Where the "ingredients to avoid" lists came from
The genre started in the early 2000s when a handful of pet food advocacy bloggers began publishing lists of ingredients they considered dangerous. Some of those original concerns were legitimate. BHA and ethoxyquin have real regulatory signals behind them. Artificial colors genuinely serve no purpose in dog food.
But over time, the lists grew. Each blogger added their own concerns, often based on human health data extrapolated to dogs, or rodent studies at doses hundreds of times higher than what appears in commercial pet food. The lists got shared, copied, and expanded without anyone going back to check the citations. Ingredients like sodium selenite, meat meals, and beet pulp got added because they sounded bad, not because anyone had demonstrated they were bad.
The result is a pet food culture where many conscientious dog owners are avoiding ingredients that are perfectly safe while paying premium prices for marketing terms like "natural" and "holistic" that have no enforceable nutritional meaning.
Sniff's approach is different. Every ingredient verdict is assigned one of three tiers, each with a published basis type.
Active concern
Documented canine harm at commercial levels, or banned/restricted by a major regulatory body, or IARC-classified carcinogen pathway. Score penalty: -15 to -25. Triggers a hard cap.
Worth monitoring
Plausible concern from non-canine evidence, regulatory asymmetry between jurisdictions, or transparency issue. Score penalty: -5 to -10.
Evidence does not support the concern
The internet says it is bad, the peer-reviewed literature says otherwise. No penalty. Sometimes a small bonus for the premium alternative.
The 21 that actually warrant concern
Artificial preservatives
BHA (butylated hydroxyanisole). Flag
The International Agency for Research on Cancer classifies BHA as a Group 2B possible human carcinogen. California lists it under Prop 65, requiring cancer warning labels. The FDA announced a reassessment in 2025. The EU permits it only up to 200 mg/kg of fat content. Rodent studies show forestomach tumors at chronic high doses. Dogs lack a forestomach, which weakens the direct extrapolation, but the regulatory weight plus the existence of safe natural alternatives (mixed tocopherols, rosemary extract) makes this an easy flag. There is no reason to use BHA in dog food in 2026 when better options exist and cost pennies more.
BHT (butylated hydroxytoluene). Watch
Less evidence than BHA. IARC classifies it as Group 3 (not classifiable as carcinogenic). The FDA included it in a 2025 Request for Information alongside BHA. The safety profile at typical pet food levels (100 to 200 ppm) is not alarming, but it loses to natural antioxidants on every dimension: safety margin, consumer trust, and regulatory trajectory. No compelling reason to choose it.
Ethoxyquin. Flag
The EU suspended ethoxyquin in 2017 and it remains suspended in 2026. The suspension was based on the inability to establish a safe acceptable daily intake and concerns about a genotoxic impurity called p-phenetidine. Ethoxyquin was originally developed by Monsanto as a rubber stabilizer. Canine 90-day studies showed elevated liver enzymes and pigment changes. The FDA requested voluntary reduction to 75 ppm in 1997. Most premium US brands removed it years ago. If your dog's food still contains ethoxyquin, that tells you something about the manufacturer's priorities.
TBHQ (tert-butylhydroquinone). Watch
Prohibited in Japan. Less studied than BHA or BHT, but emerging immunotoxicity signals in rodent models. The regulatory asymmetry (legal in the US, banned in Japan) combined with the availability of natural alternatives earns the watch. Not as strong a signal as BHA, but no reason for a manufacturer to use it.
Propyl gallate. Watch
Often used alongside BHA and BHT. Limited standalone data in dogs. Flagged primarily because it tends to appear in formulations that already contain other synthetic preservatives, which compounds the exposure.
Artificial colors
Red 40, Yellow 5, Yellow 6, Blue 1, Blue 2. Flag
The EU has required warning labels on foods containing these dyes since 2010, based on the McCann 2007 Lancet study linking them to hyperactivity in children. California banned six synthetic dyes from school foods in 2024. The FDA and HHS announced a phase-out initiative in April 2025.
Here is the thing that makes this an easy call: dogs are dichromats.
They see in shades of blue and yellow. They cannot perceive the red, orange, or green that these dyes create. The dyes exist to make the food look appealing to you in the store. Your dog gets zero benefit from them. A colorless food with the same formulation would be nutritionally identical. When a manufacturer adds artificial color to dog food, they are accepting a nonzero safety risk for a purely cosmetic benefit that the intended consumer cannot even see.
Added sugars
Corn syrup, high-fructose corn syrup, sucrose, dextrose. Watch
No acute toxicity. But 60% of US dogs are overweight or obese according to the Association for Pet Obesity Prevention. Adding sugar to a complete dog food is nutritionally unjustifiable. It exists to improve palatability in formulations that would not taste good otherwise. The presence of added sugar is a quality signal: if the food needed sweetening, something about the base formulation is not working.
Chemical preservatives banned or restricted elsewhere
Sodium nitrite / sodium nitrate. Flag
Documented canine death. A 2005 case report in the New Zealand Veterinary Journal describes a dog that died of methemoglobinemia from pet food containing 2,850 mg/kg of nitrite. The IARC classifies the nitrite-to-nitrosamine pathway as Group 2A (probably carcinogenic). The EU tightened residual nitrite limits in cured meats to 80 mg/kg in 2023. Added nitrite in dog food is unnecessary. There is no botulism niche to fill as there is in cured human meats.
Titanium dioxide. Flag
The EU banned titanium dioxide in food as of August 2022 after EFSA concluded it could not rule out genotoxicity. It is still permitted in US food and pet food. Its function in pet food is cosmetic whitening. Like artificial colors, this is a zero-benefit additive with a nonzero safety concern. The EU did not ban it because they enjoy banning things. They banned it because the safety data could not support continued use.
The synthetic vitamin K problem
Menadione sodium bisulfite complex (vitamin K3). Watch
Menadione is banned in over-the-counter human supplements in the US due to liver and kidney toxicity concerns at high doses. It is currently the only AAFCO-approved vitamin K source for pet food. This means even manufacturers who would prefer to use a natural alternative (K1 from green vegetables, K2 from fermentation) have no approved pathway to do so.
The toxicity threshold is more than 1,000 times the recommended dietary inclusion. At typical pet food levels (1.64 mg/kg dry matter), there are no documented cases of canine harm. The watch status reflects the regulatory asymmetry (banned for humans, required for pets) and the fact that AAFCO has not approved alternatives despite available options. If your dog's food contains menadione, it is almost certainly at a safe level. The issue is systemic, not product-specific. More on why AAFCO works this way.
Vague-source proteins and fats
The pattern in this section is consistent: every unnamed-species ingredient earns a watch because the label opacity prevents quality assessment. Named species is the differentiator. Below, two cards showing what the Templeman 2022 digestibility data finds when you actually measure protein quality.
Templeman 2022 DIAAS-like
- Chicken meal37 to 47
- Chicken by-product meal32 to 42
Organ meats are nutritionally dense, even if the term sounds bad.
Templeman 2022 DIAAS-like
- Meat meal (unnamed)22 to 31
- Animal by-product meal22 to 23
Less than half the digestibility of named chicken meal.
Same word ("meal" or "by-product meal"). Very different nutrition. The name on the label is the signal that separates them.
"Poultry by-product meal" (unnamed). Watch
Not the same as "chicken by-product meal." When the label says "poultry" instead of naming the species, you do not know whether you are getting chicken, turkey, duck, or a rotating mix. The nutritional quality might be fine. The transparency is not.
"Meat meal" (unnamed). Watch
AAFCO defines "meat meal" as rendered mammalian tissue, excluding blood, hair, hoof, horn, hide trimmings, manure, and stomach contents. But "meat" without a species name means you do not know which animal. Could be beef, pork, lamb, or something less identifiable. The protein is there. The quality and consistency are not guaranteed.
"Animal fat" (unnamed). Watch
AAFCO's definition of "animal fat" is rendered fat from mammals with no species identification requirement. This is the ingredient that has the widest quality variance in the pet food supply. Named fats ("chicken fat," "beef fat," "salmon oil") tell you the species. "Animal fat" tells you nothing.
"Animal digest" (unnamed). Watch
AAFCO definition 9.10: hydrolyzed unspecified animal tissue. Used as a palatability enhancer, essentially a flavor spray. Not toxic, but completely opaque about what is in it. "Chicken liver digest" is a meaningful upgrade in transparency.
"Animal by-product meal" (unnamed). Watch
The worst-performing protein source in the Templeman 2022 data. Less than half the digestibility of named chicken meal. When this is in the first three ingredients, the food's protein quality is fundamentally compromised regardless of the crude protein percentage on the label.
"Meat and bone meal" (unnamed). Watch
DIAAS-like value of 14 to 31 depending on source. Can be as low as 68% apparent crude protein digestibility. When this is position 1, the food is built on the cheapest available rendered mammalian tissue.
Emulsifiers and thickeners with emerging signals
Carrageenan. Watch
Important distinction: food-grade undegraded carrageenan (molecular weight over 100 kDa) is not the same as degraded poligeenan, which IARC classifies as Group 2B. EFSA confirmed food-grade carrageenan safe at use levels in 2018. However, rodent studies show a plausible colitis mechanism, and researchers have raised concerns about potential degradation during processing. No canine clinical evidence exists. Reasonable concern for dogs with IBD or sensitive GI tracts. Not justified as a hard flag for healthy dogs.
Guar gum, xanthan gum. Watch
Common thickeners in wet food. Emerging microbiome research suggests emulsifiers may affect gut barrier function, but the canine-specific evidence is thin. Minor penalty. Multiple emulsifiers stacked in the same formula earn a stronger flag than any single one alone.
Other concerns
Propylene glycol. Watch for dogs · Flag in multi-cat households
Banned in cat food since 1996 because it causes Heinz-body anemia in cats. In dogs, a two-year study at up to 8% of the diet showed no adverse effects. The canine safety profile is actually clean. The watch is because it is unnecessary in dry kibble (it is used as a humectant in semi-moist foods) and because of the cross-species risk in households with both dogs and cats.
Copper sulfate at high total copper levels. Watch 20-25 · Flag >25 mg/kg DM
This is one of the most important emerging safety signals in commercial dog food. After AAFCO changed permitted copper sources in 1997 from poorly absorbed copper oxide to highly bioavailable copper sulfate and chelates, veterinary researchers documented rising rates of copper-associated liver disease in dogs. Cornell, Tufts, and Colorado State have published data. Multiple experts called for AAFCO to set a copper maximum. AAFCO's Pet Food Committee voted against it in 2023.
Sniff treats total dietary copper from sulfate or chelate sources as a graduated concern: WATCH at 20 to 25 mg/kg dry matter, with a small CIP deduction, and FLAG at over 25 mg/kg DM, which triggers a hard score cap at 59 (D-tier) regardless of how good the rest of the formula looks. The cap exists because copper accumulation is irreversible in dogs without intervention, and the regulatory body responsible for setting a maximum has declined to do so. Full detail in the AAFCO guide.
Sodium tripolyphosphate (STPP). Watch
Strong evidence of renal injury markers in cats at supplemental phosphorus above 0.5 g/Mcal. The dog data is extrapolated but physiologically reasonable, particularly for senior dogs with subclinical kidney compromise.
The 8 people wrongly think are bad
This is the section that will get us in trouble on certain forums. That is fine. The data supports every word.
Sodium selenite. Clear
This is the single most maligned ingredient in pet food that has zero documented safety concerns in dogs.
Sodium selenite is an inorganic form of selenium, an essential trace mineral. It is found in 85% or more of commercial kibble. AAFCO sets a maximum of 2 mg/kg dry matter. Typical inclusion is 0.1 to 0.3 mg/kg. At these levels, there are no documented cases of selenium poisoning in dogs from commercial pet food. None. A 2021 systematic review in PMC found nothing.
The "sodium selenite is dangerous" narrative traces back to consumer pet food blogs citing the injection studies above. Injecting a mineral directly into tissue is a fundamentally different exposure pathway than eating it in food. The dose, the route, and the biology are not comparable.
Selenium yeast is a marginal upgrade. It is an organic form with slightly higher bioavailability and a wider safety margin. Premium brands use it, and Sniff awards a small bonus (+1) for selenium yeast as a formulation quality signal. But penalizing sodium selenite would be epistemically dishonest. We will not do it, and we will explain why every time someone asks.
Named by-product meals. Clear
By-products include organ meats, bone, and connective tissue. Liver, heart, gizzards, kidneys. These are some of the most nutrient-dense tissues in an animal. Liver alone contains more vitamin A, B12, iron, and copper per gram than muscle meat.
The term "by-product" sounds like waste. It is not. It is the parts of the animal that Americans do not typically eat but that are nutritionally superior to the muscle meat we do eat. In most of the world, these are considered the best parts of the animal.
The key distinction is named vs unnamed. "Chicken by-product meal" tells you the species. You know it is chicken organs, rendered and dried. "Poultry by-product meal" does not name the species, and "animal by-product meal" is a black box. Sniff flags unnamed by-products (transparency concern) and clears named ones (nutritionally sound).
Beet pulp. Clear
Beet pulp is the fibrous material left after sugar is extracted from sugar beets. By the time it reaches dog food, virtually all the sugar is gone. It is not a sugar source. It is a fiber source, and a good one.
Beet pulp is a moderately fermentable fiber, meaning it feeds beneficial gut bacteria and produces short-chain fatty acids that nourish the cells lining the colon. It has been studied extensively in dogs. It provides both soluble and insoluble fiber in a ratio that supports stool quality without the gas and bloating that come from highly fermentable fibers like chicory inulin at high inclusion.
The "beet pulp is filler" and "beet pulp is sugar" claims originate from the same blog-citation-chain that produced the sodium selenite myth. The name contains the word "beet" and the word "pulp," both of which sound unappetizing. The ingredient itself is a well-validated functional fiber.
Brewers rice. Clear (with context)
Brewers rice is a milling by-product: small, broken rice fragments that are separated from the whole kernel during processing. It has lower nutrient density than whole brown rice (less fiber, fewer B-vitamins, less manganese and selenium). It is not a safety concern. It is a quality signal.
If brewers rice is the primary carbohydrate source in a food marketed as "premium," that is a formulation choice that prioritizes cost over nutrition. If it is at position 8 behind named proteins, named fats, and whole grains, it is a minor ingredient providing digestible energy. Context matters. Sniff applies a small deduction (-2) not because brewers rice is harmful, but because its presence in a prominent position suggests the formulation is cutting corners.
Named meat meals (chicken meal, salmon meal, lamb meal). Clear
This one needs to be said clearly because the "fresh is always better than meal" narrative has done real damage to consumer understanding.
Chicken meal is chicken that has been rendered and dried to about 10% moisture. It is approximately 65% protein by weight. Fresh deboned chicken is about 70% water and 18% protein by weight. In extruded kibble, where everything goes through the same high-heat processing anyway, the "fresh" chicken at position 1 may contribute less actual protein to the finished product than the chicken meal at position 3.
Think of it like comparing fresh grapes to raisins. The grapes weigh more because of water. The raisins are more concentrated. In the context of kibble, meal is the raisin. It delivers more protein per gram in the final bag. A good chicken meal from a reputable supplier (low ash content, under 10%) is a strong protein source. A named fresh meat plus a named meal of the same species in the top three is the gold-standard configuration. More on this in the label guide.
The quality variable is the supplier, not the rendering process itself. Poor chicken meal (high ash, 12 to 16%) is lower quality. Good chicken meal is excellent. The label will not tell you the ash content, which is one of the data gaps Sniff's methodology tries to address through the protein quality index.
Whole corn. Clear
Corn has been treated as a villain by the pet food marketing industry for over a decade. "Corn-free" is printed on bags as if it were a health claim. It is not.
Whole ground corn is a digestible energy source that provides linoleic acid (an essential fatty acid), beta-carotene, and lutein. When gelatinized through extrusion, corn starch is over 95% digestible by dogs. It is not a top-tier carbohydrate (whole brown rice, oats, and barley score higher in Sniff's carbohydrate quality index) but it is not filler and it is not harmful.
The distinction that matters: whole ground corn is fine. Corn gluten meal used as a primary protein source is a different story. Corn gluten meal is 60% protein but severely limiting in tryptophan and lysine, two essential amino acids for dogs. When corn gluten meal replaces animal protein in the top five ingredients, that is a formulation problem. But that is a problem with how it is used, not with corn itself.
Chicken fat. Clear
"Chicken fat" sounds unappetizing to humans. For dogs, it is one of the best fat sources available.
Chicken fat is a named, species-specific fat with a favorable fatty acid profile: high in linoleic acid (an essential omega-6) and oleic acid (a monounsaturated fat). It is 90 to 98% digestible. The fact that the species is named ("chicken fat" vs "animal fat" vs "poultry fat") is itself a quality indicator: you know exactly what animal it came from.
The "chicken fat is bad" perception comes from projecting human dietary biases onto dogs. In human nutrition, we have been taught to minimize animal fats. Dogs have no cholesterol concern and metabolize fat as a primary energy source far more efficiently than carbohydrates. Named animal fat is a feature, not a flaw.
Mixed tocopherols. Clear (positive signal)
Mixed tocopherols are natural vitamin E compounds used as preservatives to prevent fat oxidation (rancidity). They are the direct replacement for BHA, BHT, and ethoxyquin.
If your dog's food uses "mixed tocopherols" or "natural mixed tocopherols" as its preservative system, that is a manufacturer that chose the safer, more expensive option. Sniff awards a bonus (+1) for natural vitamin E. The fact that "tocopherols" sounds like a chemical and gets lumped in with synthetic preservatives on internet avoid-lists is one of the clearest examples of how ingredient fear is driven by word recognition, not science.
The framework behind every verdict
Every ingredient verdict on Sniff is assigned using four filters.
Regulatory action.
Has a major regulatory body (FDA, EFSA, AAFCO, IARC, a national government) banned, suspended, restricted, or required warning labels for this ingredient? If yes, that is a strong signal regardless of whether canine-specific data exists. Regulatory bodies move slowly and conservatively. When they act, the evidence threshold has usually been exceeded.
Documented harm.
Is there peer-reviewed evidence of harm in dogs at the concentrations found in commercial pet food? Not in rats injected subcutaneously. Not in humans at 1,000 times the dietary dose. In dogs, at the levels they actually encounter. If yes, that is the strongest possible flag.
Plausible mechanism.
Is there a biologically plausible pathway from the ingredient to harm, even if direct canine evidence is limited? This is where the precautionary flags live. Carrageenan has a rodent colitis mechanism but no canine clinical data. High-legume diets have the Freeman 2025 phospholipidosis biomarker but await replication. These get WATCH status because the mechanism is credible even if the canine case is not yet closed.
The Lindy test.
How long has this ingredient been used in animal nutrition? Mixed tocopherols as preservatives: roughly 40 years of validated use. Chelated minerals: 30 years. Whole grains in dog food: thousands of years. BHA in pet food: 60 years, but with accumulating regulatory action throughout. Pea protein concentrate as a primary protein source in extruded kibble: about 10 to 15 years. The longer the track record without safety signals, the higher the prior confidence. Novel ingredients carry the burden of positive evidence.
If an ingredient fails all four filters, it is CLEAR regardless of how many blog posts say otherwise.
That is how sodium selenite, beet pulp, and named by-product meals end up defended rather than flagged. The internet is not a peer-reviewed journal, and repetition is not evidence.
How Sniff handles each tier in scoring
Penalty -15 to -25 points and a hard score cap. A food containing BHA is capped at 49 (D-tier) regardless of how good the rest of the formulation is. Two FLAG ingredients cap the score at 39 (F-tier). This is the Yuka-inspired "non-negotiable" mechanic. A premium ingredient list does not wash away a suspected carcinogen.
Penalty -5 to -10 points. Three or more WATCH ingredients in a single food trigger a cap at 64 (C-tier). Individual WATCH flags do not destroy a score, but they accumulate. A food with menadione, animal digest, and generic poultry fat is carrying three small concerns that add up to a meaningful one.
No penalty. Some earn a small bonus: rosemary extract (+2 as a natural preservative), selenium yeast (+1 as a premium selenium form), chelated minerals (+2 to +5 for superior bioavailability), and mixed tocopherols (+1).
Every verdict, every penalty, and every cap trigger is documented in the full methodology. Every product page shows which ingredients were flagged, which verdicts were applied, and what the resulting penalty was. You can trace any score back to the specific ingredient that drove it. That is not a feature we added for credibility. It is the whole point.
Frequently asked questions
Is sodium selenite safe in dog food?
Yes. At AAFCO-permitted levels (maximum 2 mg/kg dry matter, typical use 0.1 to 0.3 mg/kg), there are no documented cases of selenium poisoning in dogs from commercial pet food. The "sodium selenite is dangerous" claim traces to rodent studies using subcutaneous injection, which is a completely different exposure pathway than eating it in food. Selenium yeast is a marginal upgrade in bioavailability, and premium brands use it. But sodium selenite at normal dietary levels is safe. A scoring system that penalizes it would be misleading, and Sniff does not.
Are by-products bad in dog food?
Named by-products (chicken by-product meal, turkey by-product meal) are nutritionally excellent. They contain organ meats, which are among the most nutrient-dense tissues in an animal. The concern is with unnamed by-products ("poultry by-product meal," "animal by-product meal"), where you do not know the species or the specific tissues involved. Sniff clears named by-products and watches unnamed ones. The name matters more than the word "by-product."
Is corn bad for dogs?
Whole ground corn is a digestible energy source that provides essential fatty acids and micronutrients. Dogs digest gelatinized corn starch at over 95% efficiency. The "corn is filler" narrative is a marketing position, not a scientific one. The issue is not corn itself but how it is used: corn gluten meal as a primary protein source (limiting in tryptophan and lysine) is a formulation problem. Whole corn as a carbohydrate source alongside named animal proteins is fine.
What is the difference between chicken meal and fresh chicken?
Chicken meal is chicken that has been rendered and dried to about 10% moisture, concentrating the protein to roughly 65% by weight. Fresh deboned chicken is 70% water and 18% protein. In extruded kibble, the "fresh" chicken loses most of its water during processing. Chicken meal at position 3 may contribute more actual protein to the finished product than fresh chicken at position 1. A named fresh meat plus a named meal of the same species in the top three is the strongest protein configuration.
Does my dog's food really need preservatives?
Yes, if it contains fat. Fat oxidation (rancidity) produces harmful compounds and degrades nutritional value. The question is not whether to preserve but how. Natural preservatives (mixed tocopherols, rosemary extract, vitamin C) are effective and carry no safety concerns. Synthetic preservatives (BHA, BHT, ethoxyquin) work but carry regulatory signals that natural alternatives do not. If your food lists "preserved with mixed tocopherols," that is a positive sign, not a concern.
What is the difference between BHA and mixed tocopherols?
BHA is a synthetic antioxidant classified as a possible carcinogen by IARC, listed under California Prop 65, and under FDA reassessment. Mixed tocopherols are natural vitamin E compounds that do the same job (prevent fat rancidity) without the regulatory baggage. They cost slightly more, which is why budget brands still use BHA. The price difference per bag is pennies. The safety difference is meaningful. Sniff flags BHA and awards a bonus for mixed tocopherols.
Is "meat meal" worse than "fresh meat"?
Not necessarily. The quality of meat meal depends on the source, the species identification, and the rendering process. A named chicken meal with low ash content (under 10%) is a strong, concentrated protein source. An unnamed "meat meal" with no species identification and unknown ash content is a transparency problem. "Fresh" sounds better on a label, but in the context of extruded kibble where everything undergoes high-heat processing, the distinction between fresh and meal is smaller than the marketing suggests. What matters most is whether the species is named and whether the source is traceable.
Related reading
How to Read a Dog Food Label
The reading skills that make flagged ingredients visible at a glance, plus split-ingredient labeling and dry-matter conversion.
AAFCO Statements Explained
Why a regulatory framework with industry advisors writing its own rules cannot set a copper maximum even when researchers ask.
Grain-Free Dog Food and DCM
The pulse-stacking concern is real. The grain panic is mythology. Same divide, different ingredient category.