Hangover recovery.

1. Salted water, not plain water. 500 ml with about half a teaspoon of salt, or an electrolyte sachet. Plain water on its own mostly becomes urine.
2. Ibuprofen 400 mg with food. Not paracetamol/Tylenol — see the headache section, this one genuinely matters.
3. Eat something with salt, carbohydrate and protein. Eggs on toast. Soup. It doesn't need to be clever.
4. Coffee if you normally drink it, one cup, alongside the water — skipping it adds caffeine withdrawal to your morning.
5. Go back to sleep if you can. Nothing on this list beats it.

Sauna, cold plunge, a hard workout, hair of the dog, or ordering a supplement that arrives on Tuesday. Three of those are covered below because they're actively worth avoiding, not just useless.

You will feel better in about 45 minutes and it won't be dramatic. Nothing reliably cures a hangover. Time does. Everything here is about shaving the edges off — roughly, taking a 7/10 morning to a 5/10 morning. Anyone selling you more than that is selling.

Alcohol is broken down to acetaldehyde, then to acetate. The entire supplement category is built on clearing acetaldehyde faster. But when researchers actually measured it: hangover severity tracked blood ethanol, not acetaldehyde. Inflammatory markers tracked ethanol too — and not acetaldehyde. Acetaldehyde is undetectable by 12 hours, which is roughly when people feel worst.

What does correlate is inflammation. TNF-α at four hours predicts hangover severity strongly (r ≈ 0.68). IL-6 and CRP follow the same pattern. Your immune system reacts to a large dose of ethanol, and the sickness behaviour that produces — the aching, the fog, the misery, the low mood — is most of what a hangover is.

• Sleep. Alcohol wrecks the second half of the night — REM suppressed, then rebound and fragmentation. A large part of how you feel is simply bad sleep, and no pill fixes that.
• Dehydration. Alcohol suppresses vasopressin so you lose more water than you take in. Real, but a contributor: it drives the thirst, dry mouth and dizziness, not the malaise.
• Acetate and adenosine. The metabolite after acetaldehyde drives adenosine accumulation, which is implicated in the headache. Note that this is also not acetaldehyde.

If inflammation and sleep are the main drivers, then: an anti-inflammatory has a plausible mechanism, rehydration fixes one specific symptom cluster, sleep is the highest-value intervention available to you, and an acetaldehyde scavenger is aimed at the wrong target — which is most of the shelf.

• Total ethanol. The dominant factor by a distance. Everything else is a rounding error next to how much you drank.
• Congeners. Dark spirits (bourbon, brandy, red wine) contain more of these fermentation by-products than clear ones (vodka, gin). Bourbon reliably produces worse hangovers than vodka at matched dose — but the difference is smaller than the difference between four drinks and seven.
• Sleep length. Going to bed at 2am rather than 4am is worth more than any supplement in this guide.
• Drinking speed and food. Same total, faster peak, worse morning.

• DHM (dihydromyricetin) — a flavonoid from the Japanese raisin tree and vine tea, used in Chinese medicine for alcohol complaints for centuries. It speeds up the two liver enzymes that clear alcohol and acetaldehyde, and blocks the GABA-A receptor alcohol acts on. The single most common ingredient in the category. Full entry →
• L-cysteine — an amino acid that reacts with acetaldehyde and turns it into a harmless compound you excrete. Real chemistry, tiny trial. Full entry →
• Prickly pear — cactus extract, anti-inflammatory, and the only ingredient here with a decent randomised trial. Full entry →
• Milk thistle (silymarin) — the liver-support ingredient in almost every formula, on thin evidence. Full entry →

Near enough every product in this category is a variation on: clear acetaldehyde faster, and add B vitamins and electrolytes. DHM speeds the enzymes that clear it, L-cysteine binds it chemically, ZBiotics engineers a gut bacterium to break it down before it's absorbed. Different clever routes to the same target.

A 2025 survey of the 46 hangover products on sale in the US found DHM in 52.6% of them, then vitamin B1 (42.1%), B12 (42.1%), sodium (42.1%) and B6 (39.1%). The same paper's conclusion: efficacy has not been demonstrated for any product on the US hangover market. Systematic reviews say the same thing in more words.

Not quite, and the distinction is worth holding. The mechanisms are real chemistry — cysteine genuinely binds acetaldehyde, DHM genuinely upregulates those enzymes in animals. This isn't homeopathy. The two problems are that (a) the human trials are small, short, self-reported and mostly funded or run by people with an interest, and (b) even if they work perfectly, they're clearing a metabolite that the inflammation evidence suggests isn't driving your symptoms.

Call it a plausible mechanism aimed at a doubtful target, with weak evidence. That's not fraud. It's also not worth $3 a night to most people.

1,200 mg vine tea at 98% DHM (so ~1,175 mg actual DHM) plus 450 mg L-cysteine, milk thistle, prickly pear and B vitamins. About $2.92/dose (12 doses, $34.99).

The best of the bunch, for one specific reason: it tells you the doses. You can check them against the trials. It also contains prickly pear, which is the one ingredient here with a decent RCT behind it. Still no trial on the finished product.

Vitamins B, C and E, magnesium, zinc, potassium, and a 383 mg ‘Liver Support Blend’ containing silymarin, taurine, DHM and L-glutathione. About $2.95/serving (40 chewables, 20 servings).

Do the arithmetic: four ingredients sharing 383 mg means the DHM dose is somewhere under 383 mg and probably well under — against a Cheers dose of ~1,175 mg and trial doses of 300–600 mg. You cannot know what you're taking, and with a dose-dependent mechanism that's the number that matters. The included glutathione is orally near-useless (see below). Same price as the one that discloses.

A genetically engineered Bacillus subtilis (strain ZB183) that produces acetaldehyde dehydrogenase in your gut, where your liver's enzyme doesn't reach. $6.45–$12 per bottle — two to four times the others.

Genuinely novel and a real piece of engineering. But: the enzyme is proven in a test tube, not shown to work in a human gut; the safety study on this strain is in rodents; and there is no randomised trial on hangover symptoms. Note their own language is careful — ‘feel better after drinking’, not ‘prevents hangovers’. The most expensive option with the thinnest outcome data.

Worth separating from the rest, because it isn't a supplement at all. Blowfish is 500 mg aspirin + 60 mg caffeine in two effervescent tablets — an FDA-registered OTC drug labelled specifically for hangover, not a ‘dietary supplement’ making structure-function noises. About $0.63/dose — a fifth of the DHM products.

And notice what it targets: aspirin is an anti-inflammatory, and caffeine blocks adenosine. Those are the two mechanisms the evidence in this guide actually supports. Effervescent means faster absorption and gentler on a raw stomach than a tablet. It is the most defensible product on this page — and it is also, transparently, aspirin and a coffee, which you can have for about eight cents. You're paying for buffering, speed and the fact that it's already in the cupboard.

The aspirin caveats are real: not on an empty stomach, not if you have ulcers, reflux or a bleeding disorder, and see the headache section below on why aspirin specifically has a wrinkle here.

The rest recombine the same short list: DHM, cysteine or NAC, milk thistle, prickly pear, B-complex, electrolytes, sometimes ginger for nausea. If you've read this far you've read the category. Judge a new one on two questions: does it state its doses, and is there a trial on the finished product? For essentially all of them the answers are partly and no.

Buy the one that states its doses and take it before or during drinking, never the morning after — every mechanism in the category is metabolic and the ethanol is long gone by then. Expect a modest reduction in severity at best, and know you're paying roughly $3 a night for a maybe.

You can assemble it for well under a dollar: prickly pear extract before drinking (the best trial in the category), salted water before bed, ibuprofen and food in the morning. That is most of the value of a $3 sachet, and the ibuprofen is doing more work than the DHM.

Alcohol depletes liver glutathione. Glutathione is what the liver uses to neutralise reactive metabolites. Oral glutathione is largely destroyed in the gut. Therefore: inject it, bypass the gut, top up the liver. Every step of that sounds right and the last one doesn't follow.

Two facts settle this. Injected glutathione has a plasma half-life of about 14 minutes. And it does not enter cells intact — the enzyme γ-glutamyl transpeptidase cleaves it in the extracellular space, and cells then take up the pieces and rebuild glutathione internally. In one study, IV glutathione raised plasma cysteine from 8.9 to 114 µmol/L.

So a glutathione shot is, mechanistically, a fast, short, expensive way of delivering cysteine. That's not nothing — cysteine is the rate-limiting precursor. But it means the honest comparison isn't ‘shot versus nothing’. It's ‘shot versus a cheap oral cysteine donor’.

No, and this is the part worth being clear about, because the subQ shot is now more common than the drip. The cleavage happens in the extracellular space, so it happens whatever route the molecule arrives by. Injecting under the skin doesn't route around γ-glutamyl transpeptidase; it just gets to the bloodstream more slowly, through tissue and lymph where some of it oxidises on the way.

If anything the case is weaker. IV glutathione at least has published pharmacokinetic data. Subcutaneous glutathione has essentially none — absorption fraction, peak timing and local tolerability have not been characterised in peer-reviewed studies, and the 600–1,200 mg protocols circulating in peptide communities are extrapolated from IV work rather than validated. You are self-administering an unapproved compound at a dose nobody has tested by that route.

NAC is exactly that: well-absorbed oral cysteine, pennies a dose. It was put through a randomised, double-blind, placebo-controlled crossover trial in 49 people for hangover. No difference in overall hangover score (median 10 vs 13, p = 0.45). A subgroup signal suggested women improved and men didn't (p = 0.04) — a post-hoc subgroup on a null primary, which is a hypothesis for the next study, not a result.

So the delivery route that gets cysteine in cheaply, orally and reliably didn't beat placebo. That's the strongest argument against the injection, and it's an argument from the same mechanism the clinics use to sell it.

Probably because it does something — just not necessarily what's on the label. Take these seriously rather than dismissing them:

• Hangover is subjective, self-limiting and improves hourly. Anything you do at 10am is followed by feeling better at noon, whatever you did.
• Expectancy is enormous for subjective symptoms, and it scales with cost, effort and needles. A shot you paid for and administered yourself is close to a perfect placebo delivery system — that is a statement about how symptom perception works, not about anyone's honesty.
• If it was an IV, the litre of saline is doing real work and getting none of the credit.
• Cysteine delivery is real. You are getting cysteine, just expensively and briefly. If that's the active ingredient, there are cheaper routes — which is exactly what the NAC trial tested.

None of that means your friends are wrong that they felt better. It means the glutathione is the least likely part to be responsible.

Injectable glutathione is not FDA-approved for any use. Regulators have warned about it repeatedly in its main market (cosmetic skin-lightening), citing liver, kidney and neurological toxicity, and there are no published dosing guidelines. Compounding quality is a real hazard: in one incident a pharmacy made injectables from oral-grade powder and hospitalised patients with endotoxin reactions. If you're getting a shot, the sterility of the preparation is a bigger question than the molecule.

Fair question, and ‘doesn't enter cells intact’ is not the same as ‘does nothing’. Extracellular cleavage IS the body's own delivery system. Glutathione made in the liver is exported, broken down at other tissues' cell surfaces, and rebuilt inside them — the γ-glutamyl cycle. Injecting it isn't fighting physiology, it's paying premium for one step of it.

So what you actually buy is: a short spike in extracellular antioxidant capacity, and a large delivery of cysteine. Both are real. Neither needs an injection.

Its one use with genuine traction is skin lightening, which is where nearly all the demand and nearly all the safety warnings come from — and which plausibly works at the melanocyte surface by shifting melanin synthesis, an effect that doesn't require getting inside the cell. Beyond that, trials in conditions like Parkinson's have been small and mostly disappointing. For the liver, after alcohol, there is no route advantage at all: the liver makes its own glutathione from cysteine, and swallowing cysteine gets it there without a needle.

Skip it. The molecule is cleaved before it reaches a cell, its oral equivalent failed a proper trial, it's unapproved and unstandardised, and whatever benefit people report is best explained by the saline it arrives with. If you want the IV fluid because you can't keep water down, that's a legitimate medical reason to go — ask for fluid, not glutathione.

Alcohol suppresses vasopressin, so you produce a lot of dilute urine — you lose proportionally more water than salt. Drinking plain water into that dilutes your remaining sodium further and a good share goes straight back out. Adding sodium is what makes fluid stay in you. That's the entire principle of oral rehydration therapy, one of the best-evidenced interventions in medicine.

• Sodium — the one doing the work. 500–1,000 mg alongside 500 ml–1 litre of fluid.
• Potassium — modest losses from one night. Worth having, not worth chasing, and genuinely risky to over-supplement.
• Magnesium — heavily marketed here; losses in a single night are small. Real value if you drink often, little for one occasion.
• Everything else on the label — zinc, calcium, trace minerals. Marketing.

Which means the expensive sachet and half a teaspoon of table salt in a litre of water are doing much the same job. Buy the sachet for convenience and taste, not chemistry.

Before bed is the highest-value moment, then on waking. Sipping steadily beats a pint in one go, which mostly becomes urine. Don't overdo it — drinking litres to ‘flush it out’ is uncomfortable, doesn't speed alcohol clearance, and in extreme cases is dangerous.

No evidence they beat drinking. Clinical view is that an IV doesn't rehydrate you better than oral fluid and doesn't speed hangover recovery; IV fluid is indicated when you can't keep liquid down. It is faster — 30–60 minutes versus a few hours — so you're buying speed, at $150–$300, for something a litre of salted water gets you by lunchtime.

The hangover headache runs on prostaglandins and adenosine — both inflammatory-ish targets, both hit by NSAIDs. Ibuprofen 400 mg with food is the sensible default. Naproxen 250–500 mg lasts longer if you're writing off the day. Older trials of tolfenamic acid (an NSAID) before drinking and at bedtime showed milder hangover headaches, which is the mechanism working as advertised.

Take it with food and water. Alcohol has already irritated your stomach lining; an NSAID on top of that on an empty stomach is how people give themselves gastritis.

This is the one hard rule in the guide. Paracetamol is cleared by a pathway that produces a toxic intermediate (NAPQI), which the liver neutralises using glutathione. Binge drinking depletes exactly that glutathione. In the lab, alcohol plus paracetamol produced markedly more liver damage than paracetamol alone.

Note the irony: the glutathione depletion that makes people want a glutathione shot is precisely what makes paracetamol dangerous here. The shot won't fix it; not taking the paracetamol will.

Aspirin works and is what Blowfish uses. Two caveats: it's harder on the stomach than ibuprofen, and it inhibits alcohol dehydrogenase, which can raise blood alcohol if taken while still drinking. Morning-after aspirin is fine; aspirin as a pre-drink strategy is not.

Have your normal amount, not more. It constricts the dilated vessels and blocks adenosine, which helps the headache a little — and if you're a daily coffee drinker, skipping it stacks caffeine withdrawal on top of a hangover, which is the actual reason coffee ‘helps’. More than usual worsens the jitteriness and won't help the sleep debt.

Alcohol suppresses hepatic gluconeogenesis — your liver is busy metabolising ethanol and stops making glucose properly. Add a night of not eating and low glycogen, and part of the shakiness, weakness and fog is straightforward low blood sugar. Fasting extends that. There is no detox argument for it: your liver clears alcohol at a fixed rate, roughly one standard drink an hour, and nothing you eat or don't eat changes that rate.

If you're a committed intermittent faster, this is the morning to break the rule. One skipped fasting window costs you nothing; a day of hypoglycaemic misery costs you the day.

• Salt + fluid + carbohydrate together. Broth, miso soup, pho, congee. This is why every drinking culture has a soup for the morning after — they all independently arrived at salted liquid carbohydrate.
• Eggs. Genuinely useful: high in cysteine, plus protein and choline.
• Simple carbohydrate early — toast, rice, banana, honey. Restores glucose fastest when you're shaky.
• Fruit and juice for fluid, potassium and sugar. Orange juice, watermelon, banana.

• The full fry-up as a first move. A big fat load into a stomach that's already irritated and slow to empty is how people end up vomiting. Eat it at lunch if you want it, not at 8am.
• More alcohol. ‘Hair of the dog’ works by re-anaesthetising you and postponing the same hangover, plus it's a pattern worth not building.
• Very sugary drinks alone — a spike then a crash without the salt or fluid that actually help.

This is the strongest ‘no’ in the guide. Alcohol and heat both cause vasodilation and drop blood pressure; alcohol impairs thermoregulation and prolongs the QT interval. Together that means fainting, heat stroke and arrhythmia. In Finland — a country that knows about saunas — alcohol is a factor in around half of sauna-related deaths, and the major clinical review of sauna use lists alcohol as a circumstance in which you should not use one.

Less acutely dangerous once you're sober, but still not clever: you're dehydrated, your resting heart rate is up, and a sauna dehydrates you further. If you're going to, rehydrate properly first, keep it short, and know that it isn't doing anything special. You cannot sweat out alcohol — the liver clears roughly 90% of it and sweat is a rounding error.

Safe enough when sober, and the noradrenaline hit genuinely makes people feel more alert and better for an hour or two. It is not clearing anything or reducing inflammation in a way that matters here. A cold shower gets you most of it. Never while still intoxicated — cold shock plus impaired judgement around water is its own category of accident.

Expect roughly 11% less aerobic capacity, heart rate 10–15 bpm higher at any given pace, and 30–90 seconds/km slower than your normal easy pace. Coordination, reaction time and proprioception are still impaired even after the alcohol has cleared — which is an injury risk, not just a performance one. Heavy drinking also raises arrhythmia risk for up to 48 hours, and that risk is highest during exertion.

So: a walk outside is the best exercise for a hangover — light movement, daylight, fresh air, no risk. If you want more, easy zone-2 cardio at conversational pace is fine once you've rehydrated. Skip heavy lifting, skip anything requiring balance or skill, skip anything competitive, and don't set a PB attempt for the morning after a wedding.

The interventions that reduce alcohol-related organ damage, in order of effect size: drink less, drink less often, have alcohol-free days, and don't binge. Nothing in a capsule is within an order of magnitude of these, and a supplement that makes you comfortable drinking more is a net negative however good its ingredient list.

In nearly every hangover formula. Reasonable antioxidant and membrane-stabilising mechanism; the Cochrane-level evidence in alcoholic liver disease is unconvincing, with no clear mortality or histology benefit. For a single night it does nothing you'd notice. Harmless, cheap, mostly there for the label.

The most defensible thing in this section, for a reason that has nothing to do with hangovers: it restores glutathione, and it's the actual hospital antidote for paracetamol poisoning. If you drink regularly it's a rational daily supplement. It did not beat placebo for hangover symptoms in a 49-person trial, so don't take it expecting to feel better tomorrow.

If you drink heavily and regularly, thiamine is genuinely important — alcohol impairs its absorption and storage, and depletion causes Wernicke's encephalopathy, which is a medical emergency and is under-diagnosed. Take it daily, 100 mg, not on hangover mornings. For occasional drinkers it does nothing, which is why its presence in 42% of hangover products is decoration.

Less to say, which is itself the point. Healthy kidneys handle occasional drinking; the useful interventions are the same fluid and salt already covered. The genuine kidney risk in this guide is NSAIDs plus dehydration — safe in a rehydrated person taking one dose, a real problem if you're volume-depleted and taking them repeatedly. Rehydrate first, then take the ibuprofen.

1. Drinking less. Total ethanol dominates every other variable. Unhelpful as advice at 8am, decisive at 8pm.
2. Sleep. More of it, and going to bed earlier. The single biggest lever you still control after the drinking is done.
3. Time. Genuinely the cure. Everything else is edge-shaving.

1. An NSAID with food — the most reliable symptom relief available, ~30–45 min, pennies.
2. Salted fluid, before bed and on waking — fixes thirst, dry mouth and dizziness within the hour.
3. Food with salt, carbs and protein — fixes the shakes and fog.
4. Prickly pear extract, 5 hours before drinking — the only supplement here with a decent RCT (CRP −40%, severe hangovers roughly halved). Requires planning you probably won't do.

1. DHM products (Cheers, H-Proof et al.) — plausible chemistry, weak human evidence, aimed at a target the mechanism data doesn't support. ~$3/night.
2. L-cysteine — one small positive trial with borderline p-values.
3. ZBiotics — the cleverest science, the least outcome evidence, the highest price.
4. Ginger for nausea — decent evidence for nausea generally, untested here specifically, and harmless.

1. Paracetamol/acetaminophen. Actively risky with a depleted liver.
2. Sauna while still drunk. The one genuine danger in this guide.
3. Glutathione injections. Cleaved in minutes, oral equivalent failed a proper trial, unapproved and unstandardised.
4. Hair of the dog. Postponement, not treatment.
5. Hangover IV drips as a routine, unless you can't keep fluid down — then it's legitimate medicine.