In-process bioburden control is a foundational element of biologics manufacturing, directly informing sterility assurance strategy and the acceptability of a drug substance or drug product at critical manufacturing steps. For regulatory and CMC teams preparing or responding to a Biologics License Application, understanding how FDA microbiology reviewers evaluate bioburden acceptance criteria—including alert and action levels, hold-time limits, and pre-filtration specifications—is essential for anticipating Information Requests and structuring robust CMC packages.
This analysis examines FDA CBER and CDER review documents for BLAs approved from January 2021 onward, drawing on publicly available microbiology and CMC review memoranda from Drugs@FDA and the CBER BLA review set. It characterizes the elements FDA has scrutinized, the language reviewers use when raising deficiencies, and the types of data and justifications that have satisfied agency reviewers across roughly two dozen BLAs in the review window.
Want Rhizome's help on your own question? Try it for free.
FDA assessment of in-process bioburden limits in BLA reviews (2021 to 2026)
What "in-process bioburden limits" means in a BLA review
In a BLA review, in-process bioburden limits are the microbiological acceptance criteria applied to in-process material (intermediate bulk, drug substance in-process bulk, drug product pre-filtration pools, and hold-step samples) to demonstrate that the manufacturing process keeps microbial load under control before sterile filtration and aseptic fill. FDA microbiology reviewers assess these limits as part of the overall contamination-control strategy, alongside sterile filtration validation, aseptic process simulation (media fills), environmental monitoring, endotoxin control, and hold-time studies 13813932.
The recurring FDA review framework
Across BLA reviews in this period, FDA microbiology reviewers evaluated in-process bioburden control against a consistent set of expectations. When any element was missing, it generated an Information Request.
1. Pre-established limits at all critical manufacturing steps. FDA repeatedly asked sponsors to provide pre-established bioburden (and endotoxin) limits at critical steps rather than reporting results without predefined acceptance criteria. This request recurred for COLUMVI (glofitamab-gxbm, BLA 761309) 6, VABYSMO (faricimab-svoa, BLA 761235) 23, NEXVIAZYME (avalglucosidase alfa-ngpt, BLA 761194) 15, LEQEMBI (lecanemab-irmb, BLA 761269) 1622, IDACIO (adalimumab-aacf, BLA 761255) 14, BKEMV (eculizumab-aeeb, BLA 761333) 17, and EPYSQLI (eculizumab-aagh, BLA 761340) 20.
2. Defined sampling points, with bioburden sampled before 0.2 µm filtration. FDA specified that bioburden sampling should occur prior to any 0.2 µm (sterilizing-grade) filtration step, so the pre-filtration microbial load challenging the sterilizing filter is actually characterized. This appeared explicitly for PYZCHIVA (ustekinumab-ttwe, BLA 761373) 1018, IDACIO 14, YUSIMRY (adalimumab-aqvh, BLA 761216) 7, NEXVIAZYME 15, and as "sampling points and in-process limits for bioburden and endotoxin" for POMBILITI (cipaglucosidase alfa-atga, BLA 761204) 5 and SIMLANDI (adalimumab-ryvk, BLA 761299) 25.
3. A pre-sterile-filtration bioburden ceiling of NMT 10 CFU/100 mL. Where FDA stated a numeric expectation, the recurring criterion was that pre-sterile filtration bioburden should not exceed 10 CFU/100 mL. FDA applied this to POMBILITI 5, PYZCHIVA 1018, and SIMLANDI 25. This reflects the standard expectation that the microbial challenge to the sterilizing filter is limited and validated.
4. Method qualification, validation, and suitability supporting the limit. FDA assessed whether the bioburden test method was validated and suitable for the specific in-process matrix, including recovery/suitability studies showing no product inhibition of microbial growth. For HEMGENIX (etranacogene dezaparvovec-drlb, BLA 125772), FDA reviewed the bioburden method as a lot-release method and confirmed method suitability was performed and compliant, with no product inhibition on microorganism growth, concluding the method was appropriate under actual conditions of use 78114. Where qualification data were absent, FDA issued deficiencies (see NEXOBRID below) 663.
5. Hold-time studies and endotoxin paired with bioburden. FDA evaluated whether hold times at in-process steps were supported by data and whether endotoxin was controlled alongside bioburden. FDA requested hold-time and method-qualification data supporting the limits for COLUMVI 6, LEQEMBI 1622, VABYSMO 23, and YUSIMRY 7. For CARVYKTI (ciltacabtagene autoleucel, BLA 125746), FDA reviewed an endotoxin hold-time study in which drug product spiked with known endotoxin was stored at -120°C to evaluate masking, confirming a maximum hold time with no impact on endotoxin content 35. For BEQVEZ (fidanacogene elaparvovec-dzkt, BLA 125786), FDA identified bioburden, sterile filtration, and aseptic filling as critical steps with control ranges tabulated, and noted in-process controls including bioburden limits, hold times, and minimum total particulates as acceptable 30313334.
6. Microbial control of intermediates and in-process bulk, not just the final filter. FDA expected microbial limits on intermediate and in-process bulk to reflect the microbial quality of the process as a whole. This was the central issue in the NEXOBRID review (see below) 6521.
Product-specific examples (BLAs approved from 2021)
| Product (BLA) | What FDA reviewed on in-process bioburden | Citation |
|---|---|---|
| NEXOBRID, anacaulase-bcdb (761192) | Missing microbial limits for BSP intermediate bulk; inadequate bioburden limits for DS in-process bulk; missing method validation/qualification and the <10 cfu/10 mL detection limit and criterion for the gel-vehicle carbomer; bioburden method suitability for DS did not meet acceptance criteria | 6521663 |
| POMBILITI, cipaglucosidase alfa-atga (761204) | Sampling points and in-process limits for bioburden and endotoxin; pre-sterile filtration bioburden not to exceed 10 CFU/100 mL | 5 |
| COLUMVI, glofitamab-gxbm (761309) | Monitoring of bioburden and endotoxin at critical steps; pre-established limits plus hold-time and method-qualification data | 6 |
| PYZCHIVA, ustekinumab-ttwe (761373) | Bioburden sampling prior to any 0.2 µm filtration; pre-established bioburden and endotoxin limits; pre-sterile filtration bioburden NMT 10 CFU/100 mL | 1018 |
| IDACIO, adalimumab-aacf (761255) | Pre-established bioburden and endotoxin limits at critical steps; sampling prior to 0.2 µm filtration | 14 |
| VABYSMO, faricimab-svoa (761235) | Pre-established bioburden and endotoxin limits plus supporting PPQ, hold-time, and test-method qualification data | 23 |
| NEXVIAZYME, avalglucosidase alfa-ngpt (761194) | Pre-established bioburden and endotoxin limits, including sampling before 0.2 µm filtration | 15 |
| LEQEMBI, lecanemab-irmb (761269) | Pre-established bioburden and endotoxin limits plus in-process method qualification and hold-time data | 1622 |
| YUSIMRY, adalimumab-aqvh (761216) | Pre-established limits at critical steps, including before 0.2 µm filtration; PPQ lot, hold-time, and method-qualification data | 7 |
| BKEMV, eculizumab-aeeb (761333) / EPYSQLI, eculizumab-aagh (761340) | Bioburden and endotoxin monitoring at critical steps with pre-established limits | 1720 |
| SIMLANDI, adalimumab-ryvk (761299) | Sampling points and in-process limits; pre-sterile filtration bioburden NMT 10 CFU/100 mL | 25 |
| BEQVEZ, fidanacogene elaparvovec-dzkt (125786) | Bioburden, sterile filtration, aseptic filling identified as critical steps; control ranges tabulated; bioburden limits, hold times, minimum total particulates found acceptable | 30313334 |
| CARVYKTI, ciltacabtagene autoleucel (125746) | Endotoxin hold-time/masking study at -120°C; sample hold times and maximum hold time before testing established; contamination control via in-process and release testing | 3546 |
| HEMGENIX, etranacogene dezaparvovec-drlb (125772) | Bioburden method suitability performed and compliant, no product inhibition, method appropriate for use; supplier bioburden/endotoxin testing of components | 7811453 |
| ADSTILADRIN, nadofaragene firadenovec-vncg (125700) | Sufficient in-process controls at critical DP stages, including bioburden limits, hold times, minimum total particulates, judged acceptable | 3334 |
| MNEXSPIKE, COVID-19 vaccine mRNA (125835) | In-process hold conditions established; microbial controls during sterile filtration tabulated; media fill used to assess maximum hold times with no contaminated containers and no microbial excursions during PPQ | 2952 |
| REGENECYTE, HPC Cord Blood (STN 125764) | Media simulation initially insufficient (aliquots only); FDA required full-volume incubation to show absence of contamination across the process; resolved with full media fill and environmental monitoring | 36375051 |
| PENMENVY, meningococcal ABCWY vaccine (125819) | Bioburden assessed throughout the process and DP process as part of contamination-control strategy | 138139 |
| BREYANZI, lisocabtagene maraleucel (125714) | Microbial contamination controls (cleanroom, gowning, environmental monitoring, line clearance) and aseptic process validation meeting acceptance criteria | 3239 |
Common deficiencies FDA cited
The most frequent shortcomings that drew FDA Information Requests were: (1) no pre-established bioburden and endotoxin limits at critical steps 623151614; (2) bioburden sampled after, rather than before, the sterilizing filter 10147; (3) missing method validation or qualification for the in-process bioburden method, including failure to demonstrate the stated detection limit and acceptance criterion 66365; (4) bioburden method suitability failing acceptance criteria and needing to be repeated 12; and (5) microbial limits on intermediates or in-process bulk that did not reflect the microbial quality of the process 6521. For cell and gene therapy and cord-blood products, FDA additionally scrutinized whether aseptic process simulations captured the full process, requiring full-volume media incubation in at least one case 36375051.
Bottom line
For BLA approvals from 2021 to 2026, FDA's assessment of in-process bioburden limits was consistent and centered on five things: pre-established numeric acceptance criteria at every critical step; correctly located sampling (bioburden taken before 0.2 µm sterilizing filtration, commonly held to NMT 10 CFU/100 mL); a qualified and suitable test method proven free of product inhibition; hold-time data supporting the limits; and endotoxin controlled in parallel with bioburden 51025786. Numeric limits are often redacted in the public reviews, so the granular acceptance values usually are not visible even where FDA confirms it reviewed and accepted them.
Suggested follow-ups to ask Rhizome
- Pull the specific redaction-permitting language and any disclosed numeric bioburden limits for a single product's CMC review.
- Compare CBER (cell/gene therapy) versus CDER (mAb/biosimilar) expectations for pre-filtration bioburden sampling.
- Map how these BLA-review expectations align with FDA sterile-drug and aseptic-processing guidance and USP microbial enumeration chapters.