What Are Manganese and Vitamin C Doing in My Dog's Joint Supplement?

The fourth line on the panel is manganese ascorbate — one compound supplying two of the things you read, because ascorbate is vitamin C. The amount on the label traces to a patent filed in 1993, not to anyone measuring what a dog needs. Practically, three things are worth knowing. One Cosequin tablet carries 3 mg of manganese, three times what a full daily serving of the same company's human product delivers. A dog over 60 lb on the first six weeks' loading dose is taking 9 mg a day, against an AAFCO food minimum of about 2.2 mg a day for a dog that size. And the vitamin C riding along with it — somewhere between 12 and 20 mg a tablet, most likely near the top of that range — is a nutrient dogs manufacture in their own livers and that AAFCO's dog food profile does not list at all. None of this is a reason to throw the tub away. It is a reason to step down to the maintenance dose on schedule rather than leaving a dog on the loading dose indefinitely. The dosing table says "initial 4–6 weeks"; nothing on the bottle marks the date that ends.

What the four ingredients on the panel actually are#

If you have typed the whole ingredient list into a search box, it is probably because the tub would not explain itself. Here is the panel most of these products carry, using the Cosequin chewable tablet as the reference — its figures were re-read on cosequin.com on 1 September 2026.^[1]

Three of those four rows are the thing you thought you were buying. The fourth is a mineral, and it is the only row that is not an ingredient in its own right so much as a salt — a delivery form. Manganese ascorbate is one manganese ion bound to two ascorbate molecules. Ascorbate is vitamin C. The formula weighs in at 405.17 g/mol against manganese's 54.94, so the compound is about 13.6% manganese by mass and about 86% vitamin C.^[2]

That is why an owner reading a label can end up searching for "manganese and ascorbic acid." They are not two decisions. They are one ingredient, and the label only names half of it.

dogorthopedic.com

Where three of those four rows came from#

The MSM row varies from formula to formula. The other three — glucosamine, chondroitin and the manganese — travel together wherever the manganese appears at all. And it is worth saying up front that it does not appear everywhere: of the joint supplements this site covers, the Cosequin tablet lines and Dasuquin Advanced print a manganese row, while the retail Dasuquin with MSM tub carries none at all. We have not surveyed the store brands, so we are not going to tell you what the whole aisle does. What we can tell you is where the row on your panel came from, and it is not convergent research. It is a patent.

US Patent 5,364,845, "Glucosamine, chondroitin and manganese composition for the protection and repair of connective tissue," was filed on 31 March 1993 by Robert W. Henderson and assigned to Nutramax Laboratories. It granted on 15 November 1994. Its claims cover exactly this combination — glucosamine or its salts, chondroitin sulfate, and a manganese salt, with manganese ascorbate named as the preferred form.^[3]

The patent also states its reasoning, which is worth reading because nobody puts it on a tub. On the manganese: it "is required for enzyme activity of glycosyltransferases" — the enzymes that link sugars into glycosaminoglycans and add sulfate to amino sugars — and so "acts as a cofactor for catalyzing the conversion of the glucosamine into hyaluronic acid." On the choice of salt: manganese ascorbate is preferred because it "is a soluble salt which also provides ascorbic acid needed for collagen synthesis."

Read that last clause again. The ascorbate half was not an accident of chemistry. It was chosen as a second active ingredient — vitamin C, for collagen. That is sound reasoning in a human, who cannot make vitamin C. It is much weaker reasoning in a dog, who can.

To be fair to the document, and against the easy version of this story: this is not a human-only patent. It carries case studies in dogs, cats and horses — a nine-year-old Rottweiler with difficulty rising in the rear, a nine-year-old cat with a limp, a thoroughbred with ankle chips — and it sets species-specific doses, "approximately 38 mg" of manganese ascorbate for humans and small mammals against "approximately 120 mg" for horses and large mammals.^[4] So dogs were in scope from the start. What was never done, then or since, is the study that would show the manganese contributes anything on its own.

The patent also does something no modern dog label does: it prints the manganese and the ascorbate as separate rows for the same capsule. (Nutramax's human label still does; its dog panel does not.) Its human/small-animal formulation lists "Manganese (as Ascorbate) 5 mg" beside "Ascorbate (as Manganese Ascorbate) 33 mg"; the equine one lists 16 mg against 104 mg.^[4] Both are a ratio of about 6.6 parts ascorbate to one part manganese. The compound's own composition predicts 6.4 — close enough that the patent is plainly working from the salt, and far enough apart that we will give you both numbers rather than pretend they are identical. Hold on to that ratio — it is what lets us work out something the dog tub will not tell you.

The patent expired long ago. We cannot show you a document proving that no manufacturer has since re-derived the manganese amount for dogs, so take this as the inference it is. What we can show you is narrower, and it is checkable. No published trial has ever tested the manganese separately. And on the one product line where the amount can be tracked across those thirty years, it did not hold steady — it fell, and you have to be careful about the basis or the fall looks smaller than it is. The patent's human and small-animal formulation lists 5 mg of manganese per capsule, at one to six capsules a day. Nutramax's current human Cosamin DS declares 1 mg in a three-capsule daily serving, which is 0.33 mg a capsule. Per capsule that is a fifteenfold drop; per day it is 5 to 30 mg then against 1 mg now. Nobody publishes what moved it, and nobody publishes what set the dog tablet's 3 mg.

How much manganese is your dog actually getting?#

Here is the arithmetic, with every input named so you can check it.

Cosequin's own dosing table for a dog over 60 lb is three tablets a day for the first four to six weeks, then one to two tablets after that.^[1] At 3 mg a tablet that is 9 mg a day on the loading dose and 3 to 6 mg a day after it.

For comparison, AAFCO's Dog Food Nutrient Profiles set the adult-maintenance minimum for manganese at 5.0 mg per kg of dry matter, equivalently 1.25 mg per 1,000 kcal of metabolisable energy. AAFCO's own note on that figure records that the minimum "was not changed from the previous amount of 5.0 mg/kg DM, the amount being slightly more than the 2006 NRC RA of 4.8 and slightly less than the FEDIAF Guidelines of 5.6 mg/kg DM."^[5] For a 70 lb (32 kg) dog, the pet-dog equation from a meta-analysis gives a maintenance requirement of about 1,790 kcal a day — and note that this is the paper's pet-dog subgroup equation, drawn from 16 treatment groups, 23% of the 70 it pooled, rather than its headline all-dogs equation.^[6] So a food sitting exactly on that floor delivers about 2.2 mg of manganese a day. Put the other way round, and with no estimate needed at all: for a dog to get 9 mg from food at the AAFCO minimum, he would have to eat 7,200 kcal a day.

That last row is the one worth carrying away. Nutramax's human product, Cosamin DS, declares 1 mg of manganese and 4 mg of vitamin C in a complete three-capsule daily serving.^[7] One dog tablet carries three times what a whole daily serving of the human product does. A dog over 60 lb on the loading dose is taking nine times that serving.

And here is the comparison that cuts the other way, which we would rather print than leave for someone to find. Human intakes are far better characterised than canine ones: the NIH puts adult manganese Adequate Intake at 2.3 mg a day for men and 1.8 mg for women, with a Tolerable Upper Intake Level of 11 mg a day.^[8] So the dog's 9 mg loading dose is roughly four times a typical human dietary intake and still sits below the human ceiling, in a considerably smaller animal — the 11 mg is a flat adult figure, not a per-kilogram one, so a 32 kg dog reaching 9 mg is closer to that ceiling per unit of body weight than the raw numbers suggest. We are not going to put a number on how much closer, because scaling a human UL to a dog by body weight is not a validated thing to do. That is not a canine limit and it is not evidence of safety in dogs. It is the nearest published ceiling that exists for this element, and it is not an alarming place to be standing.

Two caveats. First, most dog foods sit well above the AAFCO minimum and very few publish their manganese content, so we cannot tell you what your bag contains — if yours does publish it, add the two numbers, and that is the method rather than a number we can give you. Second, none of the above is a toxicity claim. It is a statement about which container in your kitchen is the dominant source, and the answer during loading is the tub.

dogorthopedic.com

Is there an upper limit? The blank is the finding#

AAFCO's dog profile is the clearest place to see the gap. In the minerals block, calcium, phosphorus, iodine and selenium all carry a figure in the maximum column. Manganese carries none — a minimum, and then nothing.^[5] The absence of a ceiling is not permission. It means the study that would set one has not been done in this species.

The peer-reviewed literature says the same thing in its own words, and it is short enough to give you in full. As of today it is three reports covering five dogs, all of them mass accidental ingestions.

Borchers opens by stating the gap directly: manganese "is a ubiquitous, essential trace element and a common ingredient of joint supplement tablets," and "little information is known about the inherent toxic potential if ingested at higher doses." Jaffey, writing ten years later, notes that the whole veterinary literature at that point was "only 2 reports containing 3 dogs."

Two things fall out of that table, and they point in opposite directions, so here are both.

It is not uniformly fatal. Three of the five dogs recovered, and Wolf 2023 is specifically the first published account of manganese toxicosis in a dog being successfully managed. Which is to say: anything written on this subject before 2023 was describing a canine record that consisted of one Pug, euthanised.

And the doses involved are nowhere near a label dose. Jaffey's two dogs took 2,062 mg of manganese between them — call it a thousand milligrams each. A 70 lb dog on Cosequin's loading dose takes 9 mg a day. One dog in that report ate something over a hundred days of loading dose in a sitting. Note also that Jaffey's product was a human supplement, and not the kind of human joint supplement discussed above: 2,062 mg across roughly 75 capsules works out at about 27 mg of manganese per capsule — our arithmetic from the report's own figures, not a number it prints. Nutramax's human Cosamin DS declares 1 mg in a full daily serving. "Human supplement" spans both, which is exactly why the milligrams on the panel matter more than the species on the box.

So the honest read is narrow. Nothing here is evidence that the label dose harms a dog. It is evidence that the container is worth keeping shut, and that no one has established where the ceiling is.

The practical consequence is small and worth doing anyway. The loading dose is a four-to-six week dose. It is not a permanent dose, and nothing on the bottle marks the date it should end. The only prompt in the system is the dosing table, and only if you go back and read it. The wider joint-supplement overdose literature — the cases where the glucosamine load rather than the manganese was implicated — is set out in our piece on giving Dasuquin and Cosequin together.

The vitamin C the label never prints#

The manganese cannot arrive without the ascorbate, because it arrives as a salt of it — and ascorbate is about 86% of that salt by mass. So how much vitamin C is in a tablet? Nutramax does not say — the dog panel has no vitamin C row, verified absent again on 1 September 2026.

Two routes give the same answer, though they are not independent of each other — the second is a label printed from the first:

• The compound's own composition. 3 mg of manganese implies about 22 mg of manganese ascorbate, of which roughly 19 mg is ascorbate.^[2]
• The patent's own paired rows, 5 mg manganese to 33 mg ascorbate — a ratio of 6.6, which applied to 3 mg gives about 20 mg.^[3]

A third figure disagrees, and it is the maker's own. Nutramax's current Cosamin DS panel prints 4 mg of vitamin C against 1 mg of manganese — a ratio of 4, not 6.6 — which applied to 3 mg would give about 12 mg.^[7] A third-party retailer page reproducing an earlier version of the same label, read the same day, prints 6 mg against 1 mg — which does fit. It is a retailer's reproduction, not a Nutramax page, and we have not been able to confirm it against the maker.^[12] We could not reach several other retailers carrying this label on 31 August, so treat that as a single data point rather than a survey.

So the honest answer is 12 to 20 mg a tablet, and we think the top of that range is the better estimate. The chemistry and the patent agree at 6.6 — and they are not independent of one another, so on their own they would be one piece of evidence rather than two. What makes the difference is that a retailer's reproduction of an earlier version of the same human panel lands there too, at 6 mg to 1 mg. We cannot confirm from Nutramax that this was ever its own printing — it is a third-party page, and we say so — but it lands at a ratio of 6 by a route that runs through neither the chemistry nor the patent. It is still a reproduction of Nutramax's own earlier printing, so treat it as a weak third data point rather than an outside check. The only thing pointing at 12 is the current panel, which disagrees with its own patent and with the composition of its own ingredient. For a big dog on the loading dose, that is somewhere between 36 and 60 mg a day. It is a calculation rather than a printed figure, and we have labelled it as one.

For almost every dog, this is a non-event, and the reason is the interesting part. Dogs are not one of the species that lost the ability to make vitamin C. The losses are well mapped: they occurred in teleost fishes, anthropoid primates, guinea pigs and some bat and Passeriformes bird species, and in every case studied the cause is mutation of the L-gulono-γ-lactone oxidase gene that catalyses the last step of ascorbate synthesis. Dogs are not on that list.^[13] The regulatory paperwork follows: search AAFCO's Dog and Cat Food Nutrient Profiles for "ascorbic" and you get zero hits. The dog's vitamin list runs A, D, E, thiamine, riboflavin, pantothenic acid, niacin, pyridoxine, folic acid, B12 and choline. There is no vitamin C row at all, because a complete and balanced dog food is not required to contain any.^[5]

So the second active ingredient in the patent — ascorbic acid, chosen for collagen synthesis — is, in dogs, being supplied to an animal that was already making it.

“Because your dog can't tell you how much pain he or she is in.”

The one group of dogs for whom the ascorbate is worth a sentence#

There is one real, narrow exception.

For dogs who form calcium oxalate bladder stones, dietary vitamin C is a named consideration. The Ontario Veterinary College's pet nutrition team, writing on nutritional management of canine bladder stones, states it plainly, and without offering a mechanism: "Excessive intake of calcium and oxalic acid, as well as the supplementation of vitamin C should be avoided."^[14]

A 2026 clinical review in Today's Veterinary Practice, by two ACVIM diplomates, puts the breed picture like this: "Inherited tendencies play a significant role, and breeds such as the miniature schnauzer, bichon frise, and Yorkshire terrier are consistently overrepresented." Note the "such as" — that is a list of the most overrepresented breeds, not a closed one.^[15]

And now the part that keeps this proportionate, because it argues against the point we just made. That same clinical review does not mention vitamin C anywhere — not as a risk factor, not as a dietary consideration. Two credible veterinary sources, one advising against supplementation and one not considering it worth naming. We are publishing the disagreement rather than resolving it silently.

What to take from that: if your dog has formed a calcium oxalate stone, mention the manganese ascorbate to your vet, because it is a source of ascorbate you would otherwise not think to declare and it is nowhere on the tub. If he has not, this is not something to act on.

Does the manganese actually do anything? What the two trials tested#

The patent gives a mechanism. A mechanism is not a result. So here is every human and canine trial of this specific three-part combination that we could find, read at source.

In people. Leffler and colleagues ran a 16-week randomised, double-blind, placebo-controlled crossover trial in Military Medicine in 1999, giving glucosamine HCl 1,500 mg/day, chondroitin sulfate 1,200 mg/day and manganese ascorbate 228 mg/day. Thirty-four men from the US Navy diving and special warfare community with chronic pain and radiographic degenerative joint disease were randomised. Knee symptoms improved: summary disease score −16.3% (p = 0.05), physical examination score −43.3% (p = 0.01). Running times did not change. The authors were explicit that the study "neither demonstrated, nor excluded, a benefit for spinal DJD," and called for a larger data set.^[16]

Note that dose. 228 mg a day of manganese ascorbate — the salt, which is how the trial and the patent both report it. Modern labels report the element instead, and the two are not the same unit, which is the single most confusing thing on this whole topic. Converting with the patent's own 6.6 ratio, today's retail Cosamin DS at 1 mg of manganese a day is about 7 mg of the salt a day — roughly one-thirtieth of what the trial gave. And the patent is worth putting beside those two on the same daily basis: it specifies 38 mg of the salt per capsule for humans and small mammals, at one to six capsules a day, so its own top dose is 228 mg a day. That is exactly the trial's dose. The patent and the trial agree with each other; it is the product on the shelf that has moved away from both. We flag that conversion as ours; the point of publishing it is that the product carrying the trial's name no longer carries anything close to the trial's dose of this ingredient, and no label anywhere puts those two numbers side by side.

Run that conversion the other way and it produces the number this article would be accused of hiding, so here it is. The trial's 228 mg of the salt is roughly 31 mg of elemental manganese a day — about three times the 11 mg human upper limit quoted above — given in a trial that randomised 34 men, with the authors reporting side-effect frequency similar to baseline and no haematologic effects. Be careful with the exposure, because the abstract does not give it: the sixteen weeks is the whole crossover, split between active and placebo periods, and how long any one man was on the compound is not reported.^[16] That is a small, short pilot and not a safety study, and it does not license anything in a dog. But it is the largest documented human exposure to this specific ingredient we could find, it ran well over the ceiling, and it is against our own framing, so it belongs on the page rather than off it.

In dogs. Johnson and colleagues published the one canine study of this combination in Osteoarthritis and Cartilage in 2001. Sixteen adult dogs underwent unilateral cruciate ligament transection and were split four ways — groups of three, three, five and five. Synovial fluid was sampled at 0, 1, 3 and 5 months and assayed for the chondroitin sulfate epitopes 3B3 and 7D4. Treated dogs showed significantly elevated 3B3 (p = 0.029), 7D4 (p = 0.036) and 7D4/GAG (p = 0.007) versus controls at three months. The 7D4 and 7D4/GAG elevations also "remained significantly elevated (P=0.012)… over the 5 month period," so the signal was not a single time point. But — and this is the sentence the marketing never quotes — "when epitope concentrations were expressed as a ratio of CCL-transected to contralateral non-operated knee, treatment effect of CS–G–M was no longer significant."^[17]

So the canine evidence for this exact mixture is: sixteen surgically induced dogs, a biomarker rather than a limp, groups as small as three, and an effect that vanished under the study's own normalised analysis. The authors' conclusion was appropriately careful — these compounds "may act to modulate articular cartilage matrix metabolism in vivo."

And the finding that matters most is a negative one. In neither study, nor in any other we could locate, was the formula compared with manganese against the same formula without it. Nothing in the evidence base attributes any part of any result to the manganese. Its presence in the formula traces to a patent; its amount on a dog panel traces to nothing either maker or literature publishes.

That is worth holding next to what we already published about the sibling product: in the one blinded, placebo-controlled canine force-plate trial of Dasuquin with MSM, the formula found no significant advantage over placebo^[18] — and Dasuquin with MSM contains no manganese at all.^[19] The full comparison is in Dasuquin vs Cosequin.

What we don't know, stated plainly#

• Whether the manganese contributes anything. Never isolated in a trial. Not once.
• How much manganese your dog gets from food. Most brands do not publish it. If yours does, add it to the tub's figure yourself.
• What a safe upper limit would be. AAFCO sets no maximum for manganese in dog food, and the case literature is five dogs. The study that would set a ceiling has not been done.
• The exact ascorbate content of a dog tablet. Bracketed at 12–20 mg by three methods that do not agree, published by nobody.
• Whether any of it changes how a dog walks. That question is answered by measuring your own dog, which is what is my dog's limp getting better is for.

So should you stop giving it?#

Not because of the manganese. Possibly for a different reason.

Nothing here says the tub is harmful at label doses. What it says is that one row on the panel got its amount from a patent rather than from a canine dose study, has never been tested in dogs on its own, and is delivered at its highest during exactly the six weeks that nothing prompts you to end.

The reason to stop is the ordinary one, and it applies to every product in this aisle including the ones we card. Give it six weeks against something you measured beforehand, and if it did not earn its place, stop buying it. The canine evidence for this specific mixture is a sixteen-dog biomarker study whose effect disappeared under its own normalised analysis; the sibling formula came up null on a force plate. That is not a body of evidence that entitles anything here to a standing order. Your own dog, scored on the same walk before and after, is better evidence than either study, and it is free.

With that said, the practical advice is unglamorous:

Step down on schedule. Write the date six weeks from the day you started on the lid. Over 60 lb, that means going from three tablets to one or two. That single act cuts the manganese by a third to two-thirds and costs you nothing — it saves money.

Read the panel before you add a second tub. Dasuquin with MSM contains no manganese;^[19] Cosequin has 3 mg a tablet; Dasuquin Advanced has 1 mg or 3 mg a chew depending on size, at two large chews a day for a dog over 60 lb during loading.^[20] Full accounting in can I give Dasuquin and Cosequin together.

If you are continuing, buy the right bottle size — and check which line it is. A dog over 60 lb eats three tablets a day for six weeks, which is 126 tablets, so the 75-count we link is 25 days of loading — $24.99, or about $0.333 a tablet, verified on Amazon on 1 September 2026.^[21] The cheapest per-tablet route Nutramax publishes is the 250-count of the plain-MSM line, which our own product page has recorded at about $0.24 a tablet; it sells direct from store.cosequin.com, on which we earn nothing. We did not re-verify the direct price today, so compare it yourself before you buy. Note the tradeoff honestly: the plain-MSM tablet has the same 600/300/250 base and the same 3 mg of manganese, but not the 6 mg of sodium hyaluronate.^[22] One caution from today's check: an Amazon listing titled "MSM+HA, Chew Tablets, 250ct" was selling at $59.99 on 1 September 2026 — the same $0.24 a tablet —^[23] while cosequin.com lists the with-HA tablet only at 60, 75, 120, 150 and 180 counts.^[22] Our own product page, from an earlier check, records the Amazon 250-count at $62.99 and $59.99 as the direct price — so the listing has moved as well as being hard to place. We cannot reconcile those from published pages, so read the panel on the bottle when it arrives.

And do the free thing first. In a five-year study of 48 Labrador Retrievers, littermates fed 75% of what their pairmates ate had less frequent and less severe radiographic hip osteoarthritis at five years.^[24] That is a controlled canine trial running five years with an intervention anyone can perform, which is more than this entire supplement aisle can say — and to hold it to the standard we just held Johnson 2001 to, its endpoint is radiographic, a structural read rather than a limp. It costs nothing.

The reason this product carries the article is not that we think the manganese earns its place — this whole article says nobody has shown that it does. It is that Cosequin prints the row. Three milligrams, named as manganese ascorbate, on a public page with a dosing table beside it. That is what let us do this arithmetic at all. You cannot audit a panel that is not printed, and this one is. Its real weaknesses are elsewhere and we have listed them: Nutramax publishes no inactive-ingredient list on the product's own page — we previously printed one sourced to retailer reproductions and have since marked it unverified — so a dog on a strict elimination diet cannot be cleared for it, and no calorie figure exists on any surface we can find.

If having read all of this you would simply rather not add manganese at all, the retail Dasuquin with MSM tub is the answer, and we should be clear that this is a preference rather than a finding — it is a different formula with ASU, Boswellia and green tea, not a manganese-free version of the same thing, it costs more, and it is the formula that came up null on a force plate. Said plainly: if the manganese is your only reason to switch, this article has not given you one, and buying neither tub is a defensible reading of everything above.

dogorthopedic.com

What to do this week#

Take the tub off the shelf and do three things that together take five minutes.

Find the manganese row and note the milligrams. Find the dosing table and check whether you are still on the initial four-to-six week dose months after you started — if you are, step down today. Then write on the lid the date you started and the date the loading dose was supposed to end, so the next tub does not repeat it.

That is the whole intervention, and it is worth doing for a reason that has nothing to do with manganese. A supplement you cannot describe the dose of is a supplement you cannot evaluate, and one you cannot evaluate is one you will keep buying for years without ever knowing whether it did anything. The delay that costs a dog the most is rarely a diagnosis nobody made. It is the months spent on a plan nobody was measuring.

If you want to set that measurement up properly before the next six weeks go by, start with how many joint chews your dog actually needs and then dosing by weight.