Fixed-ratio antibody cocktails and multi-protein co-formulations raise CMC questions that single-component biologics do not: how to identify and quantify each component in the mixture, how to justify and control the ratio, how to show potency for each component, and how to show the components stay compatible over the shelf life. No single FDA or EMA guideline answers all of these, so CMC and regulatory teams planning a co-formulated program have to piece the expectations together from several sources.
The analysis below covers how FDA and EMA define co-formulated products and how those definitions differ from bispecifics and co-packaged presentations. It then reviews the general monoclonal antibody and ICH quality guidance that applies, the FDA drafts written specifically for cocktails, and the product-quality assessments for approved co-formulations. It focuses on identity, ratio control, potency, stability, and compatibility.
CMC expectations for co-formulated biologics: how FDA and EMA approach multi-antibody and multi-protein products
Neither FDA nor EMA has issued a standalone CMC guideline for biologics that combine two or more antibodies or proteins in one presentation. The expectations come from three places: general monoclonal antibody (mAb) quality guidance and ICH Q6B, a small number of FDA cocktail-specific drafts, and the product-quality assessments for approved co-formulations. Read together, these sources point to a consistent control strategy. Each component must be identifiable and quantifiable in the mixture. The ratio must be justified and controlled from lot to lot. Potency must be shown for each component where the mechanism requires it. The components must be shown to be compatible with each other over the shelf life.
What counts as a co-formulated product
Regulators distinguish between product types that are often grouped together loosely:
- Antibody cocktail (co-formulation). FDA's 1997 Points to Consider in the Manufacture and Testing of Monoclonal Antibody Products for Human Use defines a cocktail as two or more mAbs administered at a fixed ratio 83. FDA's bispecific guidance describes a cocktail as different antibodies mixed during manufacturing 85. FDA's 2023 draft potency-assay guidance notes that a fixed-ratio cocktail may be filled in one vial or in separate vials, but for that draft's purposes it limits the term to two or more mAbs filled in a single vial 304.
- Coadministered antibodies. These are separate products used together, each potentially on its own regimen, or combined only at administration 85.
- Bispecifics. These are single engineered molecules with two distinct binding domains, not mixtures 305. FDA's Bispecific Antibody Development Programs guidance (final, May 2021) states that it does not address antibody cocktails, polyclonal antibody products or combinations of mAbs 307584.
- Co-packaged products. In the EU, Ronapreve supplies its two antibodies in separate, co-packaged vials, one per antibody 5253. Nidlegy was proposed as a combination pack of two separately formulated vials (L19IL2 and L19TNF) to be mixed only in a syringe at administration; its marketing authorisation application was withdrawn after CHMP raised major objections 482483.
This distinction drives the CMC package. A true single-vial co-formulation needs analytical methods that work in the mixture. A co-pack needs per-vial controls plus in-use compatibility data for the admixture.
FDA expectations
The 1997 Points to Consider baseline
The 1997 Points to Consider document is still FDA's only general statement on mAb cocktails. It says the rationale for combining the products should be clear and grounded in the clinical context or prior experience with the individual products. It says lack of interference among the mAbs should be shown and synergistic or additive effects characterized. It also calls dose-ranging for each component "highly desirable", with dose and ratio potentially set from preclinical or clinical data showing that a particular dose and ratio is necessary or superior 83. The same document allows mAbs produced and controlled by similar procedures in the same facility to be documented, in some cases, in one Master File 395. It does not set cocktail-specific lot-release or stability requirements 58.
2023 draft: potency assays for antiviral mAbs and cocktails
The draft Potency Assay Considerations for Monoclonal Antibodies and Other Therapeutic Proteins Targeting Viral Pathogens (March 2023) is the most direct FDA statement on cocktail release testing 583. It states: "For mAb cocktails, release testing methods should include an identity method that demonstrates the presence of each individual mAb and a quantitative method verifying the ratio of the individual mAbs. The sponsor should ensure the ratio is consistent from lot to lot." 573
Its general potency recommendations also apply to cocktails:
- A binding assay is generally sufficient early in development. Later, more comprehensive methods that address the proposed mechanism(s) of action should be added to drug substance and drug product release testing and to stability protocols, and should be described, justified, qualified and validated to support a BLA 571.
- Lots should be compared against an appropriately qualified in-house reference material, with activity expressed as a percentage of that reference 572.
- Potency assays should include relevant positive and negative controls 604.
2021 draft: rabies mAb cocktails
The draft Rabies: Developing Monoclonal Antibody Cocktails for the Passive Immunization Component of Post-Exposure Prophylaxis (July 2021) is indication-specific, but it gives the most detailed FDA thinking on cocktail design and ratio control 582577:
- The mAbs may be combined at the formulated drug substance stage or during drug product manufacture 579.
- Each mAb's variable region should be screened for posttranslational-modification sites that could impair antigen binding. FDA recommends engineering out susceptible residues that reduce potency unless they are crucial to binding 580.
- Neutralizing activity should be evaluated for the cocktail, for each constituent mAb and for an HRIG comparator against a diverse strain panel 294. The clinical ratio should typically also be used in toxicology studies 581.
- Potency testing of the individual mAb drug substances and of the cocktail (as formulated drug substance or drug product) typically includes a binding ELISA and RFFIT 606. The RFFIT reference should be an international standard, or an in-house standard qualified against one, so that each mAb's potency can be expressed against the same standard 607.
- The ratio "may be based on mass or potency". Because individual mAbs may differ in potency, sponsors should justify the ratio and develop an assay that can demonstrate lot-to-lot consistency 288.
EMA expectations
EMA has no cocktail-specific quality guideline. The Guideline on development, production, characterisation and specification for monoclonal antibodies and related products (Revision 1, 2016) is the governing document 129. Its provisions map directly onto multi-antibody products, even though it does not name mixtures:
- Identity (section 7.1). Identity tests should be highly specific and based on unique structural or other properties. Because antibodies share similar constant domains, more than one physicochemical, biological or immunochemical test may be needed, and the tests should distinguish other antibodies manufactured in the same facility 196.
- Potency (section 7.3). A relevant potency assay belongs in the drug substance and/or drug product specification and should ideally reflect clinical activity. A justified binding assay may suffice where activity depends only on binding or neutralisation. Where effector functions matter, the assay should address them, and binding and cell-based assays are not necessarily interchangeable 197.
- Specifications (section 7). Specifications are product-specific and justified by characterisation, manufacturing consistency, stability and clinical data. Relevant product-related variants need individual and/or collective acceptance criteria, assessed with orthogonal methods 198199.
- Related products. The guideline, in its original version (effective 2009) and in Rev 1, extends its principles beyond intact mAbs (for example, to bispecifics) on a case-by-case basis. Neither version expressly addresses mixtures of separate antibodies 25.
For both agencies, ICH Q6B supplies the underlying logic. Drug product identity tests should be highly specific, and some products need more than one test 518519. Potency measures biological activity, not mass. Where an appropriate potency assay is used for the drug substance, an alternative method may suffice for the drug product if a rationale is given 520521522. Quantity is normally determined as protein content by mass 523. Q6B does not say whether a multi-protein product needs a separate identity, potency or quantity assay for each protein. That has to be justified product by product 519517.
Precedents from approved and assessed products
Approved products show how the general principles are applied in practice.
| Product | Composition | Key CMC findings |
|---|---|---|
| Phesgo (EU and US) | Pertuzumab and trastuzumab with rHuPH20 in a single SC vial. EU: 1,200/600 mg loading dose (2:1) and 600/600 mg maintenance dose (1:1) 535 | EU: Antibodies made as separate active substances, then thawed and compounded with rHuPH20 before filtration and aseptic filling. Formulation-robustness studies supported concentration and ratio ranges 536537538539. Identity by RP-UHPLC Lys-C peptide mapping. Each antibody's potency controlled by its own HER2-epitope-specific ELISA, because in the mixture one antibody masks the other's response in the anti-proliferation assay. Anti-proliferative and ADCC activity are controlled at active-substance level 540541. Shelf life of 18 months at 2 to 8°C. Recommendations covered polysorbate-20 degradation and fatty-acid particles, and a shipping-agitation study 542545546. US: OPQ recommended approval with 18-month dating 558563. Review noted degradation and visible particulates in clinical lots on long-term storage 563564. Leachables PMC required 565566 |
| Opdualag (EU and US) | Nivolumab and relatlimab in one vial 50, a 3:1 protein-mass fixed-dose combination 233234 | EU: Drug product process of thawing, mixing, filtration and aseptic filling with defined mixing speed, duration and hold times. Mixing studies supported the parameter ranges 398399. Specification covers identity, purity and potency, with in-house reference standards for both antibodies 400. Interaction studies between the two antibodies supported co-formulation. A pH effect on charge variants was noted 401. Shelf life of 36 months at 2 to 8°C 596597. US: PMC to reassess cell-based relatlimab potency release and stability criteria (drug substance and drug product) after 30 commercial lots 237238. Low endotoxin recovery led to interim rabbit pyrogen testing 239240 |
| Inmazeb (US) | Atoltivimab, maftivimab and odesivimab, 1:1:1, 241.7 mg each per 14.5 mL vial 214215 | Each drug substance produced individually in CHO cells 214. Drug product potency expressed relative to a combined three-antibody reference standard, using pseudovirus neutralization (atoltivimab, maftivimab) and ADCC reporter (atoltivimab, odesivimab) assays 216217. Component-specific blocking ELISAs showed each antibody stayed functional in the cocktail and that the antibodies did not compete for binding 218220. Protein quantity and identity listed as drug product quality attributes 592. Dating of 30 months at 2 to 8°C 593 |
| Rituxan Hycela (US) | Rituximab with recombinant human hyaluronidase, 120 mg/2,000 U per mL 267 | Separate drug substances, formulated and filled at the drug product site 267268. Concentration of each component treated as a drug product CQA at release and on stability 269. FDA issued an information request on identity testing under 21 CFR 610.14 270. Rituximab potency by CDC assay, plus a hyaluronidase reference standard for enzymatic activity 271272. 30-month dating. No CMC PMCs 276279 |
| Nidlegy (EU, counterexample; MAA withdrawn) | Two fusion proteins in separate vials, mixed in a syringe 482483 | CHMP found the finished product process and control strategy incomplete and kept a major objection on manufacture 487488. Control of identity, potency and purity was judged insufficient 489490491492. In-use mixture data did not establish compatibility. CHMP asked for a study with compatibility- and stability-indicating methods 493 |
Recurring control-strategy elements
Across the guidance and precedents, reviewers look for the following elements:
- Separate drug substances with a defined compounding step. Phesgo, Inmazeb and Rituxan Hycela each manufacture components as separate drug substances 536214267. Mixing parameters, hold times and mixing studies should be validated, as in the Opdualag EU assessment 398399. FDA's rabies draft allows combining at either the formulated drug substance or the drug product stage 579.
- Ratio justification and control. FDA's drafts expect a quantitative method verifying the ratio, with lot-to-lot consistency 573288. The ratio may be mass- or potency-based if justified 288. In Phesgo, CHMP accepted formulation-robustness data supporting the ratio ranges 539.
- Identity of each component in the mixture. FDA's 2023 draft expects an identity method showing each mAb is present 573. EMA's mAb guideline expects identity tests that can distinguish closely related antibodies 196. Phesgo used peptide mapping 540.
- Component-specific potency where the mixture defeats a combined assay. The Phesgo precedent shows that interference between components can justify epitope-specific ELISAs at drug product level, with functional assays kept at drug substance level 540541. Inmazeb shows the alternative, a combined reference standard with mechanism-matched assays 216217.
- Reference standards. Opdualag used in-house reference standards for each antibody 400. Inmazeb used a combined three-antibody standard 216. FDA's rabies draft favors a common international standard so each mAb's potency is reported on the same scale 607.
- Interaction and compatibility data. This expectation dates back to the 1997 Points to Consider requirement to show lack of interference 83. Recent assessments cite physical compatibility studies between the actives 401, antibody-excipient and antibody-rHuPH20 interaction studies 538, and functional non-competition data 220.
- Stability with attention to mixture-specific degradation. Phesgo's charge shifts, particle formation and polysorbate degradation 542545563, together with Opdualag's pH-dependent charge variants 401, show that degradation behavior can differ from the single-agent products.
- Specification lifecycle commitments. Limited commercial experience can lead to commitments to re-evaluate acceptance criteria, as with relatlimab potency in Opdualag 237238.
FDA and EMA compared
| Topic | FDA | EMA |
|---|---|---|
| Dedicated co-formulation guidance | None general. 1997 Points to Consider 83, plus two indication-specific drafts 582583 | None. mAb guideline Rev 1 applied case by case 12925 |
| Identity of each component | Explicit in the 2023 draft 573 | Implied by the requirement that identity tests distinguish antibodies 196 |
| Ratio control | Explicit quantitative method and lot-to-lot consistency 573288 | Assessed through formulation robustness and content testing in precedents 539 |
| Potency | Mechanism-relevant assays in release and stability 571. Individual and cocktail potency in the rabies draft 606 | Relevant potency assay in drug substance and/or drug product specification 197. Component-specific approach accepted for Phesgo 540 |
| Interference and compatibility | Lack of interference to be shown 83 | Interaction studies cited in Opdualag and Phesgo assessments 401538. Inadequate compatibility data raised in Nidlegy 493 |
Open questions sponsors usually take to the agencies
The public record leaves several points to product-specific discussion. These include numerical ratio acceptance ranges (not disclosed in the Phesgo or Opdualag assessments 539400), whether a combined or per-component reference standard is preferable, and how to qualify a combined potency assay when one component masks another. FDA's 2023 draft itself directs sponsors with novel mechanisms of action to consult the Agency on potency requirements for release and stability testing 605.