Why Dogs and Cats Metabolise Green Tea Extract Differently
Glucuronidation, acetylation and methylxanthine clearance — the three pathways that make a single inclusion rate unsafe for two species.
Quick answer. Dogs are poor acetylators but glucuronidate phenolic compounds normally. Cats are the reverse: UGT1A6 is a non-functional pseudogene, UGT1A9 is absent, and glucuronidation of phenolic substrates runs roughly tenfold lower than in dogs — and cats also lack NAT2. Because catechins are cleared mainly by glucuronidation, the canine NOAEL of 40 mg EGCG/kg body weight/day cannot be read across to cats, which have no established ceiling at all.
1. Two species, two different bottlenecks
The most common error in pet supplement formulation is treating "dog and cat" as one target. The two species fail at different steps, and green tea extract happens to sit directly on the step where cats are weakest.
| Pathway | Dog | Cat | Relevance to green tea extract |
|---|---|---|---|
| Glucuronidation (UGT1A) | Ten functional isoforms; normal phenolic clearance | UGT1A6 is a pseudogene, UGT1A9 absent; roughly tenfold lower activity | Catechins are cleared mainly this way — the primary route is compromised in cats |
| N-acetylation | Poor acetylators; slow acetylation pathway | Lack NAT2; NAT1 present at low activity | Affects sulfonamides, not catechins |
| Methylation (COMT / TPMT) | Normal | Low thiopurine methyltransferase activity | Secondary catechin route also reduced in cats |
| Methylxanthine clearance | Caffeine half-life ≈4.5 h; theobromine ≈17.5 h | Theophylline half-life ≈7.8 h vs ≈5.7 h in dogs | Green tea carries caffeine, not theobromine |
Sources: Court MH, Feline drug metabolism and disposition, Vet Clin North Am Small Anim Pract 2013;43(5):1039–1054; Merck Veterinary Manual, methylxanthine toxicosis.
Read the table row by row and a pattern appears: the dog's known defect (acetylation) is irrelevant to catechins, while the cat's defects — glucuronidation and methylation — are precisely the two routes catechins depend on. The canine safety dataset therefore does not transfer.
What is known about dogs supports the contrast. In canine studies the data point to predominantly biliary excretion of EGCG and epicatechin gallate, while the two non-gallated catechins, epicatechin and epigallocatechin, were conjugated before being excreted in urine. In other words, the dog is running the conjugation routes that the cat performs poorly — the same routes that, when they fail in cats, produce the well-known sensitivities to paracetamol, propofol and permethrin.
2. Why the feline gap is a plant-chemistry problem
There is an evolutionary reason for the feline deficit, and it maps almost exactly onto green tea. Cats are obligate carnivores, and a standard explanation in comparative drug metabolism is that they almost completely lost the ability to glucuronidate xenobiotics because their ancestral diet never required them to clear readily glucuronidated plant flavonoids and alkaloids.
Catechins are plant flavonoids. Green tea extract is, in effect, a concentrated dose of the one chemical class the feline conjugation system is least equipped to process. That does not prove toxicity, but it removes any basis for assuming a cat behaves like a small dog.
The data gap is explicit. Beynen's 2020 review of green tea extract in petfood states that "for cats there is no available information about maximum safe intake and metabolic handling of EGCG." The single published feline datum — Isogai et al. (2008), 29 cats fed 333 mg total catechins (142 mg EGCG) per kg diet for 45 days — was an uncontrolled gingivitis trial, not a safety study. Working from that datum and from a conventional tenfold uncertainty factor on the canine NOAEL gives a defensible interim cap of about 140 mg EGCG per kg complete dry food, roughly one seventeenth of the canine figure.
3. Where the methylxanthine argument misleads buyers
Suppliers often justify decaffeination by invoking chocolate toxicosis. That argument does not survive contact with the numbers. Chocolate is dangerous to dogs mainly because of theobromine, whose canine half-life is roughly 17.5 hours; green tea contains negligible theobromine. Caffeine, the methylxanthine green tea does contain, has a canine half-life of about 4.5 hours, well inside the normal human range.
Cats are not uniformly more sensitive, either — which is exactly why extrapolation fails in both directions. Reported minimum lethal doses for caffeine are 140–150 mg/kg in dogs and 100–150 mg/kg in cats, a marginal difference; but the theophylline LD50 is approximately 250 mg/kg in dogs against roughly 800 mg/kg in cats, meaning cats are considerably more tolerant of that particular methylxanthine. Species differences here are compound-specific, not a general "cats are fragile" rule.
At realistic inclusion rates the methylxanthine load from green tea extract sits two to three orders of magnitude below those thresholds even for a fully caffeinated 8% extract. For dogs, EGCG is the binding constraint, not caffeine.
4. What this means for your specification
Three consequences follow. First, one premix cannot serve both species: a premix dosed for a dog product will overshoot a defensible feline cap by more than an order of magnitude. Plan a separate dilution or inclusion step.
Second, treat any supplier document that supports a feline claim with a canine NOAEL as a red flag. The two species fail at different metabolic steps, so a dog number is not evidence about cats — it is evidence about dogs.
Third, specify the two lots differently. Both need EGCG by HPLC with a named marker list and a stated reporting basis, but the feline lot warrants a tighter catechin window, a lower numeric caffeine limit and tighter batch-to-batch variation, because at an inclusion rate near 237 mg/kg a ±10% assay swing carries far more weight than it would in a dog formula.
Finally, be explicit in your documentation about what the numbers are. A canine NOAEL of 40 mg EGCG/kg body weight/day is a defensible point of departure for a dog product; the feline figure is an interim screening cap derived from one short uncontrolled study, not a safety determination. Labelling the latter as though it were the former is the single most common misrepresentation in this category, and it is the one a technically literate buyer will catch first.
