Pegcetacoplan
Pegcetacoplan (Empaveli, Syfovre), PEGylated cyclic peptide complement C3 inhibitor
Written by Aaron CuhaReviewed Sep 2026
Also known as: Empaveli, Syfovre, Aspaveli, APL-2
The same molecule holds two unrelated approvals: subcutaneous Empaveli for paroxysmal nocturnal haemoglobinuria, where it beat eculizumab by 3.84 g/dL of haemoglobin at 16 weeks, and intravitreal Syfovre for geographic atrophy, where it slowed lesion growth by 16 to 22 percent and showed no difference on any visual function endpoint at 24 months. European regulators refused the eye indication on exactly that point.
Overview
Pegcetacoplan is a symmetrical construct of two identical 13-residue cyclic peptides joined by a 40 kDa linear polyethylene glycol linker. Most of its mass is the PEG. It inhibits complement component C3, which sits upstream of the C5 target of eculizumab and ravulizumab.
The level of the cascade is the entire clinical rationale in blood. Blocking C5 stops intravascular haemolysis but leaves C3 fragments coating red cells, which the liver and spleen then clear extravascularly, so a proportion of patients on eculizumab remain anaemic. Blocking C3 addresses both routes. The FDA approved Empaveli on 14 May 2021 under NDA 215014 for paroxysmal nocturnal haemoglobinuria in adults, and the indication has since expanded to C3 glomerulopathy and primary immune-complex membranoproliferative glomerulonephritis in patients aged 12 and over, to reduce proteinuria and loss of kidney function.
The eye is a different story with a different verdict. The FDA approved Syfovre on 17 February 2023 under NDA 217171 for geographic atrophy secondary to age-related macular degeneration, on the strength of two 24-month phase 3 trials, OAKS and DERBY. Those trials measured the growth of the atrophic lesion on fundus autofluorescence imaging. Pegcetacoplan slowed that growth. DERBY missed the 12-month primary endpoint, reaching only 12 percent slowing with P equal to 0.062 for monthly dosing, though both trials reached significance by 24 months. The published paper states, in one sentence, that there were no differences in key secondary visual function endpoints at 24 months.
That sentence is why the European Medicines Agency's committee adopted a negative opinion in June 2024 and confirmed it on re-examination in September 2024. The committee's reasoning was that slowing lesion growth did not lead to clinically meaningful benefits for patients and that a treatment's benefits should affect everyday functioning. Two regulators looked at the same two trials and drew opposite conclusions about whether a slower-growing lesion counts as a benefit. That disagreement is the most useful thing on this page.
Mechanism of action
Binding to complement C3 and its activation fragment C3b, preventing cleavage of C3 and therefore blocking the common step shared by the classical, lectin and alternative complement pathways. Because C3 is upstream of C5, inhibition removes both the terminal membrane attack complex that lyses red cells in the circulation and the C3b opsonisation that tags them for removal by the liver and spleen. In the retina, the alternative pathway is implicated in the chronic local inflammation that accompanies geographic atrophy, and intravitreal injection delivers the drug to that compartment. The breadth of the blockade is also the risk: taking out C3 removes more innate immune function than taking out C5, and the label warns about serious infections caused by encapsulated bacteria.
Human evidence
Two completed randomised phase 3 programmes in two unrelated diseases. In paroxysmal nocturnal haemoglobinuria the effect on haemoglobin and transfusion need is large and clinically obvious. In geographic atrophy the effect is on a lesion measured by imaging, and no visual function benefit was demonstrated over 24 months.
- PEGASUS, 80 patients: haemoglobin difference 3.84 g/dL at week 16 in favour of pegcetacoplan, P below 0.001.
- Transfusion avoidance 85 percent versus 15 percent on eculizumab.
- PEGASUS failed to show non-inferiority on lactate dehydrogenase, the classic marker of intravascular haemolysis, which is the predictable cost of moving upstream from C5 to C3.
- OAKS and DERBY, 1258 patients: lesion growth slowed 16 to 22 percent at 24 months. DERBY missed the 12-month primary endpoint.
- No differences in key secondary visual function endpoints at 24 months, including best-corrected visual acuity, reading speed and functional reading independence.
- New-onset exudative age-related macular degeneration occurred more often on pegcetacoplan than on sham in both eye trials.
What this does not tell you: PEGASUS enrolled only patients already anaemic on eculizumab, so it says nothing about treatment-naive disease, and it was open-label with 80 patients, which matters for patient-reported outcomes such as fatigue. The eye trials measured an imaging endpoint; a slower-growing atrophic lesion is a surrogate for vision, and vision itself did not differ. Blocking C3 removes more of the innate immune system than blocking C5, and long-term infection risk is not fully characterised. Both programmes were funded by the manufacturer.
Reading the research record
This page exists to show what a surrogate endpoint dispute looks like when two competent regulators disagree in public.
The FDA approved Syfovre because lesion growth slowed. The European Medicines Agency's committee refused it because, in its own words, benefits should impact patients' everyday functioning and that was not demonstrated. Both were reading OAKS and DERBY. The disagreement is not about the data, it is about whether a measurable change on a retinal scan counts as a benefit when reading speed, functional reading independence and visual acuity did not move over two years. The re-examination was decided with dissenting votes, so the committee itself was not unanimous.
Two further facts belong in any honest account. DERBY, one of the two trials, missed its primary endpoint at 12 months. And conversion to neovascular age-related macular degeneration, a distinct and serious complication, was more frequent on pegcetacoplan than on sham in both trials, which is a real trade-off rather than a background rate.
The blood indication is a different matter and should not be tarred with the same brush. A haemoglobin difference of 3.84 g/dL and transfusion avoidance rising from 15 to 85 percent are outcomes patients feel directly. Even there, the trial reported one endpoint it did not meet, and reporting that is the standard this archive holds itself to.
The evidence, charted
Fig. 1 · evidence composition
4of 4 citations (100%) are in people
Every citation cited here is Human work; no other study type is cited on this page. This count is of our own citation list and understates any literature larger than the sources we cite.
Fig. 2 · evidence over time
Evidence spans 3 distinct years, 2021 to 2025, counted from the citation list on this page.
Fig. 3 · legal status at a glance
US
Approved
UK
Approved
AU
Approved
CA
Approved
Approved in all four jurisdictions shown. A jurisdiction's classification is a regulatory fact, not a verdict on the science; see Legal status below for the full text and any notes.
Key studies & citations
- Human2021
Pegcetacoplan versus Eculizumab in Paroxysmal Nocturnal Hemoglobinuria
PEGASUS. Phase 3, open-label, randomised, in adults with paroxysmal nocturnal haemoglobinuria and haemoglobin below 10.5 g/dL despite eculizumab. 41 to pegcetacoplan monotherapy, 39 to eculizumab. Primary endpoint, change in haemoglobin at week 16, favoured pegcetacoplan by an adjusted least-squares mean difference of 3.84 g/dL, P below 0.001. 35 of 41 (85 percent) on pegcetacoplan avoided transfusion versus 6 of 39 (15 percent). Non-inferiority was shown for change in reticulocyte count but NOT for change in lactate dehydrogenase. Injection site reactions 37 versus 3 percent and diarrhoea 22 versus 3 percent; breakthrough haemolysis 10 versus 23 percent. No meningitis in either group.
New England Journal of Medicine - Human2023
Pegcetacoplan for the treatment of geographic atrophy secondary to age-related macular degeneration (OAKS and DERBY): two multicentre, randomised, double-masked, sham-controlled, phase 3 trials
1258 patients aged 60 and over. At 12 months OAKS met its primary endpoint, with lesion growth slowed 21 percent monthly (P equal to 0.0004) and 16 percent every other month (P equal to 0.0055), while DERBY missed it, at 12 percent (P equal to 0.062) and 11 percent (P equal to 0.085). By 24 months both were significant, 22 and 18 percent in OAKS and 19 and 16 percent in DERBY. The paper states there were no differences in key secondary visual function endpoints at 24 months. New-onset exudative age-related macular degeneration at 24 months: 11 and 8 percent on pegcetacoplan versus 2 percent on sham in OAKS, and 13 and 6 percent versus 4 percent in DERBY. Funded by Apellis Pharmaceuticals.
The Lancet - Human2025
Efficacy and Safety Maintained up to 3 Years in Adults with Paroxysmal Nocturnal Hemoglobinuria Receiving Pegcetacoplan
Longer-term data out to three years in adults with paroxysmal nocturnal haemoglobinuria. This is an open-label extension, not a randomised comparison, so it describes what happens to patients who stayed on the drug rather than what the drug does relative to an alternative.
Advances in Therapy - Human2021
Study to Evaluate the Efficacy and Safety of APL-2 in Patients With Paroxysmal Nocturnal Hemoglobinuria (PNH)
The registry record for PEGASUS. Status COMPLETED, actual enrolment 80. Confirms the small size of the trial that changed practice in this disease, and that enrolment was restricted to patients already anaemic on eculizumab.
ClinicalTrials.gov
Frequently asked questions
Why does the same drug have two brand names?
Because it holds two unrelated approvals with different routes and doses. Empaveli is a subcutaneous infusion for paroxysmal nocturnal haemoglobinuria and now also C3 glomerulopathy. Syfovre is an intravitreal injection into the eye for geographic atrophy. Same molecule, different product, different prescriber, different specialty.
Does Syfovre improve vision?
No. The trials measured the growth rate of the atrophic lesion on retinal imaging, and that slowed by 16 to 22 percent at 24 months. On the visual function endpoints, including visual acuity, reading speed and functional reading independence, there were no differences at 24 months. That distinction is the whole argument about this drug.
Why did Europe reject it when the FDA approved it?
The European Medicines Agency's committee adopted a negative opinion in June 2024 and confirmed it on re-examination in September 2024, on the reasoning that slowing lesion growth did not lead to clinically meaningful benefits and that benefits should affect a patient's everyday functioning. The FDA accepted the imaging endpoint. Same trials, different judgement about what counts as a benefit.
Is pegcetacoplan better than eculizumab?
In the specific population PEGASUS studied, adults still anaemic on eculizumab, it produced 3.84 g/dL more haemoglobin at week 16 and lifted transfusion avoidance from 15 to 85 percent. It did not meet non-inferiority on lactate dehydrogenase. It has not been tested against eculizumab in patients who have never had complement inhibition.
What are the main risks?
Blocking C3 removes more of the innate immune defence than blocking C5, so the label warns about serious infections caused by encapsulated bacteria and requires vaccination. Injection site reactions occurred in 37 percent in PEGASUS and diarrhoea in 22 percent. For the eye product the warnings include endophthalmitis, retinal detachment, retinal vasculitis and vascular occlusion, and conversion to neovascular age-related macular degeneration was more common than on sham.