Annual report, CBE-30, or prior approval supplement: how sponsors classify post-approval manufacturing changes, and when FDA pushes back

Every marketed drug and biologic keeps changing after approval. Suppliers get swapped, sites get added, processes get scaled, specifications get tightened. US law (21 CFR 314.70 for drugs, 601.12 for biologics) forces the applicant to sort each of those changes into one of three reporting buckets before implementing it: a prior approval supplement (PAS) that must be approved before distribution, a changes being effected supplement (CBE-30 or CBE-0) that can be implemented on a shorter clock, or a routine annual report notification. The bucket a sponsor picks determines how long the change is held up and how much data has to travel with it, so the classification decision is itself a regulatory judgment that FDA reviewers audit and, when they disagree, override.

This article walks through how FDA frames the choice, what real submissions have actually reported in each category, and the recurring situations where the Agency told a sponsor it had picked the wrong one.

The risk-based logic FDA applies

FDA's operating rule is proportionality: the more likely a change is to affect product quality, safety, or effectiveness, the higher the reporting category ²⁷. That maps to the statutory tiers:

Minor changes (low risk) go in the annual report ²⁷.
Moderate changes (some potential impact) go in a CBE-30, or in some cases a CBE-0 ²⁷.
Major changes (substantial potential for adverse effect) require a PAS ²⁷.
Depending on the circumstances, a change may even be better handled in an original application rather than as a supplement at all ²⁷.

The same principle is applied at the level of individual change types. For a change in glass vial supplier with no change in glass type, coating, or container/closure dimensions, FDA accepts an annual report ²⁷. But for an alternate stopper supplier, if there are differences in stopper properties or in the manufacturing process between the approved and proposed supplier, FDA says the increased risk warrants a higher category (PAS); only where those differences are minimal can a lower category be considered ²⁷. That single example captures the whole logic: the reporting category tracks the delta between what was approved and what is proposed, not the label on the component.

The core reference documents a reviewer will cite are the umbrella guidance "Changes to an Approved NDA or ANDA" ¹²⁸¹²⁹, the companion "CMC Postapproval Manufacturing Changes To Be Documented in Annual Reports" ³⁴, the SUPAC family (for example SUPAC-IR and its Q&A) ³³³⁵, and change-specific guidances such as "Container Closure System and Component Changes: Glass Vials and Stoppers" ²⁷.

What actually goes in an annual report

The annual report is for changes with minimal ability to disturb the approved product, plus post-approval follow-up data. Chemistry reviews show sponsors using it for:

First-batch and later-interval stability after a process change. In the Plendil (felodipine, AstraZeneca) supplement, a mixing process change was treated as a Level 2 change requiring a CBE supplement, but the long-term stability data on the first batch was committed to the annual report ³². The same review classified a move to alternate coating equipment of the same design and operating principle as a Level 1 change ³².
Ongoing analytical investigations. For CellCept (mycophenolate mofetil, Roche), the sponsor committed to submit melting points from both analytical facilities for all lots produced over one year in the annual report while it investigated a site-to-site discrepancy, and to report analytical results for recovered solvents there as well ²⁸.
Accumulating stability data. Upneeq (oxymetazoline, RVL) proposed submitting additional later-interval stability data in the next annual report ³⁰³¹, and Sustol (granisetron, Heron) reviewers accepted additional drug-release and stability data post-approval in the annual report ²⁹.

The SUPAC guidances add the classic "minor" manufacturing changes that belong in an annual report: moves of secondary packaging, labeling, ink imprinting on solid oral dosage forms, or manufacture of drug-substance intermediates other than the final intermediate to a site with satisfactory CGMP status ³⁴¹²⁸¹²⁹; relocation within the same facility with no scale-up, process, equipment, or component change ³⁴; changes in non-release-controlling excipients within specified ranges; a change in excipient supplier at the same technical grade and specification; and changes in release-controlling excipients up to 5% of the total in a modified-release product ³⁴.

What goes in a CBE-30 (and when a CBE-0 fits)

The CBE-30 is the moderate-change workhorse: file it, wait 30 days, then implement (unless FDA objects). Real examples from chemistry reviews:

Adding a drug-substance manufacturing site. Horizant (gabapentin enacarbil, Azurity) proposed a CBE-30 to add a new drug-substance manufacturing site ⁶⁰.
Adding a packager. For Actos (pioglitazone, Takeda), adding a drug-product packager was treated as a moderate change suitable for a CBE-30 ⁷¹.
Container-closure changes with supporting protocols. Exparel (bupivacaine, Pacira) proposed CBE-30 changes to the container closure system plus release/stability specifications ⁴⁹, and Evamist (estradiol, Padagis) submitted a comparability protocol and proposed a CBE-30 for a planned drug-product container change ⁶⁴.
Supplier/site changes conditioned on data. For Symbicort (budesonide/formoterol, AstraZeneca), FDA agreed that if data on a new actuator supplier/site were available by approval, they could come in as a CBE-30 ⁶⁵. Zaltrap (ziv-aflibercept, Sanofi) carried post-marketing commitments to submit container-closure integrity and dye-ingress study data as CBE-30 supplements (with a shipping-qualification study committed as a CBE-0) ⁵⁰.

The CBE-0 (implement on submission) shows up for lower-risk or already-vetted changes. Tri-Sprintec (Barr) used a CBE-0 to bring alternate packaging sites and an alternate norgestimate manufacturing site into effect after final approval, having earlier run the same changes through the 30-day CBE-30 clock ¹¹⁸¹¹⁹. Benlysta (belimumab, Human Genome Sciences) was directed to submit capper qualification and container-closure-integrity (helium leak) validation in a CBE-0 ¹²⁵. A recurring chemistry-review rule for dissolution specifications illustrates the CBE-0/PAS line: use a CBE-0 where final specifications are tighter than interim specs (or unchanged), but a prior approval supplement otherwise ¹²²¹²⁶.

What requires a prior approval supplement

PAS territory is major change: anything with substantial potential to affect identity, strength, quality, purity, or potency, sterility assurance, or modified-release performance ¹²⁸¹³¹. Drugs@FDA reviews are full of them:

New or alternate drug-substance manufacturer / API supplier. Fludarabine Phosphate Injection (Areva) ⁶, Torsemide Tablets (Apotex) ²¹, and Chlorhexidine Gluconate Cloth 2% (Sage) ²⁰²⁴ each filed a PAS to add an alternate drug-substance manufacturer; Procanbid (King Pharmaceuticals) combined a new manufacturing/testing facility with a change in active-substance supplier ⁷¹⁰¹⁶.
New drug-product manufacturing site. Herceptin (trastuzumab, Genentech) filed a PAS to add the Penzberg, Germany facility as an additional trastuzumab drug-substance site ²³; Cardizem CD (diltiazem, Valeant) filed a PAS for an alternate drug-product manufacturer described as a new formulation and process change ¹⁸.
Formulation and packaging-configuration changes. Hectorol (doxercalciferol, Sanofi) filed a PAS for a change in formulation and packaging configuration plus a new site ¹; Ibuprofen Tablets 200 mg (LNK) filed for a change in excipients in the formulation ²³.
Container-closure / delivery changes affecting sterility or process. Restasis / Restasis Multidose (cyclosporine, AbbVie) filed a PAS for a new multi-dose applicator with associated sterilization/process validation and stability data ⁵; BiCNU (carmustine) added an alternate diluent manufacturing/testing site along with container, process, expiry, storage, and specification changes ⁴⁸.
Combined post-approval packages. Vitrakvi (larotrectinib, Bayer) bundled a drug-product site change, an API supplier change, and a new blue imprint ²⁵.

Where FDA disagreed with the sponsor's chosen category

This is the part that matters most for risk planning: the classification is the sponsor's to propose, but FDA's to confirm, and reviewers do reclassify.

Sponsor called it "changes being effected"; FDA called it major. The clearest case is Ibuprofen Tablets 200 mg (LNK International). The sponsor filed the excipient change as a "Supplement - Changes Being Effected." FDA rejected that category outright: "Rather, this change is a major manufacturing change that requires approval of a supplement before a product made with the change can be distributed... Please do not implement the proposed changes until you receive notification that the supplemental application is approved" ². The change was forced up to a PAS.
Site transfer cannot ride along in the NDA. For Exparel (Pacira), FDA held that "since a site-transfer for product is considered a major change, qualification of this alternate site without a comparability protocol, or comparability data in the NDA is not acceptable. This site may be submitted post-approval in a post-approval supplement" ⁹¹. The Agency both fixed the category (major) and dictated the vehicle (post-approval PAS or a comparability protocol).
A new strength at a new site needs a supplement, not the current pathway. In the Ditropan XL (oxybutynin, Janssen/Alza) review, FDA stated the sponsor should be informed that to manufacture the 15 mg strength at the Mountain View site, a supplement would need to be submitted, per SUPAC-MR ¹¹³.
A tradename change is a major change. For Clindagel (clindamycin phosphate, Bausch), FDA wrote that the reporting category for a tradename "qualifies as a major change requiring a prior approval supplement per 21 CFR 314.70(b)(3)" and that a labeling supplement would be required ⁷⁶.
Changes outside the approved design space are not annual-report matters. For Votrient (pazopanib, Novartis), FDA stated that changes outside the approved design space "should be implemented according to 21 CFR 314.70," i.e., through the appropriate supplement rather than as a routine notification ⁸⁹.

The enforcement side of the same principle appears in warning letters, where firms are cited for implementing changes without the correct supplement:

Biotest Pharmaceuticals was cited for not reporting a change to use multiple filters in the bulk filtration step for Bivigam, a change that "required approval of a supplement prior to distribution" under 21 CFR 601.12(b), and for failing to report changes to maximum process times (centrifugation and UF/DF) under the BLA change-reporting rules ⁷⁸.
Sanofi Pasteur was told that a future change such as a reduction in stability testing should be filed as a prior approval supplement under 21 CFR 601.12(b) because it could adversely affect potency ⁸¹.

The pattern across these examples is consistent. FDA disagreements almost always run in one direction, upgrading the category, and cluster around a few triggers: site transfers of finished product, drug-substance/API sourcing, excipient/formulation changes, sterility-relevant process changes, and anything that moves outside an approved design space. A sponsor that defaults to CBE or annual report for any of those should expect a challenge.

Tools that let sponsors lock in (or lower) a category

FDA guidance also gives applicants ways to pre-negotiate the category rather than argue it after the fact:

Comparability protocols (CPs). Under "Comparability Protocols for Postapproval Changes to the CMC Information in an NDA, ANDA, or BLA," an approved CP can specify that later, in-scope modifications be filed in a CBE supplement or annual report instead of a PAS, effectively pre-authorizing a reduced category for those changes ⁴⁴. In Exparel, FDA itself pointed to a comparability protocol as the acceptable route for the site transfer ⁹¹.
Established conditions and ICH Q12. A product lifecycle management document can define established conditions and propose reporting categories for future changes to them; if the applicant does not propose a category, regional guidance applies, and the justification belongs in the relevant ICH Q12 subsection ⁴³. FDA notes that successful comparability demonstrations can be used to leverage tiered reporting categories and commitments, including for stability-related changes ⁴².

Practical takeaways for RA teams

The reporting category is a data-supported risk argument, not an administrative label. Three points fall out of the record above:

Classify on the delta, not the component. The same nominal change (a stopper supplier, a site, an excipient) can be an annual report or a PAS depending on how much the proposed condition differs from the approved one ²⁷. Build the category around the difference and the supporting data.
Assume upgrade risk for site, source, formulation, sterility, and design-space changes. These are exactly where FDA has reclassified sponsors upward or cited firms for skipping approval ²⁷⁸⁸¹⁸⁹⁹¹¹¹³. When in doubt on one of these, a PAS or a pre-approved comparability protocol is the defensible choice.
Use CPs and ICH Q12 established conditions to fix the category in advance. Negotiating a lower category through an approved protocol converts a future dispute into a settled expectation ⁴²⁴³⁴⁴.

A deeper dive into any single change type (for example, the full SUPAC level tables for a specific dosage form, or the sterility-assurance PAS triggers for a given biologic) is a natural next question to put to Rhizome directly.

Post-Approval Manufacturing Changes: Annual Report, CBE-30, and Prior Approval Supplement Classification in FDA Submissions