FDA CMC and Impurity Questions for Synthetic Peptide ANDAs Referencing rDNA-Origin Liraglutide and Teriparatide

FDA CMC and impurity questions for synthetic peptide drugs, including ANDAs referencing rDNA-origin liraglutide and teriparatide

Generic peptides such as liraglutide and teriparatide take an unusual path. The reference listed drugs (RLDs) were made by recombinant DNA (rDNA) technology, but most generic applicants make the peptide by chemical synthesis. FDA's review questions follow from that difference. Can the applicant show that a synthetic peptide is the same active ingredient as a recombinant one? Are new peptide-related impurities small enough, and well enough characterized, that they do not add immunogenicity risk? Do aggregation behavior and innate immune stimulation match the RLD? This article covers the guidance framework, the 2026 product-specific guidances (PSGs), and what FDA review records show about specific liraglutide, teriparatide and glucagon applications.

Key takeaways

FDA's final 2021 guidance lets an ANDA pathway be used for synthetic versions of five rDNA-origin listed peptides: glucagon, liraglutide, nesiritide, teriparatide and teduglutide. Its scope is synthetic versions; generic peptides made by rDNA methods fall outside this guidance ⁶⁷.
Under that guidance, each new specified peptide-related impurity should be characterized and should not exceed 0.5% of the drug substance. The applicant must justify why the impurity will not affect safety or effectiveness, including immunogenicity, and FDA may ask for a lower level ³³⁹³⁴⁰⁵³¹.
New impurities should not introduce a sequence with greater affinity for MHC (a potential T-cell epitope). The proposed product must also show no increase in aggregation propensity and no greater or different innate immune stimulation than the RLD ⁵³²³⁴¹.
The July 2026 draft PSGs for liraglutide and teriparatide set out a structured impurity scheme. Impurities above 0.1% are reported and those above 0.5% are identified. A shared impurity may not exceed the RLD level or 1.0%, whichever is greater. Applicants should work to keep new impurities below 1.0%, and total impurities may not exceed the RLD total ³¹⁸¹¹⁵.
In review records for the first synthetic liraglutide and teriparatide ANDAs, FDA's main deficiencies were active-ingredient sameness, impurity method adequacy, and aggregation comparability. FDA also asked for labeling that names the peptide as chemically synthesized ⁴⁴⁴⁴⁴⁶⁴⁹⁵⁴⁵³⁴⁹⁶.

The regulatory framework

The 2021 synthetic peptide ANDA guidance

FDA's final guidance, ANDAs for Certain Highly Purified Synthetic Peptide Drug Products That Refer to Listed Drugs of rDNA Origin (May 19, 2021), sets out when an ANDA may be appropriate for a synthetic peptide that references an rDNA-origin listed drug. It names glucagon, liraglutide, nesiritide, teriparatide and teduglutide. FDA says the guidance does not list every study an ANDA needs and points applicants to other guidance, including PSGs ⁶⁷.

The guidance's main CMC expectations are:

Drug substance description. Submit a full description of the drug substance: physical and chemical characteristics, stability, synthesis or isolation and purification methods, and specifications for identity, strength, quality and purity. Propose and justify impurity limits ³³⁸.
Shared impurities. A peptide-related impurity also found in the RLD should be present at the same or a lower level ³³⁹³⁴⁰.
New specified impurities. Characterize each one. None should exceed 0.5% of the drug substance, with a justification that it will not affect safety or effectiveness, including immunogenicity risk. FDA may ask for a lower level after review ³³⁹³⁴⁰⁵³¹.
T-cell epitope risk. For each new impurity, show that it contains no sequence with increased MHC binding affinity, and that impurity differences do not change physicochemical properties, biological activity or immunogenicity risk ⁵³².
Aggregation and innate immunity. Show no increase in aggregation propensity, and no change in the kind of aggregates formed, compared with the RLD, especially under stress. Show no impurities or contaminants that cause greater or different innate immune stimulation. FDA may recommend further nonclinical immunogenicity studies ³⁴¹.

These impurity and aggregation criteria apply to the specific generic synthetic peptides the guidance covers. They are not universal specifications for every synthetic peptide drug substance ³⁴⁴.

Active-ingredient sameness and immunogenicity guidance

FDA's draft guidance Sameness Evaluations in an ANDA: Active Ingredients (November 8, 2022) says peptide sameness can generally be shown through physicochemical characterization and biological evaluation. It recommends comparative testing against the RLD even when a compendial standard exists. Orthogonal methods should cover primary sequence, physicochemical properties, secondary structure, oligomer and aggregation states, and biological activity ⁸². The same draft notes that peptide-related impurities can change immunogenicity or otherwise affect safety and effectiveness. It also says the right application pathway depends largely on what data are needed to evaluate impurity differences ⁹.

FDA's draft Clinical Pharmacology Considerations for Peptide Drug Products (December 14, 2023) generally recommends an immunogenicity risk assessment for peptide products. The assessment covers product, process, patient and use factors, including impurities and aggregates. FDA encourages early discussion of how the assessment will shape any clinical anti-drug antibody evaluation ¹⁰.

The July 2026 draft PSGs for liraglutide and teriparatide

Liraglutide

The draft liraglutide PSGs (recommended July 2026) cover the 18 mg/3 mL (6 mg/mL) subcutaneous solution. They reference two RLDs, NDA 022341 (Victoza) and NDA 206321 (Saxenda). Their stated purpose is to support active-ingredient sameness, impurity assessment, and a request to waive in vivo bioequivalence (BE) studies ³¹³³⁰⁹³¹⁴.

Element	Liraglutide PSG recommendation
Sameness and characterization	Same primary sequence and physicochemical properties. Compare higher-order structure (secondary and tertiary structure and aggregates) with the RLD where applicable ³¹⁵.
Batch design	At least three generic batches at or near release and at the end of proposed shelf life. At least three RLD batches aged before expiry under worst-case label storage conditions ³¹⁵.
Biological assays	May be omitted with justification if the formulated peptide has no functional secondary or higher-order structure ³¹⁷.
Reporting and identification	Report each impurity above 0.1% of drug substance. Identify each impurity above 0.5% ³¹⁸.
Shared impurities	Not more than the RLD level or 1.0%, whichever is greater ³¹⁸.
New impurities	Efforts should be made to keep them below 1.0%. Total impurities above 0.1%, including new ones, should not exceed the RLD total ³¹⁸.
Out-of-range impurities	Do not automatically rule out an ANDA, but need a safety and efficacy justification and are subject to FDA review ³¹⁸.
Innate immune activity	Assess through aggregates and non-active-ingredient-related impurities. If those impurities meet compendial and toxicological limits and aggregation matches the RLD, in vitro innate immune testing may not be scientifically necessary ³⁴⁵³⁴⁶.
Recombinant products	Only if the generic is made by a recombinant process (not synthetic ANDAs): demonstrate and justify control of host-cell proteins and residual DNA ³⁴⁶.
BE waiver	Q1/Q2 the same as the RLD, with inactive ingredient concentrations within ±5% ³⁴⁷³¹³.

Teriparatide

The draft teriparatide PSG (recommended July 28, 2026) covers the subcutaneous solution and lists NDA 021318 (Forteo) and NDA 218771 as RLDs ¹¹¹¹¹⁰. Its sameness, batch-design and biological-assay recommendations match the liraglutide PSG ¹¹³¹¹⁴. So do its impurity scheme (report above 0.1%, identify above 0.5%, shared impurities at or below the RLD level or 1.0%, new impurities below 1.0%, totals at or below the RLD total) ¹¹⁵, its innate immune and host-cell impurity language ¹¹⁶¹¹⁷, and its Q1/Q2 waiver conditions ¹¹⁸¹¹⁹.

How the PSGs relate to the 2021 guidance

The two documents use different numbers. The 2021 guidance sets a 0.5% ceiling for each new specified peptide-related impurity ³³⁹³⁴⁰. The 2026 draft PSGs use 0.5% as an identification threshold and ask applicants to try to keep new impurities below 1.0% ³¹⁸¹¹⁵. The PSGs also suggest that in vitro innate immune testing may not be needed when impurity and aggregation profiles are comparable ³⁴⁵¹¹⁶. Sponsors with ANDAs in development should check which version of FDA's expectations governs their submission.

FDA issued other peptide PSGs on the same July 28, 2026 date, and they appeared in searches for peptide-impurity and immunogenicity language. They cover semaglutide, tirzepatide, glucagon, dasiglucagon, vosoritide, calcitonin salmon and pegcetacoplan ²⁹²²⁹³²⁹⁴²⁹⁵²⁹⁶²⁹⁷²⁹⁸²⁹⁹³⁰⁰³⁰¹³⁰²³⁰⁶.

What FDA asked in specific applications

Generic liraglutide: Hikma ANDA 215503

Hikma's ANDA 215503 was approved December 23, 2024, ahead of the other liraglutide ANDAs found in Drugs@FDA ¹²⁴. The review record shows FDA's main concern was whether the chemically synthesized liraglutide was characterized well enough and comparable to Victoza:

Impurity method. FDA found the drug substance impurities method unacceptable. Early reviews found too little evidence of active-ingredient sameness. As of February 6, 2024, FDA was still waiting for responses to deficiencies on API impurities and aggregation ⁴⁴⁴⁴⁴⁶⁴⁴⁷.
Pre-ANDA advice on comparability. FDA recommended side-by-side impurity profiles of the test product and the RLD, with test product samples near release and at the end of shelf life. It also recommended comparing aggregation under heat, agitation, light, freeze-thaw and humidity stress, using at least two orthogonal validated methods for each class of comparability study ⁴⁴⁵.
Applicant response. Hikma said it had submitted comparisons with age-matched RLD batches. These covered higher-order structure, oligomer and aggregation state, impurity profiles, biological activity and immunogenicity, and included three exhibit batches at the end of shelf life ⁴⁴⁹.
Effect on BE review. FDA found the formulation Q1/Q2 the same as the RLD. However, the API sameness deficiencies at first prevented a pharmaceutical equivalence finding and contributed to denial of the BE waiver at that stage. A later BE review found the application adequate from a BE standpoint, while noting that OPQ's sameness decision was still pending ⁴⁵⁰⁴⁵¹⁴⁵².
Labeling. FDA asked that the package insert describe the liraglutide as chemically synthesized and note its 97% homology to native GLP-1 (arginine substituted for lysine at position 34). Hikma revised the text, and FDA accepted it ⁴⁵³.
Resolution. The integrated assessment lists the immunogenicity and aggregation-state consults as adequate in 2023, and the mass spectrometry and impurity method verification consult as adequate on April 22, 2024 ⁴⁵⁴⁴⁵⁵.

Drug products cleared later include liraglutide ANDAs from Nanjing King-Friend (April 2, 2025), Lupin (July 22, 2025), Teva (August 27, 2025), Orbicular (January and February 2026), Biocon (February and March 2026), Fresenius Kabi (March 30, 2026), Mylan Institutional (June 15, 2026) and Sandoz (July 1, 2026) ¹²⁰¹²⁶¹²⁸¹²¹¹²²¹²³¹²⁷¹³¹¹²⁹¹³⁰. Public review text for most of these was not available in the records reviewed.

Generic teriparatide: Apotex ANDA 211097 and Teva ANDA 208569

FDA approved the Apotex ANDA 211097 and Teva ANDA 208569 teriparatide applications on November 16, 2023 ⁴⁵⁴⁷.

In the Apotex review, FDA noted that the RLD was made by rDNA technology while Apotex's product was chemically synthesized. FDA applied the synthetic peptide criteria: shared specified impurities no higher than in the RLD, new specified impurities no more than 0.5% with justification, and comparisons of primary sequence, secondary structure, oligomer and aggregation state, and biological activity ⁴⁹²⁴⁹³⁴⁹⁴. FDA classified the drug product deficiencies as major, partly because the drug substance sameness data were insufficient. The referenced drug substance DMF was reviewed on May 5, 2021 and found adequate, with minor deficiencies ⁴⁹⁵. The same major-deficiency assessment also cited insufficient drug/device compatibility data and failed accelerated device stability data. Real-time data were needed to support an extended expiry ⁴⁹⁹.

Apotex's file also raised a compendial issue. FDA asked for labeling that identified the teriparatide as chemically synthesized. It also asked Apotex to petition USP either to remove the monograph's reference to a recombinant source or to add chemical synthesis as a second source. Apotex reported filing the petition on July 7, 2019 ⁴⁹⁶⁴⁹⁷⁴⁹⁸.

For Teva's ANDA 208569, the record shows a complete response letter on May 13, 2022, followed by amendments. The deficiencies in that letter are not described in the available text ⁴¹¹. FDA later approved the ANDA and rated the product therapeutically equivalent to Forteo ⁴¹². Amphastar's teriparatide ANDA 213641 was approved December 12, 2025 ²⁶.

Teriparatide 505(b)(2) comparator: Bonsity (NDA 211939) and NDA 218771

Bonsity (PF708, NDA 211939, approved October 4, 2019) is not a synthetic peptide. Its drug substance is made with a recombinant Pseudomonas fluorescens strain ³⁵³⁶⁵³⁶⁶. Even so, its quality review shows what FDA expects in a peptide-to-RLD comparison:

Methionine oxidation. FDA noted a drug substance exception to the USP limit for total oxidation impurities (Met8, Met18 and Met8,18), which is NMT 0.5%. The drug product related-impurity limits met the USP injection monograph ³⁵⁵.
Impurity profiling against Forteo. Three Bonsity lots were compared with three Forteo lots by RP-HPLC, CEX-HPLC, SE-HPLC and SDS-PAGE, and purity was found highly similar ³⁵⁶.
Structure and function. The comparison used intact mass and peptide mapping (primary structure), far-UV circular dichroism (secondary structure), intrinsic fluorescence (tertiary structure), and receptor binding and cAMP bioassays (function) ³⁵⁷.
Aggregates. A teriparatide dimer was the main aggregate. Aggregates were measured by SE-HPLC under a revised specification. FDA noted that the RLD drug product specification did not include aggregate testing ³⁵⁸³⁶⁰³⁶¹³⁶².
Impurity qualification. A four-week rat bridging study compared the to-be-marketed formulation with Forteo for toxicity, toxicokinetics and immunogenicity ³⁶³.

NDA 218771 (Almaject, approved June 4, 2024) relied on NDA 211939 for its CMC information and added immunogenicity and analytical bridging data for a separate therapeutic equivalence evaluation ¹⁸³⁹⁵³⁹⁶³⁹⁷. Its drug substance specification controls total methionyl sulfoxides against the USP limit of NMT 0.5%. Teriparatide succinimide (30) and rhPTH (1-30) appear as USP-specified drug product impurities ³⁹⁸³⁹⁹⁴⁰⁰. Labeling reports anti-teriparatide antibodies in 2.2% of patients in each arm of a 24-week trial ⁴⁰²⁴⁰³. The one information request found in the review concerned the pen, not impurities: FDA asked for injection force and dose-button pull force comparisons against Forteo ⁴⁰⁴⁴⁰⁵.

Glucagon

Amphastar's glucagon ANDA 208086, which referenced Lilly's Glucagon for Injection, received a complete response letter on October 28, 2019 and was approved December 28, 2020 ⁵⁰⁷⁵⁰⁸. The available records do not describe the CMC questions in that letter. In a separate case, the February 9, 2024 complete response letter for Gvoke NDA 212097/S-012 found the bridge to the relied-upon glucagon product inadequate for IV use. FDA asked for comparative data on total impurities, glucagon degradants and aggregates, plus a justification that formulation differences, including DMSO and hypertonicity, would not affect in vivo performance ⁴⁵⁶⁴⁵⁷⁴⁶⁰.

Recurring themes for sponsors

Sameness comes first. In both the Hikma liraglutide and Apotex teriparatide reviews, insufficient sameness data was the lead deficiency, and it held up related decisions such as pharmaceutical equivalence and the BE waiver ⁴⁴⁶⁴⁵⁰⁴⁹⁵.
The impurity method must hold up on its own. FDA rejected an impurities method outright in the liraglutide review. It later relied on a separate mass spectrometry and impurity method verification consult ⁴⁴⁴⁴⁵⁴.
Compare aged RLD lots under stress. Pre-ANDA advice and the 2026 PSGs both call for end-of-shelf-life test batches, aged RLD lots, and stress conditions, measured with orthogonal methods ⁴⁴⁵³¹⁵¹¹³.
Immunogenicity is assessed at the impurity level. The 2021 guidance asks for T-cell epitope analysis of each new impurity and an innate immune comparison ⁵³²³⁴¹. The 2026 PSGs allow in vitro innate immune testing to be dropped only when impurity and aggregation profiles are comparable ³⁴⁵¹¹⁶.
Labeling and compendial status should state the synthetic source. FDA asked for "chemically synthesized" labeling for both liraglutide and teriparatide. For teriparatide, it also asked the applicant to address the USP monograph's recombinant-source wording ⁴⁵³⁴⁹⁶⁴⁹⁷.
Device and stability issues come with the peptide questions. Pen compatibility, device stability and actuation-force comparability were raised alongside the peptide CMC deficiencies ⁴⁹⁹⁴⁰⁴.

Open questions worth pursuing

The public review text found here does not name individual new impurities, give measured levels, or report in vitro T-cell or innate immune assay results for the synthetic liraglutide and teriparatide ANDAs. Several of these details are redacted or not described in the documents reviewed ⁴¹³⁴¹⁶⁵³⁴. Areas for follow-up include:

how later liraglutide ANDA reviews (Teva, Lupin, Biocon, Sandoz) handled the same questions once their review packages are posted
how the 2026 PSG impurity scheme will be applied to applications filed under the 2021 guidance's 0.5% ceiling
whether the same impurity and immunogenicity framework appears in the semaglutide, tirzepatide and glucagon PSGs