Is a Glass of Wine a Night Really Harmless? What Alcohol Does to Sleep, Nerves & Your Ability to Detoxify

She came in with a familiar list. The weight had not moved in eight months despite a diet she could defend to a nutritionist. She was falling asleep fine and waking at 2:40 a.m. most nights, wired, and then dragging until noon. And her feet burned — that particular, hard-to-describe burning and pins-and-needles that small nerve fibers produce when they are unhappy. She had been told by another clinician that her nightly glass or two of wine was not a problem. “It’s within the guidelines,” she was told. “It’s heart-healthy. It’s not the wine.”

I disagreed, and I want to explain exactly why — not with slogans, but with mechanism and with data.

“It’s only one or two glasses. Surely that can’t be what’s holding me back?”

Here is the short version. Alcohol is not a food, a nutrient, or a neutral beverage. It is a small, fat-and-water-soluble solvent that your body has no storage compartment for and therefore must destroy immediately, in a two-step reaction that produces a highly reactive aldehyde as its first product and a large, hours-long shift in cellular redox chemistry as its price. That reaction takes priority over almost everything else the liver was doing. It burns through NAD+ and glutathione. It reroutes fuel metabolism. It fragments the second half of the night. It is directly toxic to peripheral nerve axons independent of any vitamin deficiency. And for a person whose liver is already fully occupied with the background work of modern life — mycotoxins, metals, plastics, pesticides, the residue of an old mold exposure, a chronic infection — it is a line-cutter at a counter that already has a queue out the door.

None of that is controversial biochemistry. What is contested — and I will treat that dispute honestly rather than pretending it doesn’t exist — is the population-level question of whether light drinking shortens the average life. In 2025 two federally commissioned expert reviews looked at overlapping evidence and reached opposite headlines [98,89]. I will show you both, explain why they diverge, and then explain why I think that particular argument is, for my patients, largely beside the point.

Because the question that matters in a clinic room is not “does one drink a day change average population mortality by three percent?” The question is: given everything this specific body is already trying to repair, does adding a nightly dose of a solvent and its aldehyde help or hinder? On that question, the evidence is not ambiguous at all.

And I want to state my position at the top rather than burying it, because you should know where this article is going before you invest an hour in it.

One drink can undo several days of the work you are doing to get well.

You can take the supplements, train consistently, sleep on a schedule, eat paleo or carnivore or whatever protocol you have chosen with real discipline — and a single evening’s drinking will suppress your fat oxidation for the next fifteen hours, strip the REM and deep sleep out of that night, flatten your heart rate variability, burn ATP and NAD+ that your repair systems needed, and leave you measurably impaired the next day at a blood alcohol of zero. That is not a rounding error against a good week. It is a meaningful fraction of the week, spent going backwards.

So the position is straightforward: we strongly discourage alcohol on the healing journey, and we discourage it generally for anyone pursuing optimal wellness. Not because of moral disapproval — alcohol is not a character question — but because when you look at the whole picture rather than one endpoint at a time, there is no dose at which the benefits outweigh what it costs a body that is trying to repair, perform, and age well. Even in a person with nothing wrong with them, alcohol is not building anything. It is only ever subtracting, and the only variable is how much.

This is a long article. Use the contents to find the part that applies to you, and come back for the rest.

First, What “One Drink” Actually Means

A US standard drink is 14 grams of pure ethanol: 12 oz of 5% beer, 5 oz of 12% wine, or 1.5 oz of 40% spirits. That is the unit every study in this article uses, and it is almost never the unit that appears in a real glass.

Modern wine is frequently 13.5–15% alcohol, not 12%. A generous home pour is commonly 8–9 oz. Do the arithmetic and a single “glass of wine” can easily be 2 to 2.5 standard drinks. When a patient tells me “two glasses,” the honest translation is often four to five standard drinks — 56 to 70 grams of ethanol — six nights a week, or roughly 350–420 grams weekly. For reference, the largest pooled analysis of alcohol and mortality ever conducted — 599,912 current drinkers across 83 prospective studies — put the threshold for lowest all-cause mortality at about 100 grams per week, and found that people drinking above 350 g/week lost an estimated four to five years of life expectancy compared with those below 100 g [96].

The reason this matters is that nearly every reassuring claim about alcohol — the guideline limits, the “heart-healthy” studies, the J-curve — is anchored to the 14-gram standard drink. If your real intake is double what you report to yourself, you are not in the part of the curve you think you are in.

The Biochemistry: Alcohol Becomes an Aldehyde Before It Becomes Anything Else

Your body has no way to store ethanol and no receptor whose job is to welcome it. It has only one strategy: oxidize it, fast, in the liver. That happens in two steps.

Step one: alcohol dehydrogenase (ADH) strips two hydrogens from ethanol and hands them to NAD+, producing acetaldehyde. Step two: aldehyde dehydrogenase (ALDH2), inside the mitochondria, oxidizes acetaldehyde to acetate, again consuming NAD+. Acetate then leaves the liver and is burned as fuel by muscle, heart, and brain [2,1].

Two additional facts change everything about how this plays out in real life.

First, the kinetics are zero-order. Unlike most substances, which you clear as a percentage of what is present, ethanol is cleared at a roughly fixed rate — about one standard drink per hour for most adults. Drink faster than that and the excess simply waits in your bloodstream and tissues. There is no way to speed it up: not coffee, not a cold shower, not sweating in a sauna. Only about 2–5% of an ethanol dose leaves unchanged through breath, urine, and sweat; the rest must be oxidized [1].

Second, women generally produce less gastric first-pass ADH activity than men, so a larger fraction of the same dose reaches the systemic circulation — an effect that is on top of, not explained by, differences in body water [3]. This is one reason women reach higher blood levels from an identical pour and show alcohol-related organ injury at lower lifetime doses.

Acetaldehyde is the part nobody markets

Acetaldehyde is a small, electrophilic, extraordinarily reactive molecule. It does not simply pass through. It forms covalent adducts — permanent chemical attachments — with proteins and with DNA.

• Protein adducts. Acetaldehyde-modified liver proteins are demonstrable by immunohistochemistry in the centrilobular zone of the livers of people who drink, and are essentially absent in non-drinkers. Critically, the immune system recognizes these modified proteins as foreign and generates antibodies against them — an autoimmune-flavored mechanism operating in ordinary drinkers [12].
• DNA adducts. Acetaldehyde forms N²-ethylidene-deoxyguanosine and related lesions. In animal models, hepatic adduct burden rises about fourfold in ALDH2-heterozygous animals and tenfold in ALDH2-null animals compared with wild type, on identical alcohol exposure — showing that adduct load tracks directly with how long acetaldehyde lingers [10]. In humans, a single controlled alcohol dose reaching 0.11% blood alcohol produced roughly a 160-fold rise in ethyl-DNA adducts in oral cells [11].
• Local exposure in your mouth. Oral bacteria carry their own alcohol dehydrogenase. After a 30-second rinse with an alcoholic beverage — swallowing nothing — salivary acetaldehyde averaged 353 µM, far above the ~100 µM threshold associated with DNA damage in vitro, and was still elevated five minutes later [13,14]. This happens at the first sip, at any dose.

The International Agency for Research on Cancer classifies not only alcoholic beverages but acetaldehyde associated with the consumption of alcoholic beverages as a Group 1 human carcinogen — the same evidentiary tier as tobacco smoke and asbestos [9].

The third pathway that turns on when you drink regularly

Beyond ADH there is a second, inducible system: CYP2E1, the microsomal ethanol-oxidizing system first characterized by Charles Lieber [5]. At low, occasional intake it contributes little. But CYP2E1 is inducible: in human volunteers, measurable induction appeared after roughly 40 grams of ethanol daily for one week, and chronic drinkers can carry up to tenfold higher CYP2E1 activity than non-drinkers [4].

CYP2E1 is a leaky enzyme. It generates reactive oxygen species as a byproduct of its own catalytic cycle, which is why chronic drinking creates oxidative stress even between drinks [16]. And — this is the part that matters enormously for the rest of this article — CYP2E1 is the enzyme that bioactivates a long list of other chemicals into their toxic forms. We will return to that under “the detox queue.”

Is oxidative stress detectable at ordinary doses? Yes. In healthy volunteers given escalating single oral doses of ethanol, urinary F₂-isoprostanes — the gold-standard in vivo marker of lipid peroxidation — rose in a clean, stepwise dose-response fashion from a baseline of 116 pg/mg creatinine to 491 pg/mg at the highest dose, with clear intermediate rises at each moderate dose along the way [15]. There was no threshold below which nothing happened.

The NAD+ Crash and the Fat-Burning Shutdown

Both steps of ethanol oxidation consume NAD+ and generate NADH. Metabolize a couple of drinks and you have shoved the hepatic NADH:NAD+ ratio sharply toward the reduced end for hours. NAD+ is not a minor cofactor; it is the currency your mitochondria use to run the citric acid cycle and beta-oxidation, and it is the required substrate for the sirtuins (SIRT1, SIRT3) that govern mitochondrial biogenesis through PGC-1α.

The consequences cascade [2,27]:

• Fat oxidation stops. With NAD+ scarce, beta-oxidation stalls and fatty acids are re-esterified into triglyceride — the origin of the fat droplets that appear in liver cells long before any diagnosis.
• Gluconeogenesis is inhibited, which is why alcohol on an empty stomach can drop blood sugar and why the 3 a.m. wake-up is so often a glucose event as much as a nervous-system one.
• Lactate accumulates, and lactate competes with uric acid for renal excretion.
• Uric acid rises — and not only from the lactate effect. Ethanol accelerates the breakdown of adenine nucleotides (ATP → ADP → AMP → ... → uric acid) independent of dietary purines [17]. This is the biochemistry behind a gout flare after a night out, and it is also a window into the ATP problem described in the next section.
• Sirtuin signaling is constrained, blunting the very pathways we are usually trying to upregulate in metabolic and longevity work.

How long does fat burning actually stay off?

This was quantified precisely, in a metabolic chamber, in 1992. Adding ethanol to the diet at 36% of energy suppressed whole-body lipid oxidation by 36 ± 3%; substituting ethanol isocalorically for other macronutrients suppressed it by 31 ± 7% — and the suppression persisted for approximately 15 hours [18].

Read that again, because it reframes the whole calorie conversation. It is not primarily that wine has calories (it has about 7.1 kcal per gram of ethanol, plus residual sugar). It is that for most of the following day, your body is preferentially burning acetate and is not burning fat. Have a drink most nights and you have installed a nightly switch that keeps the fat-oxidation machinery mostly idle.

Acetate itself is not inert either. In heavy drinkers, brain acetate concentration was about 80% higher than in light drinkers, with a significantly elevated cerebral metabolic rate of acetate — the brain adapts to running partly on alcohol’s exhaust [19]. And in an elegant isotope-tracing study, alcohol-derived acetate was shown to be incorporated directly into histone acetylation in the hippocampus, altering gene expression in memory and reward circuitry — a molecular mechanism by which drinking rewrites the brain’s epigenetic settings [20].

The Energy Drain You Can’t See: Endogenous Fructose and ATP

Here is a mechanism most people have never heard of, and it explains the “I slept eight hours and I’m still flat” complaint better than anything else.

When you drink, the liver activates the polyol pathway: aldose reductase converts glucose to sorbitol, and sorbitol dehydrogenase converts sorbitol to fructose. Your body manufactures its own fructose, even if you ate no sugar at all. That endogenous fructose is then phosphorylated by ketohexokinase (fructokinase), an enzyme with a peculiar and consequential property: unlike glucose metabolism, it has no negative feedback and it consumes ATP rapidly, driving intracellular phosphate and ATP down and generating uric acid as a byproduct.

The evidence has moved from hypothesis to demonstration. Elevated fructose and uric acid, driven by aldose reductase, were shown to contribute to alcoholic liver disease in both experimental models and human liver tissue [21]. And in 2025, hepatocyte-specific deletion of ketohexokinase in mice prevented ethanol-induced steatosis, inflammation and fibrosis — while also reducing voluntary alcohol intake and alcohol-associated reward behavior, implicating the same node in both the liver injury and the pull to drink again [22].

Now put the pieces together. Your brain uses roughly 20% of your resting energy budget. Dopamine neurons, which drive motivation and focus, are among the most energy-hungry cells you own. Peripheral nerve axons are absurdly long and depend entirely on mitochondrial ATP delivered down that axon. When the cellular energy charge drops and stays down — ATP consumed, NAD+ depleted, fat oxidation off, mitochondria under oxidative load — the symptoms are not mysterious. They are exactly the ones people describe: flat, foggy, unmotivated, sore, slow to recover.

Why Alcohol Cuts the Detox Line

This is the argument I make most often in clinic, and it is the reason I am stricter about alcohol than most clinicians. I want to make it precisely, because the version of it that circulates online is partly right and partly overstated, and I would rather give you the accurate version.

The correct version of the claim

Your liver does not have unlimited parallel processing. It has a finite pool of shared resources: NAD+, glutathione, cysteine, ATP, methyl groups from the one-carbon cycle, and a set of enzymes and transporters that must be shared among everything you are exposed to. Ethanol does not politely wait its turn. Because there is nowhere to store it and because it is toxic at rising concentrations, it is oxidized preferentially and immediately. Everything else — hormone clearance, xenobiotic conjugation, fat handling — is done with whatever capacity is left.

Here is the part with the strongest human evidence, and it is more alarming than the general claim: alcohol doesn’t merely occupy capacity. It induces CYP2E1, and CYP2E1 is the enzyme that converts a long list of other chemicals into their more toxic forms.

The most dramatic human demonstration of the same principle comes from occupational toxicology. Among vinyl chloride workers — vinyl chloride is bioactivated by CYP2E1 — high exposure alone carried risk, and alcohol alone carried risk. But high vinyl chloride exposure combined with alcohol above 60 g/day produced an odds ratio for hepatocellular carcinoma of 409, with 85% of the cancer burden attributable to the interaction rather than to either exposure alone [24]. That is an industrial-level exposure and the magnitude should not be generalized to a household. But the principle it demonstrates — that alcohol converts a tolerable chemical exposure into a far more dangerous one — is exactly the principle that concerns me in a patient carrying a chemical burden.

The same logic applies to a documented food-toxin interaction: aflatoxin B1 exposure raises hepatocellular carcinoma risk specifically in the setting of alcohol consumption or hepatitis C [25].

The glutathione and methylation drain

Glutathione is the workhorse of Phase II conjugation and the primary intracellular defense against the reactive species CYP2E1 generates. Alcohol depletes it, and depletes the mitochondrial pool disproportionately [16]. Glutathione is also the molecule we lean on hardest when we are moving mycotoxins, metals, and solvents — which is to say, alcohol drains exactly the tank we are trying to fill during a detoxification protocol. It is not a coincidence that the one interventional trial worth citing here found that adding N-acetylcysteine — a glutathione precursor — to steroids in severe alcoholic hepatitis reduced one-month mortality (8% vs 24%), even though the six-month primary endpoint did not reach significance [26].

Methylation takes a parallel hit. Alcohol impairs intestinal folate absorption, increases renal folate loss, and directly inhibits methionine synthase — lowering SAMe, raising homocysteine and SAH, and constraining every methylation-dependent reaction downstream, from DNA methylation to catecholamine and estrogen handling [28,27]. If you carry MTHFR variants or already have a marginal methylation picture, nightly alcohol is working directly against you.

And NAD+ — already discussed — is not just an energy cofactor. It is the substrate for PARP-mediated DNA repair and for the sirtuins. Alcohol-driven oxidative stress activates PARP-1, which consumes still more NAD+, further constraining SIRT1 [27]. Repair capacity and detox capacity draw from the same well.

Where I have to correct the popular version

The Backlog: What Your Liver Was Already Carrying

The reason detox capacity matters more for my patients than for the average person is that my patients are, almost by definition, already at capacity.

Start with the ambient chemical picture. The EPA’s TSCA inventory listed 86,862 chemicals as of its August 2025 update, of which 42,578 are designated “active” in US commerce [29]. The CDC’s National Report on Human Exposure to Environmental Chemicals has been measuring hundreds of these in Americans’ blood and urine continuously since 1999, across metals, VOCs, PFAS, flame retardants, phthalates, pesticides and combustion byproducts [30]. This is not a fringe claim; it is federal biomonitoring data. Nobody in a modern country has a clean baseline.

Layer onto that the specific burdens that bring people to a functional medicine practice: a water-damaged building and the mycotoxins that came with it; mercury from amalgams or fish; lead and cadmium accumulated over decades; a tick-borne infection; the residue of glyphosate and plastics. Then consider two additional facts that come as a surprise to most wine drinkers.

What I can and cannot claim about mold illness specifically

I want to be scrupulous here, because this is where clinical conviction most often outruns published evidence.

There is no published study examining alcohol consumption in patients with CIRS, mold-related illness, or chronic biotoxin illness. None. Anyone who tells you there is has not looked. What exists is a chain of well-documented mechanisms — CYP2E1 induction and cross-bioactivation, glutathione and NAD+ depletion, one-carbon disruption, gut permeability and endotoxin translocation into an already TLR4-primed innate immune system, plus the ochratoxin content of the beverage itself — each link of which is real, and which together make a strong mechanistic case. That is an inference, and I will call it one.

But here is the thing about inference in clinical medicine: we act on it constantly, and we should, when the mechanism is solid, the downside of the intervention is zero, and the alternative is asking a struggling patient to keep paying a metabolic tax for a beverage. Removing alcohol costs nothing but habit. There is no deficiency syndrome from not drinking.

We are asking the liver to do the hardest work of its life. It seems unreasonable to hand it a solvent every evening while we do.

The same reasoning applies to chronic infection. Alcohol arrests NK cell maturation, reduces NK numbers and cytotoxicity, depletes mucosal MAIT cells, and dysregulates the innate lymphocyte populations that provide surveillance against persistent pathogens [35]. There is no study of alcohol and Borrelia control, and I will not pretend there is. But if we are asking an immune system to do sustained work against a chronic infection, nightly suppression of its innate arm is a strange thing to permit.

Sleep: Where You’ll Feel It First

If you take only one section from this article, take this one — because it is the fastest, most personally verifiable, and most consequential effect, and because alcohol’s reputation as a sleep aid is exactly backwards.

Alcohol is a sedative. It shortens the time it takes to fall asleep and increases slow-wave sleep in the first half of the night. That is the part people notice, and it is why the myth persists. But sedation is not sleep, and the second half of the night pays for the first.

What the meta-analyses show, by dose

The most recent and most useful synthesis pooled 27 studies and did something previous reviews had not: it separated the effects by dose [36].

That last row deserves emphasis. The reason people believe alcohol helps them sleep is that they are measuring the wrong thing. Duration is preserved; architecture is not. Older polysomnography work found the same split: at a peak blood alcohol of 0.10%, first-half slow-wave sleep rose significantly while first-half REM density fell significantly, and second-half light-stage fragmentation increased [38,37].

Why you wake at 2:40 a.m.

Three mechanisms converge on the same hour.

• Neurochemical rebound. Alcohol acutely potentiates GABA-A (inhibition) and suppresses glutamate/NMDA signaling (excitation). As blood alcohol falls through the night, that balance snaps back the other way — a mini-withdrawal, with glutamatergic hyperexcitability arriving precisely when you should be in your deepest consolidation window [68,69]. This is also the neurochemistry of “hangxiety”: the anxious, jangled feeling the next morning.
• Melatonin suppression. In a double-blind randomized crossover trial, ethanol suppressed nocturnal melatonin by 41% at midnight and 33% at 1 a.m., in a dose-dependent way, with the higher dose also raising plasma norepinephrine [41].
• Autonomic activation. Rather than the parasympathetic quiet that repair requires, the nervous system stays subtly switched on all night (below).

Hormones and breathing

Two more effects deserve mention because they hit precisely the goals people are usually pursuing.

Growth hormone. In a classic controlled study, alcohol suppressed nighttime plasma growth hormone by 70–75% — total integral, mean hourly rate, and peak alike — returning to normal only on the withdrawal night [42]. Nocturnal GH is a principal driver of overnight tissue repair and body-composition maintenance. If you are lifting weights and trying to preserve lean mass in midlife, this is not a small detail.

Breathing. Alcohol relaxes upper airway dilator muscles. A meta-analysis of 21 studies found a 25% higher relative risk of obstructive sleep apnea in drinkers (RR 1.25, 95% CI 1.13–1.38) [43], and pooled cohort data showed alcohol increased the apnea-hypopnea index by 3.98 events per hour and lowered the night’s minimum oxygen saturation by 2.72% [44]. If you already have mild sleep apnea — and a great many people with fatigue and brain fog do — an evening drink converts “mild” into something meaningfully worse, every night.

The next day, at zero blood alcohol

A systematic review and meta-analysis pooling 1,163 participants found significant next-day cognitive impairment at zero or near-zero blood alcohol: short-term memory (Hedges’ g = 0.64), long-term memory (g = 0.59), sustained attention (g = 0.47), and psychomotor speed (g = 0.66) [45]. Those are moderate effect sizes on exactly the faculties people are trying to reclaim when they tell me they have brain fog.

What the Wearable Data Shows — and Why It’s So Persuasive

For fifteen years the argument about “one drink” was hard to settle because the effect is invisible to the drinker. That changed when millions of people started wearing continuous heart-rate and HRV monitors.

Controlled laboratory work confirms it is not an artifact of who drinks. In a dose-response study, nocturnal heart rate ran about 4% faster than placebo on a low-dose night (one drink for women, two for men) and 14% faster on a high-dose night; SDNN and RMSSD — the standard HRV metrics — were significantly reduced on both nights, and baroreflex sensitivity was suppressed across all six monitored hours at the high dose [39].

The clinical significance is this: heart rate variability is our best non-invasive proxy for parasympathetic (rest-and-repair) tone. A night with suppressed HRV is a night in which the body was sedated but was not recovering. Multiply by six nights a week and you have removed most of the recovery windows from the week — which is precisely what patients describe when they say they feel like they are running on fumes despite “doing everything right.”

Training, muscle, and recovery

Athletes have known the direction of this effect for a long time; the mechanisms are now well characterized, and the doses matter.

• Muscle protein synthesis. Alcohol consumed after a training session suppressed the peak rise in myofibrillar protein synthesis by 24% when co-ingested with protein and 37% when consumed with carbohydrate instead of protein, with blunted mTOR signaling at 2 and 8 hours [46]. Honest note: that study used a binge-level dose (1.5 g/kg, roughly twelve drinks). It shows the mechanism dramatically; it does not prove the same magnitude at two drinks.
• Force recovery. After eccentric muscle damage, a higher dose (~1 g/kg) meaningfully worsened torque recovery at 36 hours, while a low dose (0.5 g/kg) did not differ from alcohol-free [47,48]. There is a real dose threshold here and I will not overstate it.
• Hydration. Post-exercise fluid retention fell progressively with beverage alcohol concentration — 59% at 0% ABV down to 41% at 4% ABV — although measured vasopressin did not differ, so the mechanism is not simply ADH suppression [49].
• Cortisol up, testosterone down across the reviewed resistance-training literature [48], compounding the growth hormone suppression above.

Put the recovery findings next to the sleep findings and the picture is coherent: the acute muscular effects of two drinks are modest, but the nightly loss of deep sleep, REM, growth hormone, and parasympathetic recovery time is not modest at all — and it compounds.

My Own Six-Month Experiment

I should tell you that I did not arrive at this position from the literature. I arrived at it from my own data, and then went looking for the mechanism.

I was never a heavy drinker. A casual drink, a cocktail with dinner, one or two a week — the kind of intake that every guideline in the world would call harmless and that I would previously have described as a non-issue. But I wear a tracker, and trackers are unsentimental. Every time I had a drink, the same pattern appeared: less deep sleep, more fragmentation, a lower recovery score in the morning, a resting heart rate that stayed elevated overnight. And the next day was measurably worse — harder to focus, less capacity in training, a fatigue that did not match the amount of sleep I had recorded.

For a long time I explained it away. One or two drinks a week cannot possibly matter. Then I actually looked at the data across months rather than nights, and the pattern was too consistent to argue with.

So for the past six months I have had no alcohol at all. What changed:

• Deeper, less fragmented sleep, with recovery scores that stay in a range I used to hit occasionally
• Higher VO₂ max
• Better muscle mass and easier weight maintenance
• Noticeably better training capacity and faster recovery between sessions
• More stable focus and energy through the day

This is an n of 1, uncontrolled and unblinded, and I would not publish it as evidence. But it is worth telling for two reasons. First, it is exactly what the mechanisms above predict — which is what made me take the mechanisms seriously. Second, it demonstrates the central practical point of this article: you cannot know what alcohol is costing you until you remove it and look. I thought one or two drinks a week were free. They were not free. They were simply cheap enough that I never noticed the bill.

Two Patients Who Changed My Mind Further

Details are altered to protect privacy, but the clinical arcs are exactly as they happened. I include them because the pattern in both is the same: alcohol was the variable nobody was counting.

Case one: the low-calorie cans

A man doing everything right — training hard, taking his supplements, showing up for follow-ups — could not lose weight. His drink of choice was the canned mojito: low-calorie, refreshing, genuinely tasty, and marketed as the sensible option. He was having six or seven a day.

Two things surfaced when we actually did the arithmetic with him. First, the sugar: about 20 grams per can, which at his intake was roughly 120–140 grams of sugar a day — something close to a kilogram of sugar a week, entirely invisible to him because the can said “light.” Second, the ethanol underneath it, six or seven servings a day, every day.

His picture made complete biochemical sense once you lay it against the mechanisms in this article. A history of gout and gout attacks — predictable from two directions at once: the fructose in the cans and the endogenous fructose generated by the alcohol, both burning ATP through ketohexokinase and generating uric acid, plus lactate from the NADH shift blocking renal urate excretion [17,21]. Rising liver enzymes and, eventually, fatty liver — predictable from NAD+ scarcity stalling fat oxidation while fructose drove new fat synthesis. Weight that would not move despite real effort — predictable from fat oxidation being suppressed for most of every day [18], appetite being driven upward, and sleep being degraded nightly.

What actually changed his mind was not my lecture. It was the ultrasound and the enzymes. Seeing fatty liver on a report made the abstraction concrete, and he quit. He switched to Athletic Brewing and to non-alcoholic mixed drinks and mocktails — and notably, he did not have to give up the ritual, only its contents.

What followed was fast and, by now, familiar: he felt better almost immediately, slept better, put on muscle mass, lost weight, and his liver enzymes came down. Nothing else about his program had changed. We had simply stopped subtracting.

Case two: the ferritin nobody could explain

A woman came to us with persistently elevated liver enzymes. We had her on a supplement protocol. When she saw her internist, she was referred to a hepatologist — entirely appropriate — who told her to stop all supplements and told her that three or four glasses of wine a week was almost certainly not the issue.

She stopped the supplements. The liver enzymes stayed elevated.

Then an ultrasound showed fatty liver. And there was a third finding that kept recurring and that nobody was connecting to anything: her ferritin kept climbing. It rose high enough that she underwent therapeutic phlebotomy to bring it down.

It was only after we sat down and walked her through what low-dose alcohol actually does — the same material in this article — that she agreed to stop entirely. She is now doing much better: feeling better, sleeping better, losing weight.

In retrospect, I think the sequence was misread from the beginning. The supplements were never the likely cause — and stopping them, which was the one intervention actually tested, changed nothing. The far more parsimonious explanation is that the alcohol was driving the inflammation, the hepatic fat, the enzyme elevation, and the ferritin all at once.

Nerves: Why Alcohol and Neuropathy Don’t Mix

This is the part of the conversation that made me want to write the article. A patient with burning feet was told alcohol was irrelevant to her nerves. It is not irrelevant. It may be the single most modifiable variable she has.

The old model was wrong, and the correction matters

For most of the twentieth century, alcohol-related neuropathy was taught as a nutritional disease: heavy drinkers eat poorly, become thiamine deficient, and their nerves suffer. Correct the vitamin, correct the nerve. That model let clinicians reassure moderate drinkers who were eating well.

Then Koike and colleagues did the study that should have ended it. They compared 64 patients with alcoholic neuropathy — 36 of whom had normal thiamine status — against patients with pure thiamine-deficiency neuropathy, with sural nerve biopsies in 56 of them. The two conditions were clinicopathologically distinct. Pure alcoholic neuropathy was sensory-predominant with selective small-fiber (pain and temperature) loss, and it occurred with thiamine levels that were entirely normal [51,52].

Alcohol damages nerves directly. It does not require a vitamin deficiency to do it.

The small-fiber selectivity was confirmed with skin biopsy: intraepidermal nerve fiber density at the calf was 4.85 fibers/mm in the alcohol group versus 8.83 in controls (p<0.0001), with the proximal thigh spared — the classic length-dependent, dying-back pattern — and with sweat gland nerve fiber density also reduced [53]. Small fibers are the ones that produce burning, tingling, and temperature and autonomic symptoms. They are the first to go and the hardest to see on standard nerve conduction studies, which is why so many patients are told their tests are “normal.”

How common is this? A meta-analysis of 87 studies found neuropathy on clinical examination in 44.2% of people with alcohol use disorder and on nerve conduction studies in 46.3%, with painful neuropathy in 42% [50]. In another cohort, 58% had polyneuropathy, with mixed, large-fiber, or isolated small-fiber patterns [54].

The mechanisms

• Direct axonal toxicity from ethanol and acetaldehyde, with adduct formation and oxidative injury in dorsal root ganglion neurons [55].
• Protein kinase C epsilon. In ethanol-fed rats, C-fiber mechanical thresholds fell and PKCε rose about 50% in the dorsal root ganglia; blocking PKCε reversed the pain phenotype [56]. (Animal data — a mechanism, not a human proof.)
• Impaired mitochondrial function and axonal transport in the longest, most energy-dependent cells in the body — which connects the ATP story from earlier directly to the feet.
• Schwann cell and myelin disruption, consistent with the segmental demyelination seen on biopsy [55].

Nutrient depletion still matters — it is additive, not alternative

The direct-toxicity finding does not mean nutrients are irrelevant. It means they stack on top.

• Thiamine. Chronic alcohol reduces intestinal thiamine absorption by lowering transporter Vmax, and impairs hepatic phosphorylation to the active form [58,59]. Thiamine pyrophosphate is the required cofactor for transketolase, pyruvate dehydrogenase, and α-ketoglutarate dehydrogenase — the enzymes that let a neuron make ATP at all. α-KGDH is itself inactivated by oxidative stress, creating a self-amplifying loop of energy failure and free-radical damage in vulnerable neurons [62].
• Vitamin B6. Acetaldehyde — not ethanol — accelerates degradation of pyridoxal-5-phosphate. In 66 alcohol-using subjects without liver or hematologic abnormality, 53% had plasma PLP below 5 ng/mL [60].
• Magnesium. Pooled prevalence of hypomagnesemia in alcohol use disorder was 44.4%, driven predominantly by renal wasting rather than poor intake [61].
• B12 and folate transporter activity are both reduced with chronic alcohol exposure [59].

And alcohol is the second most common cause of distal symmetric polyneuropathy in the United States after diabetes — with essentially non-overlapping mechanisms, meaning the damage from the two is expected to be additive [57]. If you have diabetic or prediabetic neuropathy and you drink nightly, you are running two independent nerve-damaging processes at once.

Recovery data is also thinner than I would like. Clinical series report that months to a few years of abstinence produce clinical and electrodiagnostic improvement, though residual neuropathy is common — but no rigorous controlled longitudinal trial has tracked recovery with skin biopsy or nerve conduction endpoints. In practice, when we remove alcohol and simultaneously repair thiamine, B6, B12, and magnesium status, control glucose, and support mitochondrial function, patients frequently report a reduction in burning within weeks to a few months. I present that as clinical observation, not trial data.

The Brain: Fog, Mood, and Measurable Volume

The old reassurance about the brain was that moderate drinking might even be protective. That claim has now been dismantled by two independent lines of evidence: high-resolution imaging in very large cohorts, and genetic (Mendelian randomization) analysis.

Imaging: the effect starts at one drink

A separate 30-year longitudinal cohort found that people drinking 14–21 units/week — squarely within older “moderate” guidance — had 3.4-fold higher odds of right hippocampal atrophy versus abstainers, rising to 5.8-fold above 30 units/week, with faster decline in lexical fluency. Light drinking showed no protective effect [65]. And moderate intake above roughly 7 units/week was associated with higher brain iron in the basal ganglia, which in turn tracked with slower executive function [66].

Genetics: the “protective” signal disappears when you remove the confounding

The most important recent study on this question combined 559,559 adults from the Million Veteran Program and UK Biobank with Mendelian randomization drawn from genome-wide data on 2.4 million people [67]. The observational analysis reproduced the familiar U-shaped curve, with non-drinkers and heavy drinkers both at higher dementia risk than light drinkers. The genetic analysis did not. It showed a monotonic increase in dementia risk with greater alcohol consumption — a 1 SD increase in log drinks per week associated with 15% higher dementia risk.

And the authors identified the reason for the discrepancy directly in their data: people who went on to develop dementia had been reducing their drinking over time. The apparent protection of light drinking was, in significant part, early cognitive decline causing people to drink less — reverse causation, visible in the longitudinal data once you look for it.

Why your thinking feels slow the day after

Beyond structural change, three acute mechanisms explain the fog:

• Energy. Neurons, and dopamine neurons in particular, are exquisitely ATP-dependent. The NAD+ and ATP effects described earlier land hardest on the most metabolically demanding tissue you own.
• Neuroinflammation. Alcohol activates TLR4 signaling on microglia; in TLR4-deficient animals, ethanol-induced glial activation, inflammatory mediator production, and apoptosis were abolished [70]. Gut-derived endotoxin (next section) feeds the same receptor [71]. This is rodent work, and I flag it as such — but it is a coherent mechanistic account of “brain fog” as a neuroimmune state rather than a psychological one.
• Dopamine. PET imaging shows reduced striatal D2 receptor availability in people with chronic heavy alcohol use [72]. Less receptor availability plus less ATP to run dopamine signaling equals exactly the flat, unmotivated, procrastinating state people describe — and which they are often tempted to treat with another drink.

That last loop is worth naming explicitly. Alcohol reliably produces a next-day state of lower mood, higher anxiety, and lower resilience. That state increases the appeal of the thing that caused it. This is not weakness of character; it is pharmacology with a feedback loop, and understanding it is often the thing that lets people step out of it.

The Gut Barrier, Endotoxin, and Mast Cells

Alcohol’s effect on the gut is one of the more clinically useful parts of this story, because it explains why people with bloating, food reactivity, histamine symptoms, and flushing so often improve when alcohol comes out — even when nothing else changes.

Barrier and endotoxin

The intestinal epithelium is one cell thick and depends on ATP-hungry tight junction complexes to stay sealed. Acetaldehyde disrupts these directly: it increases tyrosine phosphorylation of ZO-1, E-cadherin and β-catenin, pulling tight junction proteins away from the cell junctions, and alcohol upregulates iNOS, generating peroxynitrite and further cytoskeletal damage [74].

When the barrier loosens, bacterial lipopolysaccharide translocates into portal blood and activates Kupffer cells via CD14/TLR4, driving TNF-α and hepatic inflammation — and the same TLR4 signal reaches the brain [71].

Histamine, mast cells, and MCAS — with an important correction

Many patients with mast cell activation syndrome notice that wine is one of their worst triggers. That observation is valid; the usual explanation for it is wrong.

The common claim is that alcohol inhibits diamine oxidase (DAO), the enzyme that degrades histamine in the gut. A careful study tested exactly that and found that relevant concentrations of ethanol, acetaldehyde and acetate do not inhibit recombinant human DAO; an apparent effect from disulfiram turned out to be assay interference [79].

The better-supported explanation has two parts. First, many alcoholic beverages — wine especially, and some beers — contain preformed histamine from fermentation, so you are ingesting the offending molecule directly. Second, and more importantly: acetaldehyde triggers mast cells to degranulate. In human airway mast cells, acetaldehyde directly induced histamine release [78]. Which means the histamine problem is not only what is in the glass, but what your own liver makes from it.

For an MCAS patient, that is a decisive difference: it means low-histamine wine, sulfite-free wine, and DAO supplements do not solve the problem, because the trigger is generated endogenously during metabolism.

Weight, Blood Pressure, and the Metabolic Story

I will be honest with you about this section, because overselling it is how credibility gets lost: the epidemiology on alcohol and body weight is genuinely mixed. Not every nightly drinker gains weight. Some large studies find weak or inconsistent associations.

But “the population association is noisy” is a very different statement from “it isn’t affecting you.” Here is what is not in dispute.

• Fat oxidation is suppressed by about a third for roughly fifteen hours after drinking [18]. Drink most evenings and your body spends most of its life not burning fat.
• Alcohol contributes 7.1 kcal/g, and those calories are obligatorily oxidized first — they cannot be stored, so everything else in the meal is spared for storage.
• It drives eating. In a mechanistic study, alcohol activated hypothalamic AgRP hunger neurons, and blocking those neurons abolished alcohol-induced overeating — a specific neural mechanism for the “drunk munchies” that is not merely disinhibition [80]. (Mouse data.)
• It wrecks the sleep that regulates appetite. The REM and deep-sleep losses documented earlier reliably shift ghrelin/leptin balance and next-day food choice.
• It suppresses nocturnal growth hormone by 70–75% [42], the hormone most responsible for overnight body-composition maintenance.

For a patient with insulin resistance, disrupted sleep, midlife hormonal change, and a sluggish metabolism — which is nearly everyone who comes to me about stubborn weight — that is five independent headwinds, every night.

Blood pressure: linear, with no threshold

No threshold means the curve does not start at “heavy.” It starts at the first drink and rises from there. If your blood pressure has been creeping up alongside stubborn weight and poor sleep, alcohol is a bigger lever than most people assume.

Atrial fibrillation: the cleanest causal evidence in the whole article

Most of what I have cited is observational or mechanistic. Here is a randomized controlled trial. 140 regular drinkers with paroxysmal or persistent atrial fibrillation were randomized to abstain or continue. The abstinence group cut intake from 16.8 to 2.1 drinks/week; controls went from 16.4 to 13.2. Over six months, AF recurred in 53% of abstainers versus 73% of controls (hazard ratio 0.55, 95% CI 0.36–0.84, p=0.005), and AF burden — the proportion of time in AF — was significantly lower (median 0.5% vs 1.2%, p=0.01) [85].

Randomize people to stop drinking and their arrhythmia improves. That is as close to proof of causation as clinical medicine gets.

Estrogen and hormonal clearance

In 1,864 postmenopausal women in the Women’s Health Initiative, women on menopausal hormone therapy drinking 7+ drinks/week versus under 1 had 26% higher unconjugated estrone and 26–29% higher estradiol; among never/former hormone users, liquor was associated with 19–32% higher parent estrogens and metabolites [86]. The authors interpreted this as increased estrogen production rather than impaired clearance — a nuance worth preserving. Either way, for a woman managing estrogen dominance, fibroids, endometriosis, or breast cancer risk, this is directly relevant.

Your Liver’s Early Warning System

Patients often tell me their liver enzymes are “fine, so the wine must be fine.” Two things are wrong with that.

First, normal is not the same as optimal, and reference ranges were built from populations that drink. GGT is the most alcohol-sensitive of the standard enzymes and typically rises with regular intake before ALT or AST move at all. An ALT that has drifted from 14 to 28 is still “normal” and is still a signal.

Second, and more importantly, the fat comes first. Steatosis — those triglyceride droplets accumulating because NAD+ scarcity stalled beta-oxidation — develops long before enzymes rise, and enzyme elevation is a comparatively late event.

The 2023 international nomenclature consensus is worth knowing about, because it created a diagnostic category that describes many of my patients precisely. Under the new system, steatotic liver disease is subdivided by cause: MASLD (metabolic dysfunction-associated) applies below roughly 20 g/day of alcohol in women and 30 g/day in men; MetALD is the new intermediate category for people with cardiometabolic risk factors who drink in the range of roughly 20–50 g/day (women) or 30–60 g/day (men); above that, it is alcohol-associated liver disease [81].

Look at those numbers against the pour arithmetic from the beginning of this article. Two generous glasses of 13.5% wine is roughly 51 g. A woman drinking that six nights a week, with any degree of insulin resistance, is not in a gray zone — she is in MetALD, a named liver disease category, while being told her drinking is moderate.

Ferritin: the marker that gets misread

One lab deserves its own note, because I have now seen it send more than one patient down the wrong diagnostic road (see case two above). Ferritin rises in drinkers for two separate reasons, and neither is hemochromatosis.

It is an acute-phase reactant, so hepatic inflammation pushes it up regardless of iron stores. And alcohol genuinely increases iron absorption: alcohol-metabolism-driven oxidative stress down-regulates hepcidin, the hormone that restrains iron uptake, increasing duodenal iron transporter expression [82], and alcohol intake tracks with higher indices of iron stores including ferritin in population data [83]. Free iron then catalyzes further oxidative injury, which raises inflammation, which raises ferritin again.

So a rising ferritin in someone who drinks is usually a signal about the alcohol, not an independent iron disorder. Check transferrin saturation before assuming overload, and remove the alcohol before committing someone to phlebotomy.

The Cancer Question and “No Safe Level”

I include this not to frighten anyone but because it is the single largest change in the alcohol evidence base in the past decade, and most people have not heard about it.

Alcoholic beverages are classified by IARC as a Group 1 human carcinogen — the highest evidentiary category, shared with tobacco and asbestos — with sufficient evidence for causal links to seven cancers: oral cavity, pharynx, larynx, esophagus, liver, colorectum, and female breast [9,90].

The dose-response starts below “moderate”

The definitive dose-response meta-analysis, pooling 572 studies, quantified risk at light drinking — defined as up to one drink a day [87]:

The breast cancer figure deserves a moment. A relative risk of 1.04 at light drinking sounds trivial — and for one woman on one night, it is. But breast cancer is common, so a small relative increase applied to a large baseline produces a large number of cases, and the effect is already measurable at the lowest exposure category studied. The confidence interval excludes 1.0. There is no observed threshold below which the association disappears.

Globally, an estimated 741,300 new cancer cases in 2020 were attributable to alcohol — 4.1% of all new cancers. Of those, heavy drinking accounted for 346,400 cases and risky drinking 291,800 — but moderate drinking (under 20 g/day) accounted for 103,100 cases (13.9%), and drinking up to just 10 g/day — less than one standard drink — accounted for 41,300 cases [88].

Biological aging

There is also a signal at the level of aging biology itself. In two Mediterranean cohorts, a DNA-methylation score for alcohol exposure and GGT were both significantly associated with multiple epigenetic aging clocks, and alcohol consumption correlated with shorter telomere length [103]. The authors are careful to note these reflect cumulative exposure rather than proven acceleration of biological aging, and I will be careful too — but it is consistent with everything above.

The 2025 Surgeon General’s Advisory

This is where international guidance has been heading for several years. The WHO stated in 2023 that “there is no safe amount that does not affect health,” noting that half of all alcohol-attributable cancers in the European region arise from light and moderate drinking [90]. Canada’s 2023 guidance replaced a single limit with a risk continuum: 1–2 drinks/week is low risk; at 3–6 drinks/week the risk of several cancers, including breast and colon, increases; at 7+ per week cardiovascular risk rises significantly [91].

But What About Red Wine? The J-Curve, Honestly

For thirty years the dominant story was a J-shaped curve: abstainers do worse than light drinkers, heavy drinkers worse still, therefore a drink or two a day is protective. Resveratrol, the French paradox, the cardiologist who recommends a glass of red. Let me explain, carefully, why that curve appears in observational data and why it largely evaporates under better methods.

Problem one: who is in the “abstainer” group?

Most cohort studies compare drinkers against “non-drinkers.” But that reference group is contaminated. It contains people who quit because they became ill (the “sick quitter”), people who never drank because of chronic disease, frailty or medication, and former heavy drinkers. Those people have elevated mortality for reasons that have nothing to do with their current abstinence — which artificially inflates the reference group’s risk and makes moderate drinkers look protected.

Problem two: moderate drinkers are different people

In Western cohorts, moderate drinkers tend to be wealthier, better educated, more socially connected, more likely to exercise and to see a doctor. Every one of those independently predicts lower mortality. Statistical adjustment helps, but never fully removes, that confounding.

The method that solves both problems

Mendelian randomization uses genetic variants that affect alcohol metabolism — ALDH2 and ADH1B in East Asian populations, or polygenic scores — as natural randomizers. Your genotype was assigned at conception, is unrelated to your income or education, and cannot be changed by getting sick. It is, in effect, a randomized trial nature ran for us.

• In over 500,000 Chinese adults, conventional analysis showed the familiar protective association for moderate drinking and stroke. The genetic analysis showed the opposite: alcohol raised blood pressure and stroke risk dose-dependently, and the authors concluded that “the apparently protective effects of moderate alcohol intake against stroke are largely non-causal” [94].
• In 371,463 UK Biobank participants, genetic instruments showed that alcohol consumption of all amounts was associated with increased cardiovascular risk, with light intake carrying minimal and heavier intake exponential increases — a monotonic curve, not a J [95].
• The same pattern appears for cancer in Chinese cohorts using ALDH2/ADH1B instruments [100], and for dementia in the 559,559-person study described earlier [67].

And the two enormous population analyses that used conventional methods but with careful design landed in a consistent place: lowest all-cause mortality risk at about 100 g/week [96], and the Global Burden of Disease analysis of 195 countries concluding, in its own words, that “the level of alcohol consumption that minimised health loss is zero” [97].

Two footnotes for completeness. Resveratrol’s doses in the cell studies that made it famous are orders of magnitude beyond what a bottle of wine delivers; you would need to drink hundreds of glasses. And the polyphenols that actually are beneficial are available in grapes, berries, olive oil, and tea — without the ethanol.

The 2025 Dispute: Why Two Federal Reviews Disagreed

Here I have to be careful and fair, because if I only told you the evidence above, I would be leaving out the strongest counterargument that exists — and you deserve to see it.

In 2025 two congressionally mandated reviews of alcohol and health were completed, and they produced opposite headlines.

Both reviews were competent. Why the divergence?

The fault line is methodological, and it is exactly the one described in the previous section. NASEM’s mortality conclusion rests on pooled observational cohort studies comparing moderate drinkers with never-drinkers. That is precisely the comparison that Stockwell’s bias-correction work [92,93] and the Mendelian randomization literature [94,95,67] identify as systematically biased toward showing benefit. NASEM graded its own confidence as “moderate” — not high — and it explicitly declined to make dietary recommendations. It reported associations. It did not claim causation.

Meanwhile the final 2025–2030 Dietary Guidelines for Americans, released in January 2026, resolved the dispute by declining to engage with it. Its alcohol section now says only “consume less alcohol for better overall health,” and it contains no numeric drink limits at all — the previous “up to 1 drink/day for women, 2 for men” framework was removed — and no mention of cancer risk [99].

So Where Does That Leave “Moderate”?

I want to propose a reframe, because I think the entire public conversation asks the wrong question.

“Is moderate drinking safe?” is a population question. It averages a healthy 35-year-old with no exposures, good sleep, and intact detox genetics together with a 58-year-old with insulin resistance, a mold exposure, small-fiber neuropathy, and a marginal methylation cycle. The average of those two people describes neither of them.

The right question is not “is this dose safe?” but “what is this dose costing this system, given what it is already carrying?”

Alcohol’s cost is not fixed. It scales with the state of the body receiving it:

You may tolerate alcohol socially and still not tolerate it metabolically. Those are two different questions, and only one of them can be answered by how you feel at the table.

Our position, stated plainly

Notice what happens when you read down that table. Every row where the body is doing something — healing, detoxifying, rebuilding, adapting, defending — the answer is zero. The rows get shorter and the reasoning gets thinner as the body gets healthier, but the direction never reverses. There is no row where alcohol is building something.

That asymmetry is the whole argument, and it is why I do not think “how much is safe?” is answerable in the way people want it to be. Every other input we discuss in clinic has a dose-response curve with a beneficial region: protein, sleep, training load, sunlight, minerals, even stress. Too little is a problem, too much is a problem, and somewhere in between is where you want to live. Alcohol has no such region. There is no deficiency state. Nobody has ever presented with symptoms of insufficient ethanol. The curve starts at zero and only goes one direction, and the honest question is not where it becomes harmful but how much harm you are willing to accept in exchange for the taste and the ritual.

So here is the recommendation this clinic makes, without hedging:

• If you are on a healing journey — recovering from mold, Lyme, chronic fatigue, neuropathy, autoimmunity, gut dysfunction, or anything we are actively treating — the target is zero. Not less. Not weekends only. Zero, for the duration of the work. You are asking your body to do the hardest repair of its life; do not hand it a solvent every evening while it does.
• If you are healthy and pursuing optimal wellness, performance, or longevity, we still discourage it. The mechanisms do not switch off because your labs are normal. You simply have more buffer to absorb them — and what you are spending that buffer on is the recovery capacity you were trying to build.
• If you drink anyway, you have not failed at anything. You have made a trade, and this article exists so that you can make it with your eyes open rather than on the strength of a slogan about red wine.

The single most common thing I hear from patients three months after they stop is not “I miss it.” It is: “I did not realize how much better I could feel.” You cannot know what you were paying until you stop paying it.

What Happens When You Stop

This is the encouraging part, and it is where I have the most direct evidence to offer, because someone ran exactly the study I would have wanted.

Look at that insulin resistance number again. A 26% improvement in HOMA in thirty days, from removing one variable, in people who were not heavy drinkers. I have watched patients spend a year of disciplined effort for less. And VEGF and EGF are growth factors implicated in tumor angiogenesis and proliferation — the fact that they fell by 42% and 74% respectively suggests the cancer-risk pathway is not merely cumulative and irreversible, but at least partly dynamic.

The behavioral data is encouraging in a different way. In the research on time-boxed abstinence months, participants showed lower drinking and lower AUDIT scores at six-month follow-up — and notably, that held even for people who did not complete the full month [102]. The framing of a bounded experiment, rather than a permanent vow, appears to be part of why it works.

An evidence-anchored timeline

“Won’t my anxiety get worse?”

This is the most common fear and it is usually backwards. In a non-dependent drinker, a large portion of baseline anxiety is the accumulated residue of nightly mini-withdrawals — the GABA/glutamate rebound described earlier [68,69]. Remove the cycle and anxiety typically improves over the following weeks, sometimes markedly.

Supporting the Liver: An Honest Supplement Review

Two ground rules before the list. First: no supplement makes drinking safe. Anything sold as “liver protection so you can enjoy your wine” is selling permission, not physiology. Second: I am going to grade these honestly, including the ones I would rather were stronger.

Notice the shape of that table. The two strongest interventions are free: stop drinking and move your body. Everything else is adjunctive.

What I actually use, in sequence, for someone coming off alcohol while in an active detox or recovery protocol:

1. Restore the glutathione substrates first. NAC and Glycine Powder (1–3 scoops at bedtime — glycine is also calming and supports sleep, which is the other thing we are trying to rebuild). Liposomal or S-acetyl glutathione where indicated.
2. Repair the B vitamins and minerals that alcohol depletes — thiamine (benfotiamine if there is neuropathy), B6, B12, folate in an active form, magnesium and zinc — guided by labs rather than guesswork.
3. Support bile and elimination, since Phase II conjugates leave through bile. Ox Bile 500 mg or BiLEMIN, plus adequate fiber and genuinely daily bowel movements.
4. Rebuild the gut barrier that alcohol has been loosening — see our leaky gut protocol; MegaMucosa and Tributyrin-X are frequent components.
5. Then resume the detox work — binders, drainage, and antimicrobial or antifungal strategy — on a system that now has the capacity to do it. See our detox protocols and binders articles.
6. Protect sleep aggressively in the first two weeks, since the sleep architecture is what recovers first and drives everything else. Glycine, magnesium, and where appropriate L-Tryptophan at night.

Anything we don’t stock in the shop can be ordered through our practitioner dispensary at Fullscript.

Socializing Without Alcohol

The objection I hear most is not about health. It is: “what do I do at dinner, at the wedding, at the work thing?”

Ten years ago that was a fair concern. It is not anymore. The non-alcoholic beverage category has gone from a punchline to a genuine industry. Athletic Brewing, founded in 2017 in Connecticut, went from roughly $2.5 million in revenue in 2019 to nearly 400,000 barrels produced in 2024, became the number one non-alcoholic beer brand by US grocery sales, rose to the eighth-largest craft brewer in the country by volume, and its Free Wave IPA became the first non-alcoholic beer to win Supreme Champion at the International Beer Challenge. There are now credible alcohol-free options in essentially every category.

Practical notes from actually doing this:

• Hold a glass. Most of the social friction people fear comes from having empty hands and looking like you are abstaining. With a beer-shaped bottle or a wine glass in hand, in my experience, nobody asks and nobody notices.
• Alcohol-free beer is generally better than alcohol-free wine. Beer can be brewed with arrested fermentation or dealcoholized while retaining body; wine develops most of its structure and aroma during full fermentation, and removing the alcohol — which carries aromatic compounds — tends to flatten it. Expect to try several before finding one you like.
• Read the sugar label. Many alcohol-free wines and most mocktails replace the body of alcohol with juice concentrate or added sugar. If you are here for metabolic reasons, a 30-gram-sugar mocktail is not a win. Sparkling water with bitters, citrus, and herbs is usually the better move.
• Watch kombucha. It is a live ferment; commercial products must stay under 0.5% ABV, but home-brewed and long-aged kombucha can exceed that.

What About CBD and THC Drinks?

This question now comes up in nearly every conversation about cutting out alcohol, and it deserves a real answer rather than a reflex. Hemp-derived THC seltzers and CBD beverages have gone from novelty to genuine category: 1.6 million case-equivalents sold in 2025, up 133% in a single year [110]. Meanwhile alcohol is in retreat — Gallup’s August 2026 survey found only 54% of US adults drink at all, tying the lowest reading since the question was first asked in 1939 [111]. Beer and spirits companies are entering this space precisely because they can read those numbers.

So: is a THC seltzer a better choice than a glass of wine? The honest answer is more complicated than either the industry or the alarmists would like.

What these products actually are

The 2018 Farm Bill defined hemp as cannabis containing no more than 0.3% delta-9 THC by dry weight. That is a concentration limit measured against dissolved solids, not a dose limit measured against the drink. Because a sweetened 12-ounce beverage contains tens of grams of sugar and flavoring, a can can legally carry 5, 10, or more milligrams of genuinely psychoactive THC — the water, which is most of the can, barely counts in the math. That is the loophole the category was built on.

Under that umbrella sit very different things: CBD-only beverages (non-intoxicating, typically 5–25 mg — the category we would point you toward); hemp-derived delta-9 THC seltzers (intoxicating, typically 2.5–10 mg); delta-8 and similar isomers (usually chemically synthesized rather than extracted, with the worst contamination record); and dispensary cannabis beverages (a separate legal channel, higher-dose but actually tested).

Does swapping actually reduce drinking?

The literature is genuinely mixed — which is not what you will hear from anyone selling either product.

A widely reported 2026 survey found cannabis-beverage users reporting a drop from about 7 drinks a week to about 3.4 [115] — but it was a retrospective self-report survey of 438 people recruited partly through dispensaries, in a sample where roughly half did not know what was in what they were drinking. Hypothesis-generating, not proof.

Pointing the other way: the NIAAA’s own review concludes there is “compelling evidence for both substitution and complementary effects” — in many populations cannabis use accompanies more drinking, not less [116]. Research on simultaneous use is more consistent still: people who use both on the same occasion reliably experience more negative consequences than people who use either alone [117]. And state alcohol sales after legalization went down in Colorado and Oregon but up for spirits in Washington, leading the authors to conclude the two are “not clearly substitutes nor complements” [118].

The harm-reduction case rests entirely on replacing. It collapses the moment you are doing both.

Dose: why 10 mg is the number that matters

Ten milligrams has become the effective standard dose in commercial hemp delta-9 beverages, and the research explains why it is also the threshold worth knowing about.

That delay is where people get into trouble. Beverages are marketed as fast-onset, so a first-timer who feels nothing after twenty minutes opens a second can — and then both doses arrive together. In a review of nearly 10,000 cannabis-related emergency visits in Colorado, edibles accounted for about 10.7% of cannabis-attributable ED visits while representing only 0.32% of cannabis sold by THC weight — roughly thirty times their market share — skewing toward acute psychiatric and cardiovascular presentations [120].

What you buy may not be what the label says

The foundational study of CBD products sold online found only 30% were accurately labeled, with some marketed as THC-free containing detectable THC [121]. More recent independent testing of 21 hemp delta-9 beverages reported only 4 of 21 accurately dosed, 57% containing under 80% of the labeled THC, and one 5 mg product containing 0.19 mg — though that testing is not peer-reviewed, so weigh it accordingly [122]. Regulated dispensary products test far more accurately [123]. You cannot give careful dosing advice about a product whose real dose is unknowable at purchase.

Does THC do to sleep what alcohol does?

The costs that are real

• Dependence is not a small risk. The CDC estimates roughly 3 in 10 cannabis users develop cannabis use disorder, and a meta-analysis found a withdrawal syndrome in 47% of people with regular or dependent use [127,128]. A nightly beverage is exactly the pattern that drifts toward daily use.
• Anxiety often worsens — in precisely the people most likely to try this. THC-dominant cannabis significantly increased state anxiety; CBD buffered that effect only in people with low baseline trait anxiety, not in those with high trait anxiety [129]. Anxious, autonomically fragile patients are overrepresented in mold, Lyme, and MCAS populations.
• Cardiovascular signal. A 2025 meta-analysis found elevated cardiovascular mortality (RR 2.10), acute coronary syndrome (RR 1.29) and stroke (RR 1.20) [130]. The evidence is observational, but the acute tachycardia is not in doubt — which matters with POTS or any dysautonomia.
• Psychosis risk scales with daily, high-potency use — odds ratio 3.2 for daily use and 4.8 for daily high-potency use in a large case-control study [131].
• Impairment outlasts the feeling. Driving and cognitive impairment persisted up to 4.5 hours, and people had poor insight into their own impairment [132].

What about CBD on its own?

CBD is meaningfully safer than THC on essentially every endpoint above: no intoxication, no psychosis signal, no meaningful dependence risk, no acute cardiac stress. It also has a legitimate evidence base — an FDA-approved anticonvulsant, with controlled trials showing reduced anxiety. Two honest caveats, and the first may be the most useful thing in this section.

Where that leaves us

1. The best choice remains a genuinely non-intoxicating drink — alcohol-free beer, sparkling water with bitters and citrus, a mocktail without the sugar load. Nothing to metabolize, monitor, or titrate.
2. If you want something with an effect, CBD is the lower-risk choice, and we would rather you have a CBD beer or gummy than a glass of wine. Just know a beverage-level dose is likely below the threshold that does anything, and that higher-dose CBD is a real enzyme inhibitor worth telling us about.
3. THC is at your own risk, and we mean that literally. If you are THC-naive, a 10 mg beverage will very likely be psychoactive [119]. Start lower if you start at all, wait a full two hours before considering more, and never on a night you might need to drive.
4. Never both on the same occasion. Everything defensible about substitution depends on cannabis replacing alcohol; same-night co-use is consistently associated with more harm [117].
5. Don’t build a nightly habit out of it. Swapping a nightly wine ritual for a nightly THC ritual solves the ethanol problem while leaving the pattern — the reaching for something at 6 p.m. — entirely intact.
6. Tell us what you are taking — especially with medications, an active detox or antimicrobial protocol, a cardiac or dysautonomia diagnosis, anxiety or panic, or any personal or family history of psychosis. This is information we need, not a confession.

The 60-Day Experiment: How to Actually Run It

I am not asking you to make a lifelong vow. I am asking you to run a decent experiment, because that is the only way you will ever know what alcohol is costing you. Sixty days — long enough for the one-month biomarker changes to show up on labs, plus a month to see what body composition and stamina do.

Before you start

1. Get baseline labs. GGT, ALT, AST, platelets (for FIB-4), fasting insulin and glucose (for HOMA-IR), lipids, uric acid, ferritin, hs-CRP, HbA1c. If neuropathy is in the picture, add B12 with methylmalonic acid, B6, folate, RBC magnesium, and a fasting glucose/HbA1c to rule in or out a glycemic contribution.
2. Record baseline blood pressure — morning, seated, three days running.
3. Capture two weeks of baseline wearable data if you have a tracker: deep sleep, REM, overnight resting heart rate, HRV, and recovery score. This is often the most persuasive dataset you will ever see about your own body.
4. Write down your symptoms and rate them 0–10 today: sleep quality, morning energy, afternoon energy, focus, mood, bloating, burning or tingling, joint stiffness. Do not trust your memory in sixty days; you will have adapted and forgotten.
5. Take a photo and a waist measurement.

During

• Zero, not “less.” A tapering experiment produces uninterpretable data. This is a washout, and washouts have to be clean.
• Stock the alternative before day one. The failure mode is standing in a kitchen at 6 p.m. with nothing to reach for.
• Replace the ritual, not just the liquid. The glass at 6 p.m. is usually marking a transition — work ends, evening begins. Keep the transition; change its contents. A walk, sparkling water in the good glass, tea, a shower.
• Expect the first week to be unremarkable or mildly worse. Sleep often takes 3–7 nights to reorganize. Do not judge the experiment from day three.
• Tell two people. Framing it as “I’m running a 60-day experiment with my doctor” deflects almost every social question, because it is a project rather than a confession.

At 30 and 60 days

• Repeat the labs at day 30 (the Mehta study’s endpoint — expect GGT and HOMA-IR to move) and again at 60 if the first round was informative.
• Repeat the blood pressure, the symptom scores, the measurements, and the photo.
• Compare wearable averages: 14 days pre versus days 45–60.
• Then, and only then, decide. You will have something almost nobody has: your own controlled data.

Frequently Asked Questions