Extractables and leachables data in recent FDA drug and combination-product reviews

Extractables and leachables (E&L) has become one of the most consistently negotiated chemistry topics in FDA review of injectables, biologics, and drug-device combination products. The pattern across recent Drugs@FDA and BLA reviews is clear: FDA expects a risk-based extractables study of the container closure and device components, followed by leachables studies run on stability batches under actual use conditions, with every result above a defined analytical evaluation threshold identified, quantified, and toxicologically qualified. Where a submission falls short, the Agency either issues a postmarketing commitment or, more often, sends an information request or deficiency asking the sponsor to redo the work. This overview summarizes the framework FDA is applying and what it has actually asked for, product by product.

The framework FDA is reviewing against

FDA's expectations track a small set of documents that reviewers cite repeatedly. The centerpiece is the draft ICH Q3E Guideline for Extractables and Leachables, which lays out a risk-based program: an adequate extractables study using worst-case manufacturing and storage conditions and multiple complementary analytical techniques, followed, when warranted, by a leachables study representing actual manufacturing and intended storage/in-use conditions over shelf life ⁹⁰⁹¹. Q3E frames the decision logic around thresholds: an analytical evaluation threshold (AET) to decide what must be identified, quantified, and reported; a safety concern threshold (SCT) below which leachables present negligible mutagenic and non-mutagenic concern; and the threshold of toxicological concern (TTC) and qualification threshold (QT) used to derive those values ⁹⁰⁹³⁹⁵¹⁰⁶¹¹⁰. For products with local toxicity concerns, Q3E says the SCT should be the lowest of the mutagenic, non-mutagenic, and local-toxicity thresholds ¹⁰⁴.

Reviewers layer several other documents on top of Q3E depending on dosage form and delivery system:

Container Closure Systems for Packaging Human Drugs and Biologics, together with USP <1663> (extractables) and USP <1664> (leachables), is the reference set FDA points sponsors to for study design ³³⁵⁷.
For combination products, the CGMP guidance lists leachables and extractables testing among the design verification/validation activities expected for a prefilled syringe, alongside biocompatibility and drug/syringe compatibility studies ³⁰. FDA also advises sponsors to generate adequate characterization and safety-assessment data for extractable and leachable compounds from the container closure or device constituent, and to submit those data earlier in development (even during IND) when there are specific concerns ³³.
For therapeutic proteins, the immunogenicity guidance recommends a comprehensive E&L laboratory assessment using multiple analytical techniques, because leached materials can affect protein quality and immunogenicity, a concern the guidance flags specifically for syringes and vials ³¹.
For inhalation and nasal products, the Nasal Spray and Inhalation CMC guidance expects material selection that minimizes leachables, determination of recurring leachables through end of shelf life, and, where possible, correlation of leachables to extractables profiles so that routine stability testing for leachables may be reduced once a valid correlation is shown (extractables levels generally exceeding leachables levels) ³⁵³⁶³⁷.
For topical ophthalmics, FDA's quality guidance asks for extractables data, leachables data, and a separate toxicological risk assessment for the container closure system ⁹⁷⁹⁹.

What FDA asks for in parenteral and injectable reviews

For sterile injectables, the recurring request is a complete extractables profile of the primary container closure system, extraction studies on each packaging component with solvents stronger than the drug product, validated analytical methods, and leachables analysis on stability samples with at least 12 months of data and continued monitoring through expiry. FDA laid out exactly that scope for pemetrexed disodium injection (Sandoz, NDA 214657), extending the requirement to leachable/extractable data for all product-contact parts of the manufacturing equipment ⁸⁵.

Recent reviews show how the Agency handles gaps:

Glycopyrrolate injection (Fresenius Kabi, NDA 214919): FDA judged the leachables data suggestive of low risk but not sufficient to fully characterize the E&L profile of the container closure, and imposed a postmarketing commitment to run an extractables study with a defined AET and a leachables stability study at ICH Q1A(R2) timepoints ⁶⁵⁵⁴.
Brixadi (buprenorphine, Braeburn, NDA 210136): the most prescriptive package in the set. FDA required extractables studies using rigorous USP <1663> techniques, methods able to detect compounds at or above 1.5 mcg/day, leachables studies on at least three batches of the to-be-marketed weekly and monthly products at beginning, middle, and end of shelf life, testing of any secondary container closure under the same sterilization method, an AET able to detect any compound at 1.5 mcg/day (or an LOQ-based evaluation if not achievable), and a toxicological risk assessment qualifying leachables at or above 1.5 mcg/day for genotoxicity and 5 mcg/day for general toxicity ⁶⁷⁶⁹⁷⁴⁸³. FDA had found the original submissions lacked an adequate leachables evaluation for the proposed container closure ⁶⁷⁷⁴⁸³.
Fentanyl citrate injection (Exela, NDA 215870): extraction on primary container-closure and selected manufacturing components, evaluated for elemental, volatile, semi-volatile, and non-volatile compounds, using harsh conditions (pulverized components refluxed in 25 mM buffer at pH 1.8, 2, 6, and 12 for at least 8 hours), with leachables run at 6 and 12 months on six batches ⁹.
EXBLIFEP (cefepime/enmetazobactam, Allecra, NDA 216165): extractables on the vial, rubber stopper, and flip-off cap across aqueous buffer at pH 2.5 and 9.5, 50/50 isopropanol/water, and dichloromethane reflux, with nine organic compounds above the AET, and leachables on three batches stored upright or inverted for 24 months at 2 to 8°C and 25°C against an SCT of 5 mcg/day (1.67 mcg/vial) ⁷¹.
EPYSQLI (eculizumab-aagh, Samsung Bioepis, BLA 761340): extractables on a glass vial, rubber stopper, and flip-off cap in 5% IPA, basic water, 100% IPA, and 1% nitric acid to capture VOC, SVOC, NVOC, and trace metals, with above-AET extractables monitored as leachables; most detections were below 5 µg/day and trace metals below ICH Q3D limits ⁷⁸.

Thresholds: how FDA applies AET, SCT, QT, and TTC

The most technically contested part of these reviews is the derivation of the AET and the qualification thresholds. FDA has repeatedly rejected sponsor approaches it considers inappropriate and specified its own. A few patterns:

For parenterals, FDA reminds sponsors that the AET must be able to detect leachables at 5 mcg/day and higher, that the 1.5 mcg/day figure applies to potentially carcinogenic/genotoxic compounds under ICH M7, and that meeting the mutagenic threshold does not substitute for general toxicology qualification above the 5 mcg/day qualification threshold. FDA found the E/L assessment for Combogesic IV (acetaminophen/ibuprofen) inadequate on exactly this point and asked for toxicological risk assessments for all leachables above 5 mcg/day, using the maximum daily dose (4000 mg/day) and volume (400 mL/day, 4 vials) plus an uncertainty factor ¹⁵⁶.
For acute-use nasal products, thresholds shift. For RiVive (naloxone nasal spray) FDA accepted an AET of 25 mcg/mL derived from a 5 mcg/day qualification threshold and a 200 µL/day dose volume, noting a 120 mcg/day SCT for acute-use genotoxic/carcinogenic impurities and a 5 mcg/day QT for non-genotoxic impurities, with the lower value driving the AET ²¹⁴.
FDA has corrected the choice of framework itself. For Tofidence (tocilizumab-bavi, BLA 761354) the sponsor appeared to use an ICH Q3D(R1)-style elemental AET, which FDA said was inappropriate for organic leachables; the Agency required AETs based on a 1.5 µg/day SCT (calculating an AET of 0.0375 µg/mL) and individual risk assessments for compounds above the 5 µg/day QT ¹²²³. For Hikma acetaminophen injection, FDA rejected a PDE-based selection of targeted leachables that could miss compounds above 5 mcg/day and required an AET built on a 3 mcg/day SCT with a 50% uncertainty factor ³⁴⁷¹⁰.
Where sponsors qualified above-threshold leachables properly, FDA cleared them. Imbruvica (ibrutinib) used an AET of 0.25 µg/mL from a 1.5 µg/day SCT and a 6 mL/day dose, with nothing above the AET ⁸; Aponvie (aprepitant emulsion) used the QT as the SCT for acute exposure and found levels below it ¹³; and Dapzura RT (daptomycin), initially deficient because above-SCT leachables were not qualified for general toxicity, was resolved on resubmission with primary toxicology data, route-to-route extrapolation, surrogate compounds, NOAELs/LOAELs, and compound-specific PDEs ²⁵. For Zevtera (ceftobiprole), FDA had to derive reviewer PDEs when a full toxicological assessment for three leachables above 5 mcg/day was not provided ¹⁶.

The practical message: FDA expects the AET to be derived from an appropriate SCT for the route and dose, expressed so that the analytical method's LOQ sits at or below the AET, and it will not accept mutagenic-only qualification as a proxy for general toxicity.

Combination products and device constituents

For combination products, FDA reviews the primary container closure the same way it reviews a vial, but it adds scrutiny of device-specific components and materials. The division of labor is now routine: CDER addresses primary-contact leachables/extractables while CDRH handles skin-contacting biocompatibility, as spelled out for Idacio (adalimumab-aacf, BLA 761255), where the final device components excluding primary packaging were found not to contact drug product ⁴³⁴⁴. For Idacio, extractables covered the syringe barrel and plunger stopper of a 1 mL glass syringe with a 29G staked needle, under long-term (5 ± 3°C, 4/24/36 months) and stressed (40°C/2 weeks) storage with the syringe held horizontally to maximize stopper contact; six organic compounds were found for the drug-substance CCS and nine for the prefilled syringe, all below PQRI thresholds (5 mcg/day non-genotoxic, 1.5 mcg/day genotoxic) and below compound-specific PDEs ³⁹⁴⁰⁴².

Other combination-product reviews show what FDA looks for and where it pushes back:

Ebglyss (lebrikizumab-lbkz, Eli Lilly, BLA 761306): E&L assessed for both the drug-substance CCS and the drug-product prefilled-syringe-with-needle-safety-device or autoinjector, under exaggerated conditions using HS-GC-MS, GC-MS, HPLC-PDA-MS, and ICP-MS/ICP-OES, with the AET tied to an SCT and the maximum recommended human dose; FDA concluded there were no safety concerns for either presentation ⁸⁰⁸⁴.
Zurnai (nalmefene autoinjector, Purdue, NDA 218590): FDA found the extractables studies unconvincing, citing only three non-volatile extractables, unusual semi-volatile extractables from the glass barrel with needle, very few nonvolatiles in the rubber stoppers, and flawed sample prep because components were not cut into pieces ⁴⁸. It asked the sponsor to re-run extractables to capture non-volatile, semi-volatile, and volatile species and then repeat leachables at early, middle, and late timepoints across shelf life ⁶⁰⁷⁵⁸¹.
Brixadi (CAM2038 prefilled syringe): FDA said the applicant did not adequately address potential metals/elements from the container closure and that the extraction conditions were inadequate; an unidentified compound was tolerated only because later leachables testing did not detect it and the method assured less than 5 mcg/day ⁴¹⁴⁹.
Yuflyma (adalimumab-aaty, BLA 761219): FDA noted extractables testing was done only on the plunger stopper, with no extractables assessment for the glass syringe with needle or the staked needle with rigid needle shield; the Agency accepted this only because the leachables data were negative ⁴⁶.
Dawnzera (donidalorsen autoinjector, NDA 219407): for a staked-needle glass prefilled syringe with rubber plunger stopper assembled into an autoinjector, FDA requested finalized E&L study reports and a safety assessment covering all primary container closure components in contact during storage plus any secondary components with migration potential, per USP <1663> and <1664> ⁵¹⁵⁷.
Prevduo (glycopyrrolate/neostigmine, NDA 216903): E&L on a 3 mL prefilled syringe identified one leachable above the SCT, which was properly qualified ⁶³.
Enflonsia (clesrovimab-cfor, BLA 761432): the prefilled syringe barrel assembly and plunger stopper met ISO 10993 criteria, and three leachables above the PQRI 1.5 µg/day threshold were cleared through pediatric-specific risk assessments with acceptable margins of exposure ⁵⁵.
Tyenne (tocilizumab-aazg, BLA 761275) and Yesintek (ustekinumab-kfce, BLA 761406): FDA reviewed E&L for both PFS and autoinjector (Tyenne) and PFS and glass vial (Yesintek) presentations, with no toxicologically significant extractables or leachables identified ⁵⁶⁵⁹⁶²⁷⁹.

Across combination products, the draft Q3E even includes a dedicated leachables scenario for prefilled syringes, signaling that the Agency treats PFS leachables as a distinct concern ³². Device-specific materials that draw attention include rubber stoppers, needle shields, glass barrels and staked needles, adhesives, silicone oil, and tungsten residues, though not every review addresses each material explicitly.

Where submissions most often fall short

Reading the deficiencies together, a short list of recurring problems emerges:

Insufficient extraction rigor: components not pulverized or cut, too few extraction media or pH conditions, or conditions not exceeding worst-case use, leading FDA to request re-run extractables (Zurnai, Brixadi) ⁴⁸⁴⁹⁶⁰.
Incomplete container-closure coverage: testing only some components (for example, the plunger stopper but not the glass syringe or staked needle), which FDA accepts only when leachables come back negative (Yuflyma) ⁴⁶.
Wrong or insufficiently sensitive AET: using an elemental-impurity framework for organic leachables, or a PDE-based target list that could miss above-threshold compounds, prompting FDA to specify its own SCT-derived AET (Tofidence, Hikma) ³⁴¹²²³.
Missing general-toxicity qualification: relying on the 1.5 mcg/day mutagenic threshold without qualifying leachables above the 5 mcg/day general-toxicity threshold (Combogesic IV, Dapzura RT, Zevtera) ¹⁶¹⁶²⁵.
Leachables stability gaps: too few timepoints or missing early/mid/late sampling across shelf life, and incomplete assessment of secondary container closure and manufacturing-equipment contact surfaces (Brixadi, Zurnai, pemetrexed) ⁶⁷⁶⁰⁸⁵.

Practical takeaways for sponsors

The direction of travel is toward the draft ICH Q3E model, and recent reviews already read as if it were in force. A submission is far more likely to clear review when it presents a rigorous, worst-case extractables study of every product-contact component (primary and, where migration is plausible, secondary), a route- and dose-appropriate AET whose derivation is shown and whose analytical LOQ sits at or below it, leachables data on at least three registration batches at multiple stability timepoints through expiry, and a toxicological risk assessment that qualifies each above-threshold leachable for both mutagenic (1.5 mcg/day) and general-toxicity (typically 5 mcg/day for parenterals) endpoints, with compound-specific PDEs where needed ⁶⁷⁶⁹⁸⁵⁹⁰⁹¹. For combination products, the same package applies to the device's fluid path, coordinated with CDRH biocompatibility for patient-contacting surfaces, and should give explicit attention to stoppers, needles, needle shields, silicone oil, and tungsten ³⁰⁴³⁴⁴. Where a sponsor cannot fully close the E&L picture before approval, FDA's fallback has been a postmarketing commitment rather than a blanket acceptance, so the data will be expected either way ⁵⁴⁶⁵.