (81%) 31 (63%) Estimated OR (95% CI) vs placebo ·· 1·309 (1·041 to 1·647) 1·308 (1·041 to 1·644) NA p value ·· NA¶ NA¶ NA Endpoints are presented in order of the prespecified statistical testing strategy. 95% CI widths were not adjusted for multiplicity and should not be used in place of hypothesis testing. 95% CIs were derived from the linear mixed model for repeated measures model. LS=least-squares. MG-ADL=Myasthenia Gravis—Activities of Daily Living. MG-QOL15r=Myasthenia Gravis Quality of Life 15-item (revised). MGC=Myasthenia Gravis Composite. NA=not applicable. OR=odds ratio. QMG=Quantitative Myasthenia Gravis. RR=relative risk. *Placebo-adjusted differences, RRs, and corresponding p values for pozelimab monotherapy versus placebo are not presented as this comparison was not in the scope of NIMBLE study; the pozelimab monotherapy group was included in the study to evaluate the contribution of components only. †Defined as a ≥3-point reduction (improvement) in MG-ADL total score from baseline at week 24. ‡Defined as a ≥5-point reduction (improvement) in QMG total score from baseline at week 24. §Not statistically significant at prespecified final α=0·0459; the remaining α was used in an interim analysis. ¶p values not reported as secondary endpoints were subject to hierarchical testing and the previous endpoint was not statistically significant. ||Defined as an MG-ADL total score of 0 or 1 at week 24. **Defined as a ≥2-point reduction (improvement) in MG-ADL total score on two or more consecutive assessments spanning 4 or more weeks during the double-blind treatment period. Table 2: Primary and secondary efficacy endpoints (modified intention-to-treat primary analysis set) Articles reason and the lowest total number of days of Despite its clinical ecacy, cemdisiran monotherapy hospitalisation for any reason versus the other treatment showed a lower percentage of terminal complement groups. All participants who were hospitalised in the inhibition at week 24 than the combination therapy placebo group had at least one admission related to (figure 2G). Mean percentage change from baseline to myasthenia gravis, while the cemdisiran, combination, week 24 in CH50 was –76·6% (SD 15·9) in the cemdisiran and pozelimab groups each had one hospitalisation group, –99·6% (2·4) in the combination group, and related to myasthenia gravis. 4·1% (25·7) in the placebo group. In participants who The least-squares mean dierences in change from received cemdisiran monotherapy, a change of baseline in MG-ADL total score at week 24 were 0·2 approximately –55·8% (15·2) from baseline in CH50 (SE 0·6; 95% CI –1·0 to 1·4; p=0·74) for combination inhibition was observed by week 2 and reached a peak therapy versus cemdisiran monotherapy (assessed in the eect of around –76·8% (14·6) change from baseline by mITT analysis set) and –0·9 (0·7; –2·3 to –0·5; p=0·22) week 4, which was sustained throughout the double- for combination therapy versus pozelimab monotherapy blind treatment period. In participants who received (assessed in the mITT-PC analysis set). Thus, combination combination therapy, an approximately –83·5 (18·0) therapy showed no added benefit over cemdisiran change from baseline in inhibition of CH50 was observed monotherapy. by week 2 and progressed to around –98·9 (4·6) by A MG-ADL total score B QMG total score 0 0 –1 –1 –2 –2 –3 –3 –4 –4 –5 –5 –6 –6 Baseline2 4 8 12 16 20 24 Baseline2 4 8 12 16 20 24 Study week Study week Number of patients Placebo group Pozelimab group Combination group Cemdisiran group (Figure 2 continues on next page) 1720 enilesab morf egnahc naem SL C Proportion of participants with improvement in MG-ADL total score from baseline to week 24 ≥10 ≥9 ≥8 ≥7 ≥6 ≥5 ≥4 ≥3 ≥2 100 80 60 40 20 0 20 40 60 80 100 Proportion of participants (%) )noitcuder( tnemevorpmi tniop mumuniM Placebo group Pozelimab group Combination group Cemdisiran group 59 59 58 55 54 52 54 53 59 59 58 55 54 52 54 52 49 48 46 48 47 47 48 48 49 48 46 48 47 47 48 48 67 64 67 67 66 64 64 64 67 64 67 67 66 64 64 64 64 64 64 64 64 62 63 64 64 64 64 64 64 62 63 64 D Proportion of participants with improvement in QMG total score from baseline to week 24 Placebo (n=59) Placebo (n=58) 6 6 · · 3 0 1·7 Combination (n=67) ≥10 12· 9 5 ·0 5·2 Combination (n=67) Cemdisiran (n=64) Cemdisiran (n=64) 10 7 ·4 ·8 1·7 ≥9 1 1 6 4 ·4 ·1 8·6 1 1 4 4 · · 9 1 5·1 ≥8 2 1 0 9 · · 3 4 8·6 23 1 · 8 9 ·8 8·5 ≥7 31·3 22·4 15·5 3 3 7 4 ·3 ·4 15·3 5 4 1 9 ·6 ·3 23·7 ≥6 42·2 29·9 17·2 68·8 53·7 30·5 ≥5 48·4 35·8 19·0 t R h e r s e p s o h n o d ld er 76·6 65·7 44·1 R th e r s e p s o h n o d ld er ≥4 57·8 49·3 22·4 81· 7 3 4·6 50·8 ≥3 64 5 ·1 8·2 29·3 100 80 60 40 20 0 20 40 60 80 100 Proportion of participants (%) Articles E MGC total score F MG-QOL15r total score 0 0 –1 –1 –2 –3 –2 –4 –3 –5 –4 –6 –7 –5 –8 –6 –9 –10 –7 Baseline 2 4 8 12 16 20 24 Baseline 2 4 8 12 16 20 24 Number of patients Placebo group Pozelimab group Combination group Cemdisiran group Figure 2: Patient-reported outcomes, performance measures, and CH50 during the double-blind treatment period (modified intention-to-treat primary analysis set) (A, B) Change from baseline in MG-ADL total score (A) and QMG total score (B) over time; error bars are SEs. (C, D) Proportion of participants with improvement in MG-ADL total score (C) and QMG total score (D) from baseline by week 24. (E, F) Change from baseline in MGC total score (E) and MG-QOL15r total score (F) over time; error bars are SEs. (G) Percentage change from baseline in CH50 over time; error bars are 95% CIs. CH50=total complement haemolysis activity. LS=least- squares. MG-ADL=Myasthenia Gravis—Activities of Daily Living. MGC=Myasthenia Gravis Composite. MG-QOL15r=Myasthenia Gravis Quality of Life 15-item (revised). QMG=Quantitative Myasthenia Gravis. week 12, which was sustained throughout the double- 65 (81%) of 80 in the combination group, 40 (82%) of blind treatment period. The ecacy benefit was 49 in the pozelimab group, and 54 (77%) of 70 in the seemingly uncoupled from the degree of complement placebo group (table 4). Most adverse events were mild inhibition even at the earliest timepoints—ie, the ecacy to moderate in severity. The most common adverse of cemdisiran monotherapy matched or exceeded that of event in the cemdisiran group was upper respiratory the combination at week 2 in most measures, despite tract infection (nine [12%] of 78 patients), which showing a smaller change from baseline in CH50 occurred at a similar incidence in the placebo group inhibition (figure 2). (eight [11%] of 70). Overall, the rates of infection during Total C5 concentrations are presented in appendix 1 the double-blind treatment period were 21 (27%) of (pp 19, 27). 78 in the cemdisiran group, 30 (38%) of 80 in the The proportion of participants with at least combination group, and 28 (40%) of 70 in the placebo one adverse event during the double-blind treatment group. There were no meningococcal infections in any period was 54 (69%) of 78 in the cemdisiran group, group during the double-blind treatment period. enilesab morf egnahc naem SL Study week Study week G CH50 10 –10 –20 –30 –40 –50 –60 –70 –80 –90 –100 Baseline 1 2 4 12 24 Study week Number of patients Placebo group Pozelimab group Combination group Cemdisiran group enilesab morf egnahc egatnecrep naeM Placebo group Pozelimab group Combination group Cemdisiran group 59 59 57 55 54 52 54 52 54 54 54 52 51 49 50 49 48 47 45 47 46 46 47 47 46 45 43 45 45 44 45 45 67 64 67 67 66 64 64 64 64 61 64 64 63 61 61 61 63 63 63 63 63 60 62 63 60 59 60 60 60 59 60 60 58 52 52 51 46 40 41 37 39 40 40 37 66 63 61 59 58 48 60 58 54 56 53 44 Articles 23 participants had an adverse event of special interest, Placebo Combination Cemdisiran Pozelimab among which the most common were hypersensitivity (n=66) (n=75) (n=75) (n=49) and injection-site reaction in the cemdisiran and Participants with at least one myasthenic 9/66 (14%) 6/75 (8%) 1/75 (1%) 0/49 combination groups. All hypersensitivity and injection- crisis (including impending crisis) site reactions were non-serious and mild to moderate in Participants with at least one day of any 10/66 (15%) 7/75 (9%) 2/75 (3%) 5/49 (10%) hospitalisation post-randomisation* intensity, and none led to treatment discontinuation Total days of hospitalisation post- 117 60 10 50 (appendix 1 p 23). One participant who had a history of randomisation autoimmune hepatitis and primary biliary cholangitis at Participants administered rescue therapy† 6/59 (10%) 4/67 (6%) 1/64 (2%) 1/49 (2%) baseline in the cemdisiran group had a liver alanine aminotransferase elevation, which was not related to the Data are n/N (%) or n. *All ten participants in the placebo group had myasthenia gravis-related hospitalisations; one participant in each of the combination, cemdisiran, and pozelimab groups had myasthenia gravis-related study treatment. hospitalisations. †The modified intention-to-treat primary analysis set is presented for this endpoint only as rescue The proportion of participants with at least one serious therapy was considered as an intercurrent event in efficacy analyses. adverse event during the double-blind treatment period Table 3: Exploratory endpoints during the double-blind treatment period (modified intention-to-treat was two (3%) of 78 in the cemdisiran group, seven (9%) analysis set) of 80 in the combination group, and ten (14%) of 70 in the placebo group (table 4). Three serious adverse events occurred in two participants in the cemdisiran group: Placebo Combination Cemdisiran Pozelimab uterine leiomyoma and adenomyosis (one participant); (n=70) (n=80) (n=78) (n=49) and decreased oxygen saturation (one participant). Any adverse event 54 (77%) 65 (81%) 54 (69%) 40 (82%) These events were severe in intensity and were Mild 26 (37%) 35 (44%) 31 (40%) 12 (24%) considered not related to the study treatment. No Moderate 21 (30%) 22 (28%) 21 (27%) 23 (47%) serious infections occurred in the cemdisiran group Severe 7 (10%) 8 (10%) 2 (3%) 5 (10%) during the double-blind treatment period. In the Severe adverse event 7 (10%) 8 (10%) 2 (3%) 5 (10%) placebo group, most serious adverse events were related Serious adverse event 10 (14%) 7 (9%) 2 (3%) 5 (10%) to myasthenia gravis crisis, impending crisis, or Adverse event of special interest* 2 (3%) 12 (15%) 8 (10%) 1 (2%) myasthenic exacerbations. Serious adverse events Treatment-related adverse event 16 (23%) 32 (40%) 23 (29%) 13 (27%) considered related to study treatment were pulmonary embolism and deep vein thrombosis in the combination Adverse event leading to study 2 (3%) 0 0 1 (2%) treatment discontinuation group (one patient), cellulitis in the placebo group Adverse event leading to dose 3 (4%) 3 (4%) 1 (1%) 1 (2%) (one patient), and urinary tract infection in the placebo interruption or reduction group (one patient). Adverse events of infections and 28 (40%) 30 (38%) 21 (27%) 25 (51%) No deaths occurred during the double-blind treatment infestations by SOC† period. After the double-blind treatment period, Serious infections and infestations 3 (4%) 2 (3%) 0 2 (4%) two deaths occurred. One death due to pneumonia— Adverse events occurring in ≥5% of participants in any treatment group, by preferred term assessed as related to the study treatment by the Upper respiratory tract infection 8 (11%) 6 (8%) 9 (12%) 6 (12%) investigator but not related to the study treatment by the Headache 7 (10%) 9 (11%) 4 (5%) 5 (10%) sponsor—occurred in a participant in the cemdisiran Nasopharyngitis 3 (4%) 2 (3%) 4 (5%) 2 (4%) group during the extension treatment period. The other Rash 1 (1%) 2 (3%) 4 (5%) 2 (4%) death was due to septic shock—assessed as not related to Urinary tract infection 2 (3%) 5 (6%) 4 (5%) 4 (8%) the study treatment by both the investigator and the Injection-site reaction 1 (1%) 6 (8%) 3 (4%) 1 (2%) sponsor—and occurred in an older participant in the Diarrhoea 5 (7%) 11 (14%) 2 (3%) 6 (12%) combination group during the open-label treatment Arthralgia 1 (1%) 5 (6%) 1 (1%) 1 (2%) period. Additional details on the two deaths are described Cough 1 (1%) 4 (5%) 1 (1%) 1 (2%) in appendix 1 (p 19). Myasthenia gravis 12 (17%) 4 (5%) 1 (1%) 4 (8%) Transient treatment-emergent antidrug antibody low Pain in extremity 1 (1%) 4 (5%) 1 (1%) 0 titre responses to pozelimab and cemdisiran were Pruritus 0 4 (5%) 0 0 observed in one (2%) participant who received combination treatment and one (2%) who received Data are n (%) of patients with at least one event. No adverse events leading to death were recorded during the double-blind treatment period. ALT=alanine aminotransferase. SOC=system order class. ULN=upper limit of normal. cemdisiran. *Defined as any hypersensitivity reactions potentially related to study treatment, suspected Neisseria infection, any Mean concentration of total C5 in plasma at week 24 is injection-site reactions, and any of the following abnormalities: ALT ≥3 × ULN, if baseline ALT was less than the ULN; presented in appendix 1 (p 27). Mean concentrations of ALT ≥2 × baseline, if baseline ALT was at or above the ULN; or total bilirubin ≥2 × ULN considered hepatic in origin by the investigator. †No cases of suspected Neisseria (meningococcal) infection were recorded. cemdisiran and its major active metabolite, AS(N-2)3’ cemdisiran, in plasma were quantifiable within 1–4 h Table 4: Adverse events during the double-blind treatment period (safety analysis set) after administration and fell below the lower limit of quantification in subsequent samples at week 12 (appendix 1 pp 28–29). Concentration-time profiles of pozelimab are presented in appendix 1 (p 30). 1722 Articles Discussion infections were numerically lowest with cemdisiran This phase 3 trial showed the ecacy and safety of monotherapy. The finding that cemdisiran monotherapy cemdisiran (a novel GalNAc-conjugated siRNA targeting was also associated with lower rates of infection than the C5 mRNA) as a monotherapy administered combination group might be attributed to the finding subcutaneously every 3 months, as well as in combination that it only partially inhibited complement compared with a low dose of pozelimab (a fully human antibody with the combination therapy. targeting the C5 protein) administered subcutaneously The dierent degree of ecacy found for cemdisiran every month, for the treatment of generalised myasthenia monotherapy in generalised myasthenia gravis compared gravis. The cemdisiran monotherapy and combination with that previously seen for cemdisiran monotherapy in groups both met the trial’s prespecified primary and key PNH16,23 might be the most surprising aspect of our secondary endpoints of improvement in the patient- findings. In PNH, cemdisiran monotherapy had little reported outcome of MG-ADL and the physician-reported activity, whereas the combination therapy, tested head-to- outcome of QMG versus placebo by week 24. The head versus ravulizumab in a small exploratory cohort of proportions of patients with MG-ADL and QMG an ongoing phase 3 study (by lowering the C5 load and responses by week 24 were also significantly greater with then also adding a blocking antibody) seemed to have cemdisiran monotherapy and combination therapy than better ecacy than current C5 inhibitors.24 These with placebo. The clinical benefits were rapid, sustained, findings would suggest that PNH requires complete and clinically meaningful, with no waning of ecacy complement inhibition for optimal disease control (as near the end of the respective dosing intervals. Exploratory achieved by the siRNA–antibody combination) while analyses also indicated that occurrence of generalised generalised myasthenia gravis might benefit from myasthenia gravis crisis (including impending crisis), substantial but only partial complement inhibition (as use of rescue therapy, and hospitalisations were lower achieved by the siRNA alone). This could be due to the with cemdisiran monotherapy and combination therapy diering ways in which complement is activated and than with placebo. mediates damage in these two diseases. In PNH, the A unique aspect of this study design allowed erythropoietic lineage is increasingly overtaken by assessment of the contribution of each treatment mutated clones missing endogenous complement component, which revealed that combination therapy regulators normally found on the cell membrane, leading had no additional benefit over cemdisiran monotherapy. to haemolysis; the clinical data suggest that complete As expected, the low dose of pozelimab monotherapy in complement inhibition might be required to prevent this trial had little ecacy. Cemdisiran monotherapy intravascular haemolysis.16 By contrast, in generalised reached a clinically meaningful, least-squares mean myasthenia gravis, autoantibodies activate complement placebo-adjusted dierence of –2·3 points on the and cause damage at the neuromuscular junction in the MG-ADL (the primary endpoint), which was greater than setting of normal endogenous regulators; the clinical the dierence seen in the combination group (–1·7). No data presented herein suggest that substantial but formal test of cemdisiran monotherapy superiority over incomplete complement blockade might be sucient to combination therapy was conducted in this study; prevent neuromuscular junction damage, performing at however, cemdisiran monotherapy results were generally least as well as complete complement blockade. numerically favourable for most clinical endpoints at The underlying mechanisms explaining the diering almost every timepoint assessed. These ecacy findings degrees of complement inhibition required for ecacy were surprising because cemdisiran monotherapy led to in generalised myasthenia gravis versus PNH remain to a lower proportion of terminal complement inhibition be elucidated. Peripheral CH50 might not fully reflect (~77%) than combination therapy (~99%), as measured complement activity at the neuromuscular junction, by CH50 assay; these relative degrees of complement where disease pathology actually occurs. We also inhibition were consistent with previous cemdisiran measured total C5 (appendix 1 p 27) and modelled free clinical trials.16,35 C5 (pharmacokinetic and pharmacodynamic modelling; Cemdisiran monotherapy was generally well tolerated, unpublished data), which showed free C5 reduction with no adverse events leading to treatment of ~98% with cemdisiran. Based on the mechanism of discontinuation, and no new safety concerns were action, cemdisiran suppresses the expression and identified. After the double-blind treatment period, there secretion of C5 but does not eliminate the circulating was one possible treatment-related death, as assessed by residual free C5. This residual C5 might still be activated, the investigator, in a patient in the cemdisiran group generating C5a (anaphylatoxin) and C5b. Because C5b is with diabetes and use of immunosuppressive therapies, necessary for the assembly of the C5b–9 membrane which should be noted; the potential risk of infections attack complex, even a small amount of residual free C5 was monitored carefully in patients receiving this (~2%) could be sucient to support approximately 24% treatment. There were no serious infections in the residual complement activity in the CH50 assay. By cemdisiran monotherapy group and no meningococcal contrast, the two components used in combination oer infections in any group. Overall, adverse events and a dual mechanism of action: cemdisiran reduces C5 Articles production, while pozelimab blocks the remaining Declaration of interests circulating C5, thereby fully depleting free C5 available TV has served as the University of South Florida Site Principal for complement pathway activation. Investigator for myasthenia gravis clinical trials sponsored by Alexion/ AstraZeneca Rare Disease, Amgen, argenx, Cartesian Therapeutics, The ability to achieve compelling ecacy with COUR, Dianthus Therapeutics, EMD Serono, ImmunAbs, Immunovant, cemdisiran monotherapy, while maintaining residual Johnson & Johnson, NMD Pharma, Novartis, Regeneron classical complement activity, could also help explain the Pharmaceuticals, Sanofi, VOR, and UCB; has served as a speaker for favourable safety profile observed in this study for Alexion/AstraZeneca Rare Disease, Amgen, argenx, CSL Behring, and Johnson & Johnson; has served on advisory boards and/or received monotherapy as compared with combination therapy. If consulting fees from Alexion/AstraZeneca Rare Disease, argenx, confirmed in further studies, this would have important Amgen, Dianthus Therapeutics, Johnson & Johnson, ImmunAbs, NMD implications for the risk–benefit ratio in patients—ie, if Pharma, Takeda, and VOR; and has participated in continuing medical disease control can be achieved without completely education events sponsored by AcademicCME, CMEO, PER, and MedLive. AAH has received research grants from Alexion/AstraZeneca, blocking complement activity, then the residual argenx, UCB, Immunovant, Regeneron Pharmaceuticals, CabalettaBio, complement activity could also allow for lower infection Amgen, Novartis, Arcellx, Merck, Mkarta, NMDPharma, Janssen/ risk. This finding suggests there might be an optimal Johnson&Jonhson, Cour Pharmaceuticals, and Kyverna; has received amount of residual complement activity for other consulting fees from Alexion/AstraZeneca, argenx, UCB, Regeneron Pharmaceuticals, Amgen, Novartis, Merck, NMDPharma, Grifols, autoantibody disorders characterised by aberrant Janssen/Johnson & Johnson, and Kyverna; has received honoraria from complement activity, which should be explored on a per- Alexion/AstraZeneca, argenx, UCB, Amgen, Janssen/Johnson & disease basis. Johnson; has received support for attending meetings and/or travel from A limitation of this trial is that it was not designed to Alexion/AstraZeneca, argenx, UCB, Amgen, Janssen/Johnson & Johnson; has participated in a data safety monitoring board or advisory evaluate switching from target therapies to cemdisiran board for Alexion/AstraZeneca, argenx, UCB, Regeneron monotherapy. However, as shown in an ongoing phase 3 Pharmaceuticals, Amgen, Novartis, Merck, NMDPharma, Grifols, study of PNH in which ravulizumab-treated patients were Janssen/Johnson & Johnson, and Kyverna; and received equipment, switched to cemdisiran-based therapies,24 there is no materials, drugs, medical writing, gifts or other services from Alexion/ AstraZeneca, argenx, and UCB. SJ has served as an advisory board safety issue expected upon switching therapies. member for Alexion, Alnylam, Argenx, Immunovant, Johnson & Additionally, the current findings are constrained by Johnson, Merck, Novartis, Regeneron Pharmaceuticals, and UCB; served insucient long-term safety and ecacy data, which will as an expert panel for argenx; and is a trustee of Myaware. HM has be available through extended monitoring up to 120 weeks received research grants from Japan Ministry of Health and Labour and Welfare; has received consulting fees from Alexion, argenx, and UCB; upon study completion. Another limitation is that the and has received honoraria for speaking from Novartis, Johnson & study was not powered and analysed to definitively assess Johnson, and Merck. JV has received grants/contracts from Novo the superiority of cemdisiran monotherapy over Nordisk foundation; has received consulting fees from Sanofi, Dyne combination. Subgroup analyses were not presented due Therapeutics, and Roche; has received honoraria from Edgewise Therapeutics, UCB, Alexion, and Janssen; and participated in a data to the small sample size in some subgroups (eg, history safety monitoring board for Dyne Therapeutics. ASh is stockholder of of positivity for anti-LRP4 antibodies; Black or Regeneron Pharmaceuticals. AM has received grants/contracts from African American race), which warrants future German Research Foundation, Alexion, argenx, Octapharm, and UCB; has received consulting fees from Alexion, argenx, Janssen, Novartis, investigations. Finally, the magnitude of the placebo and UCB; has received honoraria from Alexion, argenx, Amgen, Axunio, eect observed in this study was large, but is consistent Desitin, Genpharm, Janssen, Neopharm, Novartis, Sanofi, and UCB; with that in other phase 3 studies in generalised has participated in a data safety monitoring board or advisory board for myasthenia gravis.7–9 Alexion, argenx, Amgen, Janssen, and Merck; and has served as a Chairman of the Association for Research into Myasthenic Syndromes In summary, this phase 3 trial showed that cemdisiran in Germany and a member of the advisory board for German monotherapy led to rapid and sustained improvements Myasthenia Gravis Society. PS has received consulting fees from Biogen versus placebo in the treatment of symptomatic and Sanofi; has received honoraria from Alnylam, AstraZeneca, Biogen, generalised myasthenia gravis, with an acceptable safety Novartis, and Roche; and has participated in a data safety monitoring board or advisory board for Biogen and Sanofi. SL has received grants/ profile in the majority of patients. Cemdisiran contracts from National Key Research and Development Plan of the subcutaneous dosing every 3 months was ecacious National Natural Science Foundation of China. RP, UC, NJ, MD, WM, while preserving residual complement activity, with no KAM, MEB, SK, SS, DP, NS, ASo, ASi, and LP are employees and meningococcal infections and no evidence of increased stockholders of Regeneron Pharmaceuticals. GDY is an employee and stockholder of, has issued and pending patents within, and serves a infective risk compared with placebo. Cemdisiran thus leadership role in Regeneron Pharmaceuticals. JFH has received has the potential to oer a novel therapeutic option in the consulting fees from Regeneron Pharmaceuticals. RM, TL, YH, AA-R, JI, treatment arsenal for generalised myasthenia gravis. and AG declare no competing interests. Contributors Data sharing TV, HM, JV, AM, RP, UC, MD, ASh, LP, GDY, and JFH were involved in Qualified researchers can request access to study documents (including the conceptualisation of the study and interpretation of data analysis. the clinical study report, study protocol with any amendments, blank TV, RP, UC, NJ, MD, WM, KAM, MEB, SS, SK, DP, NS, and ASo case report form, and statistical analysis plan) that support the methods accessed and verified the data. All authors had access to all the included and findings reported in this Article. Individual anonymised participant data and were critically involved in the interpretation of the data. RP and data will be considered for sharing under the following conditions: once UC wrote the original draft. All authors reviewed and edited the the product and indication have been approved by major health manuscript, reviewed the final version of the manuscript, and accept authorities (eg, FDA, EMA, PMDA, etc) or development of the product responsibility for the decision to submit the manuscript for publication. has been discontinued globally for all indications on or after April, 2020, 1724 Articles and there are no plans for future development; if there is legal authority 16 Jang J-H, Wong RSM, Hartford C, et al. Safety, ecacy, and patient- to share the data; and if there is not a reasonable likelihood of participant reported outcomes from a phase 2 randomized trial of pozelimab re-identification. Requests should be submitted at https://vivli.org/. and cemdisiran combination in patients with paroxysmal nocturnal hemoglobinuria. eJHaem 2025; 6: e70095. Acknowledgments 17 Barratt J, Liew A, Yeo SC, et al, and the Cemdisiran Phase 2 Study This study was supported by Regeneron Pharmaceuticals. We thank the Investigators and Collaborators. Phase 2 trial of cemdisiran in adult trial participants and their families, the investigators, sta at the clinical patients with IgA nephropathy: a randomized controlled trial. sites (including subinvestigators and study coordinators), and members Clin J Am Soc Nephrol 2024; 19: 452–62. of the independent data monitoring committee; Caryn Trbovic and 18 An G. Pharmacokinetics and pharmacodynamics of GalNAc- Hannah H Chang from Regeneron Pharmaceuticals for their assistance conjugated siRNAs. J Clin Pharmacol 2024; 64: 45–57. with writing and developing the manuscript; Eric Bachman for his 19 Macdonald LE, Karow M, Stevens S, et al. Precise and in situ critical review and guidance; Dimittri Delevry from Regeneron genetic humanization of 6 Mb of mouse immunoglobulin genes. Pharmaceuticals for support with analysis and interpretation of patient- Proc Natl Acad Sci USA 2014; 111: 5147–52. reported outcomes; and Kuan-Ju Lin for support with analysis regarding 20 Murphy AJ, Macdonald LE, Stevens S, et al. Mice with megabase dose selection. Formatting and copy editing of the submitted manuscript humanization of their immunoglobulin genes generate antibodies was provided by Rachel McGrandle and Elke Sims from the Prime as eciently as normal mice. Proc Natl Acad Sci USA 2014; Group of Companies (Knutsford, UK), funded by Regeneron 111: 5153–58. Pharmaceuticals, according to Good Publication Practice guidelines. 21 Regeneron Pharmceuticals. VEOPOZ® [prescribing information]. For the Good Publication 2024. https://www.regeneron.com/downloads/veopoz_fpi.pdf References (accessed Nov 25, 2025). Practice guidelines see https:// 1 Gilhus NE, Tzartos S, Evoli A, Palace J, Burns TM, Verschuuren JJGM. www.acpjournals.org/ 22 Ozen A, Chongsrisawat V, Sefer AP, et al. Evaluating the ecacy Myasthenia gravis. Nat Rev Dis Primers 2019; 5: 30. doi/10.7326/M15-0288 and safety of pozelimab in patients with CD55 deficiency with 2 Hehir MK, Silvestri NJ. Generalized myasthenia gravis: hyperactivation of complement, angiopathic thrombosis, and classification, clinical presentation, natural history, and protein-losing enteropathy disease: an open-label phase 2 and 3 epidemiology. Neurol Clin 2018; 36: 253–60. study. Lancet 2024; 403: 645–56. 3 Claytor B, Cho S-M, Li Y. Myasthenic crisis. Muscle Nerve 2023; 23 Gaya A, Munir T, Urbano-Ispizua A, et al. Results of a phase 1/2 68: 8–19. study of cemdisiran in healthy subjects and patients with 4 San PP, Jacob S. Role of complement in myasthenia gravis. paroxysmal nocturnal hemoglobinuria. eJHaem 2023; 4: 612–24. Front Neurol 2023; 14: 1277596. 24 Jang JH, Patriquin C, Taneja D, et al. Ecacy and safety of 5 Howard JF Jr, Bril V, Vu T, et al, and the ADAPT Investigator Study pozelimab plus cemdisiran versus ravulizumab in patients with Group. Safety, ecacy, and tolerability of efgartigimod in patients paroxysmal nocturnal hemoglobinuria who received prior C5 with generalised myasthenia gravis (ADAPT): a multicentre, therapy. Blood 2025; 146 (suppl 1): 1420. randomised, placebo-controlled, phase 3 trial. Lancet Neurol 2021; 25 Wolfe GI, Herbelin L, Nations SP, Foster B, Bryan WW, Barohn RJ. 20: 526–36. Myasthenia gravis activities of daily living profile. Neurology 1999; 6 Antozzi C, Guptill J, Bril V, et al, and the Vivacity-MG Phase 2 52: 1487–89. Study Group. Safety and ecacy of nipocalimab in patients with 26 Janssen MF, Dewilde S, Wolfe GI, Muppidi S, Phillips G. generalized myasthenia gravis: results from the randomized Psychometric properties of MG-ADL items and MG-ADL score: phase 2 Vivacity-MG study. Neurology 2024; 102: e207937. an assessment of distributional characteristics, validity and factor 7 Vu T, Meisel A, Mantegazza R, et al. Terminal complement structure in two large datasets. J Neurol Sci 2024; 463: 123135. inhibitor ravulizumab in generalized myasthenia gravis. NEJM Evid 27 Barohn RJ, McIntire D, Herbelin L, Wolfe GI, Nations S, 2022; 1: EVIDoa2100066. Bryan WW. Reliability testing of the quantitative myasthenia gravis 8 Howard JF Jr, Bresch S, Genge A, et al, and the RAISE Study Team. score. Ann N Y Acad Sci 1998; 841: 769–72. Safety and ecacy of zilucoplan in patients with generalised 28 Burns TM, Conaway MR, Cutter GR, Sanders DB, and the Muscle myasthenia gravis (RAISE): a randomised, double-blind, placebo- Study Group. Construction of an ecient evaluative instrument for controlled, phase 3 study. Lancet Neurol 2023; 22: 395–406. myasthenia gravis: the MG composite. Muscle Nerve 2008; 9 Howard JF Jr, Utsugisawa K, Benatar M, et al, and the REGAIN 38: 1553–62. Study Group. Safety and ecacy of eculizumab in anti-acetylcholine 29 Burns TM, Grouse CK, Conaway MR, Sanders DB, and the MG receptor antibody-positive refractory generalised myasthenia gravis Composite and MG-QOL15 Study Group. Construct and concurrent (REGAIN): a phase 3, randomised, double-blind, placebo-controlled, validation of the MG-QOL15 in the practice setting. Muscle Nerve multicentre study. Lancet Neurol 2017; 16: 976–86. 2010; 41: 219–26. 10 Lee JW, Sicre de Fontbrune F, Wong Lee Lee L, et al. Ravulizumab 30 Burns TM, Sadjadi R, Utsugisawa K, et al. International clinimetric (ALXN1210) vs eculizumab in adult patients with PNH naive to evaluation of the MG-QOL15, resulting in slight revision and complement inhibitors: the 301 study. Blood 2019; 133: 530–39. subsequent validation of the MG-QOL15r. Muscle Nerve 2016; 11 Howard JF Jr, Nowak RJ, Wolfe GI, et al, and the Zilucoplan MG 54: 1015–22. Study Group. Clinical eects of the self-administered subcutaneous 31 Muppidi S. The myasthenia gravis—specific activities of daily living complement inhibitor zilucoplan in patients with moderate to profile. Ann N Y Acad Sci 2012; 1274: 114–19. severe generalized myasthenia gravis: results of a phase 2 32 Vissing J, Jacob S, Fujita KP, O’Brien F, Howard JF, and the randomized, double-blind, placebo-controlled, multicenter clinical REGAIN study group. ‘Minimal symptom expression’ in patients trial. JAMA Neurol 2020; 77: 582–92. with acetylcholine receptor antibody-positive refractory generalized 12 Janssen Pharmaceutical Companies. IMAAVYTM [prescribing myasthenia gravis treated with eculizumab. J Neurol 2020; information]. 2025. https://www.janssenlabels.com/package-insert/ 267: 1991–2001. product-monograph/prescribing-information/IMAAVY-pi.pdf 33 Costabile M. Measuring the 50% haemolytic complement (CH50) (accessed Nov 25, 2025). activity of serum. J Vis Exp 2010; 37: 1923. 13 Alexion Pharmaceuticals Inc. SOLIRIS® [prescribing information]. 34 Rubin DB. Multiple imputation for nonresponse in surveys. 2025. https://solirispro.com/pdf/Soliris_USPI.pdf (accessed New York: John Wiley & Sons, 1987. April 9, 2026). 35 Badri P, Jiang X, Borodovsky A, et al. Pharmacokinetic and 14 Alexion Europe SAS. Ultomiris [prescribing information]. 2025. pharmacodynamic properties of cemdisiran, an RNAi therapeutic https://www.ema.europa.eu/en/medicines/human/EPAR/ultomiris targeting complement component 5, in healthy subjects and (accessed Feb 2, 2026). patients with paroxysmal nocturnal hemoglobinuria. 15 UCB Inc. ZILBRYSQ® [prescribing information]. 2023. https:// Clin Pharmacokinet 2021; 60: 365–78. www.ucb-usa.com/ZILBRYSQ-prescribing-information.pdf (accessed Nov 25, 2025). --- [PDF原文](https://sci-net.xyz/storage/7932541/a87fa72dfa89455045aacf5f92e5e70caaf012af07f8a0f10fcd6814eac2aa0c/Efficacy-and-safety-of-cemdisiran-siRNA-in-myasthenia-gravis-NIMBLE-a-double-blind-randomised.pdf) DOI: 10.1016/S0140-6736(26)00690-2