FROM THE EDITOR
Three warning letters hit the FDA enforcement database in the first week of October. Each one named a different type of contractor — an independent analytical testing lab, a nonclinical CRO, and a contract drug manufacturer. Each one carried the same underlying message: the GC’s company is legally responsible for what these contractors did, regardless of what the contract says.
That’s the non-delegable obligation doctrine, and it’s the thread running through this issue. FDA doesn’t pierce it for good contracts. It doesn’t pierce it for qualified vendors. The FD&C Act puts the obligation on the sponsor, and warning letters like these are the reminder that the obligation follows the sponsor wherever it outsources.
The October cluster landed close enough together to read as a signal worth studying. This issue walks through each letter, what the legal exposure looks like for sponsors who used these contractors, and what the contract testing agreements should have said — and still can say — going forward.
John LaRocca
THE OCTOBER ENFORCEMENT CLUSTER
FDA’s warning letter database posted three enforcement actions in the first week of October 2026, each targeting a different link in the pharmaceutical outsourcing chain. Read together, they describe a pattern biopharma GCs need to understand: FDA treats the sponsor’s product as adulterated when the contractor fails, whether the failure occurred at a testing lab, a nonclinical CRO, or a contract drug manufacturer.
Eurofins SF Analytical Laboratories (Horsham, Pennsylvania, FEI 3014704316)
received a CGMP warning letter dated October 6, 2026, citing violations of 21 CFR 211.194(a) — the requirement for complete laboratory records. FDA investigators found analysts recording microbial testing results incompletely: total colony counts recorded as “too numerous to count” without documenting actual plate counts; dried and cracked pour plates recorded as meeting specifications; and zero-CFU results entered without required second analyst verification, with FDA’s investigator finding a visible colony on a plate logged as 0 CFU. FDA cited the tested products as adulterated under FD&C Act Section 501(a)(2)(B).
Lovelace Biomedical Research Institute (Albuquerque, New Mexico)
received a GLP warning letter dated October 2026, citing violations of 21 CFR 58.35(b)(6) — the obligation of the Quality Assurance Unit to verify that final study reports accurately reflect raw data and protocols. Following an April 2025 inspection, FDA found two specific study protocols with systematic omissions: in Protocol 1, abnormal clinical observations were missing from the final report for 20 of 34 test animals; in Protocol 2, abnormal stool observations recorded during the study for all 48 animals were absent from the final report. Sponsors who relied on Lovelace nonclinical data for pending IND or NDA submissions face a material data integrity question.
Silgan Unicep Packaging, LLC (Spokane, Washington, FEI 3011913386)
— a Silgan Dispensing Systems subsidiary — received a CGMP warning letter dated October 5, 2026, following an inspection from March 26 to April 2, 2026. FDA found the firm averaging out-of-specification yeast and mold counts across lots to produce a passing microbiology result — a practice FDA said was inconsistent with USP <61> — and then releasing those lots into manufacturing. The violations cited were 21 CFR 211.192 (inadequate OOS investigation) and 21 CFR 211.22 (quality unit failures). A linked Class I recall for potential fungal contamination was already underway when the letter issued. FDA also cited the firm’s product labeling as containing disease claims without an approved NDA, triggering liability under FD&C Act sections 301(d) and 505(a) for marketing unapproved new drugs.
WHY THE ANALYTICAL LAB IS DIFFERENT FROM THE CDMO
The Eurofins SF letter requires its own legal analysis, because the sponsor’s exposure runs through a channel most GCs underestimate: FDA does not need to find a defective product to treat the sponsor’s goods as adulterated.
Under FD&C Act Section 501(a)(2)(B), a drug is adulterated if it is manufactured, processed, packed, or held under conditions that do not conform to CGMP. When Eurofins SF generated the testing data for a sponsor’s product, it was operating within the CGMP umbrella for that sponsor’s manufacturing process — analytical testing of a drug product is part of CGMP compliance, not separate from it. The plate-recording failures FDA cited are failures in that CGMP process. Every lot tested at Eurofins SF during the relevant period is presumptively adulterated, regardless of whether a single patient was harmed or a single product failed a specification.
That is the analytical lab’s unique legal relationship to the sponsor. A packaging CGMP failure at a contract manufacturer is serious, but the chain of liability runs through a contract that typically includes lot acceptance, certificate of analysis review, and identity testing by the sponsor. The analytical lab sits at the point where that verification chain ends — the sponsor relies on the lab’s data as the evidence of conformance. When the data is fabricated or incomplete, the conformance it was supposed to demonstrate never occurred.
THE SPONSOR’S LEGAL POSITION
FDA’s enforcement posture on contract testing is grounded in two principles that GCs should have in writing before the first lot is tested.
First, the sponsor’s CGMP obligation is non-delegable. The holder of the NDA, ANDA, or BLA is responsible for ensuring that every step in the manufacture of its product — including analytical testing performed by a contractor — conforms to CGMP. A quality agreement, however detailed, does not transfer that obligation. It creates private remedies between the parties; it does not limit FDA’s enforcement reach against the sponsor.
Second, the Responsible Corporate Officer doctrine — affirmed in United States v. Park — establishes that senior executives can be personally liable for CGMP violations within their organization’s sphere of authority, including those committed by contractors they directed or engaged. The RCO doctrine does not require the executive to have known about the violation. It requires only the authority to prevent it and the failure to do so.
For GCs, these two principles together answer the question of why the contract testing agreement matters beyond indemnification. The agreement is not just a risk-transfer document. It is the GC’s record that the sponsor exercised reasonable oversight — that it set standards, required access, demanded testing data in usable formats, and had documented remediation rights. Without that record, the sponsor’s position with FDA is that it trusted a contractor it couldn’t supervise.
THE GLP APPLICATION INTEGRITY QUESTION
The Lovelace letter creates a specific problem for sponsors with pending regulatory submissions that include Lovelace nonclinical data.
FDA’s GLP regulations require that studies used to support INDs, NDAs, and BLAs be conducted in compliance with 21 CFR Part 58. When FDA finds that a QAU failed to verify that final reports accurately reflect raw data — as it found at Lovelace — the evidentiary value of those studies becomes questionable. FDA has authority under 21 CFR 58.35 to disqualify a testing facility, which would render its submitted data unacceptable in pending applications without replication.
GCs with Lovelace studies in any pending submission should be asking three questions now:
1. Are any of our pending IND or NDA submissions supported by nonclinical studies conducted at Lovelace Biomedical between the April 2025 inspection date and the October 2026 warning letter?
2. Do those studies fall under the two protocols specifically cited in the warning letter — and if not, does the QAU failure FDA identified extend to other studies not yet enumerated in enforcement documents?
3. What is our documented basis for believing the data in those studies accurately reflects what actually happened to those animals?
The third question is the hard one. If the QAU wasn’t checking, the sponsor doesn’t have a record of independent verification — and rebuilding that record from raw data, assuming it exists and is recoverable, is a multi-month exercise that delays submissions on its own.
WHAT THE CONTRACT TESTING AGREEMENT SHOULD HAVE SAID
The October cluster is a useful checklist for auditing existing contract testing agreements. Seven provisions that should appear in any agreement with an analytical testing lab or nonclinical CRO:
Explicit CGMP/GLP compliance certification — Not just a representation, but a defined obligation with a specific right to audit on 48-hour notice.
Data integrity standards by name — Cite 21 CFR 211.194(a) for analytical labs; 21 CFR 58.35(b)(6) for GLP facilities. The specific CFR section should appear in the agreement so neither party treats it as implicit.
Unredacted access to raw data — Not just the certificate of analysis or the final study report, but underlying instrument records, analyst logs, second-verification signatures. The Eurofins SF violations were all in the underlying records, not the final results.
Mandatory disclosure of FDA observations — The contractor must notify the sponsor within 24 hours of receiving a Form 483 observation related to any sponsor product or study, and within 24 hours of receiving a warning letter.
Lot hold rights — The sponsor’s right to place a hold on any lot pending resolution of an FDA observation is a right that should exist before the warning letter arrives, not one you negotiate under pressure.
Survival provisions for regulatory remediation — The agreement’s data access, audit, and remediation provisions should survive termination for the period covered by any pending regulatory submission that used the contractor’s data.
Qualified Person bridge — For sponsors with EU operations, the agreement should expressly identify which party’s QP takes responsibility for certification of lots tested by the contractor, with a mechanism for QP notification on any data integrity finding.
THE EUROFINS NETWORK QUESTION
Eurofins SF Analytical Laboratories is one subsidiary in the Eurofins Scientific network — approximately 900 laboratories across more than 60 countries. The warning letter addressed to FEI 3014704316 in Horsham, Pennsylvania covers one facility. It does not extend on its face to Eurofins subsidiaries in other locations.
GCs whose contract testing portfolio includes other Eurofins facilities face a judgment call the letter itself doesn’t answer: whether the plate-recording practices FDA cited at Horsham reflect site-specific behavior or are symptomatic of a broader analytical culture. That judgment requires a conversation with the Eurofins quality team and, ideally, a documentation review at any Eurofins facility currently testing active sponsor products.
FDA’s enforcement action at one subsidiary does not trigger automatic liability for products tested at others. But it is notice. A GC who knew about the Eurofins SF letter and took no steps to review other Eurofins relationships has a harder time explaining the omission if a different Eurofins subsidiary generates a similar finding twelve months later.
WORKSTREAM 1 — IMMEDIATE
Map the Lot Exposure Before the Quality Team Does
The first legal question after a contract testing warning letter is not “what does our agreement say” — it is “which lots were tested there.” Quality teams will build that list. The GC should see it at the same time they do, not as a report afterward.
For Eurofins SF: identify every lot of your products tested at FEI 3014704316 in Horsham during the period FDA’s investigators reviewed. That list defines the population of potentially adulterated product under 501(a)(2)(B). Once you have it, the questions become: were any of those lots distributed? Are any still in the supply chain? Was any of the testing data submitted to FDA in a pending application?
For Lovelace: identify every nonclinical study conducted at Lovelace that has been submitted to FDA in a pending IND, NDA, BLA, or sNDA. Cross-reference against the specific protocols cited in the warning letter, then make a conservative judgment about whether studies outside those protocols were conducted under the same QAU regime.
For Silgan Unicep: if any of your products use Silgan Unicep as a contract drug manufacturer, identify which lots were manufactured during the March–April 2026 inspection period and determine whether any are subject to the Class I recall already underway.
BUSINESS IMPACT: The exposure map determines the scope of every decision that follows — disclosure to FDA, voluntary recall, submission hold, or none of the above. Building it slowly is the costliest possible response.
WORKSTREAM 2 — NEAR-TERM
Audit the Contract Testing Agreement Against Seven Provisions
Once you know which contractors you use and which products are affected, pull every contract testing agreement for analytical labs and nonclinical CROs. Run each one against the seven provisions described above.
The audit will likely find three categories of agreements: those that have most of the provisions and need targeted updates; those that have a CGMP/GLP representation but nothing about data access, audit rights, or FDA observation notification; and those that are effectively just purchase orders. The second and third categories are the legal exposure — relationships where FDA could argue the company didn’t exercise the oversight the non-delegable obligation requires.
The Eurofins SF and Lovelace letters are also leverage. Any analytical lab or CRO that pushes back on data access rights or audit provisions in the next six months is taking a position that will be harder to defend with an active FDA warning letter in the same industry vertical.
BUSINESS IMPACT: A contract that doesn’t give you real-time access to raw data is a contract where you learn about the problem from FDA, not from the contractor. Getting to audit rights now, before the next warning letter, is cheaper than getting to them in response to one.
WORKSTREAM 3 — STRATEGIC
Build a Contract Testing Lab Oversight Framework
The October cluster is the case for a formal framework that doesn’t currently exist at most biopharma companies: tiered oversight of contract testing facilities calibrated to the regulatory risk of the product and the criticality of the testing.
Tier 1 — Critical Path Testing: Analytical labs and CROs generating data for pending NDA or BLA submissions, or testing product lots currently in distribution. Annual on-site audits, 48-hour FDA observation notification obligations, unredacted raw data access, and a GC review of the contract testing agreement annually.
Tier 2 — Supporting Testing: Labs generating stability, compatibility, or development testing for programs not yet in late-stage submission. Biennial on-site audits, FDA observation notification obligations, certificate of analysis review plus raw data access on request, and a GC review of the agreement on any scope change.
Tier 3 — Commodity Testing: Low-criticality testing where results don’t directly support regulatory submissions. Qualification documentation on file, FDA observation notification obligations, and a spot-check audit on any change of ownership or change of site.
The framework doesn’t require new contracts for all three tiers at once. It requires a documentation standard — knowing which tier every testing relationship sits in, and having a record of the oversight commensurate with that tier.
BUSINESS IMPACT: A tiered oversight framework gives the GC documented evidence that the company exercised oversight proportionate to risk. That evidence is the difference between a sponsor who acted reasonably in the face of a contractor failure and a sponsor who deferred to a contractor it didn’t supervise.
FDA’s treatment of Eurofins SF-tested products as adulterated under 501(a)(2)(B) does not require a defective drug — it requires a defective process. The data integrity failures are the CGMP failure. The adulteration follows as a matter of law.
For interested readers, I have created a Contract Testing Lab Oversight Toolkit — six practical tools a GC can use immediately. It can be found at https://gumroad.com/products.
ABOUT THE AUTHOR
John LaRocca is the founder of LaRocca Advisory Services and the author of The GC Playbook. He served as CLO at Aerie Pharmaceuticals (Aerie/Alcon, ~$930M), Y-mAbs Therapeutics (SERB, $412M), and Eagle Pharmaceuticals, with 37 years of legal experience and 26 years at the GC/CLO level in specialty pharma and biotech. The GC Playbook is published on Substack at laroccaj.substack.com.


