Seven Days of Oolong Extract Changed How Mice Cleared Acetaminophen

The Rule of this Order is arithmetic. Ninety degrees, not "hot." One hundred and twenty milliliters to five grams, not "a good pour." Seven infusions, counted, because the count is the claim. Everything the Order has ever written about oolong has been a measurement of what the leaf does to the cup, and through the cup, to the person sitting still in front of it.

Here is a measurement of something else. Not what the tea does to the palate. What the tea does to the machinery that clears a drug out of the body.

What was actually dosed, and to what

In May 2023, Miaogao Zhang and Zhenguo Qiu of Zhongkai University of Agriculture and Engineering in Guangzhou published a study in Frontiers in Nutrition on the effects of aqueous Baiyedancong-Oolong extract on cytochrome P450 activity, P-gp and OAT transport, and transcription levels in mice¹. The tea is Baiyedancong (白叶单枞), a Phoenix Dancong cultivar out of Guangdong. Familiar leaf. Many novices of this Order have it in the cupboard.

The study ran on two separate mouse cohorts, 180 animals in total, specific-pathogen-free grade, on the same dosing scheme. The larger cohort, 96 mice (48 male, 48 female), carried the CYP450 and P-glycoprotein work; a second cohort of 84 mice (42 male, 42 female) carried the kidney transporter work. Tea powder was extracted in boiling water and diluted to two working concentrations, one freeze-dried preparation and one hot-air-dried, which produced closely similar results. Delivery was oral gavage, 20 mL/kg of body weight, twice daily at 8am and 6pm, for seven consecutive days. The high dose ran 1454 to 1464 mg of extract per kg per day. The low dose ran 291 to 293.

The effects were dose-dependent, larger at 1454 to 1464 mg/kg/day than at 291 to 293, which is the pattern that separates a real pharmacological signal from noise.

In the liver, CYP3A activity rose significantly in a dose-dependent manner; at the top dose it reached roughly the induction level of rifampicin, the drug used in the study as the reference CYP3A inducer. The paper does not give liver CYP3A's own fold-change in the text, only in a figure, so that comparison is the precise claim to make here. Liver CYP2E1 did carry stated numbers: 2.82-fold to 6.81-fold over control across the four dose groups. Liver CYP2C37 rose 1.62-fold to 2.65-fold. In the intestine, CYP3A activity rose 3.73-fold to 6.01-fold.

The transporters moved the other way. P-glycoprotein ATPase activity fell to between 49% and 72% of control, and small-intestine MDR1 mRNA, the gene encoding that pump, dropped significantly. In the kidney, PAH uptake, a functional readout of organic anion transporter activity, fell to between 23% and 32% of control, with OAT1 and OAT3 mRNA significantly reduced.

Then the authors gave the animals a test dose of acetaminophen and measured it in serum. It ran higher in the dosed mice than in the controls. The enzyme numbers are laboratory readouts. That last one is a drug, in blood, at a different concentration because of the tea.

The authors' own conclusion is stated without drama: daily or high-dose Baiyedancong oolong consumption "may affect drug absorption, metabolisms, and excretion."

What those four names do

An enzyme rising fivefold means nothing to a reader who does not know what it was doing at onefold. Four checkpoints, briefly.

CYP3A is the workhorse. It sits in the liver and the intestinal wall and chemically dismantles a very large share of prescription drugs, by many estimates the majority of them. Raise its activity and a drug is broken down faster, which generally means less of it reaches the blood and less of it stays there.

CYP2E1 handles smaller molecules, ethanol among them, and is the enzyme that converts acetaminophen into the metabolite responsible for its liver toxicity at high doses. CYP2C37 is a mouse-specific member of the same family, a rough counterpart to the human CYP2C enzymes that clear warfarin and several others.

P-glycoprotein is not an enzyme but a pump. It sits in the intestinal lining and pushes drug molecules back out into the gut before they can be absorbed. Weaken it and more of a swallowed drug gets in.

OAT1 and OAT3 are the kidney's excretion transporters. They pull certain drugs out of the blood and into urine. Weaken them and the drug stays in circulation longer.

Note that these do not all point the same way. Induced CYP3A means faster breakdown. Suppressed P-gp means greater absorption. Suppressed OAT means slower excretion. The net effect on any particular drug depends on which of those checkpoints that drug actually passes through, which is why "tea changes drug metabolism" is not a usable statement and the specific readout on acetaminophen is.

The dose, honestly converted

A mouse mg/kg dose is not a human mg/kg dose. Metabolic rate scales with body surface area, not mass, and the standard correction is the Km factor. Nair and Jacob's practical guide to dose conversion between animals and humans³ gives the mouse factor as 12.3: human-equivalent mg/kg is the mouse mg/kg divided by 12.3.

Run it. The high dose of 1454 to 1464 mg/kg/day comes to roughly 118 to 119 mg/kg/day human-equivalent. For a 70 kg adult, roughly 8,300 mg of extract per day, about 8.3 grams. The low dose of 291 to 293 mg/kg/day comes to roughly 23.6 to 23.8 mg/kg/day, or about 1,660 mg for the same adult, about 1.7 grams. These are method-based approximations, not figures the paper reports.

Now the question every reader is holding. How many cups is that?

The Order will not tell you, because the Order cannot. Converting grams of extract into grams of dry leaf requires the extraction yield, the fraction of the leaf that ends up in the powder, and the paper's materials and methods do not report it. Without that number, any cups-per-day figure would be manufactured. It would also be the most quoted line in this piece, which is exactly why it will not be written here.

This is a limit, and the Order states limits rather than covering them. What can be said is that the low dose is not an absurd quantity of anything, and that the effects were already present at it.

The precedent that makes this credible

One paper in mice is one paper in mice. What keeps it from being dismissible is that the mechanism class is not speculative.

Grapefruit juice is the textbook case: an ordinary beverage that measurably changes drug levels in humans through CYP3A4, with clinical consequences serious enough that the FDA requires warning labels on some prescription drugs against it⁴. Its direction is opposite to this finding. Grapefruit inhibits intestinal CYP3A4, slowing clearance and pushing drug levels up. The oolong extract induced CYP3A, which would push in the other direction, faster clearance and potentially lower levels of a drug that needs to stay above a threshold to work. Different sign, same category of event.

The same Guangzhou group published a companion paper on the freeze-dried Baiyedancong extract² reporting the same direction of effect across CYP450, P-gp and OAT. A replication by the original group is weaker evidence than an independent one, and it is still better than a single unrepeated result.

What to do about it

No human clinical trial of oolong-drug interaction has been conducted. This is a mechanistic finding in an animal model, on one cultivar, over seven days. It is a reason to treat the interaction class as real. It is not a measurement of your risk.

So: do not stop drinking oolong on the strength of a mouse study, and do not decide the study is irrelevant because the subjects had tails. If you take a drug with a narrow therapeutic index, where the gap between too little and too much is small, say warfarin, digoxin, lithium, an immunosuppressant, an antiepileptic, tell your physician or pharmacist that you drink oolong daily and roughly how much. That sentence costs nothing and belongs in the same conversation as grapefruit and St John's wort.

The Order measures water to the degree and leaf to the tenth of a gram. A pharmacist measures the same way, on a different substrate, and is the one holding your other numbers.