Chocolate poisoning in dogs: where the thresholds actually come from, and why no calculator fits an Indian wrapper

How much chocolate is toxic to a dog, and where that number comes from

Chocolate is toxic to a dog at about 20 mg of methylxanthines per kilogram of body weight for mild signs, 40 to 50 mg/kg for cardiac effects and 60 mg/kg and above for seizures. Those three numbers run through every veterinary reference, every online calculator and every clinic's phone script, and almost nobody states where they came from.

They came from one page. Sharon Gwaltney-Brant, DVM, PhD, then at the ASPCA Animal Poison Control Center, wrote a toxicology brief for Veterinary Medicine in February 2001 which says: "Based on ASPCA Animal Poison Control Center (APCC) experience, mild signs occur in animals ingesting 20 mg/kg of theobromine and caffeine, severe signs are seen at 40-50 mg/kg, and seizures occur at 60 mg/kg", and attributes that sentence to "ASPCA/APCC Database: Unpublished data";.

Read the attribution rather than the numbers. The thresholds are twenty-five years of accumulated poison-centre case experience, written down by the clinician who held it. They are not the output of a dose-response study, there is no denominator behind them, and no confidence interval has ever been published around them. The Merck Veterinary Manual reprints all three figures today, in a page fully reviewed in September 2024 by Cristine Hayes, DVM, of the same ASPCA Animal Poison Control Center.

This is not a reason to ignore them. A threshold distilled from tens of thousands of real exposures is better evidence than most things in small-animal toxicology. It is a reason to understand what kind of number it is: a conservative line at which a poison centre starts paying attention, not a prediction that your dog will be ill.

The thresholds are an action line, not a prediction

The best test of that reading comes from the only large case series published. In a retrospective study of 156 chocolate ingestions in dogs, 112 dogs had no clinical signs at all and 44 did. In the 112 that were entirely well, the median calculated worst-case theobromine dose was 22.4 mg/kg. The median dog that sailed through without a symptom had, on the study's own arithmetic, eaten more than the published mild-signs threshold.

The dogs that were unwell had eaten roughly three times as much: a median calculated maximum of 70.8 mg/kg of theobromine and 21.7 mg/kg of caffeine. So the thresholds do separate the two groups, just a long way above where the 20 mg/kg line sits.

Theobromine and caffeine are two toxins counted as one dose

Theobromine and caffeine are both methylxanthines, both are in chocolate, and the published thresholds apply to the two of them added together rather than to theobromine alone. Getting that wrong is the commonest arithmetic error on this subject, and it runs in both directions.

Chocolate holds three to ten times as much theobromine as caffeine, and both contribute. Both work the same way: they competitively inhibit cellular adenosine receptors, which produces central nervous system stimulation, diuresis and tachycardia; they raise intracellular calcium, which increases the strength and contractility of skeletal and cardiac muscle; they compete for benzodiazepine receptors in the central nervous system, inhibit phosphodiesterase, and may raise circulating adrenaline and noradrenaline.

Gwaltney-Brant's instruction was explicit: "Add the theobromine and caffeine amounts to determine the total amount of methylxanthine ingested";. Her worked example does exactly that, taking dark chocolate at 130 mg of theobromine plus 20 mg of caffeine per ounce and treating the sum as the dose.

The conflation you will meet within three clicks

Merck's table is headed "Methylxanthine Content of Various Types of Chocolate" and gives semisweet and sweet dark chocolate as 5.3 to 5.6 mg/g. That is theobromine plus caffeine. Veterinary Partner, revised in August 2025, says "Milk chocolate contains 44 mg per ounce of theobromine, semisweet chocolate contains 150 mg per ounce, and baking chocolate contains 390 mg per ounce";. The 150 figure is Gwaltney-Brant's combined methylxanthine number for dark chocolate, 130 plus 20, presented as theobromine alone.

The two pages also disagree about milk chocolate by a third. Gwaltney-Brant's table gives 58 mg of theobromine per ounce; Veterinary Partner gives 44 Brooks W, Chocolate Toxicity in Dogs. Neither page is careless. They are drawing on different compilations of a product that genuinely varies, which is the whole problem in one sentence.

For an owner on the phone, the practical consequence is small and worth knowing anyway. Give the clinic the product, the quantity and your dog's weight, and let it pick the figures. Do not compute a number from one page and a threshold from another, because you will be dividing a theobromine value into a methylxanthine threshold or the reverse.

Every calculator in the world runs on one table from October 2000

The methylxanthine figures in the Merck Veterinary Manual, in the online calculators and in almost every chocolate-toxicity page are the same table, and that table names its own sources at the bottom: two veterinary texts and a Hershey Foods Corporation consumer nutrition page retrieved in October 2000.

That matters less for the shape of the numbers than for their reach. The ranking is certainly right: cocoa powder is the most concentrated thing in a kitchen, baker's chocolate next, dark next, milk a long way below, white chocolate negligible. What a twenty-five-year-old American confectioner's nutrition page cannot tell you is what a bar made in Gujarat or a tin bought in Kochi contains.

There is also a row that got lost on the way. Gwaltney-Brant's table carried both "Dry cocoa powder" at 737 mg of theobromine per ounce and "Instant cocoa powder" at 136 mg per ounce. Converted, that is 28.5 mg/g of methylxanthines against 5.3 mg/g, a difference of more than five times. Merck's modern table keeps the dry cocoa powder row and has no instant cocoa powder row at all.

An owner whose dog has got into a tin of sweetened drinking chocolate and looks up "cocoa powder" therefore multiplies by roughly five more than the product deserves. That is the single most common over-estimate in this subject, and in India it is the likeliest exposure in the house, because a tin of malted or sweetened cocoa drink sits in a cupboard where the bar sits in a fridge.

What Indian law requires a chocolate to contain, and why the label hides the toxin

Indian chocolate is governed by standard 2.7.4 of the Food Products Standards, and it sets its minimums in a way that separates the two halves of cocoa. Theobromine lives in only one of them, and that is the half the front of the pack never prints.

Cocoa solids come in two parts. Cocoa butter is the fat, and it carries essentially no alkaloid. Fat-free cocoa solids, also written as non-fat cocoa solids, are the brown part, and that is where the theobromine is. The relationship is so reliable that public analysts measure theobromine in order to work out how much fat-free cocoa a product contains, then multiply by a conversion factor, which is how cocoa-content declarations are enforced in the first place.

Now read the Indian standard with that in mind. Dark chocolate in India "shall contain, on a dry matter basis, not less than 35 per cent. total cocoa solids, of which not less than 18 per cent. shall be cocoa butter and not less than 14 per cent. fat-free cocoa solids". Couverture chocolate "shall contain, on a dry matter basis, not less than 35 per cent total cocoa solids of which not less than 31 per cent shall be cocoa butter and not less than 2.5 per cent fat-free cocoa solids".

Those two products carry the same headline number and differ by five and a half times in the fraction that carries the toxin. Both are legal, both may be described as 35 per cent cocoa, and nothing on either wrapper distinguishes them to a frightened owner reading a label at eleven at night.

Turning an Indian label into milligrams, with the arithmetic shown

A survey of 191 cocoa liquor samples from around the world, run for the UK Food Standards Agency and reported by Alan Richards and Brian Wailes in 2012, gives the conversion as fat free dry cocoa in g per 100 g equals theobromine in g per 100 g multiplied by 40.7. Run that backwards and a product's theobromine is its fat-free cocoa solids divided by 40.7.

1. Take the fat-free cocoa solids figure. For Indian dark chocolate at the legal floor that is 14 per cent of dry matter.
2. Divide by 40.7. Fourteen divided by 40.7 is 0.34 per cent theobromine, which is 3.4 mg per gram.
3. Repeat for couverture at its floor of 2.5 per cent fat-free cocoa solids. Two point five divided by 40.7 is 0.061 per cent, or 0.61 mg per gram.
4. Multiply by the grams eaten and divide by your dog's weight in kilograms, which gives the dose per kilogram the thresholds are written in.

Treat that as a sense of scale rather than a calculator, for three reasons the authors state themselves. The conversion factor came out at 40.7 against the 35.9 enforcement laboratories had been using, so the published factor has itself moved. The 191 samples ranged from 19,803 to 39,168 mg/kg of theobromine, a twofold spread, and the authors warn that "considerable caution must be taken when reporting on cocoa content of foods when the composition of the cocoa liquor used in their preparation is unknown". And their sampling plan was deliberately weighted towards African cocoa, with the note that "where the country of origin of the sample is known, then alternative, location-specific, factors may be more appropriate";.

Three more things the Indian standard allows that a foreign page assumes away

• Milk chocolate in India has no total cocoa solids minimum. The standard's own table sets cocoa solids on a moisture-free and fat-free basis at 2.5 per cent for milk chocolate against 12 per cent for plain chocolate, and total fat at 25 per cent for both. The Codex standard, by contrast, requires milk chocolate to carry not less than 25 per cent total cocoa solids. The bases are different and cannot be compared digit for digit, which is the point: an Indian milk chocolate is governed by a different rule from the one a borrowed calculator assumes.
• White chocolate is defined as cocoa butter, milk solids and sugar, with a dash in the cocoa solids row, so the standard sets it no fat-free cocoa requirement at all. That is consistent with the measured 0.04 mg/g.
• Vegetable fat other than cocoa butter may be added up to 5 per cent of the finished product, and when it is, the pack must declare in bold "CONTAINS COCOA BUTTER EQUIVALENT / VEGETABLE FAT IN ADDITION TO COCOA BUTTER";. The permitted fats include sal, kokum, mango kernel, mahua, dhupa and phulwara, which is a thoroughly Indian list and tells you nothing about theobromine, because none of them carries any.

Cocoa powder, drinking chocolate and the jar at dog height

Cocoa powder is the most concentrated methylxanthine source in an ordinary kitchen at 28.5 mg/g, and the Indian standard splits it into three grades whose order will surprise you: the lower the fat, the more theobromine per gram.

Indian cocoa powder must carry not less than 20 per cent cocoa butter on a dry basis. Fat reduced cocoa powder, labelled medium fat, carries 10 to 20 per cent. Highly fat reduced cocoa powder, labelled low fat, carries not more than 10 per cent.

Removing fat does not remove theobromine; it removes the part that was diluting it. Gram for gram, a low-fat Indian cocoa powder is the most concentrated methylxanthine source a household is likely to own, and it is sold as the healthier option. Nothing on the tin says so, and no chocolate calculator has a row for it.

Drinking chocolate is a quarter to a third cocoa, not cocoa

Drinking chocolate is a separate Indian standard, 2.7.6, covering dry mixtures of cocoa and sugars intended for direct consumption. A product may be called Sweetened Cocoa, Sweetened Cocoa Powder or Drinking Chocolate when its cocoa powder content is not less than 25 per cent by mass and the cocoa butter is 20 per cent or more. Chocolate Powder is defined separately as a mixture of cocoa powder and sugars containing not less than 32 per cent cocoa powder. Below 20 per cent cocoa powder, the permitted name drops to Sweetened Cocoa-flavoured Mix.

So a tin of Indian drinking chocolate is between a fifth and a third cocoa powder by mass, which puts it in the region of the instant cocoa powder row that Merck's table no longer carries. A dog that eats fifty grams of drinking chocolate has not eaten fifty grams of cocoa powder, and saying so on the telephone is worth doing, because the clinic's default assumption will otherwise be the worst case.

Two Indian storage habits follow from this and are worth acting on today. Cocoa powder, drinking chocolate and baking slabs live in jars and tins in a cupboard, not in sealed wrappers, because a humid kitchen through the monsoon cakes anything left open. An unwrapped, unmeasured, highly concentrated powder is the worst combination on this page, and it is the one nobody secures. The bar in the fridge is not the problem.

Nobody has measured theobromine in Indian chocolate

No published measurement of theobromine in chocolate sold in India exists that we could find on 22 October 2026, and the honest consequence is that every figure on this page, including ours, is foreign data applied to an Indian pack.

This is not for want of Indian cocoa. The primary growing regions of cocoa in India are Kerala, Andhra Pradesh, Karnataka and Tamil Nadu, and the estimated area under cocoa during 2020 to 2021 was 1,03,376 hectares producing 27,072 tonnes. India grows the crop, processes it and eats it. What it has not done, as far as the literature goes, is measure the alkaloid in the finished bar.

Measurements exist elsewhere and they are useful as a yardstick. A 2023 study analysed twenty commercial chocolates in Ankara, forty samples in all, by ultraviolet-visible spectrophotometry. Bittersweet dark chocolate labelled 60 to 80 per cent cocoa measured 8.19 mg/g of theobromine, semisweet dark labelled 45 to 50 per cent measured 6.40 mg/g, milk chocolate measured 2.75 mg/g and chocolate spreads 2.72 mg/g, with an overall range across all samples of 1.9 to 9.6 mg/g.

Set those against the table every reference work reprints and something uncomfortable appears. Merck gives semisweet and sweet dark chocolate as 5.3 to 5.6 mg/g of theobromine plus caffeine. The Ankara samples measured 6.40 mg/g of theobromine alone. Merck gives milk chocolate as 2.3 mg/g of methylxanthines; the measured theobromine alone was 2.75 mg/g. On these samples the canonical table under-reads modern chocolate rather than over-reading it, which is the direction that matters.

An eighty per cent bar with less theobromine than a seventy per cent bar

The same study tested whether the number on the front of the pack predicts the alkaloid behind it. Across all dark chocolates the correlation was strong and significant, with r of 0.823. Within the semisweet group alone it was not significant, at r of 0.523 and p of 0.081.

The individual samples show why. The bar labelled 80 per cent cocoa measured 0.89 per cent theobromine. The bar labelled 70 per cent measured 0.96 per cent. A reader who assumed the higher number meant the higher dose would have had it backwards on that pair.

We would rather publish this gap than paper over it. If an Indian food-science department has measured methylxanthines in chocolate sold here, that dataset would improve this page more than anything else on it.

Onset, half-life, and how long a dog has to be watched

Signs of chocolate toxicosis usually appear 6 to 12 hours after ingestion, and in severe cases they persist for up to 72 hours, because theobromine's half-life in the dog is 17.5 hours against 4.5 hours for caffeine.

Two mechanisms make the drug outstay its welcome. Theobromine and caffeine are readily absorbed from the gastrointestinal tract and widely distributed, and after metabolism in the liver they undergo enterohepatic recycling, which means a portion is excreted into bile, delivered back to the gut and absorbed again. Methylxanthines are then excreted in urine as metabolites and as unchanged parent compounds, and they can be reabsorbed from the bladder, which is why a clinic places a urinary catheter in a non-ambulatory patient specifically to minimise that reabsorption.

A dog is therefore not simply waiting for a dose to wear off. It is recycling it. A dog metabolises theobromine far more slowly than a person does, which is the whole of the species difference: the same bar that a human clears in an afternoon keeps working on a dog for most of a second day.

Three sources, three different endpoints

Merck says clinical signs may persist for up to 72 hours in severe cases. Gwaltney-Brant said the same in 2001, citing the long half-life. Veterinary Partner says "It takes nearly four days for the effects of chocolate to work its way out of a dog's system";. The BMJ letter puts the onset earlier and the recovery in the same place, saying most symptoms begin within two hours but can take as long as 24 hours to appear, and up to three days for recovery.

The disagreement is three days against four, and it is not resolvable from the published material. Plan for the longer one. A watch that ends on the evening of day two because a page said 72 hours is a watch that ends in the middle of the range.

One delay is specific to the way chocolate reaches a dog. Gwaltney-Brant noted that large quantities of chocolate can coalesce in the stomach into a mass that is not easily vomited or lavaged, and that where a dog has eaten wrapped sweets the effects of methylxanthines may be delayed by several hours or days. A box of individually wrapped chocolates, which is what arrives in an Indian house at Diwali and at Christmas, is the exposure most likely to surprise everyone on day two.

What actually happens, tier by tier

Chocolate toxicosis runs gastrointestinal first, then neurological and cardiovascular together, and the symptoms at each tier are well described. The frequencies below come from the 44 dogs in the 156-dog series that actually had signs, which is a better guide than a list with no denominator.

Initial signs may include polydipsia, vomiting, diarrhoea, abdominal distention and restlessness. These may progress to hyperactivity, polyuria, ataxia, rigidity, tremors and seizures. Tachycardia, premature ventricular contractions, tachypnoea, cyanosis, hypertension, hyperthermia, bradycardia, hypotension or coma may occur, and hypokalaemia contributing to cardiac dysfunction may appear late in the course.

That is the complete list of what can happen. The table below is what one clinic actually recorded.

Read the first row twice. The commonest sign of chocolate poisoning in dogs is not vomiting. It is restlessness, in three out of four affected dogs, and restlessness is the easiest thing in the world to explain away as a dog being a dog on a festival evening.

Death is generally due to cardiac arrhythmias, hyperthermia or respiratory failure. If your dog already has a diagnosed heart condition, say so on the first call, because the cardiac tier is where its margin is smallest. Our pages on mitral valve disease in dogs and on epilepsy in dogs cover the two pre-existing conditions that change this conversation most.

How often chocolate is actually fatal to a dog

Chocolate is rarely fatal to a dog that reaches a clinic. In the largest published series, 156 chocolate ingestions produced one death, and 112 of the 156 dogs had no clinical signs whatsoever.

C Weingart, A Hartmann and B Kohn reviewed the records of dogs presenting to a single institution over four years after chocolate ingestion, including only dogs whose owners had directly observed them eating it British Small Animal Veterinary…. Of 156 dogs, 112 presented without clinical signs and 44 with them. Twenty-eight of those 44 had eaten dark or bitter chocolate. After decontamination and symptomatic treatment, 43 of the 44 survived, and the authors describe the prognosis as good with a mortality rate of less than 3 per cent.

In the 112 dogs with no signs, physical examination was normal in every case, and where laboratory work was done it showed no abnormalities at all. Fifty-four per cent of them had eaten milk chocolate. Among the 44 that were unwell, 64 per cent had eaten dark or bitter chocolate. Type did most of the sorting.

The one dog that died, and what it had eaten

The single death was a dog that had eaten 100 g of dark chocolate twelve hours before it was presented, giving a calculated theobromine dose of 64 mg/kg and caffeine of 19.7 mg/kg. It arrived with pronounced sinus tachycardia, mild hyperthermia, vomiting and a seizure, and died four hours after symptomatic treatment began.

Twelve hours. That is the part to carry away. A calculated dose of 64 mg/kg sits only just above the published seizure threshold of 60 mg/kg, below the median of 70.8 mg/kg for the symptomatic group as a whole, and 43 of those 44 dogs lived. What distinguished this case was the delay rather than the dose.

Three honest limits on the reassurance

• It is one institution over four years, not a national dataset, and every dog in it was brought to a veterinary hospital. Dogs that died at home are not in the denominator.
• The methylxanthine dose was calculated rather than measured, and where the proportions of a mixed box were unknown the authors used the chocolate with the highest methylxanthine content, which they note may have overestimated some cases. The 22.4 mg/kg median for the symptom-free dogs is therefore a worst-case figure, and the true doses were probably lower.
• No Indian series exists. We searched on 22 October 2026 and found German, British, Czech and Turkish material and no Indian denominator for canine chocolate exposure at all, so nothing here is an Indian figure.

What to do, and what this page deliberately will not tell you

Telephone a veterinary surgeon with your dog's weight in kilograms, the wrapper, the amount missing and the clock time. Everything after that is a clinical decision, and two of those decisions are ones this page will not make for you.

Working out the dose from those four numbers is a method, and it is set out step by step on our page on foods toxic to dogs, along with what to take to the clinic and the finding that India has no 24-hour animal poison helpline. Read that page for the first hour. This one is the toxicology.

Do not induce vomiting. Whether to empty a dog's stomach is a decision a veterinary surgeon makes for that exposure, in that dog, in that hour, and the same sibling page sets out why, including the five different hydrogen peroxide doses published by five reputable sources and the fact that the bottle an Indian chemist sells is twice the strength all of them assume. Merck's own chocolate protocol considers emesis only in a patient that remains clinically normal and only with a recent exposure, within about two hours. That is a clinic's judgement, not a kitchen's.

Activated charcoal, where the reference work reversed itself

In 2001, Gwaltney-Brant advised that because methylxanthines undergo enterohepatic recirculation, "giving repeated doses of activated charcoal is usually beneficial in symptomatic animals";. Merck in 2024 says the opposite for chocolate specifically. Chocolate exposure causes dehydration through the diuretic effect of methylxanthines and its own sugar load, electrolyte abnormalities such as hypernatraemia have occurred with activated charcoal particularly in chocolate cases, and "A single low dose of activated charcoal (1-2 g/kg, PO, once) should only be considered with lethal exposures to chocolate as long as the benefits outweigh the risks";.

That is a genuine reversal within the same lineage of veterinary advice, over twenty-three years, and it is worth knowing for one reason only: if your clinic decides against charcoal, it is not being casual. It is following current guidance that an older page you may have read online contradicts.

There is no antidote. Treatment is decontamination where the window allows, then stabilisation and support: anti-emetics for vomiting, methocarbamol or diazepam for tremors and mild seizures, beta-blockers such as esmolol, propranolol or metoprolol for tachyarrhythmias, lidocaine for refractory ventricular tachyarrhythmia, fluid therapy to promote diuresis, electrocardiographic monitoring, and correction of acid-base and electrolyte derangements. In the 156-dog series, 23 per cent of the symptomatic dogs needed esmolol and 14 per cent needed forced diuresis. Bonbeno names no drug or dose for you to give, which is why this page carries the medical disclaimer.

Four things to read off an Indian wrapper before you telephone

1. Read the product name exactly as printed, because Indian law attaches different minimums to dark chocolate, plain chocolate, couverture, white chocolate, cocoa powder and drinking chocolate.
2. Read the cocoa percentage if one is declared, and say whether the pack calls it total cocoa solids or just cocoa.
3. Read whether the bold declaration about cocoa butter equivalent or vegetable fat appears, because that tells the clinic it is looking at a compound rather than a couverture.
4. Read the net weight on the pack, count or weigh what is left, and report the difference as the worst case rather than the likely one.

Ask one more question before you set off, wherever you live. Ask whether the clinic can run a continuous electrocardiogram and give an intravenous beta-blocker, because that is what a symptomatic chocolate case needs. In Mumbai, Delhi, Bengaluru, Chennai, Hyderabad, Pune, Kolkata and Ahmedabad you have the best chance of finding a second hospital if the first cannot; outside those cities the answer to that question decides where you drive. Our guide to what a vet visit costs in India covers what an after-hours emergency consultation runs to.

Finally, the prevention, which is duller than the toxicology and matters more. Chocolate arrives in an Indian house in bulk twice a year and in gift boxes at weddings in between, and most of it sits at waist height in a drawing room full of visitors. Chocolate mithai is covered on our page on which mithai is toxic, sugar-free versions on xylitol poisoning in dogs, and the broader habit of handing a dog whatever humans are eating on why biscuits are not dog food. The tin of cocoa in the cupboard is the one worth moving tonight.