Vitamin D Uniformity Controls That Prevent Pet Food Recalls

A manufacturing quality perspective on what goes wrong with micronutrient dosing — and how to stop it before FDA gets involved.

When a company recalls a product because of "variable levels" of a nutrient — some batches too low, some too high — that phrase tells you almost everything you need to know about what went wrong. The problem is not that they used the wrong ingredient. The problem is that their manufacturing and testing systems let the same formulation produce wildly different outcomes across production runs. That is a process control failure, and it is entirely preventable.

Revival Animal Health's voluntary recall of Breeder's Edge® Foster Care® Canine and Shelter's Choice® Canine Milk Replacers, flagged by FDA for variable Vitamin D levels, is a useful case study in exactly this kind of failure. I'm not writing about the recall to pile on a company doing the right thing by pulling product. I'm writing about it because the underlying quality gap — micronutrient non-uniformity — shows up across the pet food and animal health industries more often than most people realize, and the fixes are well understood.

If you manufacture, co-pack, or supply animal food products regulated under FDA's animal food rules, this is worth your time.

What "Variable Levels" Actually Means in Manufacturing Terms

Variable nutrient levels in a finished product almost always trace back to one of three root causes: a bad premix, a blending failure, or an analytical gap that let both slip through undetected.

Premix quality issues happen when the vitamin or mineral concentrate you receive from your supplier is itself out of spec, either over-potent, under-potent, or heterogeneous. If you accept a Vitamin D premix lot without incoming verification testing, you are trusting that your supplier's certificate of analysis reflects reality. Sometimes it does not.

Blending failures happen when the premix is fine but the mixing process doesn't distribute it uniformly throughout the batch. Vitamin D is a fat-soluble micronutrient added at very low inclusion rates — often fractions of a percent of the total formula. At those concentrations, poor blending creates "hot spots" and "cold spots" in the same batch, which means finished product from one end of a run can have dramatically different nutrient content than product from the other end. This is the "variable levels — both low and elevated" pattern that shows up in the Revival recall language.

Analytical gaps happen when finished product testing either doesn't happen, happens too infrequently, or uses a sampling method that isn't designed to catch non-uniformity. A single composite sample pulled from one point in a batch can pass perfectly while significant variation exists elsewhere in the same lot.

Each of these has a corresponding control. The fact that a recall happened tells us the controls either weren't in place or weren't working.

The Regulatory Framework: What FDA Actually Requires

Animal food manufacturers operating in the United States are subject to FDA's Current Good Manufacturing Practice (cGMP) requirements for animal food under 21 CFR Part 507, which was finalized under the Food Safety Modernization Act (FSMA). Subpart B of Part 507 establishes the cGMP requirements; Subpart C establishes hazard analysis and preventive controls requirements for facilities that are not otherwise exempt.

A few specific regulatory hooks are directly relevant to micronutrient uniformity failures:

21 CFR 507.25 — Receiving, Inspecting, and Storing Operations requires that ingredients and raw materials be inspected on receipt and held under conditions that protect against contamination and deterioration. For vitamin premixes, this includes incoming lot verification — you cannot simply accept a supplier's CoA without a defined procedure for confirming it.

21 CFR 507.34 — Manufacturing Operations requires that manufacturing be conducted under such conditions and controls as are necessary to minimize the potential for contamination or adulteration. Blend uniformity is a manufacturing control. If your SOP does not define mixing time, equipment parameters, and loading sequence for premix addition, you have a regulatory gap.

21 CFR 507.47 — Process Preventive Controls (for facilities subject to preventive controls) requires that you identify and implement controls for process steps where a hazard requiring a preventive control has been identified. Vitamin D at excessive levels is a known health hazard in dogs — FDA's concern in this recall is exactly that. An elevated Vitamin D level in a product intended for neonatal puppies or shelter animals is not a minor label deviation; it is a safety issue. A competent hazard analysis under 21 CFR 507.33 should have identified Vitamin D over-fortification as a hazard warranting a preventive control, with a corresponding monitoring procedure, corrective action plan, and verification protocol.

21 CFR 507.55 — Verification Activities requires that process preventive controls be verified, which means more than just doing a batch record check. Periodic testing of finished product — at a frequency and sampling plan sufficient to detect non-uniformity — is expected verification activity.

In my view, the gap most commonly exploited in pet food micronutrient failures is not the absence of a rule; it is the absence of a risk-based testing frequency. Companies do some finished product testing but do not design their sampling plan around the specific uniformity risk posed by low-inclusion-rate fat-soluble vitamins.

Specific Controls That Would Have Prevented This

Here is what a solid quality system looks like for this exact hazard. These are not theoretical — they are the controls I help clients implement, and they work.

Incoming Premix Verification

Every lot of vitamin premix should be tested — or at minimum verified against a supplier CoA using a defined skip-lot protocol with clear escalation rules — before it enters production. For Vitamin D specifically, the test method matters. You need a method validated for the concentration range in your premix, not a generic vitamin panel. A supplier qualification program under 21 CFR 507.25 should include:

• Approved supplier list with defined qualification criteria
• Incoming lot sampling and testing requirements by ingredient risk tier
• CoA review procedure with defined acceptance criteria
• Escalation protocol for out-of-spec or borderline results

If you are relying entirely on your supplier's CoA without any independent verification, you are one supplier error away from a recall.

Blend Uniformity Testing (BUT)

Blend uniformity testing is the most underused control in small-to-medium animal food manufacturing. The concept is straightforward: before you discharge a batch from the blender, pull samples from multiple points in the blender (typically 10 locations at minimum for a pharmaceutical-style protocol, though a pragmatic risk-based approach for food applications is acceptable) and test them for one or more marker analytes. If the relative standard deviation across your samples exceeds your acceptance criterion, the batch is not uniform and should not proceed to packaging.

For a product like a milk replacer — where Vitamin D is a critical nutrient for the target population and where over-fortification poses a direct safety hazard — blend uniformity testing is not optional in any quality system worth having. It should be a defined process preventive control with a specific monitoring procedure, a validated sampling method, and a documented corrective action for failing results.

Finished Product Testing with a Uniformity-Sensitive Sampling Plan

Finished product testing is necessary but not sufficient on its own. The sampling plan has to be designed to catch non-uniformity, which means:

• Pulling samples from the beginning, middle, and end of a packaging run (not just a single composite)
• Testing a sufficient number of units per lot to have statistical power to detect meaningful variation
• Defining acceptance criteria based on the nutrient specification, not just "does it pass or fail"

A single composite sample pulled from a finished pallet can average out exactly to your target level even when individual units vary by 200%. Compositing hides non-uniformity. If detecting non-uniformity is your goal, your sampling plan has to be designed for it.

Micronutrient Specification Design

This one is worth naming because it gets skipped. When you set your Vitamin D specification, what is your upper limit, and is it based on safety data? Many companies set their specs around regulatory minimums and label claims without explicitly incorporating a safety-driven upper limit. For Vitamin D in canine products, the Association of American Feed Control Officials (AAFCO) and NRC have established both minimum requirements and maximum safe levels. Your internal specification should include a maximum that accounts for analytical variability, label claim, and a reasonable safety margin. If your upper limit is "label claim + 10%," that is not a safety-based specification. It is a marketing-based specification.

What a Hazard Analysis for This Product Should Have Flagged

Under 21 CFR 507.33, a facility subject to preventive controls is required to conduct a hazard analysis that identifies known or reasonably foreseeable hazards associated with each type of animal food manufactured. For a canine milk replacer, a competent hazard analysis would include:

A hazard analysis that identifies Vitamin D overfortification as a hazard requiring a preventive control will then drive the downstream controls: the monitoring procedure, the corrective action, the verification testing, the records. The entire system runs downstream of the hazard analysis. If the hazard analysis misses it, everything else misses it.

In my experience working with 200+ clients across food, drug, and animal health regulated spaces, the hazard analysis is the document most likely to be technically present but functionally weak. It exists because the regulation requires it. But it was written to pass an inspection, not to actually drive the control system. You can tell the difference quickly — a working hazard analysis gets updated when something changes (new supplier, new formulation, new equipment), and the controls it requires are actually implemented and monitored. A compliance-theater hazard analysis is a static document that nobody references between audits.

The Supplier Control Dimension

One thing that frequently goes unexamined in micronutrient failures is the supplier side of the equation. Under 21 CFR 507.25, you are responsible for the ingredients you receive. That responsibility does not end at placing a purchase order with an approved supplier.

A practical supplier control program for vitamin premix suppliers should include:

• Supplier qualification audit — either a first-party (you conduct it) or second-party (industry group or shared audit) assessment of the supplier's manufacturing controls for the premix
• Specification agreement — a written agreement defining the CoA requirements, acceptable test methods, shelf life, and notification requirements for any manufacturing changes
• Annual performance review — tracking lot-by-lot CoA results over time to catch drift before it becomes a failure
• Change notification clause — suppliers frequently change raw material sources, processing aids, or equipment without proactively notifying customers; a contractual notification requirement doesn't guarantee you'll be told, but it establishes the expectation and gives you recourse

The premix supply chain for pet food is not as tightly controlled as pharmaceutical API supply, and that gap creates real risk. Several notable pet food recalls over the past decade have traced to supplier-side issues that the manufacturer's incoming controls didn't catch. The Revival recall does not have a publicly confirmed root cause at the time of this writing, but the "variable levels" language is consistent with either a premix homogeneity issue or a batch-to-batch potency variation from the supplier — both of which should be caught by incoming verification.

Practical Timeline: Getting Controls in Place

If you are an animal food manufacturer currently without robust micronutrient controls, here is a reasonable implementation sequence. FSMA's animal food preventive controls rule (21 CFR Part 507 Subpart C) has been in effect since 2017 for most facilities, so the regulatory deadline is not coming — it has already passed. The urgency is operational, not regulatory. A gap inspection or a recall is the cost of waiting.

None of this is exotic. It is the kind of systematic work that quality professionals do every day. What it requires is prioritization and, in many cases, an outside perspective to see the gaps that internal teams have become habituated to.

What This Means for Co-Packers and Contract Manufacturers

If you manufacture animal food products under a customer's label — as a co-packer or contract manufacturer — your regulatory obligations under 21 CFR Part 507 are the same as if the product carried your own brand. The fact that the formula, label, and brand belong to your customer does not shift the cGMP obligation off your facility.

This matters for micronutrient control because co-packing relationships frequently involve divided responsibilities: the customer specifies the formula and the vitamin premix, and the contract manufacturer follows the formula. But if the premix is out of spec when it arrives at your facility, or if your blending process doesn't achieve uniformity, the recall lands on both parties — and the FDA Form 483 observations come to your facility.

In my view, contract manufacturers should maintain their own incoming verification requirements even for customer-supplied ingredients, and should require customers to provide formulation specifications with defined acceptance criteria for critical micronutrients. A customer saying "use this premix, add X% to the formula" is not a specification. It is an instruction. The specification has to define what acceptable looks like in the finished product.

A Note on Neonatal and Vulnerable Populations

The Revival recall involves canine milk replacers — products designed for puppies who cannot nurse, often neonates in their first days of life. This population has essentially no margin for error on micronutrient dosing. A neonate cannot compensate for nutritional imbalance the way an adult animal can. Vitamin D toxicity in dogs produces hypercalcemia, which can cause kidney failure, cardiac abnormalities, and death — and the dose-response relationship is steeper in very young animals.

When you are manufacturing for vulnerable populations — neonates, geriatric animals, animals with specific medical conditions — the hazard analysis should explicitly account for the heightened vulnerability of the intended end user. The risk profile of a Vitamin D overfortification event in a milk replacer is different from the same event in a dry adult maintenance diet. Your preventive controls should reflect that difference.

Certify Consulting works with animal food and veterinary product manufacturers at exactly this kind of intersection — where regulatory compliance, quality systems, and product-specific risk come together. If your hazard analysis hasn't been updated recently, or if your vitamin premix controls feel more like paperwork than actual prevention, that's a conversation worth having.

Key Takeaways

The Revival Animal Health recall is a clear reminder that micronutrient uniformity failures are not rare or mysterious. They follow predictable patterns and respond to predictable controls. The regulatory framework under 21 CFR Part 507 has been in place since FSMA — the tools are there. What tends to be missing is a quality system that takes the hazard analysis seriously enough to drive real controls, tests at the right points and with the right sampling design, and holds suppliers accountable for the ingredients they supply.

Three things, in my view, would prevent the majority of micronutrient-related pet food recalls: a hazard analysis that explicitly identifies overfortification as a safety hazard, a blend uniformity testing protocol for low-inclusion-rate micronutrients, and an incoming verification program that goes beyond accepting a supplier's CoA at face value.

None of this is complicated. It is just disciplined.

Jared Clark, JD, MBA, PMP, CMQ-OE, CQA, CPGP, RAC is the Principal Consultant at Certify Consulting, where he has helped 200+ clients across food, drug, and animal health achieve and maintain regulatory compliance — with a 100% first-time audit pass rate.

Source: FDA Recall Notice — Revival Animal Health, LLC

Last updated: 2026-05-02