Food Allergen Control

Food Allergen Control connects two different responsibilities: keeping unintended allergens out of food and accurately declaring the allergens used to make it. A bakery can fail either responsibility independently. The correct recipe in the wrong bag creates a labeling problem; the correct bag does not resolve sesame transferred from a previous product. Effective operating practice therefore follows both the food and its packaging, rather than treating an allergen statement as the whole program.

The Food Allergen Labeling and Consumer Protection Act of 2004, or FALCPA, established major-allergen labeling provisions for most packaged FDA-regulated foods, including dietary supplements. FDA’s explanation excludes USDA-regulated meat, poultry, and egg products; alcoholic beverages subject to TTB labeling regulations; raw agricultural commodities; highly refined oils; drugs; cosmetics; and most retail or food-service foods not prepackaged with a label. These are boundaries of this labeling framework, not a declaration that other products have no applicable safety obligations.

Retail is not a blanket exemption. FDA states that FALCPA labeling requirements extend to retail and food-service establishments that package, label, and offer products for consumption. They do not apply to food placed in a wrapper or container following a customer’s order at the point of purchase. A bakery selling both prepacked cookies and counter-service pastries should distinguish those activities. See FDA’s food allergen labeling overview.

As a scoping practice, document the destination market, product category, packaging situation, and applicable facility requirements before building a checklist. Other markets identify different allergens; check the destination market rather than exporting a US-only matrix unchanged. Certification schemes have their own requirements. Neither a certification checklist nor a label review should replace the facility-specific applicability assessment.

The nine major food allergens identified by US law are milk, eggs, fish, Crustacean shellfish, tree nuts, peanuts, wheat, soybeans, and sesame. FALCPA originally identified eight. The FASTER Act added sesame as the ninth, effective January 1, 2023. FDA states that the labeling and manufacturing requirements applicable to major food allergens apply to sesame from that date. An old eight-allergen questionnaire is therefore not an adequate working reference for current US production.

For covered foods, the label must identify the food source of major allergens used to make the food. FDA explains that an ingredient’s common or usual name may already identify its allergen source. Otherwise, the source can be identified parenthetically, such as whey (milk), or through a Contains statement immediately after or next to the ingredient list. A separate Contains statement is not universally required. The specific tree nut type and fish or Crustacean shellfish species also require identification.

As a label-development practice, review the complete formula rather than only ingredients whose names look allergenic. Examine compound ingredients, coatings, fillings, toppings, flavors, and approved rework. FALCPA expressly subjects a flavoring, coloring, or incidental additive that is, or contains, a major food allergen to its allergen labeling provisions. Do not assume that a small addition or a processing function automatically removes the need for review.

FDA announced final allergen labeling Q&A guidance, Edition 5, on January 6, 2025. Its announcement describes questions covering tree nuts, sesame, milk, eggs, incidental additives, highly refined oils, dietary supplements, and specific packaging situations, including units within multiunit packages. Guidance communicates FDA’s current thinking; it is not a new statute. Consult the relevant answer for a particular labeling question rather than inferring detailed exceptions from the topic list.

A useful review practice is to compare three things side by side: approved ingredient information, the formula actually authorized for production, and the proposed label. Resolve disagreements before releasing artwork. Record which ingredient versions support the decision so a supplier change can be assessed against the correct starting point.

Practice: build an allergen map that separates intentional ingredients from potential cross-contact information. Combining both into one undifferentiated yes/no column makes decisions harder. A supplier’s advisory statement does not tell the bakery that the allergen is an intentional ingredient, while an ingredient specification alone does not describe every handling exposure inside the bakery. Preserve the distinction and investigate uncertainties rather than converting missing information into an absence claim.

Start with purchasing and receiving. Match the material identity, supplier, specification revision, and delivered lot. Review blends and compound ingredients at the component level where needed to understand their allergen sources. Include ingredients used outside the main mixer, such as release preparations, dusting materials, glazes, and finishing decorations. A purchasing substitution that seems technically equivalent may need a separate allergen and label assessment.

The table is a fictional mapping example, not a declaration about every material with a similar name. Actual supplier information governs the assessment. Useful fields include the material code, specification date, intentional allergen sources, unresolved questions, storage location, affected products, and approval owner. Keep obsolete information available for historical investigation without allowing it to masquerade as the current production instruction.

For record design, the concepts in ingredient traceability software evaluation can help frame questions about ingredient-to-batch relationships. Ask whether your chosen recordkeeping approach can reconstruct which specification and material lot supported a particular batch. This is a buyer evaluation question, not an assumption about any product’s capabilities. Test the relationship using an actual receiving record and a finished batch before relying on it during an investigation.

For facilities subject to the applicable preventive-control requirements, 21 CFR 117.130 requires a written hazard analysis, regardless of its outcome. It evaluates known or reasonably foreseeable hazards to determine which require preventive controls. Food allergens are included among chemical hazards. The assessment considers severity and probability in the absence of preventive controls, not simply whether an allergen appears somewhere on the site.

Where food allergen preventive controls are required, 21 CFR 117.135(c)(2) addresses both protection from cross-contact during storage, handling, and use, and finished-food labeling. Preventive controls must be written. The regulation is risk-based: it does not prescribe one universal bakery layout, production sequence, or inspection frequency for every product and facility.

As an operating practice, walk this flow during production, including temporary staging and interrupted work. Look for routes omitted from the formal process diagram: a borrowed scoop, dough returned from a divider, trays parked under topping equipment, or labels brought back from another line. These details can connect products that appear separate on a scheduling screen.

For each selected control, define the responsible role, the observation or check, the response to failure, and the record that demonstrates execution. Part 117 provides for monitoring, corrections or corrective actions, and verification as appropriate to the preventive control. The practical objective is a decision someone can execute at the workstation, not merely a statement that allergens are managed.

The applicable CGMP provisions address cross-contact independently of the decision to establish a particular preventive control. 21 CFR 117.20 addresses plant design, separation options, ventilation, dust, and air-blowing equipment. Section 117.40 addresses cleanable, maintained equipment. These provisions support a practical question: how could allergen-containing material move from its intended location onto another food or food-contact surface?

Fictional bakery practice example: wheat flour is tipped near a bench used later for a product formulated without wheat, while sesame topping is applied near uncovered plain rolls. Investigate the tipping and topping activities, not just the distance between benches. Observe airborne material, spilled seeds, clothing, scoops, rack routes, and the timing of exposed production. Assess changes such as contained handling, relocation, effective dust control, or separation in time against the actual transfer route.

Shared ovens deserve a specific assessment, but an oven is only one part of the route. Trays, pan rims, liners, loading tools, racks, cooling belts, and unloading practices may connect products before and after baking. As a conservative operating practice, do not credit ordinary baking as an allergen-removal step without a justified basis. A clean-looking finished loaf does not establish what happened to its tray or handling equipment.

Dedication and sharing involve trade-offs. Dedicated trays can simplify identification, but storage and handling still matter. Shared trays conserve equipment and space, but place more responsibility on cleaning, inspection, and release. Color coding is a useful identification aid only when employees understand it and the physical arrangements support it. A colored handle cannot compensate for a scoop stored in the wrong ingredient bin.

When mapping bakery operations, include maintenance and nonroutine activity: equipment opening, temporary hoses, replacement belts, and tools moved between areas. Practice: observe a real product transition and an interruption, not only a steady run. Document the transfer route, the chosen barrier, and how its condition will be checked. This gives the team a specific control to evaluate instead of a vague instruction to keep allergenic materials separate.

Practice: sequence products using their allergen profiles and the equipment they share, not simply the number of allergens each contains. A wheat-and-milk product and a wheat-and-sesame product are not interchangeable steps on a simple low-to-high ladder. Moving between them can introduce an allergen absent from the next recipe. Scheduling may reduce difficult transitions, but it does not establish that a transition is adequately controlled.

Under 21 CFR 117.35, food-contact surfaces must be cleaned as frequently as necessary to protect against allergen cross-contact and contamination. Low-moisture food-contact surfaces must be clean, dry, and sanitary before use; the rule also addresses drying after wet cleaning. Select cleaning practices for the equipment, residue, and production environment rather than importing a universal wash cycle.

Distinguish regulatory validation from verification. 21 CFR 117.160(c) says food allergen controls and sanitation controls do not need validation under that section. This does not remove applicable verification duties: section 117.165 requires verification that preventive controls are consistently implemented and effective. As practice, retain a reasoned basis for the cleaning approach and evidence appropriate to checking its performance.

Part 117 also calls for testing where necessary to identify sanitation failures or possible cross-contact. It does not turn a generic swab count into a universal bakery requirement. FDA’s food-allergies page states that it has not established a threshold level for any allergens. Do not convert a laboratory detection limit into a legal safe level. Define the purpose, method suitability, interpretation, and response before collecting samples.

Rework deserves ingredient-level attention. 21 CFR 117.80 requires allergen-containing rework to be identified and held in a manner that prevents cross-contact, and work-in-process and rework to be handled protectively. As practice, record the source batch, allergen profile, quantity, condition, and approved destination. A container marked only dough does not provide enough information for a reliable destination decision.

Fictional worked example: a bakery makes 240 kg of sesame milk-roll dough containing 6 kg of sesame. Its intentional major allergens are wheat, milk, and sesame. Before forming, it reserves 12 kg of clean, suitable dough as approved rework. For this mass-balance example, assume the dough is uniformly mixed, all quantities are measured before baking, and the receiving recipe and rework use have been approved. These assumptions are not evidence that an actual rework lot is suitable.

The original batch leaves 228 kg for forming and 12 kg in the rework container. The next compatible batch uses that 12 kg plus 228 kg of fresh dough. Recording both movements prevents the same rework from remaining available in inventory after use. The receiving batch retains wheat, milk, and sesame in its allergen assessment, with its label checked against the approved complete formulation.

In this example, the site does not authorize that rework for plain wheat rolls whose formula lacks milk and sesame. Using a smaller quantity would not solve the identity and labeling problem. The calculation tracks ingredient mass; it does not establish a safe exposure threshold or prove uniform allergen distribution in real production. Also, matching allergen profiles alone does not establish compatibility: formulation, product condition, other hazards, and intended use still need assessment.

Practice: treat label approval and pack-off verification as separate activities. Artwork approval asks whether the design is correct for the approved product. Pack-off verification asks whether the actual food is entering that correct package now. A technically correct file does not prevent an operator from loading an older roll, selecting a neighboring SKU, or leaving previous-product packaging at the line.

At startup, compare the product identity and authorized recipe with the package actually installed. Check the ingredient and allergen information, package code or revision, and relevant production identification. Include outer packs or assortments where they affect the intended presentation. Use a physical finished pack as evidence, rather than checking only a printer screen or the carton in which labels arrived. This is a suggested operating practice, not a prescribed universal inspection format.

At a label change, confirm removal or control of the previous packaging, identify the replacement, and check the first resulting pack before routine continuation under the site procedure. Consider partial rolls, printer restarts, emergency replacements, and changeovers involving visually similar products. Decide additional checks from the operation’s risks; there is no one sample size or interval established here for every bakery.

Fictional example: sesame milk rolls are being packed when a replacement label roll intended for plain wheat rolls arrives. A label-change check should compare the declared allergens with the sesame milk-roll product identity, not merely confirm that the replacement label prints legibly. If the wrong labels have been used, stop and establish the potentially affected production interval from reliable evidence. Do not assume the last carton alone defines the affected quantity.

Precautionary statements are a separate issue. FDA explains that advisory statements such as may contain are voluntary and may address unavoidable cross-contact only after good manufacturing processes and every precaution to avoid cross-contact. They should not substitute for CGMP and must be truthful and not misleading. See FDA’s explanation of advisory allergen labeling. As practice, evaluate the residual situation explicitly; do not add a blanket warning to compensate for an unresolved wrong-label or sanitation problem.

An allergen program needs evidence that can be connected, not simply a collection of completed forms. Practice: choose a finished lot and reconstruct its ingredients, rework, shared-equipment transition, and packaging. If those records use incompatible identifiers, a reviewer may know that cleaning occurred somewhere without knowing whether it supported this batch. A record relationship should answer a specific production question.

For applicable records, 21 CFR 117.305 requires actual monitoring values and observations, accuracy, legibility, and creation concurrently with the activity. Records must contain appropriate identifying information, including the person performing the activity and, where appropriate, the product and lot. An end-of-shift reconstruction from memory is not equivalent to recording what was observed when the check occurred.

As practice, preserve the versions that were valid at production time. Replacing an old recipe or label with its newest revision should not erase the basis for an earlier batch. Record exceptions and their disposition alongside routine evidence, so an investigation does not stop at a passed checkbox while overlooking a later packaging substitution.

Paper, electronic records, and combinations can support these relationships. Evaluate your approach by tracing in both directions: from a suspect ingredient or label to affected lots, and from a finished lot back to its inputs and checks. The useful question is whether the evidence is accurate, connected, and retrievable, not whether a particular technology is present.

Practice: define who can stop the operation, place material on hold, assess the extent of a problem, and authorize disposition. A sesame spill beside exposed plain rolls and a missing label-change check are different events, but both need an evidence-based assessment. Preserve what is known, separate it from assumptions, and avoid allowing production pressure to determine the affected boundary.

Where 21 CFR 117.150(a) applies, corrective-action procedures address correcting the problem, reducing recurrence when necessary, evaluating affected food for safety, and preventing affected food from entering commerce when its compliance cannot be ensured. The regulation also distinguishes certain timely corrections from corrective actions. Apply the relevant procedure; do not assume every discrepancy needs identical paperwork or that every quick fix resolves product safety.

A practical investigation follows the last reliable evidence of control, the event itself, and all subsequent movements. Consider work-in-process, packed stock, rework containers, and material already transferred elsewhere. Record why particular lots were included or excluded. If the extent cannot be established, do not narrow the hold merely because a smaller quantity is easier to manage.

Trade-offs should be explicit. More dedicated equipment may reduce some shared-equipment decisions while increasing storage and identification work. Longer production campaigns may reduce changeovers while complicating inventory and rework management. Assess these choices across the whole route. After a deviation, confirm that the corrective change addresses the failure mechanism rather than adding another signature to an unchanged process.

Practice: begin with one representative product family and its most consequential transitions, then extend the method across the operation. Include purchasing, receiving, production, sanitation, packaging, maintenance, and the person responsible for food safety decisions. Assign ownership of the handoffs: who accepts ingredient information, who authorizes rework, and who confirms that a package change matches the product? Unassigned handoffs can undermine otherwise clear procedures.

Applicable personnel requirements in 21 CFR 117.4 include qualification for assigned duties and food hygiene and food safety training appropriate to the food, facility, and duties. The provision includes temporary and seasonal personnel, and requires records of the specified training. As a practical supplement, ask employees to demonstrate the task: identify the correct scoop, reject an incompatible rework destination, or recognize a wrong label.

Build change assessment into routine work. A new filling, supplier substitution, different tray coating, revised cleaning access, or packaging redesign may alter the assumptions behind existing controls. For facilities subject to the relevant requirements, section 117.170 requires reanalysis for specified circumstances, including significant changes that create a new hazard potential or significantly increase an existing hazard, and ineffective controls. Do not wait for a calendar review to investigate a material change.

For a broader food safety workflow discussion, use these handoffs as neutral evaluation questions: how will observations be recorded, exceptions assessed, and decisions connected to affected lots? Readiness is demonstrated by a workable process and its evidence. A completed checklist is a starting point for checking implementation, not a guarantee that every product and transition is controlled.