How Your Liver Actually Detoxifies
A fat-soluble drug or toxin cannot leave your body in urine — so a liver cell rebuilds it in three steps: Phase I (cytochrome P450) bolts on a reactive handle and often makes it more toxic, Phase II glues that handle to a water-soluble partner such as glutathione and makes it harmless, and Phase III pumps the result out into bile or blood. Watch molecules run the gauntlet, then break it: induce Phase I, drain the glutathione tank the way a paracetamol overdose does, give NAC and watch the cell recover — and try the “detox tea” button to see it change precisely nothing.
Try this: start on Normal, then hit Glutathione depleted and watch the green tank empty, the red reactive intermediates pile up and ALT climb past 1,000 U/L — then hit NAC.
Live hepatocyte readout
What's happening
The Science in Plain Language
Fat-soluble is the whole problem
Your kidneys excrete things that dissolve in water. A fat-soluble (lipophilic) molecule — most drugs, pesticides, solvents, combustion products, many hormones — simply diffuses back across the kidney tubule wall and returns to your blood. Left alone it would loop around your body indefinitely, and because it dissolves in fat it would accumulate in membranes and adipose tissue. Your liver's answer is not to “flush” it. Its answer is chemistry: rebuild the molecule until it is water-soluble, then pump it out. That is what biotransformation means, and it runs continuously, in every hepatocyte, whether or not you buy anything.
Phase I: cytochrome P450 makes it more reactive, not less
In the smooth endoplasmic reticulum sits a family of about 57 human cytochrome P450 enzymes. A handful do most drug work: CYP3A4 alone handles roughly half of all prescription drugs; CYP2E1 handles ethanol, paracetamol and many solvents; CYP1A2 handles caffeine and the polycyclic aromatic hydrocarbons in smoke and char.
What they do is oxidation: they bolt a reactive handle — usually a hydroxyl (–OH) group — onto the molecule. This is the step people get backwards. Phase I is not detoxification. It is functionalisation: it creates a chemical grip point so Phase II has something to hold on to. For many drugs the Phase I product is simply inactive. But for a long list of substances — paracetamol, benzo[a]pyrene from smoke, aflatoxin B1, vinyl chloride, some pesticides — the Phase I product is an electrophile: a hungry, unstable molecule that is more toxic than what you started with. This is called metabolic activation, and it is why the middle of the diagram glows red.
The P450 cycle is also leaky. It burns oxygen and NADPH, and it “uncouples” — spilling superoxide and hydrogen peroxide into the cell. CYP2E1 is the leakiest of the lot, which is one reason alcohol is hard on the liver: it induces the enzyme that both produces reactive intermediates and throws off the most oxidative stress.
Phase II: conjugation is the step that actually makes it safe
Phase II grabs the handle Phase I created and glues a bulky, charged, water-loving partner onto it. Six main routes do this, and the diagram shows all six:
- Glutathione conjugation (GST) — the emergency service for reactive electrophiles. Glutathione (GSH) is a tripeptide of glutamate + cysteine + glycine; its cysteine thiol grabs the electrophile before it can hit anything important.
- Glucuronidation (UGT) — quantitatively the largest route, adding glucuronic acid derived from glucose.
- Sulfation (SULT) — fast, high-affinity, but low-capacity: it saturates early, which matters enormously in overdose. It needs sulfate, which ultimately comes from sulfur amino acids.
- Amino-acid conjugation — glycine, taurine or glutamine are attached (benzoic acid + glycine → hippuric acid, excreted in urine).
- Acetylation (NAT1/NAT2) — genetically variable; “slow acetylators” clear drugs such as isoniazid and hydralazine more slowly.
- Methylation — a methyl group donated by SAMe, which depends on folate, B12 and B6.
Notice what every one of those routes is made of: amino acids, glucose, sulfate, methyl groups. Phase II runs on raw materials you eat. A protein-poor diet genuinely limits conjugation capacity. That is a real, mechanistic nutrition link — and it is not what a juice cleanse supplies.
Phase III: two exits, both already open
Once conjugated, the molecule is water-soluble and can no longer drift back across a membrane — it has to be pumped. That is Phase III, and it uses ATP-driven efflux transporters:
- Into bile: MRP2, BSEP and P-glycoprotein push conjugates across the apical membrane into the bile canaliculus → bile duct → gut → stool. Large conjugates (roughly > 500 Da) favour this route.
- Into blood: MRP3 and MRP4 push conjugates back out into the sinusoid, where the kidney filters them into urine. Smaller, more polar conjugates favour this route.
Both exits run all the time. Nothing has to be “opened” or “unblocked”.
The enterohepatic loop: why fibre is the part that actually matters
Getting a conjugate into the gut is not the same as getting it out of your body. Gut bacteria produce β-glucuronidase, an enzyme that snips the glucuronide back off. The molecule reverts to its original fat-soluble form, is reabsorbed through the intestinal wall, travels up the portal vein and lands back in the liver — which has to do the whole job again. This is enterohepatic recirculation, and it is a real, measurable phenomenon (it is why some drugs have a long, bumpy half-life).
What shortens the loop is unglamorous: dietary fibre and transit time. Fibre binds bile and its cargo, adds bulk, and moves the whole lot toward the exit before the bacteria have time to undo the liver's work. Press Low fibre and watch conjugates get cut open in the gut and float back around to the liver, while the “cleared/min” number falls. This is the single most defensible thing anyone can say about “helping your body eliminate toxins” — and it involves vegetables, beans and whole grains, not a tea.
When Phase I outruns Phase II: paracetamol, NAPQI and NAC
Here is the point of the whole page. Phase I and Phase II are two taps and one drain. If Phase I runs faster than Phase II can conjugate, reactive intermediates accumulate. They do not wait politely. They form covalent adducts on cellular proteins and mitochondria, trigger oxidative stress, and the cell dies. More Phase I is not better. Balance is better.
The textbook example is paracetamol (acetaminophen). At a normal dose, ~90% of it is conjugated directly by glucuronidation and sulfation — no Phase I needed — and only about 5–10% goes through CYP2E1 to form NAPQI (N-acetyl-p-benzoquinone imine), a savagely reactive electrophile. Glutathione mops up that small trickle instantly and it leaves in urine as a mercapturate. You never notice.
In overdose (a single ingestion above roughly 150 mg/kg, or about 7.5–10 g in an adult), sulfation and glucuronidation saturate. The excess is shunted down the CYP2E1 route, so NAPQI production rises steeply at exactly the moment the other routes have quit. Glutathione is consumed faster than the cell can resynthesise it; once hepatic GSH falls below roughly 30% of normal, unquenched NAPQI starts binding cell proteins. The result is centrilobular (zone 3) necrosis — zone 3 hepatocytes have the most CYP2E1, which is why the damage is concentrated there. ALT, an enzyme that leaks out of dying hepatocytes, climbs from a normal ≤40 U/L into the thousands, typically peaking around 72 hours. (ALT clears slowly — its plasma half-life is around 47 hours — so in real life it takes days to fall, not the seconds this animation compresses it into.)
N-acetylcysteine (NAC) is the antidote, and it works for an entirely mechanistic reason: it supplies cysteine, the rate-limiting ingredient for making glutathione. Restock glutathione and Phase II capacity comes back above Phase I flux — the intermediates get caught again. NAC also scavenges NAPQI directly and supports sulfation. It is on the WHO Model List of Essential Medicines, it is given in hospital by IV or orally, and it is near-completely protective when started within about 8–10 hours of ingestion (still worth giving later, including in established liver failure). Treatment decisions are guided by a timed blood paracetamol level plotted on the Rumack–Matthew nomogram. If you ever suspect a paracetamol overdose — including a slow accumulation from taking a bit too much for several days — go to hospital immediately; the early hours are silent and asymptomatic, and that is precisely when treatment works.
This is the one true, hospital-grade “detox” intervention on this page. It is a specific antidote, for a specific poison, on a specific clock.
So do “detox teas” and juice cleanses work?
No. Press the Detox tea / juice cleanse button and watch the animation carefully: nothing changes. Phase I flux, Phase II capacity, transporter output, clearance per minute — every number stays exactly where it was. That is not a joke at your expense; it is the honest answer. There is no controlled human evidence that any commercial cleanse, tea, foot pad, coffee enema or juice protocol measurably increases the elimination of any named toxicant. A critical review of detox diets published in the Journal of Human Nutrition and Dietetics (Klein & Kiat, 2015) found the field essentially devoid of rigorous clinical trials. “Toxins” in the marketing are almost never named, and if a product cannot name the molecule, it cannot show it removed it.
What these products do contain is often a stimulant laxative such as senna and/or a diuretic herb. You get bowel movements and water loss, which feel like something is being purged and register as a couple of pounds on the scale. Neither increases hepatic biotransformation. Chronic laxative use risks low potassium and dehydration. Worse, herbal and dietary supplements are now one of the leading causes of drug-induced liver injury in the United States — high-dose green tea extract is a documented cause of hepatotoxicity. A “liver cleanse” can put you in a hepatology clinic.
What genuinely supports the machinery in the diagram:
- Drink less alcohol. The biggest lever most people have. Ethanol induces CYP2E1 (more reactive intermediates, more ROS) and depletes glutathione — it pushes both sides of the balance the wrong way.
- Eat enough protein. Glycine, cysteine, glutamine, taurine and methionine are the literal substrates of Phase II conjugation. Prolonged fasting or a juice-only regimen restricts exactly these.
- Cruciferous vegetables. Broccoli, sprouts, cabbage and kale contain glucosinolates that yield sulforaphane, which activates the Nrf2 pathway — a transcription factor that genuinely upregulates Phase II enzymes (GSTs, NQO1, UGTs) and glutathione synthesis. This is a real, well-mapped mechanism, and human broccoli-sprout trials have shown increased urinary excretion of conjugates of airborne pollutants. It is food, not a cleanse.
- Fibre. Shortens the enterohepatic loop, as above.
- Normal hydration. Supports the urinary exit. Drinking excess water does not accelerate it and, in extremes, causes dangerous hyponatraemia.
- Check your medicines. Real, large effects on Phase I come from pharmacology, not tea: grapefruit juice inhibits CYP3A4 and raises drug levels; St John's wort induces CYP3A4 and P-glycoprotein, and has caused transplant rejection and contraceptive failure.
And milk thistle (silymarin)? Be honest about it: it is traditional, it is generally well tolerated, and intravenous silibinin is used for Amanita phalloides mushroom poisoning on observational evidence. But Cochrane reviews of silymarin in alcohol-related and viral liver disease have not found convincing evidence of benefit for mortality or liver histology. It is not a detox agent, and it is not a licence to drink.
Your liver and kidneys are already doing this, right now, without being asked. The useful question is never “how do I detox?” — it is “what am I asking my liver to process, and have I given it the raw materials?”